Predicting-rock-mechanical-properties-of-carbonates-from-wireline-logs-A-case-study-Arab-D-reservoir

Predicting rock mechanical properties of carbonates from wireline logs (A case study:Arab-D reservoir,Ghawar field,Saudi Arabia)Mohammed S.Ameen a ,*,Brian G.D.Smart b ,J.Mc.Somerville b ,Sally Hammilton b ,Nassir A.Naji aaStructural Geology and Rock Mechanics Group,Geological Technical Services Division,Exploration Technical Services Department,Saudi Aramco,P.O.Box 2817,Dhahran 31311,Saudi Arabia bInstitute of Petroleum Engineering,Heriot-Watt University,Riccarton Campus,Edinburgh EH144AS,Scotland,UKa r t i c l e i n f oArticle history:Received 15October 2007Received in revised form 21January 2009Accepted 23January 2009Available online 31January 2009Keywords:Carbonate reservoirs Arab-DRock mechanical logs Geomechanics Ghawar field P-wave velocity S-wave velocity Elastic modulia b s t r a c tFour hundred plug samples from the Arab-D carbonate reservoir,Ghawar field,were tested for acoustic and mechanical properties at increasing triaxial stress.The results show that the rock mechanical parameters are primarily functions of porosity and,to a lesser degree,of mineralogy,texture and pore fabric (in order of degree of impact from higher to lower).The rock mechanical parameters of the intact matrix rock show no significant changes with stress under reservoir condition.The study enabled the generation of general and layer-specific correlation formulae of porosity with P-wave velocity,S-wave velocity,static and dynamic constants and the angle of internal friction.The formulae were then applied to a key well in the Ghawar using the compressional wave slowness from a Multipole Array Acoustic Log (MAC Ô,Baker Atlas)to derive rock mechanical pseudo-logs on reservoir-scale (referred to here as general)and on individual reservoir zone scale (referred to here as layer-specific).Comparison made between the general and the layer-specific pseudo-logs shows good agreement for each of the elastic constants with matching peaks and troughs throughout the logs.In addition the laboratory derived rock mechanical constants show a good agreement with the pseudo-logs.Where there is some difference between the general and layer-specific pseudo-logs,the layer-specific log shows better correlation with the laboratory derived constants.It can be concluded that the porosity correlation is an accurate,representative and cost effective method of obtaining a rock mechanical profile of the Arab-D reservoir.The derived formulae have been implemented as predictive tools in reservoir development and management (e.g.hydrofracturing and underbalanced drilling)and new prospect evaluation.The rock mechanical layering scheme shows higher resolution in the prolific part of the Arab-D reservoir than the Saudi Aramco conventionally used zonation (Each of Zones 2B and 3A consists of two rock mechanical layers).Furthermore the least prolific zones (lower part of 3A,the whole of 3B and 4)form one rock mechanical layer.Ó2009Elsevier Ltd.All rights reserved.1.IntroductionPredictive tools for rock mechanical parameters are essential for reservoir development,management,and prospect evaluation in exploration areas with very sparse or no borehole-based rock mechanical data.The need for such predicative methods is partic-ularly critical in carbonate reservoirs which are not as well understood or studied as clastic reservoirs.The most direct way of determining the rock mechanical data is from laboratory tests on plugs.A core-based test for the whole reservoir interval in each well is expensive,and requires extensive coring.Furthermore,directplug measurements cannot provide a continuous strength estimate as the plugs are taken from discrete points,every few feet,over a small section of the well in question.Therefore,there is a need to develop a quick and cost effective approach for rock mechanical characterization.As rock mechanical properties cannot be determined directly from logging tools,an indirect method must be introduced.Such method correlates the widely available V p (P-wave velocity)log,with the laboratory derived rock mechanical parameters,from representative core samples to produce a set of pseudo-logs.Due to the lack of theoretical models,most of the attempts to estimate rock mechanical properties are based on empirical correlations.In this study a combination of laboratory tests and empirical corre-lations of the results with porosity are used to establish rock mechanical pseudo-logs for the Arab-D.