POOO3-96

Determination of Initial Fluid Saturations -A Key Factor in By-Passed Pay Determination

D.B. Bennion, F. B. Thomas, R. F. BietzHycal Energy Research Laboratories Ltd.

Abstract Introduction

Proper evaluation of initial water saturation is essential forproper reserves evaluation, informed decisions on which zonesto complete to obtain water-free production, and influences avariety of productivity and formation damage issues. The mostcommon methods of initial water saturation determination,electrical log resistivity measurements and direct saturationmeasurements on in-situ core samples, are described and theadvantages and disadvantages of these techniques discussed inthis paper A study presented illustrates the effect of a varietyof properties, such as saturation exponent (n), Archie exponent(a), cementation exponent (m), water resistivity (Rw) andformation resistivity (~) on water saturations calculated fromelectrical log data. It illustrates that some formation valuesmay significantly deviate from the standard assumed values for"a", "mH and "nH of 1, 2 and 2, and that this can significantlyover or underestimate the value of the initial water saturation.Various techniques for the correct detennination of correct HaH,Hm" and HnH values in a controlled laboratory environment foractual reservoir rock from the formation under consideration,for a greater degree of confidence in log calculated watersaturations, are described.

Initial fluid saturations, the fraction of the interstitial space ina pore system occupied by oil, water and gas, are key factorsin determination of initial reserves and dominate reservoir flowproperties due to the influence they exhibit on relativepermeability. Surprisingly, in many cases, initial fluidsaturations are virtually unknown or improperly measured,resulting in gross over or under estimation of oil or gasreserves in place and greatly affecting potential for formationdamage due to phase trapping, resulting in a poor appraisal ofdeliverability. In many cases, improper saturationmeasurements can lead to the bypassing of potentiallyproductive pay zones resulting in significant l~t reserves, orthe erroneous completion of ineffective pay resulting in lostrevenue.

The Hazards of Insufficient Knowledge of Initial FluidSaturations

Improper detennination of the initial oil, water or gassaturations which exist in porous media may often lead toexpensive mistakes in the development of a field. In somecases, large amounts of capital are invested where minimalreserves are present, or marginal flow is obtained. In othercases, viable pay is overlooked due to a perceived belief, fromimproper saturation evaluations, that the pay will be wet ornon-productive. The hazards involved with an inadequateunderstanding of initial saturation conditions can generally begrouped into three categories:

Various types of coring programs, including low invasioncoring, traced fluid coring and analysis techniques, spongecoring, gel coring and sidewall coring are described. Theadvantages and disadvantages of each method and techniquesutilized to obtain the most accurate possible initial saturationdata are discussed.

2 DETERMINAnON OF INmAL aUlD SA11JRAnONS . A KEY FACfOR IN BYPASSED PAY DETERMINAnON P0003-96

Poor Initial Reserves Evaluation. Oil or gas in place isbased on a simple volumetric calculation of hydrocarbonvolume present in the effective porosity of the system. Sincethis value is usually assumed to be I-Swi' it can be seen thatover or underestimation of the initial water saturation cangrossly affect the perceived amount of oil or gas in place inthe reservoir. An underestimation of Swi could result in costlydevelopment of a field which has much less oil or gas in placethan anticipated, resulting in limited or negative return onultimate investment. Conversely, if Swi is overestimated, apotentially viable and highly lucrative pay zone may beabandoned.

program penetrated effective pay, and a multi Bcf gas field,which had been overlooked for 20 years, was brought ontoproduction.

Poor Completion Zone Selection. Figure I illustrates atypical set of water-oil or gas-water relative permeabilitycurves. It can been seen that underestimation of ilie initialwater saturation may result in completion of zones wiili highpotential water phase relative permeability which will result inimmediate high water cuts and poor oil or gas production rates(i.e. ilie water saturation was believed to be at point II A H,when, in fact, it is actually at point "B" on Figure I).Conversely, over estimation of initial water saturation mayresult in the decision to not complete a zone which couldconceivably produce at economic rates, resulting in lostopportunities and revenue (i.e. ilie water saturation was ilioughtto be at point HB" on Figure I in this case when it was actuallyat point H A H).

A recent example is a field where log derived saturation valuesindicated water saturations of between 40-50%, yet uponcompletion of the wells, they surprisingly produced no freewater. Subsequent traced coring programs indicated that the logcalibration constants were in error and that the true in-situwater saturation was, in fact, in the range of 20-22%, resultingin a 56% increase in the amount of oil present in the field anda significant alteration in the economics of the project.

l41ow Mechanics. Multiphase flow is governed by the relativepermeability which exists between the individual phases.Relative permeability is, in turn, governed directly by therespective saturations of the individual fluid phases which arepresent in the porous media. Proper knowledge of both therelative permeability characteristics of the porous media (whichare determined by appropriate experimental tests in thelaboratory) and the initial saturation conditions will allowinflow calculations to determine if, in the presence of mobileor immobile water or gas saturations, economic productionrates of oil or gas will be feasible.

