Yacimientos Reservoir Description

Determination of hydrocarbon in place, reserves, and production potential requires an accurate physical description of the reservoir. The basic elements of such a description are depicted in Figure 1 .

Figure 1

Areal Extent The area of the reservoir is needed for calculating the hydrocarbon in place, for selecting the proper locations of wells, and as input data for reservoir simulation studies. Physical Properties of the Productive Formation Physical properties include formation thickness, porosity, water saturation, and permeability. These four parameters are needed in practically all aspects of reservoir

engineering calculations. Preparation of contour maps for these properties constitutes the first and most important step in preparing a data base for reservoir engineering calculations. Structural Dip Reservoirs with a high angle of dip are good candidates for gravity drainage production. For secondary recovery projects in such reservoirs, one locates waterinjection wells downdip and gas-injection wells updip for maxi mum recovery. Thus the angle of dip is an important factor in formulating a recovery plan. Continuity of Strata and Stratification Continuity or lack of continuity of the productive zone determines the pattern of depletion. Identification of separate zones or communicating zones, and the degree of communication, is necessary for establishing the optimum number of wells during primary production and EOR operations. Fault Patterns The location of faults and their effects as barriers to flow define the boundaries of the reservoir and help determine the locations of production and injection wells. Fault patterns strongly affect the design of the field development plan. The number and orientation of faults strongly influence the number of wells (and, in the case of offshore, the number of platforms) required for development. Fluid Contacts Determinations of oil-gas, oil-water, or gas-water contacts are needed for a complete description of the reservoir. Without such information, the hydrocarbon in place cannot be determined to a reasonable degree of accuracy and a proper recovery plan cannot be developed. Aquifer Size The size of the aquifer relative to the hydrocarbon reservoir is important in predicting recovery under primary depletion. Furthermore, this measurement has a strong bearing on the planning of a secondary or tertiary operation. Reservoir Models Reservoir engineering calculations require the formation of a mathematical model for the reservoir. This model should be based on the physical model that emerges from data obtained from the geological, geophysical, petrophysical, and log information. It is evident that in the majority of reservoirs the complexity is so great that it is not practical to expect a faithful mathematical description. Furthermore, it is impossible to obtain a physical description of the reservoir that is 100 percent accurate. One knows the physical properties of the reservoir to a high degree of accuracy only at well locations. In between the wells, or in the part of the reservoir for which no subsurface data are available, the physical description can only be deduced. The more drilling, the better the definition of the reservoir.

However, 3-D seismics and cross-well seismic tomography can provide information about the portions of the reservoir that lie between wells. 3-D seismics employs large amounts of closely spaced data and improved migration techniques to provide volumetric reservoir interpretatios, while cross-well tomography applys highfrequency seismic waves, in which both source and recever are located in existing wellbores. These tools give the geophysicist an active role to play in modeling the reservoir. The mathematical representation of the reservoir can range from a very simple model, the tank-type (or zero-dimensional) model, to a highly complex set of equations that require numerical techniques and computers for their solution (the reservoir simulation approach). In the tank-type approach, the engineer assumes that the reservoir can be described with average values for properties such as thickness, porosity, and fluid saturations. While this approach may be satisfactory for simple problems, it may not be sufficient for other purposes. For instance, the tanktype model or a variation of it is normally used in volumetric estimation of the initial oil or gas in place. In some reservoirs it may also be satisfactory for material balance calculations. However, in other reservoirs such a model might be totally unsatisfactory and the engineer would have to resort to reservoir simulation. Generally speaking, as the heterogeneity of the reservoir in creases, so too does the required complexity of the mathematical representation. Reservoir Simulation As the complexity of a reservoir increases, the need for a more complex mathematical representation arises. The engineer must use a reservoir simulator to predict the performance of the reservoir under various development schemes. Modern reservoir simulation is based on the tank type model, which forms the basis of reservoir engineering. However, rather than considering the reservoir as one tank unit, the simulation divides the reservoir into many tank units that interact with each other. The number of tank units, or cells, depends on many factors, including the heterogeneity of the reservoir, the number of wells, and the field development scheme. Heterogeneous reservoirs require a larger number of cells. The basic reservoir engineering equations that have been used to describe the reservoir when represented by one tank unit are used in reservoir simulation. In the single-cell representation, no oil or gas crosses the boundary of the tank (i.e., reservoir). However, in a simulation with many cells, each cell interacts with its neighbors. Fluids may enter a cell from adjacent cells or may leave a cell and go to the cells neighbors. This fluid movement is governed by a well established flow equation, known as Darcys law. Keeping an inventory of the fluids in each cell is a rigorous bookkeeping operation, well suited to computers. The advent of the modern computer has increased the reservoir engineers simulation capabilities. The rock and fluid data required for reservoir studies using the one-tank model representation are required for each unit cell in a simulation study. The effort required to prepare such data and input it to the simulator is a significant part of the cost, which can range from tens to hundreds of thousands of dollars, depending on the size, complexity, and purpose of the model. Reservoir Boundaries and Heterogeneities

Boundaries A reservoir may have closed or open boundaries, or both. If the reservoir is completely bounded by sealing faults or pinchouts, it is closed. Some reservoirs are completely surrounded by an aquifer, thus their boundaries are open to water movement into the hydrocarbon zone. Still other reservoirs may be bounded by faults or pinchouts along part of their boundary and by an aquifer along the remaining part. Most reservoir engineering calculations require an accurate knowledge of the boundary conditions of the reservoir. This knowledge may establish the possible existence and extent of an aquifer activity in the reservoir. Heterogeneities All reservoirs are heterogeneous, varying only in their degree of heterogeneity. This means that the physical properties of the rock change with a change in location. One of the very important heterogeneities that needs to be considered in reservoir engineering calculations is stratification. Many reservoirs contain layers (strata) of productive rock that can be communicating or non communicating. These layers can vary considerably in permeability and in thickness. A good description of the layers and their respective properties is critical in planning many EOR operations. Fault System Another common heterogeneity in reservoirs is the fault system. Faults can be completely or partially sealing. Well locations for both production and injection are affected by the fault pattern and its effect on fluid communication. Faults are normally defined from geological, geophysical, and production data. Permeability Permeability is another directional property. When permeability measurements vary depending on the direction in which theyre measured, we say that the reservoir is anisotropic with respect to permeability. Permeability anisotropy is important in determining well spacing and configuration, as well as in considering the option of horizontal wells.

Reservoir Pressure
Reservoir pressure is one of the most important parameters of reservoir engineering calculations. Whether the calculations involve the tank type model or a more sophisticated reservoir simulator, accurate pressure values are required. However, there is an important difference between the requirements of the two models. The unit tank model relies on material balance equation calculations, and requires the average pressure for the whole reservoir as a function of time or production. In reservoir simulation studies, however, it is strongly desirable to have available buildup pressure values for individual wells as a function of time. These values represent the average pressure for the drainage volumes of the wells, and are needed for the history-matching phase of the simulation study, which is performed to validate the accuracy of the model built to represent the reservoir (Matthews et al. 1954). History matching is an essential step in "tuning" a reservoir model before conducting a predictive study. Reservoir engineering calculations require a value for the pressure in the reservoir, away from the wellbore. To obtain this value, the well must be shut in and the pressure increase with shut-in time must be recorded. We refer to this as a pressure buildup test (Matthews and Russell 1967). From these data the average pressure value is calculated. Another way of obtaining average values is to record the pressure in a well in which Production has been suspended. if such a well exists, and it is not very close to a

producer or an injector, a pressure-measuring device can be used to continuously record the pressure, without interrupting production or injection operations. For the single-tank model, an average value for the whole reservoir is required. This is normally obtained by a volumetric averaging of the pressure values from different wells. The equation for this purpose is

(3) where: = average pressure for reservoir Pi = average pressure for Well i Vi = the drainage volume of Well i Thus, if there are three wells with pressures p1, p2, and p3, and drainage volumes V1, V2, and V3, then Equation 3 becomes:

Matthews et al. (1954) and Matthews and Russell (1967) have shown that the well-drainage volume Vi is proportional to its flow rate, qi Substituting qi for Vi in Equation 3 gives (4) Equation 4 is the more Practical equation because the flow rate is usually available, while it may be more difficult to estimate the drainage volume. A very useful plot is that of the average pressure values obtained on several wells versus the total oil production of an oil reservoir, or total gas production of a gas reservoir. The pressures are plotted on the Y-axis. If there is continuity in the reservoir the Pressures from the various wells should plot close to each other. If the pressures for a well plot are consistently higher or lower than the other values, it may indicate that the well is not in good communication with the reservoir or that it is in a separate reservoir. This may point out the need for more wells to effectively drain the isolated portion of the reservoir. Furthermore, the data from the isolated well should not be lumped in with the data from other wells in material balance engineering calculations. Before comparing the pressure values measured in wells at various depths in a reservoir (very thick and/or steeply dipping reservoirs), they should be referred to a datum depth ( Figure 1 ).

Figure 1

Normally the depth of the volumetric midpoint of the reservoir is taken as the datum depth. This is determined by constructing a plot of depth versus cumulative pore volume ( Figure 2 ).

Figure 2

The depth corresponding to 50% pore volume is the volumetric midpoint depth. If a particular pressure value is obtained at a different depth than the datum, it is adjusted to the datum by Padj = p + 0.433 Padj = p - 0.433 H where: p = the pressure at any elevation, psi = specific gravity of fluid H = the vertical distance between the point at which the pressure was measured and the datum depth, ft Equations 5 and 6 apply when the point at which the Pressure was determined is, respectively, above and below the datum depth. H (5)

(6)

When an aquifer is associated with the reservoir, the Pressure behavior as a function of time at the hydrocarbon-water contact (or as close as possible to it) is needed for water influx calculations. If this is not available, one usually uses the average reservoir Pressure and adjusts it to the hydrocarbon-water contact depth. The average reservoir pressure is needed in many reservoir engineering calculations. In the case of miscible EOR techniques, for example, the average reservoir pressure determines whether miscibility will occur when CO2 or other gases are injected. This in turn affects overall recovery and the economic feasibility of the project. Reservoir pressure is a topic of significance in reservoir engineering because it is one of the critical pieces of data required by the reservoir engineer for an effective analysis of a reservoir. obtaining reliable pressure data should be a primary goal of any reservoir management program. Reservoir Temperature The calculation of primary recovery relies on the reasonable assumption that the reservoir temperature stays constant. Thus, hydrocarbon recovery during this phase is considered to be an isothermal process. This is so because as fluids are Produced any change in temperature due to Production is compensated for by heat from the cap or base rocks, which are considered to be heat sources of infinite capacity. The average reservoir temperature is needed for laboratory analyses that are made at reservoir conditions. Determining fluid properties, such as viscosity, density, formation volume factor, and gas in solution, requires a value for reservoir temperature. Reservoir temperature is usually measured at the bottom of the well or wells in a reservoir using a wireline temperature gauge. If a variation in temperature is detected across a reservoir after correcting for depth, an average value can be used for the constant reservoir temperature. For EOR techniques such as chemical and miscible processes, temperature affects the phase behavior of injected and produced fluids, and thus the recovery. The feasibility of these processes must be determined by laboratory tests carried out at reservoir temperature. In EOR processes that employ heat injection, such as steam or in-situ combustion, the reservoir temperature is not constant and hydrocarbon recovery is not an isothermal process. Therefore, in mathematical formulations of such processes, it is necessary to write an energy balance over the entire reservoir. From an operations standpoint, reservoir temperatures need to be measured continuously at monitoring wells. These measurements indicate the heat fronts pattern of movement. Normally, a uniform movement is desired, but the heat-front pattern can be altered by changes in injection and/or production schedules.

COMPLETION SELECTION AND DESIGN CRITERIA

Well completion designs will vary significantly with: gross production rate; well pressure and depth; rock properties; fluid properties; well location. Typical ranges for various classes of completions and the design implications are presented in Table 1. This table, of course, represents a partial list of well parameters; there are many other variables that figure into a given completion design. Given the variety of production conditions around the world, definition of the thresholds is naturally somewhat nebulous (a low production rate in a Middle Eastern well would be considered a very respectable rate in many North American fields). However, this table gives a general idea of the range of design considerations.

Table 1: Completion Design Considerations

Well Parameters
High Production Rate: (1500-10,000 B/D liquid [16016,000 m3/d]; 35-140 MMSCF/d gas [1 - 4 106 m3/d]). Low Production Rate: (<30 B/D liquid [5 m3/d]; < 1 MMSCF/d gas [30 103 m3/d]).

Design Implications
Significant frictional pressure losses; Large diameter tubing (>2 7/8 in. or 73 mm); Large diameter casing (>5 1/2 in. or 140 mm); Special artificial lift equipment; Thermal contraction/expansion equipment; Erosion control equipment Artificial lift required; Paraffin buildup problems; Special attention to operating costs required.

Very High Pressure: (10,000-25,000 psi [70-175 MPa]

Special stress checks required during completion; Highstrength tubulars required; Special high-performance packers/accessories required; Problems with H2S aggravated by high pressure requiring special tubular steel Flanged, rather than threaded, wellheads required; Wellkilling capabilities required

High Pressure: (3000-10,000 psi [20-70 MPa]

Low Pressure : (< 1000 psi [< 7 MPa]

Threaded wellheads may be used; Artificial lift required; Greater risk of damage/fracturing during completion process

Deep Wells: (> 10000 ft [ >3000 m]

Problems associated with high pressures; Tubular weight/tension must be considered; Casing size/liner usage must be considered; Hydraulic piston pumps or gas lift more likely to be used as artificial lift; External corrosion of tubulars may be a problem due to higher pressure and temperature Acid wash required upon completion; Difficulty identifying water contact--need formation or drillstem tests Fracturing required upon completion May need fracturing upon completion Little benefit from fracturing; Matrix acidizing may be necessary; Moderate pressure drawdown across perforations Lost circulation a problem; Sand strength may not be great enough to support high velocity flow; Easily damaged Sand control (screens or gravel pack) probably required Sand control possibly required; Minimize drawdown to prevent sand production; Maximize sand exposed to flow; Selective perforation required; Difficult to fracture successfully

Carbonate Reservoirs:

Very Low Permeability (<1 md): Low Permeability (1-50 md): Moderate Permeability ( >50 md):

High Permeability ( >1000 md ):

Unconsolidated sandstone: Partially consolidated and friable sandstone: (acoustic log reads >100s/ft [328s/m]; compressive strength <1000 psi [<7 MPa]; (poor sidewall core recovery Hydrogen Sulfide (H2S) present:

Special HSE regulations/procedures; Corrosion inhibitors may be required; Gas usually considered sour if H2S partial pressure is 70.05 psia (0.3 kPa) Consider inhibitor or special steel if CO2 partial pressure is >10 psi (70 kPa) Scaling and/or corrosion may be a problem; Special artificial lift equipment may be required Consider oxygen corrosion prevention requirements; Consider backflush requirements Offshore--Special HSE regulations; Subsurface safety valve requirments; Well servicing and access constraints Urban/populated areas--Special HSE regulations; Noise and height limits Mountainous areas--Potential for wellhead damage due to landslides

Carbon Dioxide (CO2) present: Water production : Water injection : Well location :

Functional and Well Service Requirements

Definition of the functional and well servicing requirements at the outset can considerably simplify selection of preliminary completion concepts and will highlight the key trade-offs needing further evaluation. Table 1 is a checklist for identifying the critical concerns for a completion design; it illustrates the use of such a checklist in designing a specific subsea oilwell. The completions engineer relies on experience and judgment to prepare the initial input at the concept stage. However, as development plans become more clearly defined, it is often possible to quantify the requirements, based on the results of the initial wells or of detailed design or field studies.

Completion Considerations Rates High Moderate w/chokes Low Variable Pressures

Importance or Need

Completion Design Implications

None High Possible Critical favors two small tubing strings

High Low

None Probable artificial lift required

Producing Characteristics

Multiple zones Minimize costs Access difficulty Uptime

Possible Moderate High High

stack completions review costs TFL/new technology minimize difficulty of future workovers

Rate control Rate stability Long life Density of kill fluid Safety during vessel reentry Wellhead damage

Critical Critical Unlikely Moderate Critical Possible

chokes needed wellhead chokes needed carbon steel sufficient kickoff w/gas lift 2 SSSVs and kill system annular SSSV

Monitoring

Test frequency

High

critical choke bean or dedicated flow line

Pressure measurement Special BHP surveys Log contacts Production logs Tubing investigation Artificial Lift

Moderate Some needed Critical Some needed High

TFL access for downhole tools TFL access for downhole tools vertical access required vertical access required TFL access &/or vertical access

Intermittent w/maintenance

High

gas lift is optimal method via TFL and vertical access

Continuous Increasing gross rate Pressure depletion Kick-off

Possible High Possible

Initial completion Routine operations Depleted conditions High water cut Critical rate Frequency Gas supply volume Gas supply pressure Repairs

Moderate High Possible High High High Moderate Design variable

use gas lift system gas compressor supply required

GLV maintenance system

gas compressor special requirements

Cement

High

future concurrent production and workover operations; easy access; robust tubing joints

Gravel pack SSSV Tubulars New interval Recompletions

Critical Probable Low Possible multizone completion design

Uphole Deepen Sidetrack Function change

Moderate None Possible Moderate

large casing preferable limit depth of rathole maximize casing size large CSG preferable

Well Kill

Frequency Difficulty

High or low Mod.high

operations procedure alternate methods

Production Problems

Sand control Paraffin Emulsions Water cut Scale Corrosion

Critical Possible Possible High Possible Moderate

gravel pack required TFL access for scraping chemical injection capability artificial lift required TFL access carbon steel & downhole chemical inhibitor injection

Erosion Fines GP failure

Low Probable Moderate frequent acid jobs required TFL w/annular kill valve

Table 1: Subsea oilwell functional requirements. It is important for the completion design engineer to have some appreciation for the relative impact of production revenue, capital costs, and operating costs on project economics. In a high tax environment they are usually in the order of importance listed above, with the revenue stream being the most critical. Installation costs are only significant to the extent that special completion requirements have a significant impact on the overall drilling and completion time. The actual cost of the completion equipment is often relatively insignificant compared to the value of incremental production from improved potential or increased uptime. However, production engineers must not take this argument too far. It is important to remember that, in most cases, downtime only results in deferred production. (An exception is the case of competitive production along lease lines.) Nevertheless, for subsea developments in hostile environments, it is reasonable to assume that a premium can be paid for minimizing the frequency of reentry and for equipment reliability and durability.

To a large extent, reservoir, geological, and economic considerations will dictate the functional requirements of a completion and the relative significance of major and minor workovers. These requirements have to be anticipated at an early stage since the techniques to be employed (wireline, service rig reentry, TFL, coiled tubing, etc.) are limited by the tubing design and packer/tubing configurations of the completion. The completion design of a well is also influenced by the well service requirements.The general term "well servicing" covers a broad range of activities, which can be broken down into five major functions: 1. routine monitoring (e.g., being able to run production logs, shoot fluid levels, etc.) 2. wellhead and flow line servicing (e.g., designing components for easy isolation) 3. minor workovers (e.g., through-tubing operations, wireline work, TFL) 4. major workovers (e.g., tubing-pulling operations) 5. emergency situations (e.g., well-killing operations) While to some extent these apply to all oil and gas developments, their relative importance, frequency, complexity, and cost are functions of reservoir conditions, governmental regulations, operating philosophy, and geographic and environmental considerations. For example, it should be self-evident that the options for reentry of subsea wells in deep water are limited and are going to be expensive. This is true to a certain extent for any offshore well. The designer must therefore look carefully at the functions that can be built into the completion and wellhead to minimize well service requirements. It is probable that at least three different generic types of systems will be involved in well servicing: those with functions built into the producing facilities; service units; and workover rigs. From a completion design viewpoint, it is also important to appreciate what capabilities are already inherently available. For example, all wells have the potential for "bull-heading" kill or treatment fluids through the tubing, although it becomes more difficult to control the operation and ensure an efficient displacement as the tubing size and deviation increases. Similarly, with relatively shallow dry gas wells, it should be possible to estimate the bottomhole pressure fairly accurately from tubing head pressure measurements, avoiding the need to run bottomhole surveys. Another built-in function in all offshore wells is the ability to achieve a subsurface shut-off using the government-regulation-required subsurface safety valve. As completion designs become more sophisticated, they can provide an increased number of integrated service functions, up to the ultimate multizone, full TFL completion with downhole pressure monitoring capability. The economic and technical justification for this type of completion must be based on a detailed functional analysis of the reservoir, completion lifetime, and well service economics. Moreover, increased sophistication also introduces higher risks of completion problems or subsequent failures, requiring improved quality control and materials selection.

What Goes Into Reservoir Simulation?

The basic tool for conducting a reservoir simulation study is a simulator. The development of this tool requires a good understanding of the physical processes occurring in reservoirs and a high level of sophistication and maturity in advanced mathematics and computer programming. Although simulation engineers generally do not get involved in actual software development, the effective use of reservoir simulators requires that they at least appreciate what goes into this development. Like any tool, a reservoir simulator has its strength and limitations, and it is well to keep in mind the various assumptions that factor into its development. This is not to suggest that all simulation engineers must be expert programmers; rather, they must be intelligent users. Therefore, knowledge and understanding of the simulation process are necessary precursors to a reservoir simulation study. At first, a simulation study might look like a once-and-for-all exercise. In truth, however, it is an evolutionary process, throughout which we continually refine our conceptual understanding of the system. While we cannot overemphasize the importance of accurate reservoir description in a good reservoir simulation study, we must at the same time acknowledge that the data needed for an accurate description is seldom available; invariably, studies start out with less than complete data. However, by carefully analyzing and interpreting the voluminous information generated during the study, we should be able to refine and extend the input data base. Such refinement should lead to a better understanding of the system and, ultimately, to a better reservoir description. Of course, this requires some agility and creativity; there is no such thing as a casual simulation engineer. It is, therefore, apparent that there are three basic interwoven components that go into a simulation study. These are: The tool: reservoir simulator The intelligent user: simulation engineer The pertinent information: reservoir description Figure 1 depicts the necessary interactions among the simulation engineer, the simulator and the reservoir description.

Figure 1

The engineer is clearly the prime mover in conducting the simulation study, and must be in control of other study components. This control involves:

being cognizant of the simulators limitations and the assumptions that go into its development being able to adequately describe the reservoir being fully conversant with the analytical techniques involved in interpreting the results.
Based on the initial results, it is not uncommon for the simulation engineer to revisit the appropriateness of the reservoir description through concept refinement.

Why Do We Need Reservoir Simulation?

The information we obtain from a newly discovered field is scanty at best. It is also disjointed to a certain extent, because bits and pieces of information are emanating from different parts of the field. Our first task is to integrate these pieces of information as accurately as possible in order to construct a global picture of the system. A reservoir simulation study is the most effective means of achieving this end. As field development progresses, more information becomes available, enabling us to continually refine the reservoir description. Once we establish a good level of confidence in our reservoir description, we can use the simulator to perform a variety of numerical exercises, with the goal of optimizing

field development and operation strategies. We are often confronted with questions such as

what is the most efficient well spacing? what are the optimum production strategies? where are the external boundaries located? what are the intrinsic reservoir properties? what is the predominant recovery mechanism? when and how should we employ infill drilling? when and which improved recovery technique should we implement?
These are but a few of the critical questions we may need to answer. A reservoir simulation study is the only practical laboratory in which we can design and conduct tests to adequately address these questions. From this perspective, reservoir simulation is a powerful screening tool.

What are the Simulation Approaches?

The complexity of the problem at hand, the amount of data available, and the studys objectives invariably dictate the choice of reservoir simulation approach, granted that we have already taken into account the appropriate computational environment (both in terms of hardware and software). Broadly classified, there are two simulation approaches we can take: analytical and numerical.

The analytical approach, as is the case in classical well test analysis, involves a great deal of assumptionsin essence, it renders an exact solution to an approximate problem. The numerical approach, on the other hand, attempts to solve the
more realistic problem with less stringent assumptionsin other words, it provides an approximate solution to an exact problem. From here on, we use the term simulation rather loosely to refer only to the numerical approach. The domain of interest can form another level of categorization for simulation approach and model selection. For instance, a study may focus on a single well and its interaction with the reservoir within its drainage area (i.e., single-well simulation in radial-cylindrical coordinate system). The other extreme case may be the study of an entire field (field-scale simulation in rectangular coordinate system) in which the overall performance analysis of the field is called for. In between these two extremes comes the case where only a section of the reservoir is targeted (window-study).

Figure 2

Figure 2 schematically represents these three simulation approaches.

What are the Basic Steps of a Simulation Study?

In general, a reservoir simulation study involves five steps:

Setting objectives Selecting the model and approach Gathering, collecting and preparing the input data Planning the computer runs, in terms of history matching and/or
performance prediction

Analyzing, interpreting and reporting the results

A critical step is selecting the model and the approach. We must decide at the outset how many dimensions will be adequate. Such decisions depend on the flow dynamics involved and the amount of detail required. There may be cases when a two-dimensional representation is sufficient to describe a thin reservoir, whereas a three-dimensional model is unavoidable in the case of a thick reservoir. The expected flow structure dictates the choice of the models coordinate system. For example, in most single well studies, we may use radial-cylindrical flow geometry. However, if a well is vertically fractured, we should assign an elliptical-cylindrical flow geometry. We also have to decide how best to represent physical phenomena. For instance, if compositional variation is important, we may have to employ a compositional simulator rather

than the more commonly used "black oil" model. Similarly, if we intend to study a coalbed methane reservoir, we must use a specialized model that accounts for the desorption process. One of the most labor-intensive aspects of reservoir simulation study is data gathering, collection and preparation. Oftentimes, this requires collaboration among technical personnel with varying levels of expertise. For instance, geological and geophysical data are extremely crucial and need to be processed in the form that is useful for reservoir description. Data will be often be sparse or incomplete. In such situations, statistics or other tools can prove quite helpful. Because of the large volume of data being processed at this stage, and the likelihood of internal inconsistencies in the data, the engineer must have strong organizational skills and sound judgment. In simulation studies, time (both the engineers and the computers) is of the essence. A typical simulation study requires a large number of runs, which must be carefully orchestrated to yield the desired information. As we make each run, we must carefully analyze the results, and appropriately label them to avoid confusion and costly duplications. In addition, we must avoid runs that yield no new information. It is pertinent, therefore, to take into account the inferences from the previous runs in planning the next suite of runs. Perhaps the most important step in a simulation study is analyzing and interpreting the results. It is at this stage that our creative and discerning abilities are put to the test. As tempting as it may be to do so, we should not view every number that comes out of the computer as the absolute answer. Instead, we must always be asking questions such as, "what if? what then? why? so what?" In this way, we bring into play our experience, common sense, and perhaps sometimes extraneous knowledge. A simulation studys ultimate objective is to forecast reservoir performance. If we have selected the correct model, adequately prepared our data, conducted the appropriate computer runs, and made good, informed analyses, we should be confident of our ability to predict performance. Any mistakes we make in the previous steps will have a cumulative impact on performance prediction. We must communicate study results in an appropriate manner to other technical personnel and to management. This should be in the form of a comprehensive technical report with sufficient details for others to assess the studys quality.

How are Reservoir Simulators Used?

A reservoir simulator can be an effective tool for screening, analysis and design. The thought process that goes into appropriately using a simulator for these purposes, however, is quite involved. Figure 3 illustrates the interactions inherent in this thought process.

Figure 3

The cornerstones of a reservoir simulator are the mathematical model, laboratory investigation (laboratory data), field observations and the computer code. In using reservoir simulators, these cornerstones generate signals, which propagate and interact with each other such that a continuous feedback takes place for the mutual benefit and enhancement of all the parts. For example, laboratory investigation, field observations and the computer code can highlight the need for improvement in the mathematical formulation. Similarly, a computer code originated from a robust formulation, together with pertinent field observations, may shed light on the validity of the experimental approach taken in the laboratory. This dynamic interaction illustrates the self-enhancing nature of reservoir simulators.

What Does a Simulation Study Require?

A simulation study is a challenging and demanding task, loaded with opportunities to learn more about the reservoir. To reap the full benefits of this powerful tool, it is imperative to recognize the proper roles of the engineer and of the reservoir simulator. In a successful study, neither of these can afford to dominate the other. The engineer should not demand from the simulator what it is not meant to do, but neither should he or she become overly dependent on the simulator. In a nutshell, the success of a simulation study hinges on a combination of a good engineer and the right simulator.

Steps in a Simulation Study

There are five basic steps in conducting a reservoir simulation study:

setting concrete objectives for the study

 selecting the proper simulation approach preparing the input data

planning the computer runs (including the order in which they occur)

analyzing the results

Setting the Objectives

Setting objectives is the most important step in conducting a simulation study. Clearly defined objectives help us obtain the best information at the lowest cost and in the least amount of time. Improperly set objectives can take the study on a long, roundabout journey which leads to nowhere. There are a number of factors that help us define appropriate objectives. The most important of these are data availability, the required level of detail, availability of technical support and available resources. In setting objectives , we use all of these factors to determine how to proceed. For example, it is unrealistic to attempt threedimensional simulation when the available geological data gives no information about the presence and description of the various formation layers present in the reservoir. In the broadest sense, when we consider all these factors, we will arrive at one of two types of objectives. These are sufficiently distinct that they affect the entire planning process of the simulation study. One type of objective is fact-finding, while the other is to establish an optimization strategy.

Fact-finding involves answering questions about a system or process that is

already in place. For example, a simulation study that matches well test data for the purpose of determining the damaged zone around a wellbore is a fact-finding mission.

Optimization involves developing a number of plausible scenarios for

a process (e.g., waterflooding) and studying the system response in an attempt to determine the optimum scenario. In this case,we must design a suite of numerical exercises, being careful to avoid waste on exercises that may not significantly contribute toward the goal.

Choosing the Simulation Approach

In choosing the simulation approach, we need to consider three basic factors:

reservoir complexity fluid type alcance(scope) of the study

While reservoir complexity and the scope of the study determine the simulators dimensions and coordinate geometry, the fluid type (together with the processes involved) dictate whether we should use a black-oil model or a more specialized model. For example, predicting well performance in a gas condensate reservoir will require a compositional rather than a black oil simulator. Furthermore, if the reservoir is thin and unlayered, it will be sufficient to use a onedimensional radial flow geometry. Carrying out such a study with a three-dimensional compositional simulator will require additional computational resources whose added benefit cannot be justified. In any case, we must exercise our judgement and ingenuity in selecting the most appropriate simulation approach.

Preparing the Input Data

Because simulation studies usually require large volumes of information from a wide range of sources, preparing the input data can be a laborious task. However, the time spent in ensuring that data are properly prepared is worthwhile, in that it can prevent a great deal of headaches and waste later on in the study. Often, we discover data input errors only after a problem surfaces during the run, which wastes both time and computing resources. It is our responsibility to ensure internal consistency in the data. Because data come from different sources, internal inconsistencies are not uncommon. We should resolve inconsistencies during the data input preparation. When data inconsistencies are present, they can lead to an ill-posed problem. Even worse, they could go undetected. With an ill-posed problem, we may be able to find the inconsistency by the failure of the simulator to run; but in the case of buried inconsistencies, the simulator may run and yield erroneous solutions. Pre-processing capability, particularly for the commercial codes currently available, can facilitate data preparation. Sometimes these processors have internal checks to flag any detected inconsistencies in the data. While data preparation is the simulation engineers job, input from other supporting personnel is extremely important. If inconsistencies appear in the data, or even if some data appear doubtful, it is imperative to resolve the problem with the help of the geologist, geophysicist and perhaps the production engineer. In summary, there is no overemphasizing the importance of adequate data preparation prior to making a simulation study. The payoff is exceptionally good.

Planning the Computer Runs

Planning computer runs is deceptively simple. To understand the necessity and the complexity of this planning, we only need to imagine a simulation study as a complex road map where the traveler knows the point of origin and the destination (these are clear enough from the objectives of the study). However, just as a traveler requires careful mapping out of the route that will get him or her to the destination in the best time possible, we must carefully map out the type and number of computer runs that will achieve the set objectives at a minimum cost. In so doing, we must account for several factors, which are usually problem dependent. We should consider the number of parameters to be examined, the duration of prediction, and the type of information needed to answer the pertinent questions.

Careful planning of computer runs includes not only determining their order, but also establishing a systematic labeling procedure for them. This is particularly important because of the large number of runs usually required and the voluminous amount of information invariably generated for analysis.

Analyzing the Results

When we have analyzed the results of the simulation study and made pertinent inferences from it, we can evaluate its success. This step caps all the efforts previously discussed. Considering the amount of effort that we expend on the simulation study up to this point, it is tempting to become a biased arbiter of the results. On the contrary, this is the time to ask critical questions and even ponder over the implications of the results. In other words, we must not become easy subscribers to our solutions. The mode of analysis and the presentation of results will depend very largely on the audience for whom they are meant and the post-processing capability available. The graphics capabilities currently available on most computers makes this process easier and even more inviting. It is now not uncommon to display information using three-dimensional graphics. In addition, graphics features, such as image rotation and animation, enhance our interpretation and inferential ability.

Reservoir Rock Properties

The essential rock properties in reservoir simulation are those that govern the rocks storage capacity and spatial distribution; its ability to conduct fluids; and its spatial and directional distributions

Porosity
Porosity is a measure of a rocks storage capacity. In reservoir simulation, we are primarily interested in interconnected pore space. From here on, therefore, we shall understand porosity to mean effective porosity. Effective porosity is a dimensionless quantity, defined as the ratio of interconnected pore volume to the bulk volume. In an idealized arrangement of grains of uniform size the maximum porosity value is 47.64% for cubic packing, and the minimum is 25.96% for rhombohedral packing ( Figure 1 ). However, naturally-occurring reservoirs do not conform to these theoretical limits, due to syn- and post-depositional processes that have taken place (in addition to non-sphericity of the grains). Their porosity can vary widely.

Figure 1

In the flow equations used in reservoir simulation, porosity appears as one of the parameters that scales the volume of fluids present in the reservoir at any time. During production, this volume is depleted, and reservoir pressure drops. The higher the reservoirs porosity, the less this pressure decline will be over time. The special case in which porosity does not appear in the flow equation is the single-phase incompressible flow system. As we will discuss later, in such a flow system, there is neither accumulation nor depletion, and so porosity vanishes. In the other extreme, there are reservoirs in which porosity changes with pressure, and so appears in the equation as a function of pressure rather than as a constant value.

Permeability
Absolute permeability is a measure of a rocks ability to transmit fluid. For a hydrocarbon reservoir to be commerical, it must not only be porous, but also permeable. Permeability is analogous to conductivity in heat flow. Since it is a measure of resistance to flow, a higher permeability reservoir experiences less pressure drop than a corresponding low permeability reservoir. The dimension of permeability is length squared [L2], and its practical field unit is darcy or millidarcy. One darcy is approximately equal to 10-8cm2. Permeability varies widely in naturally occurring reservoirs, from a fraction of a millidarcy to several darcies. Similar to porosity, the permeability of a reservoir could be a function of pressure. Permeability is a key parameter controlling the propagation of transients created by conditions imposed at the well. It does not determine ultimate recovery, but rather the rate of this recovery.

Homogeneous vs. heterogeneous systems

Homogeneous systems feature uniform spatial distribution, while heterogeneous systems exhibit non-uniform distribution. For simplicitys sake, we often assume homogeneity in reservoir calculations, even though many reservoirs are heterogeneous. This is where numerical reservoir simulation becomes a very

powerful tool, because it allows us to incorporate property variation in the system. It is important that when we describe a reservoir as homogeneous, we specify the property of reference (e.g., "this reservoir is homogeneous with respect to porosity but heterogeneous with respect to permeability"). Although reservoir simulation equations can accommodate property variation within the domain of interest in significant detail, such detailed information may be unavailable, or at best, very sketchy. In such cases, we must employ interpolation techniques and history matching exercises.

Isotropic and anisotropic systems

Some parameters used in reservoir simulation exhibit directional dependency. A reservoir exhibits isotropic property distribution if that property has the same value regardless of the direction in which we measure it. On the other hand, if a propertys value does vary with direction, then the reservoir is anisotropic with respect to that property. One should be careful to note that only those properties that are not volume-based can exhibit directional dependency. Porosity, for example, is a volume-based property by definition. It utilizes all three dimensions, and therefore has zero degrees of freedom in terms of directional variation. Permeability, by contrast, has the dimension of area, leaving one direction in which it can vary. Figure 2 shows all possible permutations of isotropic, anisotropic, homogeneous and heterogeneous systems for two-dimensional cases.

Figure 2

to align these axes with the principal flow directions, so that we may eliminate the six off-diagonal elements of the permeability tensor, and be left with three diagonal elements in a three-dimensional system (two for a two-dimensional system). Otherwise, an incorrect representation of the system results, as shown in Figure 3 .

The existing anisotropy determines the orientation of a coordinate systems principal axes. In most applications, reservoir simulators employ orthogonal coordinate systems, where all the axes are mutually perpendicular. It is imperative

Figure 3

Reservoir Fluid Properties

Fluid properties, like rock properties, significantly affect fluid flow dynamics in porous media. Unlike rock properties, however, fluid properties exhibit significant pressure dependency. Therefore, it is often necessary in reservoir simulation to estimate these properties using correlations and/or equations of state.

Gas properties
In calculating gas properties such as density, compressibility and formation volume factor, we often use the real gas law as our basis. For more rigorous calculations, we might use a modern engineering equation of state such as the Peng-Robinson equation of state. Invariably, these calculations express density as a function of pressure and temperature. The properties of interest in the gas flow equation are density, compressibility factor, compressibility, formation volume factor and viscosity. Density appears in the gravity term, and it is often neglected. The compressibility factor introduces an important non-linearity, in that it appears in the formation volume factor. Gas viscosity is also strongly dependent on pressure, and needs to be calculated as pressure varies spatially and temporally. Figure 4 summarizes the equations and correlations necessary for determining gas properties.

Figure 4

Oil properties
Oil properties that appear in the governing flow equations for the oil phase are density, compressibility, formation volume factor, viscosity and solubility of gas in oil. In the absence of gas, these oil properties can be treated as constants, because the compressibility of gas-free oil is very small. However, the presence of dissolved gas in oil necessitates the use of appropriate correlations to determine the variation of these properties with pressure and temperature. A recent review of the available correlations has been provided by McCain (1991). We can also use modern equations of state to calculate these properties.

Theoretically, an infinite amount of gas can dissolve in oil, provided that adequate pressure is available. Accordingly, if pressure is available, it is conceivable that there will be no free gas (undersaturated reservoirs). If pressure is not sufficient some of the gas will exist in the free state (saturated reservoirs). A typical simulation calculation may traverse saturated and undersaturated conditions. Most reservoir simulators implement variable bubble-point algorithms to handle these situations. Figure 5 shows the qualitative variation of several of these properties.

Figure 5

Water properties
McCain (1991) provides correlations for estimating such water properties as density, compressibility, formation volume factor, viscosity and gas solubility. Since gas solubility in water is very small compared to oil, for most practical cases, we assume constant values for these properties that come into play in the water flow equation.

Reservoir Rock/Fluid Interactions

Reservoir fluid flow is governed by complex interactions between the fluids and the reservoir rock. These interactions become more complicated when, as is often the case, two or more fluids are present in the same pore. When a driving force acts on such systems, the fluids compete in motion, because their movement is mutually dependent. To appropriately describe the simultaneous flow of two or more fluids in a porous medium requires a good understanding of both the fluid-fluid and rock-fluid interactions.

Wettability and interfacial tension

The principal fluids in a petroleum reservoir are water, oil and gas. When they exist as free phases, they are generally immiscible (Note: this discussion does not consider emulsions or dissolved gases). When these immiscible fluids co-exist in the reservoir pore space, their interactions with one another and with the containing rock control their spatial distribution and movement. The two principal properties used to quantify these interactions are wettability, which pertains to rock-fluid interactions, and interfacial tension, which relates to fluid-fluid interactions. When two immiscible fluids co-exist in the same pore space, one preferentially adheres to the rock surface. This phenomenon is known as wetting, and the fluid that is preferentially attracted is referred to as having a higher wettability index. The parameter which determines the wettability index is called adhesion tension, and it is directly related to interfacial tension. Interfacial tension is a measure of the surface energy per unit area of the interface between two immiscible fluids. Examples of such interfaces include the junction between water and crude oil and the junction between oil and gas. Figure 6 depicts an oil-water interface.

Figure 6

The study of surface energy phenomena is very important in recovery processes, in that many EOR processes are based on altering the surface energy so as to favor oil recovery. They work on the principle that all interfaces existing under equilibrium conditions have some free energy associated with them. As a two-phase system approaches equilibrium, the interface assumes a configuration that tends to minimize the free surface energy, unless otherwise constrained by external forces. Examples of this behavior abound in nature. A rain drop falling in a vacuum assumes a spherical shape, since this is the geometrical shape that minimizes surface area and,

therefore, surface energy. A rain drop falling through the atmosphere would do the same thing, except that an external drag force constrains it from doing so. For this reason, it has a "tear-drop" shape. Interfacial tension is the surface energy per unit area of a fluid interface. It has units of force per length. For any fluid having contact with a solid surface, the contact between the fluid and the solid has a certain value of interfacial tension associated with it. What differentiates fluids from one another in terms of relative wettability are their values of fluid/solid interfacial tension. The lower the solid-fluid interfacial tension, the lower the surface energy and the higher the tendency for the fluid to wet that surface. For two immiscible co-existing fluids in porous media, the one with the lower interfacial tension is the wetting phase, while the other is the non-wetting phase. There is a definite relationship between the solid/wetting phase interfacial tension, the solid/non-wetting phase interfacial tension, and the wetting/non-wetting phase interfacial tension. The Young-Dupres equation (Equation 2.1) expresses this relationship. SN - SW = WN(cos) (2.1) where

SN = solid/non-wetting phase interfacial tension SW = solid/wetting phase interfacial tension WN = wetting/non-wetting phase interfacial tension
= angle of contact between the fluid and the rock

Relative permeability
When two or more immiscible fluids flow simultaneously through a porous medium, they compete and do not move at equal velocity. This results on the one hand from interactions between the fluids and the rock, and on the other from interactions among the fluids themselves. As previously mentioned, this manifests itself in interfacial tensions. Interfacial tensions are not transport properties, and so we cannot use them directly to qualitatively characterize relative motion. We can, however, observe the relative ease with which each of the two competing fluids go through the porous medium that is, we can measure the relative permeability. Although relative permeability is not a fundamental property of fluid dynamics, it is the accepted quantitative parameter used in reservoir engineering. Relative permeability appears prominently in the flow equations used in reservoir simulation. By definition, relative permeability is the ratio of the effective permeability, when more than one fluid is present, to the absolute permeability. Effective permeability is

the measured permeability of a porous medium to one fluid when another is present. The effective permeability depends on the relative proportion of the two fluids present, or fluid saturation. Therefore, relative permeability is also a function of fluid saturation. Although there are models for predicting relative permeability, they are all empirically formulated from measured data sets. Figure 7 shows typical relative permeability curves for an oil-water system.

Figure 7

Figure 7 depicts relative permeability as a function of water saturation for a twophase system. We therefore refer to it as two-phase relative permeability. If a third phase is present, then each fluid has its own relative permeability, which differs from the corresponding two-phase relative permeability. Because it requires a three-dimensional representation, three-phase relative permeability is often shown on ternary diagrams, with isoperms displayed at various saturation combinations. Leverett and Lewis (1941) were one of the first to use this representation. Figure 8 shows a typical relative permeability curve for a three-phase oil/gas/water system.

