Louisiana State University LSU Digital ommons LSU Doctoral Dissertations Graduate School 2002 Log-derived cation exchange capacity of shaly sands : application to hydrocarbon detection and drilling optimization Gamze Ipek Louisiana State University and Agricultural and Mechanical ollege, gipek@lsu.edu Follo this and additional orks at: http://digitalcommons.lsu.edu/gradschool_dissertations Part of the Petroleum Engineering ommons Recommended itation Ipek, Gamze, "Log-derived cation exchange capacity of shaly sands : application to hydrocarbon detection and drilling optimization" (2002). LSU Doctoral Dissertations. 2727. http://digitalcommons.lsu.edu/gradschool_dissertations/2727 This Dissertation is brought to you for free and open access by the Graduate School at LSU Digital ommons. It has been accepted for inclusion in LSU Doctoral Dissertations by an authorized graduate school editor of LSU Digital ommons. For more information, please contact gcoste1@lsu.edu.
LOG-DERIVED ATION EXHANGE APAITY OF SHALY SANDS: APPLIATION TO HYDROARBON DETETION AND DRILLING OPTIMIZATION A Dissertation Submitted to the Graduate Faculty of the Louisiana State University and Agricultural and Mechanical ollege in partial fulfillment of the requirements for the degree of Doctor of Philosophy in The Department of Petroleum Engineering by Gamze Ipek M. S., Louisiana State University, 1999 B. S., Technical University of Istanbul, 1996 May, 2002
AKNOWLEDGEMENTS The author ould like to express her most sincere gratitude to Dr. Zaki Bassiouni, hairman of the raft and Hakins Department of Petroleum Engineering, for his insightful guidance, and genuine interest in this project. The author also ishes to thank Dr. John R. Smith and Dr. Robert Doner for their suggestions and assistance. Special thanks go to Dr. Andre Wojtanoicz, Dr. Julius Langlinais and Dr. Inone Masamichi. Finally, the author thanks the Baker Hughes, Inteq, BP-Amaco and the Petroleum Engineering Department for the financial support, hich made this study possible. The author is deeply indebted to her husband, Guncel Demircan, and her parents; M. Nihat Ipek, Gulseren Ipek, Dr. Omer Demircan and Gonul Demircan for their support. ii
TABLE OF ONTENTS AKNOWLEDGEMENTS... ii LIST OF TABLES... v LIST OF FIGURES... vi ABSTRAT... ix HAPTER 1. INTRODUTION... 1 Page 2. REVIEW OF PETROPYHSIAL MODELS FOR SHALY SANDS... 4 2.1 General oncept of Resistivity Models... 4 2.1.1 General oncept of lean Formation Model... 4 2.1.2 General oncept of Shaly Sand Formation Models... 5 2.2 V sh Shaly Sand Models... 7 2.3 Log Derived lay Volume Indicators... 8 2.4 ation Exchange apacity, E, Models... 9 2.4.1 Waxman and Smits Shaly Sand Model... 12 2.4.2 Dual Water Shaly Sand Model... 15 3.EARLY LSU SHALY SAND MODELS... 19 3.1 Silva-Bassiouni Shaly Sand Model... 19 3.1.1 S-B onductivity Model... 19 3.1.2 S-B Membrane Potential Model... 22 3.2 The LSU (Lau-Bassiouni) Shaly Sand Model... 24 3.2.1 The LSU (Lau-Bassiouni) onductivity Model... 24 3.2.2 The LSU (Lau-Bassiouni) Spontaneous Potential Model... 25 3.3 Advantages and Shortcomings of Early LSU Models... 27 4. NEW LSU SHALY SAND MODEL... 29 4.1 Ne LSU onductivity Model... 29 4.2 Ne LSU SP Model... 31 4.3 Estimation of m f m c... 34 4.4 Estimation of m eff... 35 4.5 Examples of Estimation of m eff m f, m c... 36 4.6 Advantages of the Modified LSU Model... 39 5. VALIDATION OF LSU MODELS... 40 5.1 Measured ation Exchange apacity Data... 40 5.1.1 MI 622 #6... 40 5.1.2 Baker Experimental Test Area (BETA)... 41 5.2. Statistical Validation Method... 43 5.3 Validation of the Ne LSU Shaly Sand Model... 45 5.3.1 MI 622 #6... 45 5.3.2 Baker Experimental Test Area (BETA)... 47 iii
5.4 Validation of the Perfect Shale Model... 49 5.4.1 MI 622 #6... 50 5.4.2 Baker Experimental Test Area (BETA)... 51 5.5 Validation of the Early LSU Shaly Sand Model... 54 5.5.1 MI 622 #6... 54 5.5.2 Baker Experimental Test Area (BETA)... 56 5.6 omparison of the Models... 58 6. APPLIATION TO HYDROARBON DETETION... 61 6.1 Well A... 61 6.2 Well B... 66 6.3 Well... 66 7. APPLIATION TO DRILLING OPTIMIZATION... 73 7.1 Introduction... 73 7.2 Literature Revie on Lo Penetration Rate... 74 7.2.1 haracteristic Symptoms of Lo Penetration Rate... 74 7.2.2 The Possible auses of Lo Penetration Rate... 75 7.2.3 Field Experiences and Researches of Lo Penetration Rate... 77 7.2.4 Previous LSU Researches on Poor PD Bit Performance in Deep Shales... 79 7.3 Validation of oncepts... 83 7.3.1 Normalized Rate of Penetration... 84 7.3.2 Specific Energy... 89 7.3.3 Force Ratio... 93 7.3.4 Depth of ut... 95 7.4 Detection of Pending Bit Balling... 95 7.4.1 MI 622 # 6 Bit Run # 8... 97 7.4.2 MI 636 # 1 Bit Run # 13... 102 7.5 onclusion... 105 8. ONLUSIONS AND REOMMENDATIONS... 111 REFERENES... 113 APPENDIX A: MEASURED AND ALULATED E FOR BH-1... 118 APPENDIX B: EQUATIONS REQUIRED IN NEW LSU SHALY SAND MODEL... 119 APPENDIX : SAS OUTPUT OF REGRESSION ANALYSIS FOR MI 622 #6 BIT RUN # 8... 122 APPENDIX D: SAS OUTPUT OF REGRESSION ANALYSIS FOR VALIDATION OF MODELS... 151 VITA... 157 iv
LIST OF TABLES Page 2.1 Log-derived clay volume indicators... 8 2.2 Shaly sand models for hydrocarbon-bearing formations... 11 5.1 Measured E corresponding to depth for MI 622 # 6... 41 5.2 Measured E corresponding to depth for BH-1... 42 5.3 Paired measured and log-derived ne LSU shaly sand model Es ith the differences for MI 622 #6... 46 5.4 Paired measured and log-derived ne LSU shaly sand model Es ith the differences for BH-1... 48 5.5 Paired measured and log-derived perfect shale model Es ith the differences for MI622#6... 50 5.6 Paired measured and log-derived perfect shale Es ith the differences for BH-1... 53 5.7 The log-derived E from early shaly sand model versus measured E ith the differences for MI 622#6... 54 5.8 Paired measured and log-derived early shaly sand model Es ith the differences for BH-1... 57 5.9 P-values and t-test of testing null hypothesis µ = 0 for MI 622 #6... 58 5.10 Regression analysis results ith zero intercept of models for MI 622 #6... 59 5.11 P-values and t-test of testing null hypothesis µ = 0 for BH-1... 59 5.12 Regression analysis results ith zero intercept of models for BH-1... 59 7.1 haracteristics of PD Bit Performance in Shale, MI 623 Field... 81 7.2 Measure E corresponding to depth... 85 d d v