Many very low penneability gas reservoirs exist in conditionsof HsubirreducibleH initial water samration6.7.8. If relatively littlefluid invasion has occurred during the drilling and completionprocess, this can often be detected by significant separationbetween the shallow and deep induction logs and extremelyhigh resistivities in the deep, uninvaded portion of theformation (sometimes in the hundreds or thousands ofohmmeters). Due to the adverse capillary nature of these verylow permeability systems, they tend to both spontaneouslyimbibe and pennanently retain invaded water based filtrate inwhat is known as an Haqueous phase trapH or "water block:'8.If deep invasion occurs, this can result in the creation of azone of extremely high water samration about the wellbore,which may be misinterpreted as wet. Some tight gas sandswhich are subirreducibly saturated may exhibit trappedsaturations of 70% or greater, when the initial samration maybe actually be quite low (perhaps 10-15%). In this situation,both reserves and productive potential may be grosslyunderestimated. An example of this nature was a shallow tightsand zone that had been routinely penetrated for years whiletargeting deeper oil bearing strata. Conventional gel chemicalbased muds with high fluid losses were typically used to drillthe uphole sections which had inevitably resulted in significantfluid losses to this zone. Gas shows were common whenpenetrating the horizon, but subsequent logging showed highwater samrations (due to deep invasion) and testing indicateduneconomic or no gas production rates. Close evaluation oflogging suites, after many years of drilling through this zone,suggested that a zone of low water saturation, and potentiallyhigh gas productivity and reserves existed beyond the nearwellbore damaged region. A gas based (COJ fracturing

Saturation Determination Techniques

Initial saturations are commonly detennined using thefollowing techniques:

. Log based saturation evaluations

. Direct saturation measurements on in-situ samplesBoth techniques have pros and cons.

Log Based Saturation Evaluations. Water saturationdetenninations in clean (non-shaley) formations are based onArchie's equation.:

S; = (FRw) /R, (1)

wheres -wn-

R.-Rc-F-

Fraction of pore space occupied by waterSaturation exponentFormation water resistivityTrue water resistivityFormation factor

The fonIlation factor is given by the equation:

p(XM))-96 DB. BeOOON. F.B nIOMAS. R.F. BI£IZ 3

computed water saturation can vary radically. Lower "0" valuesthan the standard 2.0, usually used, will result io lower Swivalues (8.2% SwI' for example, at an "0" value of 1.0) andhigher "0" values. which are often ~iated with thetransition to a more oil wet state of a unifOrDl pore system.will cause increases in true Swj (to 73%. for example, at an "n"value of 8). The data of Table 2 bas been plotted forillustrative p~ and appears as Figure 3.

(2)F = alc/l"

wherea -m-4> -

Archie constantcementation exponentPorosity fraction

Therefore, combining equations I and 2 yields: Accurate detennination of the saturation expooent value istherefore of essential importance for the calculation of "trueHinitial water saturation from logs. Saturation exponents arenorD1ally deterDlined experimentally in the laroratory on coresamples of die actual formation under consideration to obtaina more realistic evaluation of die "nH value. Due to the factthat the Hn" value varies with both lithology and wettability, asuite of "nH value measurements is often conducted on samplesfrom a range of permeabilities, porosities and lithologies whichmay be present in the formation. Most laboratory HnH valuedeterminations have been conducted in the past using waterand gas on cleaned core samples using the base assumptionthat all formations are water wet. This is a grossoversimplification as many researchers 3.4-' have demonstlated

that, on average, only approximately 50% of sandstone oilbearing formations and less than 10% of carbonate oil bearingformations exhibit strongly water wet behaviour. This ispossibly why the standard Hn- va1~ of 2.0 used so widely inthe industry has gained acceptance, as die vast majority of themeasurements conducted have used cleaned, artificially waterwet, porous media which typically do exhibit saturationexponent values near 2.0.

S; = (aRw) / (~Rt) (3)

Potential saturation determination errors may be associatedwith variations in:

Saturation Exponent Value. The saturation exponent value isa function of both pore system geometry and fonnationwettability. Although a value of 2.0 is commonly assumed forthe saturation exponent. this value can vary comiderably fromformation to formation and may result in over or underestimation of water saturation in many situations. Since thesaturation exponent is derived as the slope of a plot ofres~vity index vs water saturation, it assumes that acontinuous slope is obtained (i.e. the relationship is a linearone on a semi-logarithmic plot). This assumption is generallysatisfied for water wet porous media, where in the wetting,water saturation fonDS a cootinuous film on the surface in theporoos media, allowing conductive paths for currenttransmission even down to very low saturation levels.Conversely, for oil wet porous media. the water saturation iscontained in discontinuous droplets in the central portion of thepore space. Saturation exponents for strongly oil wettedsystems generally are similar to water wet systems with valuesaround 2.0 at high water saturation levels, but as watersaturation is reduced, the slope of the RI/S. relationship canincrease radically. Saturation exponent values as high as 8-10have been measured at low water saturation conditions (whichapproximate the initial reservoir saturation condition we areattempting to detennine through logging) in strongly oil wetsituations. Figure 2 provides a comparative illustration of theeffect of variation in -n- value with wettability. If too low ofa saturation expooent value is used, it will result in asignificant underestimation of initial water saturation andpotentially the erroneous completion of wet zones in themistaken belief that they cootain mobile oil or gas saturatiom.

Saturation exponent measurements are conducted in thelaboratory using the following four methods:. Bulk porous plate method where a large group of samples

are simultaneously desaturated from a condition of 100%water saturation on a large porous plate to generate a rangeof gradually lower water saturation levels. Successivelyhigher gas pressures are used to achieve this desaturation.After each higher pressure, the apparatus is disassembledand the discrete samples are removed and subjected toindividual resistivity measurements. This technique suffersfrom deficiencies, due to gntin losses and capillaryredistribution of the flui<k, because of the necessity tocontinuously be removing and re-inserting the samples inthe porous plate cell as well as fluid losses due toevaporation by continual cyclic disturbance of theequilibriwn envirooment.