Figure 8

Successful simulation of a multiphase system hinges on adequate relative permeability information. Since relative permeability is a function of saturation, which varies over a reservoirs life, the best way to get adequate information is to incorporate relative permeability models into the reservoir simulator. Several models are available (Honapour, et al. 1986), each claiming varying degrees of merit. The simulation engineer must determine which model is appropriate. Table 1 lists some common relative permeability models.

Table 1

Capillary pressure
In everyday experience, water levels in two or more connected containers have the same level if exposed to the same atmosphere. But when it comes to spaces of capillary size (like those we encounter in porous media), we cannot take this rule so literally. To illustrate, consider what happens when a glass tube of capillary size is dipped in a larger container filled with water ( Figure 9 ).

Figure 9

The water in the capillary tube rises above the water level in the container to a height that depends on the capillary size. Although strictly speaking, the water still finds its level, it does so in such a way as to maintain an overall minimum surface energy. In this situation, the adhesion force allows water to rise up in the capillary tube while gravity opposes it. The water rises until there is a balance between these two opposing forces. The differential force between adhesion and gravity is the capillary force. This force per unit area is the capillary pressure. As we might surmise from these observations, there is a relationship between capillary pressure, Pc , and the interfacial tension between the two fluids (in the case of Figure 9 , water and air).

<
where Pc = capillary pressure

(2.2)

wn = wetting/non-wetting phase interfacial tension

R = radius of the tube

 = angle of contact between the solid surface and liquid Note that the exact opposite happens if the fluid is the non-wetting phase with respect to the tube material. The classic example is oil and mercury in a glass capillary tube, where, instead of capillary rise, there occurs capillary fall. Capillary pressure is important in porous media flow description because of the saturation distribution in the capillary-like pore spaces. A plot of capillary pressure versus water saturation has a typical shape, as shown in Figure 10 .

Figure 10

It is interesting to note the hysteresis between the capillary curves for a drainage process (where wetting phase saturation is decreasing) and an imbibition process (where wetting phase saturation is increasing).

Microscopic Properties
Reservoir fluid flow is a fundamentally complex process. Fluid movement depends not only on the fluids themselves, but also on how the fluids interact with the porous medium, which in effect is a huge capillary network. Then, there is the pore structure itself to worry about. On a scientific level, understanding these microscopic properties is a must if we are to capture the essence of the system. However, engineers thrive on approximation, and reservoir simulation engineers are no

exception. Using global properties such as permeability, porosity, relative permeability, and capillary pressure, one attempts to procure the best information possible. As research progresses into reservoir characterization, improvement in the modeling approach is inevitable. Meanwhile, we must learn all we can through approximate models.

Flow Geometries and Dimensions

The same factors that dictate our choice of coordinate systems play a dominant role in our deciding how many dimensions to assign to a problem. Engineering, to a large extent, is a marriage between pure science and practical reality. In other words, selecting a higher number of dimensions to represent a system may be scientifically correct, but we may lack the information or the computational overhead needed to assign this many dimensions. So, we assign fewer dimensions and settle for a lessthan-ideal problem definition. Such compromise may seem drastic. But in reality, for most engineering problems, we can generate an adequate amount of information even within these limitations.

Rectangular flow geometry

Rectangular geometry is the one that is most familiar to us, as engineers, dating back to our high school calculus. In reservoir modeling we often use this familiarity to our advantage, since most field-scale multi-well studies are done in this coordinate system. As Figure 1 illustrates, we can consider the reservoir to be a rectangular box with the fluid particles moving in straight lines, perhaps at different speeds in different directions and locations.

Figure 1

In this case, the streamlines are parallel to the three principal axes (x, y, and z), which are orthogonal. Figure 1 also shows the partition of the box into many smaller boxes, which are rectangular prisms. Each of these rectangular prisms represents a certain portion of the reservoir, about which we can procure information through simulation studies. We use this smaller element of dimensions, (x, y, z) as a control volume to set up and discretize the governing equations. We should emphasize that a fluid particle entering an elemental volume in one direction does not necessarily exit in the same direction; by the same token, the fluid particle leaving the elemental volume in one direction did not necessarily enter it in the same direction. Formation characteristics, (such as heterogeneity, permeability contrasts and the force fields imposed by the conditions at the boundaries) dictate the flow path once the element enters the control volume. This is the essence of flow multi-dimensionality. Figure 2 illustrates the concept of one-dimensional flow along the x-direction. Although it is difficult to find real-world examples of truly one-dimensional flow, many types of analyses and systems do lend themselves to description as one-

dimensional.

Figure 2

The flow structure shown in Figure 2 precludes flow in any other direction, which implies there is no property variation along the y and z directions. Therefore, if we take a section perpendicular to the indicated flow direction, there will not be any property variation across the plane. Similarly, any cross section taken in the x-z or x-y planes (parallel to the streamlines) will reveal the uniformity of the flow structure. More explicitly, the pressure profiles of flow paths will be similar. A long, skinny reservoir that is confined between two closely spaced, parallel faults fits this description. The next level of description used in reservoir simulation is two-dimensional flow. Many reservoir simulation studies employ two-dimensional Cartesian coordinate systems. This makes sense when we consider the large lateral extent of most reservoirs compared with their thicknesses. Figure 3 illustrates a two-dimensional flow structure along the x and y directions.

Figure 3

This precludes flow in the z-direction. Therefore, any slice taken parallel to the x-y plane will not show any variation in terms of property and fluid distribution such as porosity, permeability and saturations. The introduction of the second dimension allows us to describe a wide variety of problems. We can, for instance, account for directional permeability variation and lateral well distributions. Moreover, a twodimensional approach allows us to represent a variety of well completion strategies (e.g., vertical wells, horizontal wells, stimulated wells). Thin, blanket sands that tend to display large areal coverage are ideally suited for description by a two-dimensional model. The best representation of flow is the three-dimensional model, because it allows us to procure the most information about the reservoir. Unfortunately, it also requires the largest amount of input information and a higher level of computational power and overhead. Still, incorporating a third dimension gives us the latitude we need to include all the property variations in all three spatial directions. This means that if we take two parallel slices perpendicular to the third dimension, they will exhibit property and flow differences. Figure 4 illustrates a three-dimensional flow structure.

Figure 4

A three-dimensional representation allows us to accommodate a wide variety of problems of practical interest, such as layered reservoirs (with or without crossflow), partially penetrating wells, multi-layered production schemes, and thick reservoirs where gravitational forces could be significant. A three-dimensional model makes it possible for us to come up with more realistic representations of drive mechanisms (or any combination thereof), such as gas cap expansion, bottom water drive, and so forth. In spite of a three-dimensional models many advantages, it is less often used in practice than we might expect. This is because we have to weigh such factors as cost, data availability and marginal utility. In many cases a three-dimensional model may be a luxury that we can ill afford. On the other hand, there are certain problems in which it is a technical necessity. Consider, for instance, a thick reservoir with no significant property variation in the vertical direction; while a two-dimensional model would appear adequate based on this description, it may turn out that the gravitational field contribution may be so significant as to require a third dimension.

Radial-cylindrical flow geometry

The radial-cylindrical coordinate system is particularly appealing for describing single-well problems. Figure 5 shows the principal directions of this flow geometry and its elemental volume.

Figure 5

The three principal flow directions are radial (r), vertical (z) and tangential (). To visualize this flow structure, imagine a single well located in the center of a circular reservoir, such that the wellbore and the reservoir boundary are two concentric circles. If we assume a reservoir of uniform thickness, then the system becomes two concentric cylinders of the same height. A particle moving in a three-dimensional radial-cylindrical flow geometry can be illustrated as in Figure 6 .

Figure 6

A typical one-dimensional, radial-cylindrical flow model is the classical representation used in well test analysis. In this case, flow is constrained to the r-direction such that streamlines are rays converging towards the center of the well ( Figure 7 ).

Figure 7

Studying the problem along one trajectory is sufficient because of symmetry. Any particle located on any of the trajectories will experience similar forces. By neglecting the flow in angular (q) and axial (z) directions, we introduce a series of assumptions, such as no permeability gradation along the q-direction and no gravitational effect along the z-direction. As we can imagine from looking at Figure 6 , one-dimensional flow representations in the q- and z-directions have no practical significance in reservoir studies. The two-dimensional (r-z) representation is appealing for single-well problems where gravity and/or layering effects are significant ( Figure 8 ). This r-z plane can be taken at any q location without changing the problem because of its axi-symmetric nature.

Figure 8

The three-dimensional flow structure in radial-cylindrical coordinate system admits property variation in all three directions. Figure 9 shows this system.

Figure 9

Elliptical-cylindrical flow geometry

We sometimes use elliptical-cylindrical flow geometry in single-well studies when a strong permeability contrast exists in two principal directions on the lateral plane. Another common application of this coordinate system is when a vertical well is intercepted by a vertical, high-conductivity fracture (theoretically presumed to be of infinite conductivity). Under these conditions, the normally concentric equipotential contours degenerate into confocal ellipses. Similarly, the streamlines become distorted into confocal hyperbolas. Figure 10 depicts this flow structure.

Figure 10

Spherical flow geometry

Although not commonly used for general simulation, the spherical coordinate system provides a good representation of some specific reservoir engineering problems. Two examples are partial penetration to a thick formation by a production well, and flow around perforations. The principal flow directions in spherical coordinates are radial (r), tangential () and azimuthal (), as shown in Figure 11 .

Figure 11

Curvilinear flow geometry

The most generalized coordinate system is curvilinear. In fact, all of the coordinate systems previously discussed constitute a subset of the curvilinear system. A curvilinear coordinate system allows a better representation of the flow geometry, as well as the boundary geometry where the latter dictates the former. With the flow geometry more accurately represented, the results obtained with a curvilinear coordinate system do not get distorted by grid orientation effects, as often happens with other coordinate systems. Another advantage is that curvilinear systems may help reduce the number of grid blocks needed for the same level of accuracy. Figure 12 shows the areal implementation of curvilinear coordinates to a five-spot injection/production pattern.

Figure 12

Note that the streamlines and equipotential contours define the curvilinear elemental volume. Although curvilinear coordinate systems offer attractive advantages, their use is limited because of the added mathematical and interpretational complexity they introduce. Choosing the appropriate coordinate system and number of dimensions is not only paramount to a simulation studys success, but also to its relative simplicity. It is thus essential that we use sound engineering judgment and perform thorough analyses throughout this process. We must answer questions pertaining to the reservoirs approximate geometry, possible drive mechanisms, well and completion configurations, level of detail required, type and amount of data available, and so on. As far as reservoir simulation is concerned, bigger is not necessarily better. We must exercise good engineering judgment in establishing the scope of our study. We need to avoid overkill, but at the same time, understand that under-representing the needed details is dangerous. Simply put, we must strike a balance.

Single-Phase Flow Equations

Single-phase flow in petroleum reservoirs is rare in practice. There are only a limited number of casesdry gas reservoirs, for examplewhere conditions exist for single-

phase flow. But we do apply single-phase flow assumptions (predominantly in well test analysis) as a means of simplifying problems and rendering them analytically tractable. In numerical reservoir simulation, we may relax these types of simplifying constraints because of the more versatile nature of numerical schemes over analytical methods. Moreover, even for single-phase systems, an appropriate description of real problems usually involves considering non-linear phenomena, which are often ignored or approximated for ease of mathematical handling. Again, numerical schemes are not limited by these constraints. Therefore, this discussion on single-phase flow problems has not only a pedagogical basis, but a practical one as well.

Darcys law and the concept of flow potential

Darcys law is central to describing fluid flow in petroleum reservoirs. It distinguishes flow in porous media from flow in other domains (such as pipes and conduits), and represents a kind of constitutive relationship between the pressure field and the velocity field. Although Henry Darcy derived this relationship empirically in 1851, it has since been proven that Darcys law is a special form of the Navier Stokes equation, which is commonly used in fluid mechanics. Simply put, Darcys law expresses a functional relationship between the fluid velocity and the porous media and the potential gradient.

(3.1) This law appears in different forms. Table 1 summarizes the forms used in this text, along with their appropriate units.

Table 1

Velocity is the flow rate divided by the cross-sectional area perpendicular to flow. Equation 3.1 expresses the direct proportionality between flow rate and potential gradient, with the proportionality constant reflecting properties of both the flowing fluid and the porous medium. Note that as written, Equation 3.1 is for reservoir conditions. To express flow in surface conditions (or standard conditions), we must incorporate a formation volume factor into the proportionality constant. The flow potential , defined by M. King Hubert and usually referred to as Huberts potential, is basically a combination of pressure and gravitational fields, as shown in Table 2 .

Table 2

(3.1) where q : volumetric pressure A : cross-sectional area perpendicular to flow k : permeability : viscosity / x : potential gradient in the flow direction x : unit conversion factor

(3.2) where : Huberts potential P : pressure : fluid density g : local gravitational constant

gc: universal gravitational constant D : depth with respect to datum, taken as positive downward In practical application, g/gc is set to 1.0 and hence Equation (3.2) as

(3.3) The negative sign in Equation 3.1 shows that the potential gradient is negative in the flow direction. In Equation 3.2, the sign convention is such as to give the appropriate addition or subtraction of gravity from pressure. In this text, we use the positive downward convention for depth, as shown in Figure 1.

Figure 1

This figure shows a sloping reservoir, where the datum is at sea level (the datum could be any fixed elevation, although sea level is the most widely accepted convention).Using the positive downward convention, the depth to point 1 from the datum is positive; the depth to point 2 is negative. We can calculate the potentials at points 1 and 2 as:

(3.4)

(3.5)

Assuming uniform pressure within the reservoir (P1 = P2 ), if we subtract from ,

(3.6) we obtain a positive , which reflects the hydrostatic head exerted by the fluid column of density and height D1 + D2. It is the gradient of potential, rather than the potential itself, which appears in Darcys law (Equation 3.1). It is therefore clear that if we take the derivative of Equation 3.3, we obtain

(3.7) In resolving Equation 3.7, we assume that density is constant. Furthermore, note that we can write Equation 3.7 for the other principal flow directions as well, yielding depth gradients not only along the x-direction but also along the y- and z-directions (i.e., D/y, D/z). If the depth gradient along any of the flow directions does not exist, then the potential gradient becomes equal to the pressure gradient. Equation 3.7 then becomes

(3.8) Figure 2 shows four reservoirs with varying orientations relative to the datum plane.

Figure 2

Part (a) of Figure 2 shows a reservoir where the x-y plan is parallel to the datum surface. In this case, there is no depth gradient along the x- and ydirections. In part (b) of Figure 2, the reservoir is tilted so that the x-y plane is
no longer parallel to the datum, but the edges along the y-axis remain parallel. Thus, depth gradient along the y-direction is still non-existent, whereas the depth gradient now exists in the x-direction.

The reservoir in part (c) of Figure 2 is similar to the one in part (b), except now the edges of the reservoir along the x-direction are parallel to the datum surface, yielding a vanishing depth gradient in the xdirection, but not in the y-direction. Part (d) of Figure 2 represents the most general case, in which
neither the x- nor the y-direction edges are parallel to the datum surface. This means that the depth gradient is non-zero in both the xand y-directions.

In all four of these cases, the reservoir configurations are such that the formation thickness is always measured parallel to the direction that depth is measured. This is why, in each of these cases, the depth gradient along the z-direction is always unity. We usually formulate the differential equations governing fluid flow in porous media based on the continuum assumption, in which we consider a differential element of the system and take balances over a conserved quantity of interest. When the quantity is mass, the resulting equation is the mass balance equation or the continuity equation. Figure 3 shows a representative element (control volume) of the reservoir in Cartesian coordinates.

Figure 3

Note that while the control volume will differ according to the coordinate system chosen, the basic strategy is the same .

Conservation of mass
The conservation of mass principle simply says that over a fixed time period, [Mass in] - [Mass out] = [Net change in mass content] Applying this principle to the system in Figure 3 , we obtain the continuity equation shown in Table 3 .

Table 3

The porosity term in the right-hand-side of the Equation (3.9), if treated as a constant, will come out of the differential operator. This is a reasonable assumption for a reservoir with low rock compressibility. With appropriately defined terms and parameters, Equation 3.9 is general and can be used for any system. To specialize it to porous media, we must invoke Darcys law. Substituting Darcys law, written in terms of velocity as:

(3.12) we obtain the flow equation for porous media. In a rectangular coordinate system, this equation becomes ecuacion de calor o de difucion

(3.13) In a radial co-ordinate system, it takes the form

(3.14)

Equation (3.13) does not take into account the existence of wells and possible variations in formation thickness along the flow directions. In practice, we must incorporate both of these factors into the equations. In reservoir simulation, we treat the well (which could be a producer or injector) as a source or sink within the system. In this text, we follow the convention of treating injection as positive (source) and production as negative (sink). Therefore, injection or production wells are represented in the same fashion in the flow equation, except for the sign. With this in mind, Equation (3.13) becomes:

(3.15) In formulating these equations, we have made no assumption about the nature of the fluid. Of course, this will come into the picture when we take into account that the reservoir fluid can be treated as incompressible, slightly compressible or compressible. We shall now specialize the continuity equation for these fluid types.

Incompressible flow equation

For an incompressible fluid, density and viscosity are constant and formation volume factor is equal to unity. In addition if we assume that porosity does not vary with pressure, we obtain: ecuacion de continuidad

(3.16) Equation 3.16 is written for heterogeneous and anisotropic formations. For such a formation, and without injection or production (q=0), Equation 3.16 simplifies to:

(3.17) which is the well-known Laplace Equation. The following field units apply to Equations 3.16 and 3.17: A [ft2]; k [perms]; [psi]; x, y, z [ft]; qsc [STB/day]; [cp].

Slightly compressible flow equation

For a slightly(levemente) compressible fluid, density, viscosity and formation volume factor exhibit weak dependence on pressure. Furthermore, for a slightly compressible fluid, we usually assume that compressibility does not vary within the pressure range of interest.

(3.18) For slightly compressible fluids the changes in viscosity and formation volume factor with pressure are negligible and they can be treated as constants. Furthermore, if we assume that we are dealing with homogeneous and isotropic porous media with no well, Equation 3.18 reduces to a simpler form, which is known as the diffusivity equation.

(3.19) The group in Equation 3.19 is the hydraulic diffusivity constant for the reservoir fluid system. Note that the transport phenomenon described by Equation 3.19 is not a diffusion process, but a laminar flow problem. Equation 3.19 is analogous to the diffusivity equation in heat and/or mass transfer. Note also that in Equations 3.18 and 3.19, we assume that the depth gradients are negligible. Field units for these equations are as follows: A [ft2]; k [perms]; P [psi]; x, y, z [ft]; qsc [STB/day]; [cp]; c [psi-1]; B [bbl/STB]; [dimensionless]; Vb [ft3]; t [day]; and =5.615

Compressible flow equation

Compressible fluid flow is the norm for gas reservoirs. While the concepts that apply to incompressible and slightly compressible flow also apply here, compressible fluid flow involves additional considerations. The highly compressible nature of gas makes certain gas properties (i.e., viscosity, density, formation volume factor and compressibility factor) strongly dependent on pressure. Since we cannot assume that these properties are constant, they introduce non-linearities to the flow equations. The numerical handling of the flow equations becomes more challenging as the degree of non-linearity increases. However, one aspect we do not have to worry about is the gravitational component. This is because the low density of gas makes the gravitational contribution negligible in most cases. Therefore, the potential gradients are readily replaceable by the pressure gradients. The governing equations for compressible fluid flow are summarized as follows:

(3.20) where

(3.21) Substituting Equation 3.21 into Equation 3.20 will accentuate the inherent nonlinearity of compressible flow.

(3.22) In Equation 3.22, the sources of non-linearity are the terms P/mZ in the spatial derivatives and 1/Z in the temporal derivative. Field units for Equations 3.20, 3.21 and 3.22 are as follows: A [ft2]; k [perms]; P [psi]; x, y, z [ft]; qsc [STB/day]; m [cp]; c [psi-1]; B [bbl/SCF]; [dimensionless]; Vb [ft3]; t [day]; Z [dimensionless]; T [R]; and =5.615 It is sometimes expedient to linearize the non-linear equations of compressible fluid flow in porous media (Equation 3.22). The two common approaches are called the P2 method and the pseudo-pressure method. In the approach we simply recognize that Pdp = 1/2[d(P2)] and assume that the gZ product at low pressures is constant. The dependent variable in the resulting equation is now in P2 rather than in P. The pseudo-pressure approach (real gas potential) uses the transformation

(3.23) By implementing this transformation, Equation 3.22 is linearized and the dependent variable is now P. The assumptions made in formulating the approach are not as drastic as we might think when we observe the plot of gZ versus P ( Figure 4 ).

Figure 4

At low pressures, the gZ product is essentially constant; at intermediate pressures (40 to 130 atmospheres), it exhibits some non-linearity, and at high pressures, it becomes linear with pressure. In fact, at high pressures, if we treat the P/ gZ group as constant as depicted in Figure 4 , Equation 3.22 becomes similar to the slightly compressible fluid flow equation. This is not surprising, because gases do start to behave like liquids at higher pressures. The linearized forms of the compressible fluid flow equation are summarized as follows: P2 form:

(3.24) where gi and Z are calculated at some average pressure P. P* form:

(3.25) where gi , Zi and ci are calculated at the initial pressure. Field units for Equations 3.24 and 3.25 are as follows: A [ft2]; k [perms]; P [psi]; x, y, z [ft]; qsc [SCF/day]; [cp]; c [psi-1]; f [dimensionless]; Vb [ft3]; t [day]; Z [dimensionless]; T [R]; =5.615; P* [psi2/cp]

Multiphase Flow Equations

Multi-phase flow equations are based on the same principles that govern singlephase flow, except that they must account for interactions between simultaneously flowing phases in porous media. The main parameters that we use to characterize these interactions are relative permeabilities, saturations and solution gas-liquid ratios. The reader will recall the process of formulating the single-phase flow equations. Basically, we obtain the flow equation by substituting Darcys law into the continuity equation. For multi-phase systems, we write the continuity equation

for each of the phases. Then we use an appropriate form of Darcys law, which accounts for the presence of multiple-fluid flux terms (left-hand side of the equation) to characterize the transport part of Equation 3.9. At the same time, we adjust the phase accumulation term using phase saturations. The number of partial differential equations depends on how many phases are present. This development is summarized in Table 1 .

Table 1

Two-phase (oil-water) equations

In reservoirs where two-phase flow of oil and water phases dominates (typically in dead oil reservoirs with no gas), we need to write Equation 3.28 for the oil and water phases separately. We can do this easily by setting the subscript f first to o for oilphase and then to w for water phase. The complete set of equations for two-phase oil-water transport problems, together with the unknowns to be solved, is summarized below. In these equations, we have used pressure gradients rather than potential gradients; in other words, we have neglected depth gradients. Oil flow equation (f defined as o):

(3.29) Water flow equation (f defined as w):

(3.30) In Equations 3.29 and 3.30, there are four unknowns: oil-phase pressure, Po , waterphase pressure, Pw , oil-phase saturation, So , and water-phase saturation, Sw. To solve the system, we need two more equations. These equations, called the auxiliary equations, are Capillary pressure relationship: Pcow (Sw) = Po - Pw (3.31) Saturation relationship: So + Sw = 1 (3.32) With these last two equations, we now have a well-posed problem (four equations in four unknowns). Field units for Equations 3.29 through 3.32 are as follows: A [ft2]; k [perms]; kro krw [fraction]; P [psi]; x, y, z [ft]; o, w [cp]; Bo, Bw [bbl/STB]; qosc, qwsc [STB/day]; So, Sw [fraction]; [dimensionless]; Vb [ft3]; t [day]; =5.615.

Two-phase (oil-gas) equations

In a volumetric reservoir, there are three phases present (oil, water and gas); at irreducible water saturation, the dominant flow is oil and gas. In representing flow in this kind of reservoir, we must account for both the free gas and the gas dissolved in the oil phase.

Figure 1 depicts a two-phase (oil-gas) system, considering flow only along the xdirection for illustrative purposes.

Figure 1

Note that while there is only one flow term for the oil phase, the gas phase has two flow terms, which describe free gas flow and dissolved gas flow. The two gas flow models must also be taken into account in the source and accumulation components of the governing flow equation for the gas phase. The final equations for the twophase flow of oil and gas are shown below. Oil Flow Equation (f defined as o):

(3.33) Gas Flow Equation (f defined as g)

(3.34) In the flow terms of Equation 3.34, the second term in each bracket represents the contribution from the gas dissolved in oil. Similarly, qgsc represents the free gas produced (injected), while the product Rsoqosc represents the dissolved gas produced along with oil. Finally, the second term under the temporal derivative represents the accumulation (depletion) of gas dissolved in oil. Again, the auxiliary equations necessary to complete this formulation are the capillary pressure and saturation relationships. Pcow (Sw) = Po - Pw (3.35) So + Sw = 1 (3.36)

Field units for Equations 3.33 through 3.36 are A [ft2]; k [perms]; kro, krg [fraction]; P [psi]; x, y, z [ft]; o, w [cp]; Bo [bbl/STB]; Bg [bbl/SCF]; qosc [STB/day]; qgsc [SCF/day]; So, Sg [fraction]; [dimensionless]; Rso [SCF/STB]; Vb [cf]; t [day]; a=5.615.

Two-phase (gas-water) equations

In an aquifer-driven gas reservoir, simultaneous flow of gas and water takes place, requiring us to formulate two-phase gas-water equations. The formulation is very similar to that of the oil-gas flow equation, except that we replace the oil phase with the water phase. In most practical cases, the solubility of natural gas in water is quite small, allowing us to neglect the terms of the flow equation that arise from the dissolved gas. One notable exception is the case of geopressured aquifers, where dissolved gas is significant because of the prevailing high pressures. Gas flow equation (f defined as g):

(3.37) Water flow equation (fw):

(3.38) Auxiliary equations: Pcgw (Sw) = Pg - Pw (3.39) Sw + Sg = 1 (3.40) Field units for Equations 3.37 through 3.40 are A [ft2]; k [perms]; krw, krg [fraction]; P [psi]; x, y, z [ft]; g, w [cp]; Bw [bbl/STB]; Bg [bbl/SCF]; qw [STB/day]; qg [SCF/day]; Sw, Sg [fraction]; [dimensionless]; Vb [cf]; t [day]; =5.615.

Three-phase (oil-water-gas) equations

The widest application of reservoir simulation is for three-phase oil-water-gas systems, in which all three phases are active in the production process. In fact, a large majority of petroleum reservoirs fall into this category. For this class of problems, we need to write Equation 3.28 for each of the three phases. Oil flow equation (f defined as o):

(3.41) Water flow equation (f defined as w):

(3.42) Gas flow equation (f defined as g; ignoring the solubility of gas in water):

(3.43) Auxiliary equations: Pcow (Sw) = Po - Pw (3.44) Pcgw (Sg) = Pg - Po (3.45) So + Sw + Sg = 1 (3.46) In Equations 3.41 through 3.46, the dependent variables (unknowns) are the phase pressures (Po , Pw and Pg), and phase saturations (So, Sw and Sg). The parameters that appear in the coefficients, and which are functions of these unknowns (mo, krw, etc.), are not treated as the problems principal unknowns, but are specified as part of the data input. However, their dependence on the principal unknowns introduces non-linearities of varying degrees. Field units are A [ft2]; k [perms]; kro, krw, krg [fraction]; P [psi]; x, y, z [ft]; o, w, g, [cp]; Bo, Bw
[bbl/STB]; Bg [bbl/SCF]; qosc, qwsc [STB/day]; qgsc [SCF/day]; So, Sw, Sg [fraction]; Rso [SCF/STB]; f [dimensionless]; Vb [ft3];

t [day]; =5.615.

In all of the multi-phase formulations summarized in Equations 3.29 through 3.46, each phase is characterized by its own pressure. We achieve closure of a set of equations by using the capillary pressure and saturation expressions. In the capillary pressure relationships, capillary pressure is defined as the difference between the

pressure of the non-wetting phase and that of the wetting phase. In some cases, there are incentives for assuming equality between the phase pressures, thus allowing us to assign a single pressure for all the phases. The formulation presented in Equations 3.41 through 3.46 allows interphase mass transfer, so that gas can either come out of or go into solution. However, we assume bulk transfer such that there is no compositional difference between the dissolved gas and the free gas. By the same token, the composition of the free gas remains unaltered. Changes in the oil phase density and viscosity are handled using PVT data, which take into consideration the amount of gas dissolved in the oil. This approach is commonly known as black-oil modeling, and it differs significantly from compositional modeling.

Flow Equations Based on Individual Components

Until now, we have emphasized the bulk properties of the mobile phases in the porous media without considering their compositions and compositional changes. While this is adequate for many systems, there are others in which compositional effects are significant and do affect both the transport mechanisms and bulk transport of the phases involved. Examples of such systems include volatile oil reservoirs, gas condensate reservoirs and enhanced oil recovery applications such as chemical flooding, miscible flooding, steam injection and so on. In all these cases, the systems phase behavior plays a dominant role in the transport and fluid distribution mechanisms. Numerical simulators for handling these specialized problems are usually more complicated and require a higher degree of numerical maturity and sophistication.

Isothermal compositional formulation

The strategy for formulating the governing equations of isothermal compositional systems is the same as for the black-oil model, except the focus is on each component within the control volume rather than the bulk phase. However, we normally assume that there is no self-diffusion of a component within the phase, and hence the component velocity assumes the value of the phasial velocity. As previously mentioned, since the systems phase behavior plays a dominant role, procurement of the necessary PVT data is crucial. In modern computations, we often obtain these PVT data by using an appropriate equation of state. The common ones used in the petroleum industry include the Peng-Robinson equation of state (or any of its modifications) and the Soave-Redlich-Kwong equation of state. One of the major considerations in using these equations of state is that by virtue of their cubic form, they generally require much less computational overhead than other, more complex relationships. This is especially significant when we consider the repetitive calculations that are required in reservoir simulation. The calculations involved provide such critical information as distribution of components between the phases, density variation of each of the phases, phase viscosity and other thermo-physical properties. Isothermal flow is a special case where existing temperature gradients within a given reservoir are negligible. Therefore, in isothermal compositional formulation,

temperature distribution does not come into play. Hence, all the phase behavior and transport calculations are performed under isothermal conditions. For an N-component system, we can write N equations of the form of Equation 3.47. Thus, we will have N equations in (3N+6) unknowns. Table 2 summarizes the auxiliary equations (2N+6) needed for closure.

Table 2

Non-isothermal compositional formulation

For systems in which either the intrinsic temperature gradient or the externally imposed ones are significant, the energy balance will play an important role in the flow of fluids and their distribution in the porous media. A highly dipping reservoir may be subjected to a substantial geothermal temperature gradient. Reservoirs where such EOR processes as steam flooding and in situ combustion are taking place will definitely experience severe temperature gradients. These systems call for consideration of the energy balance. Usually, we use additional equations to account for enthalpy balance in the reservoir. In any case, the additional equation is sufficient to achieve closure in view of the additional dependent variables. Needless

to say, with the additional equations comes increased computational overhead. The energy equation usually accounts for heat transfer by heat conduction and convection as well as heat storage. The energy and molar balance equations are related by the various coefficients that appear in both of them. Boundary and Initial Conditions Mathematically speaking, the differential equations that describe flow in porous media have an infinite number of solutions, only one of which will describe a particular problem. We can obtain this solution by imposing additional constraints known as boundary and initial conditions. In actuality, we impose these conditions as part of the known data about the system under consideration. For instance, for a reservoir with strong edge-water drive, we can set the pressure at the boundary to a constant equaling the initial reservoir pressure, which implies that the reservoir is in direct communication with an infinitely large aquifer. The systems physical boundary is divided into two groups: the inner boundary (usually the location where the well is physically coupled to the reservoir) and the outer boundary (usually the limits of the reservoir). The conditions that are specified at these boundaries constitute the boundary conditions. The most often used boundary conditions can be grouped into two main categories, Dirichlet and Neumann type. In a Dirichlet type boundary condition, the values of the dependent variables are specified at the boundaries, whereas it is their gradients that are so specified in the Neumann type. Depending on what we know at the wellbore (either flow rate or pressure), we can specify the Dirichlet-type boundary condition (i.e., pressure) or Neumann-type (i.e., flow rate). Similarly, at the outer boundarydepending on the physical characteristicswe may be able to specify either Neumann or Dirichlet-type boundary conditions. For instance, if we have a sealed boundary that allows no flow into or out of the reservoir, we specify the vanishing Neumann-type boundary condition (i.e., zero pressure gradient across the boundary). On the other hand, a strong edge water drive enables us to specify the Dirichlet-type boundary condition (i.e., pressure specification). Numerical handling of the inner and outer boundary conditions is a crucial component in conducting reservoir studies. Similar to the boundary conditions, it is necessary to describe the original state of the system before the process under consideration begins (for instance, before initiating a production or injection process). The conditions that describe the values of the dependent variables at the pre-set time (such time is usually set to zero) are known as the initial conditions for the problem. The existing hydrostatic head distribution determines the initial saturation and pressure distribution. With the imposition of the boundary and initial conditions, the formulation of the problem is now complete and ready for solution. What we have is a well-posed mathematical problem which guarantees the existence of a unique solution. Analytical vs. Numerical Solution Methods There are two techniques for solving mathematical reservoir models: analytical and numerical. Each of these has certain strengths and limitations.

 Analytical, or closed-form techniques offer the advantage of providing exact solutions (when they can be found); furthermore, those solutions are continuous throughout the system. The types of problems that are amenable to analytical solution, however, are very limited. Analytical methods fall short when we start dealing with varying formation thickness, non-uniform porosity and permeability, and changing fluid properties, and other such conditions that describe most real reservoirs. To find analytical solutions for the type of system that "Mother Nature" generally provides, we have to modify the problem sometimes quite drasticallyto make it plausible for handling analytically. What we end up doing is providing an exact solution to an approximate problem (e.g., a classical well test analysis model). A numerical solution involves discretizing, or approximating the mathematical modelthat is, using a numerical tool such that continuous forms of the partial differential equations are written in a discrete form. We perform this discretization process not only on the partial differential equation, but also on the physical systems. This means that we divide the physical system into a number of subdomains that are coupled to one another. The clear advantage of the numerical approach is that it allows us to assign representative properties to as many parts of a system as we have information for. However, we must not forget that we inevitably lose some measure of accuracy in discretizing the partial differential equations. The net result is that in using a numerical approach, we are providing an approximate solution to an exact problem. Numerical Models: Grid Systems Numerical models provide approximate solutions to exact problems. The degree of exactness depends largely on the discretization scheme, the heart of which is the type and size of the grid. Therefore, we must pay adequate attention to selecting the appropriate grid system. Although this discussion is limited to rectangular coordinate systems, the same basic principles apply to other coordinate systems. Body-centered grids In body-centered (or block-centered) grids, the discrete points are located at the center of each cell ( Figure 1 ).

Figure 1

There are no discrete points at the reservoirs external boundaries. In setting up the grid system, we must be especially careful in placing the grid with respect to the wells. First, the well should be located at the center of its host grid block. Second, each grid block should have no more than one well within the same grid block. In cases where these conditions are difficult to meet, there are techniques for handling them. For example, if it is difficult to locate the wells such that they coincide with the grid center, we can shift them slightly to satisfy this condition. Also, if we cannot avoid having more than one well in the same grid block, we can mathematically replace them with one well of equivalent strength. Figure 2 shows the same multi-well reservoir in which two wells are combined into one equivalent well (wells W1 and W2) located in the center of a new block.

Figure 2

In doing so, the number of blocks along the x- and y-directions is decreased by one, which results in an overall decrease of 6 blocks. Note that in using a finite number of grids to develop a numerical model, the reservoirs physical boundary cannot always coincide with a grid-imposed boundary. The finer the grid, the closer the two boundaries will be. It is necessary to compromise between how much boundary information needs to be captured (via finite grids) and the amount of computational overhead involved. Obviously, as the number of grid blocks increases, so do the memory and the CPU time requirements. Mesh-centered grids In mesh-centered (or point-distributed) grids, the discrete points are located at the grid line intersections . In contrast to body-centered grids, there are discrete points located on the reservoir boundaries ( Figure 3 ).

Figure 3

As is true for body-centered grids, it is permissible to slightly relocate the wells so that they lie at the intersection of the grid lines. As shown in Figure 3 , the discrete points in a mesh-centered grid are not necessarily located at grid block centers. In mesh-centered grid systems, the discrete points that are located at the corners and edges of the grid represent only a certain portion of the block they are associated with. For instance, a discrete point located on the corner may represent 1/4 or 3/4 of a block, whereas the discrete points located on the edges of the grid system represent only 1/2 of a full block. The finite difference approximations of partial differential equations are independent of the grid system used. There are some practical advantages to using a bodycentered grid in the case of no-flow boundaries. For systems with constant pressure boundaries, mesh-centered grids offer better accuracy, since the discrete points (at which pressure is specified) exist at the reservoir boundaries. Numerical Models: Grid Systems Size and number of grid blocks Grid size and the number of grid blocks are not independent of each other. In a fixed system (i.e., a defined reservoir), specifying the grid size determines the number of grid blocks. There is no hard and fast rule for selecting the grid size for a simulation study. This does not mean that we have unlimited freedom in selecting the grid size. The degree of freedom is often limited by the amount of input data available, the information we want to gain from the study, and the investment we are willing to make in terms of computational overhead. In any case, there are a number of factors that impose lower and upper bounds on our choice of grid block size. Within these limits, however, we do have some degree of freedom. Table 1 summarizes the

most salient factors that we need to consider.

Table 1

We must always be aware that the quality of output depends on the quality of input. Each block added to the system demands information in terms of accurate property representation. A 40-foot blanket sand without significant porosity and permeability variation, or without property information in the vertical direction, does not require subdivision into several layers. By a similar token, in terms of areal coverage, it is unnecessary to divide a reservoir into blocks beyond where information is known, or where there is significant lateral property variation. In these two examples, the only consideration on which we must base our selection is mathematical requirements, which we will discuss later. A principal goal of reservoir simulation is to define the level of information or detail desired. We must define a priori what this goal is and what questions we need to answer at each phase of the study. For instance, if we need to further develop a certain portion of a producing field, we will need to focus on that area, perhaps by imposing a larger number of blocks to better represent saturation and pressure distributions. Similarly, if a production well requires stimulation work that involves perforating and/or fracturing, then we must focus attention in that section by using finer grids. In some reservoir studies, the quality and detail of the information we need may require us to use much finer grid blocks. A case in point is the simulation of an EOR process, where we need to track a fluid front in order to better control the outcome. In instances like steamflooding or miscible flooding, where changes are abrupt and the interactions between the phases are strong, a fine grid system will help capture the details of the changes taking place.

Reservoir geology is often structurally and stratigraphically complex. Representing this complexity with a grid system sometimes requires the use of a large number of grid blocks. In other words, it is not possible to construct a grid system for a lenticular sand with several pinch-outs without imposing finer grids that capture peculiar features of this system. The number of wells in the reservoir plays a dominant role in grid size selection. We have already noted that ideally, each block should host only one well. We also have to consider that the presence of wells in the reservoir further complicates an already complex system, since most of the significant activities and rapid changes that take place in the reservoir occur in the near-wellbore region. As a rule, the more wells we have in a reservoir, the more grid blocks we need. Approximating a partial differential equation using its finite-difference analogue introduces certain errors. The magnitude of these errors and the stability of the numerical algorithm depend on the grid size. In general, we can say that finer grids produce more accurate results and, in fact, an infinite number of grid blocks enables a discrete numerical model to collapse to a continuum representation of the problem. On the other hand, one must not lose sight of the fact that more grid blocks means higher computational overhead. As the number of grid blocks increases, so do the memory and CPU time requirements. It is the responsibility of the simulation engineer to strike an acceptable balance between these factors. Grid orientation The first step in constructing a numerical model is the placement of the axes. Sometimes, for obvious practical reasons, an inexperienced engineer may be tempted to place the axes parallel to the margins of the paper. But it isnt that simple. In orienting the grid, we have to consider two factors: Permeability anisotropy, if any Coordinate orthogonality In anisotropic formations, it is imperative to align the coordinate axes with the principal flow directions. In doing so, however, we must ensure that we preserve the coordinate systems orthogonality. When the two principal flow directions are not orthogonal, the major flow directions should be aligned with one of the axes. We can most easily illustrate the effect of grid orientation by considering an isotropic reservoir in which there is no principal flow direction. Consider the five-spot pattern in this system as shown in Figure 1 .

Figure 1

In studying this system on a unit basis, we have the option of placing the grid so that production and injection wells are positioned diagonally with respect to each other. The second option is to construct the grid on a larger unit, so that the injection and production wells are on adjacent corners. Note that although we are solving the same problem, the two different grid alignments result in two different solutions. The difference becomes more significant for cases when the mobility ratio is greater than one. It becomes even further pronounced in the case of unfavorable mobility ratio when finer grid blocks are used. However, for favorable mobility ratio, refining the grid diminishes the disparities between solutions. One plausible solution to this problem is to use a higher order finite-difference approximation (e.g., a 9-point finite-difference scheme in areal studies) which brings in the diagonal interaction between the discrete points. Local grid refinement We have emphasized that while grid refinement does yield greater accuracy in numerical reservoir simulation, a studys cost tends to increase in direct proportion to the number of grid blocks it uses. A good compromiseand a way to avoid placing blocks where they are not neededis to refine the grids around the locality where we require more detailed information, and/or where rapid changes occur. This is called local grid refinement. Figure 2 shows several refinement strategies that we can use in conducting reservoir simulation studies.

Figure 2

Part (a) of Figure 2 shows a coarse grid composed of 81meshcentered grid points. The shaded area around the well is the section from which we need detailed information. We could get this information by performing a global grid refinement as shown in part (b) of Figure 2, which involves a total of 289 discrete points. Such refinement over the entire reservoir, however, is unnecessary. Part (c) of Figure 2 depicts the conventional approach for resolving this problem. This approach results in the desired level of refinement not only around the wellbore, but also outside of the area of interest, yielding a total of 169 cells. Part (d) of Figure 2 illustrates the best solution, which limits refinement to the area of interest. This scheme uses a total of 137 grid points. We should emphasize that the type of local grid refinement shown in (d) requires specialized handling of the non-continuous grid lines at the interface between the refined and the coarse areas.

Figure 3 shows a slightly different refinement strategy.

Figure 3

In this scheme, the level of refinement increases as we move toward the well, where the most rapid changes take place. A total of four levels of refinement are shown in conventional local refinement and true local refinement schemes (Parts (a) and (b), respectively). The ultimate refinement strategy involves dynamically changing the grid size as needed. For example, in following a flood front movement, the area being refined continually changes as the front propagates. In this type of adaptive grid refinement scheme, we use a certain set of criteria such as pressure and/or saturation gradients to delineate the region that needs the refinement. We use the same set of criteria to remove the previously refined region as the front moves away from that region. Dynamic grid refinement requires a high level of numerical maturity and sophistication, and so is only available in advanced simulators.

