SPE 29594

An Integrated Geologic and Engineering Reservoir Characterization

l??!e-

Engineers
Society of Petroleum

of the Notih Robertson (Clearfork) Unit: A Case Study, Part 1

by L.E. Doublet,*Texas A&MU., P.K. Pande,* M.BClark,*J.W. Nevans,* Fins 011 and Chemical Company,
and T.A. Blasingame, Texas A&MU.

SPE Members

Sdefy ofPelrdeum
CaPm!$f1995, EBgineers, Inc.

TNe peper was pmpered for preeenlatbn al the 1995 Joint Rocky MotsMeh RegbrIsl Meeting end Low-Permaabflity Rasarvdm Symposium, Oermr, Cobrado, USA, 2022 Memfr 1995.

Thii paper was sefeded Ior preeenfefiin b an SPE Program CernmMee foflowing review of hfomdon cmfained in an ebalmct submilted by the author(s). contents of the paper, es pmaanfed,
have nof been re- by the Society of J atroieum Engineers and are s@ed 10 mredion by the atifsotts). Tfm malarial, es presented, does nof nemaeeri fymffadwvfdIbfrOf ti~of
Pefroleum Engineers, ile officers, or memlxrs. Papets presented at SPE maelinge are sbjecf to fion rakiaw by Edttil Carmlltees of fm Society of Pafrdeun Enghears. Penniaekm 10
COPY is restkl~ 10 an abstract of not more t~n 300 WOI@S. llluslra~ons may not be @. *stmdshldatain@ns@WWa-@gmeti ofnhereandbywfmmt hepeperia
presented. Wrile Libmrisn, SPE, P.O. Sox 833S36, Richardson, TX 750@3-383S, USA. Tabx, 1S3245 SPEUT.

BRIEF SUMMARY in~rwellcommunication. fheac results can be used to develop a

iti~di
=11:- uf
dIIIIIIIg
- ..,,.11.--
wuim UIIa~ .,-if -
UIIIIO. CF.m-+i-io
MJUP...rie, withnllt rro~rd In
.. . . .. . . --= -- -- ~.D &-m~!abQn_rnodelfor prediction of infill locations. III lhk
rcsemoir performance and characterization, must become a pro- work, we will demonstrate the application of reservoir
ccas of the past. Such efforts do not optimize reservoir dcvclop- surveillance techniques to identify additional reservoir pay
rnpmtac
...... . thrv -.-, --- accounL for [he ~~mp]ex
.- ---
fail tn m(um of reservoir zones, and to monitor pressure and preferential fluid movement in
hcterogcncitics present in many low pcrrncabilitycarbonate reser- the reservoir. These techniques are: long-term production and
voirs. These reservoirs arc typically characterizedby: injection data analysis, pressure transient analysis, and advanced
Large, discontinuous pay intervals
open and cased hole well log analysis.
The major contribution of this paper is our summary of cost ef-
Vcri.icaland lateral chsngcs in reservoirproperties fective reservoir characterizationand managementtools that will
Low reservoir energy be helpful to both independent and major operators for the optimal
development of heterogeneous, low pcrrncability carbonate rescr-
e High rcsiauai oii sahmatioii ~Qi~ Such-.. - ----- _R@crtson (C!earfork) Unit.
-.+ Now
as the

Low rccovc~ cfticicncy INTRODUCTION

The operational problems wc cncountcr in these lypcs of rcscr-
vous include:
Poor or inadequate completions and stimulations
Early water breakthrough
Poor reservoir sweep efficiency in contacting oil throughout
the reservoir as well as in the near-weltregions
Chamcling of injcctcd fluids duc to
prcfcrcntial fracturing caused by cxcessivcinjectionrates
Limited data availabilityand poor data quality
Infill drilling operations only need target arcaa of the reservoir
which will bc economically successful. If rhc most productive
areas of a reservoir can bc accurately identified by combining the
results of geologic, pctrophysical, reservoir pcrformancc, and
pressure transient analyses, then this integratedapproach can be
used to optimize reservoir pcrforrnancc during sccondruyand ter-
tiary rccovcry operations without resorting to blanket infill
drilling methods.
Ncw and emerging technologies such as cross-borehole tomogra-
phy, gcostatistical modeling, and rigorous decline type curve
analysis can be used to quantify reservoir qurdityand the degree of
Rcfcrcrtccsand illustrations at cnd of paper
There arc many complicated factors that will affect the successful
implementation of infill drilling programs in hetcrogcncous, low
permeability carbonate reservoirs such as the Clcarfork/Gloncta of
west Texas. Before wc began this project, we conducted an ex-
tensive literature review to gain a better understanding of tic pro-
ducibility problems we face at the North Robertson Unit (NRU).
Fortunately, these reservoirs have a long producing histow and
there is a large quantity of useful data available from case studies
for prim~, secondary, and tertiary operations in the Clearfork
and other analogous reservoirs.
In a 1974 case study conccming watcrflooding operations at the
Denver (San Andrea) Unit, Ghaun, et afl gave valuable insights
concerning reservoir discontinuity, injector-producer confor-
mance, and the effect of reservoir quality on reservoir sweep
efficiency.
Poor reservoir rock quality and the cxistcncc of discontinuous pay
between injection and producing wells resulted in a recommenda-
tion to reduce nominal WCI1 spacing from 40 acres to 20 acres. An
outcrop study on the San Andrcs was performed to verify rcser-
vou discontinuity. Injection wells were completed and stimulated
preferentially in an effort to flood only the continuous layers of
the reservoir. The original peripheral injection design was
convcrmd to inverted tic-spot paucms in an effort to decrease the
amount of water chrmnclingand early water breakthrough via the
most ~nncablc members.
In 1976, Stilcs2 summarized the difficulties encountered in water

465
2 An Intemted Geolozic irndEngineering Reservoir Characterizationof the North Robertson (Clearfork) UniC
A Case Study, Part 1
flooding npcrations at the Fullerton (Clcarfork) Unit. The author
noted lhat increasing the injection rate would never result in an
equal response at the producing WCIIS.The concept of pseudo
fill-up was introduced 10 cxpiain that Siihotigit RSCittii fiii-ii~
~,ay occur in LhC rnOSt pcrmcablc or continuous layers of the
reservoir, a large gas saturation still existed in the poorer quality
reservoir rock. For this reason, the theoretical maximum
producing rate would never be achicvcd without con~cting tie
discontinuous areas through infill drilling.
A statistical study was performed to quantify reservoir continuity
as a function of interwell distance on the basis of continuous and
discontinuous reservoir layers. Stiles maintained that injection
pressures above the parting pressure of the formation was re-
quired in order to maintain acceptable injection rates in the reser-
voir.
In a 1978 review of WCS1 Texas carbonate reservoir watcrflooding
operations,George and StilcsJ outlined their recommendationsfor
optimizing waterflood operations in the Means (San Andrcs),
Fullcrlon (Clcarfork), and Robertson (Clcarfork) Units. These
authors stressed the importance of infill drilling and pattern modi-
fication to ovcrcomc pay rock stratification, and the need for con-
tinuous interaction bctwccrr geologists and cnginccrs in order 10
produce a program of optimal reservoir development and deple-
tion.
A rock-log model was formulated for the Robertson (Clcarfork)
using a limited amount of core data and old gamma ray/neutron
logs that were available field-wide. Original oil-in-place (OOIP)
was calculated by both volumetric and matcriat balance methods.
The authors pointed out that the ratio of material balance OOIP to
volumetric OOIP should yield a qualitative rncasurc of reservoir
continuity since the material balance calculationonly considers in-
tcrvats that arc continuous or cffcctivcly completed, while the vol-
umetric calculationconsiders all payquality reservoir rock.
George and Stiles provided a method to identify ftoodable pay,
which was diffcrcntiatcd from continuous pay on the basis of the
most probable geometry of a continuous layer between an injec-
tion and producing well. Obviously, the amount of floodable pay
in the reservoir was aiways siightiy icss than the mnoiirliof coil-
tinuous pay, and both could be optimized through intlll drilling.
The authors concluded that floodablc pay must be continuous
bctwccn injection and producing WCIIS,bc injection supported,
and be cffcctivcly completed at the producing well.
In a 1980 summary of work completed at the Denver (San
Andrcs) Unit, Ghauri4 outtincd the importance of integrated geo-
logic and engineering studies in the dcvclopmcnt of the Wasson
(San Andrcs) field from primary through tertiary depletion. The
author goes into great detail describing the proccsscs that were
utilized to incrcasc sweep cfficicncy, optimize completion and
stimulation proccdurcs, and improve WCIIconforrnancc. The de-
sign and installation of automated artificial lift systems arc also
highlighted.
Barber, et a~ provided a case study in 1983 describing infill
drilling results in nine carbonate and clastic reservoirs in Texas,
Oklahoma, and Illinois. This work resulted in an cxtrcmciy im-
portant observation regarding the affect of rcduccd well spacing
on pay continuity. Using data for 20-acre wells in the Means
(San Andres) UniL a 4 pcrccnt increase in pay continuity was
*---- 4-A ...k.. ..n...ml.ml..-e~imm wiac rofillrfift !Q lo ~~~c~.
t. A~LXLU4 W1lU1l llu1lllllaI c.pati..,~ r.- . . . . . . .
However, after pay continuity was rccatculatcd on the basis of 10-
acrc welt data, it was found that the actual pay continuity incmasc
~M i4 p~rccnL The authors noted that past observations
regarding additional recovery from injill wells made prior to
drilling were probably extremely pessimistic.
In 1987, Barbc and Schnocbclcnb summarized tic results of an
aggressive infill drilling program in the Robertson (Clcarfork)
Unit. The authors found that obstacles associated with poor
reservoir continuity in hclcrogcncous, low permeability carbonate
r~scrvoirs could only be overcome through infill drilling on a rc-
duccd nominat wctl spacing. I 1991. The relevant fluid, formation, and production data are
466
iPE 29594
In addition, performance data arratysisand pressure trrrnsicnttest
results indicated a roughly cast-west directional permeability or
fraclurc orientation. This will be the most likely direction of prcf-
..fi.~;.l .~att-~...
blbaa.lul-s-. nm-wement in [k Rokr-fsorr (C]ctiork). Pay conti-
. . .... ... ... --- . --_-.
nuity was quantified using gcologicrd,reservoir pcrformancc, and
pressure lransicnt data, and aii three -. L-A..-.... .:...:1. ,-nO,
mcmuus giNV S,lllll- ,WAW.
11.
Wireline formation test results showed that individual Iaycrs had
widely
.. . different
.. w~fin formation pressures, indicating a lack of vertical
con~m~y h~e~i~fl~rk.
A further conclusion was made regarding the best locations for the
infill WC1lS.Barbe and Schnocbclcn found that the best 10-acre
and 20-acre producers were in the same anms as the best 40-acre
producers, which indicates that the identification of areas of high
quality reservoir rock is perhaps more important than finding the
areas of poor reservoir continuity when dwiding on infill WCIIlo-
cations.
The primary objcctivc of our work is not to explain the concepts
behind the initiation of a successful waterflood and subsequent
infill drilling program for hctcrogcneous, low permeability car-
bonate reservoirs as this has been effectively discussed in the
literature.l~ Rather, wc will usc the tools discussed in our work
to aid in our analyses, and wc will introduce new and existing
technologies that can bc economically implemented by all
operators.
Exarnplcs of these technologiesinclude:
1. Formulation of a rock-logmodel to identify the highest
quality pay intervals using availableopen hole well log
data for the extrapolation of core propccdcafrom a
. limited number of wells.
2. Calculation of total and movable fluid volumes and
estimation of formation flow characteristicsusing a
rigorous dcclinc type curve method for the analysis of
readily available oil production data.
3. Low-cost acquisition and analysisof pressure falloff
data from injection wells using surface pressure gauges.
4. Development of spatkd relationshipsfor reservoir
.,ns-i.hl~
,IU>m
mf intmrocf
A u-----
at IitIQ~tilPd
. . .. .~...-
(i~~rwc~)
\-
locations
across the Unit using gcostatisticaltechniques.
5. Performing 3-D reservoir simulation for history
matching, inffli drilling dcvclopmcntforecasting, and
validation.
HISTORICAL BACKGROUND
The NRU is located in Gaines County, Texas in the northern part
of the Ccntrat Basin PlaUorm of the Permian Basin (Fig. 1). The
producing horizons arc the Glorieta and Clcarfork Formations
(referred to as the Upper and Lower Clcarfork). The hydrocarbon
bearing interval extends from the top of the GIoneta to the base of
tic Lower Clcarfork, between the correlative depths of approxi-
mately 5,870-7,440 fccL
The NRU project area of 5,633 acres contains a total of 259 wells
as of January, 1995. This includes 144 active producing wells,
109 active injection WC1lS and 6 water supply wells. For the pur-
poses of this study, the Unit has been divided into three pre-
demonstration study areas (PDSA) as shown in Fig. 2.
Development and Production History
---- l-- .L-. m,--L n -l.-+.,,.- G...l
i%oitu~uonmom m~ mu UI nuw WII 8&~Ad n-o in
h-~a bti5=, ~..W.--.,
thm .wwlw
1950s with 40-acre primary depletion development. This 40-acre
primary dcvclopmcnt resulted in 141 producing WC1lS.The NRU
..- ..
WIC fntid
..... offm-livp
.. -.,... March
...--... 19X7
..-. --- -. pqNMc of irX@CItlent-
for the
ing waterflood and infdl drilling operationsto reduce normnal well
spacing from 40 acres to 20 acres. At the time of unitization, oil
production from the Unit area was approximately 670 STWD,
with a GOR of 1,550 scf/STB, and water production of 500
BW/D. Secondary rccovc~ operations were initiated after uniti-
zation and in conjunction with infii drilling. Most of the 20-acre
infill drilling was completed between unitization and the cnd of
S}E29594 L.E. Doublet, P.K. Pande, M.B. Clark, J.W. Nevrms, and T.A. Blasingame 3

