platform - press.utp.edu.mypress.utp.edu.my/wp-content/uploads/2018/12/vol-10-no-1_2014_new.pdf ·...

P L A T F O R M

VOLUME TEN NUMBER ONE JANUARY- JUNE 2014 ISSN 1511-6794

A Journal of Engineering, Science and Society

PL

AT

FO

RM

VO

LU

ME

TE

N N

UM

BE

R O

NE

JAN

UA

RY

- JUN

E 2

01

4

Volume 10 Number 1 Jan - Jun 2014

Multiple Hydraulic Fractured Vertical Wells in Gas Condensate ReservoirsMohammed Abdalla Ayoub, Abdollah Esmaeili 2

A Study on Gas Flaring Activities in NigeriaAliyu A. Sulaimon, Modupe Ogunmekan 12

Gender Diversity on Boards and Market Performance: An Empirical Investigation on Malaysian Listed CompaniesRohail Hassan, Maran Marimuthu 17

Fuzzy Linear Model For Electricity Planning In India With Focus On Cost Of Biogas PowerJebaraj S 27

Application of Artificial Neural Networks Technique for Estimating Permeability from Well Log DataMohammed Abdalla Ayoub, Abdollah Esmaeili 31

Capital Structure and the Determinants of the Banking Sector: The Case of MalaysiaMaran Marimuthu 38

A Comparative Numerical Simulation Study of Polymer and Surfactant Adsorptions on Sandstone RocksMohamed Ali Hamid, Mustafa Onur 49

Reviewers 56

What to sendSend article/manuscript in Microsoft Word (any version) via e-mail, or on a media (CD or flash drive), plus one PDF copy.

Send high resolution illustrations/images/tables/graphs with jpeg extension separately.

FormatIEEE format should be adopted in preparing the article/manuscript. Do check the available IEEE format and templates at IEEE – Manuscript Templates for Conference Proceedings.

The Manuscript1 The text should be preceded by a short abstract of 50 -100 words and a minimum of four keywords.2 The manuscript must be typed on one side of the paper, double-spaced throughout with wide margins not exceeding 8 pages (including references).3 Figures and tables have to be labelled and should be included in the text. If authors are using databases, tables , diagrams, etc from internet and print materials, permission must be sought first from the rights holder. Please attach the permission(s) with the manuscript.4 Footnotes should be kept to a minimum and be as brief as possible. The footnotes must be numbered consecutively.5 Special care should be given to the preparation of the drawings for the figures and diagrams. Except for a reduction in size, they will appear in the final printing in exactly the same form as submitted by the author.6 When submitting the article/manuscript, authors are also required to submit photographs and biography of themselves including academic background, work experiences and research interests.

Submission/Acceptance Procedures1 Articles/Manuscripts are evaluated by an editorial review board (internal and external). 2 Following initial selection, authors whose papers have been accepted for publication will be notified by email. 3 Only original and unpublished articles may be submitted.

Submit Articles/Manuscripts to: PLATFORM Editor-in-Chief Universiti Teknologi PETRONAS 32610 Bandar Seri Iskandar Perak Darul Ridzuan MALAYSIA Email: [email protected] [email protected]

ABOUT PLATFORM

PLATFORM is a biannual, peer-reviewed journal of Universiti Teknologi PETRONAS. It serves as a medium for faculty members, students and industry professionals to share their knowledge, views, experiences and discoveries in their areas of interest and expertise. It comprises collections of unpublished and original work by the academic staff and students of the university as well as industry professionals.

Guideline for Submitting a Paper to PLATFORMThese guidelines are designed to simplify and help standardize submissions. Length varies, but an average of 3,500 to 5,000 words, plus illustrations/images and captions. Maximum length should be about 8,000 words.

1VOLUME TEN NUMBER ONE JANUARY - JUNE 2014 PLATFORM

I S S N 1 5 1 1 - 6 7 9 4

Contents

Copyright © 2015Universiti Teknologi PETRONAS

PLATFORMJanuary - June 2014

Patron:Y.Bhg. Datuk Ir (Dr) Abdul Rahim B Hashim

Advisor: Prof Ir Dr Ahmad Fadzil M Hani

PLATFORM Editorial

Editor-in-Chief:Assoc. Prof. Dr Mohd Fadzil Hassan

Secretariat:Ismi Hazariah Bt Shahardin

UTP Publication Committee:

Chairman: Prof Ir Dr Ahmad Fadzil M Hani

Members:MOR DirectorsHead of Departments

Secretary:Naziri B Yusof

Layout and Production:

YY ChongMohamed Azaman IsmailUTP Press

Address:PLATFORM Editor-in-ChiefUniversiti Teknologi PETRONAS32610 Bandar Seri Iskandar, Perak Darul Ridzuan, Malaysia

http://www.utp.edu.my

[email protected]@petronas.com.my

Telephone +(60)5 368 8687

Facsimile +(60)5 365 4090

Published byUTP Press© Universiti Teknologi PETRONAS,Owned by Institute of Technology PETRONAS Sdn Bhd.2015

Multiple Hydraulic Fractured Vertical Wells in GasCondensate Reservoirs

Mohammed Abdalla Ayoub, Abdollah Esmaeili 2

A Study on Gas Flaring Activities in Nigeria

Aliyu A. Sulaimon, Modupe Ogunmekan 12

Gender Diversity on Boards and Market Performance: An Empirical Investigation on Malaysian Listed Companies

Rohail Hassan, Maran Marimuthu 17

Fuzzy Linear Model For Electricity Planning In India With Focus On Cost Of Biogas Power

Jebaraj S 27

Application of Artificial Neural Networks Technique for Estimating Permeability from Well Log Data

Mohammed Abdalla Ayoub, Abdollah Esmaeili 31

Capital Structure and the Determinants of the Banking Sector: The Case of Malaysia

Maran Marimuthu 38

A Comparative Numerical Simulation Study of Polymer and SurfactantAdsorptions on Sandstone Rocks

Mohamed Ali Hamid, Mustafa Onur 49

Reviewers 56

Copyright and Reprint PermissionAll rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means, including photocopying and recording, without the written permission of the copyright holder, application for which should be addressed to the publisher. Such written permission must also be obtained before any part of this publication is stored in a retrieval system of any nature.

The publisher, authors, contributors and endorsers of this publication each excludes liability for loss suffered by any person resulting in any way from the use of, or reliance on this publication.

2 PLATFORM VOLUME TEN NUMBER ONE JANUARY - JUNE 2014

PLATFORM - A Journal of Engineering, Science and Society

INTRODUCTION

The first step in studying the gas condensate reservoirs is to characterize the reservoir fluid. It has been found that characterization of the gas condensate fluid is strongly influenced by some main factors such as hold up of retrograde condensate in the formation resulting in excessive producing gas to liquid ratio (GLR). Figure 1 shows a common characteristic of gas condensate fluids. The liquid drop out reaches a maximum and then decreases by vaporization during pressure depletion. This behavior may imply that when the reservoir pressure decreases sufficiently, the condensate will be recovered by re-vaporization. However, by the time pressure falls below the dew point, the original phase diagram is no longer valid

as the system composition changes during the production period. The compositional analysis of gas condensate fluids is conducted generally in more details than that of oil. The compositional data are used often in phase behavior models, particularly in reservoir simulation. The fluid is commonly analyzed by flashing it at the atmospheric pressure and measuring the composition of the stabilized gas and liquid phases. The fluid heavy fraction is analyzes to identify major components , and also characterize it by extended carbon groups , as the results of phase behavior models are very sensitive to the heavy end description of gas condensate systems.

MULTIPLE HYDRAULIC FRACTURED VERTICAL WELLS IN GAS CONDENSATE RESERVOIRS

Mohammed Abdalla Ayoub, Abdollah EsmaeiliUniversiti Teknologi PETRONAS

Email: [email protected], [email protected]

ABSTRACT

Gas reservoirs can be classified into dry gas reservoirs, wet gas reservoirs and gas condensate reservoirs. In gas condensate reservoirs, the reservoir temperature lies between the critical temperature and the cricondentherm. The gas will drop out liquid by retrograde condensation in the reservoir, when the pressure falls below the dew point. This heavy part of the gas has found many application in industry and also in daily life and by remaining in reservoir not only this valuable liquid is lost but also its accumulation will result in forming a condensate bank near the well bore region which makes a considerable reduction in well productivity. This highlights the need to find an economical way to increase the condensate recovery from these reservoirs. Wells in gas condensate reservoirs usually exhibit complex behaviors due to condensate deposition as the bottom hole pressure drops below the dew point. Formation of this liquid saturation results in reduced gas relative permeability around the well bore and a loss of gas productivity. One of the several ways of minimizing the pressure drop in order to reduce liquid drop-out is hydraulic fracturing before or after the development of the condensate bank. The pressure transients are often used as a reliable evaluation of stimulation performance for field development planning. It has been shown that condensate deposits effects can be identified and quantified by well test analysis dealing with a well test composite behavior; which, in presence of hydraulic fractures becomes much more complex. But the various impacting factors of stimulation; such as fracture length, conductivity, orientation, etc. can also be observed and defined in these analyses. In this paper, modeling and interpretation of pressure transient responses of multiple hydraulic fractured vertical wells using a numerical reservoir model has been investigated.



Figure 1. Liquid drop out of gas condensate [30]

Figure 2 shows pressure temperature relationship for gas condensate fluid with constant composition. Figure 3 shows effect of composition change on phase envelope.

Figure 2. Pressure – temperature curve for gas condensate with constant composition [31]

Figure 3. Effect of composition change on phase envelope [31]

In gas condensate reservoirs, gas injection and gas recycling are practices applied to reduce the drop

out of the condensate in the reservoir. Gas injection can be implemented at the initial reservoir pressure to maintain the pressure above the dew point (full pressure maintenance) , or after the reservoir pressure falls below the dew point (partial pressure maintenance) in which injection gas vaporizes the condensate and reduces the condensate accumulation in the reservoir . Gas recycling has been implemented in gas condensate reservoir for many years, but the increasing application of the reservoir gas and as a result its price made the engineers find an appropriate replace for it. N2 and CO2 were suggested as two alternatives which are now applied in some reservoirs. Nitrogen injection is not as effective as CH4 and CO2 injection, but as it is available and non-corrosive it can be properly applied for this purpose. Figure 4 shows condensate saturation in reservoir and Figure 5 shows gas saturation versus relative permeability.

Figure 4. Condensate Saturation in Reservoir [32]

Figure 5. Gas Saturation versus Relative Permeability [32]



PURPOSE OF STUDY

This work investigates the modeling and interpretation of pressure transient response of multiple hydraulic fractured vertical wells in gas condensate reservoirs, using a fully compositional numerical reservoir model. After validating the numerical model using an analytical solution applied to a simpler reservoir/well model, complex reservoirs using the condensate PVT data are simulated and pressure transient response signatures obtained. Sensitivities of key reservoir/well/fracture properties are investigated and a result of each case is presented.

METHODOLOGY

The main aim of this investigation is to know fractured vertical wells behavior in a gas condensate reservoir. For this purpose, we use a compositional numerical model with one well to be able to predict derivative curves for fractured vertical wells with a pressure below dew point pressure. We use ECLIPSE 300 software which is compositional software. First, input data for numerical simulated model, fluids properties and initial reservoir parameters will be introduced. Then, three dimensional compositional models including one well with numbers of hydraulic fractures will be built. Finally, the results of this study will be discussed.

Cylindrical Compositional Model and Hydraulic Fracturing in Gas Condensate Reservoirs

Figure 6 shows a general pseudo- pressure logarithmic plot and its derivative plots for a vertical well in a two phase gas condensate reservoir. Figure 7 shows a fracture in a reservoir. Figure 8 shows effective well radius for a fracture with infinite conductivity. Figure 9 shows that slope 0.5 is a character for characterization fracture with infinite conductivity. Figure 10 shows that there is a little difference between steady state and non-steady state hydraulic fracture.

Figure 6. Models for a Well with Two or Three Cylindrical Compositional Regions [32]

Figure 7. Fracture in a Reservoir [33]

Figure 8. Effective Well Radius for a Fracture with infinite conductivity [33]



Figure 9. Slope 0.5 is a character for characterization fracture with infinite conductivity [34]

Figure 10. A Little Difference between Steady State and Non-steady State Hydraulic Fracture [34]

Fluid Properties Modeling

The main part of gas condensate reservoir modeling is finding a suitable fluid model. As fluid modeling and model optimization using laboratory data of fluid was not our purpose in this study, we use a prepared simulated fluid from a gas condensate reservoir and its properties as input data for simulation software. The composition of this fluid is shown in table 1.

Table 1. Composition of fluid which was used in simulation

Number Component Mole %

1 N2 0/00462 CO2 0/00613 H2S 0/04 C1 0/68645 C2 0/1396 C3 0/06897 iC4 0/00668 nC4 0/02669 iC5 0/0062

10 nC5 0/009411 C6 0/01412 C7+ 0/0348

Sum 1

Building Numerical Model for Simulation

A three dimensional model with one vertical well was built by ECLIPSE 300 software. For this simulation, we used a reservoir with thickness of 100 foot and length of 450 foot along Cartesian X-Y axes. In this model, we ignored wellbore storage effect and skin factor. Initial reservoir, rock and fluid properties are summarized in table 2.

Table 2. Initial reservoir, rock and fluid properties which was used in model

Porosity 0.2

Initial water saturation 0.349

Well radius 0.3 foot

Isothermal rock compressibility coefficient 0.000004 1/psi

Total length of reservoir 17800 foot

Total width of reservoir 16100 foot

Horizontal permeability 10 md

Vertical permeability 1 md



Numerical Gridding Analyzing

Sizes of grids especially near well bore are very important because of its effects on fluid saturation and pressure of each block. Therefore, we must optimize lengths of grids in each direction. To do this, several model with different number of grids, 4669 (7×23×29), 5887 (7×29×29), 10933 (13×29×29), 20787 (13×39×41), and 31447 (13×59×41) were tested. To study and understand gas condensate behavior near well bore clearly, we used small grid sizes around well bore and bigger grid sizes for grids which are far from well bore. Outer layers are selected with so big sizes that boundary effects on well test results can be ignored. The main purpose of selection this type of gridding was to be able to study flow regimes and accumulated condensate distribution during production or well test period more clearly and

carefully. Finally, the model with 20787 (13×39×41) was selected as the best model. Gridding system used for this model was shown in figure 11 and its grid sizes are shown in table 3.

Figure 11. Cartesian grid system for simulated model

Table 3 – Grid sizes used for simulated model

Dz k Dy j Dy j Dx i Dx i17.8 1 0.2 21 1500 1 35 21 1500 116 2 0.2 22 1500 2 35 22 1500 28 3 0.4 23 1500 3 35 23 1500 34 4 .8 24 1500 4 35 24 1500 42 5 1.4 25 1014 5 35 25 1000 52 6 2.5 62 512 6 35 62 800 61 7 5 27 256 7 35 27 400 72 8 10 28 128 8 35 28 200 82 9 20 29 64 9 35 29 88 94 10 35 30 35 10 35 30 45 108 11 64 31 20 11 45 31 35 11

16 12 128 32 10 12 88 32 35 1217.8 13 256 33 5 13 200 33 35 13

512 34 2.5 14 400 34 35 141014 35 1.4 15 800 35 35 151500 36 0.8 16 1000 36 35 161500 37 0.4 17 1500 37 35 171500 38 0.2 18 1500 38 35 181500 39 0.25 19 1500 39 35 19

40 0.25 20 1500 40 35 20



There are two methods to design grids sizes and grid refinement around well bore using ECLIPSE300 software which are Near Wellbore Module (NWM) and Template (Fig. 12)

Figure 12. Various methods for grid refinement around wellbore

One of the advantages of NWM model is that numbers of grids in a reservoir are less than a case which total reservoir is divided into smaller grids. But, as larger grids are located along side small grids, its calculation error will be increased. To use this model, first, that part of the reservoir which include well and hydraulic fractures must be departed from total reservoir, then, by selecting type of refined grids and based on local grid refinement theory, smaller grids around wellbore and fractures will be built.

Two numerical models which were built using these two methods for reservoir are both at an initial reservoir pressure above dew point (about 4500 PSIA) and have five fractures with same properties. Production program from this well is such that first, it produces for 20 days and then is closed for 20 days. After this time period, is opened again and produces with a producing rate of 10 MMSCF/D for 150 days and then is closed for 250 days. Initial reservoir pressure at the start of simulation was selected so high to prevent gas condensate formation even after 150 days production period, therefore, reservoir fluid was in a single phase. After analyzing these two methods, we concluded that template method is better for this study.

Hydraulic Fracture Parameters

We want to know the effects of fracture physical parameters on pressure derivative curves. We use template model with only two vertical fractures with a distance of 670 foot to each other. Initial parameters of fractures are summarized in table 4. With a fluid which was introduced before and an initial reservoir pressure of 3280 PSIA, a model designed such that gas condensate accommodation around well bore can be seen.

Confirming Numerical Model Using an Analytical Model

Accuracy of grid system in simulated model must be checked by comparing with an analytical model. Simulated model includes a vertical well in a homogeneous with permeability of 10 md and without skin factor. This well produced 50 days with a rate of 10 MMSCF/D and after this period closed for 50 days. Initial reservoir pressure was supposed to be 4500 PSIA such that after 50 days production period, bottom holes pressure was above dew point (about 3180 PSIA) and fluid was as a single phase fluid during testing period. Figure 13 shows logarithmic plot of pressure increasing period after shut in for this well. Radial, linear and pseudo radial flow regimes which are exist in analytical model are matched with the results of numerical model.

Figure 13. Confirming numerical model using an analytical model

Sensitivity of Hydraulic Fracture Parameters

The effect of physical parameters of fracture on



pressure derivative curves has been studied. For this purpose, we used Template Gridding model including only two vertical fractures in a distance of 670 foot with each other. Initial fracture parameters are summarized in Table 4. Using the fluid with properties which introduced before and supposing reservoir pressure to be about 3280 PSIA, modeling was done in such a manner that accommodation of gas condensate around well bore can be seen.