Finite-Difference Approximation of Reservoir Flow Equations

In addition to converting the spatial and temporal domains (i.e., the reservoir and time) into a set of coupled discrete sub-domains we need to convert the continuous form of the governing differential equations into a discrete form. To do this, we need to utilize some numerical tools. There are a number of tools that we could employ, including finite-difference, finite-element, collocation techniques and so forth. The most widely used numerical technique in reservoir simulation is the finite-difference approach. Adems de convertir los dominios espaciales y temporales (es decir, el depsito y el tiempo) en un sistema de secundario-dominios discretos juntados necesitamos convertir la forma continua de las ecuaciones diferenciales que gobiernan en una forma discreta. Para hacer esto,

necesitamos utilizar algunas herramientas numricas. Hay un nmero de herramientas que podramos emplear, incluyendo la diferencia finita, finito-elemento, tcnicas de la colocacin y as sucesivamente. La tcnica numrica lo ms extensamente posible usada de la simulacin del depsito es el acercamiento de la diferencia finita The finite-difference approach gives us a great deal of flexibility in handling the nonlinear partial differential equation, in addition to the property distribution in heterogeneous systems for which an analytical solution is not feasible. The governing equations, as well as the boundary conditions used for describing flow in porous media, have only first-order and second-order derivatives, and so we will limit our discussion to these.

First-order derivative
First-order derivatives appear in the governing equations on the right-hand side in the form of the time derivative (the accumulation term). In addition, first-order derivatives appear when the gradient is specified across a given boundary. To approximate the first-order derivatives, we use truncated Taylor series expansion. Figure 1 introduces the notation we will use throughout this discussion.

Figure 1

In this figure, we show a hypothetical pressure distribution, P, along the x-direction. The derivative is approximated at the point x (also designated as i) at which the value of pressure is Pi. The two neighboring points to the central point are x - x (also designated as i -1) and x + x (also designated as i - 1). Accordingly, at these two neighboring points, the pressure values are Pi - 1 and P i + 1, respectively.

In general, a Taylor series expansion for evaluating a function f (x) at x + x can be written

(4.1) Using the notation of Figure 1 , we can write Equation 4.1 as:

(4.2) Forward-Difference Approximation: The forward-difference approximation to the first-order derivative at point uses the values of the function at points i and i +1. We can obtain this approximation from Equation 4.2 by truncating all terms after the first-order derivative and rearranging the equation to solve for (P/x)i. Note that in forward differencing, only the + sign in Equation 4.2 is relevant. The result is

(4.3) The second term on the right-hand side of Equation 4.3 denotes the error in the approximation and is read as "order of ." The magnitude of the error is the same as the magnitude of . In practice, the error term is dropped in writing the finitedifference analogues. Thus, we write:

(4.4) which is clearly an approximation. Figure 2 shows the geometrical interpretation of this approximation.

Figure 2

The exact derivative of P at point is the slope of the tangent AB to the curve at point O. Equation 4.4 uses the slope of the secant OC to approximate this derivative. As we can conclude from the figure, the accuracy of this approximation depends on the shape of the curve, as well as the length of the interval between and . In the limit as point is moved progressively closer to the secant OC tends to obtain the same shape as the tangent AB. In reservoir simulation, we use the forward difference approximation in conjunction with the temporal derivative. The two neighboring points in this case represent the old time level and the new time level . When this derivative is written at a point i in the spatial domain it represents the rate of change of pressure with time at point i. In other words,

(4.5) In Equation 4.5, is the unknown that we want to determine, and represents the known value of pressure at the old time level. Remember, we are using a marching scheme in time, where we obtain information at the new time level from what is already known at the old time level. Backward-Difference Approximation:

With the same strategy, but now using the neighboring point , we can obtain the backward-difference approximation to the first-order derivative from Equation 4.2 using the (-) sign:

(4.6) Figure 3 shows the geometrical interpretation of Equation 4.6.

Figure 3

Central-Difference Approximation: The central-difference approximation to the first-order derivative at point i uses the two adjacent neighboring points and . We can obtain this approximation by adding Equations 4.4 and 4.6:

(4.7) Figure 4 shows the geometrical interpretation of Equation 4.7.

Figure 4

It is apparent from Figure 4 that central-difference provides a more accurate approximation to the first-order derivative than either the forward- or backwarddifference approximation does. Note that the secant CE is almost parallel to the desired tangent AB. We can explain this increased accuracy by deriving Equation 4.7 directly from Equation 4.2. In this case, when we write Equation 4.2 two times (once with the (+) sign and once with the (-) sign), and then subtract them from each other, the second-order terms drop out (not truncated) and the truncation starts with the third-order derivative, resulting in

(4.8)

Second-order derivative
The left-hand-side of the flow equation is composed of second-order derivatives representing the flux terms. To approximate these second-order derivatives, we use central-difference approximation. This involves writing Equation 4.2 once with the (+) sign and then with the (-) sign; we then add the two resulting equations together to yield

(4.9) Dropping the error term, Equation 4.9 becomes ESTA ES LA APROXIMACION DE LA ECUACION EN DIFERENCIAS FINITAS

(4.10) For heterogeneous systems where properties vary spatially, the second-order

derivatives appear in the form < from Equation 4.10:

, and the approximation is different

(4.11) Note that in Equation 4.11, the dependent variable P is expressed at the nodal points, whereas the coefficient a is expressed at the block boundaries between nodes i and i + 1 and i and i - 1. The subscripts i + 1/2 and i - 1/2 simply indicate that a needs to be calculated using some averaging technique boundaries between blocks i and i + 1 and i and i - 1 respectively.

Finite-difference schemes
There are two principal groups of finite-difference schemes: explicit and implicit. We can illustrate the governing concepts for these schemes using the classical diffusivity equation as it is written in one dimension:

(4.12) Equation 4.12 has a second-order spatial derivative and a first-order temporal derivative (a is assumed constant). Using the finite-difference approximations for these derivatives we can write the finite difference analogue for Equation 4.12. Before we do that, however, we must answer one question: at what time level do we implement the approximation to the spatial derivative? The two most common approaches are to evaluate them at the old time level (time level n) or the new time level (time level n+1). These lead, respectively, to explicit and implicit schemes. Accordingly, the explicit finite-difference analogue to Equation 4.12 is

(4.13) Since pressures at the old time level are known at all locations, the only unknown in Equation 4.13 is . In fact, we can rearrange Equation 4.13 to solve for this unknown:

(4.14) When Equation 4.14 is written for every node, we have one unknown in one equation for each node. Thus, we can solve explicitly, one unknown at a time, for the pressures at the new time level. The implicit finite difference analogue to Equation 4.12 is:

(4.15)

It is obvious that Equation 4.15 has three unknowns < . By collecting all the unknown terms on one side, we obtain the characteristic form of the implicit finite-difference scheme to Equation 4.12:

(4.16) When we write Equation 4.16 for all the nodes where pressure is unknown, we obtain a system of simultaneous equations. The solution to this system of equations provides the pressure distribution at the new time level. With the explicit scheme (Equation 4.14), we only need to solve one equation at a time for the unknown pressure at a particular node. For the implicit scheme (Equation 4.16), we must solve all the equations simultaneously. But while the explicit scheme looks more attractive in terms of its simplicity, it is less often used because of its restrictive stability constraints.

Flow equations in the finite-difference form

Having studied finite-difference approximations to partial differential equations and the differences between the explicit and the implicit formulations, we can now write a

finite-difference analogue for porous media flow equations. For illustrative purposes (and for the sake of brevity), we will develop this analogue for a single-phase flow equation. Let us consider the flow equation for the slightly compressible single-phase flow problem in three dimensions:

(4.17) To approximate the second-order derivatives that appear in Equation 4.17, we invoke Equation 4.11, and obtain

(4.18) We write Equation 4.18 in the most explicit form simply because it is conducive to recognizing one of the most important groups in reservoir simulation, the transmissibility terms. In Equation 4.18, the pressure coefficients on the left-handside are known as the transmissibility. These define the interaction between two neighboring grid blocks. Figure 5 shows two neighboring blocks (i, j, k and i+1, j, k) whose properties are different, and the transmissibility term,

Figure 5

at the interface.

calculated

The term Axkx/x is the constant part of the transmissibility group, while the term B could depend on the pressure. To calculate these two terms at the interface, we must use averaging procedures. For the constant part of the transmissibility, if we consider the two blocks as being connected in series, it becomes necessary to use harmonic averaging, such as

(4.19) The second term (B) is considered a weak function of pressure; that is, it exhibits weak non-linearity. We calculate this term at the arithmetic average of pressures at the two neighboring blocks:

Note that under multiphase flow conditions, a typical transmissibility term will include relative permeability to the fluid for which the flow equation is being written. In that case, the transmissibility group between the blocks i, j, k and i+1, j, k is

(4.20) We calculate all of the terms in this multiphase transmissibility group as singlephase, except for the new term (krf). The relative permeability term is a strong function of saturation and exhibits a high degree of non-linearity. In expressing this term at the interface of two neighboring blocks, we need to establish the flow direction to use the saturation of the upstream block. This procedure is known as single-point upstream weighting, and can be summarized as

(4.21) A more accurate representation of the relative permeability at the interface is known as two-point upstream weighting, and is expressed as

(4.22) We obtain Equations 4.21 and 4.22 by using a Taylor series expansion to expand the relative permeability function. While Equation 4.21 yields a first-order approximation. Equation 4.22 is a second-order approximation, and thus more accurate. Although we have illustrated the transmissibility groups only for the i+1/2, j, k interface, similar analyses apply to the other interfaces. In summary, it is imperative to note that the transmissibility group consists of three components, and that each is calculated differently. The constant component is calculated using harmonic averaging The weakly non-linear component is estimated using arithmetic averaging The strongly non-linear component is obtained using upstream averaging

Incorporation of boundary conditions

To obtain a particular numerical solution to a flow problem, we must impose specific boundary conditions. First, let us consider the outer boundary. Figure 6 shows a pure

Dirichlet-type boundary condition with a mesh-centered grid system.

Figure 6

The boundary conditions imposed on the system (as shown in Figure 6 ) translate to the following equations:

(4.23) where NX and NY represent the maximum number of blocks along the x- and ydirections, respectively. The implementation of Equation 4.23 will result in pressure discontinuity at the corner points (e.g., P1, 1 = PC and PD, PD PC ). Although at first glance, this may seem to pose a problem, it does not. Finite difference equations are only written at the nodes where pressures are not specified. In a conventional 5-point finite-difference scheme, when the equation is written for node (2,2), we do not write the equation for node (1,1) or the other corner points. If we use a 9-point finite-difference scheme, then the corner points come into the picture, and we simply use the arithmetic average assigned along the two conjoining edges. Although we use a mesh-centered grid system in this discussion, the implementation of the Dirichlet-type boundary conditions for a body-centered grid system follows a

similar logic. On the other hand, the implementation of the Neumann-type boundary conditions for the body-centered and mesh-centered grid systems differ. This is because different nodes are reflected with respect to the boundary. The reflection nodes (which are actually outside of the domain of interest) constitute a practical means of incorporating Neumann-type boundary conditions. In creating a reflection node, we simply need to treat the boundary as a mirror showing the images of the actual nodes that are adjacent to the boundary. Furthermore, note that these image nodes are assigned the same properties as the actual nodes that they reflect. Figure 7 shows a body-centered grid with Neumann-type boundary conditions.

Figure 7

In this case, finite-difference representation of the boundary conditions is as follows:

(4.24)

The above Equations 4.24 contain the reflection nodes (imaginary nodes), which we can calculate as

(4.25)

Figure 8 shows a similar treatment for the mesh-centered grid system.

Figure 8

The important difference is the node being reflected. The implementation of the finite-difference approximation in this case will be as follows:

(4.26)

Again if we solve for the reflection nodes, we obtain the following sets of equations:

(4.27)

Note that for Neumann-type boundary conditions, pressures are still unknown for the boundary nodes, and so must be computed. To write the finite-difference equations for these nodes, we must obtain values of the pressure at the reflection nodes. We achieve this by using Equations 4.25 and 4.27. Although introducing the reflection nodes may appear to be adding more unknowns to the problem, each reflection node brings with it a new equation, through which the boundary condition is incorporated. Finally, for the special case of no-flow boundary conditions, pressures at the reflection nodes are equal to those at the corresponding reflected nodes (simply set the constants C1 through C4 equal to zero in Equations 4.25 and 4.27. Having dealt with outer boundary conditions, we now focus on the inner boundary (i.e., the well). Figure 9 shows a typical grid block hosting a vertical well.

Figure 9

We may express the relationship between the block pressure, the sandface pressure and the flow rate as: qsc = -PI (Pi,j,k - Psf) (4.28) where PI is the productivity index, Pi, j, k is the pressure of the block hosting the well and Psf is the sandface pressure. The sign convention is such that a positive q means injection and a negative q indicates production. The PI of a vertical well, as defined by Peaceman (1983), is given in Table 1 .

Table 1

As mentioned earlier, we can specify either the flow rate at the wellbore or the flowing sandface pressure. For flow rate, we substitute the specified value into Equation 4.18 (written for the block hosting the wellfor blocks with no wells, we set flow rates equal to zero). After solving for the pressure distribution, we come back to Equation 4.28 to solve for sandface pressure. When we specify sandface pressure, we substitute Equation 4.28 into Equation 4.18, replacing the qi, j, k term. After solving for the block pressures, we can then use Equation 4.28 to obtain the resulting flow rate. There have been a number of formulations proposed for implementing horizontal wells in reservoir simulation. One simple approach is to use Equation 4.8 with appropriately defined parameters (e.g., for a horizontal well oriented along the xdirection, we replace all the x-direction related terms in Equation 4.29 with the corresponding terms in the z-direction and vice-versa; thus, kx becomes kz, x becomes z, and h becomes x). With the implementation of the inner and outer boundary conditions, the finitedifference representation is now complete and ready for solution.

Shorthand notation for the finite-difference equations

Because finite-difference analogues are usually cumbersome, even for relatively simple, single-phase flow problems, we commonly employ a shorthand notation such as the Strongly-Implicit-Procedure, or SIP (Stone, 1968). For example, we can write Equation 4.18 in rearranged SIP form as follows:

(4.33) Z, B, D, F, H and S are the transmissibility terms representing the interaction of block i, j, k with the neighboring blocks. E represents the summation of transmissibilities and the accumulation term coefficient, while Q includes all the known terms collected on the right-hand side. Figure 10 graphically shows the interaction of block i, j, k with the surrounding blocks.

Figure 10

The principal advantage of the SIP notation is its inherent flexibility. For instance, having written it for three-dimensional flow, for one- and two-dimensional flows some of the SIP coefficients simply drop out. Table 2 summarizes the possible

combinations and their corresponding SIP coefficients. Note also that for the blocks located at the boundary, we can easily implement a no-flow boundary by setting the corresponding SIP coefficient equal to zero across that boundary.

Table 2

Special considerations
In writing finite-difference analogues for partial differential equations, we must always consider a number of important points, which principally pertain to the accuracy of the resulting solution. The specific points of consideration include solution stability, consistency and truncation error. Rarely does the user of a reservoir simulator need to worry about these problems, but one should be aware of them. Stability analysis is meant to ensure that the round-off error (which arises over time due to the computers finite word length as time evolves) does not magnify in such a way as to obscure the solution. Stability analyses focus on defining the stability criterion for a given finite-difference scheme. These types of analyses provide results such as conditional stability and unconditional stability or instability. When we encounter a conditional stability, the criterion will bring in certain limitations on the block dimensions and time step sizes. Although an unconditional stability implies no restriction on the block and time step sizes,we shouldkeep in mind that the physical meaning of the solution can be lost if we assign unrealistically large block and time step sizes. The next important question to answer is whether the proposed scheme is consistent (i.e., compatible) with the original partial differential equation. In other words, does the proposed finite-difference analogue collapse to the original partial differential equation in the limit as the block dimensions and time step size approach zero? The

truncation error is the difference between the original partial differential equation and its finite-difference analogue. For a compatible scheme, we should expect that the truncation error will disappear as the block dimensions and the time step size become infinitesimally small. If this does not happen, we have a consistency problem. In other words, the proposed scheme produces a solution to a different problem (partial differential equation). It is worth noting that in general, schemes used in reservoir simulationat least the ones discussed in this textmeet stability as well as consistency criteria. If these schemes are used, the engineer need not worry about these types of problems.

Single-Phase Flow Equations: Solution Methods

One-dimensional flow system
Figure 1 shows a simple one-dimensional system composed of five blocks.

Figure 1

A no-flow boundary condition is imposed on the left end, while a non-zero pressure gradient is specified on the right. A well with a flow rate qs is located in block #2. We assume that the flowing fluid is slightly compressible. Writing Equation 4.33 with the appropriate coefficients, we have

(5.1) Equation 5.1 represents the characteristic equation for any block in Figure 1 . Pressure is not known in any of these five blocks, and so we must write Equation 5.1 for each of them (i.e., i = 1,..., 5):

Notice that although subscripts 0 and 6 (denoting "reflection" blocks 0 and 6, respectively) appear in the first and the last equations, they will be removed by the proper implementation of the imposed boundary conditions. This is done as follows:

At block #1, a no-flow boundary condition exists, implying that

or

Moving to block #5, the imposed constant pressure gradient implies that

or

These manipulations eliminate both unknowns.

and

, leaving us with five equations in five

In implementing the flow rate specified at the well block, we invoke the following definition of Q2:

Table 2a and Table 2b summarize the remaining coefficients and present them in matrix form.

Table 2a

Table 2b

The 5x5 coefficient matrix in Table 2 is a tri-diagonal coefficient matrix, which is normally written in a more compact form as follows:

When we solve this system of equations, we obtain the pressure distribution in the system. We then calculate the sandface pressure in the wellbore using Equation 4.20. The objective is to determine if the reservoir under study can support the imposed flow rate at the wellbore. Although the calculation of sandface pressure, Psf, may appear to be superfluous, it is actually an important parameter in production planning. For instance, the sandface pressure for the specified flow rate qs can become too low to support the friction losses in the production tubing and the associated peripheral wellbore devices; at this point a pump (e.g., downhole or sucker rod) may be needed to augment the flow energy. In fact, it is not impossible for the calculated sandface pressure to assume a negative value, which, of course, is unrealistic. In essence, what this tells the engineer is that the reservoir can no longer support the desired production rate.

Two-dimensional flow system

Figure 2 is a simple two-dimensional system partitioned into nine blocks.

Figure 2

The reservoir under consideration has three no-flow boundaries and one constantpressure boundary. A well is located in block (2,2), at which sandface pressure is

specified as Psf. Although there are nine blocks in the system, the constant pressure specification on the boundary blocks (1,1), (2,1), and (3,1) means that only six blocks have unknown pressures. Hence, we need to write Equation 4.33 for just these six blocks:

To implement the boundary conditions (similar to the one-dimensional case), we can write the following:

In implementing the inner boundary condition at the well block (2,2) where sandface pressure is specified, we need to invoke the proper equation for E and Q. Table 3a

Table 3a

and Table 3b summarize the resulting coefficient matrix, unknown vector and the right-hand side vector for the two-dimensional, slightly compressible fluid flow problem.

Table 3b

The matrix in Table 3 is a penta-diagonal coefficient matrix, and it is normally written in the more compact form shown below:

Solving the system of equations will yield the pressure distribution in the system. The final step to calculate of the resulting flow rate (as determined by the computed pressure distribution surface) using the equation qsc = -PI (Pi, j, k - Psf) where PI is the productivity index, Pi, j, k is the pressure of the block hosting the well and Psf is the sandface pressure. The sign convention is such that a positive q means injection and a negative q indicates production.

Three-dimensional flow system

In this third example, we illustrate logical continuity of this procedure using a threedimensional system. Figure 3 shows a three-dimensional reservoir with 18 blocks in which slightly compressible fluid is flowing.

Figure 3

The reservoir is volumetric (completely sealed), and there is a well in the center of the field, completed through only the top layer (2,2,2). Assume that a constant flow rate of qs is specified. We start with the flow equation in its finite difference form:

We then write this equation for every block at which the pressure is unknown, i.e.,

i=1, j=1, k=1 i=2, j=1, k=1 i=3, j=1, k=1 i=1, j=2, k=1

i=1, j=1, k=2 i=2, j=1, k=2 i=3, j=1, k=2 i=1, j=2, k=2

i=2, j=2, k=1 i=3, j=2, k=1 i=1, j=3, k=1 i=2, j=3, k=1 i=3, j=3, k=1

i=2, j=2, k=2 i=3, j=2, k=2 i=1, j=3, k=2 i=2, j=3, k=2 i=3, j=3, k=2

In this case, there would be a total of 18 equations. The long-hand notation used here is only for illustrative purposes; ordinarily these equations are automatically generated within the computer model. Again, in implementing the no-flow boundary conditions we should simply observe the following:

Zi, j, k Bi, 1 , k D1, j, k F3, j, k Hi, 3, k Si, j, 2

i = 1,2,3 and j = 1,2,3 i = 1,2,3 and k = 1,2 j = 1,2,3 and k = 1,2 j = 1,2,3 and k = 1,2 i = 1,2,3 and k = 1,2 i = 1,2,3 and j = 1,2,3

To implement the constant flow rate specification at the well block, we invoke the appropriate equation:

As clearly shown, a three-dimensional single-phase problem leads to a hepta-diagonal coefficient matrix whose general form is as follows:

When we solve the system of equations with the hepta-diagonal coefficient matrix structure, we obtain the reservoir pressure distribution. We use this pressure distribution, together with the flow rate specification at the wellbore, in calculating the flowing sandface pressure at the well.

Irregularly bounded reservoirs

The tri-, penta-, and hepta-diagonal matrix structures obtained for the three previous examples are quite standard. However, discretization of irregularly-bounded reservoirs may yield other matrix structures. Figure 4 illustrates a simple example.

Figure 4

For the sake of brevity, we will adopt a simple block numbering scheme, as shown in Figure 4 . The ten equations for this reservoir are as follows (cf. Equation 4.33): E1 P1 + F1 P2 + H1 P5 = Q1 D2P1 + E2P2 = Q2 E3P3 + F3P4 + H3P6 = Q3 D4P3 + E4P4 + F4P5 + H4P7 = Q4 B5P1 + D5P4 + E5P5 + H5P8 = Q5 B6P3 + E6P6 + F6P7 = Q6 B7P4 + D7P6 + E7P7 + F7P8 + H7P10 = Q7 B7P4 + D7P6 + E7P7 + F7P8 = Q7 D9P8 + E9P9 = Q9

B10 + E10P10 = Q10 The resulting 10x10 coefficient matrix is

Notice that the above coefficient matrix is hepta-diagonal instead of penta-diagonal, as would be normally expected from two-dimensional flow problems. The purpose of this example is show that some problems do not necessarily fit into the standard mode. However, there are methods for handling such special cases. Multi-Phase Flow Equations: Solution Methods Simulating multiphase fluid flow in porous media involves solving a system of coupled non-linear partial differential equations. Similar to the case of single-phase flow models, developing a computer model for these types of systems requires the use of finite-difference approximation to discretize these equations. The various solution techniques differ with respect to how we manipulate the governing partial differential equations. The following material summarizes the most prominent methods used for handling multiphase flow equations. IMPES method We can trace the origin of the Implicit Pressure, Explicit Saturation (IMPES) method back to the works of Sheldon, et. al. (1959), and Stone and Gardner (1961). The basic strategy of this method is to obtain a single equation in which the sole unknown is the pressure of one of the phases. We achieve this by combining the partial differential equations for each phase in such a way as to eliminate the saturation derivatives. Furthermore, we assume capillary pressure to be constant during any time step. By so doing, we obtain just one partial differential equation, with a phase pressure as the only unknown (this is usually the water-phase pressure). After writing the finite-difference approximation to this partial differential equation, we obtain the appropriate characteristic equation. When we apply this characteristic equation at the grid nodes, we generate a system of linear algebraic equations. The coefficients appearing in this system of equations are functions of the pressures and saturations; therefore, they are estimated using the information available at the previous iteration level. At any iteration level, when solution for the phase pressure (e.g., Pw) distribution is obtained, the next step is to solve explicitly for that phase

saturation distribution, Sw from the partial differential equation describing the flow of that phase. At this stage, we know the Pw and Sw distributions. This enables us to determine the oil phase pressure distribution using the capillary pressure relationship between the oil and water phases. Similar to the determination of Sw, after obtaining the Pw distribution, we explicitly solve the oil-phase partial differential equation for the oilphase saturation, So. With the values of So and Sw calculated, we can easily determine Sg (Sg = 1 - So Sw). Finally, using the capillary pressure relationship between the oil and gas phases, we obtain the gas phase pressure (Pg) distribution. This completes one iteration; we then repeat the whole procedure until we achieve convergence. At the beginning of each iteration, all the pressure and saturation dependent terms are updated using the most recent information available. Figure 1 is a flow chart highlighting the major steps involved in the IMPES method.

Figure 1

Simultaneous Solution (SS) method

In the simultaneous solution (SS) method, we convert the saturation derivatives that appear on the right-hand-side of the flow equations to pressure derivatives using the capillary pressure relationships. In this manner, we obtain three partial differential equations, with the principal unknowns being the phase pressures. In other words, at each grid node there are three unknowns. This does not pose any problem, however, because we have three equations to write at each node. With three equations for each node, the resulting coefficient matrix becomes significantly larger. For example, for a three-dimensional system with 10x15x3 blocks, the size of the coefficient matrix is 1350x1350. The overall structure of the coefficient matrix is similar to the corresponding single-phase structures. However, since more than one partial differential equation is approximated at each block, the resulting elements of the coefficient matrix are either 2x2 or 3x3 submatrices for two-phase and three-phase flow problems, respectively. Accordingly, for a twodimensional three-phase flow problem, we often refer to the resulting coefficient matrix as a tri-tri-pentadiagonal matrix typifying the aforementioned structure. Once we have constructed the coefficient matrix and obtained the phase pressures, using the capillary pressure relationship in an inverse manner, we obtain Sw and Sg. The calculation of So is straightforward when Sw and Sg are available. Figure 2 summarizes the basic steps of the simultaneous solution procedure.

Figure 2

Fully Implicit Solution with Newtonian iteration In the Fully Implicit Solution with Newtonian iteration method, we reduce the six principal unknowns of the three-phase flow equations to three linearly independent principal unknowns (most often one phase pressure and two saturations) by using the capillary pressure and saturation relationships. We then use finite-differences to approximate the three partial differential equations that result. By treating the coefficients implicitly (at the same level as the principal unknowns) we generate a system of non-linear algebraic equations. We can linearize these equations using the generalized Newton-Raphson procedure, such that we can implement a Newtonian iteration. In the solution process, a residual function is formed and its derivative calculated with respect to each principal unknown to construct the Jacobian. We then use a numerical differentiation scheme to obtain the elements of the Jacobian matrix. The salient points of the Newton-Raphson method are highlighted below. Consider a set of non-linear equations in two unknowns, x and y: f (x, y) = 0 g (x, y) = 0 with (xo,yo) as an initial guess to the solution. Suppose that (xo + x, yo + y) is the exact solution. Then, a Taylor series expansion can be written in the neighborhood of (xo,yo), i.e.,

Truncating the above series after the first-order terms, we obtain a system of linear equations in two unknowns, x and y.

The solution of these two equations for x and y leads to better estimates of the solution of the original non-linear equations. We repeat this process in an iterative manner until the improvements in x and y become small enough to satisfy the pre-set convergence criterion. We can express the iterative process in the matrix form as follows: which is followed by:

x(k+1) = x(k) + x(k+1) y(k+1) = y(k) + y(k+1) In the above matrix equation, the coefficient matrix is referred to as the Jacobian (J). The most stable algorithm is the Fully Implicit procedure. But at the same time, it is also the most computationally intensive and costly method. The IMPES procedure is generally the most cost-effective method (although we must qualify this by noting that its explicit treatment of the mobility and capillary terms causes it to suffer some stability limitations). The Simultaneous Solution method treats all primary variables implicitly; therefore it is unconditionally stable. However, explicit treatment of the coefficients is similar to the IMPES procedure and this does create some stability problems. The whole area of solution techniques for partial differential equations is a subject of intensive research in many fields. As a result, there are many techniques that are evolving through rigorous testing and validation processes. For example, adaptive techniques, which utilize different degrees of implicitness at different spatial and temporal locations to maximize the stability while minimizing the computational overhead, have received a great deal of attention. Solution of Matrix Equations The following material summarizes some direct and iterative procedures for numerically solving sets of linear algebraic equations. In setting the computer model, we face the problem of solving a set of n equations relating n unknowns, which are expressed in the form of a11x1 + a12x2 + a13x3 + ... +a1nxn = C1 a21x1 + a22x2 + a23x3 + ... +a2n xn = C1 (5.2)

an1x1 + an2x2 + an3x3 + ... +ann xn = Cn We can put the left-hand members of Equation 5.2 into a square array of the coefficients, known as the coefficient matrix,

(5.3) and the unknown vector,

(5.4) The right-hand members form the right-hand side vector

(5.5) We can express the set of equations summarized by Equation 5.2 in matrix notation as AX = C (5.6) In reservoir simulation problems, coefficient matrices are sparsely filled and contain a main diagonal and a number of additional diagonals. This relates to the way in which the grid blocks are ordered (i.e., the order in which the finite difference equations are written), the irregularity of the external boundaries and the number of dimensions. Ordering of grid blocks There is a one-to-one correspondence between the re-ordering of matrix elements and re-numbering of reservoir grid blocks. Therefore, if we know the coefficient matrix structure that results from a certain numbering scheme, we can employ the appropriate direct and/or iterative equation solver. Standard ordering by rows

Figure 1 shows standard ordering by rows for grid blocks in a 5x2 model, and the resulting coefficient matrix.

Figure 1

The non-zero elements of the coefficient matrix are indicated by x; positions, while zero elements are left blank. The coefficient matrix has a banded structure, and so is called a band matrix; its non-zero elements lie on parallel diagonals. The band width for the reservoir of Figure 1 is eleven (there are eleven parallel diagonals between the lowermost and uppermost diagonals, including the main diagonal). To calculate the band width, we multiply the maximum number of blocks in the direction they are ordered by two, and then add one; in this case, (5x2)+1=11.) Note that the upper and lower diagonals 3 and 4 have zero elements. Standard ordering by columns Figure 2 shows the standard ordering by columns of the reservoir from Figure 1 ,

Figure 2

and the resulting coefficient matrix with its non-zero elements. Here, the band width for the coefficient matrix is five (2x2+1=5). If the solution algorithm operates only with the elements located between the uppermost and lowermost diagonals, this ordering scheme will significantly reduce computational requirements. Checkerboard (Cyclic-2) Ordering Consider a reservoir model numbered in a checkerboard fashion, as shown in Figure 3.

Figure 3

In ordering the grid blocks, we number the shaded blocks first and then the unshaded blocks. The resulting coefficient matrix offers the advantage that, during the forward solution stage of Gaussian elimination (a common method for solving systems of linear algebraic equations), only a small portion of the matrix needs to be worked for triangularization Checkerboard (D-4) Ordering In checkerboard D-4 ordering, we number grid blocks along alternate diagonals. This alternate diagonal ordering leads to substantial savings in computational overhead (CPU time and storage). Figure 4 shows a typical coefficient matrix generated from D-4 ordering of a 4x5 matrix.

Figure 4

(D-2) Ordering The D-2 ordering scheme is also known as diagonal ordering. The resulting coefficient matrix has a relatively small bandwidth ( Figure 5 ).

Figure 5

Table 1 provides a summary of the storage and computational work requirements for different ordering schemes.

Table 1

A close inspection of Table 1 indicates that the greatest improvement in D-4 ordering is observed when J=I. Note: All non-zero elements appearing in the coefficient matrices are single elements for one-phase problems. However, they represent 2x2 and 3x3 sub-matrices for twophase and three-phase problems, respectively. Solution methods One of the more important steps of a numerical simulation study is solving the systems of equations generated by the finite-difference approximation. Our goal is always to end up with a system of linear equations, regardless of whether the original equations are linear or non-linear. Hence we need to linearize the non-linear equations before proceeding. In the broadest sense, there are two main categories of solution methods for the resulting system of linear algebraic equations: direct and iterative. Direct solution methods For a direct solution, we assume that the machine performing the computations is capable of carrying an infinite number of digits (i.e., there is no round-off error). In

this ideal situation, a direct solution method will yield an exact solution after a finite number of elementary arithmetical operations. In reality, every computer carries a finite number of digits; round-off errors do occur, resulting in non-exact solutions. As the number of operations increases, so does the cumulative round-off error. For very large systems of equations, it is possible for the round-off errors to grow uncontrollably to the point of generating unrealistic results and even instability. Still, there are situations in which a direct solution offers distinct advantages. Furthermore, a well-written direct-solver subroutine subprogram can be used on a variety of problems. This feature is particularly attractive, because it lets us test various parts of the code development without having to worry about the efficacy of the solver. The Gaussian elimination technique is the fundamental algorithm used by direct solvers. As shown in Table 2a ,

Table 2a

Table 2b ,

Table 2b

and Table 2c his technique involves a systematic set of elementary row operations that essentially transform the coefficient matrix into an upper triangular matrix.

Table 2c

This step of the solution process is known as the forward solution. Once we obtain the upper triangular matrix, the backward solution comes into effect, whereby the last unknown is directly solved for and the others are obtained by systematic backward substitution. Matrices associated with the finite difference methods are often sparse; therefore, they lend themselves more easily to iterative solutions. Remember that in computer operations, elementary arithmetic operations involving zeros take the same amount of time. Hence the sparseness of the matrix does not affect the time involved in using direct solvers. However, there are direct solution algorithms which take advantage of the structure of the coefficient matrix involved in finite-difference techniques. One of the most popular techniques is Thomas Algorithm, which is designed for tri-diagonal coefficient matrices. With this algorithm, we simply avoid performing any arithmetic operations on the zero elements of the coefficient matrix, thereby saving a significant amount of time. Table 3 summarizes this algorithm.

Table 3

Although Thomas algorithm is designed for tri-diagonal coefficient matrices, there are other direct solvers which utilize the presence of other band structures. Again, the idea is to exploit the sparse nature of the matrix by not carrying out any operations on the zeros that are outside the band. The band matrix solvers can be effectively employed to coefficient matrices with bandwidths up to 15. As the bandwidth becomes larger than 15, the efficacy of the band matrix algorithm starts to decrease because this group of algorithms treats all zeros which are located within the band as non-zeros. Iterative methods Iterative solution methods involve making an initial guess, or approximation of the solution vector, and then improving on this guess by successfully implementing the algorithm. In other words, successive applications of the algorithm lead to better approximations. This systematic repetitive application of the algorithm continues until the solution vectors generated between two successive iterations are in agreement within a specified tolerance (assuming that algorithm generates solutions that approach the correct answer by each convergent iteration).

Iterative methods offer two important advantages: less storage requirements and less computational work. The simplest and best-known iterative methods for solving systems of linear algebraic equations are grouped under the name fixed point iterative methods. To illustrate, consider the following system of equations: a11x1 + a12x2 + ......... +a1nxn = C1 a21x1 + a22x2 + ......... +a2n xn = C1 (5.11) .................................................... an1x1 + an2x2 + ......... +ann xn = Cn In the fixed-point iterative techniques, we rearrange the first equation to solve for x1 in terms of the other variables; we similarly rearrange the second equation to solve for x2, and then the last equation to solve for xn. We begin the iteration process by making guess-values for all the variables; each variable is then updated according to a certain procedure which differs from one method to another. Table 4 summarizes the algorithmic equations for the Point-Jacobi, Gauss-Seidel and Pointwise Successive Over-Relaxation methods as they apply to the solution of Equation 5.11.

Table 4

The algorithms in Table 4 are presented in the order of increasing convergence rates. In other words, the Point-Jacobi method is the slowest, while the Point-Successive Over-Relaxation method is the fastest. The convergence rates are easy to explain: since the Point Gauss-Seidel method uses the most current values as soon as they become available, whereas the Point-Jacobi method relies on the old values until the current iteration level computations are carried out on all the unknowns. The Point Successive Over-Relaxation method is essentially based on the Gauss-Seidel method. If we set the value of opt to one, Equation 5.14 reduces to Equation 5.13. The basic strategy of Equation 5.14 is to magnify the successive improvements achieved by each converging Gauss-Seidel iteration through the use of a common multiplier, opt. We refer to this common multiplier as the optimum acceleration parameter, whose value is problem-dependent and always lies between 1 and 2. Figure 6 shows a typical variation of the number of converging iterations with values of .

Figure 6

opt is that value of which yields the minimum number of iterations . We perform the same convergence check on all these three algorithms. The basic convergence criterion is indicated by Equation 5.22. The convergence tolerance, , is a pre-selected small number (e.g., for pressures 10-1 and for saturations 10-4). All of the variables must satisfy this convergence criterion. The use of a small tolerance does not always guarantee a solution that is sufficiently close to the actual solution. We must be careful to apply material balance checks in order to ensure that the converged solution preserves material balance. If the material balance check signals an unacceptably inaccurate solution, we should tighten the tolerance and continue iterations. Figure 7 shows a comparison of convergence paths for the three algorithms discussed.

Figure 7

Figure 7 reveals two important characteristics. First the solution always uniformly approaches the true solution from one side (no oscillations around the true solution). Secondly, we see that for the same problem Gauss-Seidel converges twice as fast as the Point-Jacobi method, whereas Point Successive Over-Relaxation converges twice as fast as the Gauss-Seidel method. One note of caution: if the Jacobi method does not converge on a given problem, neither will the other two. We can considerably enhance the efficiency of point-iterative methods by grouping a number of unknowns together and solving for them simultaneously at a given iteration level. One easy way of grouping is to group all the points located along the same line. This approach is called Line Successive Over-Relaxation. Once we solve the values of the unknowns for the particular line at a given iteration level, we can immediately apply them when the algorithm is executed on the next line of points. Once we are on the new line, the points of the next line are still at the old iteration level. Sometimes we can group more than one line of points together (even possibly all the points on a plane of a three-dimensional system) to form a block of points. We then solve for all the unknowns on these points simultaneously; this strategy is referred to as Block Successive Over-Relaxation. The complexity of the problem dictates the methods we should use; as the problem complexity increases in terms of the number of unknowns and the relationships among these unknowns, we must use more powerful iterative methods. In this class of methods one can include alternating direction implicit procedure, strongly implicit procedure, conjugate gradient method, and nested factorization method. Comparison of direct and iterative methods

Although the efficacy and the efficiency of direct and iterative methods are problemdependent, there are some general features which distinguish them from one another. Table 5 summarizes these features.

Table 5

Internal Checks We must always ensure that the numbers coming out of the computer are reasonable, realistic and descriptive of the problem at hand. There are three basic checks we should always perform at the end of each time step computation. These are the residual check, incremental material balance check and cumulative material balance check. In the residual check, we simply substitute the solution back into the original set of equations and observe the departure from the equation by calculating the difference between the left-hand side and the right-hand-side of the equations. We do this check at the grid block level. Any grid blocks exhibiting unexpectedly large residuals need to be scrutinized and diagnosed for the underlying reasons. The incremental material balance check simply questions whether or not mass is being conserved or not during a time step. Before we perform this check, we should pre-determine the acceptable material imbalance tolerance. Once we obtain the solution for a given time step, we calculate the reservoirs mass content using the obtained pressure and saturation values. We then compare this mass content against the mass content calculated at the end of the previous time step. The difference between these two should equal the mass produced and/or introduced into the reservoir. An unacceptable material imbalance may indicate a loose tolerance level imposed on the solution algorithm. We can correct this problem by tightening the tolerance used on equation solvers. The cumulative material balance check is similar to the incremental material balance check. This time, however, we apply conservation of mass at the end of each time step against the original mass content of the reservoir at the initial time. This check

informs us as to the extent of the computational errors accumulated so far into the run. In this way, when we report the results of the simulation study, we will be able to comment on the level of confidence. While not all of us will be involved in actually developing simulators, we still must have a general understanding of what goes into their development. Furthermore, we must be quick to recognize "red flags" so that we can exercise prudent engineering judgement. In this way, we will be able to ask the right questions for assessing the validity of the numbers being generated by the computer.

Constructing the Reservoir Model: Data Collection and Preparation Constructing a reservoir model involves physically defining the reservoir and its fluids in terms of a large set of data. These data define the fluid properties, the reservoir boundaries and the interactions with the environment. Once we put together such a physical description of the model, we use a numerical code to study its performance in the form of response to external stimuli. Data collection There is no overemphasizing the importance of data gathering in a simulation study, particularly as it relates to data quality. An adequate and appropriate reservoir description hinges delicately on the quality of the data we use. It is imperative, when reporting reservoir performance, that we measure the actual system performance. We can only be assured that were doing this if we have used the correct data set to construct the reservoir model. A close examination of the governing equation will reveal the type and extent of the data required for building a reservoir model in readiness for a simulation study. For the sake of completeness, we consider data requirements for the three-phase, gasoil-water system. Summarized below are the various data needed to build a reservoir model. Geological Information Accurate geological information is paramount to a successful simulation study. This realization has led to the development of a new area of research called reservoir characterization, which encompasses a number of disciplines. Accurate reservoir description, even in the crudest sense, requires a team effort between geologists, geophysicists, log analysts and engineers. Geologists study the depositional environment, the syn-depositional and post-depositional forces responsible for the productive formation. Geophysical studies define the stratigraphy of the environment as well as the reservoir structure. A careful logging program provides information about porosity and saturation distribution. This, coupled with core analysis, produces credible data for constructing isoporosity maps. Table 1 gives an overall picture of geological data requirements, data sources and usages.

Table 1

Rock Properties After procuring the geological information, which basically defines the reservoir boundaries and structure, we need to obtain information about the reservoirs internal structure, as typified by its fluid storage capacity, fluid distribution and ability to transmit these fluids. That is, we need to determine its porosity, fluid saturation, and permeability, respectively. As far as the reservoir simulation is concerned, these are the key macroscopic properties that control the flow. The fundamental microscopic properties that affect fluid transport in porous media, such as pore size and pore size distribution, pore throats, tortuosity, wetting affinity, mineralogy, etc., are all embedded in these key macroscopic properties. This is why these properties, even though they have the controlling effect, are not explicitly represented in the flow equations. Table 2 provides an overall summary of the rock properties used in reservoir simulation.

Table 2

Fluid Properties Understanding reservoir fluid properties is essential to formulating their overall behavior. Apart from the fact that many of these properties appear in the coefficients of the governing equations, they are also used to compute reserves in surface units. Generally, these properties are functions of the basic thermodynamic variables of temperature, pressure and composition. There are two broad categories of simulators: black oil simulators and compositional simulators. While black oil simulators use the local average properties of the fluids for characterization, compositional models rely on the accurate composition of the fluids. Similarly, while most simulators assume isothermal reservoir conditions, there are certain situations (e.g., thermal recovery operations) which necessitate the use of energy balance equations and consideration of the fluid properties. In a typical black oil simulator, we express fluid properties as functions of pressure, to different degrees of nonlinearity. These nonlinearities make the governing equations even more nonlinear. Depending upon the degree of nonlinearity, we may need to update these properties at each iteration level in order to preserve the form of the original partial differential equation. This requires us to update the transmissibility terms at each iteration level, thus having a tremendous implication on the computational overhead required--particularly if direct solvers are used. This is because at every iteration level, a new coefficient matrix is generated as properties are updated. Therefore, triangularization of the coefficient matrix has to be performed continuously throughout the iteration process. One strategy for alleviating this problem is to simply keep the properties fixed for a pre-set number of iterations

before updating them during a time step. A heuristic approach is necessary to assess that this strategy will not adversely impact the adequacy of the solution. Table 3 summarizes the relevant fluid properties.