shown in Table 1. the Unit (Fig. 4). This development of the geologic model in-
The cumulative produccciand injcctccifluid v~iurn~~a-c swnnla- ~iud~ the depositional setting and variability, depositional envi-
rizd below. Fig. 3 shows the production and injection history of ronments, diagcnctic cffccta~ lithological variadons, cffcctS of
the Unit from dcvclopmcnt in 1956 through Scptcmbcr 1994. pore systcm geometry, physicaf pro~fi~. ~d na~~ fractu~.
Np Duc to the lack of conventional core (a common problem in older
(MMSTB) &w) &w) reservoirs), there was a need 10 develop a unit-wide geologi-
cal/pctrophysicaf model so that individurd rock types could be
As Of 1987 17.5 8.2 NIA identified on the basis of well log responses in areas where core
1987-1994 7.4 20.8 52.3 was not available. The rock-log model that was developed is
At the time of unitization, the estimated ultimate reeovcry (EUR) baaed on the petrophysical and M.hologicalstudy of 4,600 feet of
for primary production was 20.5 MMSTBO, and the secondary to whole core from eight WCIISin the NRU. This work involved
primary rccovcry ratio was calculated 10 be approximately 1:1. detailed scdimcntologic observations of the whole core, thin-scc-
[ien ~na!y~c~, x-ray diffraction (XRD), special core analYsis
The current waterflood utilizes a 40-acre 5-spot pattern type with
20-acre nominal well spacing. Current Unit production rates are (SCAL), and pore geometry investigations using scanning elec-
approximately 3,200 STB/D, 1,400 MCF/D and 11,000 BW/D. tron microscope (SEM) image analysis. The relationships be-
Water injection volumes arc approximately 20,000 BWI/D, with tween pore volumes and pore throats lravc been quantified by di-
injcclion water comprised of produced water, and fresh water rect measurements of pore casts.
from WCOgallafaaquifer obtained from water supply wells within Quantitative rock type porosity-permeability relationships were
the Unit. established in order to identify the most attractive pay intervals.
The distribution of pay rock types in individual WC1lShas been
METHODOLOGY calculated and will be used to gcncratc interwelf reservoir quality
TM pai)~~wiii summarize both complctcd ~d schcdulcdwork in maps using geostatisticaf methods. The results will then be used
the following areas to describe the reservoir for 3-D simulation.
Gcologicaf/PctrophysicatStudy Geologic Modei
- Formulation of geologic model ..
enronaf Envir~
- Core analysis The int4wrc12~on~f~r dcr)o~itionticnvironmcnk (Figs. 5 and 6)
- Formulation of rock-log model
- Flow unit identification and reservoirlaycnng have been derived prirn~iy-from qualitativedescriptionsand thin-
- Continuity anatysis of reservoir pay scction studies of the whole core samples. The depositional envi-
rornnents for the Clearfork and Gloricta Formations are as fol-
Cross-Borehole Tomography lows:
- Theory/methodology
Highstand LitAofacicsTracts
- Preliminary results
- Plan for future data acquisition supratidal and subaerial deposits
- Integration of cross-borehole seismic data for intertidrdand channclizedtidal flaIs
rmcrvoir characterization open and resticted lagoons
- Opcrationafconsiderations - grainstone shoals
- subtidrd
Reservoir Performance Data Analysis open shelf
- 011production data analysis using dcclinctype curves
Transgressive Lhhofacics Tracts
(Reservoir quality maps)
- Reservoir performance bubble maps - shclfal and patch reef
- Waterflood performance anafysis Low Stand Lhhofacics Tracts
(Injection well pcrforrnanccplots) inland sabkha
Reservoir Survcillancc Plan
L9wcr Cka@s
- Water quality program
- pressure transient data acquisition and anafysis In general, the Lower Clearfork, which is defined as the portion
- step-rate tedrlg of the Clcarfork directly overlain by the Tubb (wansgressivema-
- Review of complctionkirnulation procedures
rine shaly dolostone interval), was deposited in open marine-
- Coiled-tubing workovcrs shelfal conditions and is dominated by grainstones that are
- Cased-hole WCIIlogging program thought to have developed in shallow water shoaling
- Future data acquisition environments. The shoaling areas may have coalesced and
intcrtlngcred with one another thereby resulting in more or less
GcostatisticafSimulation continuous behs (current andfor wave dominated and organized)
- Methodology of grainstone dcposita (primarily fusulinid or pcloidal) which
- Conditional simulation techniques comprise a dolograinstone reservoir facies with some preservation
- Approach at North Robertson Unit of primary porosity.
3-D Reservoir Simulation A study of historical production data using contour maps of reser-
- Methodology voir performance suggcs~ that these outer shelf grainstone nxcr-
- Reservoir performancecritcna voirs appear to be in communication with one another and the
- Selection of areas for modeling amount of compartmentalization and heterogeneity may not be as
hfl.-.l-l:..:4:1:Iifi
- lVIWG, u,,au~il
pronounced as in the Middle and Upper Clcarfork (which overly
the Tubb markcrj. Tinisconcept may also~ iitttibiited, itt prtt :G
PVT and rock-fluidinteractiondata
more widespread and uniform conditions of deposition with leas
- History match criteria cyclicity (fewer changes in sea level and/or water depth). A
GEOLOGICAL/PETROPHYSICAL STUDY structure map of the Tubb formation is shown in Fig. 7.
A geologic model was constructed on the basis of both macro-
scopic data (visual) and microscopic data (petrographic thin sec- The remainder of the Clearfork and the Glorieta sequences are
tion and scanning electron microscope) obt&md from analysis of typified by highstand lithofacics characterizedby highfycyclic de-
. . . ..l. bl.. . . I....h ,.-.- f-mm . l;w.+;ir.tl mt. mbr nf uwllc thrnl}nhnll!
avalla Iv W1lUIUVulb ,,,,, CA ,,,.,,.- ,,,,, -, v, ..-.. -.. e..., w. ---:* :--.! ---
puo,uuual :-nnmmmt#.fill.:.t;
V,lti,,uilie.d ~..eaous,g nf inner chplf~ cnhtitlal flak
v, . . .... .... . ---- ..--,