LIST OF FIGURES Page 2.1 Typical conductivity o - plot for shaly sands... 6 4.1 o vs. porosity of clean sand for MI622#6... 37 4.2 sh vs. porosity of shale for MI622#6... 37 4.3 o vs. porosity of clean sand for BH-1... 38 4.4 sh vs. porosity of shale for BH-1... 38 5.1 Log-derived ne shaly sand model E vs. measured E for MI622#6... 47 5.2 Log-derived ne shaly sand model E vs. measured E from Na concentration for ell BH-1... 48 5.3 Log-derived perfect shale model E vs. measured E for MI622#6... 51 5.4 Log-derived perfect shale model E vs. measured E from Na concentration for ell BH-1... 52 5.5 Log-derived early shaly sand model E vs. measured E for MI622#6... 55 5.6 Log-derived early saly sand model E vs. measured E from Na concentration for ell BH-1... 56 6.1 Log-data for ell A... 62 6.2 Ne LSU S vs S from other models... 64 6.3 Water Saturation of zone Y ith gamma ray versus depth for ell A... 65 6.4 Log-data for ell B... 67 6.5 Water Saturation of zone Y ith gamma ray versus depth for ell B... 68 6.6 Resistivity log data for ell... 69 6.7 Porosity log data for ell 16... 70 6.8 Water saturation estimated from ne LSU shaly sand model for ell... 72 7.1 Rate of penetration versus normalized eight on bit in Matagorda Island 623 Field... 80 7.2 Rate of penetration versus log-derived cation exchange capacity plot for ell MI 622 # 6 bit run #8... 82 vi
7.3 Normalized Rate of Penetration vs. log-derived E for Well MI622# 6 Bit run #8... 82 7.4 Normalized rate of penetration vs. log-derived E for Well MI636#1 Bit run # 13... 83 7.5 ROP vs. measured E plot for Well MI622#6 Bit run #8... 85 7.6 ROP n vs. measured E plot for Well MI622#6 Bit run #8... 86 7.7 Gamma ray vs. depth ith the chronological order of E samples for MI 622#6 bit run # 8... 87 7.8 ROP n vs. measured E plot for PD bit runs of Well MI#622#6... 90 7.9 ROP n vs. measured E plot for Well 622# 6 bit run # 12... 90 7.10 The inverse of the specific energy vs. measured E for Well MI 622# 6 bit run # 8... 92 7.11 The inverse of the specific energy vs. measured E for Well MI 622# 6 bit run # 12... 92 7.12 Force Ratio vs. Measured E for Well MI 622# 6 bit run # 8... 94 7.13 Force Ratio vs. Measured E for Well MI 622# 6 bit run # 12... 94 7.14 Depth of cut vs. measured E for Well MI 622# 6 bit run # 8... 96 7.15 Depth of cut vs. measured E for Well MI 622# 6 bit run # 12... 96 7.16 The chart for normalized rate of penetration vs. E... 98 7.17 The chart for inverse of specific energy vs. E... 98 7.18 The chart for depth of cut vs. E... 99 7.19 Gamma Ray and log-derived cation exchange capacity vs. depth, ell MI622#6 bit run #8... 100 7.20 Normalized rate of penetration and log-derived cation exchange capacity vs. depth, ell MI622#6 bit run #8... 101 7.21 The normalized rate of penetration versus log-derived cation exchange capacity for MI 622#6 bit run # 8... 103 7.22 The inverse of the specific energy versus log-derived cation exchange capacity for MI 622#6 bit run # 8... 103 vii
7.23 The depth of cut versus log-derived cation exchange capacity for MI 622#6 bit run # 8... 104 7.24 Gamma Ray and Log Derived Shaly Sand Model E vs. depth for MI 636 # 1 bit run # 13... 106 7.25 Log Derived Shaly Sand Model E and Normalized rate of penetration vs. depth for MI 636 # 1 bit run # 13... 107 7.26 The normalized rate of penetration versus log-derived E for MI 636 #1 bit run #13... 108 7.27 The inverse of the specific energy versus log-derived E for MI 636 #1 bit run #13... 108 7.28 Depth of cut versus log-derived E for MI 636 #1 bit run #13... 109 viii
ABSTRAT Researchers at Louisiana State University, LSU, have introduced several petrophysical models expressing the electric properties of shaly sands. These models, to be used for hydrocarbon detection, are based on the Waxman and Smits concept of supplementing the ater conductivity ith a clay counterions conductivity. The LSU models also utilize the Dual Water theory, hich relates each conductivity term to a particular type of ater, free and bound, each occupying a specific volume of the total pore space. The main difference beteen these models and the other shaly sand models is that the counterion conductivity is represented by a hypothetical sodium chloride electrolyte. This study introduces a modified version of early LSU models. This modified model eliminates a questionable assumption incorporated in all previous shaly sand models. Previous models use same formation resistivity factor for all terms in the model. The proposed model considers that the electric current follos the effective porosity path in the term representing the free electrolyte and follos the clay porosity path in the term representing bound ater. The differentiation beteen the to paths is accomplished by using to different formation factors one in the free ater and another in the bound ater term of the model. It also used to different cementation exponents to express formation factors in terms of porosity. The validity of the ne model as checked using cation exchange capacities measured on core samples and drill cuttings. alculated cation exchange capacities display good agreement ith the measured cation exchange capacities. The ater saturation calculated using the ne model are more representative of hydrocarbon potential of the zones of interest. ix
In addition, cation exchange capacity calculated using this modified model and log data acquired during drilling has shon potential for diagnosis of pending bit balling of PD bits drilled ith ater based mud in overpressured shale. x
