Professional Documents
Culture Documents
Mergedasldfjasdf (Asdkfasp
Mergedasldfjasdf (Asdkfasp
Abstract
A series of novel heuristic numerical tools were adopted to tackle the setback of permeability estimation in carbonate reser-
voirs compared to the classical methods. To that end, a comprehensive data set of petrophysical data including core and log
in two wells was situated in Marun Oil Field. Both wells, Well#1 and Well#2, were completed in the Bangestan reservoir,
having a broad diversity of carbonate facies. In the light of high Lorenz coefficients, 0.762 and 0.75 in Well#1 and Well#2,
respectively, an extensive heterogeneity has been expected in reservoir properties, namely permeability. Despite Well#1,
Well#2 was used as a blinded well, which had no influence on model learning and just contributed to assess the valida-
tion of the proposed model. An HFU model with the aim of discerning the sophistication of permeability and net porosity
interrelation has been developed in the framework of Amaefule’s technique which has been modified by newly introduced
classification and clustering conceptions. Eventually, seven distinct pore geometrical units have been distinguished through
implementing the hybridized genetic algorithm and k-means algorithm. Furthermore, a K-nearest neighbors (KNN) algorithm
has been carried out to divide log data into the flow units and assigns them to the pre-identified FZI values. Besides, a cross
between the ε-SVR model, a supervised learning machine, and the Harmony Search algorithm has been used to estimate
directly permeability. To select the optimum combination of the involved logging parameters in the ε-SVR model and reduce
the dimensionality problem, a principle component analysis (PCA) has been implemented on Well#1 data set. The result of
PCA illustrates parameters, such as permeability, the transit time of sonic wave, resistivity of the unflashed zone, neutron
porosity, photoelectric index, spectral gamma-ray, and bulk density, which possess the highest correlation coefficient with
first derived PC. In line with previous studies, the findings will be compared with empirical methods, Coates–Dumanior, and
Timur methods, which both have been launched into these wells. Overall, it is obvious to conclude that the ε -SVR model is
undeniably the superior method with the lowest mean square error, nearly 4.91, and the highest R-squared of approximately
0.721. On the contrary, the transform relationship of porosity and permeability has remarkably the worst results in comparison
with other models in error (MSE) and accuracy (R2) of 128.73 and 0.116, respectively.
Keywords Permeability · Carbonate Reservoir · Support Vector Regression · Harmony Search · Principal Component
Analysis · Hydraulic Flow Unit
Abbreviations Md Millidarcy
ε-SVR Support Vector Regression MSE Mean Square Error
FZI Flow Zone Index NMR Nuclear Magnetic Resonance
GA Genetic Algorithm PCA Principal Component Analysis
HFU Hydraulic Flow Unit R2 R-squared
HS Harmony Search RBF Radial Basic Function
KNN K-Nearest Neighbor RQI Reservoir Quality Index
Lk Lorenz coefficient SCAL Special Core Analysis
* H. Heydari Gholanlo
heydaripost@gmail.com
1
Well Services of Iran (Schlumberger Method), Tehran, Iran
13
Vol.:(0123456789)
Petroleum xxx (2016) 1e5
Petroleum
journal homepage: www.keaipublishing.com/en/journals/petlm
Original article
a r t i c l e i n f o a b s t r a c t
Article history: Water saturation determination in core laboratory is known as a cost and time consuming labor.
Received 16 January 2016 Hitherto, many scientists attempted to estimate accurately water saturation from well-logging data
Received in revised form which has a continuous record without losing information. Therefore, various model were intro-
6 April 2016
duced to relate reservoir properties and water saturation. Since carbonate reservoir is very het-
Accepted 8 April 2016
erogeneous in shape and size of pore throat, the relation between water saturation and other
carbonates reservoir properties is very complex, and causes considerable overall errors in water
Keywords:
saturation calculation. By increasing the usage and improvement of soft computing methods in
Water saturation
Radial basis function neural network
engineering problems, petroleum engineers have been attended them to measure the petrophysical
Genetic algorithm properties of the reservoir.
Archie model In this study, a radial basis function neural network (RBFNN) improved by genetic algorithm has
Carbonate reservoir been employed to estimate formation water saturation by using conventional well-logging data.
