Reservoir Characteristics of Low-Permeabil Sandstones in the Rocky Mountains! ALAN P. BYRNES? | Roce: Jy 6, 1996; Accoped: August 13,1986 ® Borophyscel Constant, Sanne, Karen ABSTRACT Understanding gas production from low-permeability sandstones requires an understanding of the in situ porosity, brine saturation, and effective gas permeability at reservoir brine saturation. Analysis of data from hundreds of cores from numerous western U.S. basins indicates that petrophysical properties of well-defined lithofacies (or log-facies) are often unique within a narrow range. Diagenesis in these sand- stones commonly resulted in the destruction of much of the original intergranular porosity and left dis- solved grains, clay-filled pores, and sheet-like connecting intergranular pore throats. Pore throats or chan nels that connect larger pores typically range in size from 1 to 0.1 micron and represent only a small portion of the total porosity. In most low-permeability sandstones, porosity is not significantly changed by confining stress changes, but in situ effective gas permeabilities range from 10 to 1,000 times less than routine air permeability. The influence of confining stress on permeability can be attributed primarily to the decrease in size of the thin, tabular pore throats that connect the larger potes. Under stress, pore throats decrease in diameter by up to 50% to 70% resulting in permeability decreases of 10 to 40 times. Gas effective permeabilities also decrease rapidly to less than 1% of absolute values at water saturations above approximately 40% to 50% “Irreducible” water saturations increase with decreasing porosity and permeability, and, in sandstones with less than 0,01 md permeability, “irreducible” water saturations increase dramatically. Cumulative flow and storage capacity plots indicate that very thin higher permeability intervals typi- cally yield a large percentage of the cumulative flow capacity. Increased water saturations due to drilling or stimulation result in lower effective gas permeabilities and can unknowingly be stabilized by capillary pressure forces if pore pressures are decreased. This type of formation damage can be remedied by increasing the gas pore pressure to displace mobile water. INTRODUCTION Fundamental to effective production of gas from low-perme- ability sandstones is understanding their in situ petrophysical properties, the most important of which are porosity, brine satu- ration, and effective gas permeability at reservoir brine satura- tion, False assumptions concerning values for these petrophysical properties, or the nature oftheir distribution, can result in signif- ‘cant costs due to inefficient completion, stimulation, or explo- ration, Misconceptions arise primarily because some in situ reservoir properties differ significantly from the more economic and easily available routine core analysis properties. For low-per- ‘meability sandstones, the response of the sandstone to removal from in situ chemical and stress conditions affects some proper- ties significantly while others are only affected nominally. This can be understood by studying the properties of cares under con- ditions similar to those in the reservoir. It is then possible to develop predictive equations that can be utilized with routine petrophysical data to predict insu values Numerous studies have been published conceming the reser- voir properties of low-permeability sandstones. Barly work by Vairogs etal. (1971) and Thomas and Ward (1972) illustrated hhow increasing confining stress and water saturation decreases permeability. Their work also showed that pore volume com- pressibility is small. The work of Jones and Owens (1980) pro- \ided a confirmation and quantification ofthese influences for & larger sample set of low-permeability sandstones and indicated The Moana Geologist, Vol. 34, No.1 Gaouary 1997). 38-51, The Rocky Mountain Association of GeologistsAP Byrnes that response of these rocks to confining stress could be explained by the presence ofa thin, sheet-like, tabular pore struc- ture, Ostensen (1983) provided @ comprehensive theoretical analysis ofthe relationship between grain boundary miero-crack dimensions and permeability. The present paper follows inthe footsteps of these eary sties and attempts to expand the geo- sraphic and lthologie range of low-permeability sandstones studied. It also examines some relationships between saturation and permeability in depth. The data and trends presented here represent the compilation of dozens of small- and large-scale studies involving many different Western U.S. low-peemeability sandstone formations and basins (Table 1). Although the data shown here are only from basins in the Rocky Mountain region, the trends shown are representative ofall low-permeability sand. stones studied by the author inthe U.S. and worldwide ‘The following sections review and analyze data for each ofthe major petrophysical variables including porosity, absolute