467
 An !ntegrated Geologic and Engineering Reservoir Characterizationof the North Robertson (Clearfork) Unit
A Case Study, i%rt i
restricted and open lagoons, tidal flats which arc frequently chan-
nclinxl, intertidal and supratidat sabkha flats.
The Iowcr portion of the Middle Clcarfork, which immediately
overlies lhc Tubb Marker, is charactcri~cd by a transgrcssivc
Iithofacicstract (approximately 150 to 250 fed thick). Within this
interval there is some cvidcncc of possible patch reef
dcvclopmcnts represented by the existence of non-porous, mottled
boundstoncs that contain sponge, algae, coral, and bryozoan
fragments.
Reservoir performance contour maps suggests that the dolostonc
reservoir rocks that charactcriz~ these depositional environments
have generally poorer reservoir paramclcrs. Porosity and perme-
ability arc rcduccd as these dolostoncs tend to bc silty and argilla-
CeOUSand arc frequently anhydritic. Cyclicity has resulted in a
high dcgrcc of compartmentalization and hctcrogcncity and varia-
tions in pctrophysical properties rctlcct the prcscncc of numerous
crossflow barriers which arc indicative of reservoir cornpartmcn-
talixation.
Pay quality reservoir rocks, where present, arc once again rcprc-
scntcd by grainstoncs and wackcstoncs (ooidal and/or skeletal)
that were dcposilcd in response to shoaling depositional environ-
ments. Oomoldic and biomoldic porosity maybe well developed,
however, intcrconncctivity is usually poor, resuhing in low qual-
ity reservoir rock. Structure maps for tbc Upper Clcarfork and
Gloricta arc shown in Figs. 8 and 9.
R@cl=is
Moldic porosity has been attributed to skeletal and grain dissolu-
tion by post-depositional leaching. This has taken piacc during
-. A,-.A. ah~ OUuati.sw
p,luua -.hofi.inl .-v-mnc,,r-
wnp. . . ..w. ~QIQrn~!~ ~~St~S h~Ve i&O been
leached which has resulted in the development of intcrcrystalline
porosity. Periods of leaching and dissolution arc probably related
to sea lCVC1 fluctuations and predate diagcncticdolomitization.
Tbc dolomitization may bc explained by ncomorpho~tion of syn-
dcpositional aragonitc (high-magnesium calcite) cement which
lined fcncstral pores. The fcncstral fabric is representative of
supratidal and intertidal facies. The calcite ccmcnt was subse-
quently dolomitizcd.
Pore Geometry
Reservoir qualily and continuity arc dominated by variations in
pore gcomcKy.7 Reservoir rocks having equal values of total
porosity may have significantly diffcrerii p~imeabiiity, iF3iAW2
permeability, and irreducible fluid saturation characteristics.
These discrepancies arc a result of changes in pore structure
caused by variations in pore type, size, and throat size.a,g
Extreme varialion in pore geometry is characteristic of hclcro-
gcncous, low pcmmabilily carbonate reservoirs such as the Clcar-
fork/GloncLa.
Pore types were quantitatively dctincd for tic available core using
pore cast studies and scanning electron microscope (SEM) image
analysis on the basis OR
Pore body size and shape mcasurcmcnl
- Pore size
.
- Pore shape factor (pare perimeter /4npore area)
- Length to width ratio
Pore throat mcasurcmcnt]o
- Coordination number (number of pore throats/pore)
- @cct ratio (pore body sizebore throat size)
Matrix/porearrangement and intcrcomcction in two and
tbrcc dimensions
Pore gcomclrics arc classified by shape as triangular, irregular,
polyhedral and tetrahedral. Primary intcrparticlc porosity has tri-
angular pores and tbc vuggy porosity is dcscribcd as being irregu-
lar (and sometimes elongated). The triangular pores arc generally
WC1lintcrconncctcd and arc typical of the grainstonc reservoir fa-
cies. The irregular pores arc lypical of dissolution porosiiy and
468
SPE 29594
rdthough porosity values maybe high relative to triangular pores,
intcrconncctivityis usually relatively poor resulting in lower reser-
voir rock quality.
On the basis of these analyses, seven unique pore types were
identified for usc in rock typing. Their characteristics arc summa-
rized in Table 2.
Rock Typing
A total of eight rock types have been identified of which four have
reservoir potcntird(one being limcatoncand wa~r -g). Rock
types have been identified on the basis of volume proportions of
pore types and unique lithological characteristics. The rcIative
volume proportions of each of the seven pore types in each rock
type are shown in Fig. 10. Average core values for porosity,
permeability, and estimated rccovcry efficiency are presented for
each of the eight rock types in Table 3. Recovery efficiency was
estimated using the mclhods outilned by Wardlaw and Cassan10
on the basis of the pore arrangement, coordination number, and
aspect ratio. Low aspect ratios and high coordination numbers
typically result in good reservoir sweep efficiency.
Rock types 1 and 2 make up the primary reservoir payintervals
at the NRU. These rock types consist of coarsely crystalline
dolostones that differ in terms of their pore geometries. These
rocks generally correspond to subtidal sandflats, grainstone
shoals, and open shelf depositional environments. Rock types 3
and 4 may bc productive in certain areas, but fOrthe most pm
they arc considered non-resenoir rock. These rocks are finely
crystalline dolostones that have different pore geometries and
.....11.. n~m~.~fintit~
u~bally~u,, wpu..u .V~w=r-.-.,
CIIII atidal tidal
.-. ----fla~ and ---------
.- restricted lagoon
depaitional environments. Rock type 5 is limestone and water-
bearing.
From the standpoint of oil production, the best reservoir rock is
rock type 1, which consists primarily of grainstones and has some
primary porosity preserved (average core porosity is approxi-
mately 5.4 percent). Coordination numbers and aspect ratios are
less favorable than for the other reservoir quality rock types (2,3,
and 5). However, the intcrconnectibility is much more favorable,
resulting in good fluid flow potential (average com permeabilityis
approximately 5.5 red). The less favorable reservoir rock types
have generally low coordination numbers and high aspect ratios
reflecting relatively poor flow potentials.
The non-reservoir rock types (rock types 6, 7, ~d 8) we ==n-
.:-11..:...-
LIiUIy Llllpel
lllGaulG~llu
,..l.la ..A
-a,, ha
,-..
w i-nrt.
~u..oaww.
iA-@
- tn.W-e
h v~fl~~~ fl-QW
banicrs. The presence of these rock types is a signikrmt factor in
the ICVCIof reservoir hcterogcncity and compartmentalization.
Knowledge of the distribution of these non-reservoir rock types is
esscntird to the successful implementation and operation of sec-
ondary and tertiary rccovcry programs.
Irregular porosity development and the abundance of small po~
throats in the Glorieta/Clcarfork result in poor reservoir continu-
ity, directional permeability and fracture trends, as well as poor
sweep efficiency during secondary and tertiary recovery opera-
tions. These same characteristics suggest that infiil drilling on a
reduced spacing and a vigilant reservoir survci.llanccprogram are
required.
Special Core Analysis
In order to accurately model reservoir flow conditions, we need
representative rock-fluid interaction data. The present capillary
pressure and relative permeability data sets will be augmented
with additional data from proposed infill wells. The initial
~~~iy~~~have h~l~d idenbf~ which rock t~PMwill & imporhllt
w]th regard to reservoir producing mechanisms, as well as the
rock types that will act as barriers to flow. This data will be used
as initial input data for reservoir simulation, as well as being a
guide for future data acquisition.
A total of twenty-four core plugs from two WC1lS
(NRU 207 and
3522) were used to gcncratc mercury-air capillary pressure
 L.E. Doublet, P.K. Pande, M.B. Clark, J.W. Nevans, and T.A. Blasingame
S.PE29594
CUKVCS. These data clcarly show significant differences in the
displacement chmc@rklics ofticwmoifrockty~. Although
these data arc from WC1lSin areas of the Unit with the highest
dcgrcc of reservoir continuity, wc can make some qualitative
interpretationsregarding reservoir quality and productiordinjcction
potential as a function of rock type. The capillary pressure curves
for rock type 1 arc shown in Fig. 11.
u&~g ~h~rn~thnds ~._L_.... by Thomccr,ll the data have been
----- nrcscntcd
intcrprctcd as hyperbolic functions in order to estimate composite
averages of pore throat radius, minimum entry pressure, and the
relative amount of incffcctivc porosity (porosity occupied by mer-
cury at injection pressures cxcecding 500 psia) for each rock type.
The results arc summarized in Table 4. On average, rock type 1
again appears to bc the primary reservoir rock with the largest
pore throat radius, the lowest entry pressure, and the least amount
of incffcctivc porosity. Rock types 2 and 5 arc moderate quality
reservoir rocks, and rock types 3 and 4 appear to have limited
reservoir potential duc to their smaller pore throat radii and large
percentage of ineffective porosity. Rock type$ 6,7, and 8 can be
charactcrimd as flow barriers in the reservoir.
, .... ., -- ..-._
~ mswts arc gencratca on a IouL-LIy-IuuL
The oit/water relative permeability data were taken at various
water saturations on twelve core samples from a single well in the
Unit (NRU 3522). Relative permeability curves for rock type 1
arc shown in Fig. 12. These results confirmed that rock types 1
and 2 arc the primary pay rocks in the reservoir. From the limited
data available, rock type 1 has a comparatively lower irreducible
water saturation and lower entry pressure, however, the prcdictcd
residuat oil saturation appears to bc cxtrcmcly high. The planned
acquisition and anatysis of additional data will give us the oppor-
tunity to further study the relative permeability characteristics of
the various rock types.
Rock-Log Model
TIM~~icctivc nf tic rock-!q rnodcl work is LOfacilitate the delin-
eation ~~-r-~;;o;~flo-w-units. Eight rock types have been identi-
fied of which only three can bc truly classified as potential oil-
producing pay (one of which is limestone and water wet). The
need for a reservoir model that has field-wide applicationis csscn-
tiat as ordy a limited number of wctls in the NRU have been cored
in the Gloricta/ClcarforkFormations.
As a whole, there was no simple, direct relationship between core
porosity and core permeability (Fig. 13). However, when the
data were segregated by rock types gcncratcd from core an~ys[s,
the relationship between porosity and pcrmcabiiity became Eurly
unique and simple linear relations could be used to define the per-
meability on the basis of porosity (Fig. 14). If reservoir rock
types can bc defined using core data, then the same poros-
ity/permeability relationship should also apply for wireline log
data, and the model can bc extended throughout the Unit to all
wctls with the rquisitc well log data.
Our rock-log model rquircs the fot.towingmodcm wireline logs:
Gamma ray (GR)
Photoelectriccapurrc cross-section (PE)
Compensated neutron (CNL o)
Compensated formation density (@
Dual Latcrolog (LLD and LLS - dmphhallow rcsistivitics)
Bomholccaliper
Sonic and Microlatcrologs would have been extremely useful in
isolating pay rock types, however, these WCIIlogs were run
with insufticicnt frqucncy to bc used.
The rock-log model was formulated to take advantage of the large
amount of modcm WCIIlog data (123 welts) that was available due
to the completion of a 20-acre post-unitization infill drilling pro-
gram bctwccn 1987 and 1991. A type log for NRU Well 207 is
-L----- :- l-,- *e
snuwn m rig, 12. ---
I ncsc int)rkm w-~ii
i~g sttitt~ SiklwfOra RIG=
5
comprehensive evaluation of formation properties than would be
possible using older porosityhesistivity well log suites.
The wells drilled prior to unitization do not have the requisite well
log suites to apply the model. However, since there is usuaJlyan
abundance of older wireline log data for many fields such as
Nofih Robertson Unit, we note that these well logs should yield
an cxtrcmcly good rock-log model if the welt log data arc properly
intcrprctcd. All of this, with little additional expenditure for the
operator.
The wiretine log data has been corrcctcd for wellbore environ-
ment, normalized on the basis of core porosity and the average
porosity across the inteml of interest (when required), and depth
shifted. Rock types 1-4 are differentiated from rock types 5-8 on
the basis of apparent matrix density, apparent photoelectric
capture cross-section, gamma ray response, and lithology (Fig.
16). The primary reservoir rock types can then be further
delineated using a crossplot of apparent porosity versus the
calculated cementation factor. Pcmneabilityis calculated using a
lithology-corrected crossplot porosity and the core-derived
permeability-porosityrelationshipsfor each particular rock type.
m. .. ----___--
- ,---- u.. r- . . L--:- r.-..--.,.. -1... ----
oam M Giuxqmn pUlUa-
ity, pcrrneability, and rock type in both cored wells, and non-
corcd wells which have the necessary well log suites. The in-
trawcll rock type, porosity, and permeability data can then be ex-
tended to an interwell basis using geostatistical simulation, and
reservoir quality maps can be gcncratcd.
Reservoir Layering
The formulation of a rock-log model has provided us a
mechanism to identify parlicutar flow units, which may consist of
one or more rock types, and which may be related to their
respective depositional environments. It is worth noting that rock
types are usually not unique to a particular depositional
envir~nrn~n[. ~.~w ~ni~~~r~ d~s~LMIl~nuQuS ~d reflect a high
degree of reservoir compartmentalization and heterogeneity both
laterally and verticatty.
l%-
. .. methnrlnlnov
..&.u.v.~J tic-d
in
A.. the
. ridm-rninati-m nf .flnw
. . . . ....... . . .. . . . .. nnitc
---- .invnlvd
.. . . . . . .
the following studies:
Scdimentologic descriptions of 4,600 feet of whole rock
cores from eight WCIIS
X-ray diffraction of core samples for mineralogy and
nnanlificatinn of clav minerakwv
>- ---------- -- --- J --------
-9J
Pore geometry analysis by SEM and pore cast studies
Special core analysisfor the determinationof relative
permeabilityand capilhy pressurecharacteristicsfor
each rock type
Dcvclopmcntof a rock-log model to determineporosity
and permeabilityrelationshipsunique to each rock type
The Glorieta/Clcarfork Formations (approximately 1,200 feet
gross thickness) have been iaycrcd into i6 stratigraphic units
(Fig. 17). Further work may necessitate the definition and
recognition of additional units. Each of these stratigraphic units
(potential flow units) arc bounded by potcntiaJcross-flow baniers
that are thought to bc representative of a sabkha or supratidal
depositional environment (the culmination of a shoaling upward,
fifth order cycle ranging from approximately50 feet to 200 feet in
total thickness). These cycles have been identified within the
whole cores and in some instances have gamma ray WC1llog
responses that are characteristic of various rock types.
These boundaries arc time-defined (isochronous), and the result-
ing depositional packages within the stratigraphic units are there-
fore contemporaneous in nature. Thtxc units have an origin re-
lated to rapid and frequent custatic (sea level) changes on a car-
bonate shelf or platform that was cssentialty featureless or without
any significant topography. Therefore, small changes in eustacy
469
6 An Integrated Geologic and EngineeringReservoirCharacterizationof the North Robertson (Clearfork) UniL
A Case Study, Prnt I
pro(hrccdhighly cyclic squcnccs. These parasqucnces arc car-
boniitc-domirtatcdwith insignificantclastic influences.
Swucturc maps have been constructed for the top of each strati-
graphic unit, and have been observed to stack. Isopachs have
also been made for each unit 10ascertain where changes in shclfal
accommodation for sediments has occurred.
The relative distributions of potential reservoir rock types (1,2,3,
and 5) and non-reservoir rock types (4, 6, 7, and 8) for each
stratigraphic unit will be dctcrmincd by predictive facies analysis.
This invo!vcs showing the distribution of a particular reservoir
rock as a function OC
Total thickness of reservoir rocldtotal thicknessof
reservoir and non-reservoir rock
Thickness of a particul~ reservoir rock lypc.hotal
thickness of non-reservoir and reservoir rock.
Thickness of a particular reservoir rock typdtotal
tiickncss of all reservoir rocks
There should bc a relationship(s) bctwccn the sedimentaccommo-
dation indicated by the rcsptitivc stratigraphic unit isopachs and
lhc distribution(s) of the various reservoir rock types. There
should also bc a corrcspom!cncc bclwccn the occurrence of the
reservoir rock types and the areas of iite RXTtOii with good
r~scrvoir performance and intcrconncctivity as per the reservoir
performance maps (in particular, the distribution of reservoir
Rock Type 1, the most favorable reservoir rock).
.AdditiQn~!y,LNMCd on the rock-log model, M and @ maps will
also bc constructed for each strati~raphic unit There should be
some very disccmiblc cormspondcncc between the distribution of
the reservoir rocks types, porosity, and permeability distribu-
tion(s) within the interval isopachs and the reservoir performance
maps.
The results from this work arc forthcoming and will bc incorpo-
rated within the gcostatistical analysis of tic reservoir. Areas that
have been qualitatively evaluated as very favorable, favorable and
unfavorable for infill drilling will then bc quantified. This quan-
tification will involve assessing the merits of respective areas of
the reservoir for infill drilling on the basis of the relative probabil-