The used logging and core data have been gathered from a carbonated formation from one of
oilfield located in south-west Iran, and finally their results of the proposed model were compared
with the core analysis results. By checking the testing data from another well, it showed this
method had a 0.027 for mean square errors and its correlation coefficient is equal to 0.870. These
results implied on high accuracy of this model for oil saturation degree estimation. While the
common methods like Archie, had a 0.041 mean square error and 0.720 of the correlation coeffi-
cient, which indicate a high ability of RBF model than the other usual empirical methods.
Copyright © 2016, Southwest Petroleum University. Production and hosting by Elsevier B.V. on
behalf of KeAi Communications Co., Ltd. This is an open access article under the CC BY-NC-ND
license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
http://dx.doi.org/10.1016/j.petlm.2016.04.002
2405-6561/Copyright © 2016, Southwest Petroleum University. Production and hosting by Elsevier B.V. on behalf of KeAi Communications Co., Ltd. This is an open
access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Please cite this article in press as: H.H. Gholanlo, et al., Estimation of water saturation by using radial based function artificial neural
network in carbonate reservoir: A case study in Sarvak formation, Petroleum (2016), http://dx.doi.org/10.1016/j.petlm.2016.04.002
Egyptian Journal of Petroleum xxx (2017) xxx–xxx
a r t i c l e i n f o a b s t r a c t
Article history: Well logging analysis has been recognized as the best way to characterize reservoir properties, especially
Received 25 May 2017 water saturation. Archie equation is known as the best method to estimate water saturation from well
Revised 14 July 2017 logging data. Determination of Archie parameters is an incessant challenge in advanced reservoir engi-
Accepted 30 July 2017
neering because of its complex intrinsic. Measurement of Archie parameters in core laboratory has a high
Available online xxxx
sensitive, expensive and time consuming procedure, that causes the excluding of core analysis to be
impossible for the entire depth of reservoir zone. Although some empirical and mathematical models
Keywords:
have been introduced to measure cementation factor, without the need of coring, there is not exist a reli-
Saturation exponent
Genetic algorithm
able method to estimate saturation exponent in similar way.
Sarvak Formation Genetic algorithm is excelled than other meta-heuristic global search methods, which could be used in
solving engineering problems. In this study, a genetic algorithm based strategy has been developed to
develop a model to estimate saturation exponent from a core of drilled well in Sarvak Formation at
Marun Oil Field in Iran. This model has been validated with laboratory measurements of saturation expo-
nent with the accuracy (R2) of 0.91 and the miscalculation (MSE) of 0.23.
Ó 2017 Egyptian Petroleum Research Institute Published by Elsevier B.V. This is an open access article
under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
1. Introduction core analysis [2]. The value of n can be obtained from the slope of
the best straight line on the semi-log plot of IR vs. Sw for several
Based on Archie model [1], the ratio of a sandstone sample brine saturated level of core sample. In many studies regarding
resistance saturated partially with brine, Rt, to saturated com- the prediction of water saturation, the researchers assumed the
pletely, Ro, has a logical relation to Sw as follows: value of n to be 2 [5]. This assumption is not absolutely incorrect,
and is attributable only for samples with macro-scale porous space
Rt
Sn
w ¼ ¼ IR ð1Þ and high Sw value [6]. Many researchers have studied the depen-
Ro
dency of n to some reservoir rock and fluid properties such as sat-
The ratio of Rt/Ro is famed as resistance index (IR) that is mainly uration history [4], wettability ([7,8]), capillary pressure [9],
influenced by formation water salinity [2], and n is known as sat- salinity of brine [10], and the geometry of pore space [11]. By sur-
uration exponent. Although IR is primarily considered to be unit viving in the widely studies have been done on the relationship
on Sw of 100%, many researchers [2–4] approved experimentally between the n and other petrophysical parameters, apparently wet-
that this value is not unit, and the Archie formula (Eq. (1)) had been tability has the greatest impact on n value [8]. Wettability between
revised as follows: rock and fluid systems can be highly varied from highly water-wet
rock to highly oil-wet [7]. IR value of water-wet rock is more than
b
Sw ¼ ð2Þ oil-wet rock [4]. Although, saturation exponent of water-wet rock
IRn is approximately equal to 2, it can be changed abruptly from 1.5
where b is a drift coefficient and its behavior appertains to the wet- to 3 in carbonate rock [12]. Donaldson measured it more than 8
tability of reservoir rock, temperature, pore size distribution, exis- in oil-wet rock [13]. On the other hand, wettability can also change
tence of micro- and macro fractures, and uncertainty caused by with the saturation history [8], consequently it could be concluded
that n always changes when Sw changes in the carbonate rock [14].