per- reability, effective and relative gas and liquid permeability, water saturation, and “irreducible” water saturation. The next section provides a brief discussion of important lithologic con- trols on petrophysical properties whereas the last section dis- cusses the cause and correction of water-block formation damage that is common in low-permeability sandstones. LITHOLOGIC CONTROLS ‘The “low permeability” or “tight gas” sandstone designation sroups sandstones on the basis of flow properties alone. Perme- ability in these sandstones covers a range of approximately six orders of magnitude from I md to Ix10“me. Though these sand- stones have been grouped together for regulatory reasons, the Iihotogic and petrophysical variations that exist within this group ae large. Many of these variations are due to differences in litho- logic properties causing distinctly diferent pore geometries for Unique lthofacies. Generally, well-charactrized lithofacies,log- ‘cies, an petrofacies (ie, rocks exhibiting unique petrophysical trends) exhibit high degrees of correlation. That i, the petrophys- ical properties of a well-defined lithofacies or log-facies are defin- able within a narrow range and may be distinctly different from the trends for other lithofacies. Poorer correlations result from such factors as overly broad classifications of lithofacies groups, difficulty in quantifying geologic variables, problems in calibra- tion of logs, difficulty in obtaining in sits petrophysical data, and differences in the scales of investigation involved in the charac- terization of lithofacies,log-facies, and petrofacies Peirophysical properties are influenced by ithologic variables that are shaped by provenance, depositional environment, biotar- bation, diagenesis, and burial and thermal history. Many low-per- ‘meability sandstones from Western U.S. basins can be classified as quartzarenites (295% quartz), sublitharenites, or subarkoses (275% quartz). These compositions are evident in both fluvial and ‘marine depositional environments. Important mineral affected by diagenesis, and that influence petrophysical properties aftr alter- ation, include feldspar, chert, and other lithic rock fragments, Pet- rographic properties of many western U.S. low-permeability sandstones have been summarized by Dutton etal. (1993). Provenance plays an important role by controlling sandstone composition within several basins, Sandstones of the Mesaverde Group and Frontier Formation, for example, vary widely within basins in their feldspar and lithic content as well as grain size. ‘These variations are responsible fr differences in authigenic clay content and permeability of up to an order of magnitude. Depositional environments also exert a strong influence on textural parameters including grain size, sorting, packing, and bedding type. Decreasing grain size from medium-grained sand- stones to siltstones corresponds to permeability decreases of sev- eral orders of magnitude. Environments where poorly sorted fine- (osill-sized sands are interbedded with medium to fine sands also exhibit orders of magnitude decreases in permeability. This poor sorting can be the direct result of deposition or the result of bio- turbation, Poor sorting results in diminished porosity, as well as pote size, for a given volume of sandstone. A similar reduction in permeability can result from heterogeneities introduced by small-scale interbedding of very low-permeability beds within better quality rock. Environments where massive or large-scale cross-bedded sands are deposited can produce permeabilities ‘TABLE 1. List of Western U.S. Low Permeability Sandstones Formations and Basins Studied and a Summary of Petrophysical Properties Reported for Wells Basin Intorval Porosity% Permeability ‘Saturation’ Low High Avg Low High Low High Avg Denver ‘J Sandstone 8 14 108 0.0008 00527564 Green River Frontier 2 2 12 0.0001 030 836648 AlmondiBiair 1 10 67 0.002 004 = 45835 Piceance Dakota Sandstone is 0.02 0.05 40 Mancos °B" Shale Pa O01 0.08 Mesaverde 3 2 7 0.0002 008 = 3060 San Juan Dakota 2 16 as 002” 008 = 4060 Mesaverdo 3 4 8 002 OOF) as 65 Picture Clits 10 0.003 002 «50-0 Uinta Wasatch 6 4 2 0,009 005 2570 BH Washakie Mesaverde 2 16 0.0001 060 18 «80 Wind River Frontier a7 na 0.004 0.06 50 ‘After Dutton et al., 1993 "The Rocky Mounisin Association of Gooogits 38100 0 Insitu Kinkenberg Permeability (md) 024 6 62 ww Insitu Porosity (6) Figure 1, Cross-plot of in situ Klinkenberg permeability versus in sw porosity subdivided on the basis of lithofacies. Facies A generally represents siltstones and shaly sandstones with fine ripple laminations, wavy or Flaser bedded, biotrbated, or with drape structures. Facies B generally represents very fine to fine- ‘tained sandstones exhibiting small-scale cross-bedding, planar laminations, or massive bedding. Facies C represents upper fine ‘to medium-grained sandstones that generally exhibit large-scale ceross-bedding, few clay