| PROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION:2.1 REV. DATE: 1 JUNE 1964 REV:0 PAGE:1 of 6 stat “J! 22 Composition of Wellhead Fluids | 2.1.1 Fluid Description Wellhead fluid is a mixture of hydrocarbon liquids, hydrocarbon and non-hydrocarbon gases, contaminated by other reservoir constituents, such as water and solids. To this may be added other fluids for production purposes such as gas or CO, seawater or produced water 2.1.2 Sampling and Analysis The Drill Stem Test (DST) is used to sample the reservoir fluid by producing for several hours via special drill tools into a Test Separator. Samples may be taken from downhole, at the surface or at the separator. Downhole samples are considered the most reliable but suffer from the restricted accuracy of a single, low quantity test, from an uncompleted well. Downhole sampling is more reliable when used for the sampling of single phase ofl or gas reservoirs. Production via the Test Separator permits separation of the wellhead fluid into oil, gas and produced water. Each of these streams may be separately metered, although inaccuracies will = = , occur unless specific attention is given to the calibration of + the meters. Samples from each stream are then sent to a laboratory for detailed analysis. The separation of the wellhead fluid into the respective fluid phases, permits evaluation of the Gas to Oil ratio (GOR) for an 1 oil reservoir, or the Liquid to Gas ratio (LGR) for a gas J reservoir. These ratios are open to uncertainty however, due to the limited nature of the DST and the fluctuation of flows experienced. 1 Sampling of the wellhead fluids is also required in order to i determine the concentrations of hydrogen sulphide (H2S) present. This gas is toxic, even in low concentrations, and specific iPROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION:2.1 1 REV. DATE: 1 JUNE 1984 REV:0 PAGE:2 of 6 statoil consideration must be given to the handling of the wellhead fluids froma safety aspect. The H2S concentrations will also determine material selections, due to its corrosivity in wet and warm environments. 2.1.3 Compositional Analysis The laboratory analysis of the wellhead fluid samples enables the composition of the fluids to be established. Chromatograhic analysis is used to identify the composition of the lighter molecular weight components, up to approximately Cg or C7. The residual condensate or crude oil is then subjected to a True Boiling Point (TBP) distillation analysis in order to enable determination of the heavier molecular weight fractions. Certain lighter fractions of the TBP analysis may be subject to further analysis in order to establish the Paraffin, Olefin, - Naphthene and Aromatic (PONA) content. The analyses for both the liquid and vapour portions of the wellhead fluid may be combined mathematically to predict the actual reservoir composition. 2.1.4 Physical and PVT Analysis Laboratory analysis produces a Pressure Volume Temperature (PVT) report. The primary objective of this report is to establish data for vapour liquid equilibrium (VLE) predictions. Multiple flashes are often performed to establish further the VLE data for specific applications. However, computer simulations from generalised correlations are often found to be as accurate. The physical properties of the various distillation cuts are established. These include data on viscosity, wax content, pour point.PROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION:2.1 7 REV. DATE: 1 JUNE 1984 REV PAGE:3 of 6 stat}, i j 2.1.5 Hydrocarbon Characteristics | Wellhead fluids normally contain hydrocarbons ranging from Methane (CH4) up to high molecular weight components with 40+ carbon atoms. Four primary types of hydrocarbon are found: Paraffins - saturated straight or branched chains | Naphthenes = - saturated cyclo-paraffins Olefins - unsaturated chains i Aromatics - —_ unsaturated cyclo-paraffins : The lighter components are predominantly Paraffins, the heavier components being predominantly Napthenes, Aromatics and Olefins. Gases consist of C, to C, components with trace amounts of Cet The trace quantity of C,+ is important for the prediction of vapour/liquid separation in predominantly gas streams. Calculation of the stream properties from the properties of the individual components is well established and accepted. Crude oil consists essentially of C,+ components but may contain varying quantities of C, to C, components and contaminates. Standard TBP or ASTM distillation test methods are used for characterisation of the oil by fractionating into small cuts and measuring specific gravity and molecular weight for each cut. This minimum data can then be used for generating pseudo-components to represent the heavier fractions of the oil and to generate derived properties, such as Volume Average Boiling Point or the Nelson 'K' value. The summation of individual psuedo-component characteristics permits prediction of the properties of any mixture resulting from these components. Using these techniques, almost all physical property data required in production work can be generated. Certain advanced techniques for improving the accuracy of data, particularly VLE are recommended in the detailed text. Whilst hand calculations are feasible, computer simulation is recommended for all these ; \ calculations.PROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION: 2.1 REV. DATE: 1 JUNE 1984 — REV:0 PAGE:4 of 6 statoil 2.1.6 Other Constituents and Contaminants 2.1.6.1 Water and Salt Water produced with the reservoir fluid frequently contains salt. The concentration of these contiminents in the fluid streams should be minimised as they cause problems in downstream refinery processing units and so detract from the crude value. Oilfield produced water tends to increase as production progresses, through the rate of increase can only be established by experience. Gas tends to be saturated with water vapour and estimation methods are given in the detailed text. Problems associated with produced water are generally caused by the dissolved salts and may result in scaling, plugging and Geposition. Free water may also result in hydrate formation. Water separation usually is restricted to free water, the solubility of water in hydrocarbons being low, although sepaiation from viscous crudes may be a problem. If accurate Getermination is needed it can be estimated from the detailed text. 2.1.6.2 Acid Gases Hydrogen sulphide (H,S) and carbon dioxide (CO) are the most common acid gases found, and when present, materials must be selected for corrosion protection. When levels are high or re-~ injection used it may be necessary to remove these gases. 2.1.6.3 Solids Reservoirs are usually porous rock, hence fluids can contain sand from interstitial particles or fracture of rocks. Advanced recovery techniques may produce large amounts giving potential erosion problems. Other solids such as corrosion’ products (iron sulphides) are likely to be present and should be controlled by proper treatment. Crude oils also contain varying proportions of wax, which consist of high MW paraffins which may crystallise under certainPROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION REV. DATE: 1 JUNE 1984 REV:0 PAGE:5 of LI 6 stati! conditions. Wax formed in high enough quantity to lead to flowing and deposition problems will necessitate chemical or other treatment. 2.1.6.4 Inert Gases Gas fields may contain substantial quantities of inert gases which lower the calorific value of the gas and hence its value. In themselves inert gases are not a problem. 2.1.6.5 Metals 4 Crude oils may contain various heavy metals. Whilst all are _) undesirable in downstream refinery possessing units, only Vanadium and Nickel are usually identified as a problem. 2.1.7 Simulation Techniques for Characterisation Accurate representation of the behaviour of hydrocarbon mixtures is usually beyond quick hand calculations and requires computer simulation, The distinct gaseous species C) to Cy or C) can be | identified and are well documented. Techniques have been : established to characterise the large number of heavier crude | components by a reduced number of “pseudo-components", for each of which properties may be generated as though it were a real = chemical species. Alternatively crude assay data may be used, a — computer program being used to generate the "pseudo-components”™ and their properties. It can be seen that this can rapidly become unwieldy so judicious selection of the number of “pseudo-components" is necessary. A balance must be made of the increased accuracy resulting from the use of a large number of components against the increased computer costs. Typically, 15 to 20 real and "pseudo-components" will accurately | characterise a crude for offshore gas/oil separation calculations whilst for crude oil fractionation calculations, as many as 40 - 50 components are required.PROCESS DESIGN HANDBOOK ~ © ISSUE DATE: 25 MAY 1984 SECTION: 2.1 . REV. DATE: 1 JUNE 1964 RE PAGE:6 of 6 statoil The size of each "pseudo-component” can be controlled and it is often desirable to select more, narrow range components in a particular range of interest and lump together others which do not significantly influence the calculation result or use, e.g. crude vapour pressure calculations where the heaviest components have little individual influence. In order to accurately reproduce the reservoir fluid characteristics, the properties of some of the “pseudo- components" can be adjusted. Among properties that may be adjusted are molecular weights and critical temperatures and pressures. Simulation removes the engineer from direct involvement and important results should always be checked by comparison with published data sources. 2.1.8 | Reference Sources The following references provide more detailed information on the above topics: as "Data Book on Hydrocarbons" - J.B. Maxwell - 5 "Engineering Data Book" - Gas Processors Suppliers Association ; 3. "Petroleum Refinery Engineering" - W.L. Nelson 4. "Gas Conditioning and Processing" - J.M.°CampbellPROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION: 2.2 ane state ./ REV. DATE: 1 JUNE 1984 REV:0 PAGE:1 of 7 2.2 Specification of Products and Effluents 2.2.2 General Offshore processing facilities produce two categories of out going streams as follows: Export Products By-products & Effluents Gas Produced Water Crude Oil Ballast Water Condensate Drains Gases (Flare/Vent) . Sewage Solids/Sludges etc Export products are the objective of the facility and specifications are dependent on: : Sales Specification (subject to pricing agreement) : Export method (e.g. pipeline/tanker) Byproducts and effluents are disposed of to the local environment, reinjected to the field or sent to shore. The specifications for streams disposed of locally are usually subject to government regulation. 2.2.2 Gaseous Products The gas quality specifications fall into two categories: Sales gas or Pipeline gas. A "sales gas specification" generally applies to the sale of gas from a processing facility (onshore) to a gas distribution system. A “pipeline gas specification" is used for offshore export of gas to a pipeline where further gas treatment will be carried out onshore. A pipeline specification applies only to the pipeline section for which it is written. Ww?PROCESS DESIGN BANDBOOK ISSUE DATE: 25 MAY 1984 SECTION: 2.2 * REV. DATE: 1 JUNE 1984 -REV:0 PAGE:2 of 7 statoil Typical sales and pipeline specifications are given in Table 2.2.1. 2.2.2.1 Sales Gas Specification To meet a sales gas specification the following processing may be required: 0 Hydrocarbon Dewpoint Control 0 Water Dewpoint Control : Inert/acid Gas Removal Inert/acid gas removal may be required to meet Gross Calorific Value (GCV), Wobbe Index and maximum impurities, depending on initial wellstream composition. Of the given properties, Wobbe Index (WI) is usually the controlling factor. The Wobbe Index is the ratio of the GCV to square root of the gas specific gravity. This may be estimated from gas molecular weight, corrected for N, and CO, content using Figure 2.2.2. worine nOEX i ope a cme zs ae MOLECULAR WEIOKT Figure 2.2.1 Wobbe Index Predictionsuotqeoqsroads seo TeoTdsy TZ oTAeL SALES GAS SPECIFICATIONS PIPELINE SPECIFICATION CONT'L U.K. STATPIPE STATPIPE untrs | saves | sates LEAN RICH KOFISK| SLEIPNER GAS GAS GAS GAS Gross Calorifie Value m/sm3 |38.0-43.6|38.0-44.0] 38.1-42.1 38.9-43.7 Wobbe_tndex M3/sm? [46 .6-52.1]47.5-51. 