Table 4. Initial parameters of fractures

Fracture length 300 footFracture thickness 5 inchFracture permeability 1 DarcyFracture porosity 1Fracture type vertical

Production Rate Optimization

For studying the well behavior with hydraulic fractures and accommodation of gas condensate around its well bore, selecting the best production rate is very important because as the well production rate increases, well pressure drop will be increased which causes increasing of gas condensate accommodation around well bore. For this purpose and to study condensate accommodation and pressure behavior due to production with different producing rate, supposing a constant model, we considered three different production rates, 10, 12 and 14 MMSCF/D and studied the model with these production rates.

Figure 14. Logarithmic derivative plot for different production rates

Sensitivity of Model to the Numbers of Fractures

Our main aim in this section is to study the effect of hydraulic fractures numbers on pressure derivative curves and gas condensate accommodation around well bore. For this purpose, we designed models including 1, 2, 4 and 9 hydraulic fractures with equal distances to each other. As figure 15 shows, if the numbers of fractures be increased, gas condensate accommodation around well bore will be decreased because by increasing the numbers of fractures, total pressure drop around well bore will be decreased due to increasing of production area, consequently, formation and accommodation of gas condensate will be decreased. Figure 16 shows logarithmic derivative curves for different numbers of fractures. According to this figure, in a time period less than 10 hours, there is maximum deviation between curves. When numbers of fractures increases, equivalent reservoir permeability increases, therefore, recovery will be increased and pressure drop will be decreased.

Figure 15. Gas condensate accommodation plots around well bore for different numbers of fractures

Figure 16. Logarithmic derivative curves for different numbers of fractures



Sensitivity of Model to Fractures Conductivity

Fractures conductivity has enormous effects on well production behavior. For this purpose and to see these effects on well test results and fluid saturation, we checked sensitivity of model including two fractures to multiply of fracture conductivity to fracture width for different amounts from 100 (fracture having finite permeability) to 3000 (fracture having infinite permeability). The results are shown in figures 17.

Figure 17. Log - Log pressure response curves for different fracture conductivities

As figure 16 shows, when fracture conductivity increases, considering other reservoir parameters and production rate to be constant, in a known distance from the well (for example in a distance of 16 feet), gas condensate accommodation will be decreased because in this case, connections between well bore producing area and reservoir increase and pressure drop around well bore decreases.

Sensitivity of Model to Fractures Length

Length of hydraulic fracture is one of other effective parameters which affect well pressure behavior and gas condensate accommodation around well bore. As it can be seen from figure 18, when fracture length increases, interconnections between well and reservoir increases, so invaded production area and therefore gas condensate accommodation increases.

Figure 18. logarithmic derivative curves for different fracture length

CONCLUSIONS

1. This modeling showed that gas condensate accommodation around well bore, creates a compositional behavior in reservoir and increases total skin factor of the well. When well following pressure falls below dew point, due to the condensate accommodation, a region with low conductivity will be produced around well bore.

2. Fracture parameters such as fracture length, fracture conductivity and even numbers of fractures can affect fluid flowing regime and gas condensate accommodation around well bore.

3. Results obtained from simulation, must be compared with actual well test data to be checked and confirmed.

ACKNOWLEDGEMENT

The authors wish to express their sincere thanks and gratitude to the professors, lecturers and staff of petroleum engineering department of Universiti Teknologi PETRONAS (UTP) for supporting the present research and their helps and encouragements in connection with this study. We are grateful for all their assistance.



REFERENCES

[1] R. Mott, A. Cable, and M. Spearing, “Measurements and Simulation of Inertial and High Capillary Number Flow Phenomena in Gas-Condensate Relative Permeability”, Dallas, Texas: 2000, Society of Petroleum Engineers Inc., 2000.

[2] S. Daungkaew, F. Ross, and A. Gringarten, “Well Test Investigation of Condensate Drop-out Behavior in a North Sea Lean Gas Condensate Reservoir”, 2002.

[3] I. Roebuck Jr., W.T. Ford, G.E. Henderson, and J. Douglas Jr., “The Compositional Reservoir Simulator: Case III - The Radial Geometry,” Society of Petroleum Engineers, 1969.

[4] S.B. Hinchman and R.D. Barree, “Productivity Loss in Gas Condensate Reservoirs”, Las Vegas, Nevada: 1985 Copyright 1985, Society of Petroleum Engineers, 1985.

[5] M.J. Economides, K. Dehghani, D.O. Ogbe, and R.D. Ostermann, “Hysteresis Effects for Gas Condensate Wells Undergoing Buildup Tests below the Dew Point Pressure”, Dallas, Texas: 1987 Copyright 1987, Society of Petroleum Engineers, 1987.

[6] A.M. Saleh and G. Stewart, “Interpretation of Gas Condensate Well Tests With Field Examples”, Washington, D.C.: 1992 Copyright 1992, Society of Petroleum Engineers, Inc., 1992.

[7] T. Marhaendrajana, N. Kaczorowski, and T. Blasingame, “Analysis and Interpretation of Well Test Performance at Arun Field, Indonesia”, Houston, Texas: Society of etroleum Engineers, 1999.

[8] A. Dehane, D. Tiab, and S. Osisanya, “Comparison of the Performance of Vertical and Horizontal Wells in Gas-Condensate Reservoirs,” Dallas, Texas: Copyright 2000, Society of Petroleum Engineers Inc., 2000.

[9] P. Wang and G.A. Pope, “Proper Use of Equations of State for Compositional Reservoir Simulation”, SPE Journal of Petroleum Technology, vol. 53, 2001, pp.74-81.

[10] K.H. Coats, “Simulation of Gas Condensate Reservoir Performance,” SPE Journal of Petroleum Technology, vol. 37, 1985, pp. 1870-1886.

[11] J. Liu, J. Wilkins, M. Al-Qahtani, and A. Al-Awami, “Modeling a Rich Gas Condensate Reservoir with Composition Grading and Faults”, Bahrain: Copyright 2001, Society of Petroleum Engineers Inc., 2001.

[12] M. Gondouin, R. Iffly, and J. Husson, “An Attempt to Predict the Time Dependence of Well Deliverability in Gas Condensate Fields,” vol. 7, 1967, pp. 113 - 124.

[13] P.K. Munkerud, “Measurement of Relative Permeability and Flow Properties of a Gas Condensate System during Pressure Depletion and Pressure Maintenance”, Dallas, Texas: 1989 Copyright 1989, Society of Petroleum Engineers, Inc., 1989.

[14] G.D. Henderson, A. Danesh, D.H. Tehrani, and J.M. Peden, “The effect of velocity and interfacial tension on relative permeability of gas condensate fluids in the wellbore region,” Journal of Petroleum Science and Engineering, vol. 17, May. 1997, pp. 265-273.

[15] G. Henderson, A. Danesh, D. Tehrani, and B. Al-Kharusi, “The Relative Significance of Positive Coupling and Inertial Effects on Gas Condensate Relative permeability's at High Velocity,” Dallas, Texas: 2000,. Society of Petroleum Engineers Inc., 2000.

[16] J. Wan, V.R. Penmatcha, S. Arbabi, and K. Aziza, “Effects of Grid Systems on Predicting Horizontal Well Productivity,” Bakersfield, California: 1998 Copyright1998, Society of Petroleum Engineers Inc., 1998.

[17] S.M. Blom and J. Hagoort, “The Combined Effect of Near Critical Relative Permeability and Non-Darcy Flow on Well Impairment by Condensate Drop-Out”, Calgary, Alberta, Canada: 1998 Copyright 1998, Society of Petroleum Engineers Inc., 1998.

[18] M. Bozorgzadeh and A.C. Gringarten, “New Estimate for the Radius of a Condensate Bank from Well Test Data Using Dry Gas Pseudo-Pressure”, Houston, Texas: Society of Petroleum Engineers, 2004.

[19] A.C. Gringarten, T.V. Schroeter, T. Rolfsvaag, and J. Bruner, “Use of Down hole Permanent Pressure Gauge Data to Diagnose Production Problems in a North Sea Horizontal Well”, Denver, Colorado: Society of Petroleum Engineers, 2003.

[20] M. Briones, J.A. Zambrano, and C. Zerpa, “Study of Gas Condensate Well Productivity in Santa Barbara Field, Venezuela, by Well Test Analysis,” San Antonio, Texas: Copyright 2002, Society of Petroleum Engineers Inc., 2002.

[21] R. Al-Hussainy and H.R. Jr, “Application of Real Gas Flow Theory to Well Testing and Deliverability Forecasting”, SPE Journal of Petroleum Technology, vol. 18, 1966, pp. 637-642.



[22] R. Raghavan, W.C. Chu, and J. Jones, “Practical Considerations in the Analysis of Gas-Condensate Well Tests”, SPE Reservoir Evaluation & Engineering, vol. 2, 1999, pp. 288-295.

[23] C. Lenn, F.J. Kuchuk, J. Rounce, and P. Hook, “Horizontal Well Performance Evaluation and Fluid Entry Mechanisms”, New Orleans, Louisiana: 1998 Copyright 1998, Society of Petroleum Engineers Inc., 1998.

[24] G. Tang and A. Firoozabadi, “Relative Permeability Modification in Gas-Liquid Systems through Wettability Alteration to Intermediate Gas-Wetting”, SPE Annual Technical Conference and Exhibition, Dallas, Texas: Copyright 2000, Society of Petroleum Engineers Inc., 2000.

[25] M.M. fahes and A. Firoozabadi, “Wettability Alteration on Intermediate Gas Wetting in Gas Condensate Reservoirs at High Temperatures”, Dallas, Texas: Society of Petroleum Engineers, 2005.

[26] D. Marokane, A. Logmo-Ngog, and R. Sarkar, “Applicability of Timely Gas Injection in Gas Condensate Fields To Improve Well Productivity”, Tulsa, Oklahoma: Copyright 2002, Society of Petroleum Engineers Inc., 2002.

[27] H. Mukherjee and M.J. Economides, “A Parametric Comparison of Horizontal and Vertical Well Performance”, SPE Formation Evaluation, vol. 6, 1991, pp. 209-216.

[28] M. Jamiolahmady, A. Danesh, G.H. G, and D. Tehrani, “Offshore Europe”, Aberdeen, United Kingdom: Society of Petroleum Engineers, 2003.

[29] Xiuli Wang, "Advanced Natural Gas Engineering", Gulf Publishing Company, Houston, Texas, 2009.

[30] Ali Danesh, "PVT and Phase Behavior of Petroleum Reservoir Fluids", Elsevier Science & Technology Books, 1998.

[31] Z. Novosad, “Composition and Phase Changes in Testing and Producing Retrograde Gas Wells”, SPE Reservoir Engineering, vol. 11, pp. 231-235, 1996.

[32] A. C. Gringarten, “Well Productivity in Gas-Condensate and Volatile Oil Reservoirs: Diagnosis and Enhancement (Phase 3)”, Centre for Petroleum Studies, Imperial College, London SW7 2AZ, UK.

[33] http://www.fekete.com/SAN/TheoryAndEquations/HarmonyTheoryEquations/Content/HTML_Files/Reference_Material/General_Concepts/CBM_Properties.htm.

[34] Patrick A. Hammond, “A Reinterpretation of Historic Aquifer Tests of Two Hydraulically Fractured Wells by Application of Inverse Analysis, Derivative Analysis, and Diagnostic Plots”, Journal of Water Resource and Protection, Vol.6 No.5, 2014.

AUTHORS' INFORMATION

Mohammed Abdalla Ayoub is a lecturer at the Petroleum Engineering Department, Universiti Teknologi PETRONAS, Malaysia. Previously, he was an assistant professor at the Petroleum Engineering Department at University of Khartoum, Sudan. Dr Ayoub holds an MSc in Petroleum Engineering from

KFUPM, and a PhD from UTP, Malaysia. His main interests include hybrid EOR techniques, application of Artificial Intelligence tools into petroleum engineering related problems. He is a member of the Sudanese engineering council as well as SPE.

Abdollah Esmaeili received his Diploma in Mathematics & Physics in 1990. He obtained his B.Sc and MSc. in Petroleum Engineering from Petroleum University of Technology (PUT), Iran. He is currently pursuing his PhD in Petroleum Engineering at Universiti Teknologi PETRONAS (UTP), Malaysia.

He has been working as a reservoir engineer in the oil & gas industry for the past 20 years for companies like National Iranian Oil Company (N.I.O.C), National Iranian South Oil Company (N.I.S.O.C), Aghajari Oil & Gas Production Company (AjOGPC). He also taught petroleum reservoir engineering courses in universities of Iran. He has written several papers in petroleum engineering which were accepted for presentation in international conferences. He has attended several international conferences worldwideas a speaker. He has led several international scientific master classes and workshops worldwide.



A STUDY ON GAS FLARING ACTIVITIES IN NIGERIA

1Aliyu A. Sulaimon and 2Modupe Ogunmekan 1Department of Petroleum Engineering, Universiti Teknologi PETRONAS

2Department of Petroleum Engineering, University of IbadanEmail: [email protected]

ABSTRACT

This paper presents the findings on the current trends of gas flaring in six major oil companies in Nigeria from 2007 – 2011. Data were obtained from state-owned petroleum corporation. Analysis shows that gas flaring activities is gradually decreasing with time. A multinational company flared more than 80% of its gas in 2007 while another company has maintained the best record among the six multinational companies due to consistent reduction in its flaring activities in recent years. Among others, it is recommended that an effective implementation of regulations, provision of reward packages and development of infrastructure will ameliorate environmental degradation attributed to gas flaring activities.

Keywords: Gas, flaring, production, companies

INTRODUCTION

Gas flaring is the burning of the natural gas that is associated with crude oil during production operations [1]. Considering the high baseline in 1996, gas flaring has more than halved in Nigeria. Despite this achievement, Nigeria remains the second largest flaring country in the world and emits around $1.8billion worth of gas annually [2]. Besides, the aspects of the gas flaring evaluation are the associated environmental hazards, considerable economic loss, and energy waste despite the country’s rampant energy poverty. Although, flaring has been legally banned in Nigeria since 1984, measures to enforce the ban have been lacklustre. The concerned environmental organizations such as the Climate Change Nigeria (CCN) opined that the fines being charged for the gas flaring are much too low per cubic meters to deter companies from burning gas. Under such lame prohibition, there are little incentives for entrepreneurs to make the huge infrastructural investments required for the commercialization of gas [3].

STUDY RATIONALE

The growth of the economy of a country like Nigeria is dependent on industrial productivity hence any policy that can improve the situation should be given high priority. The potentials of natural gas as a catalyst for enhanced productivity are numerous. It is a versatile, efficient, and environmental friendly fuel. It is a source of feed stock for various chemical manufacturing process, an automotive fuel and feedstock for the production of fertilizers in the countries where efficient food production remains a top priority [4]. Therefore, to achieve a zero-flare of gas and maximum utilization of produced gas, it is important to investigate the trend at which the major oil companies have achieved targets for the implementation of gas flare reduction.

Diugwu et al [6] investigated the impact of gas production, utilization, and flaring on the estimated monetary value of the goods and services produced in Nigeria (GDP), using multiple linear regression analysis. They confirmed the widely believed notion



that gas utilization has positive impacts on the nation’s GDP while gas production and flaring are negatively correlated with the GDP. They recommended more investments in infrastructure and review of the regulatory framework for effective control and management of Nigeria’s oil and gas industry.

METHODOLOGY

The population of the study constitutes all the oil companies in Nigeria though flaring activities of the six major multinationals denoted as Companies C, M, N, P, S, and T were investigated and analysed. There was no random sampling. The purposive sampling technique was used to select all the major oil companies in Nigeria for the study. The companies were selected based on their pioneering petroleum exploitation in Nigeria which spans decades; hence data concerning their activities are more readily available [5]. In addition, they have overcome foundational problems and hence are more established and production costs are more stable than newer companies. Time series data of Associated Petroleum Gas (APG) was collected through the statistical bulletin of the Nigerian National Petroleum Corporation (NNPC) for several years. Data are presented in tables and graphs and descriptively analysed.

RESULT AND DISCUSSION

By statistical estimates in 2007, Company C flared the highest volume of gas to about 86.14% of its production, followed by company N flaring 38.87%. Company M flared 30.61% while companies S and T flared the lowest volume of gas to the tune of 12% and 11.08% respectively as shown in Figure 1. (Tables 1, 2 & 3)

Figure 1. Gas Production and Flaring in 2007

Table 1 : APG Production Output (MMscf )

Company 2007 2008 2009 2010 2011

C 191 243 166 194 266

M 464 427 427 479 278

N 320 293 272 441 470

P 0 21 207 8 12

S 763 800 455 777 864

T 289 320 302 277 260

*APG: Associated Petroleum Gas

Table 2 : APG Flaring Rate (%) (MMscf )

Company 2007 2008 2009 2010 2011

C 86.14 68.08 7.80 60.88 39.53

M 30.61 30.67 28.04 25.61 23.75

N 33.87 32.81 26.11 23.29 46.55

P 0 7.61 97.35 84.09 29.88

S 12.00 12.22 17.1 13.31 9.87

T 11.08 11.01 8.86 10.99 11.08



Table 3 : APG Flaring Output (MMscf )

Company 2007 2008 2009 2010 2011

C 162 142 112 118 105

M 183 130 122 122 66

N 108 96 71 102 219

P 0 21 201 6 3

S 96 97 77 103 85

T 33 35 26 30 28

In 2008, company C was still the leader in gas flaring though the flared volume had reduced to 68.08%, companies M and N flared 30.67% and 32.8% respectively while companies S, T, and P maintained the lowest volume of gas flaring (Figure 2).


In 2009, company P flared the largest amount of gas to the tune of about 97.35 of its APG output, while companies T and C were able to reduce their contributions to between 7% - 8%. Companies M and N however maintained their respective flared quantities at 17% and 26% while company S increased its share by 5% (Figure 3).