laminae, or massive bedding. several orders of magnitude greater than those where fine, ripple laminated, low-angle laminated, or highly bioturbated sands accumulate. In general, fora region of similar burial and thermal history, subdivision of low-permeability sandstones on the basis, of depositional environment results in a similar separation of petrophysical trends. ‘Though compositional variables ae highly important, all low- permeability sandstones have undergone extensive diagenesis ‘that has resulted in destruction of much of the original intergran- ular porosity and left dissolved grains, clay-filled pores, and sheet-like connecting intergranular pores. In many Rocky Moun- tain low-permeability sandstones, a significant percentage of pre- sent day porosity is secondary and occurs in dissolved feldspar or lithic rock fragments or within areas of prior carbonate cemen- tation, Because the dissolution and reprecipitation processes may ‘occur near each other, and grain densities for clays and feldspars are similar, there may be no significant increase in total porosity fon the seale of inches, As a result, secondary porosity develop- ‘ment does not result in porosity enhancement but only porosity redistribution, An exception to this can occur where carbonate cements are dissolved and removed. ‘The pettophysical properties for all low-permeability sand- stones are dominantly controlled by the influence of diagenesis 4s ithas acted upon a given composition sediment. Diagenesis is 39 Low Permeability Sandstones, Rocky Mountains the result of the influence of pressure, temperature, and time as they are expressed through physical processes such as grain, rearrangement, plastic deformation, brittle deformation, etc., and ‘through chemical processes including quartz pressure solution, suturing, quartz and calcite cementation, dissolution of lithic rock ‘and feldspar components, and precipitation of authigenic clays, ‘Measures of pressure, time-pressure and time-temperature are typically modeled through burial and thermal history reconstruc~ tion (Byres and Wilson, 1994). Many low-permeability sand- stones had maximum burial depths as much as 4,000 to 10,000 f. ‘greater than present burial depths. Increasing time-temperature ‘exposure results in increased quartz overgrowth cementation, ‘hich reduces both pore space and, more importantly, pore throat sizes significantly. Increasing time-temperature exposure also results in increasing lithic and feldspar grain dissolution result- ing in increased clay precipitation and secondary porosity devel- ‘opment. Burial and thermal history influences can also be modi- fied by fluid movement through a given region as the result of position within a basin, Figure | illustrates the permeability and porosity trends exhib- ited by three common lithofacies. These lithologies also exhibit significantly different “ireducible” water saturation, capillary pressure, and effective or relative gas permeability properties. The differences between these lithofucies and variance within each is ‘due to many of the factors mentioned above. For the lithofacies shown in Figure 1, porosity decreases with increasing shaliness and decreasing sorting, Permeability decreases as a function of such variables as decreasing grain size and increasing hetero- ‘geneity as a result of bedding anisotropy. Increasing clay content, associated with both increasing depositional shaliness and authi- genic clay precipitation, results in increasing “imeducible” water ‘saturation and consequently decreased effective gas permeability. It's important to note that in many low-permesbility sandstone reservoirs, lithofacies frequently vary both vertically and lterally ‘on scales ranging from centimeters to meters. These lithologic ‘variations result in distinct changes in the petrophysical properties exhibited within reservoirs on these same scales. These rapid changes necessitate the need to understand how each lithofacies ‘behaves and contributes to the whole, Commonly, reservoir mod- cls of low-permeability sandstones which reflect these variations indicate that thin, better quality intervals exert significant control ‘on total reservoir performance. A single, laterally extensive, 1-foot, thick interval of 1 md sandstone within a 0,001 md, 30-foot-thick tight gas sandstone reservoir can be the difference between & pro- ductive and non-productive well. This raises questions as to how data sampled using the standard strategy of one per foot should be hhandled statistically and how the reservoir should be modeled. Modeling the influence of variables such as lithology, burial history, and thermal history on permeability and porosity trends, and prediction of reservoir properties, may seem overwhelming, However, quantitative delineation is commonly possible using ‘multivariate linear or non-linear regression analysis in which data, for all of these variables are analyzed. This generally leads to the development of sets of