*Corresponding author.Tel.:þ96638745267.E-mail addresses:mohammed.ameen@ ,ameenms@(M.S.Ameen).Contents lists available at ScienceDirectMarine and Petroleum Geologyjournal homepage :www.else /locate /marpetgeo0264-8172/$–see front matter Ó2009Elsevier Ltd.All rights reserved.doi:10.1016/j.marpetgeo.2009.01.017Marine and Petroleum Geology 26(2009)430–444The rock mechanical and other rock parameters discussed in this paper include:static Young’s modulus(E s);dynamic Young’s modulus(E d);static bulk modulus(K s);dynamic bulk modulus (K d);static shear modulus(G s);dynamic shear modulus(G d);static Poisson’s ratio(n s);dynamic Poisson’s ratio(n d);uniaxial compressive strength(s UCS);P-wave velocity(V p);S-wave velocity (V s);bulk density(r);porosity(V),and angle of internal friction (Q ).These will be referred to in the remainder of the text and figures using the symbols indicated.2.The study areaThe present study focuses on the Ghawarfield,the largest oil field in the world,which measures over250km in length and 25km in width.It is located in the Eastern Province of Saudi Arabia (Fig.1A).All of the oil production comes from the Arab-D reservoir.The Arab-D reservoir(Upper Jurassic)is an approximately200–300ft thick platform carbonate sequence resting conformably on the Jubaila Formation(Upper Jurassic)and is overlain by the Arab-D Anhydrite Member.The reservoir consists of various carbonate rock types that exhibit an overall downward decrease in porosity.Based on the porosity log characteristics,the Arab-D reservoir is divided into six zones(Fig.1B and C)by Saudi Aramco(Alsharhan and Whittle,1995;Lucia et al.,2001;Meyer et al.,1996;Cantrell and Hagerty,1999;Cantrell et al.,2001;Saner and Sahin,1999;Swart et al.,2005).3.ObjectivesAlthough carbonate reservoirs hold a significant proportion of the known hydrocarbon reserves,little work has been done to understand their rock mechanical parameters and their predict-ability.Previous studies focused largely on clastic reservoirs.The Arab-D reservoir is an ideal choice to bridge the gap in our knowledge.It is the most important and prolific oil reservoir in the world extending across Eastern Saudi Arabia in Ghawar and in other giantfields.In addition to the universal need for a better rock mechanical understanding of carbonate reservoirs,there is an operational need for such an understanding in the Arab-D reservoir.The Ghawarfield was developed by peripheral water injection.There have been continuous efforts in the following areas:optimizing well planning, drilling and completion design,minimizing formation damage and improving productivity and injectivity indices.As some parts of the field matures such issues are becoming increasingly critical Beyond the Ghawarfield,the Arab-D sequences are exploration targets in previously unexplored parts of Saudi Arabia.Prospect evaluation and delineation of discoveries are primarily based on seismic data and wirleline logs of wildcat wells.Therefore predic-tive tools that link sonic wave velocities to petrophysical and rock mechanical properties are essential for seismic data calibration,and their implementation in exploration and development.4.Previous workEmpirical correlation has been used to establish rock mechan-ical logs since the1950s.Wyllie et al.(1956,1958,1963)introduced the use of the acoustic velocity for porosity determination with the ‘‘time-average’’equation,which empirically relates acoustic transit time or velocity to porosity.Similar correlations were discussed by Gardner et al.(1974),and Raymer et al.(1980).Tixier et al.(1975), derived mechanical properties logs based on a correlation of the in-situ strength with the dynamic elastic constants computed from sonic and density logs.The Schlumberger Mechpro method (Anderson et al.,1986;Bruce,1990)went a step further by trying to predict the uniaxial compressive strength from the sonic log.It is based on the correlation between the static Young’s modulus and the uniaxial compressive strength with correction factor for the shale content.Santarelli et al.