ityy of success i.e., the qualitative geologic evaluation for relative
success of infill drilling will k quanticd and uneconomicblanket
drilling in the lCSSfavorable areas of the reservoir will hopefully
be clirninatcd.
Pay Continuity Analysis
The quantification of pay continuity based on 20-acre well log
data is in progress. Although reservoir continuity will vary for
individual WCIISdepending upon the direction in which the
correlations arc made, this is still an effective tool for the
evaluation of pay continuity, as WCI1as locating the best areas for
infdl drilling.
A study based on 40-acre WC1llog data was performed prior to
unitization on thirty-nine WCIISin Sections 5, 325, and 329 using
methods introduced by Stiles.2 Prior case studics2,Sindicate that
the analysis of the existing 20-acre, and future 10-acre WC1ldata
will show that the actual reservoir continuity is less than that prc-
,,.
dicica morn 40-aCrc wdi iUt~iySk, ~itd M by itXiWhg Wii
spacing to 10 acres wc will be contacting a much Iargcrvolume of
the reservoir. This 40-acre continuity data, and the estimated rc-
suhs of 20- and 10-acre WCI1analyses are shown in Fig. 18.
CROSS-BOREHOLE TOMOGRAPHY
The objccrivc of the cross-borehole tomography work is to obtain
intcrwcUdata conccming the spatial variability of formationprop-
crtie.s,reservoir structure, and reservoir hctcrogcncity. The cross-
borchole seismic tcchniquc has promise since it provides a mech-
anism for tmdcrstandingthe physical scale of the i.nrcrwcllvertical
and lateral continuity. This geophysical information can bc used
with gcostatistical studies to formulate an integrated reservoir de-
scription, and provides an additional method for choosing opti-
4/1)
SPE 29594
mum infill drilling locations, as WC1las a method for monitoring
flood fronts during secondary and tertiary recovery operations.12
Reservoir analysis utilizing data from cores and well logs is in-
complete since it does not include enough information concern@g
properties bclwccn individual WC1lS.Pressure transient testing
only considers portions of the reservoir that arc in communication
with the wellbore, but will not adequately delineate reservoir het-
erogeneities and flow barriers.
Cross-borehole seismic data is acquired by physically lowering
seismic source and receiver arrays down wellbores via electric
wireline and recording waves reflected off reservoir interwell fa-
cies that possess varying acoustical impedance properties. The
use of cross-borehole seismic results in higher resolution images
than are possible using surface seismic since the distances over
which the acoustic waves must travel are shorter, resulting in less
wave attenuation and allowing for the usc of a broader range of
bandwidths for interpretation.13
Seismic Travel Time Tomography Approach
Seismic travel time tomography has been in use for several
years.lz-lc Seismic energy transmitted through the formation is
expressed in terms of a source-to-rc=ivcr travel time and inverted
to a velocity ticld representation of the formation.
1- ~F~
All Ulue.. *fi
hnvn 9 cllffirient nnmk.r of rav oaths for inversion tO
. ..=. 9.. . . . -.... .. . ... . . . - ., r
the velocity field, many combinations of sourcekceiver pairs arc
recorded. The rccorciing acsign is based on RISCifWii eoiiditkrtts
and imaging requirements. A schematic diagram of the tomog-
raphic data acquisition
..-l -L_.-.:-- .L- -.,SL-. operation is shownfrnrn
.-.f . . . . math.
in Fig, 19. An ex-
th~ torn.~g~w
ample snuwlllg L1lG llU1ll UG1 U, .~, FW,Q .. .. . ..+ .
recorded between NRU wells 207 and 403 is shown in Fig. 20.
Seismic travel time inversion tomography measures changes in
velocity between the wells, for which the relationship of rock
properties to velocity is the principle concern. A decrease in ve-
locity is expected with an increase in porosity (higher fluid con-
tent).
Completed and Planned Tomography
A cross-well tomography survey using NRU wells 207 and 403
was completed during July, 1994. A future survey is planned for
pre-demonstration study area (?DSA) III of tie Unit using NRU
wells 3522, 3511, and 3528 (Fig. 21). One reason these areas
were chosen is that fuHy cored wells are available as control
points for data analysis (NRU wells 207 and 3522).
Survey parameters on the completed survey on NRU 207
(Receiver Array) and NRU 403 (Source) were:
80X 80 survey (80 source and 80 receiver positions)
800 Her@ 32 Golay
Receiver and source spacingof4meters(13. 1 feet), 1048
fed verlical distance surveyed
Interwell spacing of approximately 1040feet,
It is expected that an order of magnitude more ray paths can be
obtained in the survey for PDSA LIJas a new generation prototype
tool will bc used. It is also expected Ihat the survey will be
rccordcd at a greater range of operating frequencies, and at a re-
duced depth sampling intcrvai than the first survey. Tinisshouiti
result in better quality data for processingand interpretation.
Processing of results from onc survey is in progress and is only
partially complete. Another survey is planned for the fmt quarter
nf .-
. 1W)5
-. .Since --- lnm~g~~hy
------ Ihe .- profllc procw~g is pfCSCtlU)h-
complcte, the interwell continuity of the reservoir has been mves-
tigatcd using a simpler connectivity mapping process for the
NRU well 207/403 survey data.
Connectivity mapping is an cffcctivc imaging technique in map-
ping the continuity of the interwell formations. The technique is
based on the principles of wave guide phenomcn~ Wave guides
SW 29594 L.E. Doublet, P.K. Pande, M.B. Clark, J.W. Nevans, and T.A. Blasingame
arc formed in bcd layers possessing different acoustic velocities.
mum -.: -:-!- ;.
I m, [Jrliiclplv
.:...; 1.. in
m a. UIIItU
ih. ! nf a I_nrIIscI-I
UJ LI1mL. ~ .W.,..
rnciclivilv
.-.
- . . . . -----
fluid flow analysis, seismic waves will always follow the path of
least resistance.
This processing rncthod utilize-sspectral analysis of the specific
acoustic wave trains (channels) formed in each frequency domain
of intcrcsLin order to estimate bcd connectivity on the basis of the
quality of wave transmission through it. The advantage of this
method is that it rcqtrircs only a small portion of the full tomo-
graphic data set for processing.
There arc several data analysis slcps required for processing as
shown on the flow diagram (Fig. 22) and dcscribcd in Ref. 16.
Fig. 23 shows the connect.ivity map of the NRU 403/207 survey
based on these spectral characterizations. The connectivity map
displays the interwell continuity of the beds from the source well
to the rcccivcr WC1l.The lightest areas of the map correspond to
intervals with good connectivity. This dots not mean they are
continuous, simply that they have the sarnc acoustic propcrtiea
throughout the section.
This connect.ivitymap is wcll correlated with the existing sonic
and gamma ray logs. Interwell hctcrogcncitics can be caaily iden-
tified from the map, and arc especially apparent in the lower por-
tion of the Middle Clcarfork, which would be expected from the
geologic intcrprclation of the intcrvat. The results also indicate
that several formations show good interwell continuity. This is
not surprising, since the results of geologic and reservoir perfor-
mance work indicate this area of the Unit has relatively high re-ser-
vou continuity,
When the flow unit Iaycring schcmc (Fig. 17) is compared to the
connectivity map (Fig. 23), wc note fairly good correspondence
bctwccn the two, indicating that connectivity mapping maybe an
effective tool for layering the reservoir for flow simulation pur-
poses. It should bc noted that due 10the diagcnctic processes that
are common in carbonate reservoirs, the resulting wave guides
that form will not always follow dcposilionat boundaries, but may
a!so {Aw .1. wl..~ -.. - ...
zlfimuTnc in which pos!-dcposiha! porosity has
been created via diagencsis.
Additional proccsing of the prcscnl data set (velocity map, inver-
sion tomography, rcftcction profiling) is currently in progrcas.
These rcsutts, as WCI1as the results from the survey scheduled for
PDSA HI, will be discussed in greater detail in a subsequent
work.
Operational Considerations
Approximatelyhalf the cxpcnditurcs on the initial survey pertained
10prc- and posl-survey WC1lpreparation costa. The planning re-
quired for the usc of producing WC1lS for tomographic surveys is
much more straightforward (and economic) than for injection
WCIIS.In watcrfloods such the NRU, where the producing WC1lS
arc pumping WCIIS,flowback is generally not a conccm. The
preparatory work required for producers typically involves just
removing the pump, rods, and tubing. Injection wells generally
have to bc killed with mud or salt water to achieve static well-
bore conditions for the survey. Post-survey restimulation using
acid or other chcmicats may be required in order to restore injec-
tivity. This stimulation work on the injector adds signi.tlcantlyto
post-survey costs.
In the Permian Basin, typically sour (H2S) conditions exist in
reservoirs for which tomographic data would be of interest.
n-- . ...-- UJ1l
~Ffidtu13
4..-.
,. . . . ,:..9 Ok
LLUIIJ1ihb
e !-,m.,.
VGJS
.I.1,-,,,1,-l
--
,Xlmr
w w
k with th~ wrvice
. . ... .... . . ----
company to carefully consider the ability of the tools to withstand
sour wellbore conditions. The materials used in the tools must be
able to survive norrnat live wellbore conditions, in which gas is
likely to bc present in wellbore fluids.
Rcccivcr arrays arc generally more time consuming to move from
WC1lto WCI1than the source tool. For this reason, consideration
should be given to minimixe well-to-well moves of the rcccivcrar-
ray whcncvcr possible in multiple WCI1surveys. In addition,
consideration should bc given to the amount of wellbore
7
ratholc. In order to perform both tomography and reflection
1001.
AS irr
profiling work, there needs to bc sufficient rathole so that tool
strings can be positioned below the zones of intcrcsl to generate
seismic waves upholc, as WC1las downhole.
Scrvicc companies offering tomographic scrviccs are beeoming
more aware of the need of operators to minimize pm- and post-
survcy opcrationat costs. Efforts to reduce tool size and minimize
the possibility of tool sticking arc also in progress. Service com-
panies have also rccogntid the need to reduce operational down-
time of tools. Only if a significant number of surveys are
rccordcd with the prototype tools to ensure reliability and generate
rcvcnuc for further development will survey costs be reduced to
the point where tomographic services can become a part of
standard reservoir charactcrimtion and surveillanceoperations.
RESERVOIR PERFORMANCE
In order to best define the factors affecting reservoir producing
mechanisms, we will analyze all available long-term production
and injection data using the following reservoir performance tools:
MatcrirdBalance Decline Curve Analysisof Long Term
ProductionData
Reservoir Pcrfonnrmcc Bubble Maps
Waterflood PcrforrnanceAnalysis
Material Balance Decline Curve Analysis
In order to verify the results of the rock-log modeling, the
analysis of long-term production data is being performed using a
ngQrousmaterial balance decline type curve method.17 An initial
study of the primary 40-acre producers has been completed
utilizing the Fetkovich/McCray Type Curve.18-zo(Fig. 24). A
step-by-step procedure for the use of this technique is given irr
Ref. 17.
This method yields excellent results for both variable rate and
variable bottomhole flowing pressure cases, without regard to the
---------- VA
SUULLUIG
r.- ..
-c UIG ...*-.
lG=L
;. (.k m -au
Voli\~laaP .wi ..-,,
.;.-} .
rir the recervnir drive
. ..v.. . . .. ----
mechanisms. The use of three different type curve plotting func-
tions (rate, rate integral, and rate integral derivative) allows for the
arrrdysisand interpretation of lypical noisyfield production data.
In addition, the integral functions provide better type curve
matches than could be obtained using existing dceline type curve
matching techniques and increase confidence in our interpreta-
tions. These analysis techniques have been verifkd by evaluation
of a number of simulated and actual field data cases, with out-
standing results.
Results of these analyses include the following:
. Reservoirpropcrtictx
- Skin factor for near well damage or stimulation,s
- Formation flow capacity, h%
. In-place fluid volumes:
- Originat oil-in-place, N
- Movable oil at current conditions, NP,WV
- Reservoir drainage area, A
Wc focus on using data that operators acquire as part of normal
field operations (e.g., production rates from sales tickets and
pressures from permanent surface and/or bottomhole gauges). In
most cases, these will be the only data available in any significant
quantity, especially for older wells and marginally economic
welts, where both the quantity and quality of ~ typekof data are
limited. This approach eliminates the loss of production that
occurs when welk are shut in for pressure transient testing, and
provides analysis and interpretation of well andjieldpe$ormance
at [ittle or no cost to the operator. T?ds techniquealso provides an
additional method for localing the most productive areas of the
reservoir, and identifying any preferential flow paths that may
exist.
An example analysisonNRUWC1l3510 is provided below. For
471
v
i, ArI Integrated Geologic and Engineering Reservoir Characteri@ionof the Nofi Robc~on (Cl~ork) Unit SPE 29594
A Case Study, Part 1

Clcarfork data, wc prefer to report flow capacity instead of pcrmc- ~. (Fig. 30)
.. .l..~.- .La C,.,., 4A, mr.,.,,pmtm ., Iltm. f,.I. ~ot ~~v {nfp~~l ~~
available, WC
atrihly uuu m UJUIaLLUKILaWUICWL. V&uw .U. .,-. F-J . ..-. . -- ,..- .-. -- ..
Cinrm there nre no ~~!!~-~-h~i~ fjowing pm.$sure data

not available. Average fluid properties were used in the analysis. plot the daily oil ralc, q, versus NP to find lhc movable oiI vol-
NRU Well No. 35j.f) ume. The extrapolation of the data trend to the Np axis intercept
Fig. 25 shows the locationofNRUWCI13510 with respect to its yields a movable volume at the time when all reservoir energy has
WCI1patlcm and Section 329. This WC1lwas drilled in 1963, and been dcplctcd.
complctcd in the Lower, Middle, and Upper Clcarfork. The The calculation of movable oil volume using the q versus NP plot
Lower Clcarfork was stimulated with 2,000 gallons of acid, and yields acccptile results urdcsspwf varies significantly. Skluk[-
kA_iIIL-21~v
Llyu, mmwa.
frarlllrd
, .. -.-.-
dnwft
-- .- - -
casing wilh 21,500 gallons OfffSC-
turing oil and 60,000 pounds of 20140 sand. The Middle and cd data cases were used to verify that the q versus N pioi yieidcd
Upper Clcarfork were stimulated with 2,000 gallons of acid, and similar results to those prcdictcd by more rigorous pfots that could
fractured down casing with 86,000 gallons of frachuing oil and be made when bottomhole flowing pressure data were available.
180,000 pounds of 20/40 sand. This conclusion has also been confirmed for field data cases for
.-. -- _. tvhieh
... -.. hot
... h .
cllrfaee ----- hot
...-., and . . . lnm~de
. . . . . . --- flowing
. - prmsure data were
The well initially tested at 55 SllW/U Irom the Lower Cit%~ftlrk
and 127 STBO/D from lhc Middle and Upper Clcarfork. It had available.17
produced approximately 226 MSTB as of July 1989, when it was The estimate for primary EUR was 270 MSTB. Our results indi-
converted 10a water injection w.c1l.Scmilog and log-log produc- cate that approximately 44 MSTB of primary movable oil re-
tion plots shown in Figs. 26 and 27 indicate that there was a sig- mained in the drainage area of the WCIJwhen it was convcrtcd to
nificant ralc fluctuation af[cr a workovcrhcompletion that oc- water injection. Obviously, the actual movable oil volume will be
curred in 1968. The data set was reinitialized in time to eliminate slightly lCSSthan the volume extrapolated for production to zero
this data spike. At the time of its conversion 10injection, there oil rate.
had been no visible response 10Unit waler injection that began in
1987. Anal~
~
T: (Fig. 28 and 29) NP = 226.0 MSTB
We now consider the type curve matching of the rate, rate irrtcgral, Np,mov =270.0 MSTB (primary EUR)
and rate integral derivative functions plotted versus material bal- Recovery Factor = 7.5% (primary)
ance time on the Fctkovich/McCray type curve. The three rate The results of the type curve match and material balance analysis
functions are force matched on the Arps b=l (harmonic) dcclinc yield rigorous estimates for original oil-in-place and movable oil,
~t~in as di~tat~d bY th~OrYf~~ t!?~ ~se Of rn~!~tid hdMCC .,~~,-.,,=l;f.t;w~ - t;matac fnr Arainaof= an-a ftnw a acitv. Md Skill
~tu qu-~-~- ~-..-- .. .--.-e ---, ---- --P----.,
time,17,20and the appropriate match points arc obtained. factor.
- The primary recovery factor calculated using the value of
To obtain the best type curve match, the data was rcinitialixd at a OOIP from the type curve match is typical for weiis in &suiit
time of 2227 days. After rcinitialization, we obtained a good
match on the depletionstems and a unique match on the r~ = 800
transient stcm. From the log-log production plot (Fig. 27), wc The results of the dcdine type curve analysis were used to gener-
note that the transient flow period introduced by the well ate rcscxvoirquality maps of in-place oil, primary estimated ulti-
workover had not ended at a time of 2227 days, and the transient mate rccovcry (EUR), Hr. and estimated drainage area (Figs. 31-
match should bc valid. Using this dimensionless radius and the 34). The total original oil-in-place for analyses of all 40-acre pri-