multivariate equations or models (Byrnes, 1994; Bymes and Wilson, 1994), Because itis the intent of this paper to illustrate the broader petrophysical consistencies “The Rocky Mountain Associaton of GeologistsAP Byrnes between western U.S. low-permesbility sandstones and not to illustrate how lithologic differences subdivide this group, the pre- sentation of unique multivariate models for the numerous litho- facies and petrofacies is left to a future paper. POROSITY Porostis within low-permesbilty sandstones range from 0 to 22% though much ofthe original intergranular porosity has been largely oceluded by compaction resulting from pressure solution suturing and cementation by quartz and carbonate. Much ofthe present porosity is secondary intragranular porosity occurring within dissolved feldspar and chert grains and carbonate cements, Pores within leached grains or cements commonly range from 20 to 150 microns in diameter depending on grain size. Secondary pores may, however, be partly filled by authi- genic clays that donot significantly reduce the total pore volume, but reduce the maximum cross-sectional area to less than 10 microns (Sampath and Keighin, 1981). In adsition, many areas within pores and on the pore walls may contain interparti- cle and inter-clay erystal pores that are less than I micron in size, Pore throats or channels that connect these larger pores range in diameter from 1 t0 0.1 micron. These pore throats represent a small portion of the total porosity but exert strong control on permeability. Because features smaller than 1 to 2 microns are not observable using standard thin-section microscopy, direct examination of much of the connecting pore system using petto- graphic or image analysis sifu ‘Because much ofthe porosity in low- permeability sandstones is within intragranular pore spaces enclosed by a rigid frame- ‘work, porosities are not significantly changed by confining stress changes. Because ofthis, both downhole logging and routine core analysis porosity values can be close to reservoir values rors in porosity measurement arse from instrument errr itself and from minor changes thatthe ock undergoes when released from in sinw stresses. Understanding porosity distribution explains much ofthe low pore volume compressibility exhibited by these sandstones between unconfined routine Iaboratory eon- ditions and those in the reservoir. Porosity Measurement Most routine core analysis porosity data is measured on whole core or plugs using a Boyle’s Law type approach under no con- fining stress to avoid extensive equilibration times. Unconfined ‘measurements allow helium entry into the rock sample from all sides, which decreases equilibration times to several minutes compared with tens of minutes to hours for confined samples. ‘Theoretical and intra-Inboratory precision for helium porosime- try is generally near = 0.1 porosity units (pu) although inter-lab- ‘oratory comparisons show agreement of approximately + 0.25 pu. (Thomas and Pugh, 1987; Luffel and Howard, 1988). Correlation of core and log-measured porosities is complicated primarily by problems with unidentified or difficult quantify Tithologie and environmental influences on log measurements and differences in the scale of investigation, and to a lesser “degree by confining stress effects on routine core porosity values. ‘The Rocky Mountain Associaton of Geologists Insitu Porosity (%) ae Routine Helium Porosity (%) Figure 2. Cross-plot of routine, unconfined helium porosity, versus a Tboratory-measured, confined in situ porosity. In sine values are approximately 0.8 pu Tess than routine over the entire range of values. Lithologic and environmental influences on the principal poros- ity measuring logs (density, neutron, and sonic) include: matrix density variation, fluid invasion, borehole rugosity, borehole size, tool-horehole wall contact, gas effects, shaliness or clay effects, the presence of subhorizontal fractures (causing cycle skipping), the degree of consolidation, and matrix velocity variation. In ‘addition, error can result from tool limitations and drift from cor- rect calibration and depth error can result from snagging or cable stretch (Kukal, 1981). Accuracy and precision of log-measured porosities is approximately + 0.5 pu assuming # 2% calibration resolution of a 25 pu standard (Richards, 1988). ‘Scales of investigation are significantly different for logs and ccore, Core analysis porosities represent a scale of investigation of 1 to 6 inches vertically and a volume of investigation of 13 10 1,600 ce. Log analysis porosities representa scale of investigation ‘of approximately 18 to 24 inches vertically depending upon log- «ging speed, the source-detector spacing, and the sampling rte and volume of investigation of approximately 60,000 to 175,000 ce. Influence of Confining Stress For the range of porosity of reservirquality low-permeabil ity sandstones, the influence of confining stress on porosity is small