HAPTER 1 INTRODUTION The main purpose of ell logging is the identification and evaluation of the potential of hydrocarbon bearing formations. The potential of a zone is measured by estimating its ater saturation, S. In clean (shale free) formations, ater saturation can be calculated using the ell-knon Archie s equation. Archie s equation is based on the assumption that brine is the only electric conductor in the formation. Hoever, this is not the case in shaly sand formations here ions associated ith clay minerals also transport electricity. The presence of clay minerals results in reduction of the SP deflection, E SP and an increase in the rock conductivity, t. Hence, cation exchange capacity, hich represents the clay ability to conduct electricity, has a considerable effect on the evaluation of hydrocarbon-bearing formations. onsequently, the use of clean sand models to estimate the ater saturation results in inaccurate estimation of the potential of hydrocarbon zones. The result is usually higher ater saturation than actually present in the formation. Water saturation of hydrocarbon bearing shaly formations can be detected using available V sh or E models. V sh models assume that the shale effect is proportional to the shale volume. They can be easily misunderstood and misused. The major disadvantages are that there is no universally accepted V sh indicator and they do not consider the clay type. Therefore, V sh shaly sand models fail to consistently predict representative values of hydrocarbon saturation from ireline data 6,12. urrent E models are based on the cation exchange capacity and ionic double layer concepts. These models yield a better result than the V sh models 1
because E considers different clay types. The use of these models is impractical. Because Q v is not commonly available to the log analyst and different laboratory techniques are found to yield different Q v values for same core sample. Researchers at LSU developed a shaly sand interpretation technique using the log data from resistivity, spontaneous potential, neutron and density logs referred to herein as LSU Model. These models to be used for hydrocarbon detection are based on the Waxman and Smits concept of supplementing the ater conductivity ith a clay counterions conductivity. The LSU models also utilize the Dual Water theory, hich relates each conductivity term to a particular type of ater, free and bound, each occupying a specific volume of the total pore space. The main difference beteen these models and the other shaly sand models is that the counterion conductivity is represented by a hypothetical sodium chloride electrolyte. The LSU model is a practical approach that represents the conductivity behavior of shaly sand. Hoever, as all available models, the LSU model assumes that the electric current follo the same path in both free and bound ater part hich follos from using same formation factor for the free ater and bound ater terms. Also, this model application to field data has never been supported by E measurements. Analysis of previous field application (hapter 5) revealed that it as necessary to modify the LSU shaly sand model. In the first part of this study a ne modified shaly sand model is presented. The modification takes into account that the electric current follos the effective porosity path in the term representing the free electrolyte and follos the clay porosity path in the term representing the bound ater term. This modification results in to different formation factors one for free ater and another for bound ater. 2
In the second part of this study, the model is applied to the diagnose of bit balling using the correlation beteen the drilling parameters and cation exchange capacity. Global bit balling as identified as the primary cause of ineffective PD bit performance hen drilling shale ith ater based muds 3. Global bit balling results from cohesion beteen shale cuttings. Agglomeration of cuttings creates a ball, hich jams the space beteen the bit body and the bottom of the hole reducing bit efficiency. It as theorized that the origin of this phenomenon and its severity are related to shale electrochemical properties. Shale electrochemical properties can be represented by its cation exchange capacity, E. Drilling may be optimized if a petrophysical model is developed relating cation exchange capacity, E, to shale properties commonly measured using logging technology. Demircan, Smith and Bassiouni 7 stated that cation exchange capacity values correlate reasonably ell ith effective drilling rates for shale formations. Hoever, the scatter plot technique is used in the previous study to determine the correlation beteen the rate of penetration and E. Also, only rate of penetration and normalized rate of penetration are used as a drilling parameter to correlate ith E. The purpose of this part of study is to investigate the correlation beteen the drilling parameters; rate of penetration, normalized rate of penetration, specific energy, force ratio and depth of cut, and E, using statistical methods. An algorithm has to be developed and tested to diagnose pending bit balling for PD bits. 3
HAPTER 2 REVIEW OF PETROPHYSIAL MODELS FOR SHALY SANDS This chapter revies general concept of the shaly sand models and the petrophysical models used for the determination of the ater saturation, S in shaly sand. 2.1 General oncept of Resistivity Models 2.1.1 General oncept of lean Formation Model For clean formations, Archie 61 introduced the concept of the formation resisitivity factor, F hich is defined as; F R o (2.1) R o here R o is the resistivity of the rock hen fully saturated by an electrolyte of a resistivity R. o and are the respective conductivites. Thus, a plot of o vs. for a clean formation should yield a straight line of a slope of 1/F passing through the origin. Furthermore, as illustrated in Figure 2.1, the formation factor is empirically related to the porosity of the rock as: F a (2.2) m here the coefficient a and the cementation exponent m are generally assumed constant for a given formation. Archie 61 concluded that the resistivity exhibited by a clean formation is not only affected by the resistivity of the saturating brine and its porosity, by also the amount of electrolyte present in the pore space. This results in the Archie s resistivity equation for clean hydrocarbon formation; n t S (2.3) F 4
here S is the ater saturation expressed as a fraction of the pore space, n is the saturation exponent, and t is the conductivity of the reservoir rock under S saturation conditions. 2.1.2 The General oncept of Shaly Sand Formation Models As mentioned earlier, the conductivity of a ater bearing clean rock, o, varies linearly ith the conductivity of saturating fluid as; o (2.4) F Hoever, shaly sands exhibit a complex behavior as illustrated in Figure 2.1. At lo salt concentrations of the saturating electrolyte, the conductivity of a shaly sand rapidly increases at a greater rate than can be counted by the increase in 32. With further increase in solution conductivity, the formation conductivity increases linearly in a manner analogous of clean rocks. The magnitude of formation conductivity for shaly sand is generally larger than the magnitude of formation conductivity for a clean formation at same porosity. The excess conductivity is attributed to the presence of shaly material. A more general relationship beteen the conductivity of formation, o, and conductivity of free ater, for shaly sand formations can be described by folloing equations 32 ; For ater formation o X (2.5) F Where, o onductivity of the formation hen fully saturated ith ater onductivity of ater 5