4) 48.3-52.8 48.4-52.8 Hydrocarbon Dewpoint at 2 - 70 bara [Oc -3.0 -1.0 -3.0 37. -3.0 at 52 bara °c “11.0 Water Dewpoint ‘at 70 bara °c -8.0 -9.0 -8.0 -18.0 -10.0 Cricondenbar bara 110.9] Max. Content of Impurities ~Oxygen mole’ 0.5 0.1 o.L on 0.2 ‘carbon Dioxide | molet 2.5 2.0 2.5 2.0 2.5 420 initrogen mole’ ilo 70 ‘Hydrogen mg/Sm 5.0 5.0 5.0 5.0 5.0 Sulphide 7 -Mercaptans ng/Sm: 15.0 “Total Sulphur |mg/sm?| 150.0] 150.0 150.0 | 35ppm(v) as HLS Yearly Sulphur|mg/Sm? 120.0 e average 3 sMercury ng/Sm 5.0 Delivery Tenperaturet °c 5.0-30.0| 5.0-38.9 1.0-38.0 Delivery Pressure* bara 51.0 69.0} 121.0-173.0| 75.0 * Delivery Pressure and Delivery Point Location. Figures given are typical. ow Temperature may be subject to tgLvd SNSST a ey g 8 a 8 i j e 5 a a e orame MOOGANYE NOISHA ssgooudPROCESS DESIGN HANDBOOK ~ ISSUE DATE: 25 MAY 1984 REV. DATE: 1 JUNE 1984 REV:0 statoil 2.2.2.2 Pipeline (Offshore) Gas Specification Offshore processing is generally limited to hydrocarbon and water dewpoint control. Inert or acid gas removal is not usually economically feasible, nor carried out, and the full sales gas specification is not used. Gas treatment consists primarily of that necessary to transport the gas ashore to production facilities in the most economic manner, i.e. using minimum processing equipment. The pipeline (offshore) gas specification may be based on the final sales gas specification. The following is a summary of factors to be considered when developing an offshore gas delivery specification: Hydrocarbon Dewpoint = a. To prevent liquid formation in the pipeline, set the hydrocarbon dewpoint temperature 5 to 10°C below the minimum pipeline operating temperature for the operating pressure range. Set dewpoint in accordance with onshore sales gas specification (if available). Water Dewpoint a. To prevent free water formation and hydrates either in the pipeline or within the hydrocarbon dewpoint control system on the platform, the water dewpoint temperature should be set 5 to 10°C below the minimum operating temperature for any of the given operating pressures. b. To prevent pipeline corrosion. As for above for free water formation. c. Set dewpoint in accordance with onshore sales gas specification (if available). If onshore processing involves inert or acid gas removal, gas may be resaturated with water, and this specification will not be relevant.PROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION: 2.2 ~y REV. DATE: 1 JUNE 1984 = REV:0 PAGE:5 of 7 stal_ if a. Water dewpoint may alternatively be quoted as an equivalent water content, where typically the following specification is used: Water content: max 48mg/Sm? (31b/MMSCF). This is equivalant to a water dewpoint of minus 12° at 70 bara. However, for low pressure gas systems this dewpoint is lower than would be required. Delivery Temperature a Maximum temperature into a pipeline is limited by ; consideration of riser expansion and temperatures in thes) . splash zone region of the riser promoting external Y pipeline corrosion. Typically temperatures in the region of 30 to 50°C are used. Delivery Pressure } a. This is fixed by onshore arrival pressure requirements. 2.2.3 Crude Oil Product Crude may be stored on the platform and exported via a tanker or - | pipeline. For the safe storage and transportation of crude, the vapour pressure and maximum delivery temperature is specified typically, as follows: | 2.2.3.1 Tanker/Storage 4 Reid Vapour pressure (RVP): 0.69 bara i True Vapour pressure (TVP): 0.965 bara at 25% i Delivery Temperature: 35°C maximum. 2.2.3.2 Pipeline True Vapour Pressure: 6.9 bara at 38°C (typical). ! This figure will be set in conjunction with the operating Parameters of the pipelines and must be lower than the proposed | arrival pressure at the delivery location. Crude must be pumped to ensure pipeline is liquid phase throughout. As the vapour | pressure is fixed by the conditions of the final separationPROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION: 2.2 REV. DATE: 1 JUNE 1984 — REV:0 PAGE:6 of 7 statoil stage, the impact on the processing system should also be considered. For the maximum delivery temperature to the pipeline, the comments under Section 2.2.2 for gas pipelines apply. Care must be taken in assessing the equivalant TVP for a given RVP, the latter being an experimental procedure. Existing correlations in literature are not applicable to high GOR (i.e. high methane, ethane) reservoir fluids. It is recommended that a computer simulation’ be used for this estimation. Other specifications are fixed by the sales requirement, and as crude price is fixed by quality, the following have “desired" values: Water Content 5 2 wee Salt : 70 to 200 mg/l As higher water content reduces pipeline/storage capacity and crude price, and over longer pipelines picks up significant quantities of iron (Fe), a more stringent water content specification is often used: BS & W (Base, sediment and water) content: 0.5 vols maximum 0.1 vol & average Pour point and/or viscosity may be considered for pipeline capacity and storage problems but in general facilities are designed to accept the product rather than vice versa. 2.2.4 Condensate Condensate production when exported via gas facilities, as a two phase mixture, must meet the gas water content specifications. When exported by dedicated pipeline, to prevent corrosion or hydrate formation, a water content specification should be applied. Additionally limitations imposed by the delivery point may be specified. vapour pressure may apply if injected into a crude stream or exported by other transport means.PROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION: 2.2 “N REV. DATE: 1 JUNE 1984 REV:0 PAGE:7 of 7 stat a 2.2.5 By-Products and Effluents (See detailed system section) Disposal of by-products and effluents to the environment shall meet the requirements of the Norwegian State Pollution Control Authority, 1 October 1982, given in NPD regulations. 2.2.5.1 Produced, Ballast Water and Drains The normal limit for hydrocarbon concentration in produced water discharge is 40mg/1 measured with IR(NS 4753) calculated as a monthly average. For ballast water and drains the normal limit is 25mg/1 for each discharge source in isolation. These ») concentration levels are subject to agreement by the authority for each development, and this should be obtained at an early stage in the project. : 2.2.5.2 Gases (Flare/Vent) Flaring or venting via a suitably located flare boom. Compositions generally C, to Cg hydrocarbons, dependent on source. Sour systems require special consideration. 2.2.5.3 Sewage Maceration is only required if food waste is mixed into sewage. "2.2.5.4 Solids/Sludges etc y It is prohibited to throw overboard firm waste. This shall be : collected in suitable containers and sent to shore for disposal.1 PROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 REV. DATE: 1 JUNE 1984 REV:0 statoil 2.3 Dehydration (Gas 2.3.1 General In the processing and transportation of natural gas, the presence of water may result in hydrate formation and/or corrosion (in association with CO, or HS). If the sales gas is exported to a pipeline system, a water Gewpoint specification is usually imposed. In general, the produced gas stream will require treatment to meet this specification. Several methods of hydrate inhibition and gas dehydration are available: a Glycol (or methanol) Injection b. Glycol Contacting (absorption) c. Solid Desiccant Dehydration (adsorption) The first method (injection) inhibits the formation of hydrates by reducing the temperature at which they form, As it does not involve water removal, it has no effect on the corrosive nature of the gas stream, or the gas water dewpoint. The other two methods (both dehydration) remove water from the gas, reducing the water dewpoint, and hence prevent hydrate formation and reduce corrosion. In deciding which method to use, and specifying the requirements, the following steps should be taken: ae Identify the problem - hydrates, corrosion, export sales gas specification or a combination. b. Establish design water dewpoint c. Establish applicable gas treating methods a. Compare suitable methods on basis of space, weight, utility requirements and cost. Consider integration with other topsides facilities. 2.3.2 Design Specification 2.3.2.1 Water Content In designing the dehydration or hydrate inhibition systems, it is necessary to determine the water content of the gas at thePROCESS DESIGN HANDBOOK 25 MAY 1984 SECTION: 2.3 ISSUE DATE: REV. DATE: 1 JUNE 1984 REV:0 PAGE:2 of 27 statvui esate (OPA) ot et gee ot 1 ond HOLE obey § cre cce tem TOD 0 welt on Met wt oe 0 OF and 1.1 pte Figure 2.3.1 Water Dew Point of Natural Gas !PROCESS DESIGN HANDBOOK ISSUE DATE 25 MAY 1984 SECTION: 2.3 . REV. DATE: 1 JUNE 1984 REV:0 PAGE:3 of 27 statoil initial, downstream or other specified conditions. The dewpoint water content of natural gas is given in Figure 2.3.1. A sales gas specification is quoted as either a water dewpoint (i.e. given pressure and temperature) or water content (kg/Sm?). From this, and the water dewpoint curve (Figure 2.3.1), the required dewpoint temperature at the operating pressure in the dehydration unit can be determined. Hence the required dewpoint depression (inlet - outlet dewpoint temperature) to meet the gas specification can be determined. 2.3.2.2 Hydrate Formation Hydrates are crystalline structures which form when the following > pre-requisites exist: . Presence of free water phase 2 a _ Low temperature and high pressure : Thorough agitation of the phases. : Presence of C, through C, hydrocarbons. “go determine the degree of hydrate inhibition required for a given system, it is necessary to predict if, and at which conditions, hydrates form. As a first estimate, use Figure 2.3.2 to predict the minimum pressure at which hydrates will form for a given temperature and gas specific gravity (Air = 1). Using this estimating technique, a safety factor of at least 10°C should be included in designing the inhibitor or dehydration system. To calculate the hydrate formation point more accurately, the method presented in the ‘Handbook of Natural Gas Engineering’ by Katz et al, should be used. In this method, the hydrate formation point may be determined using the vapour-solid equilibrium constants for hydrates, i.e.: Dyy/K, = 1.0 where: Yy = mole fraction (dry) of component i in the gas. K, * hydrate equilibrium constant for component i-iF PROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION: 2.3 > REV. DATE: 1 JUNE 1984 = REV:0 PAGE:4 of 27 statui. Charts for determining K; are given in Katz, Campbell, GPSA, etc., and are not reproduced here. A Statoil in-house hydrate prediction program based on the above methods is available and should be used whenever hydrates are expected. ~ o 3 7 @ Temperature, °C 2 30 Figure 2.3.2 | Pressure - Temperature Curves for Predicting Hydrate Formation || PROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION: 2.3 . REV. DATE: 1 JUNE 1984 REV:0 PAGE:5 of 27 statoil { 2.3.3 Comparison of Methods ‘ Method Dew Point | Min. Dew | Remarks Deprn. °C | Point °C TEG 0 to 45 Minus 25 Atmospheric Regen. Min. Contacting contact temp. 15 - 260C. 45 to 55 | Minus 40 | Vacuum regeneration and gas stripping. o 55 to 90 | Minus 73 | Regen. by azeotropic distilln. (Dow "Super Drizo") Solid Minus 90 Molecular Sieves Desiccants Minus 73 | Alumina Minus 60 Silica Gel (High 7 capital/operating costs.) Table 2-3-1 Comparison of Dehydration Methods Inhibitor |Min. Temp. oc]Remarks f Methanol [minus 95 Can be regenerated and recovered from ao possibly to |liquid hydrocarbons. Significant minus 106 vapour losses above minus 25 °C. Ethylene |minus 40 Can be regenerated. Lower vapour Glycol losses and less solubility in liquid | (uz) hydrocarbons than methanol. Significan vapor losses above minus 1°C. | Diethylene| minus 10 Use only in warmer, low pressure Glycol Isystems where glycol losses are high (DEG) lor where glycol dehydration is used in conjunction with glycol injection. Table 2.3.2 Comparison of Inhibitors for Hydrate SuppressionPROCESS DESIGN HANDBOOK REV. DATE: 1 JUNE 1984 REV:0 PAGE:6 of 27 ISSUE DATE: 25 MAY 1984 SECTION: 2.3 —~e stat-y. 2.3.4 Hydrate Inhibitor Injection A typical MEG injection and regeneration system for an offshore gas processing system is shown in Figure 2.3.3. The MEG injection system is combined with the hydrocarbon dewpoint control system to inhibit hydrate formation. The achievable water dewpoint for the sales gas is a direct function of the operating temperature and pressure in the dewpoint control separator. MEG injection is required at any point where hydrates may form e.g. upstream of heat exchangers and control valves. MEG S injected into the process and condensate are separated in the dewpoint control separator and fed to regeneration system. 