Table 2 provides a illustrative set of data indicating how thevalue of the sabJration exponent can alter the calculated valueof the initial water saturation. Using the fixed -standard-values given in Table 1 for a. m, D, R. and Rr. the -true-

water saturation value is calculated at around 28.7%. Holdingall other parameteIS constant and varying only the "n" value,one can see that, in the range of -n- values from 1.0 to 8.0, the

. Centrifuge methO<k, where samples are spun at graduallyincreasing speem to reduce the water saturation andgenerate a range of saturation values at which resistivitycan then be evaluated. Like the bulk porous plate method,these techniques suffer from core disturbance, grain loss

4 DETERMlNA1l0N OF INl11AL FLUID SA'nJRAllONS - A KEY FACTOR IN BYPASSED PAY DEJERMINA1l0N P0003-96

and evaporation effects which can compromise dIe accuracy ofthe obtained saturation exponent data.

calculated water saturation as the value of "a" changes. Thisdata has been plotted and appears as Figure S. Once again thefixed parameter data of Table I is utilized holding allparameters constant except for the "a" value. Examination ofthis data indicates that variations in "a" alter the watersaturation values, but not to the extent that variations in "n"did previously. Variation of "a" from 0.4 to 1.7 resulted in arange of water saturations from 18.1 to 37.4 %, indicating thatsignificant departure from the value of 28.7% calculated witha typical "a" value of 1.0 could sti1I occur.

Individual porous plate methods consist of special cells inwhich single core samples are mounted with individualporous plates and electrodes attached to the samples tomonitor in-situ resistivity. A micro pipette is used to trackfluids produced from the core as capillary pressure isincreased to cause the water saturation reduction, and hencethe entire test can be nm without disrupting the core sampleby removing the applied pressure which can cause phaseredistribution within the core and eliminating the potentialfor grain l~es and evaporation. The technique has theseadvantages over the bulk porous plate and centrifugemethods described previously, but suffers from thedrawback of being a time-consuming method, particularlyfor samples of low permeability, where literally monthsmay be required to obtain an effective desaturation.

Formation Water Resistivity. This is an extremely importantfactor in the calculation of initial water sanuation as the ioniccomposition of the water affects its overall electricalconductivity and hence resistivity. Fresh waters exhibit highresistivity, while highly saline brines are extremely electricallyconductive, and hence exhibit low resistivity values. Logresistivities in die near wellbore region are complicated by thefact dIat during most overbalanced drilling processes, losses ofmud filtrate to the near wellbore region occur. This invadedflltrate is generally of a different composition than die in-situformation water (usually fresher in most cases). This results ina zone of invasion of altered resistivity about the wellbore, atransition zone of blended invaded and in-situ water and thennon-flushed reservoir. This phenomena is complicated withtime as dispersion continues due to capillary forces, generallyresulting in an outward imbibition of flltrates from thewellbore into the formation. Blending of fluids duringproduction will also alter the complex configuration of thefluid invasion patterns. Therefore, significant madIematicalmanipulation of the resistivity data may be necessary to obtaina hope of obtaining an accurate sanuation profue, particularlyin the case of deep invasion. Such discussions are beyond thescope of this paper, but are contained in the literature9.

The continuous injection technique is a relatively newmethod for saturation exponent determination and consistsof mounting the sample in a pressurized core cell with aporous plate at the outlet end of the core. A micro injectionpump is used to inject non-wetting phase into the sample(oil or gas) at an extremely low rate to slowly reduce thewater saturation in the porous media over a period of time.Resistivity measurements are obtained on a continuousbasis and average fluid saturations are obtained by anaccurate material balance of fluid produced from the coresample. In this manner, a complete RI vs Sw curve withliterally hundreds of points can be constructed, covering theentire range of saturations and detecting any subtle changesin saturation exponent with water saturation. The techniquehas all the advantages of the individual porous platemethod, plus it is generally much faster and can easily beconducted at full reservoir conditions of temperature,pressure, overburden pressure and using live reservoir fluidsif required.

Fonnation water resistivities are commonly detennined usingthe following techniques:

Figures 4a-4d provide schematic illustrations of the bulkporous plate, centrifuge, individual porous plate and continuousinjection techniques respectively.

Archie Constant Value. Archie constant values are measuredexperimentally for a series of formation factors determined ona range of porosity value samples for a given lithologyexpected to exist in a specific formation. Generally, the lowerthe degree of consolidation, the lower the value of the Archieconstant. A value of 0.62 for the Archie constant was derivedby Humble for poorly consolidated sands. The value tendstowards 1.0 for compacted sands and may exceed 1.0 as thedegree of compaction becomes extreme.