Table 3

Rock-Fluid and Fluid-Fluid Interaction Properties It is normal to expect that interactions between the fluid and the reservoir rock will affect the systems dynamic behavior. In the presence of multiple fluids, the dynamic behavior of the individual fluid phases will be likewise affected. We quantify these effects in terms of interaction properties, namely relative permeability, interfacial tension, wettability and capillary pressure. Some of these interaction properties exhibit their influence on a microscopic scale, and so do not appear explicitly in the governing equations. They manifest themselves, however, in other macroscopic properties such as relative permeability. Although very commonly used and taken for granted, relative permeability is not a fundamental property, but rather a global property, in which microscopic properties, (interfacial tension and wettability) are implicitly embedded. It is currently the most practical way of distributing these microscopic parameters over the entire reservoir domain.

We always express relative permeability as a function of fluid saturation, and this function is strongly nonlinear. Therefore, it introduces the same degree of nonlinearity into the transmissibility terms. The importance of this becomes even more accentuated in a displacement process, where the displacing front is being propagated as a sharp discontinuity. In this type of situation, we can often achieve a better numerical description of the problem by using a numerical weighting scheme where we assign a greater weight to the points upstream than to those downstream. The classical examples of this treatment are single-point and two-point upstream weighting procedures for relative permeabilities. In actuality, these procedures are analogous to expanding the relative permeability functions in terms of Taylor series. Production and Pressure History Advances in seismic and other geophysical techniques notwithstanding, the wellbore remains the only true window through which we can glimpse what is happening in the reservoir. We do this primarily by keeping account of production and pressure data and using this database to refine the models descriptive and predictive capabilities. Pressure data from well test analyses are useful in monitoring the reservoirs response to certain impulses. Such data may be used to infer reservoir properties. Wellhead pressures taken over a period of time can also be used for this purpose. Reservoir production history, both in terms of rate and cumulative flow, is essential not only in validating the simulators predictive capability, but also for tuning the simulator for future predictions. As a rule of thumb, we can achieve a well-tuned simulator after the reservoir has produced about 10 percent of its ultimate recovery. With this level of tuning, we can develop the confidence to use the tuned simulator in a predictive mode. Production data include not only the production rates of individual phases, but also the gas-oil ratio (GOR) and the water-oil ratio (WOR). Data preparation Now that we are familiar with the types of data we need for a simulation run, we can focus on how to put this data into usable formats. Current technology in the areas of image digitization and graphics visualization has considerably enhanced several aspects of this process, making it easy to translate maps and graphical representations to numerical values. In making these translations, we create digitized databases. For these numerical values to be useful, we must sort them out in a proper format. At this point, we should scrutinize the data for consistency and conformity. If we find an anomaly in some of the data which we cannot substantiate, we should reject these data. Figure 1 shows a set of hypothetical data for a variety of property variations, with typical functions superimposed on each.

Figure 1

It is unusual for measured data to lie on a smooth continuous curve; data scattering is almost always present. We may have to smooth the data using statistical techniques such as least squares analysis, analysis of variation and regression. All of these techniques are a part of preparing the acquired raw data for a reservoir simulation study. After we have acquired, analyzed and smoothed the data, we must put them into the form of a simulator input. There are two basic methods of doing this. We can have data input already expressed in terms of analytical functions, or in tabular form. In obtaining an analytical equation, two approaches are commonly taken: approximating functions by least square analysis and interpolating functions such as cubic splines. These functions are coded as function subprograms, which have to be called each time the property value is needed. Since this involves a great deal of data transfer between the different segments of the code, some measure of inefficiency in terms of time is expected.

 A tabular data form, usually referred to as the table look-up method, makes the data transparent to the subroutine that needs it. This can save a substantial amount of time. In a typical table look-up routine, we need to bracket the data points and then use a simple interpolation method (often linear) to obtain the data of interest. As Figure 1 demonstrates, some properties lend themselves to being represented with analytical equations, while others are not so readily conducive. Properties that fall in the latter category are invariably obtained by the table look-up method.

Reservoir Simulation and the Use of Pseudo Functions

Selecting the number of dimensions (1D, 2D, or 3D) to adequately represent a reservoir simulation problem calls for prudent engineering judgement. We need to balance the amount of additional information that an extra dimension could provide (assuming the necessary data are available) against the increased computational overhead it would require. In general, we should always try to use the least number of dimensions that provide sufficient information for engineering analysis. For some problems, however, we need greater detail, and so we must use higher dimensions. Methods have been developed over the years to try to get the best of the two worlds--that is, to use a smaller number of dimensions and still attain a high degree of detail. One prominent method involves the use of pseudo functions. The concept of pseudo functions dates back to the early 1960s. It developed in response to the need for accommodating a third dimension within the limits of the computers then available. Of course, we could argue that these limitations have more or less disappeared. Still, three-dimensional computations are expensive, not only in reservoir simulation but in other process simulations as well. This is true both in terms of computational overhead and the shear amount of data required. It is not always easy to procure this type of data, particularly in reservoir simulation. As a result, there remains even today a strong interest in using pseudo-functions to reduce problem dimensionality and other numerical difficulties (Zagalai and Murphy, 1991). Pseudo functions are broadly divided into two categories, based on where they are used in the reservoir simulation: interblock pseudo functions and well pseudo functions.

Interblock pseudo functions

The purpose of interblock pseudo functions is to describe flow between grid blocks. Since interblock flow is controlled by relative permeability functions, which in turn depend on saturation, generating interblock pseudo functions involves averaging the saturations in the blocks of interest. There are two basic kinds of interblock pseudo functions: analytical and dynamic. Analytical pseudo functions involve volumetric averaging of water saturation over the thickness of the interval. We then assign this volume-averaged saturation as the saturation for the grid block. For the case of segregated flow, we neglect capillary forces. This implies a complete gravity segregation. Thus, in calculating pseudo

relative permeabilities, we simply need to use the equality of the horizontal flow calculated between the blocks using the pseudo-relative permeability (computed at the volume-averaged saturation) and the actual relative permeability (computed at the actual water saturation in the uninvaded zone). Similarly, the use of appropriate pseudo capillary pressure functions either partially or totally eliminates the need for vertical gridding as long as such functions maintain the integrity of the initial fluid distribution, fluid movement and pressure distribution. Another concept, which is more widely used in analytical pseudo functions, is that of vertical equilibrium . Here, we assume that the potential gradient of all the phases in the vertical direction is zero. As depletion takes place, gravity and capillary forces attain equilibrium in the vertical direction in every grid block of the reservoir model. This implies that the differences between the phase pressures are totally balanced by capillary pressure, thereby fixing the vertical saturation distribution. In effect, this reduces a three-dimensional problem to a two-dimensional areal problem, and a two-dimensional cross-sectional problem to a one-dimensional problem. While the vertical equilibrium pseudo function has some inherent error in describing the vertical dimension, it is quite adequate for many reservoir problems. It is imperative to pre-test the applicability of any set of pseudo functions by comparing results obtained from them with models describing detailed vertical segmentation using actual data. While the basic assumption of vertical equilibrium is reasonable for many reservoir problems, it is not valid in reservoirs with poor vertical communication. In such cases, the time it takes for any perturbation to dissipate in the vertical direction compared to a mean residence time for any fluid particle moving in the lateral direction causes significant vertical potential gradients, thus violating the premise of vertical equilibrium. One suggested procedure is to assume a set of vertical equilibria in time; but even this approach breaks down in many cases. The approach for generating the dynamic pseudo functions follows essentially the same procedure as in the case of analytical pseudo functions, except that areal weighting is used crosssectionally. This is done by partitioning each block into finer grid blocks. The end result is that a pseudo function is generated for each vertical column of blocks.

Well pseudo functions

Well pseudo functions attempt to extend the concept of interblock pseudo functions to describe multiphase flow into the wellbore from adjacent blocks. This entails extending the concept of pseudo relative permeability functions to the vertical performance of individual wells. There are certain reservoir problems in which this application becomes extremely useful. Practical examples include water coning and gas cusping, in which multiphase flow in the near-wellbore region plays a dominant role and usually requires more details than usual. Ordinarily, a three-dimensional model is most appropriate for such problems, but the use of well pseudo functions can serve as a good compromise. The most direct way to handle these problems is to use source representation, which couples the wellbore with a numerical description in the near-wellbore region. Pseudo-relative permeability and pseudo-capillary pressure functions have been found to be satisfactory. Often, the pseudo functions can be expressed in a field-scale numerical

model in terms of analytical solutions, correlations based on numerical models and pseudo functions.

Special Purpose Reservoir Simulators

Black oil models encompass as much as 80% of all reservoir simulation applications. But there may be times when we find ourselves dealing with totally different physical environments--environments that (in terms of processes or the type of porous media) are not amenable to black-oil modeling. For example, we cannot adequately represent a coalbed methane reservoir with a conventional gas-water simulator, even though it is a gas-water (two-phase) system. This is because the conventional model does not account for adsorption/desorption phenomena. Neither can a black oil model track the processes that govern miscible flooding in volatile oil and condensate reservoirs, because pressure and temperature changes have a tremendous impact on the phase behavior involved and hence the fluid properties. As we encounter increasingly complex reservoir problems, our data requirements become more stringent, and the numerical solution schemes we employ become more sophisticated. Furthermore, we have to keep track of significantly more variables than usual and interpret their trends. In situations where the governing physical and chemical principles differ from what we see in black-oil reservoirs, we may resort to special purpose simulators. The question often arises as to whether we could develop a general purpose simulator that is capable of handling all the arrays of reservoir problems. Theoretically (considering that every simple problem is a merely a subset of a more complex one) the answer is a resounding yes! In a practical sense, however, such an all-encompassing model is a virtual impossibility; CPU and storage requirements, along with compilation overhead, would be prohibitive, and in most cases, redundant for the problem at hand. This is why special purpose simulators are a rule rather than an exception.

Water Coning Simulators

Water coning, a premature invasion of the wellbore by water, usually occurs in bottom-drive reservoirs. Basically, the oil-water contact, which is normally horizontal, becomes distorted near the wellbore, assuming an upward concave posture. Coning usually occurs if the well is completed close to the oil-water contact and produced at a high rate, creating a sharp pressure gradient near the wellbore and resulting in excessive water production. This water invades the pore spaces around the perforations inhibits the flow of oil towards the wellbore. A similar phenomenon often takes place near the gas-oil contact where the gas cap becomes distorted as the gas cusps downward and causes excessive gas production. Figure 1 illustrates these two phenomena.

Figure 1

When using a reservoir simulator to study water coning or gas cusping, we must take extra care to capture the rapid changes that take place within the immediate vicinity of the wellbore. To enhance the description of these near-wellbore changes, we resort to single-well modeling most of the time. A radial-cylindrical grid ( Figure 2 ), with smaller grid spacing along the radial-direction and dense spacing along the vertical direction, captures the movement of the saturation front.

Figure 2

As shown in Figure 2 , a coning study obtains the formation of the cone as axisymmetric with respect to the wellbore. Therefore, it is common to study the problem in two-dimensional radial-cylindrical coordinates (r-z directions only) to decrease the storage requirements and increase the computational speed. This class of simulators is always demanding to develop and to run. The main difficulty arises from the convergent nature of the flow toward the wellbore. As fluid flows toward the wellbore, the cross-sectional area perpendicular to the flow direction becomes progressively smaller, resulting in higher velocities near the well. This necessitates the use of the smaller time steps to ensure numerical stability While the data requirement for a coning study is the same as for a full field model, the grid system is significantly different. Sometimes, we superimpose rectangular geometry with a hybrid grid system over the computational domain (part a of Figure 3 ).

Figure 3

In some cases, if simulators can handle locally refined grids, then we can use local refinement techniques effectively. We can place these locally refined sections around the wellbore (part b of Figure 3 ), or they can be dynamic such that they move together with the oil-water contact (part c of Figure 3 ). Note that not every simulator has these options, and that implementing these options to existing models can sometimes be extremely demanding.

Dual Porosity/Permeability Simulators

With more and more reservoirs being identified as naturally fractured, there is a growing interest in simulating such systems. We generally characterize naturally fractured reservoirs as dual porosity systems because of the presence of two continua: the rock matrix and the fractures. We usually represent these continua using two different sets of porosity and permeability functions, referring to the matrix porosity as primary and the fracture porosity as secondary. Fluids are stored mainly in the rock matrix, and transmitted through the fracture network; the permeability of the fracture network is much higher than that of the matrix. Porosity and permeability are not the only discontinuous functions in a dual-porosity system.

Other properties, such as the capillary pressure and relative permeability functions, also differ as we move from one subdomain of flow to the other. There are two general approaches to simulating dual-porosity systems: dualporosity/single-permeability and dual-porosity/dual-permeability.

In the dual-porosity/single-permeability approach, we write flow equations for only one of the two continua, i.e., the fracture network. We account for matrix flow by imposing a source term on the flow equations for the fracture continuum. We calculate this mass transfer term between the matrix blocks and the fractures by using a transfer function, which is based on either unsteady-state or pseudosteady-state computations. In the dual-porosity/dual-permeability approach, we write flow
equations for both continua. We subdivide matrix blocks into grid blocks, and solve flow equations within each of these subdomains (matrix blocks). Again, we calculate the transfer of fluid between the matrix and the fracture using the existing pressure gradients at the interface. In practice, natural fractures occur randomly in the porous medium (part a of Figure 4 ).

Figure 4

However, in simulating the fracture system, we idealize the system as shown in part b of Figure 4 . Usually, a computational block consists of several matrix blocks. For instance, in part b, each computational block is made up of eight matrix blocks. Although the matrix blocks in the idealized dual-porosity system (part b of Figure 4 ) are shown as cubes, we could have shown them as spheres, cylinders, or slab elements. Studies have shown that there is no significant disparity between results obtained by using any of these geometrical configurations. The main difference usually results from the handling of the matrix/fracture fluid exchange and the types

of the capillary pressure and relative permeability functions used within the fracture network.

Thermal Recovery Simulators

Thermal recovery processes are designed to raise the temperature of reservoir oil, thereby decreasing its viscosity and enhancing its flow characteristics. The primary differences among various thermal recovery methods are in the heat sources used to raise oil temperature. The two most popular methods are steamflooding and in-situ combustion. Of all enhanced oil recovery (EOR) processes, steamflooding is among the most successful. The major difference between thermal recovery simulators and other types of models is the need for the energy balance equation. Temperature distribution is the main driving factor in thermal recovery, and so must be adequately predicted--especially since viscosity is a strong function of temperature. The energy equations are usually highly nonlinear and strongly related to the mass balance equations. In addition, two types of heat transfer mechanisms are accounted for: conduction and convection. The behavior of the energy equation depends on which of the two mechanisms dominates. If conduction is dominant, the equation exhibits parabolic behavior (diffusivity equation) but if convection dominates, it exhibits hyperbolic behavior (shock wave equation). This translates to sharp thermal fronts, which leads to a numerical smearing of the front. In this case, we must pay special attention to the numerical scheme in order to avoid numerical instability. One other feature of thermal simulators is the need to calculate the heat loss to the surrounding formations. Whereas in formulating fluid flow equations, we consider the reservoir an isolated system with no fluid exchange with the surrounding, thermal equations allow heat exchange with the surroundings, at least by conduction. This is more critical for the overburden and underburden because of the large contact area relative to the adjacent formations in the lateral directions. Steamflooding enhances oil recovery by transferring heat from the steam to increase the temperature of the reservoir section adjacent to the wellbore. This temperature increase reduces the oil viscosity and hence the flow resistance in the wellbore vicinity. There are basically two types of steam injection schemes: cyclic, where injection and production are done sequentially using the same well, and continuous, which involves separate injectors and producers. The primary effects being simulated in steam injection are the temperature increase and the resultant decrease in the oil viscosity. There are two basic types of steamflooding simulators: compositional and non-compositional. Non-compositional simulators are simpler and usually adequate for most cases. But in cases where we suspect that the distillation of light components significantly affects the stimulation process, we should use a compositional steamflood simulator. Such a simulator requires a phase behavior description for the oil/steam system. An additional complication arises when three-phase relative permeability data are considered temperature-dependent.

The main computational challenges arise from the strong inherent nonlinearities in the energy equations, the discontinuity resulting from phase changes (condensation, vaporization), and a strong coupling between the fluid movement and the energy transfer. These peculiar characteristics manifest themselves in numerical problems such as instability and grid orientation effects. In-situ combustion uses basically the same principle as steam injection, i.e., using heat to reduce viscosity of oil. In this case, however, the heat comes from injecting air into the reservoir and igniting part of the oil to start a combustion front. The primary consideration in simulating in-situ combustion, apart from all the other factors mentioned in the discussion of steam flooding, is the combustion reaction kinetics. The temperature dependence for the kinetics equations is usually described by Arrhenius type rate expressions. These strongly temperature-dependent functions introduce a new level of nonlinearity to the energy equations. We must be aware of possible severe stability problems in in-situ combustion simulators. Excessive computational time is not at all uncommon, since it is not unusual for a typical in-situ combustion run to be two orders of magnitude larger than a typical black oil problem applied to the same reservoir. In thermal simulators, apart from mass balance checks, we must also ascertain energy balance as part of the internal checks.

Compositional Simulators
Compositional reservoir simulators account for multiphase flow and interfacial mass transfer of each component in a hydrocarbon system. This implies that at any given time, the simulator tracks fluid movement and establishes the state of equilibrium of the reservoir fluids at the discrete points. At each node, phase pressure, phase saturation and overall composition are computed as a function of time. Compositional simulators are particularly useful in describing gas condensate reservoirs, volatile oil reservoirs, gas cycling processes and in some thermal recovery processes, in which compositional changes are important. The distinguishing feature of a compositional simulator is its full coupling of the fluid phase behavior model with the flow equations and perhaps the energy equation. In this case, rigorous flash calculations are performed using either tabular equilibrium ratios or analytical equations of state. All the volumetric and thermal properties are computed as being composition-dependent. There are certain limitations inherent in using compositional reservoir simulators. The two principal ones are the excessive CPU time requirements and the problem of adequately describing the fluid phase behavior.

Generally, CPU time requirements increase exponentially as the number of fluid components increases. We can alleviate this problem by lumping the system into a few pseudo-components. This requires rigorous testing to ensure that the pseudo-system mimics the original systems phase behavior within the range of temperature and pressure being simulated.

 In compositional simulation, we generally treat petroleum as a mixture of limited discrete components. But in fact, it is a continuous mixture. The standard method of handling this problem is the use of C7+. However, difficulties often arise from assigning the needed parameters to this pseudo-component.
One of the methods of avoiding these problems is to apply the concept of continuous thermodynamics. This area is still undergoing research, however, and its practical use is currently limited.

Miscible Displacement Simulators

A miscible displacement process involves two or more fluids that are mutually miscible when they come into contact in all proportions. When complete miscibility takes place, there is no interface formed between components. The miscibility between two components can take place in two different ways: first-contact miscibility and multiple-contact miscibility. In the first case, the displacing fluid is immediately miscible with the displaced fluid, while in the second, miscibility occurs after a series of equilibrium contact stages. Examples of miscible displacement processes include chemical flooding (e.g., miscible carbon dioxide injection, polymer flooding, micellar flooding) and displacement of oil by solvents. Miscible displacement simulators can be multicomponent and multi-mechanistic models. Multi-mechanism indicates that flow is taking place due to convection and dispersion. We can represent multi-mechanistic flow by describing the velocity of a component as the sum of the velocities due to different mechanisms. In a miscible displacement simulator, we generally assume single-phase flow. This assumption implies the presence of full miscibility and lets us avoid difficult vaporliquid equilibrium computations of multiple-contact miscibility. Furthermore, we usually consider two components (possibly oil and the solvent), and emphasize flow by dispersion. in the construction of these simulators. We calculate phase properties such as viscosity and density using mixing rules. This type of formulation assumes no volume change upon mixing.

Chemical and Polymer Flooding Simulators

Chemical flooding simulators are much more demanding to develop than other special purpose simulators. This is simply because the physics involved in a typical chemical flood are much more complex, and require consideration of the extensive microscopic phenomena that are taking place at the fluid-fluid and fluid-rock interfaces. Chemical floods employ several different fluids, and therefore form several fluid banks. Interfacial phenomena, phase behavior of the complex systems, adsorption and desorption of certain chemical agents to and from the rock grains make the problem even more complicated. Most chemical flood simulators are developed to study certain phenomena in the laboratory, where it is much easier to control process variables.

Polymer injection is a complex process. It involves simultaneous multiphase fluid flow with interphase mass transfer of water between the polymer and water phases when there is adsorption of polymer on the rock of the porous medium. In a polymer injection simulator, the polymer and water phases form the aqueous phase. Usually water is allowed to transfer from the polymer phase to the water phase as a result of polymer slug deterioration caused by polymer adsorption on rock. However, most of the time, water is not allowed to transfer from the water phase to the polymer phase (i.e., the polymer slug cannot be diluted by water). It is necessary to consider the polymer adsorption on the rock and its effects on the permeability of the rock to the existing phases. We also need to account for transverse dispersion of the polymer component within the aqueous phase. Most polymer injection simulators treat the adsorption of polymer on the rock as permanent (i.e., there is no desorption). However, the rock adsorptive capacity limits the maximum polymer adsorption on the rock. We can use chemical and polymer flood simulators, like other simulators, as screening tools to select optimal patterns. We can also use them to determine optimal slug sizes, and to analyze the increased production and profitability of a full scale chemical flood under several operating strategies. Finally, we can use them to predict the effect of fluid and rock properties on the oil recovery and flood performance.

Coalbed Reservoir Simulators

The petroleum industry classifies methane from coalbed reservoirs as unconventional natural gas. Unconventional resources offer significant potential, both now and in the future, in terms of large volumes of reserves, low production costs, and relatively simple development techniques. One of the more important characteristics of coal seams is their dual-porosity nature, characterized by well-defined macropore and micropore structures. The natural fracture networks of coal seams are uniformly distributed, and composed of two fracture systems which are almost orthogonal to each other (face and butt cleats). While the face cleat is continuous throughout the reservoir and capable of draining large areas, the butt cleat is discontinuous and ends at the face cleat. Thus, the anisotropic nature of a coal seam originates from this cleat system in which the permeability in the direction of a face cleat is considerably larger than that in the direction of a butt cleat. The micropore system, as a primary porosity matrix, has a size in the order of molecular dimensions to a few Angstroms. In general, we assume that the openings are not accessible to water, and that they contribute the major portion of gas storage in areas in which gas is stored in adsorbed and free states. Quantitatively, as much as 2000 SCF of methane can be stored in a ton of coal by adsorption. When the coal seam is in virgin conditions, the volume of the free gas in the micropores is almost negligible compared to the volume of the gas in the adsorbed state. Again, for coal seams in virgin conditions,we assume that the cleat system is fully saturated with water. As water is removed from the macropores, gas is desorbed from the micropore surfaces toward the macropore structure. This process represents a distributed source mechanism over the macropore structure. As the desorption process continues, the free gas saturation within the fracture network increases.

Most coalbed simulators use Langmuir sorption isotherms to describe the release of methane from the adsorbed state. They achieve this by solving a first order kinetic sorption model. Along these lines, two approaches have been developed: equilibrium sorption isotherms (pressure dependent) and non-equilibrium sorption isotherms (pressure and time dependent).

In the equilibrium sorption isotherm approach, we assume that the gas adsorbed onto the micropore walls is in a constant state of equilibrium with the free gas phase in the macropore system. Models based on this approach are essentially single-porosity models altered for coal seams by either the inclusion of a pressure dependent source term or by the modification of the storage term. These models generally predict optimistic results, since the adsorbed gas is assumed to instantaneously enter the macro-pore system. A slightly more sophisticated approach involves treating matrix sorption as a quasi-steady state model, in which the desorption rate is proportional to the difference between the gas concentration at the external matrix surface and the average concentration contained within the matrix. A more realistic approach is the non-equilibrium formulation. The
non-equilibrium (unsteady-state) formulation takes into account the time lag experienced during transport through the micropores. The overall structure of coalbed reservoir simulators is similar to the conventional dualporosity/single-permeability models. With the aid of these models, production performances of coalbed reservoirs have been thoroughly examined. If the simulator has the options to accommodate ongoing mining activities, it can be used to predict the methane emission rates into the active mine working area. The first-generation coalbed reservoir models consider the only existence of a single component gas (methane) and water. In some coalbed reservoirs, components other than methane, (e.g., such as carbon dioxide), may play an important role in the coals sorption characteristics. Similarly, enhanced recovery of coalbed methane through nitrogen or carbon dioxide injection is under consideration as a viable process to increase the rate of methane recovery. In studying these types of applications, it will be necessary to use second-generation coalbed reservoir models which use a compositional approach in modeling the selective adsorption/desorption of different gas components.

PROBLEMAS Write the finite difference analog of the following partial differential equations of fluid flow in porous media: a.

(i) Consider kx, Ax, , and B as constant properties.

(ii) Consider kx, Ax, , and B as non-constant properties. b.

<

=0

First, substitute for the potential gradients, then approximate. None of the reservoir properties are constants. c.

Use the implicit scheme.

SOLUCION

a. < represents the flow of a single-phase, incompressible fluid in one-dimensional rectangular coordinates. (i) Since kx, Ax, , and B are to be treated as constants,

< or

< The finite difference approximation will be

< or

< or

< (ii) If kx and Ax distributions are heterogeneous, we can write the governing partial differential equations as

< or

< The finite difference approximation to the above equation will be

< or, in more general form,

< b.

< Potential gradient is defined as (D positive downward)

< Similar definitions will hold along the y and z directions. Substituting these definitions into the given partial differential equation, we obtain

< we can rearrange the above equation as follows: <

The finite difference analogue of this equation is

<

<

c.

< The above equation represents the single-phase flow of a slightly compressible liquid in two dimensional rectangular oordinates. The finite-difference representation using the implicit scheme is <

OTRO

Consider the mathematical formulation of the two-phase flow of oil and gas. Identify the unknowns of the problem and give the equations that need to be supplied for the solution. Assume that sandface pressures are specified at the well locations. SOL We can express the two-phase flow of oil and gas using the following equations (for onedimensional flow): Oil equation:

< Gas Equation:

< where

< Unknowns of the problem are Po, Pg, So and Sg. Therefore, we need two more equations. These two auxiliary equations are Saturation relationship: So+Sg=1.00 Capillary pressure relationship: Pcgo(So) = Pg-Po The two remaining unknowns are the flow rates qo and qg. Since sandface pressures are specified, we can use the well equations to calculate the flow rates for each phase:

<

OTRO The following partial differential equation describes a specific fluid flow problem in a porous medium with kx = 10 md, = 2 cp and Vb = 10,000 ft3. After examining the

given mathematical formulation, describe the flow problem and he porous media in the fullest extent.

SOL

immediately signals that we are dealing with an incompressible The equation < fluid (note the absence of time dependency). Thus, we can start with the most general form of the single-phase, incompressible fluid flow equation in one dimension. Assuming no depth gradient,

For a homogeneous system,

or

Comparing with the given equation, we can write

For = 2 cp, k = 10 md, Vb= 10,000 ft3 and B=1, q= 300 STB/D (injection rate OTRO You are given the function f(x)=3x2 + 6x + 5. a. Calculate the value of at x = 1 using forward, backward and central difference approximations using a grid spacing of X = 1. Compare your results against the exact value of b. Calculate the value of at x = 1.

at x = 1 using central difference.

Compare your result against the exact value of SOL

at x = 1 and comment.

Given the function f(x)=3x2+ 6x-5, we can calculate the exact value of =12

at x = 1 as

= 6x+6

a. forward-difference approximation for

at x = 1, with x=1:

backward-difference approximation for

at x = 1, with x=1:

central-difference approximation for

at x = 1, with x=1:

b. From part (a), we can see that the central difference approximation gives the same result as the exact differentiation. This is not surprising, because centraldifference approximation is second-order accurate, and the higher-order derivatives of the given function (which is a quadratic equation) vanish. That is,

Therefore, there is no error introduced with the central-difference approximation when a second-order function is approximated.

OTRO Give a complete formulation of two-phase (gas-water) flow applicable to a twodimensional domain with no depth gradients. Assume the solubility of gas in water and the capillary pressure between the phases to be negligible. Treat the porous medium as homogeneous, isotropic and incompressible. Reduce the equations to their simplest forms and identify the principle unknowns and the equations that are necessary to solve the problem. SOL Assumptions specified: Rsw=0 Pcgw=0 (Pg=Pw) No depth gradient (=P) 2-D flow (x and y directions only) Homogeneous property distribution Isotropic property distribution (kx=ky=k) Water equation:

Gas equation:

Since we have homogeneous and isotropic property distributions, we can further simplify the two equations above:

In these two equations, there are three principal unknowns: P, Sw and Sg. We therefore need one more equation to close the system. This last equation will be the saturation equation, which states that pores are always 100% filled with water and gas. In other words, Sw+Sg=1.00 For the flow rates, it will be necessary to use the well equations if sandface pressure is specified. If one of the rates is specified, then the other flow rate will be treated as an unknown. OTRO

You are given the formulation below:

a. Describe the fluid flow problem and the porous media to the fullest extent. b. What are the principal unknowns of the problem? Also, provide the necessary auxiliary equations to solve the problem.

SOL.

a. The two equations represent two-phase (oil and water) flow in a one-dimensional porous medium. Although the reservoir has uniform thickness and width (i.e., Ax is constant), it has heterogeneous permeability distribution. The reservoir is inclined, and therefore we consider the depth gradient in the x-direction. The formation porosity is a function of pressure, which is why it is not carried to the front of the derivative as a multiplier. We ignore the capillary pressure between the oil and water phases, which is why there is no subscript assigned to P. b. The problems principal unknowns are P, So and Sw. In addition to the two partial differential equations given above, we need to use the saturation relationship So+Sw=1.00 The flow rates will be calculated in conjunction with the well equations.

OTRO Consider the one-dimensional flow of an incompressible fluid in a homogeneous reservoir which is inclined with respect to the datum plane as shown in the figure. Note that flow is taking place in one-dimensional mode parallel to the bedding plane. (Flow direction is indicated as x-direction.). Figure 1 a.

Write the partial differential equation which describes the flow problem outlined above. SOL.

a. < Assuming that Ax and kx are constant, and since m and B are constants, we can simplify this equation as

< For this problem, dD/dx is constant. Therefore, the equation reduces to

<

; integrating a second time, we get P=C1x+C2. Integrating once, we get < Using boundary conditions C2=3000 and C1=-1, we obtain P = -x+3000. At x=300 ft, P=2700 psia. Again at x=300 ft,

<

OTRO
Consider the two-dimensional body-centered grid with the boundary conditions as indicated in the following figure. Figure 1

Figure 1

Show all the reflection nodes and their respective equations that you will need to solve this problem.

SOL.

Figure 1

For reflection nodes: 1x, 3x, 5x, 9x, 2x and 4x (no-flow boundary),

For reflection nodes: 1y, 2y , 9y , 10y , 11y , 12y (no-flow boundary),

For reflection nodes: 8x and 12x,

For reflection nodes: 7y and 8y,

OTRO

Solve < for the two grids shown using zero boundary conditions. First, use Gauss-Seidel, = 1 . Then repeat with PSOR using = 1.2, 1.4, 1.6, 1.8 and 2.0. Plot number of iterations versus and determine the opt. Figure 1

Figure 1

SOL.
For configuration a, with an initial guess of 100, we obtain the following results ( =0.1): Table 1

Table 1

Again, with an initial guess of 100 and convergence criterion of 0.1, we obtain the following results ( =0.1): Table 2

Table 2

Note that an exact solution for both of the systems will entail a zero distribution throughout the systems.

OTRO
Consider the following Dirichlet-type problem given to describe the two-dimensional flow of a single-phase slightly compressible liquid in an isotropic and homogeneous porous medium with no depth gradients or wells. Figure 1 a.

Write the partial differential equation which describes this problem and simplify it. b. Write the characteristic finite-difference approximation of part (a) (Express the pressures at time level .) c. Identify the unknowns of the problem and generate the system of equations.

SOL.
a. Assuming no depth gradients, we can express the single-phase, slightly compressible flow problem in two dimensions as

Since there are no wells, and the reservoir has isotropic and homogeneous property distribution, we can further simplify the above equation (treating m and B as constants):

b. The characteristic finite difference approximation to the above partial differential equation, using the implicit procedure, is

We can rearrange this equation as (note that (x)i,j = (y)i,j = constant

c. Essentially, the problem has two unknowns (shown as 1 and 2 on figure 1 ):

figure 1

At node 1:

Which we can simplify as

At node 2:

Again, after simplifying,

In matrix form, we can write the above two equations as follows:

OTRO
Find the pressure distribution and flow rate into the well for the following 2dimensional, homogeneous and isotropic system. In solving the system of equations generated by the finite-difference approximation, use the Gauss-Seidel iterative procedure, with a pressure tolerance of 1 psia. Calculate the percent error by checking the material balance. Figure 1

Figure 1

SOL.

Figure 1

The two-dimensional, single-phase, incompressible flow equation in its general form is

For the homogeneous, isotropic reservoir specified above, we can collapse the governing partial differential equation to

The finite difference approximation to this simplified partial differential equation is

Since sandface pressure is specified at the well block, flow rate is an unknown. Therefore, it will be necessary to incorporate the flow rate equation using

Close inspection of the figure reveals that because of the existing symmetry planes, there are actually four different blocks (numbered 1, 2, 3 and 4 in the above figure) with unknown pressures. Therefore, we will only need to write four equations at these four blocks and solve them simultaneously. Block 1 (well-block) is the only one where q _ 0. Therefore, we need to calculate the following groups.

Then we can calculate the productivity index (PI) for the well as follows:

Then, q1=-5.35(P1-600) Equation for Block 1:

or -5.186P1 + 2P2 + 2P4 = -712.1 Equation for Block 2: P1+P3-4P2+P3+2000 = 0

or P1+4P2+2P3=-2000 Equation for Block 3: P4+P2-4P3+1600+2000 = 0 or P2-4P3+P4 = -3600 Equation for Block 4: P3+P1-4P4+P3+1600 = 0 or P1+2P3-4P4=-1600 We can write these four equations for the Gauss-Seidel procedure as follows:

The following table summarizes the progression of the Gauss-Seidel iterations with a convergence criterion of =1: k 0 1 2 3 4 5 6 7 8 9 10 P1 1000 909 1134 1296 1371 1403 1418 1425 1428 1430 1430 P2 1000 1227 1512 1633 1683 1707 1718 1723 1725 1726 1727 P3 1000 1497 1617 1681 1712 1727 1733 1736 1737 1738 1739 P4 1000 1356 1492 1565 1599 1614 1621 1629 1626 1627 1627 (initial guess)

As we can see from the above iterations, convergence has been obtained after 10 iterations within a tolerance of 1 psi. The flow rate into the well, then, is q1=-5.35 [1430-600] = -4441 STB/D ("-" sign indicates production). Material Balance Check: Since this is a steady-state flow problem, the volume of fluid entering the well block has to be equal to the volume leaving the block through the wellbore. Volume entering well block:

Volume leaving well block = 4441 STB/D Material balance error = 0.3%

OTRO
Consider the single-phase, incompressible fluid flow problem taking place in a onedimensional homogeneous porous medium composed of four uniform blocks as shown below. a. Write the partial differential equation which governs the flow problem described above and put it into its simplest form. b. Give a finite-difference approximation of the partial differential equation of part (a) and put it into a characteristic form. c. Generate the system of equations that need to be solved to determine the pressure distribution in the system. Assume that g/gc. Do not solve the equations. d. Can you offer an educated guess for the expected flow rate from the well located in block 1? What is your guess? Figure 1

SOL. a. Starting with the single-phase, incompressible fluid flow equation in one dimensional rectangular coordinates, we have

< For a homogeneous reservoir, we can further simplify this relationship:

< Assuming g/gc=1,

<

b.

< For Block 1, we implement the no-flow boundary condition Similarly, for Block 4,

< For the well block (Block 1), we need to use the flow rate equation:

< At this stage, we are ready to write the finite-difference equations for each block. Block 1:

< Since P0=P1 (from the given boundary condition), we can further simplify the above equation to -3.296P1 + P2 = -2886.10 Block 2:

< which simplifies to P1-2P2+P3 = 58.33 Block 3:

< which simplifies to P2-2P3+P4 = 7.77 Block 4:

< which simplifies to P3-P4 = -177.22 Now we can solve the above four equations in four unknowns simultaneously. In the matrix form, we have

< The solution vector will be P1=1279.27; P2=1330.39; P3=1439.84; and P4=1577.06 psi. d. Since this is a steady-state flow problem, the volume entering from the right end of the reservoir should be equal to the flow rate from the well. Volume entering:

< Well production:

(-5.176)(1279.27)+6418.44 = -203.06 which checks with the entering volume to within 0.1%.

OTRO
Consider the single-phase, incompressible fluid flow problem shown in the following figure. The reservoir is homogeneous and there are no depth gradients. The outer boundaries of the blocks are completely sealed. A strong bottom drive keeps blocks 1 and 4 at a constant pressure of 3000 psia and block 3 at a constant pressure of 2000 psia. Sandface pressure of the well in block 5 is controlled by the hydrostatic head of the fluid column in the well, i.e., wellhead pressure is atmospheric pressure. Figure 1 a.

Calculate the artesian flow rate at the surface. In the solution of equations use Gauss-Seidel iterative procedure with a convergence criterion of 0.5 psia. b. Perform a material balance check.

SOL The governing equation is

which simplifies to

The finite-difference approximation will be

Since (x)i,j = (y)i,j,

Finite Difference Equation for Block 2: P2+3000-4P2+2000+P5=0 -3P2+P5=-5000 Finite Difference Equation for Block 5:

which can be simplified to P2-2.807P5=-4409.92 Gauss-Seidel equations will then be With iterations; K 0 1 2 3 4 5 P2 2500 2500 2487.23 2485.71 2485.53 2485.51 P5 2500 2401.08 2457.12 2456.52 2456.51 2456.51

P2=2485.5 psia; P5=2456.5 psia Flow rate calculation: q5=(-16.375)(2456.5)+28601.7=-11623.67 STB/D b. Material Balance check: Volume leaving the system=

Volume entering the system:

% error = 0.05

OTRO

Consider the incompressible fluid flow through the two-dimensional porous media as shown below. The reservoir has homogeneous and isotropic property distribution with no-flow boundaries Both of the wells are stimulated, and they have a skin factor of -2. The shaded blocks are maintained at 3000 psia via strong bottom water encroachment. Figure 1 a.

Write the partial differential equation which describes the problem and identify the unknown blocks and unknowns of the problem. b. Write the finite difference equations for blocks 4, 6, 7, and 12 using the numerical values of the coefficients

SOL
. The governing equation is

< For the homogeneous, anisotropic reservoir described in this problem, we can write the above equation as

< Unknown pressures: P3, P4, P6, P7, P8, P9, P10, P11, P12 Other unknowns: Flow rate from well in Block #4; sandface pressure at the well in Block #6. b. First, we write the characteristic form of the finite-difference equation:

<
We can manipulate this equation as follows: <

which simplifies further to

< For blocks with no wells, we will have

< For the block with the injection well, q = 600 STB/D (Block 6): 0.2P6+0.3P6-P6+0.3P7+0.2P10=-30 For the block with the production well, we need to calculate re 4 and k4 in the flow rate equation (psf=1000 psia; Block #4).

< Then

< Upon substitution into equation for Block 4, we obtain

< For Block 7, 0.2P3 + 0.3P6 - P7+0.3 P8 + 0.2P11=0 For Block 12, 0.2P8 + 0.3P11- 0.7P12= -600

OTRO
Consider the flow of a single-phase incompressible fluid in a reservoir which is represented by the body-centered grid as shown below. Figure 1

Figure 1

Solve for the pressure distribution in the reservoir using LSOR procedure with opt = 1.00

SOL

This equations finite-difference approximation is

In the above equation,

which gives qi,j as

We need to solve the line equations separately and continue with the iterations until we attain convergence. In this problem, Line 1 has one equation. For Line 2, we need to solve 3 equations simultaneously. For Line 3, there is again only one equation to be solved separately.

OTRO
Consider the steady-state flow of a single-phase incompressible liquid as shown in the following figure. The permeability and thickness maps are provided. There are two wells in the system producing at the indicated rates. Calculate the value of the pressure gradient along the unknown boundary indicated in the figure.

Figure 1

Figure 1

SOL;
At j=4, p/y= 0.3 indicates influx. The total influx from j=4 blocks will be

Wells 1 and 2 are producing -1000 STB/D. Therefore, across the lower boundary, there must be and influx of (-1000+676.2=-323.8 STB/D). Thus,

OTRO Aki me quede

Consider the unsteady-state flow of a slightly compressible fluid in the twodimensional system shown below. Pressure throughout the reservoir is given as 3600 psia before the well is put on production. Assuming that the pressure-dependent fluid properties are constant within the pressure range of the reservoir, solve for the production rate at the end of 30 and 60 days using the backward difference approximation to the governing partial differential equation. NOTE: All external boundaries in this reservoir are no-flow boundaries. Figure 1

Figure 1

SOL

Simplifying the above equation, we get

The implicit finite-difference approximation will be

Rearrangement of this equation gives

In the above equation,

with re2 = 0.14[6002+6002]1/2 k2 = 42 md S2=0

Therefore,

Now we can write the finite-difference approximations for the five blocks using a time step size of 30 days. For Block 1, P1+P1 -4P1 +P2+P1 +0=0.064P1 -(0.064)(3600) or -1.064P1+P2=-228.81 For Block 2,

or P1 - 4.095P2 + P3 + P4= -1679.841 For Block 3, P5+P2-4P3+P3+P3+0 = 0.064P3-228.81 or P2-2.064P3+P5 = -228.81 For Block 4, P4 +P4 -4P4 +P5 +P2 + 0 = 0.064P4 -228.81 or P2-2.064P4+P5 = -228.81

For Block 5, P5 +P4 -4P5 +P5 +P3 + 0 = 0.064P5 -228.81 or P3 +P4-2.064P5 = -228.81 At the end of the 30 days, we need to solve the above five equations in five unknowns to find the pressure distribution in the system. The solution of this system of equations generates P1=1994.25 psia; P2=1893.07 psia; P3=2039.01 psia; P4=2039.01 psia; P5=2086.65 psia. Note that as expected, P3 = P4 because of the existing symmetry. The flow rate from the well at t=30 days will be q2 = (-0.61)(1893.07)+858.7 = -296.07 STB/D (production) For the second time step at t=60 days, we generate the following equations: For Block 1, -1.064P4+P2=(-0.064)(1994.25) or -1.064P4+P2=-127.63 For Block 2,

or P1-4.095P2+P3+P4=-1572.19 For Block 3, P2+-2.064P3+P5 = (-0.064)(2039.01) or P2-2.064P3+P5 = -130.49 For Block 4, P2 -2.064P4 +P5 = (-0.064)(2039.01) or P2-2.064P4+P5 = -130.49

For Block 5, P3 +P4 -2.064P5 = (-0.064)(2086.65) or P3 +P4-2.064P5 = --133.55 When we solve the above five equations in five unknowns simultaneously, we will find the pressure distribution at t=60 days. The solution vector for this time step is P1 = 1569.09 psia; P2 = 1541.88 psia; P3 = 1586.36 psia; P4 = 1586.36 psia; P5 = 1601.87 psia. Again, the symmetry between Blocks 3 and 4 are preserved. The flow rate from the well at t=60 days will be q2 =(-0.61)(1541.88)+858.7= -81.85 STB/D (production) As shown by the results of these two time-step calculations, the production rate from the well at 30 days declined by 214.2 STB/D by the end of 60 days. This number is reasonable, because the well is being produced at a constant sandface pressure of 1400 psia.