Data mining is a set of intelligent tools and machine learning
Peer review under responsibility of Egyptian Petroleum Research Institute.
⇑ Corresponding author. techniques that can explore the hidden patterns behind the
E-mail address: heydaripost@gmail.com (H. Heydari Gholanlo). existing data sets, and extract a quantity formula between the
http://dx.doi.org/10.1016/j.ejpe.2017.07.013
1110-0621/Ó 2017 Egyptian Petroleum Research Institute Published by Elsevier B.V.
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Please cite this article in press as: H. Heydari Gholanlo et al., Saturation exponent determination by using genetic algorithm in carbonate reservoirs: A case
study in Sarvak Formation, Egypt. J. Petrol. (2017), http://dx.doi.org/10.1016/j.ejpe.2017.07.013
Neural Comput & Applic (2018) 30:1657–1666
https://doi.org/10.1007/s00521-016-2770-1
ORIGINAL ARTICLE
Received: 12 June 2014 / Accepted: 29 November 2016 / Published online: 22 December 2016
The Natural Computing Applications Forum 2016
Abstract Water saturation is a key parameter in reservoir composed of pure limestone. The final results show a high
engineering to calculate the volume of hydrocarbon in exactness of the proposed model estimating cementation
reservoirs. The first attempt to estimate water saturation factor with an R-squared of 0.864 and a mean squared error
using well log data was implemented by Archie in 1942 for of 0.01899 .
a clean sandstone reservoir. This method requires labora-
tory measurement of cementation factor. This parameter Keywords Water saturation Cementation factor
has the most impression on water saturation calculation Carbonate reservoir Mamdani fuzzy inference system
compared to the other parameters, which is nearly constant Genetic algorithm Sarvak Formation
in homogenous sandstone reservoirs. However, due to high
variation of cementation factor along depth of wellbore in Abbreviations
carbonate reservoirs due to rock’s nature, it is incorrect to m Cementation factor
assign a constant value to cementation factor. On the other u Porosity
hand, experimental core analysis to determine cementation a Tortuosity factor
factor values is an expensive and time-consuming work, n Stauration exponent
and it is impossible to calculate this parameter in labora- FIS Fuzzy inference system
tory for the whole depth of a drilled wellbore. In industrial FRF Formation resistivity factor
applications, using a constant cementation factor can lead R2 R-squared
to erroneous calculations of water (and hence oil) satura- MSE Mean square error
tions. Also, previous conventional methods estimating k Permeability
cementation factor from logging data are not often sensi- MF Membership function
tive to pore system of rock and generate a massive source c Center of Gaussian and sigmoidal membership
of error. In this study, a new approach to estimate function
cementation factor using a genetic Mamdani fuzzy infer- r Width of Gaussian membership function
ence system was implemented for a case study in Sarvak a Parameter of sigmoid membership function
Formation located in Zagros Basin which is mostly GA Genetic algorithm
MD Membership degree
FMI Formation micro imager
& Hamid Heydari Gholanlo NMR Nuclear magnetic resonance
heydaripost@gmail.com
1
Department of Petroleum Engineering, Abadan Faculty of
Petroleum Engineering, Petroleum University of Technology,
1 Introduction
Abadan, Iran
2 Archie [1] derived a formula between water saturation (Sw)
Department of Chemical Engineering, Abadan Faculty of
Petroleum Engineering, Petroleum University of Technology, and formation properties for clean sandstone reservoirs
Abadan, Iran obtained from well logs with interparticle pore system
123
Abstract
Cementation factor is a critical parameter, which affects water saturation calculation. In carbonate
rocks, due to the sensitivity of this parameter to pore type, water saturation estimation has associated
with high inaccuracy. Hence developing a reliable mathematical strategy to determine these properties
accurately is of crucial importance. To this end, genetic algorithm pattern search is employed to find
accurate cementation factor by using formation resistivity factor and the porosity obtained from
laboratory core analyses with considering the assumption that tortuosity factor is not unity.