o Shaly Sand Trend lean Rock Trend Nonlinear Zone Linear Zone o = /F Figure 2.1 Typical conductivity o - plot for shaly sands 32 F X Formation factor Shale conductivity term The ratio of / o is effectively equal to the intrinsic formation factor only if shale conductivity is sufficiently small and/or is sufficiently large. Additionally, the value of X is not alays constant. The most accepted fact regarding the effect of shaliness on the conductivity behavior of a rock sample is that the absolute value of X increases ith to some maximum level after hich it remains constant at higher salinities 32. This corresponds to respectively to non-linear and linear portions of the shaly formation conductivity of Figure 2.1. For hydrocarbon-bearing formation; n t S X (2.6) F 6
Where; S =Water saturation above the irreducible ater saturation n=saturation exponent The other type of model for hydrocarbon bearing shaly sands includes another parameter, S s related to changes due to shaliness here s is a shale term saturation factor 32. n s t S XS (2.7) F 2.2 V sh Shaly Sand Models V sh is defined as the volume of the etted shale per unit volume of reservoir rock. Hossin 31 defined the shale conductivity term, X by the folloing equation; X V 2 sh sh (2.8) Hossin s shaly sand relationships involving V sh in ater-bearing formation and in hydrocarbon-bearing formation are respectively; o t V 2 sh sh (2.9) F S V n 2 sh sh (2.10) F Simondoux 39 (1963) reported experiments on homogenous mixtures of sand and montmorillonite. In his proposed equation, V sh does not correspond to the etted shale fraction like in Hossin s equation, because the natural calcium montmorillonite as not in the fully etted state. Simondoux 39 s proposed shaly sand equation for ater formation and hydrocarbon formation are, respectively; F V o sh sh (2.11) 7
t S V n sh sh (2.12) F Poupon and Leveaux 40 developed a model for Indonesia that has fresh ater formation and high degrees of shaliness. His proposed shaly sand model is; For ater formation: o Vsh 1 V 2 sh sh (2.13) F and for hydrocarbon formation: t Vsh n / 2 1 2 n / 2 S Vsh sh S (2.14) F It should be emphasized that Hossin 21 s and Simandoux 39 equations better describe the linear region. In contrast, Poupon-Leveaux 40 better describes the nonlinear region of o vs. plot. 2.3 Log Derived lay Volume Indicators Table 1. lists the equations developed for clay volume indication 43,44. Fertl 43,44 stated that one of the ays to determine clay volume is by using natural gamma ray spectral data. Th and K curves are used simultaneously by calculating the product index. The advantage of product index is that it s virtually independent of clay types. Besides the unavailability of a universal V sh equation, the disadvantage of V sh models is that the V sh parameter does not consider the effect of the mode of distribution or the mineral composition of shales. Hence, same numerical fractions of V sh may result in highly different shale effect. Table 2.1 Log-derived clay volume indicators 43 Logging urve Mathematical relationship Favorable onditions Unfavaroble onditions Spontaneous Potential V cl=1.0-(psp/ssp) Vcl=1- Water-bearing laminated shaly sand Rmf/R=1.0 Thin, Rt zones. Hydrocarbon-bearing. Large electrokinetic and/or invasion effects c1.0 as a function of (Table ontinued) 8
V cl=(psp-sp min)/(ssp-sp min) Vcl=1-c clay type Gamma Ray V cl=(gr-gr min)/(gr max-gr min) V cl=(gr-gr min)/(gr max-gr min) Only clay minerals Radioactive 1.0 frequently approximately 0.5 hen V cl40% Radioactive minerals other than clays (mica, feldspar and silt) Only potassium-deficients kaolinite present. Uranium enrichment in permeable, fractured zones. Vcl=(GR-W)/Z Where W, Z=geological area coefficients V cl=0.33(2 2Vcl -1) Highly consilated and Mesozoic rocks Radiobarite scales on casing, Severe ashouts Younger, unconsolidated rocks Older consolidated rocks Spectralog V cl=0.083(2 3.7Vcl -1) V cl=(a-a min)/(a max-a min) V cl=(a-a min)/(a max-a min) V cl=0.33(2 2Vcl -1) V cl=0.083(2 3.7Vcl -1) Tertiary clastics ondition similar to gamma ray discussion A=Spectral log reading (K in % Th in ppm) A min=minimun value in clean zone Amax= Max. Value in essentially shale zones Similar to gamma ray discussion. Hoever, uranium enrichment in permeable, fractured zones and radiobarite build up are no limitations. If Th curve is used, localized bentonite streaks should be ignored. Resistivity V cl=(r cl/r t) 1/b Where b=1.0, b=2 Lo porosity zones (carbonate marls); pay zones ith lo (S -S i) R cl/r t from 0.5 to 1.0 R cl aproaches R t High porosity ater sand; high R ct values V cl=r cl Neutron Vcl / N D High gas saturation or very lo reservoir porosity ncl is lo Pulsed Neutron V / cl N D V cl=(- min)/( max- min) V cl=( cl/)(- min)/( max- min) min can be varied Fresh ater enviroment lo porosity and gas bearing zones Density/Neutron Density/Acoustic Neutron/Acoustic V cl=( B( Nma-1)- N( ma- t)- t Nma+ ma))/ (( sh- f)( Nma-1)( ma-t) V cl=( B(t ma-t f)- t( ma- t)- tt ft ma+ mat t)/ ((t ma-t f) ( sh- t) -(t sh-t f) ( ma- f) V cl=( N(t ma-t t)- t( Nma-1)- t ma+ Nmat f)/(( t ma-t f)( Nsh-1)-( Nma-1)( t sh-t f) Less dependent on lithology and fluid conditions than density /neutron crossplot Use only in gas bearing zones ith lo S Too lo V cl in prolific gas zones. Not for use in severe hole conditions,lithology effected Badly ashed out ellbores. Highly uncompacted formation (shallo overpressure) Similar effects because of shaliness on both logs 2.4 ation Exchange apacity, E, Models The clay minerals are phylosilicates; they have a sheet of structure somehat like that of micas. The principal building elements of clay mineral are (1) a sheet of 9