and routine helium porosity values tend to bo within 952% of those inthe reservoir. To obtain a porosity value as near to reservoirconditions as possible itis necessary torestor the core to in stu stress conditions. Numerous papers have investigated the effect of confining stress on the porosity and pore volume compressibility. Some of thess include Carpenter and Spencer (1940), Fatt (1958a, b), McLatchie et al. (1958), Mann and Fatt (1960), Dobrynin (1962), Newman (1973), Newnan and Martin(1977), and Jones and Owens (1980). Unfortunately many of these papers do not describe or distinguish between rock liholo- ies. Thus, the published data exhibit significant seater due o ‘wide variations in lihologic variables In work summasized in this paper, routine unconfined helium porosity values were compared with in stu porosity values mea sured using either helium and a Boyle's Law approach or by measuring the volume of liquid expelled from a brine-satuated core during compression. For most well-consolidated, unfrac~ tured lithologies, compression ack to in stu stress levels results ina porosity value decrease of only 0.5 to 1.0 pu from zoutine values (Fig. 2). Increasing ductile or shale content increases com pressibilty and results in a greater difference between ambient and in it values tis likely that the intragranular pores undergo litte Compression and that most of the pore volume decrease ‘occurs in the thin, sheet-Hke intereonnecting pores. This is sup- ported by both the changes evident in permeability and hes lar 05 0 1.0 pu decrease exhibited by both the very low and high porosity sandstones. Assuming the pore volume ofthe intercon necting pores are similar in low and high porosity rocks, then the 05 to 1.0 puchange evident in both high and low porosity rocks represents the volume of pore space that contals flow within these rocks. In low porosity rocks the relative proportion of chin sheet-lke tabular pores to total porsiy is greater than in higher porosity sandstones. Therefore, for the same pore volume deorease stoal porosity decreases the relative perentage of the compressing pore space to total pore space increases and the per- centage change in porosity increases from values near 5% for high porosity sandstones to near 50% for very low porosity sand stones. This indieates that forthe lowest porosity rocks a high percentage of unconfined routine porosity is an artifat of core ‘decompression. However, for sandstones with porosity greater than approximately 1056, routine porosity values ae similar to, ‘or are ensily comected to inst vals, Once cores are near in sta confining stress levels (generally > 2,000 psi NES, net effective stress), addtional stress results in Tite change tothe pore volume. Compressbiity values a reser- voir conditions range from only 1*10* to 710 ® eveelps (Yol- ‘ume change/volume pore/psi change). Because of compressibil ity effects, routine helium porosity values ae always higher than ‘in situ values, For all cores studied, a simple relationship for pre- diction of in situ porosity (Oy) ftom routine helium porosity (nin 88 Shown in Figure 3, can be expressed: Gran 0S = Pie 08 For porosity values greater than approximately 10%, this equa- tion is similar to those reported by Jones and Owens (1980) and Luffel eta. (1991). However, for fower porosity values this equa- tion predicts greator than a $% decrease in porosity, ‘The relatively small response of porosity to confining stress in Jow-permeability sandstones isnot surprising given that many of these sandstones have well-cemented rigid framework systems Pores are commonly intrapartcle within dissolved feldspar or lithic grains, but may also occur as well-encasedinta- and inter- ssranular spaces. These pores experience litle compression under confining tress and hence exhibit litle pore volume decrease, at Low Permeability Sandstones, Rocky Mountains PERMEABILITY In most low-permeability sandstones, in stu high-pressure gas oF liquid permeability values range from 10 to 1,000 times less than routine air permeability values. This difference is due pri- ‘marily to the combined effects of gas slippage (Klinkenberg cor- rection), confining stress, partial brine saturation and its effect on relative permeability polar fluid-rock interaction, fines plugging, and, finally, sample handling and measurement procedures. Klinkenberg Correetion ‘Most routine core analysis methods use gas as the flowing pphase for convenience and because gas does not chemically alter the rock. To obtain equivalent liquid or reservoir pressure gas permeability values from gas flow measurements conducted at the low mean pore pressures used in the laboratory, a conection ust be made to correct for the mean-free path slippage of the ‘28s through the pores because gas molecules do not collide as ‘often as do liquid molecules, This effect is known as the Klinken- berg effect (Klinkenberg, 1941). The relationship between Klinkenberg, liquid-equivalent permeability (K,) and a routine air permeability (Kroutine) is expressed ‘where b is the slip correction factor (psi) and Pp isthe mean pore pressure (psi). The Klinkenberg effect is an inverse function of the radius of the pore, therefore, rocks with high permeabilities and large pore sizes exhibit less Klinkenberg effect than rocks having low permeabilities and small pore sizes. Klinkenberg cor- rection i also an inverse function of the mean pore pressure and the correction approaches zero as mean gas pore pressure “approaches the high pressures associated with reservoirs (.e., a8 the gas approaches a liquid-like state). Correlations of the slip factor, b, and the liquid equivalent permeability, k, take the gen- cral form: b= Gh Values for C1 and C2 generally range from 11.4to 17.8 and-0.33, 10 -0.50, respectively, as a function of the pore geometry from slit-like to citcular (Heid et al, 1950; Jones and Owens, 1980; ‘Sampath and Keighin, 1981; Ostensen, 1983; Luffel etal, 1991). Based on the general range in C1 and C2 values, the Klinkenberg correction decreases routine air permeabilities by a factor of ero to three times for permeabilities ranging from one darey down to the microdatey range respectively. Influence of Confining Stress Early studies by Fatt (1953) demonstrated the influence of confining stress on permeability. For low-permeability sand- stones, studies by Vairogs et al. (1971), Thomas and Ward "The Rocky Mouniin Association of GeologistsA.B. Byrnes: Pore Size Frequency (%) oor a + Pore Throat Diameter (um) Figure 3. Example of pore throat size distribution, as determined by mercury porosimetry, for a low-permeability sandstone sample under nominal confining stress and under simulated in situ stress. This sample exhibited an average decrease in pore throat size of 50% compared to the range forall ‘samples studied of 20% to 80%. (1972), Byrnes etal. (1979), Jones and Owens (1980), Walls et al. (1982), Sampath and Keighin (1981), Ostensen (1983), Weiet al. (1986), and Luffel et al. (1991) have shown that confining stress results in a progressively greater decrease in permeability from routine ar permesbilty values with decreasing routine pet- rmeability and with increasing confining stress. This effec is attributed primarily to the decrease in sizeof the thin, tsbular pore throats that connect the larger pores. Comparison between ‘uiconfined and confined pore throats, as determined from mer- cory capillary pressure analysis (Byes and Keighin, 1993), shows that under confining stress, pore throats of lower perme- ability sandstones decrease by a8 much as 50% to 70% (e-8- {micron o 0.5 or 0.3 microns; Fg. 3). Using a simple model of interconnected orthogonal planar eracks, permeability can be expressed as k=wL2 where k = permeability (md), w = crack width (microns), L crack length per unit area, For this model, small changes in crack ‘width due to confining stress result in large changes in perme~ ability. Assuming a decrease in crack width of 50% to 70%, as indicated by the mercury porosimetry, a change in permeability of approximately 10 to 40 times would be anticipated. This is consistent with measurements on low-permeability sandstones, Ostensen (1983) provided a more precise model relating crack asperity (roughness) to permeability. ‘Routine air and in situ Klinkenberg permeabilities were mea- sured on several hundred cores from the Rocky Mountain region, ‘The Rocky Mountain Association of Geologists a Initu Kllokenborg Permeabilty (rd) om ott Routine Ai Permeabilty (md) Figure 4, Cross-plot of in situ Klinkenberg permeability versus routine air permeability at a confining stress of approximately 400 psi. The progressively greater decrease in in sit values with decreasing routine permeability values is due to both Kinkenberg gs slip comection and to the influence of confining sitesson thin, sheet-like pore throats. ‘To messure permeabilities, the cores were placed in a Hassler- type confining pressure cell and subjected toa hydrostatic confin- ing stress of approximately 40 psi forthe routine-ar permet ity measurements. Jn sit Klinkenberg permeabilities were measured ata confining stress equal in value (ps) to half the ere depth (0 simulate in stu stresses. Klinkenberg permeabilities ‘were determined using a pressure pulse decay method. Routine air permeabilities were determined using nitrogen anda single flow ‘measurement at approximately 100 psi differential pore pressure. “The combined influence of Klinkenberg and confining stress effects becomes progressively greater with decreasing perme- ability below approximately 1 md (Fig. 4). A genera equation relating routine ai permeability (kjyjye) values t0 in situ ‘Klinkenberg permeability (kj) values can be expressed: Fay = 10124 H0 96} ‘The correlation coefficient for this relation is r2=0.88 and the standard error of prediction is a factor of 3.2 (ie., a predicted valve of 1 md may be 3.2 md or 0.32 ma, 1 standard deviation). Scatter in Figure