OTRO
Consider the one-dimensional reservoir with no depth gradients as shown in the following Figure 1. Assume the existence of single-phase, slightly compressible fluid flow dynamics. From seismic studies, it is concluded that the west boundary of the reservoir is completely sealed. However, there is a disagreement on the nature of the east boundary. Using the pressure and production data provided, answer the following. Note that pressure values reported are measured downhole at the sandface, and flow rates are surface flow rates.

Figure 1

a. b.

Determine the nature of the east boundary. It is proposed to drill a third well at t=60 days. Your engineering manager indicates that the third well should only be drilled if pressure at the location of the third well is not less than 50% of the initial formation pressure at 60 days. What would you recommend to your manager? Justify your recommendation. Use implicit finitedifference approximation in your answer.

SOL
a. The governing partial differential equation for this problem is

After simplifying, we obtain

The finite-difference approximation will be

or

In this equation, we can calculate the following groups:

and

therefore,

We also need to calculate the equivalent well block radius, re:

At t = 30 days: We can use the well flow equation to find P1 at t = 30 days.

For Well #1, we can write this equation as

Solving for P1 gives P1 = 3257.63 psia at t = 30 days. Also, P2 at 30 days is given as P2 = 3320.14 psia. Now, we can write the finite-difference equation for Block 2 at t = 30 days.

or substituting for P1n+1 and P2n+1, we can solve for P3n+1 at t=30 days. 3257.63-(2.0478)(3320.14)+P3n+1= -(0.0478)(4000) P3 = 3350.15 psia Therefore, at t = 30 days, block pressures are determined by P1 = 3257.63 psia; P2 = 3320.14 psia; P3 = 3350.15 psia. Assume that _P/_x = C1 along the east boundary: Figure 1

Figure 1

Now we can write the characteristic finite-difference equation at t =30 days for Block 3:

or

C1=-2.6310-4 At t =60 days:

0 (for all practical purposes)

Since C1 =0, we conclude that the east boundary is also a no-flow boundary.

The finite-difference approximation needs to be written for each block at t =60 days. Since we have already determined the nature of the east boundary, there should be no difficulty in writing these equations as follows: i =1:

or

i =2:

or

i =3:

or

The solution of the above three equations in three unknowns gives the pressure distribution in the reservoir at t =60 days as P1 = 1764.33 psia P2 = 1790.95 psia P3 = 1862.08 psia Clearly, P3 at t = 60 days is less than 2000 psia (i.e., 50% of the original formation pressure). Therefore, according to the managers criterion, the third well should not be drilled.

OTRO
Consider the flow of a compressible fluid in the one-dimensional heterogeneous porous medium shown in the following figure. Pressure at every point in the system is given as 3600 psia before the well is opened to production. Calculate the pressure distribution at the end of the 20th day of production. Also, calculate the wells production rate. Use backward-difference scheme with a pressure convergence tolerance of 10 psia. Use a time step size of 20 days. Figure 1

Figure 1

Repeat the solution using an available simulator for the following combinations of time step sizes and pressure convergence tolerance: a. t = 20 days; = 0.001 psi b. t = 20 days; = 1 psi c. t = 1day; = 0.001 psi d. t = 1day; =1 psi e. t = 1day; = 10 psi

SOL
Results are summarized in the following table: Table 1

ESTIMACION

Reserves Estimation: Introduction The term "reserves" means different things to different people. To the banker, reserves are the amount of capital retained to meet probable future demands. To the oil and gas operator, reserves are volumes of crude oil, natural gas, and associated products that can be recovered profitably in the future from subsurface reservoirs. Reasons for Reserve Estimates Estimates of oil and gas reserves are required for different purposes by different segments of the industry and at different stages in the life of a particular oil and gas property. Segments of the industry concerned with oil and gas reserves include companies and individuals responsible for exploration, development, and operation of oil and gas properties buyers and sellers of oil and gas properties

 banks and other financial institutions involved in the financing exploration, development, or purchase of oil and gas properties agencies with regulatory or taxation authority over oil and gas operators investors in oil and gas companies Depending on the scope of their operation, operators of oil and gas properties require reserve estimates at various stages of exploration, development, and production. Of concern are potential reserves on undrilled prospects proved, probable, and possible reserves on prospects being developed sizing and design of equipment to process reserves and transport them to market opportunities for additional profit from incremental reserves that might be attributable to stimulation of producing wells, infill drilling, equipment modifications or additions or improved recovery projects.

Methods of Estimating Reserves

Methods of estimating reserves may be classified as deterministic (i.e., calculating a single "best" estimate) or probabilistic (i.e., calculating a range of estimates reflecting the probability distributions of the basic data). While it has been recognized that there is always some degree of uncertainty in estimating reserves (see, for example, Walstrom et al. 1967) most of the published literature on reserve estimating has focused on deterministic methods. Probabalistic methods, however, have received continuing attention in the geologic literature on the evaluation of exploration "plays" (see, for example, Newendorp 1975). With the advent of the North Sea discoveries in the late 1960s and early 1970s, probabilistic methods were applied with increasing frequency to estimate reserves in these fields. This led Keith (1986) to propose adoption of a system of reserves classification based on these methods. While the need for such methods was recognized as being appropriate to situations like the North Sea, the World Petroleum Congress recommended adoption of a classification system based on deterministic methods (Martinez et al. 1987). Accordingly, the methodology discussed in this work is based on deterministic methods. As noted by Arps (1956), the methods used to estimate reserves and the accuracy of the result depend on the type, amount, and quality of geologic and engineering data available, all of which generally will depend on the maturity of development and production of the field or reservoir. Figure 1 illustrates schematically the relationships between

 the maturity of a property,

Figure 1

the general methods used to estimate reserves for the property the range of uncertainty in the reserve estimate.
Methods to estimate reserves are categorized here as

analogy/statistical (identified by Arps as the "barrels-per-acre period," or Period I); volumetric (identified by Arps as the "barrels-per-acre-foot period,"
or Period II); and

performance (identified by Arps as the "decline-curve period" or Period III).

The performance method includes material balance analysis and performance/decline curve analysis.

Analogy/Statistical Methods
Analogy/statistical methods typically are used for undrilled prospects, and to supplement volumetric methods in a field or reservoirs early stages of development and production. In addition, the method may be used to estimate reserves for undrilled tracts in a partially developed field or reservoir. The methodology is based on the assumption that the analogous field, reservoir, or well is comparable to the subject field, reservoir, or well, regarding those aspects that control ultimate recovery of oil or gas. The weakness of the method is that this assumptions validity cannot be determined until the subject field or reservoir has been on sustained production.

Volumetric Methods
Volumetric methods are used when subsurface geologic data are sufficient for structural and isopachous mapping of the objective field or reservoir. One of the objectives of this mapping is to estimate oil and gas initially in place. The fraction of oil and gas initially in place that is commercially recoverable may be estimated using a combination of analogy and analytical methods. In a relatively uncomplicated geologic setting, it may be possible to make a reasonably accurate estimate of oil and gas initially in place with relatively sparse subsurface control. In contrast, in a complex geologic settin (e.g., one characterized by extensive faulting and/or complex stratigraphy) it may not be possible to make accurate maps until the field is almost completely developed.

Performance Methods
Performance methods may be used after a field, reservoir, or well has been on sustained production long enough to develop a trend of pressure and/or production data that can be analyzed mathematically. The analysis may involve material balance calculations or it may involve curve fitting trends of oil and/or gas production, pressure, water/gas ratio (WGR), water/oil ratio (WOR), gas/oil ratio (GOR), or combinations of these performance indicators. The curve fitting procedure is based on the assumption that those factors that control the fitted trend will continue in the future. Material Balance Methods Material balance methods may be used to estimate reserves when there are sufficient reservoir pressure and production data to perform reliable calculations of hydrocarbons initially in place and to determine the probable reservoir drive mechanism. For reliable material balance calculations, the reservoir should have reached semisteady state conditions, i.e., pressure transients should have affected the entire initial hydrocarbon accumulation. Depending on reservoir fluid and rock properties and on the reservoir drive mechanism, this could require cumulative production of as much as 5 to 10% of the hydrocarbons initially in place.

Reliable application of this method requires accurate historical production data for all fluids (oil, gas, and water), accurate historical bottomhole pressure data, and pressure-volume-temperature (PVT) data representative of initial reservoir conditions. If computer simulation is being considered, historical bottomhole pressure data may be required for each well in the reservoir, depending on the degree of reservoir complexity and the purpose of the simulation study. For volumetric gas reservoirs, a graphical form of the material balance equation may be used to estimate gas initially in place and reserves. For dissolved gas-drive oil reservoirs, the material balance equation may be used to estimate oil and gas initially in place and the probable drive mechanism. To estimate reserves, the material balance equation usually is combined with gas/oil relative permeability data in one of the prediction methods, either Tamer (1944) or Muskat (1945). In partial water drive reservoirs, it may be preferable to use a computer simulation model. In high-permeability reservoirs, where reservoir pressure does not exhibit large areal variations, "zero-dimensional" or "tank" model material balance calculations usually are acceptable. In low-permeability reservoirs, where reservoir pressure may exhibit large areal variations, or in geologically complex reservoirs, it may be necessary to use a multidimensional reservoir simulation model. Performance/Decline Curve Analysis The term "decline curve analysis" refers to the analysis of declining trends of the production of oil or gas the principal products of oil wells or gas wells, respectively versus time or versus cumulative production to estimate reserves. Depending on the reservoir drive mechanism and operating practices, however, it may be possible to estimate reserves before there is a decline in the production rate of the principal product. This procedure is called "performance trend analysis." For example, in a bottom-water drive oil reservoir, wells generally begin producing water early in life. Depending on pump capacity, oil production from these wells usually can be maintained at more or less constant rates for a period of time. Frequently, a semilog plot of the fraction of oil in the total liquid production versus cumulative oil for individual wells is linear. Extrapolation of the fraction of oil in total liquid to an economic limit may be used to estimate reserves for these wells.

Combination Methods
Usually, more than one method is used to estimate reserves. Typically, in the early stages of development and production of a field or reservoir, reserves are estimated using a combination of analogy and volumetric methods. In some areas, it may be feasible to utilize seismic data to help determine reservoir or field size before there are sufficient well data to prepare reliable geologic maps. As development continues, and the early wells begin to develop pressure/production trends, reserves for those wells may be estimated using performance/decline curve analysis. Reserves for undrilled tracts in a developing area may be estimated by analogy with older wells in the same, or similar, reservoirs in the field.

Reconciliation Between Methods

Analogy and volumetric methods for estimating reserves are static methods and may yield results significantly different from performance methods, which are dynamic methods. Field case histories have attested to many situations where well and reservoir performance differed significantly from that estimated prior to initiating production. See, for example, Buchanan and Hoogteyling (1979), Cronquist (1984), Hazeu et al. (1988), Markum et al. (1978), and Van Rijswijk et al. (1981). In the North Sea Thistle Field, for example, Nadir and Hay (1978) reported that the first seven development wells encountered reservoir conditions in excellent agreement with conditions predicted by the operators predevelopment reservoir model. This model had been developed from seismic mapping, geologic modeling, and five widely spaced appraisal wells. Subsequent to initiating production, however, wellhead pressures of wells in the crestal area of the field declined faster than had been anticipated, and "it became obvious that ... production rate ... would be significantly below expectation" (Nadir 1980). Periodic monitoring of well and reservoir performance throughout reservoir life and reconciliation of differences between static and dynamic reserve estimates are essential elements of good reservoir management.

Reconciliation of Material Balance and Volumetric Estimates

The volumetric and material balance methods are independent ways to estimate oil and gas initially in place. Because the basic assumptions for each method are different, the two methods may not account for the same volume of hydrocarbons, which might lead to significant differences between estimates. The volumetric method accounts for net pay observed in wells and generally is based on the assumption that the pay is continuous between wells. The material balance method accounts for production and the resulting pressure decline and is based on the assumption that the pressure decline is caused by withdrawals from the same fluid system. Depending on the depositional environment of the reservoir rock and on well spacing, pay may not be continuous between wells (Ghauri et al. 1974; Sneider et al. 1977; George 1978). Also, intervals counted as pay for volumetric calculations but not perforated for production may be in reservoirs separate from intervals that have been perforated. The unperforated intervals may not contribute to the pressure response and may not be accounted for in material balance calculations. Because of these and other factors, there may be substantial differences between the results of volumetric and material balance estimates of oil and gas initially in place. Differences of this type are commonly reported in complex carbonate reservoirs (Stiles 1976). If such differences are observed and cannot be reconciled by careful review of the data and computational procedures, then a detailed geologic and engineering review may be warranted. Such a review may lead to the drilling of additional wells or the perforating of additional intervals for production.

Volumetric Estimates of Intial Oil and Gas in Place

Oil Reservoirs
For an oil reservoir, or for the oil column of an oil reservoir with a gas cap, we may calculate the oil initially in place as follows: N = 7758o(l - Swo)Aoho/Boi (3.1)

where:

N = oil initially in place (STB) 7758 = barrels in an acre-foot o = average porosity in the oil zone (fraction) Swo = average water saturation in the oil zone (fraction) Ao = area of the oil zone (acres) ho = average net thickness of oil zone (ft) Boi = initial formation volume factor (RB/STB) For most practical applications, oil initially in place usually is rounded to the nearest thousand stock tank barrels (MSTB). Porosity and water saturation should be volume-weighted averages in the oil zone. If there is an oil-water transition zone initially present, the average water saturation may be weighted accordingly for use in Equation 3.1. If the transition zone occupies a significant portion of the reservoir, it may be preferable to calculate oil initially in place in the transition zone separately. Solution gas dissolved in the oil at initial reservoir conditions may be calculated as follows: Gsi = NRsi (3.2)

where:

Gsi = solution gas initially in place (SCF) Rsi = initial solution gas/oil ratio (SCF/STB) For most practical applications, solution gas initially in place usually is rounded to the nearest thousand standard cubic feet (MSCF).

Gas Reservoirs
For a nonassociated gas reservoir, or for a gas cap, free gas initially in place may be calculated as follows: GFi = 43560g(l - Swg)Aghg/Bgi (3.3)

where:

GFi = free gas initially in place (SCF) 43560 = cubic feet in an acre-foot g = average porosity in the free gas zone (fraction) Swg = average water saturation in the free gas zone (fraction) Ag = area of gas cap or gas reservoir (acres) hg = average net thickness of gas cap or gas reservoir (ft) Bgi = initial formation volume factor (CF/SCF)

For most practical applications, free gas initially in place usually is rounded to the nearest million standard cubic feet (MMSCF). Porosity and water saturation should be volume weighted averages in the gas cap or gas reservoir. Condensate, or distillate, in the vapor phase at initial reservoir conditions may be calculated as follows: Ci = GFiCvi (3.4)

where:

Ci = condensate, or distillate, initially in place (STB) Cvi = initial condensate/gas ratio (STB/MMSCF) GFi = free gas initially in place (SCF) For most practical applications, condensate initially in place usually is rounded to the nearest thousand stock tank barrels (MSTB).

Sources of Data
Equations 3.1 through 3.4 require three types of information: petrophysical data, fluid data and volumetric data. Sources of these data are noted briefly below.

Petrophysical Data
The petrophysical data needed for Equations 3.1 and 3.3 include average water saturation (Sw) and porosity (). Water saturation may be estimated from log analysis, capillary pressure data, or analysis of cores taken using oil-base mud. Log analysis methodology is discussed in several publications, for example, Asquith (1982), Dewan (1983), Dresser Atlas (1982), Ellis (1987), Hilchie (1982), Jorden and Campbell (1986), Schlumberger (1987), Timur (1987), and Tixier (1987). Measurement and interpretation of capillary pressure data are discussed by Amyx, Bass, and Whiting (1960) and Keelan (1977), among others. Porosity may be determined from core and/or log analysis. Core analysis procedures are discussed by Muskat (1949), Keelan (1977), and others. In situations where data are not available, reference may be made to tabulations of reservoir data published by the Society of Petroleum Engineers (Jenkins 1987).

Fluid Data
The fluid data needed for Equations 3.1 through 3.4 include initial formation volume factor (Boi or Bgi), and initial solution gas/oil ratio (Rsi) or initial condensate/gas ratio (Cvi). (The latter two parameters are called GOR and CGR, respectively, in subsequent discussion.) These data may be obtained from laboratory analysis of fluid samples representative of the reservoir being studied.

Caution should be exercised in sampling reservoirs with fluids near critical conditions, because cases have been reported where fluid composition varied from undersaturated oil near the base of the accumulation to gas condensate at the top (Neveux et al. 1988) If laboratory PVT data are not available, parameters needed for Equations 3.1 through 3.4 may be estimated from empirical correlations that have been published by Standing (1952), Arps (1962), Glas (1980), Vasquez and Beggs (1980), and AlMarhoun (1988). Standing (1962), Beggs (1987), and Sutton and Farshad (1990) have commented on the reliability of various PVT correlations; these sources may be useful in deciding which correlation to use.

Volumetric Data
The volumetric data needed for Equations 3.1 and 3.3 include reservoir area and net pay. Early in the development of a field or reservoir, before there are enough subsurface geologic data for mapping, reserves usually are estimated on a per-well basis. Reservoir area in Equation 3.1 or 3.3 becomes drainage area for the well or wells in question. Drainage area usually is estimated by analogy with wells with similar rock-fluid properties and producing with the same reservoir mechanism. Pressure transient tests, in the form of reservoir limit tests, can also be useful in determining the drainage area. Drainage volume usually is calculated as estimated drainage area times the true vertical net pay logged in the well or some fraction of the net pay, depending on circumstances. As development proceeds, and sufficient subsurface geologic data become available, structure and isopachous maps should be made.

Analysis of Data
Raw data from field and laboratory sources should be analyzed carefully and crosschecked for reasonableness and consistency before being used for engineering calculations. Muskat (1949), Amyx, Bass, and Whiting (1960), and Havlena (1966 and 1968) provide comprehensive discussions of data analysis for reservoir engineering.

Statistical Analysis
Law (1944) and Bulnes (1946) were among the earliest investigators to note that porosity and permeability of reservoir rocks tend to display certain statistical distributions. For example, porosity in both sandstones and carbonates tends to exhibit a normal, or Gaussian, distribution. Permeability tends to exhibit a log-normal distribution i.e., the logarithm of the permeabilities tends to be normally distributed. In rocks with a progressively lower porosity, the variance exhibited by the permeability tends to be progressively larger. Determination of the statistical distribution of porosity and permeability is key to the characterization of the reservoir in question and to the proper treatment of other reservoir properties with which porosity and permeability may be correlated.

In addition, as discussed by Bennion and Griffiths (1966), Testerman (1962), and others, statistical analysis may reveal the presence of more than one type of depositional unit in the reservoir, which can have an important bearing on estimating reserves. Average water saturation (Sw) is one of the parameters needed for volumetric calculations of oil and gas initially in place (Equations 3.1 and 3.3). If there is not a significant oil-water transition zone, the average water saturation in the reservoir may be approximately equal to the irreducible water saturation (Swi). In some reservoirs it might not be possible to determine Swi directly from in-situ log measurements; indirect methods using core analysis might have to be used. Depending on rock texture, there might be statistically valid correlations between Swi and one of several different rock parameters, including (a) porosity, (b) permeability, or (c) the square root of permeability divided by porosity. Stanley (1973) and others have provided guidelines for testing the statistical validity of various correlations. If porosity is normally distributed, the best value to use for these correlations is the arithmetic average porosity. If permeability is log-normally distributed, the best value to use for some applications may be the median permeability (Warren and Price 1961). Richardson (1987), however, reported getting better agreement with well-test rates by using arithmetic average permeability. Cronquist (1984) has reported that there was better agreement between volumetric and material balance estimates of gas initially in place in a very heterogenous, Miocene-sand gas reservoir by using the geometric mean average water saturation, rather than the arithmetic average water saturation. In other reservoirs it may be necessary to use capillary pressure curves to estimate irreducible water saturation. Due to the wide variations in rock texture, there may not be a statistically valid correlation between capillary pressure behavior and porosity and permeability for the reservoir in question. In this situation, one solution is to select the capillary pressure curve from a rock sample that has porosity and permeability approximately equal to the arithmetic average porosity and median permeability for the reservoir.

Adjustment/Correlation of Data
Petrophysical and fluid data from all sources should be compared and cross-checked for consistency. Several aspects of this process for porosity, water saturation, PVT, and well-test data are discussed in the following paragraphs. Complete discussions have been provided by Amyx, Bass, and Whiting (1960), Havlena (1966), Keelan (1977) , and Muskat (1949) Porosity Depending on lithology, wellbore conditions, and logging device, porosities indicated by wireline logs may or may not agree with those determined from core analysis. The problem is especially severe in loosely consolidated sands, shaly sands, and fractured, vugular carbonates. Core recovery frequently is incomplete, and correlation of core porosity with log response may be poor. Core depths, which are determined from drillpipe measurements, seldom agree with wireline log depths, and

appropriate adjustments must be made before direct comparison is possible. A coregamma log usually facilitates correlation between core depths and log depths. Accepted procedure is to correct core depths to wireline log depths and attempt to correlate log response with core porosity at the log-adjusted depth. Frequently, this results in a poor correlation, especially in laminated sandstones and very heterogeneous carbonates. One procedure that may improve the correlation is to develop a running average of core porosity over several feet, double weighting the central values. An averaging process may come closer to the averaging process inherent in most porosity logging tools, which integrate formation response over several feet. The procedure used to develop a running average of core data should be compatible with the logging tool zone of investigation. Reference to logging company technical brochures is recommended. An alternate procedure to improve the correlation between log response and core porosity has been discussed by Havlena (1966). This procedure is called a forced fit, and treats core porosity and log response as two sets of spatially random variables, not necessarily dependent on log or core depth. For each well, core porosity and log porosity are ordered separately in ascending order, regardless of the core or log depth. The sets of values in ascending order are plotted, and a least-squares fit is used to determine the relation between log and core porosity. Usually, porosity from full-diameter cores is considered the standard against which log response is calibrated. In highly fractured or vugular carbonates, however, neither core nor log data may provide representative porosity data. In addition, core analysis of loose, unconsolidated sands may result in abnormally high porosity values, even though the cores are taken with a rubber sleeve and compacted before the analysis (Elkins 1972). Mattax et al. (1975) have discussed methods for core analyses of friable and unconsolidated sands; however, caution should be exercised in utilizing any core data from such sediments. Water Saturation For reservoir zones at irreducible water saturation i.e., where Sw = Swi, for the same rock texture there should be a reasonable correlation between logdetermined water saturation and (a) permeability, (b) porosity, or (c) square root of [permeability divided by porosity]. Frequently, this type of correlation may help identify different depositional units within a reservoir (Sneider et al. 1977), which is important because such units may have a significant influence on well and reservoir performance. Well Test and PVT Data If recombination samples have been taken for PVT analysis, the initial GORs determined from well tests those at initial reservoir pressure usually are in reasonable agreement with those determined from laboratory flash liberation tests (the so-called separator tests). If there is reasonable agreement, the laboratory formation volume factor at initial reservoir pressure (Boi), and the corresponding GOR (Rsi), should be adjusted to compensate for separator conditions in the field, before being used in Equations 3.1 and 3.2. The procedure for making these adjustments has been discussed by Amyx, Bass, and Whiting (1960) and by Moses (1986). Comparable procedures should be followed for gas reservoirs.

If bottomhole samples were used for PVT analysis, there is a possibility of nonrepresentative sampling, especially if the reservoir has low permeability and was at or below the initial bubble-point pressure at the time of sampling. In this event, GORs from initial well tests usually are significantly greater than those determined from laboratory separator tests. The recommended procedure uses empirical correlations to estimate the parameters needed in Equations 3.1 and 3.2, pending availability of representative samples or adjustment of the sample data to reflect the correct initial bubble-point pressure. If GORs from initial well tests are significantly greater than those indicated by valid PVT data, the presence of free gas in the reservoir possibly undetected during development drilling should be suspected. This could have a significant effect on volumetric calculations of oil initially in place. If PVT analyses are not available, the initial GOR from well tests should be checked against empirical correlations of bubble-point pressure and formation volume factor. The test data used in these correlations to estimate initial formation volume factor should indicate a bubble-point pressure equal to, or less than, initial reservoir pressure. When material balance or reservoir simulation calculations are made, the PVT data always should be smoothed using the Y and the _V functions, as discussed by Amyx, Bass, and Whiting (1960). The Y function is used to smooth the total relative volume data (Bt); the _V function, the relative oil volume data (Bo).

Initial Reservoir Conditions

The PVT terms in Equations 3.1 through 3.4 should represent fluid properties at reservoir temperature and initial reservoir pressure. Although highly accurate determination of reservoir temperature and initial reservoir pressure may not be critical for volumetric calculations given the uncertainties in other factors such determination is essential for PVT data because of its influence on other calculations usually made later in the life of the reservoir.

Pressure
Accurate determination of initial reservoir pressure is critical if material balance methods are to be used to estimate oil or free gas initially in place. In addition, for an oil reservoir, comparison of initial pressure with bubble-point pressure can provide useful information regarding whether an initial gas cap is a reasonable possibility. Accurate determination of initial reservoir pressure is especially important in geopressured gas reservoirs. Initial reservoir pressure should be checked against local pressure gradient information. Subnormal initial pressure might be indicative of partial drainage from other wells or reservoirs producing from the same formation and may establish a need for early pressure maintenance. Reliable application of the material balance method mandates historical bottomhole pressure data in addition to accurate initial pressure.

The methodology of measuring and interpreting bottomhole pressure data has been discussed by Plisga (1987); detailed discussion of these procedures is beyond the scope of this work.

Temperature
Accurate determination of reservoir temperature is important if laboratory PVT data are to be measured under reservoir conditions. In addition, accurate reservoir temperature is necessary for calculations of the performance of gas reservoirs. It is good practice to determine the initial temperature-depth profile in each producing well in a field using a continuous recording thermometer of the type described by Plisga (1987) Caution should be exercised in using bottomhole temperatures observed during openhole logging, because these temperatures will have been reduced to belownormal formation temperature by circulation of the drilling fluid.

Regional Correlations
In the early stages of development and production of a field or reservoir, measured bottom-hole pressure and temperature data may be too sparse to provide a reliable estimate of initial conditions. In this case, regional correlations may be helpful. Listed below are correlations for initial reservoir pressure and formation temperature for several regions in North America (Scientific Software 1975; Hitchon 1984). Reservoir temperature and initial pressure estimated using these correlations should be considered preliminary and not a substitute for actual measurements. Initial Reservoir Pressure The following correlations have been developed for initial pressure in oil and gas reservoirs in North America: Louisiana Gulf Coast (hydropressure zone) pi = 0.465Dss Texas Gulf Coast (hydropressure zone) pi = 0.45Dss West Texas pi = 0.439Dss + 1250 Illinois, Indiana,Kentucky, and Michigan pi = 0.427Dss + 210 Kansas pi = 0.475Dss + 375 Oklahoma pi = 0.439Dss + 310 (3.5d) (3.5e) (3.5f) (3.5g)

(3.5a)

(3.5b) (3.5c)

Montana, North Dakota,Wyoming, Utah, and Colorado pi = 0.47lDss + 1120 Alberta Lower Cretaceous pi = 0.36D - 133 (3.5h)

Alberta Devonian D-2 pi = 0.39D - 200 Alberta Devonian D-3 pi = 0.358D - 73 Alberta Bashaw carbonate complex pi = 702 + 0.527Dss (3.5k)

(3.5i) (3.5j)

Alberta Rimbey-Meadowbrook carbonate complex pi = 536 + 0.466Dss Alberta Windfall-Swan Hills carbonate complex pi = 947 + 0.491Dss where:

(3.51) (3.5m)

pi = initial pressure (psi) D = depth (ft) Dss = depth subsea (ft) Note: The hydropressure zone is that portion of the geologic column in which initial reservoir fluid pressure is hydrostatic i.e., the initial pressure can be attributed to the weight of the overlying static column of formation water in the zone of saturation (Gary et al. 1972). Formation Temperature The following correlations have been developed for oil and gas reservoirs in North America: Louisiana Gulf Coast (hydropressure zone) Tf = 74 + 0.0125Dss North Texas Tf = 60 + 0.01675D (3.6b) Oklahoma Anadarko Basin Tf = 66 + 0.0111D Oklahoma deep Anadarko Basin (below 21,000 ft) Tf = 66 + 0.014D Alberta Tf = 35 + 0.0201D (3.6e) Alberta Bashaw carbonate complex Tf = 32 + 0.0187D Alberta Rimbey-Meadowbrook carbonate complex Tf = 49 + 0.0167D Alberta Windfall-Swan Hills carbonate complex Tf = 32 + 0.0193D (3.6f) (3.6g) (3.6h) (3.6c) (3.6d) (3.6a)

Nichols (1947) has published an iso-temperature gradient map of the TexasLouisiana Gulf Coast region. In general, formation temperature in the hydropressure zone may be estimated from thermal gradient maps published by the USGS and the AAPG (1976), using the equation:

where:

Tf = Tsa + gG D

(3.7)

Tf = formation temperature (F) Tsa = average surface temperature (F) gG = geothermal gradient (F/ft) D = depth (ft) Volumetric Mapping Types of Traps The degree of uncertainty in a volumetric estimate of oil and gas reserves is influenced more by the degree of geologic complexity of the reservoir and by the amount and quality of geophysical and subsurface geologic data than by almost any other combination of factors. For example, regarding the Leman Field in the North Sea one of the worlds largest gas fields Craig et al. (1977) have commented that "an accurate volumetric determination of the gas in place is impossible because of seismic interpretation problems and the limited well control that results from the cluster-type development ... the situation is aggravated further by a complex fault system. It is worth noting, however, that with the progress that has been made in 3-D seismic techniques in recent years, we are in a better position to understand structurally complex reservoirs. In general, three broad categories of reservoirs may be identified: (1) those controlled by structure, (2) those controlled by stratigraphy, and (3) those controlled by a combination of structure and stratigraphy. Reservoirs on simple anticlinal closures uncomplicated by faulting and containing laterally continuous porosity are among the easiest to estimate reserves and are subject to the least uncertainty. The Santa Fe Springs Field in California and the Abqaiq Pool in Saudi Arabia are examples of simple structural traps. At the other extreme are stratigraphically complex reservoirs on highly folded and faulted structures. The Painter Reservoir Field in the U.S. western overthrust belt and the Maloosa Field in Italy are examples of complex structural traps that pose substantial difficulties in estimating reserves. It is very important throughout the life of a reservoir that the engineers involved in estimating reserves become familiar with the geologic setting in which the reserves occur and the possible geologic complications associated with that setting. Reservoir geology impacts all phases of reservoir management, including drilling and completion techniques, well testing, drainage areas, well and reservoir performance, reserve estimates, and improved recovery techniques. Levorsen (1967) and Lowell (1985), among others, have provided treatments of petroleum geology that are recommended reading. Volumetric Mapping of Reservoirs After a discovery well and one or more confirmation wells have been drilled on a prospect, seismic control (if available) should be integrated with subsurface geologic data in the preparation of structure maps on key horizons. As the field or reservoir is

developed, core, log, and well-test information should be integrated into the mapping process to delineate gas-oil, gas-water, or oilwater contacts on each reservoir, as applicable. For sandstone reservoirs, gross and net sand isopach maps should be made for each productive horizon. Fluid-fluid contacts should be transferred to the net sand isopach for each reservoir and net hydrocarbon isopachs should be prepared. Amyx, Bass, and Whiting (1960) have provided an extensive discussion on volumetric mapping. Bank-head (1962) has provided additional discussion on volumetric mapping of oil and gas reservoirs. Depending on the purpose for mapping, or on the anticipated reservoir drive mechanism, different oil-water contacts may be defined. There are several possibilities, including the 50% water level defined as the depth below which the reservoir will produce in excess of 50% water which may be delineated by testing, capillary pressure, and water-oil relative permeability data the 100% water level (analogous to the 50 water level) the free water level (the datum of zero capillary pressure) below which there frequently are no "shows" in the cores In hydrodynamic traps, the oil-water contact may be tilted, as discussed by Hubbert (1953). In low-porosity heterogeneous rocks, the 50% and 100% water levels may not be planar, but are controlled by the local capillary properties of the reservoir rock-fluid system. In a typical reservoir, however, the free water level should be planar (Knutsen, 1954). Determining Net Pay Determination of net pay is one of the most important steps in volumetric mapping to determine oil and gas initially in place. In the following discussion, the term "net pay" refers to "true vertical net pay1, i.e., logged net pay thickness corrected for borehole inclination. Snyder (1971) observed that net pay in a given reservoir may be determined for one of several different purposes. The procedure to determine net pay and the results may be different for each purpose. For example: When a well log is being evaluated to select an interval for completion, net pay generally includes only those intervals judged likely to contribute to well inflow at commercial rates. When a reservoir is being evaluated to determine total hydrocarbons in place for example, to corroborate material balance calculations net pay usually includes all hydrocarbon-bearing intervals likely to contribute to the energy balance; net pay determined for this purpose usually is greater than that determined for the first purpose.

 If a waterflood is being considered, net pay typically includes only those intervals considered "floodable." The methods used to define floodable intervals are subjective and may exclude intervals that could contribute to recovery by imbibition. Net pay determined for this purpose may be less than that determined for either of the first two purposes. If a reservoir is to be unitized and net pay is part of the unitization formula, the determination of net pay may be subject to arbitrary rules to ensure "uniformity," which may result in determining net pay unrelated to any of the considerations noted above. Depending on the purpose, methods for determining net pay also depend on the type of core, log, and well-test information available and on the lithology of the reservoir. In general, however, the following steps are involved: First, determine gross interval by establishing the top and bottom of the zone of interest. In sand-shale sequences, the inflection point of the SP curve or midway point of the gamma-ray curve usually is considered to be a zone boundary. In carbonates, it may involve using one or more of the porosity curves to establish a minimum porosity or "porosity cutoff," or it may involve using a combination of logs to determine lithologic top and bottom. Then, exclude "nonpay" intervals within the gross interval; in sandshale sequences this may be done by using a shale cutoff, or by excluding intervals where there is a specified degree of reversal in either the SP or the gamma-ray curve. In carbonates, this may be done by excluding intervals below a specified porosity cutoff, as above. In some cases for example, shaly, laminated-sand reservoirs, and many carbonate reservoirs it may be difficult to determine net pay with an acceptable degree of confidence. A net pay isopach made under this type of uncertainty may be subject to frequent correction as interpretive procedures are revised or new data become available. In this situation it may be desirable to map gross pay and apply an average net-to-gross ratio to account for nonproductive rock. The net-to-gross factor may be revised as warranted, thereby avoiding the need to remap the entire reservoir. George and Stiles (1978), Havlena (1966), Sneider et al. (1977), and Wilhite (1986) have provided additional discussions regarding net pay. Porosity-Feet-Saturation Mapping (Isovols) In reservoirs where there are wide variations in porosity and/or water saturation, it may be preferable to map these variations rather than attempt to apply average porosity or average water saturation over the entire reservoir. An isovol map is similar to an isopachous map, except that the contoured data are the product of the following equation:

(porosity)

(net pay)

(hydrocarbon saturation)

Planimetered volume is hydrocarbon pore volume (HCPV). Amyx, Bass, and Whiting (1960) have provided a discussion of various methods to determine the volume from isopachous maps. Reservoir Limits Reservoir limits include fluid contacts (i.e., gas-oil, oil-water, gas-water, faults, porosity or permeability pinchouts and unconformities and other truncations (e.g., diapiric salt or shale). In many cases, especially early in the drilling and production stages of reservoir life, it is difficult to determine these limits within acceptable confidence levels. In structurally controlled reservoirs, for example, wells may not have penetrated an oil-water contact, and the subsurface information may be limited to lowest known oil (LKO) and highest known water (HKW). In this case, only that volume down to the LKO may be classified as "proved" (SPE 1987). Depending on the horizontal distance between LKO and HKW, up to half the volume between LKO and HKW might be classified as "probable." Typically, areas involving more than two anticipated development locations beyond LKO would be classified as "possible." There is, however, no consensus on this matter. Some engineers might assign probable reserves downdip from an LKO a vertical distance equivalent to two gross pay thicknesses, and classify as possible the remaining volume to HKW. Analogous treatment would be afforded situations where only lowest known gas (LKG) and HKW had been defined. Situations also may occur in which wells have not penetrated the gas-oil contact, and subsurface information is limited to highest known oil (HKO) and LKG. In this case, the volume up to HKO may be classified as proved oil, and the volume down to LKG may be classified as proved gas. There is no consensus on the fluid type assigned to the interval between LKG and HKO. Assuming continuity of reservoir facies, depending on the volumes involved, some engineers might assign half the volume to gas and half to oil. Others might assign the entire volume to gas. In stratigraphically controlled reservoirs, the porosity or permeability limit may be only approximately delineated by dry holes around the periphery of the accumulation. The undrilled area between commercial wells and dry holes will define a "band of uncertainty" around the accumulation which may be one or more locations wide. Undeveloped reserves proved, probable, or possible might be assigned to locations in this area, depending on the following: the quality of reservoir rock observed in the commercial wells around the periphery the depositional environment of the reservoir rock the distance to the nearest dry hole the apparent rate of thinning of reservoir rock the minimum thickness of reservoir rock needed to support a commercial well.

Capillary Transition Zones If a reservoir includes a significant oil-water transition zone, it may be preferable to calculate oil initially in place separately for (a) the volume above the transition zone and (b) the volume within the transition zone. Significant transition zones may be expected in heavy-oil reservoirs, or in any oil reservoir with low permeability. Gas-oil and gas-water transition zones are negligibly small. Arps (1964) has provided a discussion of the geologic and engineering aspects of capillary transition zones. Depending on the reservoir drive mechanism, some of the oil in the transition zone may be commercially recoverable. Cronquist (1984) has provided an example of an estimate of waterflood reserves from an oil-water transition zone in a light-oil reservoir (39 API). The transition zone occupied 18% of the initial reservoir volume (NAF), but contained only 10% of the ultimate recovery. In heavy-oil reservoirs, a significant portion of the oil initially in place may be in the oil-water transition zone. On low-relief structures, the entire reservoir may be in the oil-water transition zone, and it might not be possible to make a water-free completion, even at the crest of the structure. Arps (1964) reported a structure in the Tensleep (Pennsylvanian) Sandstone in Wyoming containing 19 API oil where there was insufficient vertical closure for a commercial accumulation. It has been noted that the "best" capillary pressure curve to use for saturation calculations is from a rock sample that has arithmetic-average porosity and median permeability. The vertical distribution of reservoir rock above the 100% water level may be determined using a vertical distribution curve.

Volumetric Recovery Estimates for Oil Reservoirs

Black Oil Reservoirs
Methods are presented to estimate recovery of oil and gas from pressure depletion reservoirs (We = 0) with and without an initial gas cap and from reservoirs with water influx (We > 0). Pressure Depletion No Initial Gas Cap Two stages are recognized in the pressure depletion of oil reservoirs: (a) the undersaturated stage, i.e., reservoir pressure between initial pressure and bubblepoint pressure, and (b) the saturated stage i.e., reservoir pressure less than the bubble-point pressure. Undersaturated Stage Hawkins (1955) has shown that, in general, recovery efficiency of oil attributable to rock and fluid expansion caused by a pressure drop from initial reservoir pressure (pi) to the bubble-point pressure (pb) may be calculated as follows: [ERo]xp = ce(pi - pb)Boi/Bob (3.8)

where:

[ERo]xp = recovery efficiency of oil attributable to expansion of the rock-fluid system from initial to bubble-point pressure

ce = effective compressibility of the rock-fluid system between pi and p, defined as: (Soco + Swcw + cp)/So where: So = oil saturation (fraction) co = oil compressibility (1/psi) Sw = water saturation (fraction) cw = water compressibility (1/psi) cp = pore compressibility (1/psi) (3.9)

Boi = oil formation volume factor at initial pressure (RB/STB) Bob = oil formation volume factor at bubble-point pressure (RB/STB) If PVT data are not available for subject reservoir oil or a similar reservoir oil, co may be estimated using methods developed by Trube (1957) or Vasquez and Beggs (1980). Water compressibility may be estimated using data published by Dodson and Standing (1945), or by Osif (1988). Pore compressibility should be determined in the laboratory, placing samples under uniaxial stress (Andersen (1988). If laboratory data are not available, pore compressibility may be estimated with caution from data published by Fatt (1958), Dobrynin (1962), or Newman (1973). For naturally fractured reservoirs, Reiss (1980) has noted that effective compressibility can be expressed as:

< where: m = matrix porosity (fraction) f = fracture porosity (fraction) cpm = matrix pore compressibility (1/psi) cpf = fracture pore compressibility (1/psi) Swm = matrix water saturation (fraction)

(3.10)

and other terms are as previously defined. If fracture porosity is significantly less than matrix porosity, fracture compressibility may be ignored (Reiss 1980). In certain cases, pore and water compressibilities are negligibly small, and ce may be equated to co. Equation 3.8 becomes:

[ERo]xp = (Bob - Boi)/Bob (3.11) Comparing Equations 3.8 and 3.11, it may be observed that: co = (Bob - Boi)/Bob(pi - pb) (3.12) Because there is no free gas evolved in the reservoir during this stage of pressure depletion, the recovery efficiency of solution gas should be the same as that for oil. Saturated Stage After reservoir pressure decreases below the bubble-point pressure and the critical gas saturation is exceeded, the recovery efficiency of oil and gas will be governed by several factors, including the gas-oil relative permeability characteristics of the reservoir rock and the gravity/viscous force ratio. There are no simple analytical equations available to estimate recovery efficiency of oil and solution gas attributable to dissolved gas drive. However, several investigators, for example, Arps and Roberts (1955), have published the results o stepwise calculations to estimate oil recovery efficiency by this mechanism in which they used the Muskat (1945) form of the differential material balance equation. Arps and Roberts (1955) used PVT correlations to simulate 12 different reservoir oils. They used 6 different gas-oil relative permeability curves three to represent sandstones and three to represent carbonates. The gas-oil relative permeability curves, identified as maximum, average, and minimum recovery, are shown in Figure 1 and Figure 2 .

Figure 1

As noted by the authors,

Figure 2

the following conditions prevail:

Recovery efficiencies were calculated from bubble-point pressure to an abandonment pressure which was 10% of the bubble-point pressure. The methodology assumes complete dispersion of the liberated solution gas (no gravity segregation) with relative production of gas and oil at each stage controlled by relative mobilities of gas and oil as determined by the kg/ko curve and fluid viscosities. Connate water saturation was assumed to be 25% in the sandstones and 15% in the carbonates. Recovery efficiencies in parentheses were estimated from calculations discontinued before abandonment pressure.
In addition, the computational procedure used by the authors does not account for additional stock tank liquids recovered from rich solution gas produced from volatile oil reservoirs.