time and rate match points, we calculate values for in-place oil, mary producing wells was 259.8 MMSTB. Previous estimates of
drainage area, flow capacity, and skin. MIIF for tnc u-nithave ranged between 200 and 300 MJvRYf%i
based primarily on analogy to offsetting properties.
MatchingPararnctec r~ = 800
The estimated primary EUR for the Unit from material balance
[t~]MP = 1.0 [IMP = 10,000 days decline type curve analysis was 19.5 MMSTB, which is in C1OSC
[9~P = 1.0 [q/Ap]Mp= 0.009 STBIDipsi agrccmcnt with the estimate of 20.5 MMSTB made prior to uniti-
zation. The average estimated ultimate rccovcry for the primary
Based on our estimated value for total compressibilitywc find: wells was 142.3 MSTB, with an average primary dcclinc rate of
Net = 90.0 STB/psi 7.51 pcrccnt. Estimates for tic reservoir flow characteristicsindi-
N = 3.60 MMSTB cated that the producing intervals possessed exfrcmely low effec-
t;v-
U.
nm.meahilitv
pi... mu,,...,,
ant+ that
. ... .
the
. .
w@llc wt?re
. . -.. . . ---
nnlv
-...,
m@~~ii~
--
StirnU-

A = 43.1 acres latcd with short fracture half-length.

re = 773.0 ft
M = 12.75 md-ft
s =-1.1
A pressure buildup test was performed on WC1lNRU 3510 in
1988, and the pcrrncability to oil was estimated to bc 0.43 md,
with a calculated skin factor of -1.93. For dcclinc type curve
analysis, the calculated skin factor was -1.1, and using the esti-
mated net pay intcrvat for the WC1l(150 feet), the calculated per-
meability was 0.085 md. The large discrepancy in the pcrmcabili-
tics is duc in Iargc part to the fact that the dcdinc curve calculation
is based on twenty-five years of production history, while the
pressure buildup results reflect reservoir pressure behavior near
the WCI1over a 250 hour period. The flow capacity in the near-
wcllborc region may bc good, but duc to the fact that this WC1lis
in an area of relatively lower reservoir connectivity compared to
t.hcsurrounding WCIIS,the flow capacity calculated from dcclinc
type curve anrdysisis fairly poor.
The results of these analyses also show that the majority of the
original primary producing wells drained arm leas than 40 acres
(Fig. 34). The average drainage area for all 40-acre primary wells
based on an average net pay thickness of 250 feet was 22.7 acres.
This gives an indication of the lack of reservoir continuity, and
shows why nominal well spacing was reduced to 20 acres. The
analysis of the 20-acre WCIIproduction data will give an indication
of how effective the further reduction to 10-acreWCI1 spacing will
be. Analysis work on the 20-acre producing WCIISdrilled
between 1987 and 1991 is in progress. The analysis will most
likely be limited to calculation of oil-in-place and estimation of
flow capacity and skin factor. The estimation of secondary EUR
will be difficult since most of these wells are not on decline.
These maps vcnfy that the highest quality arcaa of the Unit am in
the northwest (Section 329), and southeast (Section 5), as was the
case for the geological/pctrophysica.lanalysis. In addition, there
also appears to be an extremely good area to the southwest
(Section 327). Each area has higher than average OOIP, M,
drainage area, and primary EUR. Due to the fact that the gco-

472
SPE 29594 L.E. Doublet, P.K. Pande, M.B, Clark, J.W. Nevans, and T.A. Blasingame
logic, pctrophysical, and cnginccnng analyses have verified onc
another, wc suggest that both the reservoir quality and contmuily
arc comparatively bcucr in lhcsc areas and they should be consid-
ered for flow modeling and subsequent infill drilling.
Due- LoUlclavt
. .I-- c..* tlla.
k-t 01.yr--
ii m ndl)clinn hw ~n
----- -- commingledfor a ]Wgc
portion of the producing life of the WC1lS thal make up the NRU,
the entire Clcarfork and Gloricta producing intervals were
analyzed as onc unit. Volumclric calculations maybe affected by
the existcncc of additional productive layers of the reservoir that
were not cffcctivcly complclcd during primary production and not
in communication with the wellbore. In addition, average values
for formation and fluid properties had to be used in (M~i to per-
form lhc analyses, and formation ncl pay thickness is almost im-
possible to dctcrminc accurately in the Clcarfork duc to extremely
long imcrvals thal arc prcfcrcntiallycompleted and stimulated.
The inability to complctc all results with a high degree of confi-
dence is not related to the analysis
. . .or interpretationmethodologies
..:~ -,4ot=with
wc used, but rather, to a lack 01 reservoir artd fhSAU . . ...
which to complctc these calculations. While the material balance
dcclinc type curve analyses have yielded extremely good results,
wc USCthis opportunity to again point out the importance of early
and complctc data collection.
Reservoir Performance Bubble Maps
An cxtrcmcly simple method for identifying areas of the Unit in
which wc want 10concentrate our cfforrs is the use of reservoir
performance bubble maps made using a commercially available
software package.zl As an example, wc have plotted the cumula-
tive volumes of oil and water production, and water injectionprior
to, and after the inception of secondary recovery operations at the
NRU (Figs. 35-39). These maps can be used as a qualitative in-
dicators of relative reservoir quality because wc can easily idcndfy
the most prolific producing arus and relate this to our previous
geologic and cnginccring itucrprctations.
Fig. 35 shows that the most productive areas of the reservoir
during primary depletion were Section 329 in the north, Section 5
in the east, and Section 327 in the southwest. High water pro-
duction was limited to a fcw wells spaced throughout the field
(Fig. 36). Secondary oil production has been extremely good
~iong ~+e~or~bcmand ~~~~h~~~~m~lcrs of tie Unit, Wil.hin-
significant contributions from the central region (Fig. 37).
Cumulative volumes of water injcctcd and produced have beat
high in both the northcm and eastern sections of the Unit, with a
fcw WC1lS making the greatest contribu~ions(Figs. 38 and 39).
Waterflood Performance Analysis
The NRU was dcvclopcd using 40-acre five-spot patterns (1:1 in-
jector/producer ratio) for optimum injectivity and pressure sup-
port. Sweep cfticicncy is still low duc to the discontinuousnature
of the reservoir and propagation of fractures along preferential
paths (cast-west) bclwccn injection WC1lS.We require some addi-
tional tools to help dclirrcatethe nature of the probkm.
Wc feel that the analysis of waterflood performance using con-
ventional techniques such as Buckley-f-cvcrctt, Stiles, and
Dykstra-Parsons will bc difficult duc to the problems associated
with obtaining accurate estimates of net pay lhickncss in the
ClcarforMGloricta. During lhc next phase of the project, wc will
determine if the.scmethods can be cffcctivcly used in the develop-
ment of a waterflood model for comparison to 3-D simulation
results. The formulation of a conventional waterflood model
without performing simulation is important because it may not be
economically feasible for smaller, indcpcndcnt operators to pur-
chase a commercial simulation package, or obtaiti the tr-anhtg re-
quired to usc it correctly.
There are many simple graphical waterflood performance evalu-
ation techniques that can be applied in order to identify the prob-
1--- ik~~.rrm~t
IvlIla tlla, ----- .~nnd
-...
dficj~nq, TIICSCdiagnostic plots CMI be
made using long-term production and injection data, which should
be available to all operators.
473
9
We can use a plot of daily injection rate and total fluid production
rate versus cumulative water injcctcd to estimate the current wa-
terflood efficiency. Figure 40 shows that UN?injcction-to-prduc-
tion ratio has steadily dccrcased since reservoir fiil-up (at approx-
imately 20 MMBW injected). Since unitization, the ratio of total
injcctcd to produced fluids is approximately 1.85:1.0. The
current ratio is about 1.35:1.0, indicating..that
. . -some
---- of
....the
.. . injected
~,,ml;t
water may still be leaving the u-ni~ tmt rnaKrecxm WaLGJ ~uga.y,
remediation, and pattern balancing work have been fairly
effective.
Jordan22provided a straightforward graphical technique for eval-
.x-- ...-.--n
uaung WmGILIOW mrfnrrnance usin Q cumulative produced ~d
A p,..-........+- -.-..= .-
injcctcd fluid volumes. A plot of cumulative total fliid production
and cumulative oil production versus cumulative water injected
can be used to characterize waterflood efficiency (Fig. 41). The
following rulescarIbe applied to the plot to evaluate waterflood
performance
., 1:- JO
: ~,
0 _,~-
imd;r=fa~
The slope of the tots.tll~as produced 1111=
of sweep efficiency
The deviation of the cumulative secondaryoil produced
line from the total fluids producedline at early time is
indicative of early water breakthrough
The slopq of the cumulative secondaryoil line is an
indicator of what the eventuaJsecondary rccovcxywill be
under current operating conditions
I%?low slope (c 45 degrees) of the total fluids curve indicates
that the NRU waterflood has a poor sweep efficiency. The devia-
tion of the cumulative oil production line from the cumulative total
fluid production line at an extremely low volume of cumulative
water injection indicates the Unit bad extremely early water
breakthrough. This line also appears to be nearing an asymptotic
value for secondary EUR that will be lower than that predicted
(20.5 MMSTB) prior to unitiution.
While these results indicate early water breakthrough, pobr SWCCP
efficiency, and low secondary EUR. it should be noted that they
better-than-average for west Texas carbonate waterflood.
Targeted infill drilling, and the optimization of completion and
-.:-..1 .,:,.. ~._.+,,wQ
Suluulauuu pAVVu . . . ..-
+nl,ld grea~y enh~~
------
tie current flood
efficiency.
Hallzs provided a straightforward graphical technique for the
analysis of long-tenrninjection well performance data. TheHall
cocfficicn~ which can be defined as the cumulative total of the
product of the average monthly injection pressure and the number
of days per month the well is on injection, can be plotted veraus
cumulative water injected to produce a diagnosticplot for monitor-
ing the behavior of injection wells. These are presently the pri-
mary tool utilized in decisions regarding water injection well
workovcrs at the NRU.
From a plot borrowed from the work of Thakurz4 (Fig. 42), we
scc that linear trends which fall above the normal line (D) indi-
cate pore plugging and a possible water quality problcm. Data
plotting below the normalline (B and C) indicate water channel-
ing or injection at pressures greater than the formation parting
pressure. Fig. 43 shows an example Hall plot of an NRU injec-
tion WC1lthat bccarnc plugged and was subsequently worked over
to remove formation scale and wellbore firl.
Wc would expect to see data fall below the normal line if injec-
6:--
UU1l ..A~ ~r- in
WtiA..l- ... rractur~
---- ~ornmunication~however, if the ncar-
WC1lregions had already been repressured at the time when frac-
ture communication began, they would be difficult to see on tJre
Hall plot. An example of this phenomenon is shown in Fig. 44.
The Hall plot for NRU WC1l301 shows an upward trend in the
lhic i,iiatappears to indicate p-r- ~.-=e-..c. However! the ICSUkS
n I=nlnroinc.
of a recent pressure falloff test indicate the well is not plugged,
10 An Integrated Geologic and Engineering Reservoir Characterizationof the North Robertson (Clearfork) Unit:
A Case Study, Part 1
but is in fracture communication with an offset injector. The
WCIISinjection rate has dcclirwd drastically, and the injection
pressure has gradually incrcascd over the past year.
Wc conclude that if injcclion WCIISarc in communication with
each olhcr, wc might WC1lscc an upward trend on the Hall plot
similar to that caused by pore plugging, indicating that it became
more difficult to maintain injectivity duc to pressure support from
an offset WC1l.In addition, a sudden change in injection pressure
or rate would also be indicative of offset pressure support.
Wc believe that in order to better identify injection well responses
such as those summarizd above, we need to develop type curve
techniques similar to those currently used in dcclinc curve and
pressure transient analysis. Since onc of the main goals of this
study is to identify useful and economical data analysis methods,
wc will work towards developing such a method for the analysis
of long-terminjection data.
RESERVOIR SURVEILLANCE
As summarized by Robertson and Kclm,~ there area number of
factors to consider when initiating a comprehensive rcacrvoir sur-
veillanceprogram. These include
Allowing for a maximum pressure differentirdto exist
. 2--.:-- .WUIIS
between the producing and iriJtM.u1g . ...11 . .:,l.
*luluut
n.,
cxcecding the formation pardng pressure
Performing early and continuous pressure buildup tests
on the producing wells to dctccl formation damage and
monitor reservoir pressure
C&.I@ out a SyStCrnatiC program of cased hole surveys
(tcmpcraturcand fluid tracer) to monitor fluid injection
profiles on a regular basis
Utilizing continuous pressure falloff testing on the
injcctio; wells to monitor the growth of v;rtical fractures
duc to continuous injection
We will add tic following as part of our surveillanceprogram:
Continuouslymonitor injection water quality to increase
injectionefficiency
Combine the results of pressure falloff test analyses with
the rcsu!ts of waterflood diagnostic plots and regular
step-ratetesting to improve mjcction efficiency
OptimizeWC1l conformance by injecting only into zones
. . ..I.~1
which arc crmdntious ix%vccrtinjeciors ~lu ..nA.
P**C
uud~e,=
Implementa continuous program of completion and
stimulationoptimization to improve sweep cfticicncy
Utilim llcrmal Decay Time (TDT) logs on a periodic
basis to monitor the movement of reservoir fluids in the
near-wellboreregions of the producing WCIIS
Water Quality Program
Injection waler quality is onc of the critical components in the im-
plcmcntalion of a successful waterflood. Unfortunately, the con-
tinuous monitoring of water qualily is still not considered part of
many operators rrxcrvoir surveillance plan. This often rcsulta in
poor waterflood cftlcicncy and numerous operational problems.
--1n~ .-LULiU
.-1 -! -:1.. : ..:*-,:,. -,, ,- ,.. ,1.e NIDI T had rhw-.-.ew! fr~m.
Uillly lUJL,b LIUU latu fuJ LIW ,.nu ,,au uvwi-e,m,v
30,000 BWUD in 1989 to 16,000 BWUD in 1992. A cost effec-
tive surveillance program was initiated to identify and resolve po-
tsn~~ WNCr rpti!y prQblC.rn.S= -At the same time, an injection well
workover program was implcmcntcd to rcmcdiatc the scaling
problems in individual WC1lS.
Duc to the fact that both fresh (Ogallala aquifer) and produced
water arc used for injection at the Uni~ both waters had to bc
tested separately for their plugging and scaling tendencies. In
addition, both waters were tested together to delerminc if any
474
SPE 29594
compatibility problems existed. The following tests were con-
ducted:
Dhvch-al
* ..J.... nrnncrlkzc
y.y -----
total dissolved solids
- pH
- particle size distribution
Filtration
suspended solids
- acid soiubks
- hydrocarbon solubles
Dissolve Gases
- Oxygen
carbon dioxide
- hydrogen sulfide
Bacteria
- Anaerobes
- acrobca
- dissolved iron
Although the injection waters were found to be compatible, both
the produced and fresh waters were found to have substantial
plugging and scaling tcndcncics. The water handling facilities
were redesigned and programs were implemented to 1) prevertt
the formation of solids, and 2) remove all remaining solids from
the system. The entire NRU water quality program is outlined in
detail in a work by Ncvans, et aLx
The Units daily injection rate subsequently increased to 26,000
BWVD, and is currently between 20,000 and 22,000 BWI/D. At
prcacnt, quartcriy tests are conducted on ititiividuai we%, aiid
field-wide tests are conducted biannually,
Pressure Transient Analysis
I At the time of unitization, a wide range of fluid bubble points ex-
isted in the reservoir. This differential pressure depletion is in-
dicative of poor pressure continuity and-is supported by the bot-
tomhole pressure data col.lcctcdjust prior to unitization and during
reservoir fill-up (Table 5). A unit-wide pressure transient data
acquisition program was initiated in the last quarter of 1994 to
provide further data for simulation history matching, to estimate
completion and stimulation efficiency, to identify the bmt areas of
the reservoir with regard to pressure support, and to identify any
other major producing probhns relamdio wakfluuu -AA~b-y
.,-- afG-
~.AS-
ciency.
The majority of the tests are pressure falloffs on injection wells,
so as to minirnizc the ioss of oil production. At present, wepiaii
to run fifty to sixty falloff tests, and tcn to fifteen pressure buildup
teats. At this time, nine rcccnt falloff tests and seventeen buildup
tests recorded just after unitization (1988) arc available for analy-
sis. We have used both pressure and pressure integral data to per-
form scmilog analysis, and log-log analysis using radial homoge-
neous and fraciurcd WC1ltype curves with wellbore storage
effects. The results of these analyses were then used to match
simulated rcsulta gcncratcd by optimizing on the formation flow
characteristics (permeability, skin factor, wellbore sloragc cocffi-
cicnt, and fracture half-length) to the raw pressure and pressure
integral data.
reqrdcd jn @tobcr and November of 1988
- --- huildun
Prtx..,m
..----_r daia