In 2010, companies P and C flared 84.09% and 60.88% of their APG respectively (Figure 4). Companies M and N similarly maintained their flaring volumes as in the previous year while companies S and T made further progress by reducing the quantities of their flared gas to 13.31% and 10.99% respectively.




Due to the zero flaring policy by 2012, there was a significant decrease in flaring volumes in 2011 for all the six oil producing companies under study in Nigeria as illustrated in Figure 5. In 2011, a total of 506 MMSCF of natural gas out of 2150 MMSCF produced was flared. This amounted to 23.53% of the total gas produced which is less than the 28.71% that was flared in 2007. Thus, on the average, less than a quarter of all gas produced is flared.


CONCLUSION

The empirical investigation focused on finding out the amount of gas flared by several major oil companies in Nigeria. The study also showed the general trends in gas flaring by the six major oil companies in Nigeria and particular trends in gas flaring by each oil company. Furthermore the gas flaring activities of various oil companies in Nigeria from 2011 to 2014 requires further investigation in order to identify which company’s flaring activities constitutes the greatest economic waste and environmental hazard through percentages and ranking. Flaring activities fluctuated but have been falling consistently in recent years such that presently only about a quarter of all gas produced is flared.

RECOMMENDATIONS

Reduction in gas flaring can be achieved if only the following measures can be implemented;

1. Accelerate the development of the domestic energy market.

2. Provide for the effective integration of energy supplies.

3. Significant scaling-up of renewable energy for electricity for both grid and off-grid distribution.

4. End gas flaring through a harmonization of gas management and electricity sector development.

5. Structured plans for the transformation away from oil & gas dependence towards renewable energy supplies.

6. Streamlined governance and increase support from the Nigerian people, and

7. Oil companies can be further motivated to reduce flaring activities by the government introducing national awards or tax holidays for companies with best records in reduction of gas flaring activities.

ACKNOWLEDGEMENT

The authors acknowledge Nigerian National Petroleum Corporation (NNPC) for providing the data used for this study.

REFERENCES

[1] Jinn (2010): “Gas Flaring in Nigeria: An Overview”, downloaded on 22nd February 2012, http://j u s t i c e i n n i g e r i a n o w. o r g / j i n n / w p - c o n t e n t /uploads/2010/04/JINN-2010-Gas-Flar ing-an-overview.pdf, downloaded on 22nd February 2012.



[2] Hassan A. and Kouhy R. (2013): “Gas flaring in Nigeria: Analysis of changes in its consequent carbon emission and reporting”, Accounting Forum, Vol. 37, Issue 2, pp. 124–134.

[3] Ebrahim Z. and Friedrichs J. (2013): “Gas flare: The burning issue”, http://www.resilience.org/stories/2013-09-03/gas-flaring-the-burning-issue; downloaded on 3rd Nov. 2013.

[4] Akpan, S. (2009): “Production and Utilization of Natural Gas Resources in Nigeria - A review”, Paper SPE-128356-MS presented at the Nigeria Annual International Conference and Exhibition, 3-5 Aug. Abuja, http://dx.doi.org/10.2118/128356-MS

[5] Madueme S. (2010): “Gas Flaring Activities of Major Oil Companies in Nigeria: An Economic Investigation”, International Journal of Engineering Science and Technology Vol. 2(4), 2010, 610-617.

[6] Diugwu I.A., Ijaiya M.A, Mohammed M., and Egila A.E. (2013): “The Effect of Gas Production, Utilization, and Flaring on the Economic Growth of Nigeria”, Natural Resources, 2013, 4, 341-348; http://dx.doi.org/10.4236/nr.2013.44041.


Aliyu A. Sulaimon (Ph.D.) joined the Universiti Teknologi PETRONAS (UTP) as a Senior Lecturer in the Department of Petroleum Engineering in 2012. He has previously taught and supervised students at both Undergraduate and Postgraduate levels at Nigeria’s Premier University, the University of Ibadan for

12 years. His research interest include but not limited to Flow Assurance and Well Testing.

Modupe Ogunmekan is a petroleum analyst. She graduated from the Department of Petroleum Engineering, University of Ibadan, Nigeria in 2012. She is a member of SPE and currently a postgraduate student at the same university.



GENDER DIVERSITY ON BOARDS AND MARKET PERFORMANCE: AN EMPIRICAL INVESTIGATION ON

MALAYSIAN LISTED COMPANIES

Rohail Hassan, Maran MarimuthuDepartment of Management and Humanities, University of Teknologi PETRONAS

Email: [email protected]; [email protected]

ABSTRACT

Lack of women participation on boards becomes an important issue recently, which in turn reflects gender diversity in the top management. This study attempts to examine the effect of gender diversity among corporate boards on market value. Theoretical framework is specially designed using concepts, measures and models. A sample of 565 large companies is considered based on their market capitalization. The largest sectors are trading & services, finance and industrial products. This study incorporates descriptive statistics, Anova, correlation testing and regression analysis. The results indicate that women participation does have a relationship with firm value.

Keywords: women, gender diversity, boards, market performance

INTRODUCTION

Developing societal, political and cultural views of corporate board members are part of demographic diversity of top management. In additional, the major factor is better corporate governance and the global desire [1]. The world faced high profile scandals like Worldcom, Enron, Adelphia and policy makers began to concentrate on the issues of corporate governance [2]. In 1997, Malaysian economy was badly affected by financial crisis and many major corporations had shut down. This corporate failure on financial crisis was the result of poor corporate governance [3]. Good corporate governance is an important part of business philosophy. Corporate governance means something broader than corporate management in view of achieving strategic goals [4]. In contrast, researchers found that gender diversity among board members could result in poor firm performance [5]. Board gender diversity has a positive relationship with organizational performance. There is ambiguity among previous researches [6]. Undoubtedly, the

Malaysian government and other regulatory bodies may benefit from the findings of study. This research is unique as we consider the relationship between women participation on boards (i.e. gender diversity among board members) with market value. A total sample of 565 large companies is considered from the population of 938 listed companies. The three largest sectors are trading/ services, finance, industrial products and comprising of 233, 43, 289 companies respectively. Descriptive statistics, graphical presentations, Anova, regression and correlation test have been used.

The remainder of the paper follows as such Sections 2 and 3 explain literature review, women participation and gender diversity issues in Malaysia. In section 4, 5 & 6, theoretical framework, objectives and focus of the study are included and section 7 explains research design of this study. Empirical results and limitations are explained in section 8 and 9. Finally, our work is concluded in the last section.



LITERATURE REVIEW

A. Theoretical Perspective on Gender Diversity

Board members are the highest authority in a firm. The board members generally believed to perform four important functions. Firstly, monitor and control to managers. Secondly, provide guidance and information. Thirdly, monitor conformance with the laws and regulations, and lastly, connecting the organization to the external environment. One basic proposition is that firm performance is partially determined by the board composition. Board composition includes gender diversity, which is directly linked with firm performance. There is no specific theory that predicts the nature of relationship between board diversity and financial performance. In this study, related theories from various aspects discussed.

B. Agency Theory

Agency theory explains the board functions of monitoring and controlling manages. It is based on the “agency problem” that arises when two parties have differing goals. The relationship between principal and agent defined as a contract where the principal engages the agent in his/her duties to the principal. Agency relationship plays a very important role in firm performance and it is influenced by composition of board [7]. Diverse boards may better monitor managers and top management teams because board diversity increase board independence [8].The agency theory particularly explains the relationship between gender diversity in boards and a firm performance. Women directors behave differently from their male counterparts and their presence changes board behavior as they are said to provide better monitoring and advisory services [9]. Having more women on boards may increase the reputation of the company[10]. The agency theory is primary concept of corporate governance that analyzes the relationships among different interested parties like shareholders, boards, managers, and employees [7]. Agency theory can show how the principals of a firm are in a weak position compared to the managers [11].

In addition, The types diversity are gender diversity [5], ethnic diversity, educational diversity and board independence [12]. Researchers have not found any conclusive findings with regard to the relationship of gender diversity and firm performance.

C. Upper-echelon Theory

The upper echelons theory is based on behavioral decision-making theories as well as concepts of organizational demography. Corporate boards are important and can be used for organizational outcomes such as firm performance and strategic achievements. This theory explains the impacts of demographic and cognitive diversity in context of firm performance. Researchers argued that diversified boards can make more effective decisions as compared to homogenous boards[13]. Organizations can attract, retain, and gain competitive edge from diverse talent to begin by increasing the diversity among top management teams [14]. This study uses the upper echelon theory as it involves the characteristics of top level management and their effects on firm performance.

WOMEN PARTICIPATION AND GENDER DIVERSITY ISSUES

Gender is a status, which is constructed through social, cultural, and psychological means; it is not based on personal traits [15]. Gender diversity is an integral of board diversity. Board diversity claims that board should reflect society and represent the gender, ethnicity, professional experience and background [16]. Diversity among boards always brings a better understanding of market place, innovation, increases creativity, leadership, better decision making and effective global relationships [17]. Gender diversity among board members has drawn the attention of researchers and issues examined includes; why fewer females on corporate boards [18] and what is the female role or how much experience that influences the corporate boards and firm performance [19].



Figure 1. Aggregate percentage of women on boards in Asian emerging economies (2009-2011)

In corporate world, female participation on boards is very low. According to Catalyst census, female participation among boards is only 12.4 per cent in the US and 6.4 in the UK [20].Women representation is less than 5 percent in Canadian boards [18]. A researcher finds that women on boards have significant positive impact on firm performance [21] but there could be no relationship between women directors and shareholder returns. Thus, there is a need to further investigate this phenomenon. Developed countries like Norway, requires at least 40 per cent women participation on board since 2008 [22]. In the view of such regulations and increasing importance of women in the corporate world, there is a need to explore and justify the impact of gender diversity on firm performance.

Women representation in the boardrooms in some of the developing countries is presented in Fig.1. This may reflect the importance of women participation on corporate boards. In relation to this, researchers suggest that a critical mass is needed to allow a board to take advantage of gender diversity (Governance Metrics International, 2011).

In relation to this, European countries have recently implemented laws for gender quotas in the business and public sectors since they believe that the presence of women in boardrooms may affect firm performance significantly. For example, Spain legislated that by 2015 women must represent 9.3 percent of seats in boardrooms, and in the Netherlands the requirement is that 30 percent of board members shall be women by 2015 (Governance Metrics International, 2011). In Wall Mart, females account for 30 percent who are holding managerial positions and in Malaysian scenario females hold only 2 percent of the managerial positions (Diversity and Inclusion Network Asia, 2012). Currently, Malaysia has laws and regulations, which are encouraging women participation quotas on boards in private sector and required companies to engage at least 30 percent females at board level (MCCG, 2012). In June 2011, the Malaysian government established a goal of 30 percent female on holding of senior positions in public sector by 2016 (SCM, 2011).

THEORETICAL FRAMEWORK OF GENDER DIVERSITY AT BOARD LEVEL AND FIRM PERFORMANCE

Fig.2 below shows the theoretical framework describes presents the relationship between diversity and firm performance. Here, variable of investigation is gender diversity. Gender diversity can be measured through dividing total females by total board members on board. The dependent variable of this study is firm performance which can be measured using market capitalization .The control variables of this study is board size. “Board size” refers to total number of directors on the board.



Figure 2. Theoretical Framework (Gender diversity at board level and its impact on firm performance)

OBJECTIVE OF THE STUDY

This study seeks to examine the effect of women participation at board level on firm performance.

FOCUS OF THE STUDY

It is of great interest for many researchers to explore the gender diversity with regard to firm performance. Listed companies have different scenarios of corporate boards with regard to women quota. The empirical examination offers more insights on gender diversity (women participation) at board level. This study also includes the extent to which the findings differ from the general expectation as argued by the previous studies. The following are hypotheses of this study

H1: Gender diversity among the board members (BODs) has a positive impact on firm (market) performance.

H2: Board size has a positive impact on firm (market) performance.

RESEARCH DESIGN

A. Sample and Data

This study is specially designed to examine the gender diversity trend among corporate boards of listed companies in Malaysia. A sample of 565 companies was considered based on the population of 965 companies. The sampling technique was judgmental in nature and a 5-year average market capitalization was used to measure the significance of the individual sectors. Hence, three major sectors were identified; they were Trading/Services, Finance and Industrial Products. As for data analysis, different statistical techniques were adopted, they include descriptive analysis, ANOVA, correlation test and bivariate regression model.

B. Model

The below modeling equation shows relationship between gender diversity (women participation) and market capitalization. Regression model to determine the relationship between gender diversity and market capitalization is as follows;

Firm Performance(Y) = α + βGender Div + Ɛ

EMPIRICAL ANALYSIS AND DISCUSSIONS

A. Market Capitalization by Sector

Figure 3. Total Market Capitalizations by Sector

Fig.3. presents the breakdown of market capitalization by sector. The leading sectors are Trading / Service, Finance and industrial Products. In this study, we focused the three major sectors to examine the effect



of women participation on boards on market value.

Table I. presents descriptive statistics of the three largest sectors in Malaysia based on market capitalization. These three sectors accounted for 565 companies listed on Bursa Malaysia. The leading sector is Trading / Services with the largest market capitalization but women participation rate in the Finance sector was the highest. Fig.4. below shows the percentage of total market capitalization controlled by the three largest sectors.

Table I. Summary of major sectors

No

No

Of

Com

pan

ies

Sect

or

Tota

l Mar

ket

Cap

(Bill

ion)

Prop

orti

on

(%)

Boa

rd S

ize

Fem

ales

on

Boa

rd

% o

f Fe

mal

e on

Boa

rd

1 233

Trad

ing/

Serv

ices 1.95 53.94 1745 134 7.67

2 43

Fina

nce

1.09 30.14 340 37 10.8

3 289

Indu

stri

al

Prod

uct

0.57 15.90 2053 148 7.20

Tota

l

565 36.08 100 4138 319 7.70

Figure 4. Market Capitalizations of Trading/Services, Finance & Industrial Products

In the context of Malaysia, conducting studies on diversity among the board members is always a great challenge. This makes the understanding of diversity in the boardrooms becoming more complex and time consuming. According to the 2009 data, the average percentage of women on boards in Malaysia is 5.9 percent. The representation of women in Malaysian boardrooms is relatively lower as compared to that of other emerging markets in Asia (Governance Metrics International, 2009).

The line graphs in Fig.5 offers more insights, while gender diversity fluctuates, board size also constantly fluctuates with some spikes and market capitalization is presented in a descending order. It can be concluded that there is a positive relationship between gender diversity and market capitalization. This involves data of all the 938 companies listed on Bursa Malaysia.

Figure 5. Graphical expression of Market Capitalization, Board Size and Women Participation

B. Descriptive Statistics

The descriptive statistics results of variables are presented in TABLE II. The largest board size is 16 and a maximum woman on a board is 16. The average women participation at board level is 0.57. Sample is N = 565. The Kurtosis score for gender diversity is 8.878.



Table II. Descriptive Statistics on Market Capitalization & Gender Diversity

Market Cap.

Board Size

Gender Gender Div.

Mean 5.7862 7.33 .57 .0763

Median 5.6500 7.00 .00 .0000

Mode 6.23 7 0 .00

Std. Deviation

.81046 1.939 .800 .11097

Variance .657 3.761 .640 .012

Skewness .783 .717 1.692 2.090

Kurtosis .809 1.019 4.428 8.878

Range 5.11 14 6 1.00

Minimum 3.32 2 0 .00

Maximum 8.43 16 6 1.00

N= 565

C. Graphical Expressions

The bar chart below Fig. 6 offers insightful information that as women participation percentage in three major sectors Trading / Service, Finance and industrial products. It can be concluded that women participation is more persistent in finance sector.

Figure 6. Bar Chart – Percentage of Women Participation in Three Major Sectors

In addition, the line graph Fig.7 presents that as gender diversity fluctuates and market capitalization constantly fluctuates with spikes. Market capitalization is presented in a descending order. It can be concluded that there is a positive relationship between women participation and market capitalization.

Figure 7. Line Chart Market Capitalization Versus Gender Diversity

The line graph Fig. 8 presents that as women participation at board level fluctuates, board size also constantly fluctuates with some spikes. It can be concluded that there is a relationship between women participation and board size. Fig.8 involves data of all selected 565 companies listed on Bursa Malaysia.

Figure 8. Graphical Expression of Board Size, Gender Diversity

The scatter diagram Fig.9 presents that gender diversity which is more concentrated between 4 & 14 members on board. It can be concluded that companies that have board size between 6 and 12



tend to have greater gender diversity.

Figure 9. Scatter Diagram between Gender Diversity, Gender (Number of Females) and Board Size The line graph Fig. 10 presents line graphs of board size, gender size and gender diversity. As the gender diversity fluctuates, board size also constantly fluctuates with some spikes. Fig. 10 involves data of all the 938 companies listed on Bursa Malaysia. In nutshell, as board size increases, the women participation also increases and indirectly it has some impact on gender diversity.

Figure 10. Graphical Expression of Gender (Number of Females), Board Size and Gender Diversity

D. Correlations Test

TABLE III presents the results of correlations test. Market capitalization has a strong relationship with board size and gender diversity at 0.05 significance level. However, there is no correlation between board size and gender diversity. We can conclude that gender influence does have impact on firm performance.

Table III. Correlation Results

Variables 1 2 3

1. Market Cap. 1 .433** .083*

2.Board Size 1 .034

3.Gender Div. . 1

** Correlation is significant at the 0.01 level (2-tailed).

* Correlation is significant at the 0.05 level (2-tailed).

E. One- Way ANOVA

Below TABLE IV presents the results of One-Way ANOVA test. We can conclude, there is no significant difference among the three sectors with regard to market performance as represented by p-value (0.162) at 0.05. This signifies that there is no sectoral influence on market performance in the presence of the gender effect.

Table IV. One Way Anova: Gender Diversity Among Sectors

Source DF SS MS F P

Category 2 0.0446 0.0223 1.82 0.162

Error 561 6.8616 0.0122

Total 563 6.9062

*S = 0.1106 R-Sq = 0.65% R-Sq (adj) = 0.29%, Individual 95% CIs



F. Regression Analysis

In addition, a bivariate regression test was conducted on the gender effect with regard to market performance. The result shows that there is a significant positive relationship between gender diversity and market performance at 0.05 as illustrated below.