is the result of both experimental error and the influence of lithologic and burial and thermal history variables. Different lithologies respond differently to confining stress. Well- ‘cemented sandstones may exhibit litle confining stress effects at permeabilities above 1 md, but with greater cementation and development of thin tabular pores exhibit increased influence. ‘Sandstones with increased shale laminae also exhibit enhanced effects due to greater compression of the higher ductile compo- ‘nents under confining stressLiquid Permeability An important characteristic of low-permeability sandstones is that they produce natural gas with very litle to no water produc- tion even at high water saturations. Ibis also commonly observed ‘that liquid permeabilities are significantly less than Klinkenberg bas permeabilities by a factor of three to ten times. The low ‘mobility of water is attributed to significant fluid-rock interaction resulting in clay swelling and fines migration and plugging Jones dnd Owens, 1980; Luffel et al., 1991). Wei et al, (1986) reported that liquid permeabilities decreased with increasing {uid polarity. They concluded from a series of tests involving dif- {erent Muids und ultrasonication of samples that much of the dif- {erence between gas and liguid permeabilities was due to disper- sion of fines particles and fines plugging Relative Permeability ‘The discussion above illustrates the influence of ga stippage and confining stress on the absolute gas permeability, that is, the single phase permeability to gas when no other phase, such as brine, is present. Many exploration and production decisions are based on absolute permeability data. However, no gas reservoir contains only gas in the pore space. As brine occupies a progres- sively sreater percentage of the pore space, te interference of gas and water flow within common flow channels esults na decrease ofthe permeability to gus and an increase to that of brine. tis often more important to know the effective permeability to either 248 or waler(elative permeability multiplied by absolute perme ability than the absolute because itis this permeability that con- trols actual production. Effective gas permeabilities at “rre- ducible” water saturations for many sandstones with absolute permeabilities below approximately 0.01 md are commonly one to two orders of magnitude less than the absolute valve. Figure 5 illustrates the approximate range of relative gas per- eability curves for low-pormeability sandstones. These steep declines represent the approximate range of curves measured by the author and also contain those that have been reported in all Jow-permeablity sandstone studies that investigated relative per- ability Thomas and Ward, 1972; Byrnes eta, 1979; Jones and Owens, 1979; Walls ea, 1982; and Ward and Morrow, 1985). It evident that gas permeability decreases rapidly at saturations ahove approximately 40% to 50%. Hence, a sandstone exhibiting 10.001 md single-phase gas permeability would exhibit an elfec- tive gas permeability of only 0.0005 md at a reservoit “inre- dlucible” water saturation of 40%. Gas production from this rock ‘would still be dry because the brine relative permeability is several orders of magnitude lower than te absolute permet The curves shown in Figure 5 illustrate gas permeability decreases with increasing water saturation fora given rock sam- ple. These curves are important in interpreting what happens to ‘8 flow if water saturation is inereased due to water invasion ‘rom driling or stimulation It is clear that if water saturations are increased inthe near wellbore or fracture region, the ability to dis- place the additional waters significantly hindered by both adras- tically diminished pas permeability, as well as the already low water permeability. Sandstones with permeabilities below a Low Permeability Sandstones, Rocky Mountains 100 g 001 ‘Gas Relative Permeability (Kegika, %) 0001 2.00001 © 10 20 90 40 60 60 70 #0 90 100 Brine Saturation (% pore volume) Figure 5. Generalized relative gas permeability curves for low- permeability sandstones, Relative permeability decreases rapidly for water saturations grester than approximately 40% to 50%, ‘They are also lower in more shaly sandstones than in fine- to medium-grained cleaner sandstones for the same water saturation (0.01 md and with greater clay or shale content are prone to exhibit the lowest relative gas permeability for a given water saturation and are, hence, the most difficult to “clean up.” Under certain cap- illary pressure conditions, discussed below, some relative perme- ability conditions can become stable and diminished gas flow can become permanent until capillary conditions are changed. ‘While the curves shown in Figure 5 illustrate relative perme- ability changes with saturation within a given sandstone sample, in many low-permeability sandstones, water saturations are at or sducible” saturation levels, Production of gas from many