_______
was available for scvcntccn producing WC1lS.At that time, the
wells had rcccived limited pressure support as the water injection
program was only initiated in the last half of 1987. Fifteen of
these wells were ncw 20-acre producing wells, and two were
original 40-acre producers which were subsequently converted 10
injection.
These surveys were recorded by measuring the shut-in surface
pressure while simultaneously making a fluid column height mea-
surement using an cchometer (automated weil sounder) device.
SF%29594 L.E. Doublet, P.K. Pande, M.B. Clark, J.W. Nevans, and T.A. Blasingame 11

The overall data quality was not extremely good, however, fifteen
of the tests were of sufficient quality to perform complete analy-
ses. The rcstrltsarc summarized below, and example analyses for
NRU wells 1008 and 3510 arc shown in Figs. 45 and 46. As
with the material balance dcclinc curve analysis, wc prefer to re-
port a value for the flow capacity rather than the pcrmcabilily due
to tic fact that the net pay thicknessesarc estimated values.
Y- rmc: ~!~~p, p~~~
Wci# k%,~ti-~l ~kiii ~~~i~i Al? ~-
date indicate that the wells may be receiving a great deal of
pressure support from offset injectors. This may also be duc to
the prcscncc of formation scale plugging or wellbore fill which
can bc identified from waterflood diagnostic plots.2~u
Pertinent falloff test data for each of the rcccntJy completed sur-
veys is shown below:
Well # qkj, BW/D Pi.j. psig Type Test Time, days

102 138 4570 fresh 65

201 7.5 -3.7 25.0 1033 295 330 4530 fresh
208 28.6 -4.1 36.0 719 3510 260 4705 prod ;:
1008 37.0 -3.1 14.2 985 3511 270 4580 fresh
4202 49.5 -4.7 71.0 507 301 420 4950 prod ;;
2217 44.6 -3.8 27.4 1350 401 520 4740 prod 49
2224 7.9 -4.1 37.0 1074 702 270 4730 fresh 49
2225 2.7 -3.6 22.0 1228 1591 125 4730 fresh
75.6 -3.2 15.0 985 3004 240 4850 prod %
%%w 14.3 -3.1 14.0 1431
2804 4.9 -5.0 90.0 1292 Individual analyses for NRU injection WCIIS301 and 3510 are
1506 32.8 +1.0 0.0 840 shown in Figs. 47 and 48. The results of all completed falloff
3iji i fii.ti -3.5 -n n
Lu. u 992 tests am summarized here for the readersconvenience
3510 64.5 -1.9 0.0 384 Well # M, md-ft Skin Factor Xf, feet Est. SIBHP, psia
3517 78.6 -3.8 26.0 939
3523 32.4 -3.1 15.0 1199 102 32.0 -5.0 89 2950
Wc see that even though the majority of these WCIISwere newly 295 130.0 -5.0 88 3730
drilled, complctcd, and stimulated at the time of the surveys, the 3510 27.0 -6.7 492 3400
hydraulic fracture treatments were for the most part unsuccessful. 3511 51.0 -6.0 250 3150
These jobs were designed to produce fracwrc half-lengths of 120 301 184.0 -5.9 232 4200
fee~ however, tbc calculated average half-length is only 27.5 fce~ 401 188.0 -5.3 126
and no fracmes rcachcd the designed length. The results indicat- 702 124.0 -6.3 323 %%
ed that two WCIIS(NRU 1506 and 3510) had no propped fracture 1591 158.0 -6.1 281 4250
3004 ?tnn
length at all, and NRU 1506 appears to~ damagco. -..
-~.~ 2!?
4!)0(?

From the analysis of NRU 3510 (Fig. 46), wc scc that as we
surmised from the material balance decline type curve analysis,
the quality of the reservoir rock in the near-wellbore region is
fairly good, however, a relatively low estimated bottomhole shul-
in pressure indicates that this WC1lis in a region of poor reservoir
connectivity as it is more pressure-depleted than the surrounding
WC1lS (NRU 3517 and 3523).
Most of these pressure buildups were not rccordcd for a sufficient
length of time to w any boundary or intcrfcrcnce effects. One of
our major goals for future buildup and falloff tests is to record
sufficient data to be able to scc these effects, and to better defiic
the reservoirs producing mechanisms. In addition, if it is eco-
nomically feasible, wc will perform pressure buildups using
downholc shut-ins in an effort to rcducc wellbore storage effects
arrdob~~n
-.. ~!t~r
.-. flllalitvdata
q-- ... --- ---
for interpretation,
---
Prcssufc Falloff Test Analvsi~
The pressure falloff data was acquired al surface in an effort 10rc-
ducc costs and dcmonswatcthal these tests can be recorded at little
or no cost 10 the operator. Results to date have been excellent,
and have helped a grcal deal in explaining some of the major
problems associated with watcrflooding a low permeability cu-
bonatc rcscmoir. Bottomhole pressure buildup tests will be done
to vcnfy the estimates of average reservoir pressure obtained from
the injection falloffs.
The pressure falloff lest is the most popular tool for monitoring
waterflood pressure performance. By using surface pressure data
acquisition, wc fed that tests can be performed easily and eco-
nomically with greater frequency. It has been shown that surface
pressure acquisition yickis data of sufiicicnt qurdity for interpreta-
tion, cycn when !ow precision pressure gauges [~ 1 psi) arc uti-
lizcd.z7
Again, in an effort to identify intcrfcrcncchoundary effects on the
injection WCI1falloff tests, wc have made an effort to let the tests
run as long as possible. Wc feel that this is especially important
in identifying the problems that may affect reservoir sweep
efficiency. The cxtrcmcly long falloff times wc have witnessed to
The shut-in bottomhole pressures were estimated using the meth-
ods of Hazcbroek, et al.~ We note that these pressures approach
the initiaJ parting pressure of the reservoir (approximately 4250
psia in the Lower Clcarfork), and are well above the initial bubble
point pressures of both the Upper and Lower Clearfork (Table 1).
The bottomhole injection pressures are all above the initial forma-
tion parting pressure.
Wc note that the type of water injected (fresh or produced) seems
to have Iittlc affect on the longevity of the tests, however, wells
with produced water injection are more likely to have pore plug-
ging and wellbore fill problems.26 The falIoff time seems to be
primarily a function of how much pressure support the wells are
receiving from offset injectors.
Comparison of tbc pressure buildup and falloff tests on NRU
3=*A1------
J>lu L-... >GVLAG
Mluws 11(JW . . ... . . *h:. menhla rtmy be. ~le b~i!dup f~!
Uua pluudn
run after the well had been hydraulically fractured (1987), and
prior to conversion to water injection (1989), indicated that the
WCIIhad no cffcctivc fracture length. The WCI1was reperforated
and acid-stimulated during conversion, bu~was not refractured.
The fafloff results show the well currcntJyhas an effective fracture
length of 492 feet, which must be the result of continued water
injection at cxccssivc bottomholepressures.
It appears that the fractures crcatcd by this excessive injection
have propagated along the preferential fracture direction in the
Clearfork, which is approximately east to wests. We see from a
falloff test run on NRU 301, that direct communication exists
with an offset injection well to the west (NRU 2601). A 200 psi
injection pressure incrcasc at NRU 2601 caused an almost
instantaneous pressure increase at NRU 301, which was on
falloff (Fig. 49). The existence of direct communication through
the iractures .bctwceninjcciors wiii hsuctily reduce he sweep
efficiency of the injection operation. Unfortunately, this is a
common problcm in west Texas carbonate waterfloods, for which
injection pressures must be kept near or above the parting
pressure of the reservoir to maintain injectivity.
If we are able to optimim future injection well ftactum treatments
by prcfcrcntially stimulating only intervals that contribute to pro-

475
12 An Integrated Geologic and EngineeringReservoir Characterizationof the North Robertson (Clcarfork) Uniti
A Case Study, Part 1
duclion in offset producing wells, and arc able to subsequently
lower bouomholc injection pressures to avoid cxccssivc fracture
propagation, then swccp efficiency can bc increased.
Step-Rate Testing
The analysis of slcp-ralc data collcctcd belwecn 1988 and 1993
for eighty-five NRU injection WC1lS indicates that the estimated
formation parting pressure has been steadily increasing from year
to year duc to reservoir fill-up (Fig. 50). The results of these tests
arc used wimarily to .sctsurface injection pressure timits for indi-
vidual i;jcction WC1lS,however, aflcr reservoir fill-up has oc-
curred their utility is limited since the reservoir pore pressure has
tin incrcascd 10the point where it is difficult to accurately esti-
mate the true parting pressure in the reservoir.
Bccausc bottomhole injection pressures at the NRU are near or
above the parting pressure of the reservoir, step ralc tests should
be used Iogcthcr with Hall diagnostic plots and pressure falloff
test analyses, not only to dctcrminc the optimum injection
pressure for individual WC1lS,but also to identify problems
aficcting injcc[ion wcii cfficicmcy. Al: of ihsc Iy=GIhUI
survcilkmcc techniquescan be applied easily and economicallyby
all operators.
Optimization of Completion/Stimulation
Previous completion cfforls have concentrated on the completion
of all intcrwds that were openin offwtting injection and produc-
ing WCIIS,without regard to rock quaiity. By utilizing the inte-
--..-..I lGSV1
gldlw
.,..,. -,,.:. ~... .:fit;fim.-C,IIIC
VU1l Ubacllpna, ..-OU.W,
mnrm
. ...
mnnhas~
...y..
on maintaining conformance between producers and injectorsonly
over intcrvats of the reservoir which effectively contribute to oil
production. In addition, efforts can be concentrated on maintain-
ing injection over the intervals that can achieve and maintain high
injectivity, instead of randomly injecting fluids into intervals with
high porosity that may, or may not be effectively conncctcd.
Additional completion and stimulation work can bc optimized and
costs can be rcduccd.
Pav Dclincatioq
The NRU was dcvclopcd using an aggressive 20-acre infill
drilling program between lhc time of unitization(March 1987)and
early 1991. During this time penod, 116 new 20-acre producing
WC1lS were added, and 107 of the 141 original 40-acre producers
were convcrtcd to water injectors.
mu-- tu- :nnpIGUILAILaLWII
rmn
-...1 ... --+.,:-., ,.:,
w
h,.. :,. GII drill;
UK. IIIU.L UIMISIS
fi.n,.in
pu~.
ramctcrs were established to identify the payquatity intcrvrdsin
each WCII.These paramclcrs included porosity, water saturation,
and bulk water volume. Any inlcrval having a combination of
porosity greater than 3.6 pcrccnt and water saturation less than 65
pcrccnt qualified as potential pay rock. These parameters were
used with only stight variations throughout the 20-acre infill pro-
gram. Duc to the fact that these simple pay cutoffs do not identify
pay rock on tic basis of reservoir rock quality or continuity, we
feel that there may be additional uncontactcd pay rock within the
reservoir.
The major concern with regard to WC1lcompletion work during
each phase of Unit dcvclopmcnt was the need for a Iimited-entry
type fracture job to ensure that the entire productive section was
being treated cquatly. The gross completion interval extends from
the top of the Gloricta to the base of the Lower Clcarfork (1200-
1500 feet). This interval was complctcd and stimulated in two or
three separate stages, depending on the characteristics of individ-
ual Wells.
0ptimii40n of the fracture treatment program has been an ongo-
ing process during Unit dcvclopmcnt. RCSUIISof the fracture
Lrcatmcntson the original 40-acre primary producers were poor
duc to the fact that the bottomhole treating pressun?could not be
maintained at a sufficiently high lCVC1for fracture propagationduc
to burst limitations on the casing.
SPE 29594
At the time of unitization, the average fracture job was approxi-
mately 1,000 barrels of fluid with 100,000 pounds of sand (l-8
pounddgallon). Since there arc no effective large-scale barriemto
fracture propagation, sufficient non-perforated intervals must be
maintained 10prevcnl communication bctwccnsuccessivecomple-
tion stages. Over the dcvclopmcnt hislory of the Unit, the number
of perforations per stage have been reduced in order to maintain a
limited-entry type of fracture. During the 20-acre infill program,
the optimum number of perforations pcr stage was determined to
be one perforation for each barrel per minute (BPM) injection rate
using a 2-D Perkins/Kcm (PKN) fracture moriei. Fraciure jobs
have been designed to create fracture half-lengths of 120 fec~
As wc have seen from the analysis of pressure buildup data on
some of the 20-acre producing wells drilled in 1987 and 1988,
these optimized stimulation treatments did still not create effective
pressure sinks at the wellbore duc to insuflicicnt fracture propaga-
tion. Future fracture jobs must be designed to create vertically
contained, fairly short, high conductivity fractures. Previous pro-
duction
._ __4:_ history
.-:4-. has ,- shown that regardless of the degree of reser-
-A .,... ,A*
volr tWnununy, lung ii C-:. .... . . .-,a llUL
dbUJL G* db
fit SLS.6U=FKY,
,.8,.
a.. .-.+
-U.-
%rt= in fart
-.- ... . . .
harmful to watcrftood sweep efficiency. If wc complete and
stimulate only the most continuous layers of the reservoir, then
In. ~m&+-arllic
.v.. . . . . . . . . . .frar-[IImc
. ..- --- am ----
- not -r@@cd.
=
Procedures
Recent efforts by service companies and operators in the Permian
Basin have resulted in a ncw fracture treatment called a pipeline
fracture. This type of fracture trcatrncnt can bc summarimd as
follows:
c~~ iM p!a@
Perforate only the top 5-10 pcrccnt at high density
Pump normal fracture job (75,000 to 100,000 lbs of
sand)
Monitor pressures and rates carefully to ensure lateral
growth and effective proppant distribution
Initialc forced closure and flow well for cleanup
Add additional perforations in payrock intcmls at
normal perforation density
By utilizing this technique, the stimulation of large perforated in-
tervals at one time is eliminated. The downward growth of the
fracture is Iimitcd and the proppant material is well distributed.
Downward fracture propagation is a major concern in the Lower
Clearfork at the NRU due to the presence of tie water-filled
-...
ill 10Q7
. . . .-w,
l-rl-
, =- Clcarfork Lime at the base of the producing interval. This tech-
nique has been applied successfully in the Lower Clcarfork by an
offset operator. Through the implementation and design of more
effective hydraulic fracture treatments such as the pipelinefrac-
ture, we hope to improve fracturing efficiency at the NRU.
Coiled Tubing Workovers
A woikovcr program ins been cicvisccito work over WIhijcckms
that show significant Iosscs in injectivity over any given six
month period. Coiled tubing has proven to be a viable method of
operation for cleaning out and stimulating injection wells at the
NRU. The numerous casing problems encountered when entering
wellbores that are now almost forty years old (original 40-acre
producing WC1lS)make conventional well workovers extremely
risky and costs prohibitive. Leaking packer seats and collapsed
casing strings were the major problems encountered. By using
coiled tubing, injection packers arc left in place, and the exposure
of the casing to corrosive simulation fluids is minimized. The
results obtained from coiled tubing treatments have been about
equal to those obtained convcntionatly, and stimulation costs have
been reduced by approximately onc-th.ird.
Cased-hole Logging
We plan to usc radioactive tracer logs in an effort to monitor the
preferential fluid movement in the near-wellbore regions of the
476
 L.E. Doublet, P.K. Pande, M.B. Clark, J,W. Nevans, and T.A. Blasingame 13
SPE29594

injection wells. These surveys yield the best quantitative results
for injection profiling work, can bc rccordcd fairly inexpensively
on a periodic basis, and will bc onc of the main components of
our reservoir swvcillancc plan. The tracer surveys will aid in the
optimization of our injcclion WC1lconformance work by verifying
that the inlcrvais wc utilix as the result of our integrated reservoir
description arc, in fac~ continuous bctwccn injectors and pro-
ducers, and by identifying any potential thie~ zones that might
still exist in the nxervoir.
The use of temperature logs would be preferable since they do not
produce Ihe radioactive residue that accompanies the use of tracer
surveys, however, they arc not especially effective in mature wa-
t.#Ifitdc w hprc Ihere May noL bc a ljiscerniblc downholc tem-
~~~~~~;~i%~n-d~c-to the voh.rmcof water injeetcd, and in this
caiw, an cxtrcmcly low initial reservoir lcmperalure (110W). The
rcsu!M Or this survey would probably only give an cxtremclY
qualitative injection profile. The usc of flowmeter surveys was
also considcrcd, however, the length of time required for well
stabilization does not allow for proper tool calibration and negates
the usc of the flowmeter for quantitative analysis, akhough
Use of all available dam to optimize hydraulic fracture
treatments
GEOSTATISTICAL SIMULATION
Geostatistics will be utilized in this project to develop spatial rela-
tionships of reservoir description variable-sof interest at unsam-
pled (interwell) locations across the Unit. Geostatistics was origi-
nally dcveloped~ for applications in mining engineering, and has
been increasingly used in reservoir engineering to characterize
reservoir properties.so In this paper, the geostatistical techniques
to be used for the NRU project are described. A subsequent
paper will describe the results.
Wc shouId be able to identify the dcgr~ of r~rvoir continuity by
weighting input data based on the method of acquisition, quaiiiy,
and scale. Reservoir properties and heterogeneities can be effec-
tively defmcd using four scrdelcvelsW31
Microscopic
micro scale data
pore and grain size distributions