Using a calculation procedure similar to that used by Arps and Roberts (1955), Wahl, Mullins, and Elfrink (1958) developed a set of nine nomographs to estimate recovery efficiency of oil attributable to solution gas drive. They used kg/ko relations calculated from empirical correlations of sandstone relative permeability data. The nine nomographs cover combinations of three different kg/ko curves and three different residual oil viscosities which range from 0.5 to 10 centipoise. The results were presented as graphical functions of bubble-point pressure (pb) formation volume factor (Bob), and solution gas/oil ratio (Rsb). The assumptions in this work are the same as those in the Arps and Roberts (1955) work, except for the following:

Recovery efficiency was calculated assuming abandonment at atmospheric pressure the Arps and Roberts (1955) calculations assumed abandonment at 10% of the bubble-point pressure.
Roberts (1955) work, Sgc was determined by the kg/ko curve.

Equilibrium gas saturation (Sgc) was fixed at 5% in the Arps and

Interstitial water saturation was varied from 10 to 50% in the Arps and Roberts (1955) work, it was fixed at 25% for the sandstones.
Use of either of these studies requires knowledge of the gas/oil relative permeability characteristics of the reservoir in question. If laboratory kg/ko data are not available, reference may be made to correlations developed by Corey (1954) and Felsenthal (1959). Wylie (1962) and Rose (1987) have provided comprehensive discussions on relative permeability. API Correlation In 1967 the American Petroleum Institute (API) published the results of a regression analysis of oil recovery efficiency from 80 reservoirs in the USA that had produced by solution gas drive no initial gas cap (Arps et al. 1967). The correlation is: [ERo]sg = 0.41815 [(l - Sw)/Bob]0.1611 x [k/ob]0.0979[Sw]0.3722[pb/pa]0.l74l where: [ERo]sg = recovery efficiency of oil attributable to solution gas drive (fraction) = porosity (fraction) Sw = water saturation (fraction) Bob = formation volume factor at the bubble-point (RB/ STB) k = arithmetic average absolute permeability (darcies) ob = oil viscosity at the bubble-point (cp) pb = bubble-point pressure (psi) pa = abandonment pressure (psi) The regression coefficient is 0.932. The standard error of the estimate is 22% i.e., there is a 68% probability the correct value is within 22% of the estimated value. The data for this equation came from 67 sandstone reservoirs and 13 carbonate reservoirs. The range of rock and fluid properties for the 80 reservoirs is as follows: (3.13)

Sandstones min k (darcies) (fraction) Sw (fraction) 0.006 0.115 0.150 20 60 639 0.095 max 0.940 0.299 0.500 49 1680 4403 0.460

Carbonates min 0.001 0.042 0.163 0.032 302 1280 0.155 max 0.252 0.200 0.350 50 1867 3578 0.207

API
Rsb (SCF/STB) pb (psig) [ERo]sg (fr)

This work was updated in 1984, and the section titled API 1984 Study should be reviewed before using Equation 3.13. Solution Gas There is little historical data available on the recovery efficiency of solution gas from solution gas drive oil reservoirs. In part, this is due to the historically low economic value of this product compared to crude oil. Also, with the increasing application of pressure maintenance techniques, fewer oil reservoirs are produced to depletion by solution gas drive. Nevertheless, prudent asset management mandates a reasonable estimate of these reserves. This is especially so when an evaluation is required to compare economics of operation by solution gas drive versus operation under waterflood. In an oil reservoir produced at pressures below the bubblepoint pressure so that solution gas is evolved in the reservoir and becomes mobile the producing GOR will be determined by the /ko characteristics of the reservoir, the degree of gravity segregation, and position of impermeable streaks relative to well completion intervals. In the absence of significant gravity segregation, about 75% of the gas initially in solution can reasonably be expected to be produced if the reservoir is produced to an abandonment pressure approaching atmospheric. Roberts and Ellis (1962), using calculational procedures similar to those used by Arps and Roberts (1955), developed a set of correlations of gas/oil ratio history versus bottomhole pressure and cumulative oil recovery. Figure 3 is an example of one of these graphs.

Figure 3

These correlations are subject to the same limitations discussed previously. They can be used to make a reasonable initial volumetric estimate of solution gas recovery, pending development of individual well or reservoir performance or history matching by computer simulation. Pressure Depletion Initial Gas Cap As observed by Muskat (1945, 1949), the recovery efficiency of oil by pressure depletion of an oil reservoir with an initial gas cap will be governed mainly by the size of the initial gas cap and the degree of gravity segregation. There are no simple analytical equations or correlations available to estimate oil recovery efficiency in this type of reservoir. However, industry publications of case histories of the performance of gas cap oil reservoirs provide some guidance. No Gravity Segregation The case of no gravity segregation-the least favorable case was studied analytically by Muskat (1945) for an oil reservoir with a bubble-point solution GOR of 534 Scf/STB and a bubble-point oil viscosity of about 1 cp. Tabulated below is calculated oil recovery efficiency down to a reservoir abandonment pressure of 100

psi for various initial gas cap sizes, where gas cap size (m) is expressed as the ratio of initial gas cap volume to initial oil column volume:

Initial Gas Cap Size, m(fraction) 0.0 0.2 0.4 0.7 1.0

Oil Recovery Efficiency(% STBIP) 21.2 25.3 28.1 30.1 33.1

Gravity Segregation In actuality, there is always some degree of gravity segregation of free gas and oil during production from a gas cap oil reservoir. The relative importance of gravity segregation may be expressed by a gravity number (NG) of the form (Smith 1953): NG = ko (sina)/o where notation is consistent with SPE (1987) standard and units are: ko = effective oil permeability (millidarcies) o = oil viscosity (centipoises) = difference in specific gravity between reservoir oil and gas (gms/cc) a = dip angle (degrees) For guidance in making a preliminary estimate of recovery efficiency of oil, Table 1 compiled from published studies of actual reservoirs may be useful.

Table 1

Note that gas was injected into some of these reservoirs; the term "minj" indicates the volume injected as a multiple of initial gas cap volume. In the presence of gravity segregation, gas cap injection has the same effect -as a large initial gas cap. In an oil reservoir with a gravity number greater than about 10, effective gravity segregation is a reasonable expectation (Smith 1953). Richardson (1989) observed that, in the absence fo an active aquifer, a gas cap of at least 0.6 times the oil-zone volume is necessary for a major part of the reservoir to produce by gas cap expansion drive. Richardson and Blackwell (1971) have published guidelines for simple computational procedures to estimate reserves from oil reservoirs subject to gas cap encroachment. Water Drive Water drive may vary from complete to partial. If the volume of water encroaching into the initial hydrocarbon bearing pore space completely replaces the reservoir voidage of oil and gas or maintains reservoir pressure above the bubble-point over reservoir life (pa > ps) water drive is considered "complete." If the volume of water is insufficient to maintain reservoir pressure above the bubble-point and

reservoir pressure continues to decline over life (pa < ps) water drive is considered "partial." Richardson (1989) observed that an aquifer pore volume of at least 100 times the initial hydrocarbon pore volume is required for a strong water drive. API Correlation The API (1967) published the results of a regression analysis of the oil recovery efficiency from 70 water drive reservoirs in the USA, as follows: [ERo]wd = 0.54898[(l - Sw)/Boi]0.0422 x [kwi/oi]0.077[Sw]0.1903[pi/pa]0.2159 (3.14)

where:

[ERo]wd = oil recovery efficiency attributable to water drive (fr) wi = initial water viscosity (cp) oi = initial oil viscosity (cp) pi = initial reservoir pressure and other notation is as previously defined. The regression coefficient is 0.958. The standard error of the estimate is 18% i.e., there is a 68% probability the correct value is within 18% of the estimated value. The data for this regression came from 70 sandstone reservoirs with the following range of rock and fluid properties:

min k (darcies) (fr) Sw (fr) API oi (cp) [ERo]wd (fr) 0.011 0.111 0.052 15 0.2 0.278

max 4.0 0.350 0.470 50 500 0.867

The validity of this correlation is discussed later in this section under the subtitle API 1984 Study. Bottomwater Drive. There is no indication in the API (1967) work as to how many of the reservoirs produced with bottomwater drive and how many produced with edge water drive. The nature of the water encroachment is an important factor in estimating reserves, and several studies of bottomwater drive reservoirs are discussed in the following paragraphs. Note that the API results are for recovery efficiency from the entire reservoir. In bottomwater drive reservoirs, recovery efficiency from individual wells may vary substantially, depending on well spacing and the ratio of horizontal to vertical permeability, among other factors.

Muskat (1947) established the theoretical foundation to estimate the performance of bottomwater drive oil reservoirs. In that work, he made a clear distinction between encroachment of bottomwater due to an active aquifer and "coning" of underlying static water in a pressure depletion reservoir. In the latter case, there is a minimum production rate at which water is not produced the so-called "critical rate." In the case of a bottomwater drive reservoir, however, the "critical rate" has no meaning, as water is eventually produced by individual wells as the encroaching aquifer rises into the well completion interval. Unfortunately, in much of the published literature the term "coning" is used to describe bottomwater encroachment; this may cause confusion. Muskats (1947) analytical treatment of bottomwater drive reservoirs was, of necessity at the time, limited to the assumptions of (a) unit mobility ratio and (b) zero density contrast between oil and water. For conditions other than unit mobility ratio and zero density contrast, analytical studies of bottomwater drive published by Hutchinson and Kemp (1956) and model studies published by Henley et al. (1961) and by Ciucci et al. (1966) provide a basis for a preliminary volumetric estimate of recovery efficiency from this type of reservoir. Figure 4 from the Hutchinson and Kemp (1956) paper and Figures 2, 4, and 5 from the Henley et al. (1961) paper are reproduced in Figure 4 ,

Figure 4

Figure 5 ,

Figure 5

Figure 6 , and Figure 7 .

Figure 7

Figure 6

Reserve estimates from these figures should be considered preliminary, pending development of performance trends in individual wells, or history matching by computer simulation. The Hutchinson and Kemp (1956) work involved analytical studies using dimensionless well spacing (aD) -of 2.25, 6 and 12, where dimensionless well spacing was defined in the same manner as in the Muskat work i.e., aD = (a/h)(kv/kh)1/2

where:

a = well spacing (ft) h = initial oil zone thickness (ft) kv = vertical permeability kh = horizontal permeability Other parameters investigated in this study included dimensionless well penetration (bD) and mobility ratio (Mwo), where dimensionless well penetration was defined in the same manner as in the Muskat work i.e., bD = b/h

where "b" was defined as the vertical distance the completion interval extends into the initial oil zone. Mobility ratio was defined consistent with the current SPE (1987) definition, i.e., the mobility of water at residual oil saturation divided by the mobility of oil at initial oil saturation. From Figure 4 , which was calculated for aD = 6, it may be observed that, at this spacing, the parameter having the greatest influence on volumetric sweep efficiency is mobility ratio.

Figure 4

These results indicate that, at a water/oil ratio (WOR) of 20 to 1, the volumetric sweep efficiency may be expected to vary from about 40% (for a mobility ratio of 2) to about 80% (for a mobility ratio of 1/3). Calculations made for aD = 12 indicate volumetric sweep efficiency (at a WOR of 20 to 1) may be expected to vary from about 10 to 25% for the same variation in mobility ratio as shown in Figure 4 . Calculations for aD = 2.25 indicate volumetric sweep efficiency (at a WOR of 20 to 1) of about 90% for mobility ratios of 1 and 2.

The above results were calculated assuming negligible gravity effects. Regarding this assumption, the authors observed that "gravity forces can affect the performance, but in most cases any significant improvement ... will be realizable only at very low rates of oil production." The Henley et al. (1961) scaled model study included water/oil mobility ratios in the range 10 to 0.1, which covers oils in the medium to high API gravity range (approximately 20 to 40 API, respectively). Again, their Figure 2, 4, and 5 are reproduced in Figure 5 , Figure 6 , and Figure 7 , respectively. The parameter "R3", was defined as the gravity to viscous force ratio: R3 = kg A/q o The Henley et al. (1961) results are in reasonable agreement with those reported by Hutchinson and Kemp (1956) for the case where R3 = 0. However, the Henley et al. (1961) results for R3 > 0 indicate progressively larger volumetric sweep efficiencies for larger values of R3. Thus, these authors observed that "gravity effects must be considered in estimating 6i1 recovery because they can be a major influence on performance." Ciucci et al. (1966) reported on a model study of bottomwater drive using a dimensionless well spacing (aD) of 1.6 and a mobility ratio of 200. This mobility ratio would be applicable to about 15 API crude. The model parameters in their study were defined in the same manner as in the Henley et al. (1961) study. However, direct comparison of these results with those of earlier studies is not possible, as the range of parameters used by Ciucci et al. (1966) does not -overlap that used by Henley et al. (1961). The results of the latter authors, however, are qualitatively consistent with those of the earlier authors. Of particular significance is the observation by Ciucci et al.(1966) that "the sweep efficiency increases remarkably as the flow rate is lowered," which corroborates the work of Henley et al. (1961). All of the published studies cover models with the ratio of horizontal to vertical permeability constant throughout the model, with the tacit assumption that the permeability level also is constant. In actuality, this type of permeability distribution is rare. Fluvial point bars typically have lower average permeabilities toward the top; in contrast, offshore barrier bars and beach sands typically have higher average permeabilities toward the top (Sneider et al. 1978). All other factors being the same, it can be expected that wells with higher average permeabilities toward the top of the oil column will exhibit higher ultimate oil recovery efficiencies due to bottomwater drive than wells with lower average permeabilities toward the top. Viscous Oils From examination of the viscosity range included in the API (1967) correlation for water drive recovery efficiency Equation 3.14, it may be noted that viscous oils are included. Additional insight into estimating oil reserves in water-drive heavy-oil reservoirs is provided by Van Meurs and van der Poel (1958). Based on model experiments, they developed an equation of the form: Npu = Swi + B[M - fwa2(M - l)]/[M - fwa(M - 1)]2 (3.15)

where: Npu = ultimate recovery (pore volumes) Swi = irreducible water saturation (fr) B = 1 - Sor - Swi fwa = economic limit water cut (fr) M = oil/water viscosity ratio (not SPE "M") This work is applicable to reasonably homogeneous, edgewaterdrive reservoirs, where the displacement is not gravity stable, but is dominated by "viscous fingering." In addition, Timmerman (1982) has published graphs that may be used to make preliminary estimates of the recovery efficiency from viscous oil reservoirs producing by edgewater drive. These graphs are based on experiments with scaled models and cover oil/water mobility ratios (Mwo) from 1 to approximately 500. API 1984 Study As noted previously, in an update to the 1967 work the API (1984) studied an expanded data base which included 376 solution gas drive and 244 natural water drive oil reservoirs in 6 states in the United States. In reviewing the 1967 work, the API (1984) observed that the reservoir data sets were weighted in accordance with the quality of the data. A reservoir considered to have good data was counted three times; one with poor data, one time. When in the course of the 1984 review the reservoirs were not weighted for data quality, the statistical reliability of the correlations decreased significantly. Regarding the 1984 study, it was concluded that "no statistically valid correlation has been found between oil recovery and definable reservoir parameters." The API cautioned "against continued use of the (1967) correlations ... to predict recovery or recovery efficiency for any one reservoir." Regarding the above "caution," in the writers experience the API (1967) correlations both Equations 3.13 and 3.14 provide estimates which, in many cases, are in reasonable agreement with observed recovery efficiency. Because these equations are empirical correlations of large amounts of data, it should not be expected that they will provide acceptable estimates in every case. If usage is tempered with good judgment and results are cross-checked with other data, there is no reason to reject the 1967 API work. In further discussion of the 1984 work, it was noted that "calculated average recoveries in a single geologic trend

Volumetric Recovery Estimates: Retrograde Gas-Condensate Reservoirs

As with depletion drive oil reservoirs below the bubble-point, there are no simple analytical expressions available to calculate recovery efficiency of gas and condensate from retrograde gas-condensate reservoirs. Given a PVT analysis, the usual procedure is to make a series of flash calculations to determine recovery to an assumed abandonment pressure, as discussed by Jacoby, Koeller, and Berry (1959) and others. Frequently, however, a PVT analysis is not available. Eaton and Jacoby (1965) published a correlation that may be used to estimate recovery of condensate from a depletion drive gas-condensate reservoir, which is in the form:

where:

lnGLu = 2.7958 ln(API) + l.3921 ln(p)- 0.65314 ln(Rt) - 20.243

(3.16)

GLu = ultimate recovery of condensate at 500 psi abandonment pressure (bbl/bbl hydrocarbon pore space) API = stock tank gravity of condensate p = initial reservoir pressure (psi) Rt = aggregate GOR from entire separator train (SCF/STB) The regression coefficient is 0.814; the standard error of the estimate is 14.1%. This correlation was developed using 27 different reservoir-fluid systems with the following range of properties:

API: 45 to 65
p: 4,000 to 12,000 psi Rt: 2,500 to 60,000 Scf/STB Tf: 160 to 290 F Calculated gas recovery efficiency (to an abandonment pressure of 500 psi) ranged from 88 to 97%, with an average of 92.6%. In reservoirs with initial pressures in the range indicated, 500 psi abandonment pressure may be too optimistic. Accordingly, it is suggested that, for an initial volumetric estimate, gas recovery efficiency be estimated at 85%, unless there are factors that would warrant a different recovery efficiency. After such a reservoir has been producing for a time sufficient to establish stabilized well performance curves, well deliverability at the economic limit may be used as a basis to estimate bottomhole pressure at abandonment. Volumetric

Recovery Estimates: Nonretrograde Gas Reservoirs

For nonretrograde gas reservoirs (those where reservoir temperature is above the cricondentherm), two gas recovery cases are discussed: (a) no water influx, and (b) water influx. No Water Influx If there is no water influx, and if interstitial water and pore volume compressibility effects may be ignored, ultimate gas recovery efficiency may be calculated as follows: [ERg]pd = 1 - paZi/piZa (3.17)

where:

[ERg]pd = recovery efficiency of gas attributable to pressure depletion (fraction)

pi = initial reservoir pressure (psia) pa = abandonment reservoir pressure (psia)

zi = gas deviation factor at initial conditions (dimensionless) za = gas deviation factor at abandonment conditions (dimensionless)

Because reservoir temperature in these reservoirs is greater than the cricondentherm, there is no retrograde loss of condensate as reservoir pressure decreases. Thus, the recovery efficiency of condensate should, in theory, be the same as the recovery efficiency of gas. This is indicated with the notation.

[ERc]pd = [Erg]pd In actuality, however, as flowing wellhead pressure declines, some types of lease separation equipment for example low-temperature separators become less effective. Under these conditions, unless the lease separation equipment is modified, the producing condensate/gas ratio (CGR) may decrease gradually over life. (While the natural gas liquids not recovered on the lease might be recovered by a gas plant, consideration should be given to possible different ownership between lease and plant in assigning reserves.)
Lamont (1963), in a study of 37 pressure depletion gas reservoirs on the U.S. Gulf Coast, reported volume-weighted average recovery efficiency of gas to be 84.1%. (Reservoir average was 82.9%.) These reservoirs generally are moderately consolidated sand, with permeabilities of several hundred millidarcies. Although not reported, well spacing probably was about 160 acres. Lamont's study is consistent with results published by Stoian and Telford (1966). In their study of 158 pressure-depleted, or nearly pressure-depleted, gas reservoirs in Canada, average recovery efficiency of gas was estimated to be about 85%.

Hale (1981), Harlan (1966), and Stewart (1970), in separate papers on gas wells in low-permeability reservoirs in central United States, reported data that indicate significantly lower recovery efficiency of gas as the transmissibility (kgh) decreases below about 100 millidarcy-feet. Recovery efficiencies calculated from data in these papers are plotted on

Figure 1 .

Figure 1

Water Influx The recovery efficiency of gas from a gas reservoir subject to partial water drive can be expressed analytically with an equation of the form (Agarwal et al. 1965; Cronquist 1980): [ERg]wd = [1 - RpZ] + [RpZEVED) (3.18)

where:

[ERg]wd = gas recovery efficiency (fr) RpZ = ratio of abandonment p/z to initial p/Z, or pazi/piza EV = fraction of the initially gas-bearing volume swept by the aquifer ED = microscopic disp1acementefficiency, i.e., (1 - Swi - Sgr)/(l - Swi) (3. 18a)

where Sgr is the residual gas saturation in the flooded portion of- reservoir; in the absence of laboratory kw/kg data, it may be estimated as (Katz et al. 1966): Sgr = 0.62 - 1.34 (3. 18b) In the development of Equation 3.18, it is assumed that gas is trapped in the flooded portion of the reservoir and is abandoned at the same average pressure as the gas in the unflooded portion of the reservoir at abandonment (pa). Depending on reservoir dynamics, gas in the downdip areas of the reservoir may be flooded out at pressures higher or lower than reservoir abandonment pressure. See, for example, Cronquist (1980). In low permeability reservoirs, where this may not be a reasonable assumption, it may be desirable to use a gas reservoir simulator to estimate reserves. Gas recovery from partial water drive gas reservoirs is rate sensitive (Agarwal et al. 1965; Cronquist 1980). Agarwal et al. (1965) advocated that these reservoirs be produced at high rates to maximize gas recovery. However, depending on the ratio of viscous to gravity forces and the degree and nature of permeability heterogeneity, high production rates may cause irregular water encroachment, which could result in less ultimate recovery than if the reservoir were produced at a lower rate (Cronquist 1980). From examination of Equation 3.18, it may be noted that the term [1 - RpZ] is the portion of the recovery efficiency attributable to pressure depletion. (It is equal to the term [ERg]pd, previously defined.) The term [RpZEVED] is the portion of the recovery efficiency attributable to water influx. If there is complete pressure maintenance by the aquifer such that pa = pi the term [1 - RpZ] becomes zero, and the recovery efficiency is the product of the volumetric sweep (EV) and microscopic displacement (ED) efficiencies. The most difficult parameter to estimate in Equation 3.18 is EV. Even with considerable experience in a given area, in the absence of well performance data for the reservoir in question, it is unlikely an accurate estimate can be made of the ultimate volumetric sweep efficiency. In moderate- to high-permeability sand reservoirs, typical of those encountered in clastic Tertiary basins, an initial volumetric

estimate of 60% recovery efficiency of gas would be reasonable for reservoirs in which water drive was anticipated. In a study of a nearly depleted major gas field in the U.S. Gulf Coast, actual recovery efficiencies of gas from partial water drive gas reservoirs were observed to be in the range 40 to 75%. Variations between individual reservoirs were attributed to differences in aquifer strength, reservoir heterogeneity, and operating practices (Cronquist, 1984).

Reservoir Heterogeneity
Reservoir heterogeneity refers to the nonuniform spatial distribution of a reservoirs physical and chemical properties. During the last 20 years or so, there has been increasing recognition that reservoir heterogeneity has a major influence on the recovery of oil and gas from subsurface reservoirs. Several developments have contributed to a growing knowledge of reservoir heterogeneity:

increasing scrutiny and infill drilling of major waterfloods in the United States,
especially of the carbonate waterfloods in the Permian Basin of West Texas-New Mexico

development of better downhole logging devices (both open-and cased-hole), and of better interpretation techniques advances in geophysical technology and interpretation techniques, including seismic stratigraphy rapid advances in computer technology that have enhanced capabilities for geophysical processing, log analysis, and reservoir simulation development and application of enhanced recovery processes, which have spurred research on reservoir characterization use of multidisciplinary "teams" for coordinated geophysical,
geologic, and engineering studies of developing fields. Heterogeneities range from microscopic to megascopic scale and may be either favorable or unfavorable to recovery processes. Two examples of favorable heterogeneities are thin, lowpermeability layers and natural fractures. Shale stringers or thin, low-permeability layers near the bottom of a hydrocarbon column underlain by water may retard bottomwater coning (Wellings 1975). Shale stringers or low-permeability layers near a gas-oil contact may retard gas coning. Richardson et al. (1978) discuss other examples and provide guidelines for calculating the effect of discontinuous shales on oil recovery. Natural fractures may be either favorable or unfavorable, depending on reservoir drive mechanism. Without natural fractures, many low permeability oil and gas reservoirs would be noncommercial; examples include the Spraberry (a siltstone) and the Austin Chalk trends in Texas. Natural fractures, however, cause major problems in waterflooding; the same natural fractures that enhanced primary recovery from the Spraberry trend were detrimental to waterflooding this reservoir.

These two examples notwithstanding, most reservoir heterogeneities tend to be unfavorable to oil and gas recovery. Heterogeneities cause two general types of problems: they make it difficult to characterize and monitor reservoir performance of reservoirs, and they tend to cause nonuniform recovery of oil and gas. Characterization of a reservoir involves determining the spatial distribution of porosity, oil and gas saturation, oil and gas properties, permeability, relative permeability, net and gross thickness, fractures, dip, rock texture and mineralogy, and other physical/chemical properties relevant to commercial oil and gas recovery. It is seldom possible to characterize an oil or gas reservoir adequately from static data alone. Adequate characterization almost always requires ongoing monitoring of dynamic data e.g., individual well bottomhole pressure, well performance, and fluid saturations. Monitoring reservoir performance involves observation of spatial and temporal changes in pressure, fluid saturation, and fluid phase as the reservoir is produced i.e., the periodic procurement of dynamic reservoir data. In addition, monitoring reservoir performance may involve periodic material balance or reservoir simulation studies. Reservoir characterization and monitoring methodology (and the degree of reliability of the techniques used) are strongly influenced by the nature and degree of reservoir heterogeneity. For detailed discussions of monitoring techniques in various reservoir settings, see, for example, Denny and HeusserMaskell (1984), Harpole (1979), Haugen et al. (1988), LaChance and Winston (1987), Posten (1983), and Talash (1988). Heterogeneities in sandstone reservoirs tend to be controlled primarily by the original environment of deposition (Le Blanc 1977). Heterogeneities in carbonate reservoirs tend to be controlled by original environment of deposition and by postdepositional diagenesis (Jardine et al. 1977).

Sandstone Reservoirs
Heterogeneities of concern in sandstone reservoirs include the following:

Different depositional units in the same reservoir (Sneider et al. 1977). Both irreducible water and residual hydrocarbon saturations are strongly influenced by rock texture, which is controlled by depositional environment. Finegrained sediments, usually characteristic of low-energy depositional environments, tend to have high irreducible water and high residual hydrocarbon saturations; coarsegrained sediments,characteristic of high-energy environments, tend to have low irredicible water saturation and low residual hydrocarbon saturation. In addition, fine-grained sediments tend to have lower permeability than coarse-grained sediments. In water-drive reservoirs, high-permeability zones tend to be swept by water before low-permeability zones; in solution gas drive reservoirs, highpermeability zones tend to pressure deplete before low-permeability zones. Lateral discontinuities and multiple reservoirs in apparently "blanket" sands (Knutsen et al.,1971; Seal and Gilreath,1975; Hartman and Paynter,1979; McCubbin,1982; Weimer et al.,1982). Typically, a deltaic environment includes many sub-environments. Although sands deposited in these subenvironments often are interconnected, in many cases they are separated by thin, impermeable shale breaks that are not always detectable by wireline logs. If these separate sand bodies are not included in well-completion intervals, significant volumes of oil and gas may be left undrained.

 Shale "breaks" of indeterminate areal extent, which may have a major influence on fluid flow, depending on reservoir drive mechanism and spatial orientation of the shale breaks (Begg et al. 1989; Hazeu et al. 1988; Richardson and Blackwell 1971)

Carbonate Reservoirs
In addition to problems similar to those discussed for sandstone reservoirs, heterogeneities of concern in carbonate reservoirs include the following:

Lateral discontinuities in pay zones (Ghauri et al. 1974; Stiles 1976). The shelf carbonates of the U.S. Permian Basin typically contain numerous thin pay zones over a gross interval of several hundred feet; many of these pay zones are not laterally continuous between 40-acre spaced wells. Recovery efficiency from this type of rock sequence has been shown to be dependent on well spacing (Driscoll 1974; Barber et al. 1983; Barbe and Schnobelen 1987; Wu et al. 1988; Godec and Tyler 1989). Recovery efficiency in shelf carbonates in other areas of the world may exhibit similar spacing dependency. Very erratic, frequently vugular, porosity (Jardine et al. 1977, Horsfield 1958). The original depositional fabric of most carbonates is quite heterogeneous because of the wide variety of material comprising these sediments. In addition, carbonates are highly susceptible to postdepositional diagenesis, which often leads to more complex pore networks (Jardine et al. 1977). Highly vugular sections frequently are difficult to core to evaluate with logs.
This section is not intended to be a complete treatment of difficulties in exploitation and reserve estimation caused by reservoir heterogeneities. The subject warrants an entire monograph. The reader is referred to specific field case histories for example, McCaleb and Willingham (1967), Wayhan and McCaleb (1969), and Lelek (1983) and to the bibliography provided by Le Blanc (1977), for more comprehensive discussions. Reserves Estimation: Material

Balance Methods
Expanded Material Balance Equation Principles
The so-called Schilthuis (1936) material balance equation is one of the fundamental relations in reservoir engineering. In "expanded" form, which includes water influx, it may be stated:

Production of oil and gas

Expansion of oil and gas initially in place

Water influx

Assuming an initial gas cap and at this time ignoring compressibility of pore volume and interstitial water, the equation may be written:

Np[Bt + Bg(Rp - Rsi)] + WpBw = N[(Bt - Bti) + mBti(Bg - Bgi)/Bgi] + We

(1)

where the left-side terms account for the reservoir volume of oil, gas, and water production, and the right side terms account for the expansion of the oil and free gas initially in place, plus the water influx. (All notation in this section is SPE standard.) Equation 1 is consistent with the formulation of Schilthuis, who ignored compressibility. However, compressibility effects should be considered for material balance calculations involving oil reservoirs above the bubble-point. Depending on the magnitude of rock-fluid compressibility compared to overall system compressibility, it may be desirable to include rock-fluid compressibility for oil material balance calculations below the bubble-point. In Equation 1, it is assumed that the reservoir can be treated as a "tank," with spatial variations in PVT properties and reservoir pressure being averaged. There are three unknowns in Equation 1:

STB oil initially in place (N); size of the initial gas cap as a fraction of initial oil zone volume (m) ;
and

cumulative water influx (We).

In application, Equation 1 is solved at the end of successive time periods, generally quarterly, using the cumulative production data at the end of each period and PVT properties evaluated at the average static reservoir pressure at the end of the period. Theoretically, given enough pressure-production history and repetitive solutions of Equation 1, it should be possible to solve for all three unknowns. In practice, this is rarely possible, mainly because of errors in measuring, and problems in interpreting and averaging, bottomhole pressures. If both a significant initial gas cap and water influx are a possibility, efforts should be made to determine initial gas cap size using volumetric methods.

Limitations
Limitations to reliable application of the material balance equation are both theoretical and practical. Theoretical limitations are imposed by assumptions necessary for a tractable methodology, which are

The assumption that oil and free gas in the reservoir are in thermodynamic equilibrium. Wieland and Kennedy (1957) report about 20 psi supersaturation in experiments conducted using East Texas and Slaughter field cores. The assumption that the PVT data, obtained from differential
liberation, replicates the liberation process in the field. As discussed by Dodson (1953) and others, both flash and differential liberation of gas may occur at various times and places between the reservoir and the

stock tank, with the differences in PVT properties between the two processes increasing with more volatile oils.

The assumption that free gas in the reservoir has the same composition as free gas on the surface, differing only in volume, as expressed by the gas formation volume factor. With progressively more volatile oils, free gas in the reservoir contains progressively more liquids in the vapor phase that are recovered as stock tank liquids but are not accounted for by the differential liberation process.
Practical limitations are imposed by data requirements and reservoir conditions. Data required for reliable application of the material balance equation include PVT analyses of representative fluid samples, accurate static bottomhole pressure history of key wells in the reservoir and accurate monthly production data for oil, gas, and water. The accuracy requirements usually exceed the routine needs for many field operations. Reservoir conditions that may limit the reliability of a material balance estimate include:

Strong water drive and/or a large gas cap which maintain reservoir pressure at
nearly initial pressure. Under these conditions the material balance equation generally does not yield stable solutions, because the small pressure drops in the reservoir are frequently of the same magnitude as the errors in the measurements.

A really extensive reservoirs with different areas at different stages of development and production. Generally, this leads to wide variations in gas saturation and reservoir pressure that cannot readily be "averaged." A really extensive reservoirs with low values of kh/. These conditions make it difficult to determine the static bottom-hole pressure reliably and often cause large areal variations in pressure that are difficult to average.
interbedded with zones of low permeability, or highly fractured reservoirs. Under these conditions the lowpermeability zones, or the matrix blocks, usually pressure deplete more slowly than the highpermeability zones, or the fractures, and it is practically impossible to determine volumetrically weighted average reservoir pressure. Some of these limitations may be overcome by using a multidimensional reservoir simulator, rather than a zero-dimensional material balance, or tank, model. Irrespective of the material balance method used tank model or multidimensional simulator it is good practice to plot all of the bottomhole pressure data versus time on the same graph for all the wells suspected of being in a common reservoir. Such a plot usually provides valuable insight of the degree of communication between wells. It may help in identifying wells that are in separate reservoirs, contrary to the current geologic interpretation.

Very heterogeneous reservoirs with zones of high permeability

Reserves Estimation: Material Balance Methods

Havlena-Odeh Applications
Havlena and Odeh (1963, 1964) have shown that the material balance equation Np[Bt + Bg(Rp - Rsi)] + NpBw = N[(Bt - Bti) + mBti(Bg - Bgi)/Bgi] + We may be written as an equation of a straight line by making the following simplifications: (4.1)

Define FpR as the reservoir volume of oil, gas, and water production:
FpR = Np[Bt + Bg(Rp - Rsi)] + WpBw (4.2) Define Eo as the expansion of a unit volume of the oil (and initially dissolved gas) initially in place: Eo = (Bt - Bti) (4.3) Define Eg as the expansion of the initial gas cap: Eg = (Bg - Bgi) (4.4) Substituting these new terms, Equation 4.1 becomes: FpR = N[Eo + mBtiEg/Bgi] + We (4.5) Depending on whether there is an initial gas cap or water drive, or both, this equation may be used in various forms to determine the amount of oil and gas initially in place and the probable reservoir drive mechanism. To estimate reserves, one of the predictive forms of the material balance equation must be used. In this section, we consider six cases: Oil Reservoirs: No Initial Gas Cap, No Water Influx

Oil Reservoirs: Initial Gas cap of Unknown Size, No Water Influx Oil Reservoirs: Water Influx, No Initial Gas Cap Oil Reservoirs: Water Influx, Initial Gas Cap Nonassociated Gas Reservoirs: Water Influx Nonassociated Gas Reservoirs: No Water Influx
Oil Reservoirs: No Initial Gas Cap, No Water Influx If there is no initial gas cap and no water influx, WpBw is dropped from Equation 4.2, and, after dropping We, Equation 4.5 reduces to (4.6) FpR = NEo A plot of FpR versus Eo should be a straight line passing through the origin, as shown by Figure 1 .

Figure 1

As noted by Havlena and Odeh (1963), "the origin is a must point; thus, one has a fixed point to guide the straight line plot." The first few points may be erratic and may be ignored in constructing the straight line. As indicated by Equation 4.6, the slope of the line is "N" i.e., STB oil initially in place (STBOIP). Oil Reservoirs: Initial Gas cap of Unknown Size, No Water Influx If there was an initial gas cap or water influx, or both, a plot of FpR versus Eo will curve upward. If the reservoir oil is saturated at initial reservoir pressure, even though free gas may not have been encountered in drilling, an initial gas cap might reasonably be suspected. If there is no water influx, Equation 4.5 becomes (4.7) FpR = N[Eo + mBtiEg/Bgi] and a plot of FpR versus [Eo + mBtiEg/Bgi] for the correct assumed "m" value is a straight line passing through the origin. As indicated by Equation 4.7, the slope is "N," STBOIP. As shown by Figure 2 , if the assumed value for "m" is too small, the plot curves upward; if it is too large, downward.

Figure 2

If there is significant water influx, the plot curves upward, regardless of the "m" value used. From examination of Equation 4.7, it may be noted that, if there is no water influx, a plot of FpR/Eo versus Eg/Eo should be linear with the slope equal to mBti/Bgi which equals G and a y intercept equal to N. Havlena and Odeh (1963) observed that the former plot is more powerful, as it must pass through the origin, but that both plots should be used in every case. Oil Reservoirs: Water Influx, No Initial Gas Cap If the oil initially in place is undersaturated, but there is reason to suspect water influx, WpBw is retained in Equation 4.2, and Equation 4.5 can be rearranged and written as follows: FpR/Eo = N + C[pQ(_tD)]/Eo (4.8) where the water influx term, We, has been written in terms of the water influx constant, C, and the van Everdingen and Hurst (1949) solution to the diffusivity equation for unsteady-state dimensionless water influx Q(tD). Use of the Hurst-van Everdingen formulation is not mandatory, and other methods to compute We may be used. (See, for example, Fetkovich 1971.) As shown by Figure 3 , a plot of FpR/Eo versus pQ(tD)/Eo should be a straight line for the correct assumed water influx parameters (tD and re/ri).

Figure 3

N is the y-intercept, and C is the slope. If the water influx term is too small, the plot usually curves upward; if it is too large, downward. (See, however, Havlena and Odeh 1964.) Stewart (1980) discussed an application of this methodology to the Piper Field (North Sea). Reportedly, he was able to obtain an acceptable straight line fit after only six months of production history; the material balance estimate 1.4 billion STB was within 10% of the volumetric estimate. In the presence of uncertainties in reservoir pressure, the Havlena-Odeh (1963) methodology may not be sensitive enough to define a linear trend from which oil initially in place can be determined. In this event, the method proposed by McEwen (1962) may work better. Oil Reservoirs: Water Influx, Initial Gas Cap If there is both an aquifer and a gas cap, Equation 4.5 may be written as

<

(4.9)

If the gas cap and the aquifer have been characterized correctly (correct "m," tD, and re/ri), then a plot of the left-side term in Equation 4.9 versus the second rightside term should be linear, with the y-intercept being N (STBOIP) and the slope being C (the water influx constant). Nonassociated Gas Reservoirs: Water Influx For a nonassociated gas reservoir (or one with a negligibly small oil rim), FpR (from Equation 4.2) may be redefined as FpR = GpBg + WpBw Equation 4.5 then may be written as FpR = GEg + C[pQ(tD)] As shown by Figure 4 , (4.11) (4.10)

Figure 4

if the aquifer is characterized correctly (correct tD and re/ri), a plot of F/Eg versus

[pQ(tD)]/Eg should be a straight line with the y-intercept equal to G, gas initially in place, and slope equal to C. Nonassociated Gas Reservoirs: No Water Influx In this case a plot of GpBg versus Eg should be a straight line through thze origin with slope equal to G. Usually, however, it is easier to use another form of the material balance equation.

Reserves Estimation: Material Balance Methods

Havlena-Odeh Applications
Havlena and Odeh (1963, 1964) have shown that the material balance equation Np[Bt + Bg(Rp - Rsi)] + NpBw = N[(Bt - Bti) + mBti(Bg - Bgi)/Bgi] + We may be written as an equation of a straight line by making the following simplifications: (4.1)

Define FpR as the reservoir volume of oil, gas, and water production:
FpR = Np[Bt + Bg(Rp - Rsi)] + WpBw (4.2) Define Eo as the expansion of a unit volume of the oil (and initially dissolved gas) initially in place: Eo = (Bt - Bti) (4.3) Define Eg as the expansion of the initial gas cap: Eg = (Bg - Bgi) (4.4) Substituting these new terms, Equation 4.1 becomes: FpR = N[Eo + mBtiEg/Bgi] + We (4.5) Depending on whether there is an initial gas cap or water drive, or both, this equation may be used in various forms to determine the amount of oil and gas initially in place and the probable reservoir drive mechanism. To estimate reserves, one of the predictive forms of the material balance equation must be used. In this section, we consider six cases: Oil Reservoirs: No Initial Gas Cap, No Water Influx

Oil Reservoirs: Initial Gas cap of Unknown Size, No Water Influx Oil Reservoirs: Water Influx, No Initial Gas Cap Oil Reservoirs: Water Influx, Initial Gas Cap Nonassociated Gas Reservoirs: Water Influx Nonassociated Gas Reservoirs: No Water Influx
Oil Reservoirs: No Initial Gas Cap, No Water Influx

If there is no initial gas cap and no water influx, WpBw is dropped from Equation 4.2, and, after dropping We, Equation 4.5 reduces to (4.6) FpR = NEo A plot of FpR versus Eo should be a straight line passing through the origin, as shown by Figure 1 .

Figure 1

As noted by Havlena and Odeh (1963), "the origin is a must point; thus, one has a fixed point to guide the straight line plot." The first few points may be erratic and may be ignored in constructing the straight line. As indicated by Equation 4.6, the slope of the line is "N" i.e., STB oil initially in place (STBOIP). Oil Reservoirs: Initial Gas cap of Unknown Size, No Water Influx If there was an initial gas cap or water influx, or both, a plot of FpR versus Eo will curve upward. If the reservoir oil is saturated at initial reservoir pressure, even though free gas may not have been encountered in drilling, an initial gas cap might reasonably be suspected. If there is no water influx, Equation 4.5 becomes (4.7) FpR = N[Eo + mBtiEg/Bgi] and a plot of FpR versus [Eo + mBtiEg/Bgi] for the correct assumed "m" value is a straight line passing through the origin. As indicated by Equation 4.7, the slope is "N," STBOIP. As shown by Figure 2 , if the assumed value for "m" is too small, the plot curves upward; if it is too large,

downward.

Figure 2

If there is significant water influx, the plot curves upward, regardless of the "m" value used. From examination of Equation 4.7, it may be noted that, if there is no water influx, a plot of FpR/Eo versus Eg/Eo should be linear with the slope equal to mBti/Bgi which equals G and a y intercept equal to N. Havlena and Odeh (1963) observed that the former plot is more powerful, as it must pass through the origin, but that both plots should be used in every case. Oil Reservoirs: Water Influx, No Initial Gas Cap If the oil initially in place is undersaturated, but there is reason to suspect water influx, WpBw is retained in Equation 4.2, and Equation 4.5 can be rearranged and written as follows: (4.8) FpR/Eo = N + C[pQ(_tD)]/Eo where the water influx term, We, has been written in terms of the water influx constant, C, and the van Everdingen and Hurst (1949) solution to the diffusivity equation for unsteady-state dimensionless water influx Q(tD). Use of the Hurst-van Everdingen formulation is not mandatory, and other methods to compute We may be used. (See, for example, Fetkovich 1971.) As shown by Figure 3 , a plot of FpR/Eo versus pQ(tD)/Eo should be a straight line for the correct

assumed water influx parameters (tD and re/ri).