Firm Performance- Market Cap. (Y) = α + βGender Div + Ɛ Market Cap. (Y) = 5.740 + 0.68Gender Div + 0.307

Beta coefficient of gender effect is 0.68 (t-value is 1.981), and the F-test (3.923) is very significant at 0.05

LIMITATIONS

This paper only focuses on gender diversity (women participation) and large listed companies. In addition, this paper incorporated only the three major sectors (trading/services, finance, industrial products) and thus, it would be quite challenging to generalize the findings but however, these three sectors’ contribution is exceeding 50 per cent of the total market capitalization.

CONCLUSIONS AND RECOMMENDATIONS

Admittedly, women participation at board level has a significant and positive relationship with market performance. Women participation should be encouraged at top management level. It should be noted that only large companies seem to serious about promoting gender diversity (women participation) at top level management. Hence, that women participation should be leveled at top management level, that in turn, can enhance the profits and perhaps improving internal operations of their companies.

Based on empirical findings, some recommendations can be suggested for encouraging women participation among board level in listed companies

of Malaysia. Obviously, a mandatory quota for women at board level should be established. Government and other regulatory bodies should ensure that necessary requirements should be imposed on companies.

REFERENCES

[1] R. Monks and N. Minow, "Corporate Governance, Vol. 3," ed: Blackwell Publishing, Malden, MA, 2004.

[2] O. f. E. Co-operation and Development, OECD Principles of Corporate Governance 2004: OECD Publishing, 2004.

[3] T. Mitton, "A cross-firm analysis of the impact of corporate governance on the East Asian financial crisis," Journal of financial economics, vol. 64, pp. 215-241, 2002.

[4] V. Bairathi, "Corporate governance: A suggestive code," International Research Journal, vol. 11, pp. 753-754, 2009.

[5] R. B. Adams and D. Ferreira, "Women in the boardroom and their impact on governance and performance," Journal of financial economics, vol. 94, pp. 291-309, 2009.

[6] N. M. Carter and C. Silva, Mentoring: Necessary but insufficient for advancement: Catalyst New York, 2010.

[7] M. C. Jensen and W. H. Meckling, "Agency Costs and the Theory of the Firm," Journal of Financial Economics, vol. 3, pp. 305-360, 1976.

[8] D. A. Carter, B. J. Simkins, and W. G. Simpson, "Corporate governance, board diversity, and firm value," Financial Review, vol. 38, pp. 33-53, 2003.

[9] I. A. G. Azmi and M. A. Barrett, "Women on boards and company financial performance: A study of Malaysian SMEs," in Proceedings of 3rd global accounting, finance and economics conference, Melbourne, Australia, 2013.

[10] M. Lückerath-Rovers and A. De Bos, "Code of Conduct for Non-Executive and Supervisory Directors," Journal of business ethics, vol. 100, pp. 465-481, 2011.

[11] J. Heath, "The uses and abuses of agency theory," Business Ethics Quarterly, vol. 19, pp. 497-528, 2009.



[12] T. W. Chamberlain, "Board Composition and Firm Performance: Some Canadian Evidence," International advances in economic research, vol. 16, pp. 421-422, 2010.

[13] D. C. Hambrick and P. A. Mason, "Upper echelons: The organization as a reflection of its top managers," Academy of management review, vol. 9, pp. 193-206, 1984.

[14] M. J. G. J. L. Raver and L. H. N. B. Schneider, "Discrimination in organizations: An organizational-level systems perspective," Discrimination at work: The psychological and organizational bases, p. 89, 2004.

[15] C. West and D. H. Zimmerman, "Doing gender," Gender & society, vol. 1, pp. 125-151, 1987.

[16] F. J. Milliken and L. L. Martins, "Searching for common threads: Understanding the multiple effects of diversity in organizational groups," Academy of management review, vol. 21, pp. 402-433, 1996.

[17] G. Robinson and K. Dechant, "Building a business case for diversity," The Academy of Management Executive, vol. 11, pp. 21-31, 1997.

[18] R. J. Burke, "Women on corporate boards of directors: A needed resource," in Women in Corporate Management, ed: Springer, 1997, pp. 37-43.

[19] D. Jamali, A. Safieddine, and M. Daouk, "Corporate governance and women: an empirical study of top and middle women managers in the Lebanese banking sector," Corporate Governance, vol. 7, pp. 574-585, 2007.

[20] V. Singh and S. Vinnicombe, "Women pass a milestone: 101 directorships on the FTSE 100 boards," 2004.

[21] N. Smith, V. Smith, and M. Verner, "Do women in top management affect firm performance? A panel study of 2,500 Danish firms," International Journal of Productivity and Performance Management, vol. 55, pp. 569-593, 2006.

[22] G. Monbiot, "The denial industry," The Guardian, vol. 19, 2006.


Rohail Hassan is currently pursuing Doctor of Philosophy (PhD) in Management Sciences at Universiti Teknologi PETRONAS, Malaysia. He received his undergraduate degree of Bs (HONS) in Business & Management Sciences from Institute of Administrative Sciences, University of

The Punjab, Lahore, Pakistan in 2010 and later, his graduate degree Masters of Philosophy (TQM) from Institute of Quality & Technology Management, University of The Punjab, Lahore, Pakistan in 2012. His research interests are Corporate Governance, Strategic Management, Financial Management, Total Quality Management (TQM), Quality Tools and Techniques, Supply Chain Management (SCM), Process Improvement (Six Sigma. FMEA, Lean Six Sigma), International Standardization & Firm Performance and his current area of research include diversity issues, corporate governance and firm financial performance

Maran Marimuthu is a Senior Lecturer at the Department of Management and Humanities, Universiti Teknologi PETRONAS. Maran Marimuthu holds a doctorate in Finance from University of Malaya in 2011. He obtained his MBA from University of Malaya in 2000 and his BBA (Hons) from Universiti Utara

Malaysia in 1994. He teaches Financial Management, Business Finance, Corporate Finance and Research Methods in Finance. To date, he has published more than fifty international refereed journals and conference proceedings (local and overseas). He has co-authored books on Financial Management, International Finance and contributed articles for both local and international magazines. Dr. Maran Marimuthu is currently holding internal and external research grants as Principal and Co-Researcher and supervising doctoral and masters students in the related area. He also has a special interest in the multidisciplinary work. Dr. Maran Marimuthu has corporate experience of four years, fifteen years of academia career.



FUZZY LINEAR MODEL FOR ELECTRICITY PLANNING ININDIA WITH FOCUS ON COST OF BIOGAS POWER

Jebaraj SDepartment of Mechanical Engineering,

Universiti Teknologi PETRONAS Email: [email protected], [email protected]

ABSTRACT

Energy is a vital input for the growth of any nation. Renewable energy sources are likely to contribute a major role in catering the future electricity requirement of a developing country like India. Development of an electricity planning model will help in the proper allocation of the energy sources in meeting the future electricity demand in India. In this paper, an attempt has been made to develop a fuzzy linear programming optimal electricity allocation model that minimizes the cost and determines the optimum allocation of different energy sources for the centralized and decentralized power generation in India with special emphasis on cost factor for the biogas based power generation. The potential of energy sources, energy demand, efficiency of the energy systems, emission released by the energy systems and carbon tax for the emissions released by each system are used as constraints in the model. The model allocates the electricity distribution pattern for the year 2020 in India, which would be helpful for policy makers in commercializing the biogas plants to the maximum extent. The results indicate that the electricity distribution pattern would be 45,520 GWh (11 %) from the biomass gasifier plants, 14,112 GWh (3 %) from the biogas plants, respectively for the year 2020. This study will help in the formation of strategies for effective utilization of energy sources in India and countries like Malaysia.

Keywords— biogas; energy planning; OEAM model; energy scenario, cost of power generation.

INTRODUCTION

Energy demand is increasing rapidly because of advancements in the agricultural, industrial, commercial, information technology and transportation sectors. The improved life style and population rise are other reasons for the increase in electricity demand. India has made rapid growth in the development of electricity generation, with a capacity increasing from nearly 1,700 MW in 1950 to nearly 243.02 GW in March 2014 [1]. Yet, the demand is continuously increasing. Considering the depleting nature and health impairment effects of commercial energy sources like coal and oil, it can be said that renewable energy can be obtained at a reasonable social cost. Even though the present cost of the power

generation from the renewable energy sources is more when compared to the conventional energy sources, there is a possibility for further reduction by achieving higher efficiencies of the energy conversion devices by using improved technologies, large scale demands, by implementing modern manufacturing methods for the energy devices and by locating suitable sites where renewable energy sources could be applied effectively. An energy planning model would facilitate the effective utilization of renewable energy sources for the electricity generation in India. A scenario based on the unit cost of biogas based power generation to improve the quality of life in India have been proposed and discussed.



REVIEWS ON ALLOCATION MODELS

A review of different kinds of energy planning models has been listed in the following sections. Bunn and Vlahos (1992) formulated a model to support adversarial decision processes for electricity planning [2]. The greenhouse gas mitigation potential of biomass energy technologies in Vietnam using the long range alternative planning system model had been formulated in the year 2003 [3]. A multi-objective fuzzy linear programming approach was used by Chinmoy Jana and Chattopadhyay for the block level energy planning for domestic lighting in the year 2004 [4]. A bottom-up approach was used for decentralized energy planning for Tumkur district in India [5]. A review on optimization methods applied to renewable and sustainable energy had been studied in the year 2011 [6]. Benjamin K. Sovacool and L.C. Bulan (2012) analyzed the energy security and hydropower development in Malaysia with the focus on the drivers and challenges facing the Sarawak Corridor of Renewable Energy (SCORE) [7]. Pierluigi Mancarella had discussed the overview of concepts and evaluation models in multi-energy systems (2014) [8]. In the present paper, Optimal Electricity Allocation Model (OEAM) has been formulated with special emphasis to cost of power generation by the biogas plants.

OPTIMAL ELECTRICITY ALLOCATION MODEL (OEAM)

The OEAM model optimizes and allocates the appropriate energy sources for the electricity generation on the factors such as cost, potential, demand, efficiency, emission and carbon tax. The objective function of the model is minimizing the unit cost of power generation. The other factors are used as constraints in the model. The mathematical formulation of the Optimal Electricity Allocation Model (OEAM) is given in the following equations:

Where C is the unit cost of the energy system, ή is the efficiency of the energy system, l is the number of energy systems for power generation =20, m is the number of system in respective resources = 19, D is the energy demand (GWh), k is the resources, P is the potential of sources (GWh), En is the emission constant (g/GWh), Tn is the target emission level (g/year), r is the carbon tax (Rs/ton), R is the projected carbon tax in 2020 (Rs/ton), X is the quantum of energy (GWh), i is the various energy systems and j is the power generation option. Among the five constraints considered in the model, the factors such as efficiency, emission and carbon tax are considered as fuzzy linear constraints while demand and potential are considered as ordinary linear constraints. Here, the constraints such as efficiency, emission and carbon tax do not have crisp values and hence they are treated as fuzzy linear constraints. On the other hand, potential and demand are considered as ordinary linear constraints since they have crisp values.

Fig. 1 shows the schematic representation of the development of Optimal Electricity Allocation Model (OEAM). The demand for electricity during 2020 in India would be 993,385 GWh, which is predicted by the artificial neural network (ANN) forecasting model. The present energy demand is 565,102 GWh. Hence; the energy gap for the year 2020 has been calculated



as 428,283 GWh. This energy gap should be met by the various energy options with special consideration to the cost of power generation by biogas plants in India. There are twenty different energy systems namely coal, diesel, gas, nuclear, hydro, wind, biodiesel, biomass

gasifier, biogas, solid waste, cogeneration, ethanol, solar PV, solar thermal, OTEC, tidal, geothermal, mini hydel, fuel cell and MHD are considered in the model for the allocation of energy to meet the electricity gap (demand) for the year 2020.

Figure 1. Schematic representation of the Optimal Electricity Allocation Model (OEAM).

RESULTS AND DISCUSSION

The OEAM model allocates the different energy sources in India for the year 2020 which is illustrated in Fig. 2. It is observed that the hydro based power generation amounts to 191,100 GWh, which is nearly 44 % of the total energy demand. Subsequently, the nuclear power plants generate around 85,400 GWh (20 %). But, the coal-based power generation accounts to only 15,800 GWh, which is 4 % of the total demand. Hence, the commercial energy sources meet nearly 68 % of the total demand. The remaining 32 % of demand is met by the renewable energy sources namely;

Figure 2. Optimal electricity distribution mix for the year 2020 in India.



wind, biomass gasifier, biogas plants, solid waste, cogeneration and mini hydel based power plants. Among the renewable energy based power generation, the biomass gasifier based power generation contributes to nearly 11 % of the total demand, which generates 45,520 GWh of electricity and 5 % (22,400GWh) of the total energy demand is met by the wind based power generation. The contribution of biogas plants, solid waste, cogeneration and mini hydel plants for the power generation is 3 % (14,112 GWh), 2 % (8,400 GWh), 8 % (33,600 GWh) and 3 % (11,970 GWh), respectively.

SCENARIO ON COST OF POWER GENERATION BY BIOGAS PLANTS

The effect on OEAM by varying the unit cost of power generation for the biogas plants is shown in Fig. 3. It was observed that there is no change in the power generation pattern either the cost of power generation for the biogas plants is decreased to any extent or increased up to 150% due to the contribution from the other plants. While the cost of power generation for the biogas plants increases to 175% and 200%, then the utilization of biogas based power generation decreases from 3% to 2.5% of total electricity demand.

Figure 3. Effect of variation of unit cost of biogas power generation on the electricity distribution mix for the year 2020.

It is found that the variation of cost of biogas-based plants doesn’t create much impact on the electricity distribution pattern. It also reveals that the cost of power generation through biogas-based plants is less in comparing with other power plants. It could be further promoted by providing subsidy by the Government.

CONCLUSION

The Optimal Electricity Allocation Model (OEAM) has been formulated for the electricity demand allocation for the year 2020. The model was developed with the objective of minimizing the unit cost of power generation subject to the constraints of demand, potential, efficiency, emission and carbon tax. The fuzzy constraints are used in the model. The extents of energy sources mix for the electricity generation would be 15,800 GWh (4 %) from the coal-based plants, 85,400 GWh (20 %) from the nuclear plants, 191,100 GWh (44 %) from the hydro plants, 22,400 GWh (5 %) from the wind turbines, 45,520 GWh (11 %) from the biomass gasifier plants, 14,112 GWh (3 %) from the biogas plants, 8,400 GWh (2 %) from the solid waste, 33,600 GWh (8 %) from the cogeneration plants and 11,970 GWh (3 %) from the mini hydel plants, respectively. A scenario is developed based on the unit cost of power generation by biogas plants and discussed. It is concluded that the OEAM model would be very helpful to policy makers for future electricity planning and to promote the usage of biogas plants in India. This model can be effectively implemented in countries like Malaysia for framing the energy policies and to promote the biogas plants for electricity generation, even though the source of power generation differs among the countries.

REFERENCES

[1] Executive Summary Report, Ministry of Power, Government of India, 2014.



[2] D.W.Bunn, and K.Vlahos, “A model-referenced procedure to support adversarial decision processes: Application to electricity planning”, Energy Economics, Vol.14, No.4, pp.242-247, October 1992.

[3] Amit Kumar, S.C.Bhattacharya, and H.L.Pham, “Greenhouse gas mitigation potential of biomass energy technologies in Vietnam using the long range energy alternative planning system model”, Energy, Vol.28, No.7, pp.627-654, June 2003.

[4] Chinmoy Jana and R.N.Chattopadhyay, “Block level energy planning for domestic lighting-a multi-objective fuzzy linear programming approach”, Energy, Vol.29, No.11, pp.1819-1829, September 2004.

[5] Rahul B. Hiremath, Bimlesh Kumar, P.Balachandra, and N.H.Ravindranath, “Bottom-up approach for decentralized energy planning: Case study of Tumkur district in India”, Energy Policy, Vol.38, No.2, pp. 862-874, February 2010.

[6] R.Banos, F.Manzano-Agugliaro, F.G.Montoya, C.Gil, A.Alcayde, and J.Gomez, “Optimization methods applied to renewable and sustainable energy: A review”, Renewable and Sustainable Energy Reviews, Vol.15, No.4, pp.1753-1766, May 2011.

[7] Benjamin K. Sovacool and L.C.Bulan, “Energy security and hydropower development in Malaysia: The drivers and challenges facing the Sarawak Corridor of Renewable Energy (SCORE)”, Renewable Energy, Vol.40, No.1, pp.113-129, April 2012.

[8] Pierluigi Mancarella, “MES (multi-energy systems): An overview of concepts and evaluation models”, Energy, Vol.65, pp.1-17, February 2014.


Jebaraj S. earned his B.E Degree in Mechanical Engineering and Masters Degree in Thermal Engineering from Bharathiyar University, Coimbatore. He started his teaching profession in the year 1996. He obtained his doctorate degree in Mechanical Engineering in 2007 from Anna University, India. He

has published 11 research papers in reputed international journals. Also, he has presented more than 35 papers in conferences in the area of energy modeling and planning. His papers received more than 725 citations. He has more than 17 years of experience in teaching in various universities. At present he is working in the Mechanical Engineering department in Universiti Teknologi PETRONAS, Malaysia. He is a reviewer for many reputed international journals. His area of research is energy planning, policy, energy modelling, renewable energy and heat transfer.