silicon (Si) and oxygen (O) atoms in a tetrahedral arrangement and (2) a sheet aluminum (Al), oxygen and hydroxyl (OH) arranged octahedral pattern. These sheets of tetrahedral and octahedral are arranged in different fashions to give the different group of clay minerals 1. In the tetrahedral sheet, tetrahedral silica (Si +4 ) is sometimes partly replaced by trivalent aluminum (Al +3 ). In the octahedral sheet, there may be replacement of trivalent aluminum by divalent magnesium (Mg +2 ). When an atom of loer positive valence replaces one of higher valance, a deficiency of positive charges results. This excess negative charge is compensated for by the adsorption onto the layer surfaces of cations that are too large to be accommodated in the interior of the crystal 1. The accumulated ions are called counterions. In the presence of ater, the compensating cations, such as Mg, Na and a, on the layer surfaces may be easily exchange by other cations, hen available in solution; hence they are called exchangeable cations. The number of these cations can be measured and is called cation exchange capacity, E, of the clay. The replacement poer of different cations depends on their type and relative concentration. There is also definite order of replaceability, namely Na<K<Mg<a<H. This means that hydrogen ill replace calcium, calcium replaces magnesium, etc 1. ation exchange capacity models result from a phenomena called the double layer. Winsaur and Mcardell 31 (1953) are the first ones that introduced the double layer model. Winsaur and Mcardell stated that the excess conductivity, double layer conductivity of shaly reservoir rocks, as attributed to adsorption on the clay surface and a resultant concentration of ions adjacent to this surface. Hill and Milburn 29,30 shoed that the effect of clay minerals upon the electrical properties of formation is related to its cation exchange capacity per unit pore volume, 10
Q v. They developed a shale effect term b, to examine the o - relationship. This approach as not developed further because o ould increase as decreases at lo salinities, in the range here corresponding to the non-linear part in figure 2.1. The most commonly used cation exchange capacity models are Waxman and Smits 8,9 Shaly Sand Model and Dual Water Shaly Sand Model 34,35. The next to sections give detailed explanation on these commonly used Shaly Sand Models. Some of the other developed shaly sand models for hydrocarbon bearing formations are given in Table 2.2. Table 2.2 Shaly sand models for hydrocarbon-bearing formations 31 REFERENE EQUATION OMMENTS m =molal concentration of exchangeable cations in formation ater m sh=molal concentration of exchangeble cation associated ith shale l.de Witte (1955) km kmsh t 15 k=conversion from msh, to conductivity S S F relates to total interconnected porosity. F F S relates to total interconnected pore space A j. Witte (1957) F=maximum formation factor 2 t S AS A=shaliness factor F S relates to total interconnected pore space Patchett & Rausch s=conductivity due to shale ( sh). F 2 (1967) relates to total interconnected porosity. S Bardon & Pied (1969) Schlumberger (1972) Juhasz (1981) Doll (unpublished) Alger et. al. (1963) t S s S F 2 t S Vsh sh S F 2 t S Vsh F(1 V ) S t t t F S 2 sh S F (1 q) 2 F q sh F sh S sh Vsh sh S Fsh 2 sh 2V sh 2 S 2 V F q(1 q)( F 2 sh sh sh ) relates to total interconnected pore space Modified Simandoux equation F relates to free fluid porosity of the total rock volume inclusive of laminated shales Normalized Waxman-Smits equation F=1/ m here is derived from the density log and corrected for hydrocarbon effects F sh= 1/ sh m here sh is derived from the density log S relates to total interconnected pore space lay Slurry model F relates to total volume occupied by volume by fluid and clay S relates to fluid-filled pore space (Table ontinued) 11
Husten & Anton (1981) Patchett & Herrick (1982) Poupon & Leveaux (1971) Poupon & Leveaux (1971) Woodhouse (1976) Raiga- lemenceau et al. (1974) 2 t S 2V F 2 Vsh sh 1 b 1 S sh F=1/ 2 t here t id total interconnected sh F b 2 t V sh V 1 V sh sh F S F sh BQ v S porosity = f S relates to total interconnected pore space 1 Laminated sand shale model 2 t S 2 F 2Vsh 2 Vsh shs 2 t S 2 F 2 2 Vsh shs 2 t S 2 F 22Vsh 2 Vsh shs t b 2 S 2 F 1.72 bs Vsh F 2Vsh Vshsh S F Vsh F F 22Vsh 1.72 b sh S 2 2 sh S b S 1.5 2 V sh=volume fraction of laminated shales only F relates total Indonesia Formula Simplified Indonesia Formula for Vsh0.5 Modification of Poupon and Leveaux equation for tar sands Dual Porosity Model b t e 2.4.1 Waxman and Smits Shaly Sand Model 8,9 Hill and Milburn s 30 ork led Waxman and Smits 8 to propose a ne conductivity model. This model assumes; 1)a parallel conductance mechanism for free electrolyte and clay-exchange cation components, 2)an exchange cation mobility that increases to a maximum and constant value ith increasing free electrolyte concentration, and 3) identical geometric conductivity constants applicable for the contributions of the both free electrolyte and the clay-exchange cation conductance to the sand conductivity. The assumption of a parallel conductance mechanism for free electrolyte and clay-exchange components results in; 12
o x y (2.15) Where c is the conductance associated ith the exchange cations. is the specific conductance of aqueous electrolyte solution, mho cm -1 and x, y are geometrical factors. Assuming identical geometry, the electric current transported by the counterions associated ith clay travels along the same tortuous path as the current attributed to the ions in the pore ater. Thus, the geometric parameters (x, y) are assumed to be same and equal to the shaly sand formation factor and this results in; x y 1 F (2.16) F Shaly sand formation resistivity factor x,y= Geometric constant Waxman-Smits 8 illustrated 1 / F as the slope of the linear correlation of the core conductivity, o vs. the equilibrating solution conductivity,., except for the loer values of the equilibrating ater conductivity. This is corresponding to the linear zone of shaly sand trend in figure 2.1. By substituting equation 2.16 into equation 2.15, equation 2.15 becomes, o 1 F e (2.17) Where o e Specific conductance of sand, 100 percent saturated ith aqueous salt solution, Specific conductance of clay exchange cations Specific conductance of aqueous electrolyte solution 13