(1991),suggested that rock strength can some-times be virtually independent of the sonic velocity,particularly in high porosity intervals.This implies that a method that relies on the traditional sonic log method would not produce satisfactory results in general.Strength estimates from sonic logs have never come close to the rock mechanical tests performed in the laboratory.An alternative approach for estimating rock mechanical properties would be to use the porosity as the primary parameter.Sarda et al. (1993),presented a direct correlation between the porosity(V)and uniaxial compressive strength(s UCS):s UCS¼258eÀ9FFarquhar et al.(1994)derived a geomechanical index for carbonates that enabled rock mechanical properties to be esti-mated using general andfield specific correlations.Edlmann et al. (1998)used laboratory measured porosity and rock mechanical parameters for North Sea reservoirs to establish direct correlations between the porosity and the rock mechanical parameters and to produce continuous rock mechanical logs.Bastos et al.(1998),established relationship between compressional and shear wave velocity and petrophysical proper-ties for an offshore Brazilianfield using laboratory tests on120 limestone samples.Widarsono et al.(2001)presented a new approach for the estimation of the elastic properties of clastic rocks in boreholes with limited log suites.5.MethodologyWe acquired four hundred core plugs from the Arab-D reservoir infive key wells,covering different parts of the Ghawarfield(Wells A,B,C,D and E in Fig.1A).The sampling rationale honors progressive changes in rock type and porosity across the six reservoir zones(Fig.1B and C).Right cylindrical plugs were cut with a diameter of37.5mm and trimmed to the required length for triaxial tests.The plugs were tested using a triaxial multistage testing technique for a range of acoustic and rock mechanical properties.A general empirical correlation between the laboratory measured V p and ambient porosity was obtained to generate porosity pseudo-logs from V p logs.Logs of static and dynamic elastic constants were generated from the porosity correlations using simultaneous laboratory measurement of static and dynamic elastic constants.The strength parameters were also determined in the laboratory and correlated with porosity.The data set from each suite of tests was then analyzed considering porosity distribution across the Arab-D to derive layer-specific correlations.The corre-lations were then applied in a key well,Ghawarfield,using the compressional wave slowness from a Multipole Array Acoustic log (MACÔ,Baker Atlas),and comparisons were made between the general pseudo-logs and the layer-specific pseudo-logs.The validity of the testing methods and the repeatability of the results were assessed from a second set of plug samples that were not used in the multistage tests and the derivation of the rock mechanical pseudo-logs.6.Derivation of porosity and rock mechanical parameters6.1.Sample preparation and ambient testsThe plugs were cleaned of salts and crude oil using the hot Soxhlets reflux technique.The cleaned samples were driedM.S.Ameen et al./Marine and Petroleum Geology26(2009)430–444431Fig.1.A.Location map of the Ghawar field and the studied key wells (oil fields:green;gas fields:red).B.Generalized Stratigraphy of the Upper Jurassic (left)and the Arab-D reservoir (center)from Cantrell et al.(2004).C.A typical porosity and flowmeter log,Arab-D,Ghawar field (right)and the rock mechanical layering established in this study.Note that rock mechanical layering scheme shows higher resolution in the main part of the reservoir than the Saudi Aramco conventionally used zonation (Zone 2B consists of two rock mechanical layers (layers 3and 4)and Zone 3A consists of two rock mechanical units (layer 5,and part of layer 6)).Furthermore the least prolific zones (lower part of 3A,the whole of 3B and 4)form one rock mechanical layer.M.S.Ameen et al./Marine and Petroleum Geology 26(2009)430–444432overnight and their dimensions and weight were measured. Porosity was determined using a Boyle’s law helium porosimeter and permeability was measured using a nitrogen gas permeameter. To determine the static