INTRODUCTION

An artificial neural network is a computer model that attempts to mimic simple biological learning processes and simulate specific functions of human nervous system. It is an adaptive, parallel information processing system which is able to develop associations, transformations or mappings between objects or data. It is also the most popular intelligent technique for pattern recognition to date. Although artificial neural networks (ANN) were introduced in the late fifties (Rosenblatt, 1962), the interests in them have been increasingly growing in recent years. This has been in part due to new applications fields in the academia and industry. Also, advances in computer technology (both hardware and software) have made it possible to develop ANN capable of tackling practically meaningful problems with a reasonable response time. Simply put, neural

networks are computer models that attempt to simulate specific functions of human nervous system. This is accomplished through some parallel structures comprised of non-linear processing nodes that are connected by fixed (Lippmann, 1987), variable (Barhen et al., 1989) or fuzzy (Gupta and Ding, 1994) weights. These weights establish a relationship between the inputs and output of each “Neuron” in the ANN. Usually ANN has several “hidden” layers each layer comprised of several neurons. If the feed-forward (FF) network (FF or concurrent networks are those with unidirectional data flow). The full technical details can be found in Bishop (Bishop, 1995). If the FF network is trained by back propagation (BP) algorithms, they are called BP. Other types of ANN are supervised (self organizing) and auto associative networks.

Monson and Pita (1997) applied neural networks to find relationships between 3D seismic attributes

APPLICATION OF ARTIFICIAL NEURAL NETWORKS TECHNIQUE FOR ESTIMATING PERMEABILITY FROM

WELL LOG DATA

Mohammed Abdalla Ayoub, Abdollah EsmaeiliDepartment of Mechanical Engineering,

Universiti Teknologi PETRONASEmail: [email protected], [email protected]

ABSTRACT

Permeability is one the most important hydrocarbon reservoir parameters. Permeability can be obtained at the laboratory from core samples by dry air injection or well testing method. These methods demand high cost and are time consuming. So, due to economic reasons and impossibility of coring in horizontal wells, core data would be available for limited number of wells. However, most wells have well log data. Artificial Neural Networks (ANN) can be useful to estimate permeability in case of there is a difficulty of obtaining represented core plugs from certain reservoirs. In this study, intelligent soft computing neural networks that nowadays are widely used in petroleum industry were used for permeability prediction in an oil field. In this study MATLAB software was used to handle neural networks core and well logs data, including permeability. These networks were developed using error-back propagation algorithm within feed-forward networks. After comparing the measured and network predicted results, the parameters of the Artificial Neural Network (ANN) were adjusted for a desired network. These results show that intelligent neural network models have predicted porosity and permeability successfully. Finally, the above mentioned networks was generalized to the third well which had no core data.



and well logs. The study provided realistic prediction of log responses far away from the wellbore. Boadu (1997) also used similar technology to relate seismic attributes to rock properties for sandstones. The use of neural networks in well logging has been popular for nearly one decade. Many successful applications have been documented (Wong et al., 1998; Brus et al., 1999; Wong et al., 2000). The most recent work by Bruce et al. (2000) presented a state-of-the-art review of the use of neural networks for predicting permeability from well logs. In this application, the network is used as a nonlinear regression tool to develop transformation between well logs and core permeability. Such a transformation can be used for estimating permeability in un-cored intervals and wells. In this work, the permeability profile was predicted by a Bayesian neural network. The network was trained by a training set with four well logs (GR, NPHI, RHOB and RT) and core permeability. The network also provided a measure of confidence (the standard deviation of a Gaussian function): the higher the standard deviation (“sigma”), the lower the prediction reliability. This is very useful for understanding the risk of data extrapolation. The same tool can be applied to estimate porosity and fluid saturations.

Porosity logs measurements require environmental corrections and are influenced by litho logy and formation fluids. The porosity derived is the total porosity, which consists of producible fluids, capillary bound fluids and clay-bound water. Permeability is a measure of fluid rock conductivity. To be permeable, a rock must have interconnected porosity. Greater porosity usually corresponds to greater permeability; however, this is not always the case. Formation permeability is influenced by pore size, shape and continuity, as well as the amount of porosity. Permeability can be determined from resistivity gradients, permeability models based on porosity and irreducible water saturation, formation tester and nuclear magnetic resonance. Perhaps, the most important feature of nuclear magnetic resonance logging is the ability to record a real-time permeability log. The potential benefits of nuclear magnetic resonance to oil companies are

enormous. Log permeability measurements enable production rates prediction and allow optimization of production completion and programs stimulation while decreasing the cost of coring and testing wells especially in heterogeneous tight reservoirs where there is considerable permeability anisotropy.

Permeability can not be measured by well logs directly. Using coring devices to take cores and permeability measurement in experimental conditions is one the most expensive and difficult methods to estimate formation permeability. The other method for measuring permeability is well testing. Because of some difficulties, high cost and long time, well testing and coring can not be done in total wells of a field but geophysical logging can be done in almost total wells during or after drilling. Therefore, permeability measurement by using obtained data from well logs and Artificial Neural Networks (ANN) is an economic and suitable method.

Generally, in this study, the wells which had well logging and coring data were been selected. Neutron, sonic, density, gamma ray and resistivity logs data were defined as input data and permeability of core was defined as output data. After three times of training and testing, we could use the modified network to well C of this field. First, we will discuss about the data preparation and correction methods, then, the relationship between neutron, sonic, gamma ray and resistivity logs data and finally permeability neutral network training and testing will be discussed.

PURPOSE OF STUDY

Our purposes in this study are: a) use of obtained data from neutron log, density log, sonic log, Gama ray log, resistivity log and core data to design an artificial neutral network to estimate and predict permeability of a reservoir b) determine the relationship coefficient between permeability which is obtained from dry air injection test of cores with permeability which is obtained from artificial neutral network c) determine the relationship coefficient between permeability which is obtained from helium injection test of cores



with permeability which is obtained from artificial neutral network.

METHODOLOGY

Using of artificial neutral network will solve most of our problems which we have in determination of reservoir parameters by coring. In this study, we use artificial neutral network to estimate porosity and permeability of an oil reservoir. To design an artificial neutral network, we need core and well log data. For this purpose, we selected three wells, well A, well B and well C of this oil field. These three wells have neutron log, density log, sonic log, Gama ray log, and resistivity log data. Wells A and B have core data but well C does not have core data.

First, we analyze the existing data of these wells to sort them and delete unsuitable data. Then, we compare them with core data and normalize them (between 0 and 1). Then, we design artificial neutral network. For this purpose, first, we select wells A and B which have core and well log data. After that, we insert well log data as input data and core data as output data into MATLAB software. After data processing and preparation, training of network has been done and its proper operation has been tested. Therefore, this network will become an intelligent network. Now, by inserting well log data of well C, we can get porosity and permeability from the output of this network.

To insert data in artificial neutral network, we must use figures, diagrams and curves in numerical form. For this purpose, we used Digitizer Software. Furthermore, some of the data which we have for a well may not be suitable to use. So these data must be deleted and not to be used in calculations. For example in some places inside a well, wash out may occur. These intervals must be known and their related data must be deleted. Well logging tools are not run at a uniform speed and time in a well. Therefore, well logs may differ a little from each other with respect to their depth. This difference in depth must be corrected. As data of neutron logs, density logs, resistivity logs, permeability logs and porosity logs are near to each

other, for better decision and networks operation, we can covert these data to logarithmic form. The best form of data for artificial neutral network is that the total data are between 0 and 1. One of the reasons for this data conversion is that transfer functions can only recognize very big amount. So to help artificial neutral network for better learning, we must normalize data. To normalize data, we insert input and output data to Excel software, then determine the minimum and maximum of each column and then convert them in a range between 0 and 1. After data normalization, we can save data in a separate sheet of Excel software. Then they will be sorted in forms of files with index "text". Final programing will be done by MATLAB software. Therefore the data will be processed and prepared to insert in artificial neutral network.

In this study, trial and error method was used to find the best and most suitable data via which network can predict the best answer. In other word, various sets of data were choused, then, they were tested in network one by one. We found that if we use sonic, density and gamma ray log data, an acceptable correlation coefficient between core permeability and predicted permeability by neutral network will be achieved. Therefore, sonic, density and gamma ray log data were selected as input data matrix and core permeability as output data matrix for this network.

Permeability Distribution in wells A and B

Formation permeability can be varied from 1 md to 10000 md (especially for fractured formations). When fractures occur, permeability will increase. The selected formation for this study is a fractured formation which its permeability can be up to 15000 md in some of its parts. Figure 1 and 2 show the permeability distribution for wells A and B in this formation.



Figure 1. Permeability distribution histogram versus depth in well A

Figure 2. Permeability distribution histogram versus depth in well B

Relationship between Network Input and Output Data

Grains size, grain shapes, sorting, cementing, mineral types, fractures and solubility are the mean factors in permeability measurements. The relationship between permeability and neutron log data, sonic log data, density log data, gamma ray log data and resistivity log data for well B of studied formation have been shown in figures 3 to 7 respectively.

Figure 3. Cross plot of core permeability versus neutron log data for well B



Figure 4. Cross plot of core permeability versus sonic log data for well B

Figure 5. Cross plot of core permeability versus density log data for well B

Figure 6. Cross plot of core permeability versus gamma ray log data for well B

Figure 7. Cross plot of core permeability versus Resistivity log data for well B

Permeability Estimation using Neural Network

The determination of permeability characteristics is labor intensive and complicated. Empirical models to predict relative permeability from rock and fluid properties have also experienced relatively limited success. Hence alternative methodologies for accurate determination of relative permeability characteristics have since been considered. Artificial neural network approach is proposed to predict an accurate permeability. It has been confirmed that multiple regression and neural network techniques perform better than empirical with neural network as the best tool.

In this study, artificial neutral network was used to build permeability neutral cell. This network was written in Work Space of MATLAB software using Feed Forward network and TrainLM. Figures 8 and 9 shows the cross plot of core permeability and predicted permeability by ANN for training and test steps respectively. This figures show a good matching of training and testing step data. Therefore the designed permeability neutral network is acceptable and can be used for well C which does not have core permeability data.



Figure 8. Cross plot of core and predicated permeability by ANN in training step for well B

Figure 9. Cross plot of core and predicated permeability by ANN in testing step for well B

CONCLUSIONS

1. Artificial neutral networks are very strong and their results are acceptable.

2. In this method, we do not need any primary assumption for data behavior, so is simple with respect to other methods.

3. Artificial neutral networks can be used for permeability prediction in wells which they do not have core samples.

4. Using artificial neutral networks, point data of wells can be related to the reservoir totally.

5. Types and sort of data are very important in artificial neutral networks, so they must be inserted in input matrix according to their importance.

6. Continuity coefficient shows the strength amount of artificial neutral networks. As this coefficient was high, we could get acceptable results and relate these results to well C which does not have core permeability.

7. As artificial neutral networks can be used for permeability estimation based on well log data, this method can be used for reservoir water saturation and other parameters estimations.

ACKNOWLEDGEMENT

The authors wish to express their sincere thanks and gratitude to the professors, lecturers and staff of petroleum engineering department of Universiti Teknologi PETRONAS (UTP) for supporting the present research and their helps and encouragements in connection with this study. We are grateful for all their assistance.

REFERENCES

[1] Al-Bazzaz,W.H et al, "Permeability modeling using Neural Network approach for complex Mauded-Burgan carbonate reservoir", Paper SPE 105337, Middle East Oil & Show and conference, Manama, Bahrain, 11-14 March, 2007.

[2] Elshafei, M. and Hamada, G.M., "Neural network identification of hydrocarbon potential of shaly sand reservoirs", Paper SPE 110959, SPEKSA Annual Conference, Dhahran, Saudi Arabia, 7-8 May, 2007.



[3] Galarza T., et al. "Pore-scale characterization and productivity analysis by integration of NMR and open hole logs-A verification study", Paper SPE 108068, Latin American and Caribbean Petroleum Engineering Conference, Buenos Aires, Argentina, 15-18 April, 2007.

[4] Lim, J.S. and Kim, J., "Reservoir porosity and permeability estimation from well log using fuzzy logic and neural network", paper SPE 88476, Asia Pacific Oil and Gas Conference, October 18-20, Perth, Australia, 2004.

[5] Riedmiller, M. and Braun, H., "A Direct Adaptive Method for Faster Back propagation: The RPROP algorithm", presented at Proceeding of the IEEE International Conference on Neural Networks, USA, 1993.

[6] Soto, B., Ferneynes, E.L. and Bejaramo, H., "Neural network to predict the Permeability and porosity of zone C of the Cantagalo field, Colombia", Paper SPE 38134, SPE Petroleum Computer Conference, Dallas, USA, 8-11 June, 1997.

[7] Al-Kharusi, A. S. and Blunt, M. J., "Network Extraction from Sandstone and Carbonate Pore Space Images", Journal of Petroleum Science and Engineering, 56: 219-231, 2007.

[8] Hollis, C., Vahrenkamp, V., Tull, S., Mookerjee, A., Taberner, C. and Huang, Y., "Pore System Characterisation in Heterogeneous Carbonates: An Alternative Approach to Widely-Used Rock-Typing Methodologies", Marine and Petroleum Geology, 27: 772-793, 2010.

[9] Reddy, G. V., Kumar, P., Naik, S. B. R., Murthy, G. N., Vandana, Kaul, S. K., and Bhowmick, P. K., "Reservoir Characterization Using Multi Seismic Attributes in B-172 Area of Heera-Panna-Bassein Block of Bombay Offshore Basin", India, 5th Conference & Exposition on Petroleum Geophysics, Hyderabad-2004, India, 716-721, 2004.

[10] Edalat, A., and Siyahkoohi, H., "Using Seismic Facies in Characterizing One of the Iranian Hydrocarbon Reservoirs", Iranian Geophysics Journal, 1(1): 37-49, 2007.

[11] Zhou, Y. J., Tao, J. Q., Guo, Y. B., Zhang, X. H., and Qiang, M., "Rock Physics Based Prestack Seismic Reservoir Characterization - Application to Thin Bed", 71st EAGE Conference & Exhibition – Amsterdam, Netherlands, 2009.

[12] Baechle, G. T., Colpaert, A., Eberli, G. P. and Weger, R. J., "Modeling Carbonate Pore Structure Effects on Velocity Using a Dual Porosity DEM Theory", SEG 76th Annual Meeting, San Antonio, Texas, 2007.


Mohammed Abdalla Ayoub is a lecturer at the Petroleum Engineering Department, Universiti Teknologi PETRONAS, Malaysia. Previously, he was an assistant professor at the Petroleum Engineering Department at University of Khartoum, Sudan. Dr Ayoub holds an MSc in Petroleum Engineering from

KFUPM, and a PhD from UTP, Malaysia. His main interests include hybrid EOR techniques, application of Artificial Intelligence tools into petroleum engineering related problems. He is a member of the Sudanese engineering council as well as SPE.

Abdollah Esmaeili received his Diploma in Mathematics & Physics in 1990. He obtained his B.Sc and MSc. in Petroleum Engineering from Petroleum University of Technology (PUT), Iran. He is currently pursuing his PhD in Petroleum Engineering at Universiti Teknologi PETRONAS (UTP), Malaysia.

He has been working as a reservoir engineer in the oil & gas industry for the past 20 years for companies like National Iranian Oil Company (N.I.O.C), National Iranian South Oil Company (N.I.S.O.C), Aghajari Oil & Gas Production Company (AjOGPC). He also taught petroleum reservoir engineering courses in universities of Iran. He has written several papers in petroleum engineering which were accepted for presentation in international conferences. He has attended several international conferences worldwideas a speaker. He has led several international scientific master classes and workshops worldwide.



CAPITAL STRUCTURE AND THE DETERMINANTS OF THE BANKING SECTOR: THE CASE OF MALAYSIA

Maran Marimuthu

Department of Management and HumanitiesUniversiti Teknologi PETRONAS

Email: [email protected]

ABSTRACT

This paper examines the determinants of capital structure of commercial banks in Malaysia. The determinants include asset structure, growth, size, risk, profitability and tax shield, liquidity and efficiency. All the nine local commercial banks are considered in this study over the period 1998 to 2007. Econometrics methodology incorporates both time series and cross-sectional data to perform regression analyses that include the pooled effect, fixed effect and random effect methods. The findings indicate that the explanatory power of the determinants tends to differ on total leverage, long term leverage and short term leverage. Almost all the determinants are correlated with the banks’ long-term leverage. It is evident that pecking order theory seems to be more applicable for the banking institutions.

Keywords—banks; capital structure; determinants

INTRODUCTION

Capital is a critical resource for all firms and capital structure accounts for corporate financing behavior [1]. Capital structure refers to the relative mix of debt and equity capital. Debt capital is an alternative source of funding to equity capital financing. Planning a firm’s capital structure is a major part and a continual process of the firm that needs to be undertaken in the promotion, growth and mature stage of the firm’s life cycle. [2] started the debate on whether capital structure decisions are important, and the consequent impact on the firm’s cost of capital. Since then, many theories and research have been developed and carried out to explain the way in which a firm chooses its capital structure.

Firms create wealth by making successful investment decisions which generate positive net cash flows. On the other hand, capital structure decision or financing decision determines the balance or relative amount of debt and equity to finance a firm [3]. The understanding of firm characteristics is also essentially

important to manage capital structure [4].

A. Overview of Local Commercial Banks in Malaysia

The current local anchor banks became nine in 2006 after Bumiputra-Commerce Bank Berhad merged with Southern Bank Berhad to form CIMB Bank Berhad. The nine anchor banks in Malaysia comprise of Affin Bank Berhad, Alliance Bank Malaysia Berhad, Ambank (M) Berhad, CIMB Bank Berhad, Eon Bank, Hong Leong Bank Berhad, Malayan Banking Berhad (Maybank), Public Bank Berhad and RHB Bank Berhad. Local commercial banks which are underlying on personal banking and business banking categories are governed by the Banking and Finance Institutions Act 1989.

RATIONALE OF THE STUDY

It is of great interest for many researchers to further explore the behaviourals or responses of banks with regard to their handling of debt levels. Unlike the



non-financial corporations, banks are expected to have different expectations in managing their debts as their major role is to connect the suppliers and demanders of funds in the financial markets. This empirical study is therefore able to offer more insights on the determinants of capital structure of banks. Specifically, signification of the determinants can be analysed in three dimensions, that include short term, long term and total debts.