2.3.4.1° Inhibitor Injection Rates The Hammerschmidt equation is used to calculate the minimum weight percent inhibitor as a function of the required hydrate depression and inhibitor molecular weight: Woe where: M * molecular weight of the inhibitor (Methanol: 32, MEG: 62, DEG: 106) : W = weight percent of the inhibitor in the solution ? = hydrate depression (or the difference between the normal and depressed hydrate point) in °C. K = constant (1297 for methanol, 2220 for glycols) The above equation applies only to typical natural gases and to solute concentration less than 0.20 mole fraction. Calculate the amount of water to be removed from the gas from the difference in the feed and product gas water contents (refer to Figure 2.3.1). \ j\ PROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION: 2.3 REV. DATE: 1 JUNE 1984 REV:0 PAGE:7 of 27 statoil REBOILER OAS OuILeT SEPARATOR cas, PROCESSING Figure 2.3.3 Ethylene Glycol Injection SystemPROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION: 2.3 —™, REV. DATE: 1 JUNE 1984 © REV:0 of 27 stal_i The minimum inhibitor injection to satisfy the inhibitor concentration required is given by: 7 (Wt. Inhibitor) (100) W (we $) We. Inhibitor + Wt. liquid water where: Wt inhibitor weight of 100% inhibitor injected Wt liquid water = weight of water injected with the inhibitor plus weight of water removed - \ from gas. 2 ? Experience has shown that concentrations should be of the order of: : Methanol - 98 wt % inlet concentration -- 90 wt % in the solution (outlet) : MEG - 70 to 80 wt & inlet concentration - 60 wt & in the solution outlet. Methanol injection rates for 100 wt % inlet methanol can be calculated from Figure 2.3.4. These rates allow for methanol vaporising into the gas strean. For MEG injection, Figure 2.3.5 shows the minimum weight percent outlet glycol concentration versus required hydrate depression and Figure 2.3.6 shows the minimum injection rate versus inlet concentration (normally 80%) with parameter of glycol outlet concentration. These are based on the Hammerschmidt equation with no design margin. As a minimum, a design margin of 50% should be added to the calculated inhibitor injection rate to allow for maldistribution. In addition, for glycol injection upstream of process equipment (e.g. heat exchangers), enough glycol should be injected to ensure complete ‘wetting’ of the equipment. Typically, a minimum injection rate of 0.5 m*/h per 10°sm?/day of gas processed will be required. ! The physical properties of methanol and glycols are given in ! physical data, Section 6.1PROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION: 2.3 : - REV. DATE: 1 JUNE 1984 REV:0 PAGE:9 of 27 statoil WORATE Tw EDUC OF HYDRATE Tew AEDUCNON,“c “CHET tee Se Ee ts Py i | Hh v0 7 } 3 J P| £ epression and 3 i gr] Withdrawal Concentration 2 1 for Ethylene o-8 Glycol : 0-8 ° 70 a 70 OLYCOL WITHORAWAL CONCENTRATION 2 BY WE GLYCOL INJECTION RATE « seg GLYCOL SOLUTION ig 10 TO BE ReSTED . eerie VS : 2 KN [Bien eter 5. l vB werlur | T1335 Z| Coe ed | E434 Figure 2.3.6 g NS = EL 5S 4 t-182t Ethylene se {4 SSF glycol el {4 3 Injection By Ty I : ParametersPROCESS DESIGN HANDBOOK ISSUE DATE: _25 MAY 1984 SECTION: 2.3 * REV. DATE: 1 JUNE 1984 © REV:0 pace:11 of 27 «Statoil To determine the reboiler duty proceed as follows: Select reboiler operating temperature, T(REB), from the boiling point of lean glycol, as follows: Concentration & 70 15 80 es T(REB), OC 116 120 125 131 Calculate duty of glycol/glycol exchanger using lean glycol side: Og vc. 7 where: Q, = Duty of glycol/glycol exchanger, W x = Glycol injection rate, kg/s . c = average specific heat lean glycol , 3/kgk T = T(REB) - T(INJ), OC A minimum injection temperature of 20°C should be used. Calculate temperature of rich glycol entering reboiler, by heat balance around the glycol/glycol exchanger. The glycol return temperature will be fixed by process conditions at the separation point. Calculate the temperature of the rich glycol, entering the reboiler, T(IN), by heat balance around the glycol/glycol exchanger. The glycol return temperature will be fixed by the process conditions at the separation point. Calculate reboiler duty as follows: Op gel aCe teexeL Qp = Reboiler duty, W 2 = Rich glycol return rate, k9/s = Ave specific heat rich glycol, J/kgKPROCESS DESIGN HANDBOOK i ISSUE DATE: 25 MAY 1984 SECTION: 2.3 se REV. DATE: 1 JUNE 1984 REV: PAGE:12 of 27 stal_ hi T = T(REB) - T(IN), OC | X = Total water removed, kg/s L = Latent Heat Water, J/kg | 2.3.4.3 Weight and Space Estimates (glycol injection) A glycol regeneration skid normally consists of reboiler, flash | drum, glycol/glycol exchanger, surge drum, filters and injection : pumps. } The regeneration skid is very similar in size and weight toa TE) i regeneration system of equivalent reboiler duty. Refer to . Section 2.3.5.3 and Figure 2.3.13 for preliminary sizing of TEG i regeneration skids. 2.3.4.4 Inhibitor Losses Very little vaporisation occurs when using glycols under appropriate conditions, and may be estimated as 0.0035m?/10°sm’ gas. Generally for first estimates use total glycol losses of 16 kg/10® sm? (1 1b/MMSCF) plus entrainment losses in the hydrocarbon phase. Typically entrainment of 200 ppm will be il achieved but with coalescers this can be reduced to 50 ppm vol. 4 For methanol, the vaporisation loss is significant and must be included when calculating the injection rate required. The vaporisation loss may be estimated from Figure 2.3.7. a | - Y\ PROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION: 2.3 . REV. DATE: 1 JUNE 1984 = REV:0 PAGE:13 of 27 statoil 1,900 700: 500: 300: 200 PRESSURE , bara gsosas 8 10 0 6 20 25 30 35 40 kg METHANOL per 106 Sm3(1 bara_& 0°C) METHANOL IN WATER PHASE Pigure 2.3.7 Methanol Vaporisation Losses 2.3.4.5 Design Considerations Generally methanol injection systems are not used for continuous injection offshore. This is due to the high methanol losses normally associated with these systems unless complex recovery schemes are used. However, a methanol injection system is normally included -as a back-up to any platform dehydration systen. Methanol is easily dispersed in the gas phase and quickly dissolves any hydrates that may form due to malfunction of the dehydration system.PROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION: 2.3 e REV. DATE: 1 JUNE 1984 REV:0 PAGE:14 of 27 stat_i Ethylene glycol injection is usually preferred to diethylene | glycol due to its lower viscosity, lower solubility in hydrocarbons and lower freezing temperature. The following points should be considered when designing a glycol injection system: a The separation between the glycol and hydrocarbon is critical to avoid excessive glycol losses and entrainment of hydrocarbon in the glycol fed to the regenerator. The separation usually requires heating of the glycol/hydrocarbon mixture to 15°C and 15 to 20 -~. minutes residence time. i Coalescers should be considered on the hydrocarbon stream leaving the glycol/hydrocarbon separator to minimise losses due to entrainment of glycol in the hydrocarbons. Generally injection systems are more favourable where only small quantities of condensate are removed in the dewpoint control separator. If large quantities of condensate are produced, the glycol/hydrocarbon separators can become very large to give the requried residence times for separation. i Poor distribution of glycol in the gas phase is a cause of many operating problems in glycol injection systens. ~~ | Care must be taken with injection nozzle design to ensure adequate pressure drop is allowed to atomise the ( glycol solution and turndown requirements can be met. ‘ Distribution of the glycol solution to each injection point should be carefully evaluated to avoid interaction Glycol solutions freeze below about 10°C at certain concentrations. . In selecting the inlet and operating concentrations for the system the freezing point curves should be checked. Generally glycol concentrations in the range 60 to 80% are suitable for most operating temperatures. | between injection points. | |PROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION: 2.3 a REV. DATE: 1 JUNE 1984 REV:0 PAGE:15 of 27 statoil 9 For recovery of glycol from wellstream fluids care should be taken over selecting materials of construction of the regeneration equipment resistant to salts in the wellstream fluid. 2.3.5 Gas Dehydration by Glycol Contacting A schematic diagram of the glycol contacting process is given in Figure 2.3.8. This process works on the principle of absorbing water from the gas stream to meet the required process water dewpoint specification. The water dewpoint achieved is primarily a function of the lean glycol concentation for a given set of process conditions. Wet gas entering the absorber column is counter currently contacted by a concentrated triethylene glycol (TEG) solution to produce a dry outlet gas stream from the absorber. The glycol regeneration system receives water rich glycol from the absorber column and regenerates the glycol to its original concentration by boiling off the water. Stripping gas is used if required to obtain a high concentration TEG solution. 7 2.3.5.1 Contactor Design The maximum allowable gas velocity, V,,,, in a bubble tray column is calculated from the following formul: Yaar = UR (OP, = pgl/ Pg)* B/s where: K 5 Constant, 0.0488 for a 24" (610mm) tray spacing and 2 inch (50mm) seal over bubble cap slots. Py = Density of TEG p, = Density of Gas 9 Knowing the volumetric flowrate of gas and maximum gas velocity, the minimum internal diameter is calculated.PROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION: 2.3 + REV. DATE: 1 JUNE 1984 REV:0 PAGE:16 of 27 statu’ TRIM COOLER Figure 2.3.8 Glycol Contacting/Regeneration System1 ( PROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION: 2.3 . REV. DATE: 1 JUNE 1984 = REV:0 PAGE:17 of 27 statoil For other than 24 inch tray spacing, the following constants (K) may be used: Tray Spacing (mm) 457 508 559 660 ni 762 Constant (m/s) 0.039 0.043 0.046 0.050 0.051 0.053 As a first approximation to contactor height (Tan/Tan) allow an extra 3 metres over the height estimated from the trays (number and spacing). Estimate number of trays from Figure 2.3.9. As a starting point use a TEG circulation rate of @i5 m°/kg water removed (3 US gal/1b). oe Figures 2.3.10, 2.3.11 and 2.3.12 can be used to calculate required TEG concentration from contact temperature and dewpoint depression, for 6, 8 and 10 trays respectively. 2.3.5.2 Regeneration For TEG, the reboiler is normally operated at 204°C (400°F) and atmospheric pressure. At these conditions, a lean solution containing about 98.4 to 98.7 wt * TEG can be produced. For any higher required concentration, stripping gas and/or vacuum is required. To calculate stripping gas rates use (Reference 3): PEG CONCENTRATION STRIPPING GAS wre SCF/GAL TEG Sn?