reservoirs involves primarily gas pressure depletion with litte or ‘no water invasion or production, Its, therefore, most important ‘to understand the trends for effective gas permeability at or near “irreducible” water saturation levels. Effective gas permeabilities ‘were measured on several hundred cores that contained brine at “reducible” saturations (procedures are described below). Fig- ture 6 shows the relationships between relative gas permeability at “ieducible” brine saturation and the “irreducible” brine satu- ration at which each permeability was measured. Similar in form to the relative permeability fora single core, relative permeabili- ties decrease as “ineducible” water saturations increase. As dis- cussed below, “irreducible” saturations increase with decreasing absolute permeability, thus, the lowest relative permeability val ues are for the lowest absolute permeability sandstones, Although low-permeabilty sandstones may exhibit high “ire , Dobryin, VM, 196, Beto overburden pressure on some properties of sand- Sones: Society f Poleum Engincers ural, December. 360.366 The Rocky Mountain Associaton of Geologists 48 Duton, $F, $4. Cit, DS. Honiton, HS, Hamlin, LF Hens, WE. Howard, MSS. Abie, and SE Laubach, 1998, Major low-pesmeabiitysadtone 8 eservirsin the continental U.S: Buren of Economic Gooly, Report €Ivesigations No. 211, p. 221 Fat 1958, Pore volume compresses of sandstone reservoir rocks: Peto eum Transactions AIME, Technical Nowe 200, 213, . 362-364, Fat, 1988, Compressibily of sandstone at ow to moderate pressures: AAPG Ballin, v2, 08.1928 Heid, 16.13. Meio, RF. Nilsen, and 8, Yost, 1950, Stay ofthe pe- ‘meablty of reks to homogencous Huds: API Driling snd Production Practices. 230-246 Jove FO, and W.W. Owens, 1980, A abosary study of lou: permeabiity gas Sands” Foumal of Peoleum Technology, September, p 1631-1640. ‘Klinkenberg 194, The permeability of porous media wo liquids and pase: ‘APLDling and Production Practices. 200213. Kula, G.C, 1981, Determination of fd omected porosity in ight gos sds ‘nd infrmaions exhibiting shallow invasion profiles: SPE Paper 9856, Proceoings ofthe 1981 SPFIDOE Symposium on Low Permeability Gas Reseris, Denver, CO, May 27-29, p. 289-200 Luff, DL, and WEE Howat, 1985, Reliability nf laboratory neasurement of. ores in tight gas sands: Seccty Petroleum Baginees Formation Evalu- in, Decomber,p 708-710. Laff, Dil, WE. Hosoda ER Hont 1981, Tes Pak core permeability ‘std oro selationships a eserwoe sre: Society Petroleum Engineers Foenation Balaton, September, p. 310-318, ‘Mana, RL, sad I Fat, 196, Efecto pore Huds onthe elastic popesies of suadsone: Geophysics, 25,No, 2, p. A334, MeLatchie, AS, ReA, Hemtock, and FW. Young, 1958, The effective com- ressibiity of reser roek dis flees on permeability: Teal Nok DOI, Transatons of AIME, v.23, p. 386-388 Newman, GH, 1973, Pore-volume compessibly of consoldited, abe nd unconsolidated eservoir rocks under hydrostatic Toaiing: Joumal Peoleum ‘Tecnology, Feb, p. 129-138, ‘Newnan, G4 and LC Marta, 1977, Bgupment and experimental metas for ‘obtaining laboratory compression characteristics of reservoir rosks ner ‘various sess and peesure conditions: SPE Paper #6835, 52nd Annual ‘Tecinial Conference of SPE, Denver, Colorado, Ocabe 9.12, 16, (Osteasen, RW, 1983, Mirorack pemeaiity i ight gas sandsione: Society ‘of Petroleum Engineers Journal, Dec, p. 919921. Piuman, .D., 1992, Reaionship of porosity and permeability to varion par ‘meters derived from mercury injeton-cpillry presse cues for sand stone: AAPG Ballet, 76, 0. 2,p. 191-198 Richa, LM, 198, Log Quality Coto: International H opment Comp.398 9. Sampath, Kan CV. Keighin, 1981, Factors affecting 288 slippage in tight Ssandsones: SPE Paper 9872, Proceedings of the 1981 SPEIDOE Sympo ‘um on Low Permeabity Gas Reservoirs, Denver, Colorado, May 27-29, 400.416. ‘Thomas, D.C, and VJ. Pugh, 1987, A statistical analysis ofthe accuracy sed epreducblityof standard core analysis: Proceedings 1987 Society of Core ‘Analysts Conference Paper Numer 8701, 7p. ‘Thom, RD, and D.C. Ward 1972, fect of overburden pres and water st ‘raion on gs petmeability of tight sandstone cores: Joual Pevoleum ‘Technology, Fetrry,p. 120-124. aiops J CL Hear, D. W, Daeg, and V.W. Rhoades, 1971, et of ek stress on gs production from low-permesblityroeks: Jounal Petroleum ‘Teetnology, September, p. 1161-1167 ‘Walls, 1.D.. AM Nur and Bourbio, 192, Bets f pressure and partial water ‘Stuzton on gas pemeability tight sands expeimental rest: Joumal ‘of Petroleum Technology, April p. 930-956 ‘Waud, 1S. and NR. Mogow, 1985, Capillary presure and gus rave penne ables of low permeability Sandstone: SPE paper 13882, Proceedings 'SPEJDOE 1985 Symposium on Low Parmeabiity Reservas, Denver, May 9.12.p Wi, KIN Mowow; and KR. Brower, 1986, Ect of Mud conning press, and tempera on absolte permeabilities of lw permeability sandstones Society Petoleum Eagincers Formation Evaluation, August, p 413423. nan Resource Devel