elastic constants electrical strain gauges were bonded to the outside of the plugs.The gauges were arranged in a rosette of two gauges aligned at90 to each other.The rosette was bonded to the surface of the plug at the mid-point of its length using epoxy resin.Diametrically opposed,a second rosette was bonded to the surface.The electrical connections of the gauges were soldered to plastic coated electrical foil,and the whole assembly coated in epoxy resin to protect the gauges during testing. The plugs were then100%saturated in Multipar,a light mineral oil, in preparation for the triaxial tests.6.2.Triaxial testsThe static and dynamic constants and strength parameters were determined over a range of effective stresses(Table1)which represent the estimated stresses experienced in-situ by the reser-voir.At thefirst stress level,the static elastic constants and P-and S-wave velocities were determined,and the stress increased to the next stress level.The measurements were repeated at each stress level until the highest stress was reached.The stress was then reduced back to the lowest stress level,and the strength parame-ters determined on steps(at each stress level)as the stress was incrementally increased to failure.6.3.Static elastic constantsStatic elastic moduli were measured using a standard Hoek cell mounted into a stiff testing machine where confining stresses (Table1)were applied hydrostatically.The strain gauges monitored the axial and radial displacement of the plugs.The axial stress (measured load divided by the cross sectional area of the plug)was plotted versus the axial and radial strain.The static Young’s modulus(E s)was determined from the tangent to the stress–strain slope.The static Poisson’s ratio,(n s)was derived from the ratio of radial to axial strain.The static bulk modulus(K s)and shear modulus(G s)were determined from the modulus of elasticity and Poisson’s ratio as follows:K s¼E s=ð3ð1À2n sÞÞG s¼E s=ð2ð1þn sÞÞ6.4.Dynamic elastic constants testsAn acoustic transmitter and receiver were placed on either side of the plug within the Hoek cell and the generated wave was recorded.The P-wave and S-wave travel times,through the plug, were picked from the recorded wave and used to calculate V p and V s.The dynamic elastic constants were determined from V p,V s and r as follows:E d¼2Âð1þn dÞr V2sV d¼0:5ÂÀV p=V sÁ2À2. ÀV p=V sÁ2À1K d¼rV2pÀ4=3V2sG d¼r V2s6.5.Multistage testingThe failure criteria for broken or intact rock are represented by a2nd order polynomial equation.However,Wilson(1980)has shown the error in assuming a linear relationship to be negligible.A non-linear failure criterion introduces complexities that make the application difficult.The failure criterion,the triaxial stress factor, and the unconfined compressive strength are determined from laboratory tests on broken or intact rock.Adequate derivation of the failure envelope and triaxial stress factor ideally requires discrete tests on a set of identical samples from the formation of interest. However rocks are inherently heterogeneous,and therefore it is very difficult to retrieve identical samples.In addition retrieving sufficient number of samples for discrete tests is usually impracti-cable due to the limited core available from reservoirs.Therefore the multistage testing technique is used.This technique enables measurements of rock mechanical parameters at successively increasing confining pressures,on the same sample(Smart et al., 1991).In the current work a series of multistage tests were performed on the samples to determine the failure criteria describing the development of rock strength with increasing confining pressure.A Mohr-Coulomb failure criterion and the angle of internal friction Q were determined from the data.7.Correlation of rock mechanical properties7.1.Correlation of V p versus V s and static versus dynamic rock mechanical propertiesThe cross plot of the laboratory measured V p and V s(Fig.2) shows a good correlation between these measured velocities.The correlation formula is essential to enable the calculation of V s pseudo-log from in-situ,V p logs.In addition well logs giveus Table1at 27.6 (MPa)y = 0.52x + 252.51R2 = 0.8710001500200025003000350040004500500020003000400050006000700080009000V p (m/s)Vs(m/s)Fig.2.Cross correlation chart and formula of V s and V p measured under reservoirpressure.M.S.Ameen et al./Marine and Petroleum Geology26(2009)430–444433a.V p and V s decrease with increasing porosity with best fit curves as simple linear least square regressions,or negative expo-nential curves (Fig.4A).b.The static and dynamic moduli (E s,K s ,G s,E d,K d,and G s ),decrease with increasing porosity,with negative exponential best fit curves (Fig.4B–D).c.The angle of internal friction,Q decreases with increasing porosity with a best fit curve as a simple linear least squares regression (Fig.4E).7.3.Correlations of rock mechanical properties with reservoir-scale porosity (layer-specific formulae)The Arab-D reservoir layers are primarily based on the porosity profile (Fig.1B and C).Therefore the test results are analyzed to assess the rock mechanical layering,and the repeatability of the general correlation criteria when derived from data in individuallayers.The results confirm the validity of general trends in indi-vidual layers.In addition,the layer-specific correlations show better correlation coefficients than those of the general correlations apart from the static elastic constants of layers 5and 6.In these two layers which are considerably tighter than the other layers,there is a sharp decline in the correlation coefficient of the static elastic constants.Table 2gives a summary of the mechanical layers established in this study versus the Arab-D zones,general and layer-specific correlations and their respective correlation coefficients.8.Derivation of the rock mechanical properties from wireline logsThe above correlations were used to derive the rock mechanical properties from wireline log porosity.This is done in two stages:at 27.6 (MPa)0100020003000400050006000700080009000Porosity (%)W a v e V e l o c i t y (m /s )at 27.6 (MPa)020406080100120140160Porosity (%)Y o u n g 's M o d u l u s (G P a )at 27.6 (MPa)01020304050605101520253035Porosity (%)S h e a r M o d u l u s (G P a )020406080100120Porosity (%)B u l k M o d u l u s (G P a )010********Ambient Porosity (%)A n g l e o f I n t e r n a l F r i c t i o n (D e g r e e s )Fig.4.Correlation charts,formulae and correlation coefficients of rock mechanical parameters versus ambient matrix porosity:A.V p and V s ;B.Static and dynamic Young’s Modulus;C.Static and dynamic shear modulus;D.Static and dynamic bulk modulus;E.Angle of internal friction in limestone,dolomite,and all rock types .M.S.Ameen et al./Marine and Petroleum Geology 26(2009)430–4444358.1.Calibration of wireline log porosity with the core-derived porosityPrior to the derivation of the rock mechanical pseudo-logs from ‘‘in-situ’’wireline porosity logs the latter were calibrated against core-derived ambient porosity logs,as follows:a.The general empirical correlation formula between the labo-ratory measured V p and the ambient core porosity was applied to the wirleline V p log to calculate core-based,ambient porosity log.b.The calculated core-based,ambient porosity log is plotted with the in-situ wireline porosity log.c.Individual,laboratory measured porosities were added to the plot as a calibration,to assess the core to wireline log depth shift,and the level of agreement of the individual laboratory measured porosities with the ambient core porosity log.Fig.5shows an example of the calibration for Well C,in which a depth shift of 5ft was applied.There is a good agreement between the core-based,calculated porosity log and the in-situ wireline porosity log,and individual laboratory measured porosity.The apparently higher core-based porosity than the in-situ wireline log porosity is due to wireline logs obtained under down-hole stresses whereas the core-based,calculated porosity log is based on the laboratory measurements under ambient condition.8.2.Implementation of the correlation formulae to derive general and layer-specific pseudo-logs of rock mechanical parameters Pseudo-logs of the static and dynamic elastic constants,and angle of internal friction were derived using the general and the layer-specific correlation formulae between porosity and rock mechanical parameters.The resulting logs were plotted against depth,with the laboratory measured parameters added for cali-bration (Figs.6–8).The results show that:a.There is a good agreement between the general and layer-specific pseudo-logs for each of the constants with matching peaks and troughs throughout the logs.b.Where there is a difference,between the general and the layer-specific logs (mostly minor differences in magnitude),the layer-specific log correlates better with the laboratory derived values than the general logs.Porosity (%)Wireline Log Porosity (%)Core-Based Porosity LogAmbient Porosity (Core) before Depth Shift0 ftR e l a t i v e D e p t h f t150 ftFig.5.Wireline porosity log and the correlated ambient core-based porosity log in well C.Individual,laboratory measured ambient porosities (from plug samples)are shown at their un-shifted and shifteddepths.Static Shear Modulus (GPa)0 ft150 ftR e l a t i v e D e p t h f tStatic Bulk Modulus (GPa)Static Young's Modulus (GPa)general layerlabgeneral layerlabgeneral layerlab 0 ft150 ftR e l a t i v e D e p t h f tFig.6.Pseudo-logs of general and layer-specific static elastic constants for Well C:A.Young’s modulus;B.Bulk modulus;and C.Shear modulus.The mechanical layers 1–6are indicated.M.S.Ameen et al./Marine and Petroleum Geology 26(2009)430–4444369.The geological impact on the derived rock mechanical properties,and the correlation formulaeThe cross correlation of the rock mechanical properties and matrix porosity show that rock samples with the same porosity have a range of values of rock mechanical properties.The range invalues of elastic moduli,V p and V s becomes progressively accen-tuated as porosity diminishes to approach zero (Fig.4A–E)with a widening gap between an upper and lower bound of the rock mechanical parameters.This indicates that as porosity declines,other rock properties (in addition to porosity)are impacting the rock mechanical parameters.Such an argument is alsosupportedDynamic Young's Modulus (GPa)GeneralLayer LabGeneral Layer LabGeneral Layer Lab0 ft150 ftR e l a t i v e D e p t h f t0 ft150 ftR e l a t i v e D e p t h ftDynamic Bulk Modulus (GPa)Dynamic Shear Modulus (GPa)0 ft150 ftR e l a t i v e D e p t h f tFig.7.Pseudo-logs of general and layer-specific dynamic elastic constants for Well C:A.Young’s modulus;B.Bulk modulus;and C.Shear modulus.The mechanical layers 1–6areindicated.Angle of Internal Friction (degrees)general layer lab0 ft150 ftR e l a t i v e D e p t h f tFig.8.Pseudo-logs of general and layer-specific angle of internal friction,for Well C.The mechanical layers 1–6areindicated.Table 3Comparison of the median values of the rock mechanical parameters in the dolomiteTable 4Correction formulate and correlation confficient (R 2)of rock machanical parameters with matrix porosity in the limestone and dolomite samples (Correlation charts are M.S.Ameen et al./Marine and Petroleum Geology 26(2009)430–4444372000300040005000600070008000Ambient Porosity (%)P Wa v e V el o c i t y(m /s)1500200025003000350040004500Ambient Porosity (%)S W a v e V e l o c i t y (m /s )102030405060708090Ambient Porosity (%)S t a t i c Y o u n g 's M o d u l u s (M P a )0.0010.0020.0030.0040.0050.0060.0070.00Ambient Porosity (%)S t a t i c B u l k M o d u l u s (G P a )0.0010.0020.0030.0040.00Ambient Porosity (%)S t a t i c S h e a r M o d u l u s (G P a )102030405060708090100Ambient Porosity (%)D y n a m i c Y o u n g 's M o u d u l u s (M P a )0.0010.0020.0030.0040.0050.0060.0070.00Ambient Porosity (%)D y n a m i c B u l k M o d u l u s (G P a )0.0010.0020.0030.0040.00Ambient Porosity (%)D y n a m i c S h e a r M o d u l u s (G P a )0.0010.0020.0030.0040.0050.00Ambient Porosity (%)A n g l e o f I n t e r n a l F r i c t i o n (D e g r e e s )Fig.9.Reservoir-scale correlation charts of rock mechanical parameters versus ambient matrix porosity for limestone and dolomite:(A)V p ;(B)V s ;(C)Static Young’s Modulus (E s );(D)Static bulk modulus (K s );(E)Static shear modulus (G s );(F)Dynamic Young’s modulus (E d );(G)Dynamic bulk modulus (K d );(H)Dynamic shear modulus (G s );(I)Angle of internal friction (Q ).Correlation formulae and correlation coefficients (R)for these charts are given in Table 4.by the apparent difference in the degree of agreement between the general and layer-specific derived properties of the reservoir in certain layers,particularly the