/m? TEG 99.20 0 ° 99.70 1 1.5 99.84 2 14.9 99.90 3 22.4 Refer to Campbell (Reference 2) for nomograph for estimating the effect of stripping gas, vacuum and reboiler temperature on the performance of a regeneration unit.PROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION: 2.3 =) REV. DATE: 1 JUNE 1984 REV:0 page:ia of 27 StQwi GLYCOL TO WATER CIRCULATION RATE me? TES /eg 20 2 201 0.07 001 9.0 095 0.06 0.0 CURVES FOR REQUIRED . DEWPOINT DEPRESSION A | AN g> ¥ ss Fess wy sae) 3 s' ure 2uF 7s Ruz'e) Figure 2.3.9 : : oF (360) [s5° F (20°C) Tray 2 Requirements 1 1 2s ¢ 8 6 7 8 8 GLYCOL TO WATER CRCULATION RATE US GAL. TEG./ LB.H20 BASIS CIRCALATION RATE + 0.0251m4/4913US gal/T) tg UMBER OF ACTUAL TRANS 60 2s 5 x Es 8 | zo Sas Figure 2.3.10 8» Concentration Requirements 2 is om 0 3 > 60 88 Go es 6 Trays | Contact TEMPERATURE ’CPROCESS DESIGN BANDBOOK ISSUE DATE: 25 MAY 1984 SECTION: 2.3 REV. DATE: 1 JUNE 1984 REV:0 pacz:19 of 27 Statoil DEWPOINT DEPRESSION ‘C DEWPOINT DEPRESSION °C " BASIS :REULATION RATE» 0.025 m37kg(9 USal/ib} NUMBER OF ACTUAL TRAYS © 20. aT *T a le = 199-7%e a Eee 60 58 5 an. 7% 45 + 4 —— 8 20 28 +30 3 40 & 50 55 60 © CONTACT TEMPERATURE ‘C BASIS: CIRCULATION RATE « 0.025 m3/kg(3 US pal/b) NUMBER OF ACTUAL TRAYS «8 ce 2 65, 60) 85 80: “ 40 *ip go ds 30 a5 co ds 80 5 G0 65 CONTACT TEMPERATURE °C Pigure 2.3.21 Concentration Requirements Trays io Pigure 2.3.12 Concentration Requirements ae Trays &PROCESS DESIGN HANDBOOK ISSUE DAT! REV. DATE: 1 JUNE 1984 RE 25 MAY 1984 SECTION: 2.3 os PAGE:20 of 27 stat_.i As a first approximation total reboiler duty Q, can be calculated from (reference 2): 557000 kJ/m? of pump circulation rate (2000 BTU/US gal of pump circulation rate) me —The reboiler may be direct fired, using fuel gas, or heated electrically or with a heating medium such as hot oil. 2.3.5.3 Weight and Space Estimates < The weight of the contactor can be calculated based on the D information given in Sections 5.1 and 5.11. The regeneration skid normally consists of reboiler, flash drum, glycol/glycol exchanger, surge drum, filters and circulation pumps. Figure 2.3.13 gives approximate plot area and weight (dry and operating) for a conventional regeneration skid based on reboiler duty. The data is based on low concentration units (no stripping gas or vacuum), direct fuel gas fired. The height of the skid is generally 4 to 5 metres, with about 7 to 8 metres to the top of the still column. For high concentration units (above 99%) using stripping gas, the skids will be larger, particularly in the height of the still ~ column. In all cases, a specialist vendor should be consulted to obtain the best design for the required duty.PROCESS Di ISSUE DATE REV. DATE ESIGN HANDBOOK 25 MAY 1984 SECTION: 2.3 e 1 JUNE 1984 REV:0 PAGE:21 of 27 statoil 3 £ = é 40 2 30. zg { & 20 i 50 o ‘tS 2 os = 10 40 = * = 0 30 < 3 20 10 2 0 200 400 600 800 1000 1200 REBOILER DUTY, kW Figure 2.3.13 TEG Regeneration Skid Sizing 2.3.5.4 Design Considerations Traditionally TEG contacting offshore has generally beén limited to relatively high water dewpoint specifications, greater than 20 mg/Sm?. Most vendors now can offer designs based on counter current stripping of the glycol solution and claim achievable water Gewpoints of 1 ppm water in the treated gas stream. There are currently two systems proposed offshore for water dewpoint specifications of around 8 ng/Sm?.PROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION: 2.3 ty REV. DATE: 1 JUNE 1984 © REV:0 PAGE:22 of 27 StQL_L b. The location of the TEG contactors should be carefully considered taking into account possible large increases in gas volume during the depletion period of gas reservoirs. Ethylene glycol and methanol injected upstream of a TEG contacting unit’ can cause potential operating problems. Ethlylene glycol will degrade at the normal TEG regeneration temperatures and methanol can build up in the regeneration still. Foaming in the contactor can cause loss of production S and excessive glycol losses. It is normally caused by carryover of heavy hydrocarbons or condensation of hydrocarbons in the absorber. To avoid condensation the lean TEG temperature should be maintained at least 6°C above the feed gas temperature. Carryover can be minimised by adequate sizing of the knock out drum upstream of the.TEG contactor or by the installation of ‘a filter separator. Installation of a double mist extractor in the contactor to reduce carryover of glycol and hence top-up requirements. Consider using the bottom section of the contactor as am knockout drum for hydrocarbon liquids. This can provide / a space saving, provided the column is still within height limitations. Por high pressure contactors, consideration should be given to installing a power recovery turbine on the rich glycol stream from the contactor. The power recovered can be used to partially drive the glycol circulation ' pump. 1PROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION: 2.3 s REV. DATE: 1 JUNE 1964 RE’ pace:23 of 27 StaQtoil 2.3.6 Gas Dehydration by Adsorption 2.3.6.1 General Several materials may be used for dehydration by adsorption, i.e. Bauxite, Alumina, Silical Gels and Molecular Sieves. O£ these, molecular sieves have achieved most widespread acceptance. This section only considers molecular sieves but the design procedure is similar for all materials. A schematic diagram of a molecular sieve adsorption process is given in Figure 2.3.14. In this process water is removed from the gas stream by adsorption onto solid molecular sieve desiccant. The molecular sieve desiccant removes essentially all water from the process = stream and the design of the unit is virtually unaffected by the - process water dewpoint specification. 2.3.6.2 Preliminary Sizing Adsorption sizing is given for molecular sieves as follows: Sieve bed quantity +: 10 kg sieve/kg water adsorbed Bulk Density, average: 722 kg/m. Bed diameter is calculated from a maximum superficial gas velocity criteria set by Figure 2.3.15 and bed height is estimated from the volume of sieve calculated above. Regeneration gas requirements can be calculated using Figure 2.3.16. Adsorption and regeneration cycle times are determined by trial and error to minimise topsides weight, space and utility requirements. Account must also be taken of the effect of the number of regenerations on bed life and desiccant adsorption capacity. Generally, a three bed system is utilised with two beds on adsorption and one on regeneration. For most natural gas dehydration systems optimum cycle times will be in the region of 4 to 8 hours.PROCESS DESIGN HANDBOOK stati’ 12.3 PAGE:24 of 27 SECTION 0 25 MAY 1984 DATE: 1 JUNE 1984 ISSUE DATE REV. 131100 Sv9 AN “SNTIOOD, waTi0MINoD 39Nan0as Figure 2.3.14 Molecular Sieve Adsorpti on Systemf r r ozt ok 8 09 dy ae 0 (o10@) gunssaua onuvusao squewertnbey ) gunnvazeWaL GF34 oo 08 kk ; seo a 0 ' uopaersusboy BE duet 108 vondiavaton - L__, Gugsorpe a4g au Suoay 28 gtre*z 2anb ya wodospe mo anes 3 unt jo uoomustas__y 3 suniels m8 ION 3 : sisva z 2 2 a ; a 2 ' 3 a - 3 ef oo we st i (a) aunimuaarat aa ARyOOTAA (9994) aunsszid ONIVEadO pepopzzedns cor 0 oe j - unayxey te*z eanbr, ster ed : ng i z z \ = ae g 3 i 3 «§ | 2 z a ae a 3 a 8 g | Soz-o4 aa z i Z azo = | tH ” n o | &z 30 szigowd oraze ye6T SNOC T ?auva “Age ’ HO WOAS €°% :NOILOgS yaeT AWW SZ gu¥d ansst fl MOOGANYH NOISSG ssa0udPROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION: 2.3 = REV. DATE: 1 JUNE 1984 RE PAGE:26 of 27 stat ~}! Because of the large number of permutations of number of beds and cycle times, it is usual to approach vendors to size the system even at a preliminary design stage. 2.3.6.3 Design Considerations Generally high utility demands and high system weight rule out solid bed dehydration systems offshore. However, for low throughputs, up to a maximum of 3.0 x 10°sm3/day solid bed dehydration systems can be competitive. Also, if low water dewpoints are required, molecular sieve systems become more ~ competitive due to the increased complexity required for TEG D regeneration. The following points should be considered when designing a solid bed dehydration system: a. The process operates on a batch basis. For this reason it is more difficult to recover heat added into the system and explains the high utility requirements associated with a solid bed dehydration system. b. Regeneration gas should preferably be at the same pressure as the feed gas to minimise pressuring and depressuring operations and to avoid condensation when operating in the retrograde region. e When dehydrating a dewpoint gas in the retrograde region, the design should take into account the om condensation of hydrocarbons on the bed due to the - pressure drop through the bed. This may require the addition of a feed gas heater upstream of the dehydration unit. a. Internal insulation may be advantageous for high Pressure designs since it will minimise the heating and cooling requirements associated with the dehydration vessels. Operating experience with internal insulation is varied, some systems have given no problems whilst others have encountered operating difficulties due to breakdown of the insulation. In any evaluation comparing internal and external insulation allowancePROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION: 2.3 i REV. DATE: 1 JUNE 1984 = REV:0 PAGE:27 of 27 statoil 2.3.7 should be made for the increased vessel diameters required with internal insulation. Regeneration gas rates and heating and cooling duties can be reduced by having two beds on series regeneration at any one time, one bed being heated while the other is cooled. This increases topsides weight and space requirements and also doubles the total number of regeneration gas switching valves required. Thus a detailed evaluation must be performed to compare utility savings against increased complexity and weight. Molecular sieves can be designed for removal of hydrogen sulphide and carbon dioxide. The general order of adsorption is water, hydrogen sulphide and carbon dioxide. Designing for simultaneous removal of water and acid gases can increase bed sizes significantly even for only trace quantities. In addition co-adsorbed acid gases will appear at a peak concentration in the regeneration cycle and may lead to off specification ‘sales gas if regeneration gas is fed into the sales gas stream. Operating experience with solid bed dehydration systems is varied. Most problems occur as a result of switching valve failures which can lead to water breakthrough and hydrate formation in downstream equipment. Other failures generally occur as a result of damaging the structure of the desiccant, e.g. by carryover of large quantities of liquid hydrocarbons, contamination by wellstream chemicals, pressuring/depressuring too quickly. References 1. Katz, et al, ‘Handbook of Natural Gas Engineering’. 2. Dr. J.M. Campbell, 'Gas Conditioning and Processing’ Vol. 2. 3. GPSA, Engineering Data Book.PROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION: 2.4 —“yt REV. DATE: 1 JUNE 1984 REV:0 ~ PAGE:1 of 11 stat i 2.4 Dewpoint Control 2.4.1 General Processing Equipment is required to meet the hydrocarbon dewpoint specification for the product gas from the platform. The processing requirements are dependent on the destination of the treated gas. The possible destinations are: a. Gas lift compressors and wells - objective is to ensure single phase flow from the gas lift manifold b. Pipeline to a further gas treating/condensate recovery ») plant. The specification should ensure single phase flow in the pipeline and take into account pressure/temperature profiles in the pipeline c Sales gas grid - typical sales gas specifications are given in Section 2.2. The sales gas specification will generally be the most severe of the options given above. This specification may be stated as a dewpoint of a specified pressure and temperature which will define a phase envelope for the product gas. A schematic of a phase diagram is shown in Pigure 2.4.1 which gives the terminology for various points: Cricondenbar : maximum pressure at which liquid may ) exist a Cricondentherm : maximum temperature at which liquid and vapour may co-exist in equilibrium Retrograde Region + that area inside phase envelope where condensation of liquid occurs by lowering pressure or increasing temperature Quality Fractions : those lines showing constant gas volume . weight percent which intersect at the critical point.PROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION:2.4 . REV. DATE: 1 JUNE 1984 REV:0 PAGE:2 of 12 statoil Figure 2.4.1 Phase Envelope Terminology The phase envelope will be determined by computer simulation. The selection of the dewpoint control system must consider interactions with the other process units. A simplified block schematic for an of1 and gas processing facility is shown in Pigure 2.4.2. This shows the usual location for the dewpoint control system. NGL produced in the dewpoint control unit is fed into the condensate/crude stabilisation system. The lighter components are separated from the condensate and back to the main gas processing stream upstream of the dewpoint control unit.PROCESS DESIGN HANDBOOK . ISSUE DATE: 25 MAY 1984 SECTION: 2.4 . 