Subsequently, particle swarm optimization (PSO) fuzzy inference system (FIS) was used for the
classification of cementation factor according to the predominated rock pore type by using the input
variables such as cementation factor, porosity, and permeability to classify the core samples in three
groups, namely fractured, interparticle, and vuggy pore system. Then, the experimental data which
was collected from Sarvak formation located in one of the Iran southwestern oil fields was applied to
the proposed model. Next, for each class, a cementation factor-porosity correlation was created and
the results were used to calculate cementation factor and water saturation profile for the studied well.
The results showed that the constructed model could predict cementation factor with high accuracy.
The comparison between the model presented herein and the conventional method demonstrated that
the proposed model provided a more accurate result with a mean square error (MSE) of around 0.024
and led to an R2 value of 0.603 in calculating the water saturation.
Keywords: Cementation Factor, Carbonate Reservoir, Genetic Algorithm, Particle Swarm
Optimization, Fuzzy Logic
1. Introduction
In order to determine the total hydrocarbon content of oil and gas reservoirs and to characterize
reservoir during field-development planning, water saturation (Sw) is known to be a key parameter.
The accurate estimation of the Sw is a critical issue in reservoir studies and has almost a high
uncertainty especially in carbonate reservoirs.
Archie (1942) derived a formula between the porosity and electrical resistivity properties of the
formation, which was obtained from well logs and water saturation; the correlation was later modified
by Winsauer, et al. (1952) as follows:
*
Corresponding Author:
Email: j.moghadasi@put.ac.ir
SPE Reservoir Evaluation & Engineering-Formation Evaluation
Manuscript ID Draft
Complete List of Authors: Heydari Gholanlo, Hamid; Shahrood University of Technology, Faculty of
rR
Keywords: Stress State, Breakout, Ultrasonic Imager Tools, Active Shear Zone
iew
Note: The following files were submitted by the author for peer review, but cannot be converted to PDF.
You must view these files (e.g. movies) online.
Figures.rar
On
ly
Journal of Petroleum Science and Engineering
Application of Signal Envelope Classification to Estimate Shear Wave Slowness from
Conventional Cement Bonding Logging Tools
--Manuscript Draft--
Keywords: Monopole Logging Tools; Shear Wave Slowness; Dipole Logging Tools; Signal
Processing; Wave Envelope
Abstract: The shear wave slowness plays a vital role in the modelling of the reservoir
geomechanics and wellbore stability in the deep underground opening, such as oil and
gas wells. In many cases, such as old wells, bad hole conditions or because of cost
controlling, dipole sonic logging tools, which are equipped with the last developed
acoustic measuring technology, are preferred not to be run. Also, the shear wave
slowness was not probably recorded in the oil wells. Although, the conventional sonic
tool, as a part of full-suit combo, provides compressional wave slowness
measurement, it is not sufficient for geomechanical reservoir modelling. The paper
suggests a novel method to generate shear wave slowness profile along depth utilizing
of conventional cement bonding tools. To that end, an algorithm has been developed
based on the signal envelope classification in Octave programming environment. A
well drilled recently in the southwest of Iran, has been selected to assess the
performance of the proposed model. The comparison of the acquired shear wave
slowness in the dipole sonic and the cement bonding tools shows the proposed model
is of high accuracy in generating shear wave slowness profile in well-cemented
intervals.
Jongkook Kim
Korea National Oil Corporation
jk.kim2008@knoc.co.kr
Similar Work
Shantanu Chakraborty
shantanu.ism@gmail.com
He is paraphysicist.
Jorge Mendoza
jsmendoz@espol.edu.ec
Similar Work
Powered by Editorial Manager® and ProduXion Manager® from Aries Systems Corporation