flowing passes could bc made and a qualitative interpretation . pore throat radius
could bc perforrncd. rock lithology
&3duc~*
Macroscopic
While the cost associated with recording a great iiurnbc~ of - core seaic data
thermal neutron capture cross-section logs (TDT) may bc cost
prohibitive for most operators, the periodic utilizationof TDT logs . permeability
is amextremely useful too~for monitoring the preferential fluid porosity

movement in the near-wellbore regions of producing wells. We saturation

will run approximately tcn surveys in order to get updated fluid . Wcttabfity
saturation values for some of the 20-acre producing WC1lS drilled
in 1987-1991 for use in the history matching segment of reservoir Mcgascopic
simulation. -
.:.....1..;,wAd hlnelr
C4U1U aml. &au ,-n
w-ale
...
&Q

Previous reservoir surveillance in Clcarfork watcrfloods has not Q wireline logs

included the use of TDT logs because they do not work WC1lin the
low porosity, low salinity conditions that exist. Advances in tool
design over the past five years have produced a tool that works
well in both fairly fresh water and low porosity formations. If we
achieve posit.ivcresults using this new generation tool, it may be-
come a part of our reservoir surveillance and monitoring program
for the NRU.
The use.of ordinary gamma ray logs to identify preferential flow
paths will also be considered. Depicted intervals and zones
through which reservoir waters have passed arc likely to be lined
with uranium salLs,which can be easily dctectcd using a natural
gamma ray tool. This phenomenon was noted on correlation logs
rccordcd prior to the tomographic survey on NRU WC1lS 403 and
207, and may be useful in delineating the reservoir flow units.
Data Acquisition - Infill Drilling
During the infill drilling phase of this project, wc hope to acquire
as much ncw data as is economically feasible to verify the results
of our previous geologic and engineering studies. This inchtdes:
Production data
- Verificationof simulation results
Additional whole core for anatysis
- Vcritication of geologic model
- Verificationof reek-log model
- Special core analysis
Ncw generation open hole well logs
- Usc of l.hcFMI tool to identify intervals with substantial
secondary porosity and possible bypassed pay
- Usc of the RFf 1001to record pressure data irreach of the
iacnt.ifiedreservoir iaycrs io cicterrnhmif tiiejjare in
communication
- Use of borehole imaging tools to confirm the preferential
fracture direction in the reservoir
Pressure transient data acquisition
seismic data
Gigaseopic
reservoir serdedata
pressure mnsient tests
geologic model
In order to utilize the different types of data and measure reservoir
properties on a common scale (such as a reservoir simulator grid
block), the affect of the support volume of each data type must be
accounted for. In addition, the volume scales of different types of
reservoir heterogeneities must also be deseribed in order to model
reservoir performance. By utilizing an integrated modeling ap-
proach, in which personnel from all disciplines of the geoseienccxs
are involved, we will obtain the most complete memoir descrip-
tion possible.
Conditional Simulation Methods
Two conditional simulation methods, simulated annealing and ge-
netic algorithm, will be used in this project. The advantages of
using conditional simulation techniquesover conventionalinterpo-
lation are as follows:
Unlike simple interpolation or extrapolation, conditional simula-
tion honors the entire sample data distribution rather than reducing
the spread of the data distribution. This is importanl for retaining
extreme values (outlicrs) irrthe sample data set, which form a very
small part of the overall sample, but which may greatly influence
the flow performance of the reservoir. An example would be a
smatl streak of high permeability, which can have sigrdflcant in-
fluence on waterflood pcrformancc, and still constitute a very
small part of the entire productive interval in terms of the total
sample distribution.
The second advantage of the conditional techniqueis that it honors