4 REV. DATE: 1 JUNE 1984 REV:0 PAGE:3 of 11 staiwi embeanon ‘cone roe Strova rection eee Corot Figure 2.4.2 Simplified Process Scheme 2.4.2. Dewpoint Control Methods The main schemes for dewpoint control are: A External refrigeration : Joule Thomson expansion (adiabatic process) O Turbo Expander (isentropic process) The selection and optimisation of the dewpoint control system is ~ achieved using computer simulation. Simplified flowschemes for the dewpoint control schemes are shown in Pigure 2.4.3. For each of the schemes the cold sales gas and condensate streams are reheated against incoming feed gas for cold recovery, providing a more thermally efficient cycle. The increased efficiency is achieved at the expense of additional i equipment items and should be subject to optimisation. ; |PROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION: 2.4 er REV, DATE: 1 JUNE 1984 REV:0 PAGE:4 of 11 statoil External Refrigeration Joule Thomson Turbo Expander Figure 2.4.3 Dewpoint Control SystemsPROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION: 2.4 an REV. DATE: 1 JUNE 1984 REV:0 PAGE:5 of 11 stat 2.4.3 Selection Method The first step in the selection of the dewpoint control system is to carry out simulations to define a product gas dewpoint curve and if appropriate the feed gas phase envelope. The possible paths can then be back calculated starting from a point on the product gas dewpoint curve. Figure 2.4.4 shows a typical phase envelope for a gas/condensate processing facility. The thermodynamic paths shown are: = . Adiabatic - expansion through a Joule Thomson 7 (isenthalphic) valve Isentropic - expansion through a turbo expander This figure shows that for a given initial pressure and temperature condition the isentropic path gives a greater temperature decrease for a given pressure change. Figure 2.4.4 also shows that for a Joule Thomson process more chilling upstream is required to arrive at the same point on the dewpoint curve. This would require more heat exchanger equipment or an external refrigeration package. Alternatively, starting at the same temperature the Joule Thomson process will reach a lower pressure on the phase envelope and will require more sales gas =~ compression. A refrigeration process could be used based on a Joule Thomson expansion across the wellhead choke valves followed by cooling. The path for such a process is shown in Figure 2.4.4. ‘The optimum pressure and temperature condition in the dewpoint separator are dependent on the gas composition, gas specification and export pressure and must be established for each case. The following guidelines can be given:PROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION: 2.4 7 REV. DATE: 1 JUNE 1984 — REV:0 PAGE:6 of 11 statoil 320 20 RESERVOIR FLUD 200 T+ | PHASE ENVELOPE me 7 HYDROCARBON DEWPOINT tno 7 peesne 0 ‘ADABAtIC Z PRocess. 160 + iSENTROME X 180 PROCESS uo 17 Vi 130 « i z wa B s204— BeLericarion, ° trea At ABIABATIC pRORESE 2 yo fH terotss fBUEE NSouson 3 B 0 E | © s0 + Probus Gas! Seana Goave_ |e *0 Is | 7 *e | POSSIBLE REFRIGERATION | PATHS so 7 so + 20 20 +t © ° mr TT ee ee aT 260 50-0 30-20-10 0 70 OH BD 60 7 80 w 100 nO 10 TEMPERATURE °C Figure 2.4.4 Joule Thomson ané Expander PathsPROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION: 2.4 ne REV. DATE: 1 JUNE 1984 © REV:0 PAGE:7 of 11 stat Z a. for a lean gas the optimum pressure condition will be above the cricondentherm pressure. Assuming the starting pressure is high this condition will give the dewpoint specification at a high pressure thereby minimising the gas export compression requirements. Optimisation should consider the compressor driver power limitations. Operation at pressures closer to the cricondenbar should be avoided because of the low temperatures involved which may impact on the water dewpoint requirements and also the increasing thermodynamic prediction uncertainty. » b. for a rich gas the optimum pressure condition will be lower on the phase envelope than for a lean gas. The low pressure condition gives a higher temperature and less light components (NGL's) avoiding high recycle rates. For a typical sales gas specification the optimum dewpoint separator pressure is expected be in the range from 50 barg to 70 barg. Every dewpoint control system is unique in terms of feed.quality and quantity, operating pressure and temperatures and gas specification. Hence it is not usually possible to preselect the best process. . ~ Advantages and disadvantages of the schemes can be summarised ana ~ are given below. 2.4.3.1 External Refrigeration a if the available process pressure corresponds to the export pressure and the dewpoint specification can be achieved at a reasonable temperature then this scheme is favourable. This method is used at low pressure where an expansion process is not feasible b. requires an external refrigeration package c. not good for very low temperaturesPROCESS DESIGN HANDBOOK Oo ISSUE DAT! 25 MAY 1984 SECTION: 2.4 e REV. DATE: 1 JUNE 1984 © REV:0 PAGE:8 of 11 statoil 2.4.3.2 Joule Thomson a. Useful where a pressure drop is available between supply and demand. b. Minimum equipment requirement. c. Inefficient at cooling gas. a. Not good for very low temperatures. 2.4.3.3 Turbo Expander a. Useful where a pressure drop is available between supply and demand. b. Involves high speed rotating equipment. c. Efficient gas cooling a. Good for very low temperatures if sufficient pressure érop is available. 2.4.4 Optimisation The critical parameters to consider for the optimisation of the Gewpoint control system are: a. wellhead flowing pressures and depletion as the reservoir is produced. Depletion of the reservoir may change the optimum dewpoint control system selection b. sales gas dewpoint and water content specifications c. sales gas export pressure a. crude or condensate specification e. flowrate and composition Hence the optimisation of the hydrocarbon dewpoint control system can not be considered in isolation from the other process systems. . The following guidelines can be given: a. the optimum process conditions should be selected to fully utilise the power available from a given sales gas turbine driver. Decreasing the dewpoint pressure can result in a requirement for additional drivers withPROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION: 2.4 stat REV. DATE: 1 JUNE 1984 REV:0 PAGE:9 of 11 To consequent large space/weight penalties (typically 300 to 400 tonnes) for a Joule Thomson process compared to an Expander process there will be a greater chilling or sales gas compression requirement. The Expander process requires additional equipment compared to the Joule Thomson Process. If the weight and space of this additional equipment exceeds the chilling/sales gas compression requirements for the Joule Thomson process then the Joule Thomson process is preferred. SS) Optimisation is usually a balance between the amount of compression power and the cold recovery from the various Process streams. A critical parameter is the temperature upstream of the Joule Thomson or Expander process. Varying this temperature will give various ‘levels of cold recovery and sales gas compression’ power. These can be investigated to arrive at the optimum process conditions. 7 for refrigeration, optimisation will generally involve adjusting the processing pressure to vary the refrigeration requirements and sales gas compression power. The depletion pressure for the reservoir will also effect the optimum processing pressure. ~~ select the dewpoint control process preliminary optimisation should be carried out for each process option. The processes can then be compared on weight, space, utility requirements and operability. Generally a detailed cost comparison will not be necessary.PROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION: 2.4 * REV. DATE: 1 JUNE 1984 — REV:0 paGe:10 of 11 StQkofr 2.4.5 | Simulation Guidelines Heat exchanger approach temperature can be set as follows: O Gas/Gas exchanger approach between hot and cold streans usec 5 Refrigeration cooler approach between cold gas and refrigerant occ streams These temperature approaches represent initial estimates and should be subject to optimisation as the design is developed. For a turbo expander/recompressor scheme typical polytropic efficiencies ar E Turbo Expander 75% Recompressor 658 other criteria for the turbo expander selection are: : maximum proven machine (approx) 5 MW 0 maximum pressure ratio 2.5:1 : Liquid content of outlet up to 408 wt For initial estimates approximate temperature reductions on expansion are as follows: Cycle Rich Gas Lean Gas °C/bar OC/bar J-T expansion 0.24 0.4 Turbo-expansion 0.48 0.8 Pressure restoration by turbo-expander recompressor may be estimated by: Pout = PinxP,K Pout - Recompressor outlet pressure Pin = Recompressor inlet pressure Py = Turbo-expander pressure ratio K = 0.35 when vapour reheated 0.50 when vapour coldPROCESS DESIGN HANDBOOK ISSUE DATE 25 MAY 1984 SECTION:2.4 Sy 1 JUNE 1984 -REV:0 pace:ilof 1. Stat il os Bis g a s 7 En = PERCENT OF OPTIMUM SPEED He er ae PERCENT OF DESIGN FLOW RATE PERCENT OF DESION EFFICIENCY E 3 PERCENT OF DESION HEAD eae go 46000 wo mo uo 180 PERCENT OF DESION FLOW Figure 2.4.5 Typical Turbo Expander Performance CurvesPROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION: 2.5 REV. DATE: 1 JUNE 1984 REV:0 PAGE:1 of 21 stato/ 2.5 Acid Gas Removal 2.5.1 General As a general rule it is not normal to remove acid gases offshore. Removal of large quantities of hydrogen sulphide and sulphur compounds is impracticable due to limitations on sulphur emissions to the atmosphere. However, to protect-gas turbine drivers, fuel gas treatment may be required to remove hydrogen sulphide to approximately 50 ppm. Bulk removal of carbon dioxide may be required under the following conditions: a. Where economics dictate i.e. where savings in compression and transportation costs outweigh the cost of the treating unit. In order to use existing pipelines which have a limitation on carbon dioxide content. For enhanced oil recovery (EOR) schemes that require injection of carbon dioxide into the reservoir. a. To avoid carbonic acid corrosion in wet pipelines. 2.5.2. Available Processes Gas treating processes can be divided into four general categories: 5 chemical solvent processes . physical solvent processes : ary-bed processes S membrane processes The major processes in each of these categories are summarised below:PROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION: 2.5 =~ REV. DATE: 1 JUNE 1984 © REV:0 PAGE:2 of 21 stat_ il Chemical solvent | Physical solvent | Dry bed Membrane Processes Processes Processes Processes fk. Alkanolamines | Sulfinol Mol Sieve Monsanto MEA Selexol Iron Sponge | prism DEA Pluor Solvent SNEA DEA Rectisol DGA MDEA MEA/DEG ~) IB. Alkaline salt Solutions Hot Pot Catacarb Benfield 2.5.2.1 Chemical Solvent Processes A. Alkanolamines These processes are characterised by: : high heats of absorption. high utility requirements. relative insensitivity of the treated gas specification and circulation rate to the absorption pressure. A typical amine unit process flow diagram is shown in Figure 2.5.1. MEA Advantages are low solvent cost, good solvent availability, high reactivity and low hydrocarbon content of the acid gas produced. Disadvantages are some irreversible degradation by carbonyl sulphide, carbon disulphide and oxygen in the feed gas, higher vaporisation losses than DEA and reversible degradation by carbon @ioxide which requires use of a reclaimer.PROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION: 2.5 REV. DATE: 1 JUNE 1984 REV:0 PAGE:3 of 21 Oo stato” INEANED ons. Pigure 2.5.1 typical Amine Treating UnitPROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION: 2.5 REV. DATE: 1 JUNE 1984 REV:0 PAGE:4 of 21 stat 7} To control corrosion, MEA solution concentrations are generally restricted to 15 to 20 wtt. At these concentrations it generally cannot compete against alternative treating processes. However, for carbon dioxide removal with no hydrogen sulphide present, it is possible to use solution concentrations of up to 30wtt by using proprietary corrosion inhibitors. At these concentrations MEA can become competive with other treating processes. Snamprogetti license an MEA/DEG design utilising a solution consisting typically of 30 wtt MEA, 50 wtt DEG and 20 wtt water. This gives simultaneous dehydration and acid gas removal. The achievable water dewpoints are, however, limited and generally an additional dehydration system is still required for most gas processing