OBJECTIVE OF THE STUDY

This study continues the effort of many researchers to further explore the capital structure issues in the financial sector. Hence, this study makes an attempt to examine the determinants of capital structure of local commercial banks in Malaysia.

LITERATURE REVIEW

Asset structure is argued to have a negative relationship with total leverage and short-term leverage, but it has the positive relationship with long-term leverage. Asset structure is the extent of the firm’s assets that comprise of tangible asset (fixed asset) that can cause firms to have a liquidation value [5]. [6], [7] indicated that asset structure has a significant positive relationship with leverage. Their results showed that tangible assets are more valuable to creditor and consistent with the greater value of tangible assets as collateral. Their findings support the argument put forth by [8], [9] that firms with greater amounts of collateral value in their assets, (which usually are fixed assets as opposed to short term assets), tend to issue more debt to take advantage of the low cost. The low cost is experienced mainly due to the lenders perceiving their investments to be of lower risk as there is the added protection from the collateralized assets. Meanwhile, growth has a positive relationship with total leverage and short-term leverage, but has the negative relationship with the long-term leverage. Applying the pecking order theory, the growing firms will place a greater demand on their internally generated funds. Consequently,

firms with higher greater growth will tend to look to external funds to finance the growth [5]. [10], [11], [6] found that growth had a negative relationship with leverage. Growth leads to higher agency costs, thus negative relationship is expected between growth and debt [12], [7]. These findings also consistent with other related studies by [13], [9], [14].

Many studies have suggested that there is a positive relationship between leverage and firm size. [15], [16], [17], [18], [19], [20] state that firm size should be positive impact on supply of debt and leverage. This is because larger firms are supposed to be less susceptible to bankruptcy as larger firms tend to be more diversified (business activities and income stream). Large firms hence are expected to experience lower volatility of earnings. In essence, the larger the firms, the higher the total and short-term leverage [5]. All forms of debt with the exception of long-term borrowings have a significant positive correlation with size. Larger and profitable firm with higher liquidity ratios might support a relatively higher debt ratio due to its greater ability to meet interest obligations [21].

[8], [9] pointed out the existence of a negative relationship between profitability and leverage. This is consistent with pecking order theory which argues that internal funding is preferable to external funding and that higher profitability will increase retained earning and reduces debt [6], [7], [11]. Banks have an incentive to employ more debt capital given that interest charges are tax deductible. Thus, successive tax increase would be associated with increasing debt capital. Therefore, his findings on total leverage and short-term leverage are consistent with the traditional capital structure theory on tax shield [5]. The increasing contingent claims does not fit the model of past due loans and show a negative relationship between tax shield and leverage [22]1.

1 Firm initially prefers to use available internal finance to finance new projects. If this source proves insufficient the firm may then amend its dividend policy to generate additional internal fund and finally firm may only resort to external financing if the source proves inadequate to firm.

[8] S. C. Myers, "The capital structure puzzle," The journal of finance, vol. 39, pp. 574-592, 1984.



Firms that have higher operating risk, tend to face a greater chance to default and thus, they are exposed to the agency and bankruptcy cost. Total and long-term leverage have a negative relationship with the level of risk while the short-term leverage has a positive relationship with the risk [21]. If banks have high non performing loans, therefore, they face higher risk, and thus, have smaller capacity to face fixed interest commitments. This then causes them to lower their borrowings. Total leverage and long-term leverage have a negative relationship with the liquidity while the short-term leverage has the negative relationship with the liquidity. Banks that have higher liquidity ratios might support a relatively higher debt ratio due to its greater ability to meet interest obligations [21].

All forms of debts have a negative and highly significant correlation with efficiency. This shows that efficient banks with high interest margins will rely less on outside debt financing but may require short-term debts [21]. [2] were the first to raise issue of capital structure irrelevance in a prefect market. According to Modigliani-Miller (MM) theory, the tax exemption of interest payments is an important tax shield for debt financing. Besides that, depreciation is another major source of tax shield. This theory suggests that there are some other taxes such as income tax that can have effect on a firm’s capital structure also. They argued that, under certain conditions, the choice between debt and equity does not affect firm value and hence, the decision is ‘irrelevant’. These conditions included assumptions about the absence of taxes, negligible transaction costs in the capital market, and information asymmetry between various market players [23]. MM model relied upon two basic assumptions which included investment and financing decision being independent and that the value of the firm is unaffected by the type or the types of capital employed in its capital structure. Thus, they took taxation under consideration and using the tax deductibility of interest argument, proposed that firms should employ as much debt as possible to achieve the optimal capital structure. Pecking order theory explains on corporate financing decisions. There is trade-off between equity and debt in finance industry and no optimum debt level for a firm. Firms

prefer internal sources better than external sources.1 In addition, higher return of capital tends to lead to greater exposure to the risk associated with the information asymmetries for the various financing choices. Thus, the firm will prefer retained earnings financing to debt, short-term debt over long-term debt and debt over equity. According to [8] managers have superior information than investors concerning the value of the firm under alternative investment strategies. In other word, they derive a pecking order theory of capital structure under asymmetric information. 2

Developing further on non-debt tax shield, [25] argue that there exist several forms of non-debt tax shields, for e.g. depreciation expenses or capital allowances and investment tax credits. This then reduces a firm’s capacity of debt tax benefit. Non-debt tax shields hence has a negative relationship on a firm’s optimal debt level, that is firms with large non-debt tax shields tend to have relatively less debt in their capital structure.

Companies can reduce leverage when they are mature and profitable but this is not conclusive. Financial success is correlated with the changes in the debts [26]. Signaling hypothesis showed that pecking order theory is quite relevant for financial managers and static trade-off theory is the need to balance gains and costs of debt financing as stated by [8]. In order to achieve optimal capital structure, static trade-off theory argues that firms will choose the equity and debt financing to balance the costs and benefits of debt. According to the static trade-off theory, firms are usually choosing their level of debt financing by trading off these bankruptcy costs and agency costs of debt against tax benefit of debt. In particular, a firm that is trying to maximize the value for its shareholders will equalize the marginal cost of debt that results from these bankruptcy costs with the marginal benefit of debt that results from tax benefits.

2 Information asymmetries exist before debt issuance and extend through repayment, creating adverse selection and moral hazard problems that raise the interest rates charged by lenders stated by [24] S. A. Johnson, "The effect of bank debt on optimal capital structure," Financial Management, pp. 47-56, 1998.



The tax benefits are created as the interest payments associated with debt are tax deductible while payment associated with equity such as dividends are appropriated from profit. The issuing of debts increase the financial distress and associate cost of firms. According to signaling theory, investors know the operation of firms and try to interpret the decision making of firms [27]. [28] explain how agency theory may be applied to determine an optimal capital structure of the firm which minimizes total agency costs of debt and equity. 3

METHODS AND PROCEDURES

This study is specially designed to examine the factors that determine the capital structure of the banking institutions in Malaysia. The factors tested in this paper are as follows; asset structure, growth, size, risk, profitability and tax shield, liquidity and efficiency. All the nine local banks in Malaysia: Maybank, Public Bank, Alliance Bank, Affin Bank, Ambank, CIMB Bank, Eon Bank, Hong Leong Bank and RHB Bank were taken into consideration and the data were collected over a period of ten years (1998-2007). The data were obtained from Bankscope and Osiris (financial databases). The entire population of local commercial banks were considered for the study.

A. Hypotheses

The following hypotheses are proposed;

Hypothesis 1: Total leverage is significantly impacted by the determinants of capital structure.

Hypothesis 2: Long-term leverage is significantly impacted by the determinants of capital structure.

Hypothesis 3: Short-term leverage is significantly impacted by the determinants of capital structure.

Specifically, a set of non-directional hypotheses involving each of the determinants of capital structure can constructed to predict total leverage, long term leverage and short term leverage. The hypotheses are as follows; H1: There is a significant relationship between asset structure and debt.

H2: There is a significant relationship between growth and debt.

H3: There is a significant relationship between size and debt.

H4: There is a significant relationship between profitability and debt.

H5: There is a significant relationship between non-debt tax shield and debt.

H6: There is a significant relationship between risk and debt.

H7: There is a significant relationship between liquidity and debt.

B. Econometrics Methodology

a) Panel Data Modeling: Panel data incorporate both time series and cross-sectional data. Panel data analysis allows identification of parameters without making any restrictive assumptions [29]. Panel data have space and as well as time dimensions [30]. Baltagi clarifies that when firms are considered over time, panel data tend to include heterogeneity; more informative data, more variability, less collinearity (among variables), more degrees of freedom, more efficiency; dynamics of change; larger sample size and thus, bias is minimized.

Let us say all variables have cross-sectional units

3 They noted three components of the agency costs of debt which are adverse incentive effects associated with highly levered firms, monitoring costs generated by theses incentive effects and bankruptcy cost. [28] M. C. Jensen and W. H. Meckling, Theory of the firm: Managerial behavior, agency costs, and ownership structure: Springer, 1979.



(referring companies- i.e. banks), thus, i = 1,2,3,…N and time period, thus, t = 1,2,3,…T. Therefore, the standard linear model is as follows;

yit = β0 + xit β + Ɛit ; xit

are the predictors and β0 and β represent intercept and slope coefficients are identical for all firms and time periods, Ɛit is the error term and yit is the dependent variable. In addition, panel data model assumes; Ɛit = ai + μit and μit denotes that homoskedasticity is assumed and not correlated over time ai is time variant and homoskedasticity is assumed across firms. The above model is also regarded as random effect model [29], [30].

In the case of fixed effect model, includes an individual firm-specific intercept term in the model as given below; yit = ai + xit β + μit ; yit is the regressand, ai (i = 1,2.3,…N) are fixed unknown constants that are estimated along with β and μit is assumed to be i.i.d over individuals and time. The overall intercept term β0 is dropped. In addition, Hausman Test and Redundant Fixed Effect Test were specially adopted in determining the most appropriate model (either pooled effect, fixed effect or random effect model) as presented in the findings section.

C. Measures

Three multiple regression models were adopted to examine the determinants of total leverage, long term leverage and short term leverage. Each dependent variable was tested using the above hypotheses using the econometrics methodology. Three regression models were considered representing three dependent variables; total leverage, long term leverage and short term leverage. As the analysis involved both cross-sectional and time series data, the adoption of the pooled, fixed and random effects were then applied for robust discussions. The details of the models and measures are as follows; General form of panel data: Yi,t = α + βXi,t + ëi,t ; i = cross-sectional dimension, t= time-series dimension, Yi,t= dependent variable, Xi,t= independent variable, ë= error term.

Models

TLi,t = β0 + β1ASi,t + β2GROWi,t + β3SIZEi,t + β4PROFi,t +

β5TAXSi,t + β6RISKi,t + β7LIQi,t + β8EFFi,t + ë .........Eq 1

LTLi,t = β0 + β1ASi,t + β2GROWi,t + β3SIZEi,t + β4PROFi,t +


STLi,t = β0 + β1ASi,t + β2GROWi,t + β3SIZEi,t + β4PROFi,t +


TLi,t (Total Leverage) = Total Debt / Total Asset,LTLi,t (Long-term Leverage) = Long–term debt / Total Asset,STLi,t (Short-term Leverage) = Short term debt / Total Asset,Asset Structure (AS) = Fixed Assets / Total Assets,Growth (GROW) = % change in total assets,Size (SIZE) = Logarithm of total assets,Profitability (PROF) = Operating Income / Total assets,Tax Shield (TAXS) = Depreciation & Amortization / Total Assets,Risk (RISK) = Standard deviation of profit before tax,Liquidity (LIQ) = Liquid Assets / Total Assets,Efficiency (EFF) = Net Interest Revenue / Total Assets[5], [31],[32], [12], [3], [33], [34], [35], [36], [37], [38], [24], [21].

FINDINGS AND DISCUSSIONS

TABLE I provides a summary of the descriptive statistics of the dependent variables (total leverage, short-term leverage and long-term leverage) and independent variables (asset structure, growth, size, profitability, tax shield, risk, liquidity and efficiency). The mean, median, maximum and minimum of each variable are shown in the table.



TABLE I: Descriptive Statistics For The Dependent Variables And Independent Variables

Variable Mean Median Minimum Maximum

Leverage 0.9210 0.9219 0.8562 0.9714

Short-term Leverage

0.8357 0.8321 0.7683 0.9117

Long-term Leverage

0.0852 0.0852 0.0266 0.1709

Asset Structure

0.0085 0.0074 0.0027 0.0209

Growth 0.1337 0.0831 -0.1515 2.1108

Size 4.6563 4.6348 3.9600 5.4094

Profitability 0.0092 0.0100 -0.0477 0.0279

Tax Shield 0.0002 0.0000 0.0000 0.0025

Risk 318.9815 239.5204 40.2537 1154.3066

Liquidity 0.2197 0.2049 0.1164 0.4224

Efficiency 0.0255 0.0256 0.0112 0.0380

Figure 1 shows the total, short term and long leverage of the local commercial banks in Malaysia over the period 1998 to 2007. Long term leverage of the banks seemed to vary from year to year. Though generally they showed an upward trend towards 2007, however, Affin, Ambank and CIMB registered a downward trend in their long term leverage.

Figure 1: Long Term Leverage of Local Commercial Banks in Malaysia

In terms of risk (changes in profit before tax), generally, banks managed to stabilize their earnings after the 1997 financial crisis (Figure 2). However, Maybank,

Public banks and CIMB registered a drastic increase in relation to their risk profile related to earnings as shown in Figure 2. Nonetheless, starting from year 2000, the banking industry started experiencing a more stable earnings. Efficiency measure (proxy of interest revenue of total asset) indicated that there was a downward trend in the banking industry till 2007 (Figure 3).

Figure 2: Risk Profile of Local Commercial Banks in Malaysia

Figure 3: Efficiency of Local Commercial Banks in Malaysia



TABLE II: Bivariate Correlation Matrix (Pooled Data)

Variables 1 2 3 4 5 6 7 8 9 10 11

1.Total leverage

1 0.2435* 0.3848** -0.4615** 0.1030 0.0139 -0.4865** -0.0453 0.1629 0.2106* -0.4066**

2.S-Term leverage

1 -0.8015** -0.3990** 0.0066 -0.1158 -0.0940 -0.1022 0.0397 0.5913** -0.1235

3.L-Term leverage

1 0.0951 0.0573 0.1188 -0.2106* 0.0693 0.0626 -0.4329** -0.1333

4.Asset structure

1 -0.0722 -0.1612 0.0172 -0.0407 -0.0583 -0.2667* 0.2773**

5.Growth 1 0.1401 0.0178 -0.0871 -0.0648 0.0721 -0.3026**

6.Size 1 0.3266** -0.0027 0.6675** -0.0174 -0.2234*

7.Profit 1 0.0493 0.1353 -0.0795 0.3718**

8.Tax shield 1 0.0375 -0.0107 -0.1681

9.Risk 1 0.0332 -0.0911

10.Liqiudity 1 -0.2691*

11.Efficiency 1

* p < 0.05, ** p < 0.01

TABLE II above shows bivariate correlation results involving all the variables considered for modeling. The significant correlation between total leverage, short-term leverage and long-term leverage (at 0.01 significance level) indicate consistency of the data series and measures used in the study. Asset structure

registered a significant negative correlation with total leverage and short-term leverage (-0.4615 and -0.3990 respectively) at 0.01. Significant correlations were also recorded among the determinants of capital structure at 0.05 and 0.01 significance level (in bold).

TABLE III: Regression Results On Pooled Effect, Fixed Effect And Random Effect

Variables Total Leverage (Model 1) Long-term Leverage (Model 2) Short-term Leverage (Model 3)

Pooled Fixed Random Pooled Fixed Random Pooled Fixed Random

Constant 0.9872*** 0.7898*** 0.9872*** 0.0776 -0.3893*** 0.0776* 0.9097*** 1.1791*** 0.9097***

AssetStructure

-1.9356***(-5.0119)

-0.9312(-1.5276)

-1.9356***(-5.9624)

0.28550.3841

3.7576***(4.2188)

0.2855(0.6019)

-2.2203***(-3.5936)

-4.6879***(-5.2349)

-2.2203***(-4.6557)

Growth 0.0058(0.8793)

0.0104*(1.7369)

0.0058(1.0461)

0.00580.4554

0.0165*(1.8938)

0.0058(0.7136)

1.68E-05(0.0016)

-0.0062(-0.7011)

1.68E-050.0021

Size -0.0085(-1.1942)

0.0277**(2.0930)

-0.0085(-1.4207)

0.01751.2736

0.0905***(4.6828)

0.0175**1.9956

-0.0260**(-2.2778)

-0.0628***(-3.2326)

-0.0260***-2.9509

Profitability -0.7772***(-4.2706)

-0.6780***(-3.9722)

-0.7772***(-5.0805)

-0.8210-2.3442

-0.4332*(-1.7270)

-0.8210***(-3.6731)

0.0438(0.1505)

-0.2540(-0.9771)

0.0438(0.1950)

Tax Shield -2.8248(-0.8249)

-0.3057(-0.0963)

-2.8248(-0.9814)

4.55670.6914

13.7939***(2.9738)

4.5565(1.0834)

-7.3812(-1.3474)

-14.1007***(-3.0236)

-7.3812*(-1.7456)

Risk 2.22E-05***(2.6440)

1.61E-05(1.1235)

2.22E-05***3.1454

-2.13E-06-0.1315

3.59E-05*(1.7136)

-2.13E-06(-0.2061)

2.43E-05*(1.8099)

-1.98E-05-0.9386

2.43E-05**(2.3448)

Liquidity 0.0042(0.1532)

0.0109(0.3655)

0.0042(0.1822)

-0.2478***-4.7158

-0.1561***(-3.5930)

-0.2478***(-7.3892)

0.2519***(5.7664)

0.1669***3.8199

0.2519***7.4707

Efficiency -0.4405(-1.1867)

0.3011(0.7154)

-0.4405(-1.4118)

-0.5856-0.8201

1.7054***(2.7736)

-0.5857(-1.2850)

0.1443(0.2430)

-1.4050*-2.2727

0.1443(0.3149)

Adjusted R2 (%) 45.97 61.82 45.97 23.91 69.01 23.91 41.92 65.39 41.92

Redundant FE/RE

40.6178*** 90.1950*** 55.9690***

Hausman Test 41.6368*** 125.8650*** 62.9568***

F-Test 10.4637*** 10.0068*** 10.4637*** 4.4960*** 13.3857*** 4.4960*** 9.0282*** 11.5117*** 9.0283***

* sig at 0.10, ** sig at 0.05, *** sig at 0.01, t-test results are given in parentheses



TABLE III shows the regression results of the three models (as given by Eq 1, 2 and 3) that examine the determinants of capital structure using the total leverage, long term leverage and short term leverage as dependant variables. Each model was derived based on the pooled effect (PE), fixed effect (FE) and random effect (RE) methods. The Hausman and Redundant FE/RE tests were employed to test the applicability of the FE and RE methods. Thus, the RE method is to be rejected in favour of the FE method at significance level 0.01 and this applies to all the three models as given in TABLE III.