Direct Measurement. If produced fonnation water samples areavailable, these can be used to obtain an exact measurement ofresistivity. Since resistivity is a strong temperature function,these measurements must either be conducted at reservoirtemperature, or corrected for temperature effects. One mustensure that the water samples produced from the fonllation are,in fact, representative of the true in-situ water. Deviations mayexist due to:. Contamination from invaded drilling, completion or kill

fluids which are still being produced back from thefonnation;

. Concurrent production of water from different zones ofvarying water composition due to multiple flowing zones orpoor cementing;

. Dilution of produced water by fresh water of condensationfrom the gas phase if high gas rates are present;Table 5 provides an indication of the expected variation of

P0003-96 D.B. BENNION, F.B rnOMAS, R.F. BIE1Z ~

Presence of a "dual water" system in the reservoir wherethe mobile water saturation has a different salinity andcomposition than the bound water locked into themicroporous clay fraction present in the porous media.

migration and desiccation effects, this results in the trappedlow water saturation exhibiting abnonnally high salinity forwhat may be typical in the regional trend. The use of nonnalcatalog Rw values in a situation such as this may lead to asignificant overestimation of initial water saturation andbypassing of potentially productive gas bearing strata.From Water Catalogues. Often extensive lists of produced

waters from various formations are compiled and arecommonly used by companies when they are evaluating similarzones in the same regional trend. This technique hinges on theassumption that the resistivity measurements conducted on theoriginal samples which are in the catalog are correct and alsothat the formation under consideration has no variation inwater chemistry from the one used in the catalog as the datasource. Verification of the Rw value from the SP curve or fromresistivity-porosity comparison is recommended in such asituation1o.ll.

Cementation Exponent. The value of the cementationexponent is highly dependant on the degree and type ofcementation in the pore system. Although a value of 2.0 iscommonly used, this value can vary highly as is illustrated inFigure T5. For poorly cemented rocks m maybe less than 2.0and in highly cemented or oolicastic rocks "m" values as highas 3.0 have been observed.

Table 5 provides a summary of how variations in "m" valueaffect calculated water saturation. This data has been plottedand appears as Figure 8. It can been seen that variations in "m"also influence water saturation calculations, albeit in asomewhat weaker fashion than "n" and Rw variations discussedpreviously. Low "m" values around 1.0 result in a reduction inwater saturation to 13.4% in comparison to a 52.6% saturationcalculated at a "m" value of 2.8. This compares to the watersaturation of 28.7%, which would be calculated using thestandard "m" value of 2.0.

From Chemical Analysis. In some cases the chemical analysisof the water is known, either from a catalog or a historicalsample. Various techniques have been detailed in die literaturefor calculating resistivity based on compositional analysiSI2.13.14.One must be aware that die composition of the water, beingused as die basis for the resistivity calculation, may be affectedin a similar fashion as described in the section "DirectMeasurement" for the direct measurement of resistivity. If theanalysis is in error this will obviously compromise theaccuracy of the calculated resistivity. Both the Archie constant and the cementation factor are

detemlined by lab measurements of formation factor (thisbeing the relationship between the resistivity of a rocksaturated 100% with water and resistivity of the water itself).Formation factor measurements are usually conducted on plugsamples in the laboratory. Obviously, as can be seen fromequation 2, the fonnation factor is influenced strongly byporosity. For this reason, the test sequence is usually conductedon a range of porosities to obtain as accurate evaluation aspossible of the relationship between a, m and porosity. Sinceporosity is influenced by overburden pressure (which willchange the pore geometry and hence alter the resistivity of the100% brine saturated matrix) formation factor measurementsshould always be conducted at the net effective reservoiroverburden pressure. In addition. for some reservoirs, theformation factor, and hence the values of a and m, may changesignificantly as overburden pressure changes. This means thatduring a typical reservoir depletion process, where reservoirpressure is dropping, that the net effective overburden pressureis increasing. Hence, a different set of a and m values may berequired for accurate field water saturation evaluation purposesas depletion proceeds. For this purpose, most sets of fonnationfactor measurements are run at a series of increasingoverburden pressures to facilitate the inclusion of this pressurevariation in the reservoir as a function of time so that accurateevaluations of initial water saturation in newly drilled wells inpartially depleted zones can still be obtained.

Resistivity from the SP (Spontaneous Potential) Log. Goodvalues of Rw can sometimes be obtained from the SP curve ifthe formation is clean and non-shaler since the static SP valueis related to the chemical activity of the formation water andmud filtrate.

Table 4 provides a summary of the effect of a variation of R,.,on calculated water saturation over a range of R,., values from0.01 to 7 ohm.m. This data has been plotted and appears asFigure 6. Water resistivity very strongly influences the valueof the log calculated water saturation. Using the baselinevalues given in Table 1, which provide a water saturation of28.7% at a Rw value of 0.1 ohm.m, it can be seen thatreducing resistivity an order of magnitude to 0.01 ohm.mresults in a reduction in water saturation to only 9%.Alternatively, as Rw increases, computed water saturation soars(rapidly to unrealistic values greater than the entire poresaturation of 1.0, as can be seen from Table 4 and Figure 6).

A major challenge in subirreducibly saturated gas bearingformations and strongly oil wet oil bearing media is the factthat due to the abnormally low initial water saturation in thesecases there is no produced mobile liquid saturation to evaluatefor accurate Rw measurements. In the case of subirreduciblysaturated tight gas sands, where the mechanism of the lowwater saturation establishment is believed to be regional gas

6 DFfERMINAllON OF INI11AL FLUID SATURAllONS - A KEY FAcroR IN BYPASSED PAY DFfERMINAllON P0003-96

Determination of Rr The basic assumption of the calculationof water saturation from resistivity logs is that the rock andany other fluids it contains (oil and gas) are perfect insulators.Oil and gas generally fulfil this requirement nicely (as they areoften used for electrical insulators in transfonners and othersituation due to their non-conductive nature); however, certainminerals, which may occur naturally in some fonnations, havea degree of electrical conductivity. These would includematerials such as pyrite (FeS or "fools gold"), galena orchalcopyrite. These materials, even when completely dry, havehigh conductivity. If present in sufficient quantity in thereservoir they can significantly lower overall resistivity. Thiswould lead to a belief that the zone under consideration is wet,when it may in fact be at an extremely low water saturationand produce water free oil or gas. Several examples of this arepresented in the literaturel6.

log calibration constants to other wells in the same field(which may not have been cored). Obviously, core from anunswept zone (by waterflooding or coring) is a requirement inthis case.