applications. DEA Advantages are good solvent availability, resistance to degradation from carbon’ dioxide, carbonyl sulphide and carbon disulphide (reclaimer is not required), and low hydrocarbon content of acid gas produced. Disadvantages are lower reactivity than MEA, higher solvent circulation rates (conventional system) and higher solvent costs, DEA is often used to treat refinery gas streams because of its resistance to degradation from sulphur compounds. ~ = SNEA (P) DEA For high acid gas partial pressures the SNEA (P) DEA process becomes highly competitive. This process uses a higher DEA solution strength, typically 30 to 40 wtt, and much higher loading of DEA. The process is licensed by SNEA (P) and was developed from operating experience gained at the Lacq gas fields in France. DGA Advantages are low circulation rates, lower utilities than MEA and DEA (conventional) and low solution freezing point. Disadvantages are poor solvent availability, higher solvent costPROCESS DESIGN BANDBOOK ISSUE DATE 25 MAY 1984 SECTION: 2.5 ° REV. DATE: 1 JUNE 1984 REV:0 PAGE:5 of 21 statoi! and reversible degradation by carbon dioxide, carbonyl sulphide, and carbon disulphide which requires use of a reclaimer. DGA offers advantages over conventional DEA and MEA processes because it can be used at higher solution concentrations, 65 to 80 wtt, without severe corrosion problems. This reduces circulation rates and utility requirements. The limited solvent availability and high cost has generally limited its use. MDEA MDEA is a tertiary amine and does not react directly with carbon dioxide. It does, however, react with carbonic acid and bicarbonate ions in solution but this results in a low overall rate of reaction. Hydrogen sulphide reacts rapidly with MDEA. The difference in reaction rates between carbon dioxide and hydrogen sulphide has led to MDEA being used to selectively remove hydrogen sulphide from gas streams in the presence of carbon dioxide. MDEA is also now being used for bulk removal of carbon dioxide. BASF have developed an MDEA process with an activator to increase the absorption of carbon dioxide. The advantage of this process is that the solution can be regenerated using a simple flash regenertion system resulting in very low utility requirements. The process design is thus similar to a physical absorption process. Also, it has an advantage over most physical solvent processes in that it does not absorb the heavy hydrocarbons from the feed gas. The disadvantage of the process is that the relatively slow rate of reaction gives high circulation rates and a large absorber. B. Alkaline Salt Solutions (Hot Pot) These processes are similar to the alkanolamine processes except that the contactor operates at a much higher temperature, 100 to 130°C. The solvent is an aqueous solution of 25 to 35 wtt potassium carbonate. These processes are generally used for bulk carbon dioxide removal but will also remove hydrogen sulphidePROCESS DESIGN HANDBOOK ISSUE DAT! 25 MAY 1984 SECTION: 2.5 ~ REV. DATE: 1 JUNE 1984 RE’ stat_il PAGE:6 of 21 provided the ratio of hydrogen sulphide to carbon dioxide is low. The advantages of ‘hot pot’ processes over alkanolamine processes are lower utility requirements, effective removal of carbon disulphide and carbonyl sulphide, lower vaporisation losses and elimination of solution/solution cross exchangers. Disadvantages are higher circulation rates, inability to meet ‘low’ treated gas specifications, irreversibly degraded by sulphur dioxide and trioxide, high contactor temperature may require large gas/gas exchangers, ‘salting out’ of the potassium carbonate solution requires careful design to avoid subsequent operating problems. . The Catacarb and Benfield processes are essentially the same as ~ the hot pot process except catalysts are added to improve the reactivity and increase the loading capacity of the solution. This gives lower circulation rates and utility requirements but involves payment of a process royalty. 2.5.2.2 Physical Solvent processes These processes are characterised by: low heats of solution. low utility requirements. : Co-absorption of heavy hydrocarbons (typically propane plus) from the gas stream. Sensitivity of treated gas specification and solvent ~ circulation rate to the absorption pressure. A typical physical solveht process flow diagram is shown in Figure 2.5.2. Generally physical solvent processes will show advantages over chemical solvent processes at high absorption pressures. The solution loading and treated gas specification are directly proportional to the acid gas partial pressure in the feed gas.stato” 2.5 SECTION: PAGE:7 of 21 REV:0 ISSUE DATE: 25 MAY 1984 REV. DATE: 1 JUNE 1984 PROCESS DESIGN HANDBOOK ie ‘svo wad, oxo “howuva nvewnt Awasonaw Wamos war009 NaRIOe ait awavnt Awasooay Mamoa mawa Heya aAoae v0 aia ej waavosay 15.2 Pigure 2 Typical Physical Solvent Treating UnitPROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION: 2.5 ay REV. DATE: 1 JUNE 1984 = REV:0 PAGE:8 of 21 stat_i Sulfinol The Shell Sulfinol process is unique in that it uses a mixture of chemical and physical solvents. The solvent is composed of DIPA (di-isopropanolamine), Sulfolane and water. Generally Sulfinol appears to be most competitive when the hydrogen sulphide to carbon dioxide ratio is 1:1 or greater and when the acid gas partial pressure is above 8 bara. Advantages of the Sulfinol process are low degradation rates from sulphur impurities, low corrosion rates, low foaming tendency, low vaporisation losses, effective removal of carbonyl sulphide, carbon disulphide and mercaptans and low heat exchanger fouling tendency. Disadvantages are high solvent cost, absorption of heavy hydrocarbons from the feed gas and reversible degradation by carbon dioxide which requires the use of a reclaimer. Selexol The Allied Chemicals Selexol process uses as a solvent dimethyl éther of polyethylene glycol. Generally, it is applicable for bulk removal of carbon dioxide and hydrogen sulphide for high acid gas partial pressures. Most commercial applicaitons have been for bulk carbon dioxide removal only. Advantages of Selexol are low utility requirements, effective removal of carbonyl . sulphide, carbon disulphide and mercaptans without solvent =~ degradation, low corrosion rate, low vaporisation losses and capability to dehydrate the feed gas if water is removed during regeneration. Disadvantages are high solvent cost and absorption of heavy hydrocarbons from the feed gas. Fluor Solvent The Fluor solvent process uses propylene carbonate as the solvent. Generally the Fluor solvent has the same advantages and disadvantages as Selexol. -Most commercial applications have been for bulk removal of carbon dioxide.Ne PROCESS DESIGN HANDBOOK ISSUE DATE 25 MAY 1984 SECTION: 2.5 REV. DATE: 1 JUNE 1984 — REV:0 PAGE:9 of 21 stato Rectisol The Rectisol process was developed by Lurgi and uses cold methanol for the solvent. The feed gas must be chilled to around” minus 30°C before entering the contactor. It has generally been used to purify town or synthesis gas in Europe. Rectisol has the same advantages and disadvantages as the other physical solvent processes. It also offers an advantage of very low freezing point of the solvent for applications in very cold climates. 2.5.2.3 Dry Bed Processes Molecular Sieves Molecular sieves are generally only used for removal of very low levels of acid gases in the feed. In general, the order of adsorption on the molecular sieve bed is water, mercaptans, hydrogen sulphide and carbon dioxide. Thus molecular sieves can be used for selective removal of hydrogen sulphide in the Presence of carbon dioxide. Most applications for acid gas removal by molecular sieves are made in conjunction with dehydration of the gas stream. Very low levels of hydrogen sulphide in the treated gas stream can be 1 achieved. Molecular sieve treating has found most use where it is necessary to remove only small quantities of hydrogen sulphide to meet a pipeline gas specification. The major disadvantage of molecular sieve adsorption is that the acid gases removed appear in the regeneration gas stream which may require additional treating. Their application is most favourable where the regeneration gas can be readily disposed of e.g. as fuel gas. i Iron Sponge This process uses wood shavings impregnated with ferric oxide in hydrated form. Hydrogen sulphide reacts with the ferric oxide to form ferric sulphide. The bed can be partially reactivated by air oxidation back to ferric oxide. Eventually the bed becomes plugged with sulphur and must be replaced. The process is iPROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION: 2.5 ™’ REV. DATE: 1 JUNE 1964 REV:0 pace:10 of 22 Stat il usually limited to treating small gas volumes with low hydrogen sulphide levels due to economics of bed replacement. The process essentially removes all hydrogen sulphide from the gas stream. 2.5.2.4 Membrane Processes Permeable membranes, developed from membranes used for water treatment, can be used for carbon dioxide removal. The process works on the relative permeation rates of gases through a semi permeable membrane. Carbon dioxide has a relatively high permeability compared to hydrocarbon gases and can be separated from them. The principle advantages of the process are that it does not wet the gas stream and is very simple in concept. However, high hydrocarbon losses are incurred unless a two or three stage processing scheme is utilised. This in turn requires large recompression equipment which can rule the process out on weight and space requirements. Generally the process has only been proven for very low capacity units, less than 1 x 10° sm°/a.PROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION: 2.5 e REV. DATE: 1 JUNE 1984 REV:0 PAGE:11 of 21 Stator 2.5.2 Licensed Processes Many of the commercially available gas treating processes are licensed by various chemical manufacturers, oil companies and contractors. A list of the available licensed processes is given below: Process Licensor Chemical Solvent Processes Alkanolamines MEA/DEG Snamprogetti DEA * | SNEA (P) (Hi-Load) DGA Fluor Corporation DIFAG Shell MDEA Phillips, Shell BASF (activated) Alkaline Salt Benfield Union Carbide Physical Solvent Processes Sulfinol — Shell Selexol Allied Chemicals Fluor Solvent Fluor Corporation Rectisol Lurgi Kohle Dry bed processes Molecular Sieve Union Carbide, Grace Membrane processes Monsanto prism MonsantoPROCESS DESIGN HANDBOOK ISSUE DATE: REV. DATE: 1 JUNE 1984 REV: 25 MAY 1984 SECTION: 2.5 stat_il PAGE:12 of 21 2.5.4 Preliminary Process Selection Although process selection will ultimately be based on economics, it is possible to narrow down the range of applicable processes by considering the following factors: : Acid gas partial pressure in the feed gas. Acid gas partial pressure in the treated gas. Amount of heavy hydrocarbons in the feed gas. . Effect of impurities in the feed gas. 2.5.4.1 Acid gas partial pressure The Partial pressure of the acid gas in the feed is the major factor in determining the solvent circulation rate. As the feed gas contacts the fully loaded rich solvent at the bottom of the absorber, there must be sufficient solvent to absorb substantially all of the acid gas from the feed, and provide a @riving force for absorption. The partial pressure of acid gas in the product determines the degree of solvent regeneration required, because the partial pressure of acid gas in the product must be higher than the equilibrium partial pressure in the regenerated solvent to ensure - a positive driving force for absorption at the top of the absorber. Preliminary process screening of available processes can be achieved by referencing Figures 2.5.3 to 2.5.6, which utilise the two pertinent partial pressures as the axes. These figures represent four separate treating cases (reference 2): Figure 2.5.3 Figure 2.5.4 Figure 2.5.5 Figure 2.5.6 Carbon dioxide removal with no hydrogen sulphide present. Hydrogen sulphide removal with no carbon dioxide present. Simultaneous carbon dioxide and hydrogen sulphide removal. Selective hydrogen sulphide removal.Cc PROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION: 2.5 REV. DATE: 1 JUNE 1984 REV:0 PAGE:13 of 21 stato’® | oo . = T T 4 Ll PANTIAL PREBSUNE OF CO2 IN FEED -PSIA es PARTIAL PRESSURE OF CDz IN PRODUGT-PSIA | Figure 2.5.3 Range of Gas Treating Process Application } CO, Removal with no H,S or other Impurities Present.