The most important aspect of the Waxman and Smits 8 model is that the conductance contribution of the clay is defined as a product of the volume conterion concentration, Q v, times the equivalent counterion conductance, B. Thus: o 1 F BQ v (2.18) B, equivalent counterion conductance at 25, hich is a function of the counterion mobility, is defined as; B 0.046(1 0.6 exp( /0.013)) (2.19) The model for hydrocarbon bearing formations has mainly to additional assumptions. First, it is assumed that the counterion concentration increases in pore ater as S decreases; Q v Q S v (2.20) here Qv is the effective concentration of cations at S conditions. Hence, Waxman and Smits 8 conductivity equation for hydrocarbon bearing shaly sand formation is; 1 t BQv / S (2.21) G * G* is a geometric factor, being a function of porosity, ater saturation and pore geometry, but independent of clay content,q v, and defined as; 1 * G n* S (2.22) F * The parameter n* is the saturation exponent for shaly sands in Waxman and Smits 8 model. 14
The ater content of formation is commonly expressed as a function of the Resistivity Index, I, and it is given by; I S BQv BQv / S n* (2.23) In terms of resistivity, equation 2.23 becomes; I S 1 R BQv 1 R BQv / S n* The Waxman and Smits 8,9 (2.24) model predicts greater hydrocarbon saturation values than those otherise calculate from clean formation models. Waxman and Smits 8,9 model is highly accepted because of its simplicity and the amount of supporting experimental ork. 2.4.2 Dual Water Shaly Sand Model Dual ater model, D-W, as first proposed by lavier, oates and Dumanoir in 1977. laiver, oates and Dumanoir 34,35 published the latest version of the model in 1984. The expose of the dual ater model in this section refers to the last version, hich is published in 1984. Waxman-Smits 8 model is simple and includes the amount of supporting experimental ork, hoever; some effects related to the adsorptive properties of the clays that had not been taken into account, namely clay ater that is a result of double layer associated ith the clay. Double layer is assumed to contain mainly positive charges and balances the negative charges on the clay surface. This diffusion layer can be considered as saltfree zone and its effect continues up to some distance from the clay surface. Hence, the pore space of shaly sand is assumed to be filled ith the clay ater and far ater. 15
Each one of these aters occupies a fraction of the available pore space that are called clay ater porosity and far ater porosity. Both dual ater model and Waxman Smits 8 model consider that the conductivity of the saturating fluid is complemented by the conductivity of a clay counterions. The basic difference of dual ater model 34,35 from the Waxman and Smits 8 model is that dual ater model considers both the far ater and the clay ater ith specific conductive properties. The D-W model characterizes the shaly sand formation by total porosity, t, formation factor, F o, shaliness parameter, Q v,, and its bulk conductivity t observed at total ater saturation, S t. The D-W model also assumes that the formation behaves as a clean rock of the same porosity, tortuosity, and ater saturation but containing an equivalent conductivity, e. The main assumption of dual model is that equivalent conductivity, e, is a mixture of the clay and far ater conductivity meaning that model geometry factors related to travel path of the electrolytes are equal. Hence, equivalent conductivity in D-W model results in; e V V (2.25) c c f here c and V c are the conductivity and volumetric fraction of the clay ater. Likeise, and V f represents the conductivity and volumetric fraction for the far ater. lay ater conductivity, c, is independent from the clay type and amount of clay, but is only given by the conductivity of the clay counterions. The fractional volume V c is proportional to the counterion concentration in terms of the total pore volume, Q v : 16
V c v Q (2.26) Q v T here v Q is the amount the clay ater associated ith one unit of clay counterions. The far ater conductivity,, is assumed identical to that of the bulk formation ater. Volumetric fraction of the far ater, V f, is the remaining of the pore space and expressed as; V f V V S v Q ) (2.27) c T ( T Q v here V is the total ater content. The conductivity e is given by the combined volumetric averages expressed in terms of e e S T 1 ( vqqvc ( S T vqqv ) ) (2.28) S T Effective ater conductivity in shaly sand Water saturation in volume fraction of total porosity vq Volume of clay-ater per counterion at 22 hen =1 cm 3 /meq (ml/meq) Q oncentration of clay counterions per unit pore volume, meq/cm 3 v c Far ater conductivity, S/m (mho/m) onductivity of clay ater, S/m (mho/m) expressed as: Using the Archie s relationship for clean rocks the conductivity of shaly sand is e n S o (2.29) t F o T Where n o is the saturation exponent in Dual Water Model; 17
Using equations 2.28 and 2.29, dual ater hydrocarbon bearing conductivity model is expressed: t no S vqq T v ( c ) Fo S T (2.30) In ater bearing formations, here S T =1, equations 2.28 and 2.30 are simplified to v Q ( 1 v Q ) (2.31) e Q v c Q v and o 1 ( vqqvc (1 vqqv ) ) (2.32) F o Water saturation in equation 2.28 is computed as a fraction of total porosity because it includes the clay ater. Far ater saturation term, S f, is defined for better calculation of ater saturation because shaly sands may have high ater saturation and still produce ater free hydrocarbon. S f V f (2.33) f Where f is effective porosity and is given f v Q (2.34) t Q v t Where t is the total porosity, fraction; and ater. Hence, free ater saturation results f is the fraction of porosity filled ith far S f S T vqq 1 v Q Q v v (2.35) 18