tighter layers (yers 5and 6,Figs.6–8).Therefore the geological impact (particularly mineralogy,rock texture,and pore type)on the rock mechanical properties is investigated.9.1.Rock mineralogyThere are two main mineral constituents of the studied samples,dolomite and calcite,occurring in various proportions and result in two main rock types,dolomite and limestone.The median values of the rock mechanical parameters for dolomite have higher values than limestone (Table 3).The relationship of the porosity and the rock mechanical parameters are analyzed considering rock type.The best fit curves show that on average dolomite has higher values of rock mechanical parameters than limestone for rocks with the same porosity (Table 4and Fig.9).9.2.Rock textureThe Dunham classification (Dunham,1962)is used to charac-terize the texture of the studied samples.The median values of the elastic moduli,V p and V s (Table 5)have maximum values in the crystalline rocks followed by packstones (mud-lean packstone for dynamic bulk modulus).Static bulk modulus differs in having the highest median value in the grainstones and mud-lean packstones followed by packstones and crystalline rocks.The rock texture of the samples was also considered in the analysis of the relationship of the porosity and rock mechanical properties.The correlation formulae are summarized in Table 6.The best fit curves of V p and V s have highest value in crystalline rocks followed by mud-lean packstones,grainstones and packstones respectively.Thedifferent static elastic moduli are impacted by texture in different ways.However the best fit curves of E d and G d have highest values in crystalline rocks and lower values in grainstones,mud-lean pack-stones and packstones respectively.The best fit curve of K d has highest value in crystalline rocks and mud-lean packstone and lower values in grainstones and packstones respectively.9.3.Rock pore typeThe median values of all the rock mechanical parameters have maximum values in rocks with dominant intercrystalline pore fabric (Table 7).The difference in the magnitude of the rock mechanical parameters between rock samples with different types of non-intercrystalline pore fabric (e.g.intergranular,interparticle,moldic,and vuggy)is considerably smaller than the contrast between rocks with intercrystalline pore fabric and those with other pore types.Samples with micropore fabric have higher static moduli than those with moldic,intergranular,interparticle and vuggy pore fabrics (Table 7).Samples with moldic pore fabric have higher dynamic moduli than those with vuggy,micropore,inter-granular and interparticle pore fabric.The trend for V p amongst the non-intercrystalline pore fabric is different from that of the V s (Table 7).The analysis of the pore fabric influence on the correlation between the elastic moduli,V p and V S with rock porosity shows no clear consistent trends.10.Verification of the resultsWe assessed the repeatability of the results and the impact of the application of multistage triaxial tests,to derive the rock mechanical parameters,in stead of the discrete failure tests.For this purpose we acquired a second set of samples beside that used to derive the rock mechanical parameters.The second set of samples was taken from Well-A and Well-D close to the plugs from the first set and cover representative range of porosities.10.1.Verification of the multistage testing techniqueWe subjected the additional set of samples to discrete tests,which involved loading the samples hydrostatically to its test stress and allowing the axial load to increase until the sample failed.The results from the discrete tests were processed in the same manner as the multistage tests i.e.Mohr-Coulomb analysis.The comparison of the results with those acquired from the multistage tests indi-cates an adequate repeatability evident from a very good correla-tion between the discrete and multistage tests,with the triaxial stress factors being within 6%(Fig.10).This verifies multistage testing as a viable technique where there is a lack of homogeneous samples from the samedepth.Table 5Comparison of the median values of the rock mechanical parameters in the testedTable 6M.S.Ameen et al./Marine and Petroleum Geology 26(2009)430–444439。