Based on the results in TABLE III, long term borrowing is found to have a significant positive relationship with the banks’ tangible assets. This is consistent with the findings for non financial firms by [6], [7], indicates that leverage of firms is positively related with asset structure. Growth is seen to have a low level of significance in relation to the total leverage and long term leverage (at 0.10) of the banks. However, the negative relationship result appears to differ from the argument in the literature in the case of non-financial firms by [11], [12], [6]. Thus, agency cost is expected to be mitigated by the increase in total and long term debts.

Bank size is found to be positively correlated with long term borrowings but negatively correlated with short term borrowings. Understandably, larger banks are able to diversify their risk and thus, are less likely to face financial distress. Therefore, we expect a positive relationship between bank size and their debts. However, larger banks tend to reduce their short term leverage and this could be due to cost effectiveness of borrowing. Also, the negative correlation between profitability and total and long term leverage reinforces the adoption of pecking order theory by the banking institutions at significance level 0.10.

Meanwhile, the relevance of tax shield remains arguable for long term leverage as we expect to have a negative relationship between the two. Tax shield obtained from depreciation is expected to be a substitute to tax saving from borrowing, thus the larger the depreciation, the lower the debts will be. Such negative relationship was found between non debt tax shield and both total short term leverage. Perhaps, the service sector (banking) does not get

optimal trade-off through the use of depreciation as fixed tangible assets are not the main component of the total assets for value creation. Instead, banks derive their profits from financial instruments, in the forms of loans to customers (long and short term) and investments in financial products. As a result, this phenomenon drives banking institutions to strive for greater tax saving via long term borrowing.

We expect borrowing to be a decreasing function of risk (earnings volatility). It seemed that risk factor could push banks for higher long term leverage. This could be argued that as banks’ operations (size) increase, their risk diversification becomes more efficient and thus, they could go for higher debts. As for liquidity, it is very much emphasized for the short term leverage and conversely, greater liquidity would lower the long term leverage. Once again, this is in line with the matching principle as liquidity of banks refers to the short term assets and this should be matched by short term financing. Also, the banks assets comprise mainly of cash and near cash items (that is assets that must be easily converted to cash easily). Finally, banks’ efficiency is essentially important for their long term leverage. Adjusted R squared is in the range of 60 to 70 per cent (FE method) and the F-test is very significant for the models at 0.01.

Admittedly, the most influential factors on total leverage of banks are profitability, followed by size and growth. The other factors were noted to be insignificant based on the FE method. On the other hand, when total leverage is decomposed into short term and long term leverage, the outcome is different. In the case of short term leverage, the important factors influencing the capital structure of banks are asset structure, liquidity, size and tax shield. For long term leverage, all factors were noted to be significant with the most significant factor is firmsize.

CONCLUSIONS

The empirical findings as presented above conclude that hypothesis testing on long term and short term leverage is partially supported, but strongly supported on the long term leverage though the directions of the hypotheses have differed. Obviously, growth and non-debt tax shield are expected to be negatively



correlated with debt level, thus these findings seem to be contradictory. The summary of the results are given in TABLE IV below.

TABLE IV: Summary Of The Regression Results

DependentVariables

Total Leverage

Short-term Leverage

Long-term Leverage

Coefficient & Significant relationship

Positive correlated

and significant

relationship

Risk Risk Liquidity

Positive correlated

and insignificant relationship

Growth Liquidity

Growth Profitability

Efficiency

Asset structure

Growth Size Tax shield

Negative correlated

and significant

relationship

Asset Structure

Profitability

Asset structure

Size

Profitability Liquidity

Negative correlated

and insignificant relationship

Size Tax shield

EfficiencyTax shield Risk

Efficiency

It was noted that all the determinants of capital structure of the banks registered a significant impact on the long term leverage. These findings are essentially important in our discussion as the long term debt component is the magnifier of firm capital structure. Greater asset structure, tax shield and efficiency are regarded as the important requirements for increasing long term borrowing. Interestingly, greater liquidity helps banks to lower their long term leverage but however, it also appears as an important requirement for their short term leverage. It should be noted that banks are also pressured to hold sufficient fixed assets as collateral for their long

term borrowing as any other non-financial firms. Specifically, banks’ financing strategy can also be identified here. The negative correlation between profitability and borrowing is indicating that they still remain conservative where their internal funds are still the main source of financing (pecking order theory).

However, the banking institutions tend to show a different pattern of relationship with regard to tax shield and risk factor and their influence on borrowing. Clearly, tax shield from depreciation may not be a good substitute to tax saving from borrowing in the case of banking industry. Also, increasing risk resulting from volatility of earnings in the banking industry could be just a temporary effect for most banks and thus, higher borrowing will be anticipated for cost saving. Lastly, it should be noted that the most important factors affecting total leverage, long term leverage and short term leverage of Malaysian banks are profitability, size and asset structure.

LIMITATIONS

Some limitations must be addressed here, each variable is dictated by different measures as given in the literature and thus, may result in inconsistent results and interpretations. For instance, firm size can be measured by total asset, total sales, total market capitalization, etc). Different accounting policies and practices tend to result in different figures in the financial statements and therefore to have different implications. These differences however were not vividly shown or explained in the data set collected. In addition, inferential studies based on relatively smaller populations will be a great challenge for many researchers.

REFERENCES

[1] L. B. Stearns, "Capital market effects on external control of corporations," Theory and Society, vol. 15, pp. 47-75, 1986.

[2] F. Modigliani and M. H. Miller, "The cost of capital, corporation finance and the theory of investment," The American economic review, pp. 261-297, 1958.



[3] T. D. K. Nguyen and N. Ramachandran, "Capital structure in small and medium-sized enterprises: the case of Vietnam," ASEAN Economic bulletin, vol. 23, pp. 192-211, 2006.

[4] M. Marimuthu, "Corporate Restructuring, Firm Characteristics and Implications on Capital Structure: an Academic View," International Journal of Business and Management, vol. 4, p. P123, 2009.

[5] M. Amidu, "Determinants of capital structure of banks in Ghana: an empirical approach," Baltic Journal of Management, vol. 2, pp. 67-79, 2007.

[6] V. M. Gonzalez and F. González, "Influence of bank concentration and institutions on capital structure: New international evidence," Journal of Corporate Finance, vol. 14, pp. 363-375, 2008.

[7] A. Antoniou, Y. Guney, and K. Paudyal, "The determinants of capital structure: capital market-oriented versus bank-oriented institutions," Journal of financial and quantitative analysis, vol. 43, pp. 59-92, 2008.

[8] S. C. Myers, "The capital structure puzzle," The journal of finance, vol. 39, pp. 574-592, 1984.

[9] R. G. Rajan and L. Zingales, "What do we know about capital structure? Some evidence from international data," The journal of Finance, vol. 50, pp. 1421-1460, 1995.

[10] N. Toy, A. Stonehill, L. Remmers, R. Wright, and T. Beekhuisen, "A comparative international study of growth, profitability, and risk as determinants of corporate debt ratios in the manufacturing sector," Journal of Financial and Quantitative Analysis, vol. 9, pp. 875-886, 1974.

[11] I. G. Sharpe, "Determinants of capital structure of Australian trading banks," Asia Pacific Journal of Management, vol. 12, pp. 97-121, 1995.

[12] S. Titman and R. Wessels, "The determinants of capital structure choice," The Journal of finance, vol. 43, pp. 1-19, 1988.

[13] C. W. Smith Jr and R. L. Watts, "The investment opportunity set and corporate financing, dividend, and compensation policies," Journal of financial Economics, vol. 32, pp. 263-292, 1992.

[14] L. Lang, E. Ofek, and R. Stulz, "Leverage, investment, and firm growth," Journal of financial Economics, vol. 40, pp. 3-29, 1996.

[15] M. Narayanan, "Debt versus equity under asymmetric information," Journal of Financial and Quantitative Analysis, vol. 23, pp. 39-51, 1988.

[16] T. H. Noe, "Capital structure and signaling game equilibria," Review of Financial Studies, vol. 1, pp. 331-355, 1988.

[17] M. Poitevin, "Financial signalling and the" deep-pocket" argument," The Rand Journal of Economics, pp. 26-40, 1989.

[18] R. Stulz, "Managerial discretion and optimal financing policies," Journal of financial Economics, vol. 26, pp. 3-27, 1990.

[19] M. Harris and A. Raviv, "Capital structure and the informational role of debt," The Journal of Finance, vol. 45, pp. 321-349, 1990.

[20] M. Harris and A. Raviv, "The theory of capital structure," the Journal of Finance, vol. 46, pp. 297-355, 1991.

[21] M. Sen, "A decompositional analysis of the liability structure of Indian commercial banks," IIMB Management Review, pp. 25-33, 2004.

[22] D. Reagle, "Back on the balance sheet: The tax effects of contingent claims in commercial banking," Review of Financial Economics, vol. 15, pp. 19-27, 2006.

[23] F. Modigliani and M. H. Miller, "Corporate income taxes and the cost of capital: a correction," The American Economic Review, pp. 433-443, 1963.

[24] S. A. Johnson, "The effect of bank debt on optimal capital structure," Financial Management, pp. 47-56, 1998.

[25] H. DeAngelo and R. W. Masulis, "Optimal capital structure under corporate and personal taxation," Journal of financial Economics, vol. 8, pp. 3-29, 1980.

[26] T. Koller, M. Goedhart, and D. Wessels, Valuation: measuring and managing the value of companies vol. 499: John Wiley and Sons, 2010.

[27] S. B. Smart, L. J. Gitman, and W. L. Megginson, Corporate finance: Thomson South-Western, 2007.

[28] M. C. Jensen and W. H. Meckling, Theory of the firm: Managerial behavior, agency costs, and ownership structure: Springer, 1979.

[29] M. Verbeek, A guide to modern econometrics: John Wiley & Sons, 2008.

[30] D. N. Gujarati, "Basic Econometrics. 4th," ed: New York: McGraw-Hill, 2003.

[31] S. Akhtar, "The determinants of capital structure for Australian multinational and domestic corporations," Australian journal of management, vol. 30, pp. 321-341, 2005.



[32] M. Suto, "Capital structure and investment behaviour of Malaysian firms in the 1990s: a study of corporate governance before the crisis," Corporate Governance: An International Review, vol. 11, pp. 25-39, 2003.

[33] B. Fattouh, P. Scaramozzino, and L. Harris, "Capital structure in South Korea: a quantile regression approach," Journal of Development Economics, vol. 76, pp. 231-250, 2005.

[34] N. Delcoure, "The determinants of capital structure in transitional economies," International Review of Economics & Finance, vol. 16, pp. 400-415, 2007.

[35] V. S. Krishnan and R. C. Moyer, "Determinants of capital structure: an empirical analysis of firms in industrialized countries," Managerial Finance, vol. 22, pp. 39-55, 1996.

[36] M. J. McCue and Y. A. Ozcan, "Determinants of capital structure," Hospital & Health Services Administration, vol. 37, pp. 333-346, 1991.

[37] C.-H. H. Tang and S. S. Jang, "Revisit to the determinants of capital structure: a comparison between lodging firms and software firms," International Journal of Hospitality Management, vol. 26, pp. 175-187, 2007.

[38] G. Huang and F. M. Song, "The determinants of capital structure: evidence from China," China Economic Review, vol. 17, pp. 14-36, 2006.


Dr. Maran Marimuthu is a Senior Lecturer at the Department of Management and Humanities, Universiti Teknologi PETRONAS. Maran Marimuthu holds a doctorate in Finance from University of Malaya in 2011. He obtained his MBA from University of Malaya in 2000 and his

BBA (Hons) from Universiti Utara Malaysia in 1994. He teaches Financial Management, Business Finance, Corporate Finance and Research Methods in Finance. To date, he has published more than fifty international refereed journals and conference proceedings (local and overseas). He has co-authored books on Financial Management, International Finance and contributed articles for both local and international magazines. He has also embarked on consultancy work involving both academic and real life decision makings. Dr. Maran Marimuthu is currently holding internal and external research grants as Principal Researcher and Co-Researcher and supervising doctoral and masters students in the related area. He also has a special interest in the multidisciplinary work. Dr. Maran Marimuthu has corporate experience of four years, fifteen years of academia career.



A COMPARATIVE NUMERICAL SIMULATION STUDY OF POLYMER AND SURFACTANT ADSORPTIONS

ON SANDSTONE ROCKS

Mohamed Ali Hamid, Mustafa Onur

Department of Petroleum EngineeringUniversiti Teknologi PETRONAS

Email: Mohamed Ali <[email protected]>

ABSTRACT

Currently, a few commercial numerical simulators offer options to include the effects of adsorption during chemical enhanced oil recovery processes with consideration of the main parameters that can affect the adsorption (temperature, salinity, and formation permeability). The main objective of this work is to conduct a comparative numerical study for comprehensive understanding and identifying the distinguished features of the adsorption models available in commercial simulators. Our results show considerable differences in the surfactant and polymer concentration profiles predicted from the three simulators considered in this study. These differences are mainly due to different adsorption models considered by each simulator.

INTRODUCTION

The main objective of an enhanced oil recovery (EOR) method is to increase the volumetric (macroscopic) sweep efficiency and to enhance the displacement (microscopic) efficiency, as compared to the ordinary waterflooding [1]. One mechanism is aimed towards the increase in volumetric sweep by reducing the mobility ratio between the displacing and displaced fluids by injecting high viscosity polymers. Since the mobility of the injected fluid is reduced, the tendency to the fingering effect is much lowered in the use of polymers. The other mechanism targeted is to reduce the amount of oil trapped due to the capillary forces (microscopic entrapment). By reducing interfacial tension (IFT) between the displacing and displaced fluids, the effect of microscopic trapping is lowered, in turn which yields a lower residual oil saturation and hence higher ultimate recovery. Surfactants are used normally for reducing the IFT between the displacing and displaced fluids. So, the final recovery factor depends upon the microscopic displacement efficiency and on volumetric efficiency of the displacement front [2]. Often these mechanisms are

used in combinations to achieve the most efficient oil recovery process. Alkaline-surfactant-polymer (ASP) flooding can be an example of which shows the combination of the mechanisms indicated. In ASP flooding, the alkaline helps to protect the surfactant slug from divalent ions, reduce surfactant adsorption onto the rock and produce surface active components that lead to a further reduction in interfacial tension (IFT). The alkali converts naphthenic acids in the crude oil to soaps. The use of alkali also reduces adsorption of anionic surfactants on sandstones because the high pH reverses the charge of the positively charged clay sites where adsorption occurs [3].

During an ASP flooding, adsorption causes a loss of a chemical agent from solution, and consequently reduces the alkaline-surfactant-polymer slug concentrations. Therefore, the efficiency of ASP flooding can be affected not only in technical aspects but also in terms of economics [4]. Adsorption is a process by which molecules or ions are retained on the surfaces of solids by chemical or physical bonding [5]/ Hence, it is important to predict the effect of adsorption on any chemical EOR process that includes



chemicals that can be absorbed on the rock. The prediction of adsorbed amounts of chemicals is crucial for both design as well as economical considerations when planning a chemical EOR process.

There are a few commercial simulators commercially and publically available that include the effect of adsorption on a chemical EOR process, particularly on an ASP flooding. In this work, a comparative numerical study is conducted for comprehensive understanding and identifying the distinguished features of the adsorption models available in commercial simulators; namely CMG-STARS [8], ECLIPSE 100 and UTCHEM [10]. The results and findings of this work should prove useful for those who will conduct simulation studies using any of these simulators for predicting and assessing the effect of adsorption of polymers and surfactants during a chemical EOR process, in particular ASP flooding.

INFLUENCES OF ADSORPTION ON CHEMICAL FLOODING PERFORMANCE

The losses (due to adsorption) reduce the injected concentrations of the chemicals. The concentration is the key parameter in the success of an ASP process particularly and chemical EOR process generally [6]. The small concentrations typically used in ASP flooding reflect the influence of losing any part of these expensive chemicals by adsorption [7]. Currently a few commercial numerical simulators offer options to include the effects of adsorption during chemical enhanced oil recovery processes with consideration of the main parameters that can affect the adsorption (e.g. temperature, salinity, and formation permeability). Next, we provide general description of the simulators; CMG-STARS [8], ECLIPSE 100 [9], and UTCHEM [10], used in this work to study the effects of adsorption on the surfactant and polymer concentration profiles.

GENERAL DESCRIPTION OF THE SIMULATORS

Three simulators, named here as A, B, and C, were

employed to predict the surfactant and polymer concentration profiles. The simulator A represents CMG-STARS [8], B represents ECLIPSE 100 [9] and C represents UTCHEM [10]. The words A, B, and C are used instead of the simulators commercial names to ease the presentation of the results in graphs as well as in the text.

Differences in the treatment of adsorption

All three software consider effect of the adsorption of surfactant, polymer, and alkaline on the conservation equations. Each simulator has a set of conservation equations based on the type of the simulator and the assumptions made in building those equations.