HAPTER 3 EARLY LSU SHALY SAND MODELS Researchers at Louisiana State University, LSU, have introduced several petrophysical models expressing the electric properties of shaly sands. These models, to be used for hydrocarbon detection, are based on the Waxman and Smits 8 concept of supplementing the ater conductivity ith a clay counterions conductivity. The LSU models also utilize the Dual Water theory 34, hich relates each conductivity term to a particular type of ater, free and bound, each occupying a specific volume of the total pore space. These models are defined in this chapter. 3.1 Silva-Bassiouni Shaly Sand Model Silva and Bassiouni 10,11,12,13, (S-B) developed a model based on a double layer of far and bound ater. Hoever, Silva Bassiouni 10,11,12,13 shaly sand model differs fundamentally from dual ater model in terms of the definition of equivalent counterion conductivity. S-B Shaly sand Model estimates the equivalent counterion conductivity from a method based on treating the double layer region as an equivalent electrolyte hose properties are derived from basic electrochemical theory. 3.1.1 S-B onductivity Model The equivalent counterion conductivity, e, is defined as the sum of conductivity of the diffused double layer dl, and the conductivity of the free equilibrate solution, es. e (3.1) dl es The conductivity of the counterions under the influence of diffuse layer, dl, is expressed as; v (3.2) dl fdl cl 19
here v fdl is the fractional volume occupied by the double layer, expressed in terms of total porosity. It can be estimated from equation 3.3; v V F Q (3.3) fdl u dl v here V u is the amount of clay ater associated ith a unit of clay counter-ions. F DL is double layer expansion factor and expressed as; 2 o 2 1 / 2 F DL ( h B n) (3.4) here h is 6.18 A, Bo is the coefficient of ion-size term in Debye-Huckel theory and n is the concentration of free formation ater in molar units. B o is empirically related to temperature (15<T<100 ) by the folloing polynomial. B o 4 7 2. 3248 1.510810 T 8.93510 T (3.5) n (ions/cm 3 ), local ion concentration expressed as; 20 n N 6.0210 (3.6) here N is the electrolyte concentration of the equilibrating solution in normality units. In equation 3.2, cl is the conductivity countributions of the exchange cations associated ith clay. Silva and Bassiouni defined cl in terms of the equivalent conductivity of the counterions in the double layer, eq, and the counterion concentration ithin the double layer, n eq, hich are electrochemical terms. n (3.7) cl eq eq The counterion concentration ithin the double layer, Q n eq is defined as: v neq (3.8) v fdl The equivalent counterion conductivity of the clay counterions, eq, is defined; 20
eq eq /( fg F( ne)) (3.9) here eq is the equivalent conductivity of the equivalent Nal solution representing the double layer, and at a temperature of 25, it is expressed as: eq 12.645 7.6725( n 11.3164n eq 1 / 2 eq ) 1 / 2 (3.10) F(ne) and fg are empirically determined correction factors. At a temperature of 25, they are given by; F 2 2 ( ne) 1 3.8310 ( 0.5) 1.76110 ( n eq n eq 0.5) 2 for n eq >0.5 mol/l(3.11) F( ne) 1.0 for n eq <0.5 mol/l (3.12) fg 1 / nc ( / ) (3.13) v fdl Q v nc (3.14) 2 0.6696 1.1796v fdl 0.14426v fdl Substituting equation 3.7 into equation 3.2 results in; dl n v (3.15) eq eq fdl Silva and Bassiouni 10,11,12,13 assumed that the electrical properties of the equilibrating solution are equal to those in the bulk solution. Therefore, the conductivity of the free equilibrate solution, es, in Equation 3.1 is defined as: ( 1 v ) (3.16) es fdl Substituting equations 3.15 and 3.16 into equation results in; e eq n eq v fdl 1 v ( fdl ) (3.17) Using the analogy of clean formation expression, the shaly sand conductivity model for ater bearing formations is expressed as: e o (3.18) Fe 21
By substituting Equation 3.17 into equation 3.18, S-B shaly sand conductivity model for ater bearing formations becomes; o eq n eq v fdl ( 1 v F e fdl ) (3.19) here F e is the formation factor of an equivalent clean formation of the same porosity, T that can be expressed as: F e (3.20) me T here me is an appropriate cementation exponent. 3.1.2 S-B Membrane Potential Model Silva-Bassiouni modified Thomas membrane potential (Em) model by introducing a correction factor,. SP Em sh Em ss (3.21) SP 2RT F m 1 m 2 SH 2RT d ln( m ) F m 1 m 2 ss eq n eq v fdl t e h na (1 v / Fe fdl ) d ln( m ) Where; Emsh Emss Electrochemical potential of shale Electrochemical potential of shaly sand R Universal gas constant F Faraday constant T Absolute temperature Tna Sodium transport number m1 and m2 molal concentrations of the formation ater and mud filtrate Mean activity coefficient 22
For n 1.28Q ( )/ v n For n 1 n Water conductivity at neutral point, 16.6 mho/m at 25 Transport number, T na is a ratio of the electric current carried by an ion to the total electric current here both pressure and concentration gradients are zero. S-B assumed that the current carried by the clay counterions is parallel to the current carried for the ater. Both currents are related to the same cell constant. Electrolyte can be treated as Nal. Thus T na is; T na h eqqv t na (1 v Q (1 v eq v fdl fdl ) ) (3.22) Where; Stokes theoretical expression for Nal Hittorf transport number at 25; t h na 1/2 50.155.402n (3.23) 1/2 126.45155.726n here; n= molar concentration of the far ater Mean activity coefficient can be determined from the folloing equation; 1/2 0.5115n log 1.75log( a ) log(1 0.27m) 1/ 2 A (3.24) 11.3065n here; a A. 99948.03959m0.0015075m 2 Membrane potential can be calculated as follos; 23
1. The concentration interval beteen the electrolytes is divided into 100 subintervals. 2. The magnitude of eq, v fdl, 3. Each T na is calculated h t na, and are determined 4. The result is then multiplied ith the constant The S-B models accurately describes the resistivity behavior and membrane potential, hoever, the application requires too many empirical correction parameters like fg, Fe(ne) and. This makes application complicated for field conditions. 3.2 The LSU (Lau-Bassiouni) Shaly Sand Model This LSU shaly sand model is a modified S-B model. Modification is done to eliminate the empirical derived correction factors so that the model can be applied to temperatures other than 25. Lau-Bassiouni 14,15,16 defined the eq, n eq and v fdl in a different ay than the ones in the S-B model, to apply the LSU model for higher temperatures and field data. 3.2.1 The LSU (Lau-Bassiouni) onductivity Model For ater bearing formation, the conductivity model is defined by; o ( n v (1 v ) / Fe) (3.25) eq eq fdl fdl Fe is calculated using density-neutron porosity crossplots and m=2, a=0.81 are assumed. For the field ell log data, Lau and Bassiouni 14,16 defined the and eq v fdl equations as a function of temperature. This results in; eq exp( 58.84.1026n.07871ln( n ).0216Ta11.85ln( Ta)) (3.26) eq eq Where; 24