Challenges which face die operator in designing a coringprogram to obtain representative initial water saturationsinclude:

Obtaining Core Material Which Has Not Been Flushed Bythe Coring Fluid. In most situations, the objective is to obtainthe initial water saturation which exists in the porous media.Typical coring operations, partiCUlary in higher permeabilityporous media, can result in significant flushing of the obtainedcore material by the coring fluid. If the coring fluid is waterbased, this will obviously result in an undesirable increase inthe measured water saturation. If the coring fluid is oil based,the initial water saturation will be unaltered, if it is at asub irreducible or irreducible value (ie - immobile), but if thewater saturation is mobile (such as in a transition zone oraquifer zone) oil based mud fIltrate may flush the core andreduce the obtained water saturation. Oil based coring fluids,therefore, usually provide a good estimate of the initialirreducible water saturation which exists in a reservoir, butmay underestimate the actual water saturation which exists insome zones if mobile water is present. Gas has also been usedas a coring fluid Due to the poor heat transfer capacity of gas,a large amount of heat is generated during the coring process.This heat, combined with the dehydrated nature and high rateof gas circulation required to clean the hole, often results indesiccation of the core and artificially low water saturationsbeing obtained.

Clays and shales may also contribute to formation conductivity .These materials exhibit natural conductivity due to the ion-exchange process which exists as a function of the activeexchange sites on the surface of the clay particles. The effectof shaliness on shaley sand conductivity is oftendisproportionately large in comparison to the amount of shalepresent in the sand and depends on the amount, type anddistribution of the shale and the composition of water trapped.Various correlations are presented in the literature to correctlog responses for the effect of conductive shalesI7.18,19;1l},21.

Table 6 illustrates the effect of varying RT on calculated watersaturation using all other parameters fIXed as given in Table 1.This data has been plotted and appears as Figure 9. Obviouslywater saturation increases with reductions in ~. It isinteresting to note how a relatively small reduction in ~, suchas might be caused by the presence of conductive materials inthe rock matrix, from 25 to 15 ohmmeters, causes an increasein the apparent calculated water saturation from 28.7 to 37.1%. This may be the difference between mobile and immobilewater in many situations and a key factor in location of apreferential zone for completion.

Flushing of the obtained core during the coring process can besignificantly reduced through the use of a "low invasion" corebit. This technology is relatively new, but is becoming quitepopular and most coring companies now offer some variant ofa low invasion core tool. The jet orientation on the bit isslightly different on a low invasion core bit, such that a vortexof high velocity and low pressure is created at the outercircumference of the bit which creates a low pressure zonewhich tends to draw the coring fluid away from the surface ofthe core as it is cut. The location of the inner core barrel isalso altered, so that it core immediately enters the core barrelafter cutting, minimizing the time in contact with thecirculating coring fluid Figure 10 illustrates the mechanism ofa low invasion coring process. The technology has beendocumented to reduce core flushing by over 90% in certainapplications and is definitely recommended for any coringprogram where in-situ saturation determination is an objective.A variant on low invasion coring is "gel" coring. Thistechnology marries a low invasion core bit with an extremelyviscous water based gel in the core barrel. As the core istaken, the viscous gel is extruded in the annular space betweenthe core and the core barrel. This isolates the core from contact

Direct Saturation Determination

It can be clearly seen from the preceding discussion that,although log techniques are commonly used to evaluate initialwater saturations, there may be significant uncertainty withrespect to the a~olute validity of the numbers obtained,particularly in the absence of any accurate actual measurementsof the electrical properties of the rock under consideration.

Another, potentially less ambiguous and more accuratetechnique is obtain a direct measurement of fluid saturationsfrom in-situ core samples obtained from the formation duringthe initial drilling process. This data, if properly obtained, canprovide representative fluid saturation measurentents which areoften used to calibrate field logs and then apply the corrected

P0003-96 D.B. BENNION, F.B rnOMAS, R.F. BIETZ rwidl dle drilling fluid, plus generally seems to improve corerecovery in some highly laminated san& or fractured zones bydle stabilizing influence of dle encapsulating gel. Somedifficulty can be encountered when trying to remove dle corefrom dle core barrel after a gel coring operation, but theseproblems can generally be handled widl experience and propertechnique.

So called radioactive tracers commonly include isotopic formsof water, these being deuterium and tritium. Deuterium, orheavy water as it is sometimes referred to, is inert and easy touse. Since it is an isotopic form of water it is removed fromthe core during a conventional extraction process and is notsubject to many of the problem associated with chemicallybased tracers. Disadvantages of deuterium include irs expense(typically approx $IOOO/litre and up to 20 litres may berequired to dope a typical mud system to a concentration of250-350 ppm), the fact that mud systems are doped in fairlylow concentrations, which decreases the detection limitsensitivity, and the point that deuterium is a stable naturalisotope of water and exists in most formation waters in aconcentration of 120-150 ppm. This natural preexistingconcentration is not an issue if produced water SaDlples fromthe field are available to allow an accurate baselinedetennination, but if samples are unavailable (such as in manylow initial water saturation formations) accuracy of the methodmay be degraded for low levels of invasion due to uncertaintyas to the baseline concentration of preexisting deuterium in thenatural formation water.