Figure 3

N is the y-intercept, and C is the slope. If the water influx term is too small, the plot usually curves upward; if it is too large, downward. (See, however, Havlena and Odeh 1964.) Stewart (1980) discussed an application of this methodology to the Piper Field (North Sea). Reportedly, he was able to obtain an acceptable straight line fit after only six months of production history; the material balance estimate 1.4 billion STB was within 10% of the volumetric estimate. In the presence of uncertainties in reservoir pressure, the Havlena-Odeh (1963) methodology may not be sensitive enough to define a linear trend from which oil initially in place can be determined. In this event, the method proposed by McEwen (1962) may work better. Oil Reservoirs: Water Influx, Initial Gas Cap If there is both an aquifer and a gas cap, Equation 4.5 may be written as

<

(4.9)

If the gas cap and the aquifer have been characterized correctly (correct "m," tD, and re/ri), then a plot of the left-side term in Equation 4.9 versus the second rightside term should be linear, with the y-intercept being N (STBOIP) and the slope being C (the water influx constant). Nonassociated Gas Reservoirs: Water Influx For a nonassociated gas reservoir (or one with a negligibly small oil rim), FpR (from Equation 4.2) may be redefined as FpR = GpBg + WpBw Equation 4.5 then may be written as FpR = GEg + C[pQ(tD)] As shown by Figure 4 , (4.11) (4.10)

Figure 4

if the aquifer is characterized correctly (correct tD and re/ri), a plot of F/Eg versus [pQ(tD)]/Eg should be a straight line with the y-intercept equal to G, gas initially in place, and slope equal to C. Nonassociated Gas Reservoirs: No Water Influx In this case a plot of GpBg versus Eg should be a straight line through thze origin with slope equal to G. Usually, however, it is easier to use another form of the material balance equation.

Reserves Estimation: Material Balance Methods

Prediction Methods
The Havlena-Odeh (1963) methodology provides an estimate of oil and gas initially in place and an indication of the probable drive mechanism. To estimate reserves, this methodology must be followed by one of the predictive forms of the material balance equation. Two such methods are available for desktop calculation; one is attributed to Tarner (1944); the other, to Muskat (1945). These predictive methods, however, are limited to solution gas drive reservoirs (no gas cap no water influx no gravity segregation) and generally have been replaced by computer simulation. The development of these two classical methods, however, is presented for completeness.

Tarners Method
Tarners method (1944) involves a series of stepwise calculations to estimate future gas and oil production from a dissolved gas drive oil reservoir (no gas cap no water influx). Each stepwise calculation involves a trial and error solution of the instantaneous producing gas/oil ratio (GOR) equation and the material balance equation. The material balance equation for a depletion drive oil reservoir with no initial gas cap is

<

(4.12)

and the following substitutions are made: Bt = Bo + (Rsi - Rs)Bg Bti = Boi Gp = NpRp (4.15) (4.13) (4.14)

Equation 4.12 then becomes

<

(4.16)

Tracy (1955) developed the following equations, which are functions of pressure:

<

(4.17)

<

(4.18)

Substituting these pressure functions and setting N equal to 1, Equation 4.16 becomes: N = Np
n

+ Gp g = 1

(4.19)

Tarners method involves the following steps:

a slightly lower pressure, pk

1. Starting from the current reservoir pressure, pj

, estimate the instantaneous producing GOR, Rk, at

2. The average instantaneous producing GOR between pj and pk is Rav = (pj + pk)/2 (4.20) 3.To determine the average oil production over this pressure decrement, Equation 4.19 is written: 1 = (Npj +Np) nk + (Gpj + RavNp) gk (4.21) 4.Solving the above equation for Np, Npk becomes Npk = Npj + Np (4.22) 5.The reservoir oil saturation at Pk is: Sok = (1 - Npk) (1 - Sw)Bo/Boi (4.23) 6.At Sok, read kg/ko from the gas/oil relative permeability ratio curve applicable to this reservoir and determine Rk from the instantaneous GOR equation: Rk = Bokgo/Bgkog + Rsk (4.24) 7.If this value of Rk agrees (to within about 1%) with the value of Rk estimated at the beginning of this method, repeat the procedure for the next pressure decrement, otherwise, make a new estimate of Rk and repeat the calculations for this pressure decrement. Originally designed for desktop calculation, Tarners method has been superceded by Muskats method, which is amenable to computer solution.

Nuskats Method
The following development of Muskats method is taken from Craft and Hawkins (1959). Stock tank barrels of oil remaining in the reservoir at any pressure are

< (4.25) where Vp is reservoir pore volume in barrels. Differentiating with respect to pressure, Equation 4.25 becomes

(4.26) < Gas remaining in the reservoir, both free and dissolved, is Gr = Vp [RsSo/Bo + (1 - So - Sw)/Bg] (4.27) Differentiating with respect to pressure, Equation 4.27 becomes

<

(4.28)

The current instantaneous GOR can be written as

(4.29) R = Gp/Np - < Substituting Equation 4.26 and 4.28, Equation 4.29 becomes

< The instantaneous producing GOR may also be written:

(4.30)

<

+ Rs

(4.31)

Equating Equation 4.30 and 4.31, and solving for dSo/dp:

<

(4.32)

To simplify handling Equation 4.32, the terms that are functions of pressure are grouped and defined as follows:

<

(4.33)

<

(4.34)

<

(4.35)

Equation 4.32 may be written in incremental form, using the groups defined above:

< Use these equations as follows:

(4.36)

1.Evaluate the derivatives d/dp of Rs, Bo, and (1/Bg )at the midpoint of the selected pressure decrement to ensure stability in solving Equation 4.36, use of numerical methods is recommended. 2.Solve the pressure functions Xp, Yp, and Zp at the midpoint of the selected pressure decrement. 3.Calculate So from Equation 4.36. 4.Evaluate kg/ko at the midpoint of the saturation decrement and calculate the instantaneous GOR from Equation 4.30. 5.Determine Np =SoN/Soi. 6.Select a new pressure decrement and repeat. Muskats method does not require trial calculations and may be programmed for computer solution. It is not self-checking like Tarners method, however, and small pressure decrements should be used, otherwise large material balance errors will occur.

Reservoir Simulation
There are several scenarios that preclude reliable use of a tank model, or zerodimensional material balance, to estimate oil and gas initially in place. Provided necessary and sufficient data are available, multidimensional reservoir simulation would be preferable for the following situations:

A really extensive reservoirs with low values of, or large areal variance in, kh/m. These conditions often lead to large areal variations in bottomhole pressure that are difficult to "average" for use in a tank model. A really extensive reservoirs at different stages of development
and/or production where there are large areal variations in reservoir pressure that may range from initial pressure to below the bubblepoint pressure. These conditions are impossible to handle with a tank model.

Very heterogeneous reservoirs with large spatial variations in permeability due to any combination of (a) natural fractures, (b) interbedding, (c) variations in depositional environment, or (d) diagenesis. Oil reservoirs with substantial vertical relief, especially those with
large gas caps or significant vertical variations in oil properties. In addition, reservoir simulation has become an almost indispensable aid for the development planning of major oil and gas fields. As discussed by Kingston and Niko (1975), extensive reservoir simulation studies were used to assist in the selection of the operating scheme for the Brent Field (North Sea). Similar applications have been reported for other North Sea fields by McMichael (1978) for the Statfjord Field, by Hillier et al. (1978) for the Forties Field, by Valenti and Buckles (1986) for the Fulmar Field, and by Till et al. (1986) for the Heidrun Field.

Uncertainties
All methods to estimate reserves are subject to varying degrees of uncertainty, with the nature and degree of uncertainty depending on, among other factors, the method. Material balance methods including multidimensional reservoir simulation utilize reservoir engineering equations to calculate future production (reserves). The reliability of material balance calculations depends on the accuracy with which the mathematical model and the available data simulate the reservoir under study. The reservoir "parameter" usually subject to the largest uncertainty and usually having the most influence on recovery efficiency (aside from economics) is reservoir heterogeneity. Reservoir drive mechanism also is important, but by the time there are sufficient performance data for a material balance estimate of oil and gas initially in place, the drive mechanism usually has been identified with a reasonable degree of certainty. Reservoir heterogeneity is manifested in material balance predictions in several ways:

Reservoir heterogeneity affects the relative permeability terms, which control the relative production rate of the driving and the driven fluids and the abandonment saturations of oil and gas.

Reservoir heterogeneity permeability stratification or natural fractures, for example usually makes it difficult to interpret bottomhole pressure surveys. Errors in bottomhole pressure (both measurement and interpretation) are among the major contributors to errors in material balance calculations. In multidimensional simulation, reservoir heterogeneity is manifested in the permeability level of each grid block, which controls the cumulative calculated recovery from the block at abandonment.
In addition, uncertainties may be caused by Insufficient history for a reliable material balance. In the case of a water drive reservoir with a limited aquifer, for example, if the pressure sink has not reached a boundary during the "history match" period, calculations may be based, mistakenly, on an infinite aquifer.

Actual relative permeability relations different from those used. This may occur if, for example, calculations of future production are made in an edge water drive reservoir in which the kw/ko relations had been matched in only a few downdip wells.

Reserves Estimation: Material Balance Methods

Volumetric Gas Reservoirs For gas reservoirs where the volume of the initial hydrocarbon pore space remains constant over life, the material balance equation may be stated as Gas production = Expansion of gas initially in place or GpBg = G[Bg - Bgi] (4.37) The gas formation volume factor, Bg, may be expressed in terms of the real gas law: (4.38) < Using the real gas law for Bg and Bgi, Equation 4.37 may be written:

< solving for p/Z:

(4.39)

<
which is in the form: y = b - mx

(4.40) (4.41)

Thus, for a volumetric gas reservoir, a plot of p/Z versus Gp will be linear with the y intercept (at Gp = 0) equal to pi/zi and the x-intercept (at p/z = 0) equal to G, as shown by Figure 1 .

Figure 1

The term Gp is the "gas equivalent" of produced gas, condensate, and water of vaporization. Both Craft and Hawkins (1959) and Smith (1962) have provided tables to estimate the gas equivalent of condensate for a range of condensate gravities. A stock tank barrel of 45 API condensate, for example, is "equivalent" to approximately 680 standard cubic feet of gas: 65 API, approximately 900 standard cubic feet. From these data, it is apparent that, for rich gas condensates, failure to include the gas equivalent of produced condensate in the term could result in significant errors in estimates of gas reserves. The vapor equivalent of fresh water is approximately 7390 standard cubic feet per barrel (Craft and Hawkins 1959). Water of vaporization, however, generally is only a few barrels per million cubic feet of gas. Thus, ignoring this correction would result in a negligible error. Moderate to High Permeability Reservoirs

In moderate to high permeability gas reservoirs, where there generally is good pressure communication between wells, p/Z versus Gp plots should be constructed for each reservoir being analyzed, using reservoir average static bottomhole pressure and aggregate reservoir production. Dietz (1965) and Matthews et al. (1954) have provided guidelines for estimating average bottomhole pressure in volumetric reservoirs. In gas reservoirs where there is good interwell pressure communication p/Z versus Gp plots for individual wells may deviate from linearity and lead to erroneous conclusions about reservoir drive mechanism and well drainage volume. It has been shown that, under semisteady state conditions, the drainage volume of each well in a multiwell reservoir is proportional to its production rate (Matthews et al. 1954). Under such conditions in a gas reservoir, the p/Z versus plot for each well should indicate the gas volume being drained by that well. If, after a period of semisteady state production, the production rate of a well is increased relative to the production rate of other wells in the same reservoir, the drainage volume of that well will increase relative to the drainage volume of other wells in the reservoir. Under these conditions, the p/Z versus Gp plot of the well deviates from the previously established trend, indicating a larger drainage volume, proportional to the new rate (Stewart 1966). Under the new semisteady state conditions, the individual p/Z versus Gp plots of the other wells in the reservoir also deviate from previously established trends, indicating proportionally smaller drainage volumes. Initial reserves for a volumetric reservoir may be estimated by extrapolating the p/Z versus Gp plot to pa/za, where pa is the estimated average reservoir abandonment pressure at the economic limit. In some cases, it may be necessary to estimate reserves for individual wells in a large reservoir, in which case p/Z versus Gp plots would be needed for individual wells. As a check on reserves estimated using this procedure, it may be desirable to use the back pressure test curves to estimate the abandonment bottomhole pressure for each well at the economic limit producing rate for the field. Water Influx If water influx is suspected, caution should be exercised in using p/Z versus Gp plots to estimate gas initially in place and reserves. Bruns et al. (1965) have shown that an apparently linear plot of p/Z versus Gp does not necessarily mean that there is no water influx. Their work showed it is theoretically possible for plots of p/Z versus to vary from concave upward to concave downward, depending on the size and permeability of the aquifer. Apparently linear plots could be generated by limited aquifers; these plots exhibited small inflections at the origin which often are ignored. Chierici et al. (1967) corroborated the Bruns et al. (1965) work with examples of several gas reservoirs in the Po Valley (Italy). These reservoirs were being invaded by water, but exhibited apparently linear plots of p/Z versus Gp, as revealed in subsequent discussion (Hurst 1967). Gruy and Crichton (1950) advocated plotting cumulative pressure (p/Z) drop versus cumulative gas on log-log paper to determine if there is water influx into, or drainage from, a gas well drainage area.

Low-Permeability Reservoirs In low-permeability gas reservoirs developed on wide spacing, where there usually is no apparent pressure communication between wells, it is advisable to maintain a p/Z versus Gp plot for each well. It is good practice to use the back pressure test curve for each well to estimate the abandonment bottomhole pressure at the economic limit flow rate. Estimates of gas initially in place (GIP) for each well not reserves should be made using volumetric methods, assuming a drainage area equal to the well spacing. These estimates should be compared with estimates of GIP made by extrapolating the p/Z versus Gp plot to zero pressure. In many cases it may be observed that the volumetric estimate of GIP is substantially larger than the p/Z versus G estimate. In such cases, in-fill drilling may be warranted to increase recovery efficiency. If infill drilling can be justified, some portion of the difference between GIP estimated volumetrically and that estimated from the p/Z versus Gp plot might be considered undeveloped reserves, pending successful infill drilling. In many cases, the transmissibility of the reservoir may be too low to determine bottomhole pressure after a reasonable shut-in period, in which case reliable p/Z versus Gp plots cannot be made.

Performance/Decline Trend Analysis

After a field, reservoir, or well has been on sustained production long enough for the producing characteristics to develop clearly defined trends, it may be possible to extrapolate these trends to the economic limit to estimate reserves. Two stages, or periods, may be recognized in performance/ decline trend analysis:

The period before there is a decline in production of the principal product oil in oil wells or gas in gas wells. There are trends in "performance indicators" such as gas/oil ratio, water/oil ratio, etc., that may be extrapolated to economic limit conditions. The period during which the production of the principal product
develops a decline trend that can be extrapolated to the economic limit production rate. There are, however, several cautions regarding this methodology.

Reservoir Performance versus Wellbore/Mechanical Problems

If performance/decline trends in wells are used to estimate reserves, the trends should be related to reservoir drive mechanism. Otherwise, extrapolation of these trends may lead to estimates of reserves that have no relationship to reservoir performance. In pumping oil wells, for example, where reservoir pressure is being maintained by water injection, the rate of total fluid production (oil plus water) should remain approximately constant. If there is a continuing decrease in the rate of total fluid

production from such a well, gradual wellbore plugging or pump wear should be suspected. In a gas reservoir producing by pressure depletion, for example, it would be reasonable to expect a trend of declining shut-in tubing pressures versus time or versus cumulative gas. If the wells are being produced at approximately constant rates, it also would be reasonable to expect a trend of declining flowing tubing pressures versus time or versus cumulative gas. If the wells were being produced against a constant back pressure, imposed by sales line or compressor intake pressure, then it would be reasonable to expect a trend of declining production rates versus time or versus cumulative gas. In contrast, if reservoir pressure is being maintained by strong water drive, declining trends in wellhead pressure or production rate might be indicative of wellbore plugging or buildup of liquids in the wellbore. Lease Curves versus Individual Well Curves Frequently, production data on individual wells are not available, and reserve estimates must be made using production data from leases that contain more than one well. On multiwell leases, there may be several different reservoirs being produced; some may be gas, others may be oil. Each reservoir may have a different drive mechanism and each may be at a different stage of depletion. Remedial operations, drilling new wells, and modifications to production equipment may influence the production history. Before attempting to estimate reserves from production trends on large, multiwell leases, the engineer should become familiar with past, current, and anticipated future operations. Routine lease operations that affect well performance may be reported only informally, if at all. Frequently, pumpers or lease fore-men maintain records of well tests, pressures, and equipment changes that are not generally available elsewhere. Depending on circumstances, the engineer responsible for reserve estimates should consider an onsite visit with field personnel to determine the nature of field operations and the possible influence of those operations on production rate. Brons (1963) has shown that, on multiwell leases the decline characteristics of the high rate wells dominate the aggregate lease decline. For example, on leases where the high rate wells are declining more rapidly than the low rate wells which is the usual case in an aggregate of depletion-drive reservoirs extrapolation of the aggregate lease trend might lead to a pessimistic estimate of reserves. Conversely, on leases where the high-rate wells are declining less rapidly than the low-rate wells (which might be observed if there were water-drive reservoirs), extrapolation of the aggregate lease trend might lead to an optimistic estimate of reserves. It may be necessary to use well-test data to construct individual well decline curves to help interpret aggregate lease production trends where there are large variations in rate between wells on the lease. On leases where development drilling is ongoing, a meaningful trend may be observed from a plot of barrels of oil (or cubic feet of gas) per month per producing well. Curtailed Wells or Leases

Production of oil or gas from leases or wells may be curtailed for several reasons, including pipeline limitations, limited market, or inability to handle all produced water. Thus, before attempting to analyze production trends, the engineer should determine whether the lease or well in question has been curtailed. Economic Limit The term economic limit may refer to a single well or to a production facility involving many wells with processing equipment that must be operated as a unit. Examples of the latter include offshore production platforms and fluid injection projects. The economic limit generally is defined as the production rate at which the total net revenue attributable to production from the well or facility equals the "out-of-pocket" cost to operate the well or facility. (State and Federal income tax and corporate overhead generally are excluded in determining these costs.) Net revenue is gross revenue less production taxes, ad valorem taxes, royalties, and transportation and treating expense, if any. Out-of-pocket costs are costs that would be saved if the well or facility is shut-in, such as power and materials. Out-of-pocket costs would include labor only if that cost would be saved by shutting in the well or facility. Performance Estimates Compared to Volumetric Estimates For each well or reservoir, ultimate recovery estimated using volumetric methods should be compared to ultimate recovery estimated using performance methods. If there is a significant difference between the two estimates, assuming each method was applied correctly, the following possibilities should be considered:

If the ultimate recovery estimated using volumetric methods is significantly greater than that estimated using performance methods: The performance data may be representative of only the transient period (when the pressure sink around wells is still expanding) and not of semi steady state conditions (when the drainage area has stabilized), which usually is more representative of well performance. The wells may need to be stimulated or equipped with highercapacity equipment. Infilling wells may be needed. Pending determination of the reason for the difference in ultimate
recovery estimated using the two methods, only those reserves demonstrated recoverable from performance methods should be classified as "proved."

If the ultimate recovery estimated using volumetric methods is significantly less then that estimated using performance methods:

 The volumetric parameters may not be representative of the reservoir. The mapped area or volume may be too small, and additional development may be warranted.
Reservoir heterogeneities often have an adverse effect on the recovery of oil and gas. A significant difference between ultimate recovery estimated using volumetric methods and that estimated using performance methods may be an indication of inefficient drainage caused by reservoir heterogeneities. Continuous monitoring of well performance is an essential element of good reservoir management. Performance Trends Frequently, one or more of the "performance indicators" of a well or reservoir exhibits a trend before the production rate of the principal product begins to decline. Depending on reservoir type and drive mechanism, these performance indicators include

water/oil ratio (WOR) water/gas ratio (WGR) gas/oil ratio (GOR) condensate/gas ratio (CGR) bottomhole pressure (BHP) flowing tubing pressure (FTP) shut-in tubing pressure (SITP).

Oil Reservoirs
Typical performance indicators for solution gas drive and water drive (or waterflood) oil reservoirs are discussed below. Solution Gas Drive In a solution gas drive reservoir, provided gravity segregation is not significant, the logarithm of cumulative gas production from the reservoir often becomes an approximately linear function of cumulative reservoir oil production (Np) as the reservoir enters an advanced stage of depletion. In a fully developed reservoir, the solution gas initially in place may be estimated with reasonable confidence by volumetric methods. Extrapolation of the trend of the logarithm of cumulative gas production (Gp) to the logarithm of solution gas initially in place (Gsi) should provide

a reasonable estimate of maximum ultimate oil recovery (Npu), as shown in Figure 1 .

Figure 1

This extrapolation is based on the assumption that the abandonment reservoir pressure can be reduced to very nearly atmospheric pressure by pumping. Under these conditions, gas remaining in the reservoir would be a very small fraction of the gas in the reservoir at initial conditions; extrapolation to the logarithm of Gsi rather than to a value slightly less would introduce negligible error. Water Drive Several methods have been used to estimate oil reserves from wells in water drive (or water flood) reservoirs. One of the most widely used procedures involves plotting fractional oil flow (fo) versus cumulative oil (Np). This and other methods are discussed briefly in the following paragraphs. fo versus Np In water drive (or waterflood) reservoirs, after breakthrough of water in individual wells the logarithm of fo frequently will be an approximately linear function

of Np for the well ( Figure 2 ). Usually, the trend may be extrapolated to an "economic limit," fo, to estimate reserves.

Figure 2

Brons (1963), however, has noted that in many cases the apparently linear trend of fo versus Np turns down as fo approaches small values. The downturn reportedly occurs at smaller values of fo for heavy (viscous) oils than for light oils ( Figure 3 ).

Figure 3

In bottomwater drive reservoirs, if the completion interval extends across a "shale break" the decreasing trend in fo may stabilize before resuming a decline ( Figure 4 ).

Figure 4

The rate of decline of fo after the stabilization period may be greater or less than the rate of decline before the stabilization period, as discussed below. After the zone below the shale break has watered out ( Figure 5 ),

Figure 5

the decline in fo usually stabilizes until the water encroaches around the end of the shale break (period "b-c" in Figure 4 ). During period "b-c" the fractional flow of oil (fo) should be approximately:

< where: ko/o = mobility of oil in the upper zone

h1 = net thickness of upper zone

(5.1)

kw /w = mobility of water in the lower zone h2 = net thickness of the lower zone If h1 > h2, the rate of decline in fo after the stabilization period usually is less than the rate of decline before the stabilization period, and vice versa.

Figure 6 ), after breakthrough of water the logarithm of total liquid production rate tends to increase linearly with time.
For wells being produced at an approximately constant oil rate (

Figure 6

This behavior is typical for oil wells in water drive reservoirs being produced by gas lift. If the oil production rate is not approximately constant, which may be the case if wells are being produced with rod pumps, the logarithm of total liquid production rate will tend to increase linearly versus cumulative oil. Oil reserves may be estimated by extrapolating the total liquid production rate to the maximum two phase flow capacity of the tubing (or to pump capacity). Generally, once this capacity has been reached, the oil production rate decreases at a rate commensurate with the rate of increase in total liquid production. Plots like Figures 2 and 6 should be prepared for oil wells producing significant water. Estimates of reserves using each of the two plots may differ slightly. Other Methods Subject to the simplifying assumption that the log of (krw/kro) is a linear function of Sw, Ershagi (1984) proposed that, for water drive or waterflood wells, a plot of (-){ln[(l/fw - 1)] - 1/fw} versus {fractional oil recovery} should be linear. In subsequent discussions (Startzman and Wu 1984), it was observed that this procedure should be compared with the "conventional" plot of log (WOR) versus Np, which should yield better results at low water cuts than the Ershagi and Abdassah (1984) method. In continuing discussion of the ErshagiAbdassah paper, Liu (1989) proposed plotting: {ln(Np/Wp) - Np/Wp} versus Np

From these papers, it seems apparent that developing the "correct" performance plot to estimate reserves in water drive or waterflood reservoirs is highly empirical. Accordingly, different types of plots should be considered, with the "best" plot being determined by regression analysis. It should be kept in mind, however, that the "best" plot at low water cuts might not be the "best" plot at high water cuts. Timmerman (1982) has provided additional discussion of the analysis of performance trends in waterflood reservoirs. Gas Reservoirs Performance indicators are discussed briefly for pressure depletion reservoirs and for water drive reservoirs. Pressure Depletion In nonretrograde gas reservoirs with moderate to high permeability, periodic shut-in tubing pressure (SITP) data on individual wells can be used to estimate shut-in bottomhole pressure which can be used to develop p/Z versus Gp plots to estimate reserves. In cases where it is not practical to obtain periodic SITP data, consideration might be given to using a procedure proposed by Lewis (1985), which involves using adjusted performance data from individual well back-pressure curves. Over life, a plot of SITP versus Gp is slightly concave up. However, late in life, the trend is approximately linear. Under constant terminal rate conditions a plot of flowing tubing pressure (FTP) should be approximately parallel to the SITP plot. After it is adjusted downward to approximate the economic limit flow rate, the FTP plot may be extrapolated to sales line or compressor intake pressure to estimate reserves. If back pressure test data representative of well performance approaching economic limit conditions are available, these data also may be used to estimate reserves. Water Drive In gas reservoirs subject to water influx, after water breakthrough in individual wells the logarithm of the producing water/gas ratio (WGR) usually is a linear function of cumulative gas from the well. Typically in such wells, the abandonment point is governed by the two-phase flow characteristics of the tubing string, and wells usually become incapable of flow before they reach the economic limit gas flow rate. In this case, reserves may be estimated by extrapolating the trend of flowing tubing pressure versus cumulative gas to the sales line pressure. In some areas, however, flow from such wells may be maintained with multistage compression. The abandonment point usually is the economic limit flow rate, which is a function of among other factors the costs of compressor operation and handling produced water. In this case, each well may have a different abandonment WGR ( Figure 7 ).

Figure 7

Reserves for each well may be determined by extrapolating a plot of q versus WGR for the well to the economic limit WGR for the facility, with the WGR for the facility generally having a trend ( Figure 7 ). (A similar procedure to determine the economic limit may be appropriate for water drive oil wells.) Production Decline Equations Background Lewis and Beal (1918) and Cutler (1924) were among the earliest to publish on the analysis of production decline curves in oil wells. As noted by Arps (1945), three types of decline curves are commonly used: hyperbolic, harmonic, and exponential. The harmonic and exponential curves are special cases of the hyperbolic curve. With certain exceptions, as discussed by Lefkovits and Matthews (1958), Fetkovich (1980), Gentry and McCray (1978), Cummings and Gentry (1986), and others, decline curves are empirical fits to observed trends of production. Unless the observed trends are related to reservoir mechanics, extrapolation of these trends to estimate reserves may lead to significant error.

Hyperbolic Declines In the following sections, the development follows Arps (1945); the notation, however, has been modified to reflect SPE (1987) standard. Mathematics Arps (1945) proposed a hyperbolic equation of the form:

< (5.2) where a is the nominal (or instantaneous) decline rate and is defined as the negative slope of the natural logarithm of the production rate versus time, or: a = (-) d(ln q)/dt (5.3) and where C is a constant which is defined under initial conditions as: C = ai/qib (5.4) The term b is called the hyperbolic decline exponent. This notation is not consistent with SPE (1987) "standard," which defines h as the hyperbolic decline exponent. Note that b = 1/h. This notation is, however, consistent with the original usage by Arps (1945) and with recent usage by Gentry and McCray (1978), and others. Note that b is equivalent to n, which was used in the Petroleum Production Handbook (SPE 1962).
After integration, Equation 5.2 becomes: q = qi(l + bait)-1/b (5.5) which is the rate-time equation for hyperbolic declines. After a second integration with respect to time, Equation 5.5 becomes: Q = qib[qi(1-b) - q(1-b)]/ai[1 - b] (5.6) which is the rate-cumulative equation for hyperbolic declines; Q is a general term to denote cumulative production, which may be either Np or Gp. Equation 5.5 may be solved to determine the time for the production rate to decline from q1 to the economic limit rate qe1:

<
Applications

(5.7)

Lefkovits and Matthews (1958) observed that, in theory, in a homogeneous oil reservoir producing by gravity drainage, wells with a gas cap should exhibit a hyperbolic decline with b = 0.5; wells without a gas cap should exhibit an exponential decline (b = 0). In investigations of actual reservoirs producing by gravity drainage they observed wells exhibited hyperbolic declines with b ranging from less than 0.1 to 2.5. Russell et al. (1966) observed that, in theory, gas wells producing early in life under constant terminal pressure conditions and cylindrical flow should exhibit hyperbolic production declines with b equal to approximately 0.5.

Stewart (1970) observed that, during the semisteady state period, gas wells completed in the very heterogenous, low permeability approximately 0.5 millidarcy Cretaceous sandstones in the Green River basin of the western United States exhibited hyperbolic declines with widely varying exponents. In his original paper, Arps (1945) observed that "most decline curves seem to be characterized by b values between 0 and 1, with the majority between 0 and 0.4." Since that time, however, b values greater than 1.0 have been reported in many reservoirs. These are reservoirs with low permeability, or low-permeability reservoirs that are naturally or hydraulically fractured (Bailey 1982; Long and Davis 1988). In addition, Gentry and Mccray (1978) have demonstrated mathematically that b should be expected to be larger than 1 in very heterogeneous reservoirs. Fetkovich (1980) observed that "rate data existing only in the transient period ... would require values of b much greater than 1 to fit the data." Exponential Declines Because of relative ease of use, extrapolation of the exponential, or constant percentage, decline curve is the most widely used technique to estimate reserves using the performance method. As noted in the applications section, however, this technique may lead to significant underestimating of reserves if applied using data measured during transient flow conditions in fractured or low-Permeability reservoirs. Examples include the Cotton Valley Formation in East Texas and the Green River Formation in the Piceance Basin of Colorado. Mathematics The exponential decline curve is a special case of the hyperbolic decline curve, where b = 0. In this case, Equation 5.2 becomes: dq/dt

< (5.8) after integration, this becomes: q = qie-at (5.9) which is the rate-time equation for exponential declines. After a second integration, Equation 5.9 becomes: < (5.10) which is the rate-cumulative equation for exponential declines.
Equation 5.9 may be solved to determine the time for the production rate to decline from q1 to the economic limit rate qel: tel = ln(ql/qel)/a (5.11) Arps (1945) noted that the nominal dedline rate a is often confused with the effective decline rate D, which he defined as

<
For exponential declines, a = (-)ln(l - D). Applications Transient flow in fractured or very heterogeneous, stratified reservoirs usually is characterized initially by a relatively steep decline in production rate versus time. These early decline trends may appear to be exponential; extrapolation of these trends to estimate reserves may result in significant underestimates. Later in life, however, production trends from such wells may exhibit decreasing decline rates, becoming exponential with relatively low a values compared to initial a values. Brons (1963) and Stewart (1970), for example, have observed that the gas flow rate from gas wells producing late in life at constant terminal pressure conditions generally exhibits an exponential decline. Also, oil wells producing at constant pressure by fluid expansion above the bubble-point pressure tend to exhibit an exponential decline (Brons 1963). Below the bubble-point, however, such wells tend to exhibit hyperbolic declines. Blasingame and Lee (1986) have presented a method, based on exponential declines, to determine well drainage area and shape factor for homogeneous and naturally or hydraulically fractured reservoirs that are producing with constant bottom-hole pressure. Harmonic Declines The harmonic decline is a special case of the hyperbolic decline where the decline rate is proportional to the production rate (b = 1). Mathematics When b = 1, Equation 5.2 becomes:

<

(5.12)

in which the decline rate is proportional to the production rate. The term C is defined under initial conditions as: C = ai/qi (5.13)

After integration, Equation 5.12 becomes:

< (5.14) which is the rate-time equation for harmonic declines. After a second integration, Equation 5.14 becomes Q = qiln(qi/q)/ai (5.15)

which is the rate-cumulative equation for harmonic declines. Equation 5.14 may be solved to determine the time for the production rate to decline from q1 to the economic limit rate tel = (qi/qel - 1)/a1 Applications It has been observed that, after breakthrough of water in wells producing from water drive oil reservoirs, the logarithm of fovs Np for individual wells tends to be approximately linearwith limitations as discussed. From examination of Equation 5.15, it may be observed that Figure 2 represents a harmonic decline. (5.16)

Methodology To estimate reserves, given a declining trend in gas or oil production rate, the following general procedure is recommended: 1. Examine the historical data, rejecting apparently "anomalous" values, e.g., production data during extended periods when wells were off production for mechanical reasons, seasonal market curtailment, etc. 2. "Fit" the remaining data with the rate-time equation that results in the minimum standard deviation. Use caution with hyperbolic or harmonic fits, because the calculated decline rate may become unrealistically low and result in projection of reserves that exceeds a reasonable volumetric estimate. It is good practice to set a minimum decline rate, consistent with experience in the area, and project production using an exponential curve after this minimum decline rate has been reached.* *Fetkovich (1980) has defined "decline curve dimensionless time" as tD = ait, and demonstrated that "any data existing before a tD of 0.3 will appear to be an exponential decline regardless of the true value of b and, thus, plot as a straight line on semilog paper. A statistical or leastsquares approach could calculate any value of b between 0 and 1." 3. Estimate the economic limit producing rate, qel, for the well or facility being evaluated. 4. Calculate reserves to be produced as the production rate trend declines from q1, the rate at the effective date of the evaluation, to qel, the economic limit. For additional details, please refer to Thompson and Wright (1984) In many areas, special methods have been developed to estimate reserves from analysis of production trends using plots other than rate-versus-time. For example, Gurley (1963) discussed a method to analyze Clinton Sand gas wells (Ohio) by

plotting "percent of best month" capacity versus time. Stricht and Gordon (1983) developed special procedures to analyze the performance of fractured, low permeability gas wells in the Piceance Basin (United States). In cases of extended or periodic shut-in, it may be preferable to analyze the trend of rate versus cumulative, rather than rate versus time. Type Curve Analysis With few exceptions, which were noted previously, decline curve analysis has been considered to be an empirical fit to observed data. In 1973, however, Fetkovich (1980) demonstrated that decline curves can be related to reservoir properties and drive mechanism. Fetkovich (1980) developed a methodology, called advanced decline curve analysis, which is similar to type curve analysis used to analyze bottomhole pressure transient data. The method involves plotting production rate versus time on log-log tracing paper with the same scales as Fetkovichs type curves, then overlaying this plot to determine the best fit to one of the type curves. Examples of the application of this method have been discussed by Fetkovich et al. (1985, 1987). Mannon and Porter (1989), in discussing application of a computer program to implement Fetkovichs method, noted that the ability to discern transient and depletion conditions is one of the strengths of this approach. Using a technique similar to that of Fetkovich (1980), Da Prat et al. (1981), presented type curves that may be used to analyze production trends from wells in dual porosity (naturally fractured) reservoirs for both finite and infinite systems. Their analysis was based on the Warren and Root (1963) model for fractured reservoirs. They observed that "he initial decline in production rate often is not representative of the final state of depletion ... type curve matching based only on the initial decline can lead to erroneous values for ... (the drainage volume) ... if the system is considered homogeneous." In addition, they noted that both dimensionless production (qD) and dimensionless time (tD) are controlled by the dimensionless fracture storage (w) and by the ratio of matrix to fracture permeability (1); thus, a different type curve would be required for different values of these parameters. Accordingly, they recommended that w and 1 be estimated using transient pressure analysis so that the correct type curve (qD versus tD) could be used to estimate reserves.

Special Problems in Estimating Reserves

Reserves Estimation involves special considerations under certain conditions, including those that involve

Remote/Frontier Areas and Harsh Operating Conditions Heavy and Extra Heavy Crude Oil Fractured/Vugular Reservoirs

 Geopressured Reservoirs Partial Water Drive (Rate Sensitive) Gas Reservoirs Thin Oil Columns (Between Gas and Water) Tight Gas Reservoirs

Remote/Frontier Areas and Harsh Operating Conditions

Remote/frontier areas are defined as areas that are so isolated from the existing oil field infrastructure that provision of supplies and support services represents major cost and logistical problems. The Timor Sea, off the northwestern coast of Australia, might be considered such an area. Harsh operating conditions are defined as environmental conditions that require extraordinary provisions to ensure safety and comfort of personnel and integrity of equipment. The Canadian Beaufort Sea might be considered such an area. Resource assessment in remote and/or frontier areas poses special problems because of two factors: (a) there usually is a high degree of geologic uncertainty and (b) the economics of developing, producing, processing, and transporting oil and gas to market are highly speculative. Harsh operating conditions impose extraordinary costs which are difficult to quantify during the early assessment phase. Under such conditions, the minimum size of potentially economic accumulations may be so large as to preclude development of all but giant fields. During their exploration in the Canadian Beaufort Sea, for example, Dome Petroleum indicated they considered minimum reserves for a commercial field to be 400 million barrels. Under conditions of substantial uncertainty in accumulation size and exploitation cost, it may be appropriate to use probabilistic rather than deterministic methods to help determine reserve volume, degree of risk, and reserve classification. Such methods have been discussed by Walstrom (1967), Newendorp (1975), and others. Keith (1986), discussing reserve estimating procedures used in the North Sea, proposed a reserve classification system based on results from Monte Carlo analyses of reservoir data. Other than the SPE (1987) definitions, there are no industry-wide standards regarding reserve estimation and classification. Under conditions of substantial uncertainty as noted previously reserves estimated and classified by deterministic methods may differ substantially from reserves estimated and classified by probabilistic methods. It has been observed, for example, that "proved" reserves estimated by deterministic methods are approximately equal to the sum of "proved plus probable" reserves estimated by probabilistic methods. Thus, caution should be exercised in comparing results of the two methodologies. Heavy and Extra Heavy Crude Oil

Background Discovered world resources of heavy and extra heavy crude oil are estimated to be approximately 4600 billion barrels, two thirds of which are in Canada and Venezuela (Briggs et al. 1988). Bitumen and tar sands are excluded from this estimate. Published data on recovery efficiency from this resource by primary drive mechanisms are sparse. Estimated primary recovery efficiency from the Lloydminster area of western Canada-where stock tank gravities range from 13 to 17 API ranges from 3 to 8% (Adams 1982). Estimated primary recovery efficiency from the Orinoco area of Venezuela where stock tank gravities range from 8 to 13 API ranges from 8 to 12* (Martinez 1987) Additional recovery attributable to waterflooding in the Lloydminster area typically does not exceed 1 to 2* (Adams 1982). Comparable data are not available on the Orinoco area. Thermal recovery methods generally steam flooding are required to develop significant reserves from this resource. For 12 mature steam flood projects, Matthews (1984) noted estimated recovery efficiencies were in the range from 39 to 68* of oil initially in place. These projects ranged in depth from 535 to 2600 feet and had reservoir oil viscosities ranging from 85 to 6400 centipoise. Well spacing for these projects generally is on the order of one to two acres per well, substantially less than what might be considered "normal" spacing. Projects were in California, The Netherlands, and Venezuela. Meyer and Mitchell (1987) estimated worldwide ultimate recovery from heavy and extra heavy crude oils to be approximately 476 billion barrels, which is 10* of the Briggs et al. (1988) estimate of the discovered resource initially in place. Fiorillo (1987) estimated ultimate recovery from the Orinico oil belt to be approximately 245 billion barrels, which is approximately 20* of the oil initially in place. Estimating Reserves Difficulties in estimating reserves from these resources may be attributed to (a) problems in characterizing the accumulations, (b) uncertainties in quantifying the response to steam injection, and (c) generally marginal economics. Problems in characterizing these accumulations include the following:

They are associated with relatively shallow reservoirs, usually containing brackish interstitial water, which causes problems in evaluating resistivity logs. Most of the resource is contained in friable sandstones, causing
problems in retrieval and analysis of cores.

Analysis of pressure transient tests is hampered by the long times required to reach semi-steady state in the viscous i.e., low-mobility oil.

 Because of the small density contrast between the reservoir oil and the interstitial water, significant fractions of these accumulations may be in oil-water transition zones.
Uncertainties in quantifying the response to steam injection are caused by the following: Within a single reservoir, the viscosity of the oil usually exhibits substantial spatial variation.

The highly empirical nature of response to steam injection makes it difficult to develop a reliable prediction model.
Marginal economics are caused by:

extra costs for complying with local air pollution requirements high energy needs to generate steam additional costs to handle these bituminous, sulfurous, and asphaltic crudes the frequent presence of heavy metals in the oil, causing problems in refining.

Fractured/Vugular Reservoirs
Background Fractured reservoirs have been observed in most producing provinces of the world, in both sandstone and carbonate reservoirs. Reiss (1980) and others have identified two broad categories of fractured reservoirs: (a) those with a porous matrix and (b) those with a nonporous matrix. In the former category, which is the most common, most of the hydrocarbons are stored in the matrix porosity; the fractures serve as the principal flow conduits. Examples include many of the Iranian fields, Ekofisk (North Sea), and Spraberry (United States). In the latter, less-common category are reservoirs in fractured metamorphic rocks, fractured shales, and fractured cherts. Examples include the Edison field (California), the Big Sandy gas field (Kentucky), and the Santa Maria basin fields (California). When they occur in carbonates, fractures tend to facilitate extensive leaching, which may result in the development of vugular, sometimes karstic, porosity. Examples include the Albion-Scipio trend (Michigan) and the Rospo Mare field (offshore Italy). Reservoir Engineering Aspects This type of accumulation poses formidable difficulties in estimating reserves. These difficulties are attributable to the extreme heterogeneity of the reservoir rock, which causes substantial uncertainties in estimates of hydrocarbons initially in place and

recovery efficiency. Estimates of reserves using volumetric methods are subject to considerable uncertainty. When feasible, such estimates should be compared with observed recovery in analogous reservoirs. The early performance of wells in these reservoirs usually is characterized by a relatively rapid decline in production rate, which is caused by transient pressure behavior. Not until wells have passed through the transient pressure period and settled into semi steady state conditions can reserves be estimated with a reasonable degree of confidence using the performance method. In general, problems are caused by the following:

Boreholes are frequently severely washed out, making log interpretation difficult or impossible. Core recovery frequently is fragmental, at best. Even in good-quality boreholes, detection of fractures and measurement of fracture porosity using logging devices is highly empirical. Well performance frequently is strongly influenced by proximity to major fractures and by completion technique. While transient pressure analysis may provide useful data, to facilitate application of modern interpretation techniques the surveys must be conducted with highly accurate quartz pressure transducers; analysis of results may not provide unique answers. The rate-time performance of wells typically is hyperbolic. While the behavior of an "average" well might be used to estimate reserves, wide variation in performance between wells should be expected.

Geopressured Reservoirs
Geologic Overview The term geopressure, first used by Charles Stuart of the Shell Oil Company, refers to reservoir fluid pressure that exceeds hydrostatic pressure and may approach overburden pressure. Hydrostatic pressure depends on the local fluid gradient and varies from 0.4 to 0.5 psi per foot of depth. Overburden pressure usually is estimated to be about 1.0 psi per foot of depth. Abnormally high reservoir pressures have been attributed to several causes, including tectonic stress (Berry 1973), under-compaction (Dickinson 1953), geochemical and/or diagenetic processes (Parker 1972), and PVT effects in sealed, subsiding gas reservoirs (Stuart and Kozik 1977).