Ta=Absolute temperature, K v fdl 2 1/ 2 (.28.0344ln( T /25)(618 B n) Q (3.27) 2 o v Where; T=Temperature, n=molarity B o 4 7 2. 3248 1.510810 T8.93510 T ln( n) 68.113.5791ln( Ta) 2.28910 2 3 Ta1.1854ln( ) 4.6761 10 As it is seen from the above equations, Q v. n eq is defined as; v fdl is a function of cation exchange Qv Ta neq (3.28) v 298 t fdl For hydrocarbon bearing formation conductivity model is defined as; ( n v (1 v ) / Fe) S (3.29) eq eq fdl fdl n In case of hydrocarbon bearing formations, the ater saturation is less than unity, exchange cations associated ith clay, Q v, become more concentrated in the pore space. Henceforth, Q v hich has the effect of hydrocarbon, called concentrated Q v is defined as; Q Q / S v v (3.30) For hydrocarbon bearing formations, Qv is used in the equations here Qv is used. 3.2.2 The LSU (Lau-Bassiouni) Spontaneous Potential Model Lau and Bassiouni 15,16 expressed spontaneous potential as; SP Em sh Em ss 25
SP 2RT F m 1 m 2 m eff m1 2RT eqneqv fdl tna(1 v fdl ) d ln( m ) d ln( m ) (3.31) F / Fe m 2 e For the field ell log data, Lau and Bassiouni 15,16 defined the equations as a function of temperature. This resulted in; log( ) log( ).5YL.5ZJ and t na 298 298 298 (3.32) Where; 1/2 298.05115n log( ) 1.75log( a 1/ 2 11.3065n a A 2 2. 999483.095910 m.0015m 289.15Ta Y 8.3147(298.15)2.3026( Ta) 1 Z 298.15Y log( Ta/298.15) 8.3147 L J 2878.6m 1m 1/2 298 3182.8m986. 5 1/2 43.5m 1m 1/2 298 72m20. 36 1/2 n=molality,mol/kg m 2/3 298 activity coefficient at 25 A m ) log(1.027m) 2/3 t na is a function of both Hittorf number, t h an and ater transport number, t. This is expressed as; t na t h na t (3.33) Hittorf number, t h an and ater transport number, t can be also related to Q v and temperature as in the folloing equations; 26
t h na 2 5 exp( 2.50891.803810 ln( m).2647ln( Ta) 1.417610 Tam (3.34) for m1.0 t 0.035m.43 (.19661LN( m).1244) (3.35) Q v For m1.0 t.036m 0.04377. 04 (3.36) 1.1 Q v 3.3 Advantages and Shortcomings of Early LSU Models The early LSU models are based on the folloing assumptions; 1. A parallel conductance for free electrolyte and the bound ater 2. The bound ater can be represented by an equivalent sodium chloride solution 3. The same formation factor affects the conductivity contributions of the free electrolyte and bound ater. 4. If hydrocarbons are present they ill preferentially displace the free electrolyte 5. When the ater saturation is less than unity, exchange cations associated ith clay, Q v, become more concentrated in the pore space The advantage of the early LSU Models are; 1. The LSU conductivity model can be used to calculate Q v and Fe of the core samples using to conductivity measurements conducted ith the core saturated ith knon to different free ater conductivity. 2. The LSU SP model can be used to determine R from SP deflection if Q v is knon. Inversely, it can be used to estimate Q v hen R is knon. 3. Membrane efficiency can be derived from SP log reading at any ater bearing zone ith an adjacent shale displaying the same petrophysical properties as the shale overlying the zone of interest. If the sand is clean, Q v =0, membrane efficiency can be calculated from simultaneous solution of the LSU SP and 27
conductivity models. If only shaly ater-bearing sands are present in the interval analyzed, membrane efficiency can be calculated by trial and error using the same set of to equations. 4. The model critical use, hoever, is that it can be combined ith conductivity model to solve simultaneously for R, Q v and S. The major advantage of the LSU model over the above-mentioned ones is that it can be easily used at any temperature ithout the need of core data and any experimental ork. The main shortcoming of the early LSU model is that past field applications for determination of Q v is not supported by core measurements. In addition, like the previous shaly sand models, it is assumed that the electric current follos the same path in free and bound ater that leads to the same formation factor representing both free and bound ater parts. Henceforth, the same formation factor affects the conductivity contributions of the free electrolyte and bound ater that may result in high porosity in bound ater part causing underestimation of E. 28
HAPTER 4 NEW LSU SHALY SAND MODEL This ne model is based on the Waxman-Smits concept of supplementing ater conductivity ith clay counterions 8,9, and the dual ater theory relating the conductivity term to a particular type of ater, each occupying a specific volume of the pore space 34,35. Like Silva-Bassiouni Model 10,11,12,13, the model also assumes that counterion conductivity is represented by a hypothetical sodium chloride solution. The main improvement of the ne LSU model is the incorporation of to different formation factors, one for bound-ater and another for free-ater. Accordingly, current follos the effective porosity path in the free electrolyte and follos the bound ater porosity path in the bound ater part. All previous models incorporate only one formation factor. 4.1 Ne LSU onductivity Model First, a parallel conductance for free electrolyte and the bound ater are assumed like in Waxman and Smits 8,9 Shaly Sand Model. Hence, ater bearing formation conductivity, o, can be ritten as cl o (4.1) Ff Fb Where; cl F f F b lay conductivity, mho/m Formation ater conductivity, mho/m mf 1 / e, free ater formation factor mc 1 / b, bound ater formation factor e Effective porosity, here positive and negative ions are equilibrium 29
b m f mc Bound ater porosity Free ater cementation exponent Bound ater cementation exponent Furthermore according to Hill, Shirley and Klein 29,30 : 1 v ) (4.2) e t ( fdl v (4.3) b t fdl Where; t Total porosity v Fractional volume of the double layer fdl Substituting equations 4.2 and 4.3 into equation 4.1 results in; ) mf mc o ( t (1 v fdl )) cl ( tv fdl (4.4) Assuming that bound ater can be represented by an equivalent sodium chloride solution 4,12, equation 4.4 becomes: o mf mf mc mc ( 1 v fdl eqneqv fdl (4.5) ) Where; eq Molar counterion conductivity of the equivalent Nal solution, (mho/m)/(mole/l) Formation ater conductivity, mho/m neq v fdl t Molar counterion concentration of Nal solution, mole/l Fractional volume of the double layer, fraction Hydrocarbon bearing formation conductivity, t, is defined as; mf mf mc mc n ( 1 v fdl eqneqv fdl S (4.6) ) 30