Some possibility of invasion, particularly in zones of highpenneability, exist even when low invasion coring is used. Forthis reason, the water saturations evaluated from a low invasioncoring process using a water-based mud may still be inquestion. This can be remedied by the use of some type oftracer in the coring fluid. If the tracer is present in the waterremoved from the core samples, it can be ascertained thatinvasion has occurred and. if flushing is not complete, theconcentration of the tracer in the central section of the core canbe used to back out the degree of invasion and hence calculatethe true initial water saturation which existed in the porousmedia. The tracer is generally some element which is easilydetectable in low concentrations and not naturally present inthe formation water. Baseline samples of drilling fluid aretaken from the circulating drilling fluid stream regularly duringthe coring process to obtain a "reference" value of the traceras a function of depth. This is due to the fact that tracerconcentration can vary due to dilution with fonnation fluids oradditions of fluid to the mud system to account for upholelosses or losses to the fonnation. Two classifications of tracersare nonnally used, these being chemically based tracers andradioactive tracers.

Tritium is less expensive than deuterium (a typical mud systemcan be treated generally for about one-tenth of the cost), isanalyzable in concentrations of down to 1 picocurie/ml (pci/ml)(in comparison to the doping concentration of 5000-10000pci/ml) and does not naturally exist in formation waters in anysignificant concentration so that the need to have a sample ofin-situ water for a baseline analysis is negated. Tritium analysisis also quick, less problematic and less expensive incomparison to deuterium analysis. Tritium is a beta emitter,with a half life of 12.3 years and concerns are often expressedabout the safety of it's use. Drinking water standard for tritiumis 1000 pci/ml and disposal standard is 1,000,000 pci/ml, soone can see that the concentration in which mud systems arenormally doped represents no hazard to humans or theenvironment. An atomic energy permit, however, is requiredto handle the concentrated tritium which is used to dope theinitial mud system at site and this process should only beconducted by experienced personnel who will conduct theappropriate site monitoring to ensure that no contamination ofconcentrated tritium has occurred. Tritium is the preferredtracer for in-situ water saturation determination for mostreservoir applications from both a cost and technical viewpoint.

Chemically based tracers would include materials not normallypresent in the formation water such as various alcohols andsoluble ions such a bromide, iodide or nitrates. Chemicaltracers are generally inexpensive and easy to analyze, but theirdeficiency centres about the fact that some chemical tracerswill complex with in-situ formation clays, affecting theanalysis, and that fact that for an accurate chemical traceranalysis a sufficient volume of water must be removed fromthe core to obtain an accurate analysis. The water must beremoved by a direct reduction technique such as centrifugingor gas displacement, as high temperature extraction (the normalmethod for cleaning most cores and removing oil and water)will result in the volatilization of the water from the systembut soluble ionic tracers will remain in the sample and not bedetected. Commutation of the core after such a process, andanalysis of the precipitated salts bas been used in somesituations, but has not tended to be a highly accurate technique.Soluble alcohols will be removed by extraction, butcondensation is often incomplete and inaccuracies are oftenpresent due to a loss of the volatile tracer from the core due toevaporation between the time the core is removed from thecore barrel and when analysis actually occurs.

Figure 11 illustrates how core is sectioned on site to detenninea tracer invasion profIle. In some cases, where only the truewater saturation is desired, only the central section of the coreis analyzed for tracer content to reduce costs and time. If anevaluation of the degree of flushing is required all threesections may be evaluated. Liquid nitrogen is generally usedeither onsite or immediately upon receipt of the core materialat the analysis lab to cut the plugs to fLX saturations in placeand avoid the introduction of any other extraneous oil or waterbased fluids into the core. The cutting process should be

conducted as soon as possible to avoid the imbibition ofdrilling mud into the core which will occur if static exposureover an extended period of time occurs if the core sitsimmersed in mud in the core barrel. Proper technique is alsoimportant and the samples should be cut in a dry environment.The condensation of water on the surface of the frozen coreafter cutting can totally disrupt the accuracy of the saturationand tracer measurements, particularly on the small outersection end cuts where surface volume of condensation is largecompared to actual pore volume of the samples. Figure 12provides an example of invasion down the length of typicalcore. Spikes in invasion are usually associated with either highpermeability streaks, vugs or natural fractures, surges in pumppressure during coring, static rotation or cessation of drillingin a single spot or at breaks between individual core barreltrips where long term static exposure of mud floods into thetop section of the first section of core to be taken once coringresumes. Figure 13 illustrates the tritium invasion profile froma typical traced core Table 7 provides a summary of typicalcorrected and uncorrected water saturation measurements fora tritium based coring program.

detemlined Sponge coring may be a useful technique toconsider in a situation where free gas, oil and water may allexist in the subject formation. Adjustments must be made forthe swelling of oil present at fomlation pressure due to thermaland gas solubility effects if oil saturation is to be determineddirectly. This can be obtained from PVT volumetric studiesconducted on the -oil. Themlal adjustment to the volume ofwater measured may also be required if high reservoirtemperatures are present. High concentrations of CO2 and H2Sin reservoir fluids may also result in high solubility in the in-situ water phase (normal hydrocarbon gases have very lowsolubility in water and do not normally appreciably affectwater formation volume factor) and corrections may berequired for water shrinkage due to gas evolution in suchsituations as the core is depressurized.