The simulator A (CMG-STARS) allows a phenomenological description of adsorption, wherein a set of constant temperature adsorption isotherms (adsorption level as a function of fluid composition) are input for each adsorbed chemical additive. Different chemicals may adsorb simultaneously, each with their individual isotherms [8]. The isotherm is specified via Langmuir isotherm coefficients. For different temperatures, enter a table of the coefficients versus T. By inputting those adsorption tables, the concentration of each chemical additive in each cell is updated accordingly [8]. The simulator A considers the effect of temperature on adsorption of chemical agents, accounts for salinity effect by an empirical relationship between adsorption and salinity (wt %), accounts for the effect of adsorption through the residual resistance factors (RRF). RRF is a measure of permeability reduction induced by polymer adsorption [15]. Adsorption properties such as RRF, inaccessible pore volume and desorption level depend upon the formation permeability. Reservoir heterogeneities can cause these properties to vary significantly within a reservoir. Therefore, equilibrium adsorption is a function of location as well as component concentration and temperature.

The simulator B (ECLIPSE 100) assumes that alkaline, surfactant, and polymer exist only in the water phase, and the input to the reservoir is specified as a concentration at a water injector [9]. Moreover, the



adsorptions of those chemicals (alkaline, surfactant, and polymer) are assumed to start immediately after contacting with rock surface. The adsorption isotherms are specified by a generic analytical adsorption model using the ADSORP keyword with POLYMER, SURFACT, and ALKALINE as arguments. This model enables the adsorption to depend on the salinity and rock permeability. In the used version of ECLIPSE the adsorption concentrations are updated explicitly for the all chemical concentrations. The simulator B considers linear relationship between adsorption and salinity (wt %), and includes the effect of permeability by a power law relationship having a power value of 0.5.

The simulator C (UTCHEM) assumes that, the general tendency for the adsorption isotherm is to reach a plateau at some sufficiently large chemical concentration [10]. For pure surfactants, this concentration is in fact the CMC. Thus, adsorption is modeled with a Langmuir isotherm that takes into account both the surfactant/polymer and electrolyte concentrations. The simulator C assumes a linear relationship between adsorption and salinity (wt %), and includes the effect of permeability by a power law relationship having a power value of 0.6.

PARAMETRIC STUDY OF POLYMER/SURFACTANT ADSORPTION

A total of four scenarios (each has been designated as Runs 1 through 4, see Table II) have been prepared to investigate on difference treatments of adsorption using the simulators A, B, and C. The reservoir and fluid properties which are common in all four scenarios are given in Table I. Note that a slug of ASP mixture contains alkali, surfactant (anionic), and polymer (anionic) with concentration of 1.5 wt%, 1 wt%, and 2000 ppm, respectively as shown in Table I. A slug size of 4 PV followed by another 5 PV of water post flush having salinities of 3 wt% for Runs 3 and 4 as given in Table II was injected, while the softened water was used in the first two runs (see Runs 1 and 2 in table II). The temperature was kept as the same for all runs at 60 °C, except for Run no. 2 (see Table II). Permeability was

200 md for all runs, except for Run no. 4 for which the permeability was taken as 600 md to study the effect of permeability change on adsorption. Note that, the same adsorption isotherm coefficients were input to all simulators, except for simulator A, coefficients were given at two temperatures (60 and 80°C).

Table I. Fluid and rock properties used in four scenarios

Property ValueReservoir size (core), ft 1*0.125*0.125

Grid block size, ft 0.0125*0.125*0.125

Number of grids 80*1*1

Porosity 0.2

Phases presence Oil and water

Initial pressure, psi 1500

Initial water saturation 0.55

Water viscosity, cp 0.34

Oil viscosity, cp 2.00

Injection rate, ft3/day 500

Production pressure, psi 1000

Alkaline concentration, wt % 1.5

Surfactant concentration, wt % 1

Polymer concentration, ppm 2000

Table II. Varied parameters in four different runs (scenarios)

Run No.

Temperature, °C

Salinity, wt %

Permeability, md

1 60 0 200

2 80 0 200

3 60 3 200

4 60 3 600

RESULTS AND DISCUSSIONS

To make a comparison between adsorption behaviors in aforementioned simulators, tracking of concentrations at the production well (outlet) for surfactant and polymer have been done. All the concentrations were normalized by initial concentration (C0) of each chemical agent. Chemical



agents started adsorbing when the ASP solution entered the porous media for the first time. Therefore, adsorption delayed chemicals breakthrough. Therefore, the C/C0 (indicating the presence of the injected chemicals in the outlet) will have value of zero until the breakthrough occurs. After the breakthrough concentrations profile start to build up until equilibrium with rock surface is achieved (indicated by maximum value of C/C0 over many injected PV). The decline in the C/C0 values indicates desorption by post flush injection.

Run No. 1 was performed in reservoir having formation permeability of 200 md, 60 °C, no salinity and with distilled water for post flushes (no effect of salinity on concentrations). The simulation results obtained from three different simulators for the surfactant and polymer concentrations are shown in Figs. 1 and 2, respectively.

Fig. 1 Surfactant breakthrough in Run No. 1.

Fig. 2 Polymer breakthrough in Run No.1.

Fig. 1 and 2 show that all three simulators produced same trend as well as identical values of the concentrations in absence of salinity effects.

Run No. 2 was conducted to see how these software handle temperature effects (Table II). The normalized surfactant and polymer concentrations computed are shown in Figs. 3 and 4, respectively. In Figs. 3 and 4 as well as in all the following figures, the concentration profile for Run No. 1 has included a reference profile (as square data points) for comparison purposes.


Fig. 4 Polymer breakthrough in Run No. 2.

As can be seen from the results shown in Figs. 3 and 4, the simulator A has responded to the temperature effects with higher concentrations at given value of PV injected while the simulators B and C produced the same concentration profiles as Run No. 1 (see square data points) which is for 60 °C. The concentration profiles predicted by the simulators B and C are expected because they do not account for the temperature effect unless static isotherms determined



from the experiments at different temperature are input to these two simulators. As noted previously the simulator A considers the effect of temperature on adsorption of chemical agents, while the simulators B and C do not include temperature dependency in adsorption calculations. In general, adsorption decreases at high temperatures [11], [12].

To investigate the influence of the salinity on adsorption Run No. 3 was performed using salinity of 3% (sea water salinity) for post flush. The normalized surfactant and polymer concentrations computed from the three simulators are presented in Figs. 5 and 6, respectively.


Fig. 6 Polymer breakthrough in Run No. 3.

Figs. 5 and 6 show that the simulator A treats adsorption in a different way than do the simulators B and C. The dependence of adsorption on salinity was modeled in the simulators B and C by linear relationships, while an empirical model (nonlinear) is used in the simulator A. Typically, adsorption is higher

at high salinities [13], [14].

Run No. 4 was conducted to study the effect of permeability on adsorption. Formation permeability of 600 md was used while all other parameters were similar to Run No. 3. The results of Run No. 4 are shown in Fig. 7 for surfactant normalized concentration and Fig. 8 for polymer normalized concentration.


Fig. 8 Polymer breakthrough in Run No. 4

The results presented in Figs. 7 and 8 show that each simulator treats adsorption in a different way. The simulator A accounts for the effect of adsorption through the residual resistance factors (RRF), this is accounted for by scaling the adsorption obtained from local concentration conditions by the factor (ADMAX(I) / ADmax), where ADMAX(I) is the maximum adsorption capacity at grid block I, and ADmax, is the maximum possible adsorption obtainable from the adsorption isotherm of the first input. On the other hand, the simulator B includes the effect of



permeability by a power law relationship having a power value of 0.6 which is reasonable for high permeability formation [9]. The simulator C includes the effect of permeability by a power law relationship having a power value of 0.5.

SUMMARY AND CONCLUSIONS

This paper presented a comparison study of adsorption treatment using three simulators A, B, and C. The behavior of adsorption was studied by tracking of the concentrations at the outlet. The effects of salinity, formation permeability, and temperature on adsorption predicted by the three simulators were compared. Based on this study the following specific conclusions can be stated:

• All three simulators predicted the same concentration profile as a function of pore volume injected if the formation water had no salinity.

• The simulator A can account for the effect of reservoir temperature on adsorption, whereas the simulators B and C cannot.

• All the simulators showed dependency of adsorption on formation permeability, but the simulators A predicts slightly higher adsorption than do the simulators B and C.

ACKNOWLEDGMENT

The authors acknowledge with thanks the financial supports of Enhanced Oil Recovery Research Center (EORC), Universiti Teknologi PETRONAS for the support of this study.

REFERENCES

[1] M. A. Bataweel, D. S. Schechter, and H. A. Nasr-El-Din, “Fluid Flow Characterization of Chemical EOR Flooding : A Computerized Tomography (CT) Scan Study,” SPE 149066, no. 1990, pp. 1–15, 2011.

[2] J. Hou, Dalian, Z. Liu, X. Yue, S. Zhang, and J. Yang, “Restudy of Main Factors of Effects on ASP Flood and Expansion of Capillary Theory in Heterogeneous Reservoirs,” paper SPE 77800, SPE Asia Pacific Oil and Gas Conference and Exhibition held in Melbourne, Australia, 8–10 October 2002.

[3] J. J. Sheng, “A Comprehensive Review of Alkaline-Surfactant-Polymer (ASP) Flooding,” paper SPE 165358, SPE Western Regional & AAPG Pacific Section Meeting, Monterey, California, USA, 19−25 April 2013.

[4] J. Wang, S. Yuan, P. Shen, T. Zhong, and X. Jia, “Understanding of Fluid Flow Mechanism in Porous Media of EOR by ASP Flooding From Physical Modelling,” IPTC 11257, International Petroleum Technology Conference held in Dubai, U.A.E., 4–6 December 2007

[5] C. T. Q. Dang, Z. Chen, N. T. B. Nguyen, W. Bae, and T. H. Phung, “Development of Isotherm Polymer/Surfactant Adsorption Models in Chemical Flooding,” paper SPE 147872, SPE Asia Pacific Oil and Gas Conference and Exhibition held in Jakarta, Indonesia, 20–22 September 2011.

[6] A. Flaaten, Q. Nguyen, G. Pope, and J. Zhang, “A Systematic Laboratory Approach to Low-Cost, High-Performance Chemical Flooding,” paper SPE 113469 Proceedings of SPE/DOE Symposium on Improved Oil Recovery, Tulsa, Oklahoma, U.S.A., 19–23 April 2008.

[7] J. J. Taber, F. D. Martin, and R. S. Seright, “EOR Screening Criteria Revisited- Part 2: Applications and Impact of Oil Prices,” paper SPE 39234, SPE/DOE Improved Oil Recovery Symposium held in Tulsa, Oklahoma, 21–24 April, pp. 189–198, 1996.

[8] Computer Modelling Group Ltd., Advanced Process and Thermal Reservoir Simulator. Technical description, Version 2011.

[9] Schlumberger, ECLIPSE Reservoir Simulation Software.Technical description, Version 2011.2, 2011.



[10] Reservoir Engineering Research Program, Utchem-9.0, vol. II., The University of Texas at Austin, 2000.

[11] Nursel Pekel and Olgun Guven, “Solvent, Temperature and Concentration Effects on the Adsorption of Poly (n-Butyl Methacrylate) on Alumina from Solutions,” Turk J Chem, vol. 26, pp. 221–227, 2002.

[12] V. Ziegler and L. Handy, “Effect of Temperature on Surfactant Adsorption in Porous Media,” paper SPE 8264, SPE 54th Annual Technical Conference and Exhibition, held in Las Vegas, Sept. 23•26, 1979.

[13] M. Puerto, G. J. Hirasaki, C. A. Miller, J. R. Barnes, and S. G. Solutions, “Surfactant Systems for EOR in High-Temperature , High-salinity Environments,” paper SPE 129675, SPE Improved Oil Recovery Symposium held in Tulsa, Oklahoma, USA, 24–28 April 2010.

[14] M. Amro, “Investigation of Polymer Adsorption on Rock Surface of High Saline Reservoirs,” paper SPE 120807, Proceedings of SPE Saudi Arabia Section Technical Symposium, Alkhobar, May 2008.

[15] Lake, L. Enhanced Oil Recovery, Prentice Hall, Englewood Cliffs, New Jersey, pages 327-329, 1989.


Mohamed Ali Hamid was born in Sudan in 1984. He received the B.Sc. degree in chemical engineering from University of Khartoum, Khartoum, Sudan, in 2006. He joined Universiti Teknologi PETRONAS, Tronoh, Malaysia, in 2008 as a researcher where he received his M.Sc. (2010) and PhD (2013)

in Petroleum Engineering. His main areas of research interest are adsorption, characterization, and mathematical modeling of chemicals during EOR process. Mr. Mohamed is a member of the Sudanese Engineering Council, Sudan; the Society of Petroleum Engineers (SPE), and the Chemical Engineering Society at University Of Khartoum, Sudan.

Mustafa Onur was holding the Schlumberger Professorial Chair at the Department of Petroleum Engineering at Universiti Teknologi PETRONAS (UTP), and a professor of the Department of Petroleum and Natural Gas Engineering at Istanbul Technical University, Turkey. Previously, he was

the Head of the EOR Centre at UTP. His expertise and research interests include well testing, pressure-transient analysis, formation testing, reservoir engineering, enhanced oil recovery methods, nonlinear parameter estimation, and geothermal reservoir engineering. He holds a BS degree from the Middle East Technical U. in Turkey and MS and PhD degrees from the U. of Tulsa, all in petroleum engineering. He has served on the editorial committees of SPE Reservoir Evaluation & Engineering and SPE Journal. He was also the recipient of the 2010 SPE Formation Evaluation Award. Onur is currently the Executive Editor of the SPE Journal.



Assoc. Prof. Dr. Azrai B Abdullah

Assoc. Prof. Dr. Chow Weng Sum

Assoc. Prof. Dr. Mohd Yussoff B Ibrahim

Assoc. Prof. Dr. Morteza Khalaji Assadi

Assoc. Prof. Dr. Srinivasa Rao Pedapati

Dr. Abdul Haque Tunio

Dr. Lai Fong Woon

Dr .Khaled Abdalla Elraies

Dr. Muhammad Ayoub

Dr. Mohammed Idrees Ali

Dr. Mohd Suhaili B Ismail

Dr. Rashidah Bt Mohd Pilus

Dr. Sia Chee Wee

Dr. Syahrir Ridha

Dr. Ting Ding Hooi

Mr Mohammad Amin Shoushtari

Mr Berihun Mamo Nagesh

Ms Noor IIyana Ismail

REVIEWERS

What to sendSend article/manuscript in Microsoft Word (any version) via e-mail, or on a media (CD or flash drive), plus one PDF copy.

Send high resolution illustrations/images/tables/graphs with jpeg extension separately.

FormatIEEE format should be adopted in preparing the article/manuscript. Do check the available IEEE format and templates at IEEE – Manuscript Templates for Conference Proceedings.

The Manuscript1 The text should be preceded by a short abstract of 50 -100 words and a minimum of four keywords.2 The manuscript must be typed on one side of the paper, double-spaced throughout with wide margins not exceeding 8 pages (including references).3 Figures and tables have to be labelled and should be included in the text. If authors are using databases, tables , diagrams, etc from internet and print materials, permission must be sought first from the rights holder. Please attach the permission(s) with the manuscript.4 Footnotes should be kept to a minimum and be as brief as possible. The footnotes must be numbered consecutively.5 Special care should be given to the preparation of the drawings for the figures and diagrams. Except for a reduction in size, they will appear in the final printing in exactly the same form as submitted by the author.6 When submitting the article/manuscript, authors are also required to submit photographs and biography of themselves including academic background, work experiences and research interests.

Submission/Acceptance Procedures1 Articles/Manuscripts are evaluated by an editorial review board (internal and external). 2 Following initial selection, authors whose papers have been accepted for publication will be notified by email. 3 Only original and unpublished articles may be submitted.

Submit Articles/Manuscripts to: PLATFORM Editor-in-Chief Universiti Teknologi PETRONAS 32610 Bandar Seri Iskandar Perak Darul Ridzuan MALAYSIA Email: [email protected] [email protected]

ABOUT PLATFORM

PLATFORM is a biannual, peer-reviewed journal of Universiti Teknologi PETRONAS. It serves as a medium for faculty members, students and industry professionals to share their knowledge, views, experiences and discoveries in their areas of interest and expertise. It comprises collections of unpublished and original work by the academic staff and students of the university as well as industry professionals.

Guideline for Submitting a Paper to PLATFORMThese guidelines are designed to simplify and help standardize submissions. Length varies, but an average of 3,500 to 5,000 words, plus illustrations/images and captions. Maximum length should be about 8,000 words.

P L A T F O R M

VOLUME TEN NUMBER ONE JANUARY- JUNE 2014 ISSN 1511-6794

A Journal of Engineering, Science and Society

PL

AT

FO

RM

VO

LU

ME

TE

N N

UM

BE

R O

NE

JAN

UA

RY

- JUN

E 2

01

4

Volume 10 Number 1 Jan - Jun 2014

Multiple Hydraulic Fractured Vertical Wells in Gas Condensate ReservoirsMohammed Abdalla Ayoub, Abdollah Esmaeili 2

A Study on Gas Flaring Activities in NigeriaAliyu A. Sulaimon, Modupe Ogunmekan 12

Gender Diversity on Boards and Market Performance: An Empirical Investigation on Malaysian Listed CompaniesRohail Hassan, Maran Marimuthu 17

Fuzzy Linear Model For Electricity Planning In India With Focus On Cost Of Biogas PowerJebaraj S 27

Application of Artificial Neural Networks Technique for Estimating Permeability from Well Log DataMohammed Abdalla Ayoub, Abdollah Esmaeili 31

Capital Structure and the Determinants of the Banking Sector: The Case of MalaysiaMaran Marimuthu 38

A Comparative Numerical Simulation Study of Polymer and Surfactant Adsorptions on Sandstone RocksMohamed Ali Hamid, Mustafa Onur 49

Reviewers 56

platform - press.utp.edu.mypress.utp.edu.my/wp-content/uploads/2018/12/vol-10-no-1_2014_new.pdf ·...

Documents