Geopressures in sand-shale sequences in Tertiary basins e.g., the U.S. Gulf Coast generally are attributed to undercompaction of thick sequences of marine shales. Geopressured reservoirs in this type of depositional sequence tend to be geologically complex and exhibit producing mechanisms that are not well understood. Both of these factors may cause considerable uncertainty in reserve estimates. Geologic complexity may cause uncertainty in estimates based on volumetric mapping. Poorly understood producing mechanisms may cause uncertainty in estimates based on pressure-production performance. Geopressured reservoirs in Tertiary sand-shale sequences frequently are associated with substantial faulting and complex stratigraphy, which may make correlation, structural interpretation, and volumetric mapping subject to considerable uncertainty. The resistivity of interstitial water in geopressured sections may approach that of fresh water, which may suppress the spontaneous potential (SP) log. Under these conditions it may be difficult to estimate net pay unless a gamma-ray log also has been run. In addition, the relatively fresh waters frequently encountered in geopressured sections complicate interpretation of resistivity logs, especially in shaley sands. Reservoir Overview There may be a substantial difference between the volumetric estimate and the material balance estimate of oil and gas initially in place in a given reservoir. Given the geologic complexity of the geopressure environment, such differences are quite likely. In view of the high costs usually associated with drilling, completing, and operating geopressured reservoirs, it is especially important to reconcile such differences to ensure cost-effective exploitation of this resource. Frequently, wells completed in geopressured reservoirs produce substantial amounts of water, even though the wells may be considerable distances from any obvious water source. If mechanical problems can be ruled out, the water may be attributable to expulsion of water from shale* zones interbedded with the productive sands. Geopressured Gas Reservoirs If gas production is attributed to gas expansion only, a plot of p/Z versus Gp will be a straight line. Because geologists considered them to be "closed" accumulations, during the early years of exploitation it was assumed that geopressured gas reservoirs would exhibit linear plots of p/Z versus Gp. While this was observed to be true in many cases see, for example, Wallace (1969) and Prasad and Rogers (l987) it soon became apparent it is not universally true. *The American Geological Institute (ed. Gary et al. 1972) defines shale as an "indurated (hardened) ... sedimentary rock formed by the consolidation ... of clay." Because geopressures in Tertiary basins generally are attributed to undercompaction, the term "proto-shale" might be more appropriate. It has been observed that p/Z versus Gp plots for many geopressured reservoirs initially appear to be linear but exhibit downward curvature as reservoir pressure

approaches hydrostatic pressure. Extrapolation of the initial part of the plot may result in an estimate of gas initially in place approximately twice that estimated using volumetric methods. The anomalously low initial slope of the p/Z versus Gp plot has been attributed to several factors, including pore volume compression (Hammerlindl 1971), connate water expansion, and partial water drive. The downward curvature of the p/Z versus Gp plot has been attributed to other factors, including depletion of a limited "shale water" aquifer (Duggan 1972) and rock collapse (Harville and Hawkins 1969). Bruns et al. (1965) have demonstrated that downward curving plots of p/Z versus Gp can be generated by partial water drive gas reservoirs, with the actual shape of the plot depending on the size and the permeability of the aquifer. Production mechanisms in geopressured gas reservoirs may vary from strong water drive to pressure depletion. In general, the producing mechanisms in geopressured gas reservoirs may include

gas expansion compressibility of the reservoir pore volume expansion of the interstitial water water influx due to water expansion from a contiguous aquifer water influx due to dewatering of interbedded "shale."
Any one, or all, of these mechanisms may be active in a given geopressured gas reservoir. There is disagreement regarding the relative importance of these mechanisms, especially compressibility of reservoir pore volume (Bernard 1987) and water influx from interbedded shale* (Wallace 1969, Bourgoyne et al. 1972, Duggan 1972, Chierici et al. 1978). As a result, several different empirical methods have been developed to adjust the p/Z versus Gp plot. As there is no generally accepted methodology, a brief discussion of work to date follows. *Fertl and Timko (1972) reported a case of an "overpressured section in the Louisiana Gulf Coast area" where a replacement well was drilled approximately 50 feet away from a well that had produced for about 22 years. The log in the new well measured increased resistivity in the shale section immediately above and below the producing sand, which was attributed to "decrease in shale porosity because some shale water ... moved into adjacent ... sands." Both Hammerlindl (1971) and Ramagost and Farshad (1981) developed methods to account for pore volume compressibility and expansion of interstitial water. Ramagost and Farshad (1981) proposed the following equation to adjust the p/Z versus Gp plot: p/Z[l -p(cwSw + cp)/(l - Sw)] = pi/zi - Gppi/GZi (6.1) This equation differs from Equation 4.40 by inclusion of a "p/Z adjustment factor," which is the left-side term in brackets. Hammerlindl proposed adjusting the "apparent gas in place" (AGIP) estimated by extrapolation of the initial part of the p/Z versus Gp plot by multiplying AGIP by the ratio: [gas compressibility]/[effective compressibility].

Both methods, however, assume that pore volume compressibility remains constant over the life of the reservoir being evaluated. In addition, there is no provision in these methods to account for possible water encroachment. The assumption of constant pore volume compressibility is contrary to work by Fatt (1958), Van der Knaap (1959), Dobrynin (1962), and others, who reported that pore volume compressibility is an inverse function of net overburden pressure. Net overburden pressure has been defined as: ([external pressure] -[0.85 < internal fluid pressure]) (Fatt 1958). If reservoir fluid pressure is reduced by production, net overburden pressure will increase. Thus, if reservoir fluid pressure decreases over reservoir life, pore volume compressibility will decrease, not remain constant. Bernard (1987) noted several discrepancies in publications discussing drive mechanisms in apparently volumetric, geopressured reservoirs. He proposed use of a catch-all approximation term, C, to account for the effects of rock and water compressibility, a small steady-state acting aquifer, and shale water influx. Incorporating Bernards proposed C in Equation 4.40 would yield: p/z[l - Cp] = pi/zi + Gppi/GZi (6.2) Using performance data from volumetric, geopressured gas reservoirs in the U.S. Gulf Coast in which gas initially in place (GIP) is "known," Bernard (1987) calculated the C factor for each of 13 different reservoirs. Reportedly, C ranged from 1.5 to 46.5 microsip and did not correlate with either depth or initial pressure gradient. The C factor was used to calculate a correction to the initial slope of the p/Z versus Gp plot so extrapolation of the early part of the plot would intersect the "known" GIP on the x-axis. Bernards (1987) correction factor ranged from about 0.96 to about 0.45. The factor appeared to correlate reasonably well with the inverse of "apparent gas initially in place," which was determined by extrapolation of the early part of the p/Z versus Gp plot for each reservoir examined. The correlation, however, is based on only 13 geopressured gas reservoirs in the U.S. Gulf Coast and may not be valid generally. In addition to gas expansion, it seems apparent that pore volume compressibility, expansion of interstitial water, and water influx also may be significant drive mechanisms in geopressured gas reservoirs. A general methodology to analyze these reservoirs has yet to be published. The following methodology is Proposed, which parallels the methodology of Havlena and Odeh (1963). Considering the factors discussed above, Equation 4.1 may be written for a gas reservoir as: GpBg + WpBw = G(Bg - Bgi) + GBgip(cp + cwSw)/(l - Sw) + We define: FpR = GpBg + WpBw (6.4) Eg = Bg - Bgi (6.5) Substituting Equation 6.4 and 6.5 in Equation 6.3 leads to: FpR = GEg + GBgi_p(cp + cwSw)/(l - Sw) + We = G[Eg + Bgi_p(cp + cwSw)/(l - Sw)] + We (6.6) Dividing by the gas expansion and rock-fluid compression term in square brackets: (6.3)

<

(6.7)

If the water influx term and the rock-fluid expansion-compression terms are "correct," a plot of the left-side term versus the second right-side term of Equation 6.7 would be a straight line, with the y-intercept equal to G, the gas initially in place. The water influx term is usually the most difficult term to evaluate, as water influx in a given reservoir may be attributable to expansion from a contiguous aquifer, dewatering of interbedded shales, or both. Conditions favoring shale water influx include high initial pore pressure gradient, considerable interbedding of shale with the gas-bearing sandstone, and a small contiguous aquifer. Opposite conditions would favor aquifer influx. Depending on the size and shape of the contiguous aquifer, We may be calculated using a limited linear aquifer model or a limited cylindrical aquifer model. If shale dewatering is suspected, a limited linear aquifer model may be used. Geopressured gas reservoirs may exhibit retrograde behavior. When reservoir pressure decreases below the dewpoint pressure, liquids condense in the reservoir pore space. This results in a change in the PVT properties of the reservoir hydrocarbon fluid from a single-phase to a two-phase system. If a p/Z versus Gp plot is used to estimate gas reserves from retrograde gas condensate reservoirs, the two-phase Z factor should be used for reservoir pressures below the dewpoint pressure. If expanded material balance calculations are made, the two-phase Bg should be used below the dewpoint pressure. In addition, condensation of liquids in the reservoir pore space may cause development of a severe "skin," which may lead to a rapid decline in well productivity.* In such cases, it may be extremely difficult to estimate average reservoir abandonment pressure and reserves may have to be determined by extrapolation of trends of declining well productivity to the economic limit. Geopressured Oil Reservoirs Oil reservoirs are encountered less frequently than gas reservoirs in the geopressured section and rarely are discussed in the literature. Comments similar to those for geopressured gas reservoirs are made *If reservoir pressure decreases significantly during production, the reservoir rock may be compressed due to the increase in net overburden pressure, as discussed previously. Compression of the reservoir rock may cause a substantial decrease in permeability, which also may be manifested by a severe "skin." regarding drive mechanism in geopressured oil reservoirs. The expanded material balance may be written:

Np[Bt + Bg(Rp - Rsi)] + WpBw = N[(Bt - Bti) + Btip(cp + cwSw)/(1 - Sw)] + We (6.8)

Most geopressured oil accumulations are undersaturated, thus, the term for gas cap expansion is excluded from Equation 6.8. As done previously, define the following: FpR = Np[Bt + Bg(Rp - Rsi)] + WpBw (6.9) Eo = Bt - Bti (6.10) Equation 6.8 may be restated as: FpR = N[Eo + Btip(cp + cwSw)/(l - Sw)) + We (6.11)

Dividing by the expansion-compression term in square brackets yields:

<

(6.12)

If the water influx term and the rock-fluid expansion compression terms are "correct," a plot of the left-side term versus the second right-side term of Equation 6.12 would be a straight line, with the y-intercept equal to N i.e., stock tank barrels of oil initially in place. Comments regarding the water influx term for geopressured gas reservoirs are appropriate for geopressured oil reservoirs. Pore Volume Compressibility There have been several studies published on the influence of reservoir pressure on pore volume compressibility, for example, Geertsma (1957), Fatt (1958), Van der Knaap (1959), and Dobrynin (1962). It generally has been recognized that the pore volume compressibility of friable and unconsolidated sandstones is dependent on the stress conditions in the reservoir, with the compressibility increasing as the stress decreases. However, it is apparent that sandstone texture also is a controlling factor, and its influence has not been established in work published to date. Newman (1973), for example, observed that "pore volume compressibilities far a given porosity can vary widely according to rock type ... the data are too widely scattered for correlations to be reliable." Thus, the correlation between compressibility and porosity published by Hall (1953) may not, in general, be valid, and should be used with caution. Hammerlindl (1971) published a correlation that showed that pore volume compressibility in geopressured reservoirs increases with reservoir depth and with increases in initial pressure gradient. In geopressured sections where the initial pressure gradient increases with depth, the net overburden pressure decreases as depth increases. Thus, Hammerlindls general relation is consistent with earlier published data that indicate compressibility increases as net overburden pressure decreases.

Reportedly, however, some geopressured sandstones have compressibilities substantially lower than those indicated by Hammerlindls (1971) correlation, with compressibilities approaching those usually associated with consolidated rock (Bernard 1987). These data, however, apparently were measured on rock samples taken from geopressured aquifers, rather than from hydrocarbon reservoirs. In the high temperatures usually associated with geopressured environments, sandstones undergo rapid diagenesis, which can result in a geologically "young" rock becoming tightly cemented. Such diagenesis is more likely to occur in aquifers, where the interstitial water is mobile, than in hydrocarbon reservoirs, where the interstitial water is immobile. These tightly cemented sandstones are generally less compressible than relatively uncemented sands. In view of the foregoing observations, it is recommended that compressibility measurements be made on samples taken from the hydrocarbon-bearing zone and not on samples from the aquifer. Extreme caution should be exercised in using compressibility data from rocks that appear to be similar to the zone of interest, or rocks that have comparable porosity and permeability. In the absence of laboratory data, industry publications may provide guidance regarding the relation between reservoir stress and pore volume compressibility. Four papers are discussed briefly: Fatt (1958), Van der Knaap (1959), Dobrynin (1962) , and Newman (1973) From limited data reported by Fatt (l958) seven sandstones with porosity in the range 10 to l5* it may be observed that, for poorly sorted sandstones, the pore volume compressibility varied from a maximum of approximately 35 microsip (at a net overburden pressure of 1000 psi) to a minimum of approximately 5 microsip (at a net overburden pressure of 12,000 psi). For well-sorted sandstones, pore volume compressibility varied from approximately 10 to 2 microsip over a comparable range of net overburden pressure. Fatt (1958) suggested the data could be fitted with a relation of the form: cp = A - B log (pon) where A and B are constants for each sample. (6.13)

For poorly sorted sandstones, the relation between pore volume compressibility and net overburden pressure (pon) for Fatts (1958) data would be approximately: cp = 100 - 10 ln(pon) where cp units are microsip. (6.14)

Van der Knaap (1962) reported pore volume compressibility data for a variety of porous media and observed an approximately linear log-log relation between pore volume compressibility and net overburden pressure. Compressibilities reported by Van der Knaap (1962), however, appear to be substantially less than those reported by Fatt (1958) over a comparable range of net overburden pressure. Data reported by Dobrynin (1962) appear to be consistent with those reported by Fatt (1958). Dobrynin (1962) reported that, "between a certain minimum (net overburden) pressure and a certain maximum (net overburden) pressure, the relation between pore (volume) compressibility and logarithm of pressure can be approximated by a straight line"; or:

where:

cp = cpxln(px/p)/ln(px/pn) cp = pore volume compressibility (1/psi)

(6.15)

cpx = pore volume compressibility at minimum net overburden pressure (1/psi) px = maximum net overburden pressure (psi) pn = minimum net overburden pressure (psi) This relationship is consistent with one proposed by Fatt (1958), or Equation 6.13, above. To determine the effective compressibility over a pressure range of interest in the reservoir, Equation 6.13 is integrated between initial net overburden pressure in the reservoir and the net overburden pressure at which Equation 6.7 or 6.12 is being evaluated. Newmans (1973) compressibility data were reported at a common pressure of 75* of initial overburden pressure. For unstressed porosity in the range 20 to 30*, compressibilities for unconsolidated and friable sandstones ranged from approximately 5 to 80 microsip. Based on the foregoing discussion, it is suggested that in the absence of laboratory data the following be used to estimate cpx, maximum pore volume compressibility, for substitution in Equation 6.15:

Rock texture Loosely consolidated to friable, poorly sorted, with 20 to 45* intergranular detritus i.e., a U.S. Gulf Coast "graywacke" Well-sorted, well-rounded grains with up to about 15% intergranular detritus i.e., a "quartzite"

cpx (microsip)* 50

10

* At pn = 200 psi and with px = 20,000 psi Partial Water Drive (Rate Sensitive) Gas Reservoirs In 1965, Agarwal et al. published a paper in which they demonstrated theoretically that, in a mathematically isotropic gas reservoir subject to edgewater encroachment, "gas recovery from certain ... reservoirs may be very sensitive to gas production rate." Their calculations suggested that lower-permeability reservoirs were more sensitive to gas production rate than higher-permeability reservoirs. They concluded by stating: "If practical, the field should be produced at as high a rate as possible, and field curtailment should be avoided. This may result in a significant increase in gas reserves by lowering the abandonment pressure."

The recovery efficiency of gas from partial water drive gas reservoirs may be estimated with Equation 3.18: (3.18) [ERg]wd = [1 - RpZ] + [RpZEV] where EV is the volumetric sweep efficiency and RpZ is the ratio of abandonment to initial p/Z. From examination of Equation 3.18 it is apparent that these two terms are of comparable significance in the determination of gas recovery efficiency from partial water drive reservoirs. Since the Agarwal and Ramey publication, it has become increasingly apparent that seemingly minor variations in permeability can have a major influence on volumetric sweep efficiency in reservoirs subject to water influx (Hartmann and Paynter 1979; Cronquist 1984). Thus, theoretical calculations for recovery efficiency of gas of the type used by Agarwal et al. should be used with extreme caution to establish production policy or to estimate gas reserves from this type of reservoir. Thin Oil Columns (Between Gas and Water) Oil columns less than about 20 feet in gross thickness generally are considered "thin." When such an accumulation is over-lain by a significant gas cap and underlain by an active aquifer, estimating recovery efficiency of oil using volumetric methods can be subject to substantial uncertainty. Typically, wells completed in this type of oil accumulation are subject to coning of the overlying gas, the underlying water, or both. This usually results in rapidly increasing GOR or WOR and relatively short economic life. Several correlations have been developed to estimate "critical rates" to produce these wells, below which in theory coning should not occur. See, for example, Muskat and Wyckoff (1935), Chaney et al. (1956), Chierici et al. (1964), Bournazel and Jeanson (1971), Richardson and Blackwell (1971), Kuo and DesBrisay (1983), Kuo (1989), and Hyland et al. (1989). From a critical review of this literature which spans an interval of over 50 years it is apparent the industry has yet to develop a general treatment of coning that includes the influence of aquifer influx and other relevant parameters. Some authors, for example, have investigated the problem of coning in the presence of an inactive aquifer, analogous to the classic coning problem first discussed by Muskat and Wyckoff (1935). Others, however, have investigated the problem in the presence of an active aquifer. It seems apparent that, in the presence of an active aquifer, the critical rate to avoid water coning would be less that the critical rate in the presence of an inactive aquifer, other factors being the same. In addition to aquifer strength, another critical parameter needed to apply these correlations is the ratio of horizontal to vertical permeability over each wells drainage area. In theory, this parameter can be determined by vertical interference testing or vertical pulse testing, as discussed by Earlougher (1977). However, the test procedure involves two sets of perforations separated by a packer, an expense many operators might be reluctant to incur. Another possible approach is reservoir simulation of wells exhibiting coning to establish the ratio of horizontal to vertical permeability that results in an acceptable match to observed behavior. However, depending on the depositional environment of the reservoir rock and the degree of lateral heterogeneity, it is unlikely that results from one such test would be applicable to all wells in the reservoir. As a practical matter, it might be easier to test

wells at gradually increasing rates to determine a maximum rate at which each well can be produced without coning. In the presence of a strong aquifer and a gas cap, the combination of water encroachment and gas cap coning may result in displacement of part of the oil column into the gas cap, resulting in usually irrecoverable losses of oil. Depending on the size of the initial gas cap and the degree to which gas cap voidage occurs, significant volumes of oil may be lost. In many cases, this loss may be avoided by injection of the produced free gas into the gas cap, thereby maintaining its size constant. Assuming acceptable operating practices, recovery efficiency of oil from this type of reservoir depends on

well spacing and completion method the ratio of horizontal to vertical permeability the size of the initial gas cap the thickness of the initial oil column whether or not produced free gas is injected into the gas cap to maintain
the size constant

the strength of the aquifer.

Tight Gas Reservoirs As defined by the United States Federal Energy Regulatory Commission (FERC), lowpermeability gas reservoirs are those with an average in-situ permeability of 0.1 md., or less. Finley (1984) and the AAPG (1986) have provided a description of the geologic and engineering characteristics of the major tight gas reservoirs in the United States. With the increasing activity in develdping and producing these reservoirs came the realization that methods to estimate gas reserves in moderate- to high-permeability reservoirs are not reliable in very low-permeability reservoirs. The problems can be attributed to (a) the geologic setting in which these reservoirs occur and (b) the completion methods required to make them commercial. In general, the geologic setting of these reservoirs is characterized by a high degree of permeability heterogeneity lateral discontinuities in apparently "blanket" sands stratigraphic, rather than structural, traps complex mineralogy, frequently with high grain density minerals randomly dispersed throughout the section

water-sensitive clays.

These attributes make it very difficult to determine porosity and interstitial water saturation by conventional log and core analysis (Brown et al. 1981; Kukal et al. 1983; Spencer 1983; Witherbee et al. 1983). These problems, plus the high degree of lateral discontinuity, lead to substantial uncertainties in volumetric estimates of gas initially in place. In many cases, it is not possible to distinguish between commercial and noncommercial intervals from log analysis alone. Drillstem tests rarely provide any useful information, as formations often are damaged during drilling, and massive hydraulic fracturing (MHF) usually is required to obtain commercial flow rates. However, despite over 35 years of experience with fracturing technology, the industry is unable to design a treatment and predict the results with a high degree of confidence (Kazemi 1982; Veatch 1983) It generally is recognized that, over life, the gas production rate from wells completed in these reservoirs typically exhibits a hyperbolic decline (Stewart 1970; Brown et al. 1981). B values generally are greater than 1. Hale (1981) has demonstrated that elliptical flow equations provide a good approximation to well behavior in these reservoirs.

RESERVOIR CLASSIFICATION
Various classification criteria have been proposed for defining fractured reservoirs. For example, Nelson (1985) modifying the classification scheme of Stearns and Friedman (1972), groups naturally fractured reservoirs into four categories, based on fracture mechanisms or origins: Tectonic fractures are those whose genesis can be attributed to tectonic events. We can identify tectonic fracture systems based on their clear relationship to faulting, folding and piercements. Contractional fractures can form from various processes that cause a reduction in bulk volume. These processes are related to internal or body forces such as dehydration, thermal contraction and diagenesis. Surface-related fractures develop from weathering and unloading of stored stress. Regional fractures are those covering extensive areas with relatively constant orientation. These fractures have shears attributed to various causes, including plate tectonics. In most cases, however, the exact cause of regional fractures may not be clear. Regional fracture systems exhibit large spacing among fractures and little change in fracture orientation. Figure 1 shows the trend of regional fractures observed in oriented cores of the Devonian shale wells in the Appalachian Basin, U.S.A.

Figure 1

Stearns (1968) describes fractures associated with folding based on the fracture direction. These include fractures parallel to dip, perpendicular to dip, perpendicular or parallel to bedding and at angles to bedding. Because faults are caused by shearing, most fault-related fractures are zero-shear fractures. From the standpoint of productivity, Nelson (1985) divides naturally fractured reservoirs into four general types: Type 1 In these reservoirs, fractures account for the bulk of porosity and permeability. Examples include the Ellenburger Fields in Texas (Bulnes and Fitting, 1945), La PazMara in Venezuela (Smith, 1956) and Amal in Libya. These reservoirs contain high fracture density, may exhibit sharp production decline (in the case of pressure depletion reservoirs) and can develop early water or gas coning. Type 2

These reservoirs consist of systems in which the fractures provide the essential reservoir conductivity. Among these types of reservoirs, Nelson lists Agha Jari and Haft Kel in Iran and the Rangely Field in Colorado. For these reservoirs, the nature of interporosity flow must be identified for infill drilling or implementation of improved recovery processes. Type 3 These reservoirs have both adequate permeability to flow and fractures that enhance overall permeability. Examples include Gachsaran in Iran, Lacg in France and Kirkuk in Iraq. Most times, the evidence of fractures is not clear in the early life of the field. Furthermore, unusual responses during pressure support by gas or water injections can be observed because of permeability trends. Type 4 In these reservoirs, fractures serve as the cause of anisotropy without providing any noticeable contribution to porosity and permeability. The above categories are important because early recognition of fracture types and fracture control during productivity can help in planning well locations, establishing a maximum efficient rate of production and selecting an appropriate, improved oil recovery process.

RESERVOIR CHARACTERIZATLON AND MODELING

Because of the importance of a simulation-based approach to reservoir management, the main purpose of reservoir characterization is to define and assign spatially varying parameters to a reservoir simulator. Many decisions especially at the early stages of development depend on being able to make rough estimates of fracture and matrix storativity and interporosity flow. In some reservoirs. fracture storativity constitutes a major portion of the overall storage capacity. Factors such as matrix porosity. permeability and the nature of matrix-fracture interaction influence the recovery of fluids from the matrix. Once we recognize that fractures are present, we need a way to express the fracture quality, distribution and pattern. The storativity of the matrix does not indicate the total accessible hydrocarbon in place. Core studies and well log analyses, as a minimum, can help us characterize fluid types and fluid accessibility in the matrix. As indicated in earlier sections, core studies can further hint at the hydraulic conductivity of the fractures. Mineralization of fracture voids can, however seriously impede the fracture conductivity.

VARIOUS CONCEPTUAL MODELS

Reviewing a reservoirs geologic history from the time of its deposition to the time of its discovery is essential to understanding the dynamics of fracture opening and closing. Fractured rocks that are subjected to solution processes can end up with enlarged caverns. Therefore, we need a porosity model that includes historical changes related to diagenetic effects such as dissolution and dolomitization.

incorporating these changes requires sequence modeling that matches the entire geologic history of the sedimentation, alteration and compaction of the reservoir rock. Much progress has been made in this regard for sandstones with conventional pore structures. But there is certainly room for more efforts to develop geologic modeling of fractured reservoirs. In the absence of such models, the industry has used some simple conceptual equivalents to handle routine interpretation of well tests and to conduct performance predictions. In the following sections, we review these approaches, while calling attention to the fact that the models in use are transitional and can be substantially improved with more data and more analytical reasoning. Naturally fractured reservoirs are among the best examples of extremely heterogeneous systems. It is unrealistic to assume we can get a deterministic image of a naturally fractured reservoir from a limited number of sampling sites. Yet to assess the hydrocarbon in place and its relative distribution in fractures and matrix, we need to either represent this complex system with a series of stochastic images, or (being even more pragmatic) apply an extremely simplistic deterministic image for rough estimation purposes. Four categories of fractures have been observed in field studies: Large fault zones recognized by seismic surveys Tectonic faults below seismic resolution Small fractures bounded by strata Microfractures

Conceptual models should include various categories of fractures and their interrelationships. Many bulk reservoir testing procedures which were developed primarily for homogeneous reservoirs, have been field tested in fractured reservoirs. To estimate lumped fracture parameters from such field tests, and to use these parameters in conceptual models, we have to make numerous assumptions about the reservoir. Numerous conceptual models of fractured reservoirs have been proposed. These models fall into two broad categories: ordered and disordered fracture distribution.

Ordered models include those which follow Euclidean geometry, such as the
dual porosity, dual permeability, triple porosity and multi-permeability models Another approach to defining fractured rocks consist of defining an equivalent porous media from discrete fracture models. Non-Euclidean models have also been recognized and are a subject of active research.

Disordered models consider some random distribution of fractures. Developments in stochastic modeling for fractured reservoirs are also expected to improve the prediction capabilities of reservoir models.

EUCLIDEAN MODELS
Because near-wellbore measurements are inherently limited in their investigative capabilities. producing bulk images of a reservoir requires tools with deeper investigative ranges. In 1963.y Warren and Root presented a technique for analyzing pressure buildup data from naturally fractured reservoirs. The response of the "sugar cube" model that they developed to represent dual porosity did fit actual pressure buildup tests. This model indicated that if we can assume such a simple representation based on interpreting various segments of a pressure vs. time plot, we can estimate the interporosity coefficient and the fracture storativity ratio. In fact, the industry has used this model and other similar representations to) describe and manage reservoirs. Figure 1 shows several idealized Euclidean models proposed in the literature for modeling of fractured reservoirs.

Figure 1

These models are expressed in terms of fracture density (AfD) and direction of flow.

Dual Porosity Pseudo Steady State Flow

The Warren and Root model assumes that interporosity flow takes place under pseudo steady-state conditions. In other words, there is no pressure gradient between any two points within a given matrix block, and the difference between average matrix pressure and fracture pressure controls the interporosity flow. Figure 2 shows the geometrical configuration of this model.

Figure 2

The two bulk parameters, and , define this reservoir configuration:

(3.1)

(3.2) DaPrat et al. (1984) reported on the application of Warren and Roots dual porosity model to well tests in Venezuela and obtained values in the range of 0.1-0.5 and values from 10-6 to 10-7. Aguilera (1985) presented a summary of values obtained by various authors.

Dual porosity gradient flow

The gradient flow model was originally proposed by Kazemi (1969) for a "layer cake" representation of fractured reservoirs. Later, researchers (e.g., DeSwaan (1976), Chen et al. (1984) and Streltsova (1983)) developed the concept for other flow geometries.

In the gradient flow model, a pressure gradient develops in matrix blocks, and the pressure within the matrix becomes space-dependent. This conceptual model is designed to represent reservoirs with equivalent large matrix blocks. Figure 3 shows the pressure distribution in matrix blocks under gradient flow conditions.

Figure 3

The quantity pm/p refers to the dimensionless pressure distribution; z is the dimensionless vertical coordinate within the matrix block; and t is a dimensionless time function.

Triple porosity model

Abdassah and Ershaghi (1986) proposed an extension of the dual porosity model to explain anomalies observed in pressure transient tests of some fractured reservoirs. In their model, the matrix contribution emanates from two distinct ranges of matrix properties as shown in Figure 4 : a very tight set of matrix blocks, which may exhibit fracture support after a very long period, and a somewhat more permeable matrix, which responds earlier.

Figure 4

Figure 5 shows an example of an observed pressure response that, in fact, fits the description of the theoretical model of a triple porosity reservoir.

Figure 5

MINC model (Multiple Interacting Continuous)

This model was originally proposed by Pruess and Narasimhan (1985). From the standpoint of reservoir modeling, matrix blocks are discretized into a sequence of nested volume elements. These elements are defined with respect to the distance from the block surface ( Figure 6 ).

Figure 6

A proximity function is defined to represent the fraction of matrix material present within a distance x from the fracture. This function allows consideration of pressure, temperature and saturation gradient, which control the inter-porosity flow. The model has been used by Gilman (1983) to study fractured hydrocarbon reservoirs. For simulation purposes, the MINC representation involves partitioning the fractured reservoir into primary volume elements. Primary grid blocks are subdivided into a sequence of secondary nested volume elements. The double porosity approximation is a special case of the MINC method.

NON-EUCLIDEAN MODELS
Euclidean models have proven useful tools for interpreting well tests and historymatching fractured reservoirs. When it comes to using them for reservoir modeling, however, they have an obvious drawback: their significant dissimilarity to observed patterns of naturally occurring fractures. To overcome the inevitable frustrations of

trying to apply Euclidean geometry to seemingly random arrays of fractures, several investigators have proposed fractal geometry. The term fractal was introduced by Mandlebrot (1983) to describe the concept of self-similarity at all scales. Brown and Scholz (1985) showed that the geometry of fractured rock surfaces is a fractal. Turcotte (1986) indicated that the fragmentation mechanism is scale invariant and, thus, fractal. In 1989, Barton and Hsieh demonstrated that fracture patterns in Yucca Mountain, Nevada, repeated themselves over a wide range of scale. Figure 7 shows the fractal plot for a number of Yucca Mountain pavements.

Figure 7

Pavement 1000 is shown in Figure 8 .

Figure 8

Sammis et al. (1992) made similar observations of observed fracture patterns on the outcrops around the Geysers geothermal field in northern California. Figure 9 shows the fractal nature of fracture patterns as examined on different scales.

Figure 9

Imperfections in rocks, when randomly distributed, can result in the formation of fractures. With an increase in the number of fragments, we observe a power law relationship between the probability of fragmentation in a given interval and the interval spacing. We can verify this power law relationship from a method referred to as box counting). We express the box fractal dimension (D) as shown in the following equation: N(r) = r-D (3.3)

where N represents the number of occupied cells and r is the corresponding scale. Figure 10 shows the box counting procedure and the plotting method to obtain the fractal dimension.

Figure 10

FRACTURE PATTERNS
Using numerical studies of fracture systems, Long and Witherspoon (1985) studied two-dimensional fracture networks, assuming constant apertures and lengths. Their study showed that for small values of fracture length and large fracture densities, permeability can be zero. To develop hydraulic conductivity, fracture length must increase while there is a decrease in fracture density. In a subsequent study, Long and Shimo (1985), using the cubic law for a fixed network of fractures, verified the coefficient of variation of hydraulic aperture and the correlation between fracture length and aperture. The results indicated that hydraulic conductivity decreased with an increase in the coefficient of variation of aperture. This reaction was attributed to the flow restrictions in the large conductors by the smaller fractures. They also predicted an opposite effect for correlation between length and aperture. In brief, using a methodology similar to the one described and given a statistical parameter set, one can generate a statistical framework for fracture patterns. The measured permeabilities can be conditioned by fracture frequencies observed at control points and other available data. Figure 11

shows examples of fracture patterns generated by a statistical approach (p = fracture frquency).

Figure 11

Geostatistical methods have been used to generate conceptual models of reservoir heterogeneity. This involves generating various realizations, conditioned to match local measurements and inferred spatial correlations. These models then constitute the basis for stochastic flow modeling and prediction. For fracture systems, there may be substantial discontinuities, and direct application of geostatistical methods may be insufficient to generate flow models. As indicated by Long (1991), such conceptual models fitting the geometry of fractures may be inadequate to predict performance. Inverse modeling such as simulated annealing and iterated function systems (IFS) have been under study and development by several researchers. Davey et al. (1988) used simulated annealing to find an equivalent discontinuum model. Their procedure involves subjecting a partially filled lattice of 1-D conductors to simulated annealing, and during several iterations, changing the configuration of lattice elements to observe field test data. The difference between the model behavior and field response is expressed as E = [ln(ho) - ln(hc)]
2

(3.4)

Random changes in the elements are made, and the E of the new configuration is changed with the previous estimate. The new configuration is maintained and subjected to further modification as long as the E is on a declining trend. The IFS system originally proposed by Barnsley (1988) has been used by Acuna and Yortsos (1991) and Long et al. (1991) to generate fracture patterns. For an initial set of points and a set of iterative functions, a fractal-like object is generated. This is referred to as an attractor. These iterated functions can be deterministic or stochastic. Figure 12 shows attractors generated by Long using several functions.

Figure 12

Figure 13 shows the development of fracture patterns after 0, 110 and 140 iterations.

Figure 13

Figure 14 shows well test data compared to model predictions. Further research is needed to extend the process to three-dimensional systems.

Figure 14

Gale et al. (1991) presented results of fracture studies on the Monterey formation, based on outcrop data. They noted that the trace length for each set of fractures, as well as spacing values, follow a log-normal distribution. They used these fracture statistics to generate stochastic images of the fracture network in two and three dimensions. Figure 15 shows examples of histograms obtained on fracture trace length and fracture spacing.

Figure 15

With these data, a fracture network code is employed to generate 2-D and 3-D stochastic images of the fracture network. In brief, Gale et als. procedure consists of generating a number of random points within the 2-D or 3-D space. At each point, a line length, characterized by statistical parameters of orientation and length, is assigned. The process continues until a prescribed fracture density is generated. After eliminating all intersections that cannot assist the flow, fracture elements between any two nodes are assigned an aperture either a fixed or a random value from log-normal distribution. Figure 16 is an example of a 3-D pattern generated using fracture statistics as described here.

Figure 16

Work presented by Chang and Yortsos (1990) and Acuna and Yortsos (1991) predicts the pressure transient response of fractal networks. Additionally, Acuna et al. (1992) demonstrates that the pressure transient tests from actual wells in the Geysers field can be explained with the predictions made for fractal patterns. They show that when pressure transient test data for a naturally fractured reservoir resemble the expected pattern for a well with a hydraulic fracture, these data may also indicate a fractal network of fractures. Figure 17 shows an example of pressure transient data from a naturally fractured reservoir exhibiting linear flow.

Figure 17

NEED FOR FURTHER RESEARCH

Many of the conceptual models and early assumptions about the nature of subsurface fractures have essentially oversimplified very complex systems. Recent applications of downhole tools for fracture detection have shed light on some important characteristics of natural fractures. For one thing, the relationship between the fracture aperture and permeability needs some re-examination. It is clear that fracture connectivity plays an important role in overall fracture conductivity. Furthermore, there are uncertainties with respect to defining an effective fracture aperture that represents aperture variations within the fracture. Also, fracture apertures measured on cores and logs do not necessarily represent the aperture variation within the fracture network. The effect of overburden on fracture closure is not deterministic and must be considered in modeling work. Lithology also affects the density and the irregularity of fractures.

Bai et al. (1993) discuss the deformation-dependent flow models of various hypothetical and conceptual fractured reservoir models. They point out the importance of incorporating solid deformation as a part of overall fluid flow modeling. A continuously changing fracture permeability pattern requires innovative approaches in compiling fracture closure with spatial variations in pore pressures.

BULK CHARACTERIZATION PARAMETERS

If we base a bulk representation of a naturally fractured reservoir on a dual porosity model of the type described by Warren and Root, we need to define an effective matrix block size Warren and Root introduced the parameters (the ratio of the fracture storativity to the total system storativity) and (the interporosity flow coefficient) to represent bulk characterization of idealized fractured reservoirs. The shape factor a is related to the dimensions of the matrix blocks. For uniformly distributed blocks,

(3.5) For blocks with dimensions of a, b and c, we may compute the effective block dimension d as shown below: n=3: d=(3abc)/(ab+bc+ac) n=2: d=(2ab)/(a+b) n=1: d=a (3.6) (3.7) (3.8)

where n is the number of normal sets of fractures. The relationship d2 is equivalent to 12, 32 and 60 for 1-D, 2-D and 3-D systems, respectively. Based on simulation studies, other values have been proposed as follows:

System 1-D 2-D 3-D

Warren and Root 12 32 60

Coats (1989) 8 24 24

The shape factor a defines the matrix block surface area per unit volume. Kazemi et al. (1976) proposed the following expression for :

(3.9)
(Lx2, Ly2 and Lz2 correspond to a, b and c in Equation 3.6.)

The above parameters help significantly during the history matching of reservoir simulation studies that employ simple sugar-cube conceptual models. In reality, block sizes and shapes need not be the same. The effective matrix block size is only meaningful in reservoirs that are not intensely fractured. For fractured reservoirs with sparse fractures, several approaches may be employed to estimate effective block size. Downhole data from fracture identification jogs, fluid entry surveys and drilling logs of intervals causing lost circulation can generate alternative measures of fracture spacing and matrix block sizes. This information, when confirmed by incorporating outcrop maps, can aid in characterizing block size distribution.

PHYSICAL MODELING
From the standpoint of fluid mechanics, fracture studies are a matter of describing flow through irregularly sized and spaced 3-D channels. Understanding multiphase flow in fractures and the interaction between matrix and fractures for various recovery processes is essential to accurately predicting reservoir behavior. Researchers have reported a limited number of physical models describing single and multiphase flow through fracture systems. These studies, which have generally been conducted on a single block basis, have addressed flow modeling, effective transport parameter determination and small scale micromodels. Fourar et al. (1992) studied the nature of two-phase flow in fractures and concluded that several flow structures (such as bubble flow, complex films and drop flow) exist that make the process similar to two-phase flow in pipes. Some issues of concern in physical modeling are:

the characteristics of fluid-fluid interfaces in fractures the nature of multiphase relative permeability the mode of multiphase flow structure in fractures the nature of capillary continuity between matrix blocks the factors affecting countercurrent imbibition, gravity drainage and diffusion effects

Depending on the experiments purpose, fractures may be natural or manmade. In general, there are two reasons for making physical measurements: (1) to determine fracture and or matrix petrophysical properties, and/or (2) to measure interporosity flow. In conducting core-level flow experiments, synthetic fractures have been cut to model the transport in the composite fracture/matrix system. Using cores with natural fractures introduces another level of complexity, in that the geometry of fracture channels needs to be characterized. Various laboratory and analytical techniques have been proposed to study the geometry of rough fracture surfaces. Several mechanisms contribute to the productivity of oil from matrix blocks:

Lowering of fracture pressures causes a pressure gradient between matrix blocks and fractures, and the resulting oil expansion expels the oil. Gradually, the gas liberated in the blocks generates a solution gas drive process. A secondary, high-pressure gas cap eventually forms around the matrix blocks, which results in oil draining from the blocks. Diffusion between the gas in the fracture and the gas in the matrix
expels the light fraction of oil from the matrix and conveys it to the fracture.

For blocks close to the water-oil contact, oil is recovered under counter-current imbibition. Both the block size and block permeability affect the oil recovery from blocks.
Imbibition in fractured reservoirs is a crucial factor in controlling recovery from a water wet matrix. Factors affecting the rate of counter-current imbibition have been studied by physical experimentation as well as numerical modeling. Mattax and Kyte (1962) related imbibition recovery to block size, fluid viscosities and rock petrophysical properties. Figure 18 shows their observed correlation of recovery versus the dimensionless parameter

Verifying the nature of these physical processes,

Figure 18

as well as their effectiveness for a particular reservoir, requires conducting physical measurements on the fractured rocks. In any case, the main concern is how well small-scale, laboratory-based experiments can simulate large-scale field conditions. The problem of scale-up is an area for further research and development. For the time being, however, we present a brief review of how these measurements are made. A number of investigators have studied flow in fractured cores. A major consideration is how to up-scale study results from laboratory scale tests to field predictions. For example, Mattax and Kyte (1962) related coreflood recoveries to reservoir scale as follows:

(3.10) Mattax and Kyte showed that Equation 3.10 applies where horizontal imbibition is the recovery process. But when effects of gravity are important, the following equation must also be satisfied:

(3.11)

From these relationships, Parsons and Chaney (1966) showed that the rate of interface rise with time (I= dz/dt) can be related between the model and the prototype:

(3.12) Babadagli and Ershaghi (1992) showed that the capillary imbibition process under fracture flow is a strong function of flow rate in fractures, matrix capillary pressure and matrix permeability. They derived an effective relative permeability curve for the composite fracture-matrix system and indicated its dependence on flow rate. The question of proper scaling still requires much additional work. Kyte (1970) suggested the use of a centrifuge test to scale the effect of both gravity and capillarity. Also, Labastie (1990) examined the issue of capillary continuity between matrix blocks. From experiments, he concludes that, in general, capillary continuity exists between matrix blocks in naturally fractured reservoirs. Gas-liquid transmissibility in fractured reservoirs has not been fully studied on physical models. Firoozabadi and Markeset(1992) presented experimental data on gas-liquid flow in fractured systems and on the nature of capillary continuity. They concluded that fracture transmissivity is insensitive to the contact area between the matrix blocks but is a strong function of fracture aperture. Horie et al. (1990) reported strong capillary continuity across horizontal fractures for a stack of matrix blocks. Thompson and Mungan(1969) conducted a series of gravity drainage experiments on Berea cores containing synthetic fractures and concluded that the displacement rate, matrix permeability and subvertical fractures affect the oil recovery. Additional experimental modeling of re-infiltration and capillary contact has been reported. Regarding the physical modeling of fractured reservoirs containing heavy oil, both combustion and steam experiments have been reported:

Schulte and De Vries (1985) conducted combustion experiments on a stack of vertically positioned, oil-saturated cores. They found that the burning process is controlled by diffusion of oxygen from the fractures into the matrix. Both thermal expansion and evaporation, followed by condensation of the oil from the matrix, constitute the recovery process from the matrix. Reis (1990) and Briggs et al. (1992) reported on steam injections
into laboratory-scaled fractured systems. They observed that steam tends to stimulate the recovery of matrix oil through imbibition, internal gas drive and differential thermal expansion.

Dreher et al. (1986) conducted laboratory experiments for hot water

and steam flooding in laboratory-scale fractured carbonate rocks. Their

observations indicated low matrix oil recovery when either hot water or steam are injected into the fractures.