Saturations on core sampl~ are generally detennined using asoxlett or dean stark extraction apparatus where hot solvent isused to remove both the water and hydrocarbons from the coresamples. The oil volume obtained from these saturationevaluations is often detennined by mass difference of thesample, whereas the water volume removed is measureddirectly. This means that the oil volume must be corrected fordensity effects. In many situations, the oil density is assumedto be 1.000 g/cc, which often results in a significantunderestimation of the oil volume in place, particularly forlight oil systems.

Degassing and Fluid Losses. For core samples which containfluids at saturation levels above the irreducible value, fluidsmay be lost from the core as the core is pulled to surface andpressurized gas contained either as a free phase in the poresystem, or as a solubilized phase in liquid hydrocarbons,evolves and displaces either oil or water from the pore system.Two techniques are usually used in this situation, these beingpressure coring and sponge coring.

Sidewall Cores

Sidewall cores are often obtained to obtain petrographic orreservoir quality data in the absence of full diameter core.Sidewall cores are obtained in the flushed zone directlyadjacent to the wellbore and hence tend to be heavily invadedwith drilling mud filtrate and, in general, may not berepresentative for saturation evaluation. If an oil based systemis used in the drilling and sidewall coring pro~, evaluationof initial or irreducible water saturation (in a mobile watersaturation system) may be determined from sidewall coresamples. Since traced water-based mud systems are notnormally used in a situation where a conventional drillingprocess (where no core is being obtained) is utilized, the useof sidewall cores in such a situation usually results inerroneously high water saturations. Even in the situation wherea traced system is used to drill the well prior to a sidewallcoring operation, unless fluid losses to the formation are verylow, near total flushing of the sidewall core region may occurwhich will make accurate evaluation of the initial watersaturation impossible.

Pressure coring involves using a special core barrel which canmaintain reservoir pressure and in some cases reservoirtemperature (d1rough the use of electrical heating elements) tomaintain the core at full reservoir conditions during thetripping operation. Liquid nitrogen is used to freeze the core atsurface, at which time the samples are removed and placed inthe extraction system complete with all fluids (including gas)intact. Pressure coring is a complex and expensive operationwith sporadic success rates due to a high number of failureswith the pressure integrity of the core barrel and is thereforeutilized infrequently.

Sponge coring involves the use of a special core barrel with anon hydrocarbon reactive sponge material placed around theinner face of the core barrel. As the core enters the core barrelit fits tightly against the sponge. As the core is brought tosurface and depressurization occurs, gas leaks off throughsmall pinholes helically placed in the outer surface of the corebarrel, but any water or oil produced is captured in the spongematerial. Analysis of the sponge material adjacent to each coresection can allow detennination of the volume of oil producedfrom the core. If a traced water based system is used to drilland initially saturate the sponge material, any significant lossesof connate water from the core into the sponge can also be

Conclusions

The determination of initial fluid saturations is a key factor inreserves detemlination and optimizing production and locationsfor well completion. Water saturation detemlinations from logs

D.B. BENNION, F.B moMAS, R.F. BIm2POOO3-96 9

may be affected as follows: 8. Bennion. D.B., HWater and Hydrocarbon Phase Trappingin Porous Media - Diagnosis, Prevention and Treatment",Paper 95-69 presented at The Petroleum Society of CIMin Banff, Alberta, Canada, May 14-17, 1995.

9. "Schlumberger Log Interpretation Principles andApplications", Schlumberger Educational Services, July,1989, Houston, Texas.

10. Segesman. F., "New S.P. Correction O1arts", Geophy,Dec. 1962,27, No.7.Accurate detemrlnation of electrical calibration properties on

representative samples of actual reservoir core material isillustrated as essential for ensuring valid log saturationevaluations. The direct measurement of liquid saturations usinga variety of coring techniques has also been described. Acombination of low invasion coring technology with an oilbased or traced water based drilling fluid system (tritium basedsystems being the most popular) have been utilized as adefinitive technique to obtain in-situ water saturation anddirectly calibrate log data.

Worthington A.E., et al, "Departure Curves for the SelfPotential Log", JPT, Jan, 1958.

11

12. Dunlap, H.F., et ai, "The Calculation of WaterResistivities From Chemical Analysis", Trans AlME(1951) 192.

Moore, E.J., et al, "Determining Fonnation WaterResistivity From Chemical Analysis", JPT, March, 1966.

13.

AcknowledgmentsMoore, E.J., "A Graphical Description of New Methodsfor Deternrining Equivalent NaCl Concentration FromChemical Analysis", Trans, 1966, SPWLA AnnualLogging Symposium.

14.The author wishes to express appreciation to Maggie Irwin andVivian Whiting for their assistance in the preparation of themanuscript and the figures and the Hycal Energy ResearchLaboratories for the funding of this work and permission topresent the data. IS. Winsaver, W.O., et aI, HResistivity of Brine Saturated

Sands in Relation to Pore GeometryH, Bull MPG, Feb.1952, 36, No.2.References

Archie, G.E., "The Electrical Resistivity as an Aid inDetennining Some Reservoir Characteristics", JPT, Jan.1942.

16. Crowell, E.C., "The Use of Petrophysical Properties toDetennine Bypassed Pay", Paper presented at the 1995Annual Technical Meeting of the Petroleum Society ofCIM, June, 1995.Acme, G.E., " Classification of Carbonate Reservoir

Rocks and Petrophysical Considerations", BuU AAPG,Feb. 1952.