| t PROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION: 2.5 ‘ey REV. DATE: 1 JUNE 1984 REV:0 pace:14 of 21 ©SStQuIl PANTIAL PRESSURE OF 117$ IN FEED -PSU, 7 ECONAMINE. ~ HIGH LOADING| DEA, SELEXOL, 2-—— on SULFINOL’ oe 3, oot ¢ e8or 2 a 6 at PARTIAL PRESSURE OF HS IN PRODUCT PSIA Figure 2.5.4 Range of Gas Treating Process Application. H,S Removal with no CO, or other impurities present. 2!PROCESS DESIGN BANDBOOK ISSUE DATE: 25 MAY 1984 SECTION: 2.5 ra REV. DATE: 1 JUNE 1984 REV:0 PAGE:15 of 21 Stato} 4.000 ' i | e ESE 41-— econamine. T] HIGH LOADING BEA, z ACTIVATED MOEA PHYSICAL SOLVENTS & iL “Qo rma z = 100 ¢ 8 s 6 T msrvsicat SOLVENTS OR ECONAMINE: w Zz 4 3 ime ‘ ACTIVATED HOT POT (HIGH CO : HS) 3 2 SULFINOL (HIGH H2S : C03) Py ‘OR AMINE 2 3 2 w = 8 zs : | zg « = ARINES On SULFINOL 2 1 : ar 2 4 e@ eu 2 « 6 Bud 2 4 6 BrOOD PARTIAL PRESSURE ACID GAS IN THE PRODUCT - PSIA Pigure 2.5.5 Range of Gas Treating Process Application Simultaneous HLS and CO> Removal with no other Impurities PresentPROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION: 2.5 7 REV. DATE: 1 JUNE 1984 REV:0 races of 22 Statuil 000 aE . necro OR 2 setSeo é 2 € wo { Bos i Es fee eee | | ! Zz, fe 5 z 2 = MOEA \ a oR i Ze aie | z z TI ‘ 2 [__ staerrono now con) on t VETROCOKE (HIGH C02) {OR ADIP- HIGH CO3i : ere imrsigiegrs 13 HS LEVEL IN PRODUCT, PPMY Pigure 2.5.6 Range of Gas Treating Process Application i Selective Hydrogen Sulphide Removal©) PROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION: 2.5 REV. DATE: 1 JUNE 1984 REV:0 PAGE:17 of 21 statoi! 2.5.4.2 Heavy Hydrocarbons In Feed Gas The effect of heavy hydrocarbons in the feed gas must be carefully evaluated. All physical solvent processes will absorb most of the propane plus components from the feed gas stream. For relatively lean gas feeds this may not prove to be a problem since the hydrocarbons can be flashed off and used as fuel gas. For rich gas feeds the hydrocarbons will normally need to be recovered from the carbon dioxide vent gas stream to meet hydrocarbon emission requirements. This will involve a complex hydrocarbon recovery scheme and will generally rule out physical solvent processes if significant quantities of propane plus components are present in the feed gas. Even for chemical solvent processes the final selection can be = affected by the presence of heavy hydrocarbons. Even relatively, small amounts of co-absorption can result in a substantial cost penalty or may require special process features to minimise the losses. 2.5.4.3 Impurities in Feed Gas Sulphur impurities in the feed gas, e.g. carbonyl sulphide, carbon disulphide, must be considered in the process selection. If the feed gas contains appreciable quantities of carbonyl sulphide and carbon disulphide MEA will not be a good choice since it is partially irreversibly degraded by these components. Also, the requirements for solvent reclaimers will depend to a large extent on the type and quantites of impurities in the feed. Processes that remove sulphur impurities can also prove advantageous in minimising or eliminating downstream treating of products. 2.5.5 Process Design Por any large scale gas treating application it is necessary: to contact Licensors even at a preliminary design stage. To achieve a competitive design, operating data and design information are required that are only available from the Licensor.PROCESS DESIGN HANDBOOK 25 MAY 1984 SECTION: 2.5 PAGE:18 of 21 statTil ISSUE DAT! REV. DATE: 1 JUNE 1984 RE For amine treating units, SSI and Chemshare have simulation Programmes which can be used to check licensor designs or, with caution, for preliminary process designs. As a guide the design information given in Table 2.5.1 (reference 3 and 4) can be used for input into these programnes. For most gas treating applications the process engineer will be required to evaluate a number of Licensor designs. To ensure a valid comparison is made, it is important that all Licensor designs are carried out to a consistent design basis. A checklist of information required by a Licensor is given below: > a. Feed gas conditions - specify feed temperature, pressure and compositon. Composition should give a detailed component analysis with heavy ends breakdown and levels of any impurities. b. Treated gas quality - specify hydrogen sulphide and carbon dioxide levels required in treated gas and maximum treated gas temperature. Ensure specifications are realistic as overspecification can cause much greater penalty to some processes than to others. ce. Acid gas quality and disposal - specify means of disposal of acid gas e.g. to vent stack and required temperature and pressure at the battery limit. Specify maximum allowable level of hydrocarbons and other impurities in the acid gas stream. a. Utilities available - specify available cooling and heating mediums with supply and return temperatures and design and operating pressures. Some processes may be limited on maximum heating medium temperature due to solvent degradation or excessive corrosion.PROCESS DESIGN HANDBOOK . ISSUE DATE: 25 MAY 1984 SECTION: 2.5 : OO REV. DATE: 1 JUNE 1984 REV:0 pace:19 of 22 «© tQtor PRIMARY AMINES | SECONDARY AMINES | MXED AMINE : mea__| oca__| dea | oA SULFINOL HOSEA Wee [fi 1 1 a Gesu DEMCAL REACTIVITY 1 1 z z 2 Design Date: Pick-up ; mol AG/mol Amine] 0-25-0-45 | 0-25-0458 035-11 05-085. 0.2-045 (2) - Weight "fe w-30 | co-75 | 240 | 25-70 a {31 Heat of Reaction HyS KU/Kg mel “6 46 Fry 8 - (Oz KJ/Kg mol 4s 45 345 365 - {4} Reboiter Duty MUA?) 280-340 | 310-370 | 250-280 5 160480 Solvent losses kg/10&Sm3 Feed 32-64 32-64 16+32 32-64 - gas Degradation temp °C (5) SOC 200°C 150°C 1eOrC 180°C Reboler temp *C voir | raeus | o-i2 | no-ns io-138 Trermat heat capacity 82°C Jig'c| 607 | 343 368 7 s% solution [50% suton Jeo solution | Lom stution Freezing point *C 55 0 a8 | 34 vories ‘Application Contociar pressure |All presna > 134 bare == Bt bore with 60°C (ean ‘amine Specification High quatity tor all | High quality only for_| High quality only for pressures pressures 13-6 bor o| pressures» 13-8 boro Reclaiming VeaieGeC 87-1890 fvecuum —|Vacuim reclaiming -, [Atmospheric |Atmospheric | “&™* — Irecialming |@ 177°C : m7'e DIPA,ISwt"/s w 11 Based on typical composition of 40 wi"/e sullolane, (Swi 2 Bosed en met AQ/mol OPA. BW Average value varies with acid ges loadings. : 4) For lean solution flowrate, 5) Maximum recommended reboiler tube wall temperature to avoid excessive degradation. Table 2.5.1 Design Data for Various Amine ProcessPROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION: 2.5 ~ REV. DATE: 1 JUNE 1984 REV PAGE:20 of 21 statTl With the above information most Licensors will be able to provide a process design ‘package’ for their treating process. The Licensor should provide the following information: a. Process flow diagram giving heat and material balance. b. Solvent circulation rate. c. Reboiler and reclaimer duties. a. Major equipment sizes and materials of construction. e. Utility summary giving heating, cooling and electrical load requirements. f. Special process design features. 3. Guaranteed treated gas specifications for a given 2 reboiler duty and circulation rate. he Anticipated solvent losses, solvent inventory and current solvent costs. ie Royalities and license fees payable. i Scope of services which can be provided. ke. Details of secrecy agreements required. 2.5.6 Design Considerations The following considerations are applicable to all gas treating processes: a. Set reflux drum pressure such that downstream treatment ~\ of acid gases is feasible. b. Review plant design concept against acceptable acid gas disposal emissions with respect to environmental standards and personnel protection (effect of wind on flared or incinerated emissions). c. Removal of carbon dioxide from natural gas will shift phase envelope - use Simulation program to predict envelope change and consequent hydrocarbon condensation effects.PROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION: 2.5 : REV. DATE: 1 JUNE 1984 REV:0 PAGE:21 of 21 stato? da. A filter separator should be provided upstream of the contactor to remove any pipeline scale and entrained liguids, which would otherwise promote foaming in the contactor. Lean solvent solution should be at least 6°C hotter than feed gas to avoid hydrocarbon condensation. References Today's gas-treating processes 1 and 2, Oil and Gas Journal, July 12, 17, 1971. Guidelines can help choose proper process for gas-treating plants, Oil and Gas Journal, Jan 10, 1977. Primary vs. Secondary amines - characteristics in gas conditioning, K.F. Dutwell, D.J. Kubek, P.W. Sigmund, Union Carbide Corp. R & D dept. DGA - a workhorse for gas sweetening, Oil and Gas Journal, Nov 8 1982.PROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION: 2.6 —~e REV. DATE: 1 JUNE 1984 REV:0 PAGE:1 of 4 stat Jl 2.6 Beat and Material Balances Reference Specification: A-SG-003 "Draughting Standard for Piping, Instrument and flow diagrams* 2.6.1 ‘Introduction The major aspects of the applicable specification are summarised in this section. All systems such as Main Process, Process Support, Utility, Fire Fighting System and Drilling shall be represented by a flow diagram. The purpose of a flow diagram is to show the basic scheme for operation and control, identify equipment and interconnecting piping in each system with information on operating conditions, instrument controls and heat and material balance. 2.6.2 Content A complete flow diagram finalized’ and approved for detailed engineering will contain the following information: a. Every item of major equipment (but groups of duplicate items may be represented by one symbol rather than a symbol for each item). b. The identifying TAG number, name, critical dimensions, performance, requirements or capacity for each item. c Main process instrumentation and controls a. Main process flow lines with lines given a stream number inside a diamond located at the line. e. All lines essential for understanding the heat, material and/or utility balance. s ~PROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION: 2.6 REV. DATE: 1 JUNE 1984 REV: PAGE:2 of 4 statoi' £ The flowrate, compositon, state, specific gravity, molecular weight and other data indicated in Section 2.6.3 below for each process stream. g. The anticipated or required conditions of pressure and temperature throughout the plant. h Labels on lines entering and leaving the drawing, giving medium and source/destination. ie Information about essential open/closed control and isolation valves. Be Drawing number ‘and title with revision status. Reference is made to the General Engineering Guideline A-SG-003, "Draughting Standard for Piping, Instrument and Flow Diagrams” for symbols for the flow diagrams. Process type subsystem specified as Vendor packages shall be treated as package units, and package limits’ indicated. Utility systems, such as heating and cooling medium shall be identified on the main process flow diagram with streams in and out and essential control valves. Flow, temperature and pressure shall be covered on separate utility flow diagrams where the same philosophy applies for the main as well as process support flows. Only essential control valves and block valves shall be shown. The instrumentation illustrated on flow diagrams shall provide a simplifed graphic picture of the primary control mode for the facilty. 2.6.3 Process Heat and Material Balance ! A process simulation computer program is generally used to | establish the Heat and Material Balance for the process system. The results of the simulation should be checked to ensure a | "balance", i.e. "feed in = products out". Additionally, if the simulation has a recycle loop, this should be checked for 1 convergence. i.e. the error on the stream change in the loop i must be within tolerance. 1PROCESS DESIGN HANDBOOK A ISSUE DATE: 25 MAY 1984 SECTION: 2.6 REV. DATE: 1 JUNE 1984 REV:0 PAGE:3 of 4 statul In’ situations where more than one operating case is being considered, it is necessary to identify those that are "design" cases, i.e. control equipment sizing. The following options should be considered: a. Extremes of feed composition b. Flowing well head pressure and temperature (FWHP and ~ FWET) c. Minimum/maximum flows for oil/water/gas a. Different delivery requirements. The Heat and Material Balances for those cases selected as 2 "design" should be presented on separate Process Plow Diagrams. The information included in the heat and material balance on main process flow diagrams shall be: ae Stream number be Phase e. Total flowrate kg/h a. Vapour flowrate, at 15.0°C sm3/h e. Vapour flowrate, at flowing condition m/h f. Liquid flowrate, at 15.0°c n/n 3. Liquid flowrate, at flow condition m3/h = he Operating pressure, barg al i. Operating temperature, °e ae Molecular weight kg/kmol ke Specific gravity, at 15.0°C Ll. Density, at flowing condition kg/m? m _Total heat flow rates, kW oor MW . (datum at 0°C) ne Viscosity cP If practical, the compositon of the molar flow rate shall be included.PROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 REV. DATE: 1 JUNE 1984 REV:0 statoi' 2.6.4 Process Support/Utilities Heat and Material Balances The heat and material balances for the process support and utility systems will, in general, be completed after a study to establish the normal and maximum usage or consumption due to each item of equipment on the platform. In particular, the number of operating cases will be given by the number of process operating cases. It is not normally required to produce these flow diagrams on a case by case basis. It will normally be sufficient to define these systems in terms of normal and maximum conditions with the addition of notes where necessary. The Flow Diagram layout should be similar to the Process Flow Diagram layout, with all equipment and users within the system presented in a simplified flowscheme, and all streams referenced to those in the heat and material balance. The information _ included in utility balance on utility diagrams shall be: a. Stream number b. Phase ce. Total Flowrate, kg/h a. Vapour flowrate, at 15.0°C sm?/h e. Vapour flowrate, at P.C. m/h £. Liquid flowrate, at F.C. n?/h 3. Operating pressure, barg he Operating temperature °c i. Density, at P.C. kg/m? ao Viscosity oPPROCESS DESIGN HANDBOOK a ISSUE DATE: 25 MAY 1984 SECTION: 2.7 i REV. DATE: 1 JUNE 1984 — REV:0 PAGE:1 of 8 statoil 2.7 Crude Oil Processing 2.7.1 Introduction Crude oil processing involves the removal of gas, water and impurities from a wellhead fluid to meet export specifications whilst maximising oil recovery. The processing system consists of a combination of the following units: : Gas/oil separation stages, with or without water removal. 2 : Heating or cooling as required before the separation stages. a Crude dehydration/desalting as required. : Cooling (if required) and pumping for export. A typical processing scheme is shown below. ne eas ees Figure 2.7.1 typical 011 Processing System A process simulation program is generally used to design and optimise a crude oil processing system to meet a given crude specification, usually vapour pressure (either TVP or RVP). Selection of a system is based on maximising the crude output whilst minimising energy (i.e. heating/cooling loads). Equipment size and weight is also a governing criterion.( PROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION:2.7 REV. DATE: 1 JUNE 1984 © REV:0 PAGE:2 of 8 statoi The offgas from each separation stage can be compressed and treated for use as fuel gas or exported, or could be flared if quantities were minimal, In designing the oil processing system, the gas compression requirements influence the total energy input. Additionally the recycle of hydrocarbon condensate from the gas compression system must be included as this will influence the performance of the system. 2.7.2 Separation 2.7.2.1 Separation Stages The optimum number of separation stages varies with flowing wellhead pressures (FWEP), reservoir composition, offgas compression requirement, and export specification for crude vapour pressure (influences final stage pressure). Without simulation data, a quick assessment can be made as follows, based on FWHP: FWEP, Bara ‘No. Separation Stages 0 - 20 lor2 20 - 70 2or 3 Over 70 3or 4 The selection of final stage separating temperature and pressure fixes the vapour pressure of the product. For export via a tanker (or to storage), and for pipeline export, the following are typical operating pressures within the final stage, and can be used as an initial estimate: Export Method TVP Specification Final Stage Operating Pressure Tanker 0.965 bara at 25°C = 1.4 = 2.0 bara Pipeline 6.9 bara at 38°C 7.0 bara Having selected an operating pressure, the required operating temperature can be chosen to meet specification. By minimising the operating pressure, the corresponding operating temperaturePROCESS DESIGN HANDBOOK © ISSUE DATE: 25 MAY 1984 REV. DATE: 1 JUNE 1984 REV:0 is minimised, thus if heating is required, the heat input will be reduced. In the case of tanker quality crude, the minimum final stage pressure is fixed by the method of disposal of the offgas. Generally it must be sufficient to feed a compression system, without the compressor inlet falling below atmospheric pressure. 2.7.2.2 Separator Sizing In an oil system, the separators are generally sized on the basis of liquid residence time as follows: oil : 3 minutes ) Liquid Residence Time . Water : 3 to 6 minutes For three phase separation (oil/water/gas), the above residence times apply. For the two phase separation of ofl and gas, a lower residence time of approximately two minutes may be used. Particular attention must be given to the foam and emulsion forming tendency of the crude oil. Data can be obtained from laboratory analysis or from previous experience. A crude oil tendency to foam will require a larger separator in order to maintain a satisfactory vapour/liquid separation efficiency. Where emulsions are formed, de-emulsifying chemicals may facilitate water removal, although the provision of a separate >, two phase (oil/water) separator may be required in severe cases. “’ ° See Section 5.1 for other separation sizing criteria and assessment of vessel weights. 2.7.2.3 Water Removal Criteria The efficiency of water removal in a three phase separator is dependent on the operating temperature, and its subsequent effect on crude viscosity, and the tendency of the crude to form emulsions. Bulk water is generally removed from the first stage of separation, in order to minimise heating/cooling of excessPROCESS DESIGN HANDBOOK - ISSUE DATE: 25 MAY 1984 SECTION:2.7 A REV. DATE: 1 JUNE 1984 REV:0 PAGE:4 of 8 statoijl liquids. Water removal from one further stage is generally required. The following outlet water in oil concentrations can be used as a first estimate: First Water Separation Stage 5 to 10& volume Second Water Separation Stage 2% volume Generally, 2& volume water in oil is the minimum that can be guaranteed from a conventional separator. If the export specification is lower than this, or the crude exhibits a strong emulsion forming tendency, crude dehydration using an electrostatic coalescer is required. Depending on reservoir characteristics and the use or not. of water injection, an increase in produced water may occur as the - reservoir is depleted. Consideration should be given at the @ésign stage of the crude oil separation train in order to cater for increased water production. The separators may be sized for the worst operating case, or alternatively, adjustments may be made to existing separator internals and level control connections in order to change the hold up times of the two phases. 2.7.3 Heating and Cooling © The stabilisation of the crude oil often requires heat to be added or removed at certain points in the processing train. Crude heating may be required for: a. Emulsion breaking and improved separation of oil and produced water b. adjustment of final product TVP For high temperature wellstream fluids, it may be necessary to cool the crude, in order that excessive vapour loss does not lea: to under achievement of the final product TVP specification and i to loss of potential product.PROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION: 2.7 a REV. DATE: 1 JUNE 1984 REV:0 PAGE:5 of 8 stal_il Particular attention should be given to the location of these heaters and coolers in the processing train, since prior removal of the bulk of the produced water will result in reduced duties. For crude oils containing wax, care must be taken in assessing skin temperatures inside coolers so that wax deposition is avoided. Skin temperatures should be at least 5°C ‘above the crude oil cloud point. When the cooling water supply temperature is below this temperature, a cooling water recycle can be incorporated to raise the cooling water inlet to the required temperature. When the minimum cooling water temperature is marginal for wax deposition, wax inhibitor injection may be 2 considered instead of a cooling water recycle system. Por heat exchanger sizing and weight estimates, reference should be made to Section 5.3. 2.7.4 Stabilisation Where four or five flash separation stages are required to achieve on specification crude, a distillation crude- stabilisation system may be considered. This would be located Gownstream of the first two stages of separation, replacing the further separation stages. This method can increase crude oil recovery and reduce associated gas recompression power. > Stabilisation is often advantageous for light crudes and condensates meeting a tanker specification. For a crude stabiliser, the approximate column diameter may be obtained from Figure 2.7.2, for the following system: Column height: 18.5 m T/T Column Operating Pressure: 9 bara Bottoms Temperature: 186°CCc PROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION: 2. REV. DATE: 1 JUNE 1984 © REV:0 PAGE:6 of 8 statoi COLUMN DIAMETER. m. | py tt I : 0 2 6 6 B 0 12 % 16 18 20 2 2 26 26 10> m3/d. OIL Figure 2.7.2 Approximate Column Diameter For column sizing and weight methods, see Section 5.11. 2.7.5 Dehydration and Desalting Dehydration and desalting are performed in identical electrostatic coalescers. Usually for complete dehydration and desalting a two stage process is used, where the bulk of the produced water is removed in the first electrostatic coalescer. This is followed by the second “desalting" stage, where wash water is injected upstream, and removed in the coalescer. The number of stages required depends on the produced water content, the inlet salt concentration and the crude salt specification. Suitable operating temperatures for dehydration and desalting are 25 to 150°C, depending on the characteristics of the crude. The electrostatic coalescers must be located after a crude degassingPROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION: 2.7 > REV. DATE: 1 JUNE 1984 REV:0 PAGE:7 of 8 statu stage, and sufficient pressure maintained to prevent vaporisation in the unit. 2.7.5.1 Preliminary Sizing Basis Unit capacity : 15 to 20 minutes residence time 100% liquid filled Vessel L/D d 4 to 6:1 For assessment of vessel weight see Section 5.1. 2.7.5.2 Dehydration Stage In the dehydration stage, up to 10% volume water in inlet oil reduces to 0.2 to 0.5% volume water in oil after coalescing. Crude oil from production separators normally contains 2% volume minimum produced water. This level should be maintained to ensure coalescence. 2.7.5.3 Desalting Stage Tanker loading salt specification = 71 g/m3 (25 1b/103 Bbl) Desalting is generally not carried out offshore as it is expensive to achieve this specification in terms of dilution water or installation of multistage desalting units, The crude export specification is usually based on water content alone. A graph of TDS (Total Dissolved Solids) in produced water vs dilution water requirement (Figure 2.7.3) to produce a salt content of 71 g/m? oil is presented. 2.7.6 Chemical Injection Crude oil and water separation may be enhanced by the injection of chemicals such as de-emulsifying and anti-foaming agents. Wax inhibitor may be used to suppress wax formation in coolers or pipelines if required. Details of the use of such chemicals are given in Section 2.3.7.PROCESS DESIGN HANDBOOK ISSUE DATE: 25 MAY 1984 SECTION:2.7 7 REV. DATE: 1 JUNE 1984 REV:0 PAGE:8 of 8 stato’ 2.7.7 References 1. Optimisation of crude stabilisation units, Oil and Gas Journal, Jan 21 1980. 2. Stabilisation process comparison helps selection, Oil and Gas Journal, Jan 28 1980. x 3 o S T | 1 | peal | S 3 ele |_| | 4. 2 8 PRODUCED WATER TDS (ppm) & ° ata =} —— 4 ° 50 100 150 FRESH DILUTION WATER, m? PER 1000 m3 OIL 6 Figure 2.7.3 TDS vs Dilution Water