17. Worthington. P.F., "The Evaluation of Shaler SandConcepts in Reservoir Evaluation", The Log Analyst, Jan-Feb. 1985.

Trieber, L.E. et al, "A Laboratory Evaluation of theWettability of 50 Oil Producing Reservoirs", SPEJ, Vol13 (4), 1973, 221.

18. PouPOn. A., et aI, "Log Analysis of Sand-ShaleSequences, a Systematic Approach", JPT, July, 1970.

4. Chillingar, A. et 81, "Oil and Gas Production FromCarbonate Rock", Elsevier, Amsterdam, 1972.

19. Poupon, A.,et aI, II A Contribution to Log Interpretations

in Shaler SandsH, jPT, June, 1954.

Anderson, W.C., "Wettability Literature Survey", JPT,Oct. 1986.

20. Poupon, A., et ai, "Log Analysis in Formations withComplex Lithologies", IPT, Aug, 1971.

6. Masters, J .A., "Elmworth - Core Study of a Deep BasinGas Reservoir", AAPG Memoir 38, 1984.

21. Waxman, W.H. and Smits, L.J.M., "ElectricalConductivities in Oil Bearing Sands", Society ofPetroleum Engineers Journal, June, 1968.

7. Katz, D.L., et aI, "Absence of Connate Water inMichigan Reef Gas Reservoirs - An Analysis", AAPGBulletin, Vol 66, No.1, Jan. 1982,91-98.

POOO3-96

Table 1Parameters Used in Water Saturation Calculations

Table 2Effect of Variation in Saturation Exponent On Calculated Water Saturation

Table 3Effect of Variation in Archie Exponent on Calculated Water Saturation

POOO3-96

Table 4Effect of Variation in Rw on Calculated Water Saturation

Table 5Effect of Variation in Cementation Exponent on Calculated Water Saturation

Table 6Effect of Error in RT on Calculated Water Saturation

\0a;-M0:5Q

..

a~~b!)e~~b!)=.-~eU"0~fIJ~

=a=.--.-~~b!)=.-fIJ

t--~~

fIJ

- =.c e

~ .-

~";-=~~u=e

.--~~=-~00~~-~

~e~"S-

a~~~

s.Q=0;.."00....0u....'=0 ...Q=!i.'OJ:

~.-..0u.-

~'0~ ...Q=.g~..(J)....cU......a..

U

'0OJ

..=

,W

I ..

..1'1

~ e';-

=5..§

; "'.-

.- =

'"

... =

c.

I-u

~:t.u ~

trn..0u...'Q

, .

e~.f/) -s-Q

.;~

--Q E

~..

E"--

~ .e;.-..-

..

s ~

'"",=

--a. ~=~

e=-

CD

'S

.§=

.-

.-.-

.. ..

=,-c,

.- r-

-

..0 ~O

J..=

.

~

~=

~O

J. E

=-

=

~E

.- ...

.~

=

~..

0 ...

!oouQ.

~~~~~u ~~

~

.VJ

..~ E

~-

=

.. E

:"':;0.-

=

'"...

Q

Co

f-U

~..'OJ ~

~fIJ

..QU

""""g

",-QO

\""""""""M

Q

M~

MQ

""...""""-~

\Q"",,,,-

~""""""..., ,

"'t'\.~~

M.~

VlO

\"'M

t'\~~

-t'\ ",O

\\O

""~"""'~

\O.V

I.,.,;,.,;,.,;,.,;,.,;,.,;,.,;,.,;,.,;...;

~$o8:on

t"-'B""~

~.;o~

\O~

-~...

"""""'N-O

\-M--

-""-""~O

IOI_--

~~

OI~

""""~_""""

'"1..-:'"1"';"'; ."'i'"1'"1

OC

-C-O

C-O

C-O

OC

-

O\f'I-r--O

\"""'~O

\M~

N~

OIO

I...OI~

OIO

"1"'!"'!"1"'!'C;~

""'1~00000,,"0000

N""~

8N

_~""O

","

0","- O

-~\o-","

-o--_o-~o-"'",

N.fO

I.OIN

r--r---ov.N

fOI9N

.N

r--f')~

"1f'),,!'C!~

f')f')f').-.,..".."..,"".-.""""...""

on\O-~

r-..Nr-.."'Q

\Q

.on-~\O

N-"'~

r-..'"!~

~

.~IC

;~,"!,"!,"!

~O

~~

~~

~~

~~

~~

~~

ooS~

~\O

~ - ~

M

- ~. -

""O\~

\O""N

~""-O

N

.O\N

\OM

N

..,..

~.-~

"'\O

~~

~

0000000000

-M"'."'-o""~

o.~

2-~"'X

"'~""~

~!~

!~!~

!~~

~

88~8""N

O9;9;8

' O

ON

'O~

~~

. .00

.00- .

. .

M""""

M

. .".

M-

-.l-M\oN

OO

-MN

-

-o."'.i

IO\M

WI\C

M

"'."'.10\-"'1010\--""-

0 .

." .

. .

. .

. .

"""""""".,.."""'.".'"

I;i~~~

I>'

---

IJ

i~s~

i~~~!~i~~~

~

~F;

iii~~~

~~

..~=

5J~

~ii~

~

Ii~~

~~

~>

'

~:

G

!

IIi...~>'z~~~

>,w

~i..

~

iilmf

~II

~I!jiI