477
14 An Integrated Geologic and Engineering Reservoir Characterizationof the North Robertson (Clearfork) Unit
the spatial relationships developed from the sample data. Many
conventional inicrpolalion methods gcncratc smooth distributions
which do not salisfy the spalial relationships established using the
sample data.
TIN last advantageof the conditional simulation method is its abil-
ity to quantify uncertainties in the reservoir description through
muhiplc, equiprobablc images of the reservoir. CondhionaJsimu-
lation alIows construction of multiple picmrcs of the reservoir, all
observing the same constraint(s).
Simulated Annealing and Genetic Algorithm Methods
The simulated annealingand genetic algorithm methods are attrac-
tive since they arc very robust and flexible and allow
incorporation of various scales of data in describing the mcrvoir
properties. This includes geological, petrophysical, cross-
borcholc seismic, reservoir performance, and pressure transient
data. Onc disadvantage of the m~uwu .I--A :. .L. * ~kn .Ionrithmc
IS U,~L u,~ ~.5v..
. . . . . -.+
slow and compulationally intensive, however, they still run faster
than conventional reservoir flow simulators with the same number
of gridblock.
These methods involve the definition of an arbitrary objective
function
... which
. must bc minimized. The more constraints (data
typcs) mm arc- mlpuwu
:--...~ U1lL
-- t...,~k;~~i;- f,,nrtinn the more
Uu,wu . .....-.., ---------
com-
putationally intensive these methods become. As wc intend to in-
corporate all available data types in order to produce the most
complctc reservoir description possible, these methods best
satisfy our rcquircmcnts. With a ncw generation of high-spud
personal computers coming on to the market, computational
demands should not be a problem, and these techniques can be
used by all operators.
Approach at North Robertson Unit
..C pLUU_Y
..2 . . . . ~m,,
n.., swwv..u-,
.rt-mdaw nav
The interwcii distribtttioii UI =-,, and
-.
non-pay rock will be generated from intrawell rock type data
using the conditional simulation techniques described above. The
project geologists will then review these rock type distributions
and determine if they honor the geologic model developed for
NRU (qualitative check). After realistic rock type distributions
have been achicvcd, pctrophysical parameters (porosity and
permeability) will bc assigned for each of the pay rock intervals
by honoring the poinl data (cores and WC1llogs), incorpora~g
tomography data 10 understand the interwell changes m
pctrophysical properties, and by honoring reservoir perforrnancc,
pressure transicn~ and other available cngincxxingdata.
The geostatistical analysis will bc extremely useful in identifying
the best infill locations within the high quality areas of the reser-
voir, however, the quality and quantity of data is probably insuf-
ficient to assign a relative probability of success to any particular
infdl well location.
RESERVOIR SIMULATION
An important objcctivc of this project is to perform 3-D reservoir
simulation for history matching, infill drilling dcvclopmcnt fore-
casting, and validation. The results of convcntionrddeterministic
simulation runs will bc compared to the results of stochastic simu-
lation runs (using gcostatisticrdrealizations)to optimize rewlts. A
L.._,_ _., .L_- -A..a
OlitCK oil, Ulrw-plhl.%
:...,.
mlllu atul
].n.
+ u. -
. ,,lili7mi
,;11
u-..-.
% we will fQcMs
three main areas:
Selectionof optimum infiil drilhg sites within the North
Robertson Unit
Predictionof future reservoir performance
Validationor comparison of prcdictcd and actual reservoir
-K.MS.t.W_
pt:va..am.v .I@+-
....~ I.-------
hi? Rcld Dcmonsuation phase of the
prOJCCt
vs. P~ .,-.
t the desirable areas for infill drilling. llIcdarker regions rcprc.umt
The types of models which arc being constructed arc pariiai-urdi
models. The areas for these partial-unit models have been
SPE29594
selected based primarily on an understanding of the reservoir
performance factors discussed below. In addition, the l~ations
of the parlial-unit models have been vcnficd by cons]dcrmg the
geologic model and the results of the dcclinc type curve analysis.
A full-unit model wiIl not be constructed duc to the large number
of WC1lS in the Unit (259 WC1lS), and the ned to focus on detailed
flow simulation in areas with the best potential for irttill drillig.
By using partial-unit models, we will be able to reduce both the
number and six of grid blocks to optimize results. A full-unit
simulation would result in a cumbersome model with a large
nurnbcr of grid blocks primarily duc to the large vertical section in
the Gloricta/Clcarfork and the large number of layers required to
reptesent this vertical hebxogcneity (approximately 15-20 layers).
Reservoir Performance Criteria
Reservoir performance factors have been considered to delineate
the areas which possess good potential for itilll drilling from
are
those with little or no potential for infill drilling. These perfor-
mance attributes for selectingsimulationareas at NW am
of high productivity
Areas
- High prima~ and secondary rccovcry.
- presence of pay rock types (rock types 1 and 2).
- Good iXXO@ant!Permeabilitycharactistics
Areas of poor reservoir continuity
. Good primary recovery but poor secondary recovery.
- Poor waterflood pattcm balance of water injected to fluids
produced
- Current production with high oil cut and relatively low
secondary production
- Primary decline much higher than normal primary and
secondarydecline (idicative of compartmentalization)
Good pattern brdanceof water injectedto fluids produced
Flat or increasingoil cut (maybe indicativeof good
waterflood sweep efficiency)
High ratio of secondary estimatedultimate recovery (EUR)
to primary EUR.
Uniform increase in pressure in surrounding areas indicating
good reservoir continuity.
Selection of Modeling Areas
A selection or scoring criteria was devised to identify the desirable
locations for infill drilling basedon the following readily available
reservoir performanceparameters
Q cmm&&.VGyluwzl,
..&-., m.t-vi,,.,inq
p-u-u.
{CPP)
,-. . ,
Curmdativcprimary to secondary recovery ratio (PSR)
Cumulativereplacementratio (CRR)
Water/oilratio (WOR)
For which the cumulative replacement ratio (CRR) is defined as
ratio
on
. ..between
.,.
the cumulative volume of water injected and the cu-
-c .-. ~ -...,4,.,..4 O.,; A.
miuauve volume WI umd pIUUUWu ~lww.
~rnr+,rtin. and inia~on gjata WSS dloeatcd to 5-spot waterflood
~~fi;~~~.&.;c-~~oir performance parameters were calculated
for each CC1l.Average Unit values for each of the scoring criteria
were calculated, and each cell was assigned onc scoring point for
having a higher than average CPP, PSR, or CRR, and a lower
than averageWOR.
The results of the selection process are shown in Fig. 51. The ar-
eas which are shaded lighter (higher score on O-4scaiej mprcsent
areas which may bc undesirable for infill drilling. Theac desirable
areas coincide with those idcntiiied by reservoir p~~f~fintii~~
maps generated from decline type curve analysis on the 40-acre
478

SPk 29594 L.E. Doublet, P.K. Pande, M.B. Clark, J.W. Nevans, and T.A. Blasingame
]rimary producing WC1!S (Figs. 31-34). In addition, these desir-
~blcinfill drilling areas arc consistent with other reservoir perfor-
mancecharactctistics dcscribcd above, such as the consideration
of primary and secondary dcclinc rates to identify areas of poor
~rvoir connectivity.
Four areas in the Unil have been selected for detailed reservoir
simulation (t%g.51). Areas 1,2, and 3 have good apparent infill
drillingpotential. Area 4 has poor apparent infill potential but will
be simulated for validation purposes. The total number of wells
in each modeling areas ranges between 18 and 32 wells.
!fljec:ors ~~rv~~ boundary WC1lSfor all the simulation areas.
This contlguration was chosen since it is more practical to allocate
injection than production in the boundary WC1lS.Also for alloca-
tkm pu~oscs, the injection rate data are more reliable than the
producuon data since water injection commcnccd more rcccntiy
than production.
Reservoir surveillance activities, consisting primarily of pressure
transient tests rccordcd to monitor reservoir pressure and forma-
tion flow characteristics, and TDT logs run 10 monitor water
movement in the reservoir, will be focused on the modeling areas
to obtain additional information for history matching and devel-
opment forecasting.
Simulation Initialization
Initialization of each simulation model requires phase behavior
(PVT) data and rock-fluid interaction parameters. Each of the data
types required for initiatiixuionarc discussedbelow:
.. . .
PVT D~
The analysis of available fluid data has conclusively established
that the fluid properties of the Upper and Lower Clearfork reser-
voir fl.uk-isare different and need to bc treated with two separate
PVT regions during simulation to properly represent the phase be-
havior interactions in the reservoir.
Some of the differences between Upper and Lower Clearfork
reservoir fluid arc illustrated in Table 6. The data on the table are
based on black oil PVT laboratory studies conducted on fluid
samples obtained from NRU 3522 and 3013 during 1991.
The feasibility of using the original PVT data from bottomhole
samples acquired on offset leases in 1947 (Lower Clcarfork) and
1958 (Upper Clcarfork) was considered. The utilization this data
(Table 7) was considered by using a phase behavior simulator to
match the original and rcccntty acquired (1991) PVT data. The
validation results indicate that the original data may be used 10
rcprcscnt PVT properties since the data were found to be consis-
tent with laboratory fluid tests conducted on the surface recom-
bined samples collcctcd in 1991.
The accurate representation of the initial fluid data, along with the
irnegratcd reservoir characterization wc have undertaken, witl
allow the physical proccsscs occumirrgin lhc reservoir to be accu-
rately modeled. By using the original data, the simulator will
automatically adjust original properties to fluid properties at any
subsequent time in the history match or forecast. The fluid prop-
erties for the wide range of depletion and repressurization paths
can be properly rcprc.scnlcd.
Rock-Fluid Inlcraction Data for Initiali-zation
Existing
._. special
, core data which will bc used for the simulation
arc pnmamy rciadvc pCimGduJIJIY ..l.:l :,.. ,im,.,
uau.
A ,mlml nf !$;limw
A Wwl1
CtfWtiV.
. .. .. ...--...
state displacements for two cored WC1lS(NRU WC1lS207 and
3522) have been conducted. The displacementsinclude data from
th. I~~-.,
Inm-r .Miri~le and ~wc~ ~l~ork. Additional spcciid COrC
. .... .-, ----
work will bc complctcd when ncw WC1lSarc drilled during the
ticld demonstration phase of tic project.
hrkt~iy Miikh Crkeria
Allocaicd production data (oil, water, gas) have been determined
to be reliable and will bc the primary history match criteria.
Prrxsure data arc available for the primiuy depletion phase of the
history match (1956- 1987). The initial reservoir pressure is
479
15
known, and pressures prior to water injection and during
reservoir fill-up arc available for 45 wells throughout the Unit
(Table 5). Additional pressure data will be available from
pressure falloff and buildup tests that are currently being
recorded.
In the initial simulation and history match it will be assumed that
there is no flux across the areas to be modeitedsince at] of the
model areas lie in multi-pattcmcd waterflood areas of the Unit. It
is possible that this simplifying no-flow bound~ assumption is
invalid due to suspected interwell fracture communication from
pressure falloff test results. Tlds assumption may need to be re-
.--.: ~.-.,~..U1l =. ..,-.h~.eaa
~ulhlu~kGU WWU-UJ
~&s @ ~ch Of the model areas if it
is difficult to obtain a history match and it is determined that the
flow boundary assumption has an important mechanistic role in
the
u.
r+icnlam?ment
.y. ---------
nmccss.
~-------
This project is very much a work in progress. The results of cur-
rent and future analyses, specifically, geostatisticaland 3-D reser-
voir simulation, will be summarized in great detail in a subsequent
work (Part 2) during the field demonstration(inffl drilling) phase.
SUMMARY AND CONCLUSIONS
As a result of this study we hope to identify useful and cost effec-
tive measures for the exploitation of the shrdlow shelf carbonate
reservoirs of the Permian Basin. The techniques that are outlined
in this work for the formulation of an integrated reservoir descrip-
tion apply to all oil and gas reservoirs, but are specitlcrdlytailored
for usc in the heterogeneous, low permeability carbonate reser-
voirs of west Texas.
1. A detailed reservoir characterizationcan be performed with a
- minimum of core data, as long as a competent geologic model
has been constnrctcd, and there is sufficient wireline log,
pressure transient, and historical production data available for
malysis.
2. Aside from the cross-borehole seismic and geostat.isticalsim-
ulation work, all of the data acquisition and analysis tech-
niques used for this integrated reservoir description are read-
ily and economicallyavailableto all operators.
3. The matcriat balance decline type curve techniques summa-
rized in this work give exccllcnt estimates of reservoir vol-
umes (total and movable), and reasonableestimates of forma-
tion flow characteristics. Using this method to analyze and
interpret long-term production data is relatively straightfor-
ward and can provide the same information as conventional
pressure transient analysis, witiout the associatedcost of data
acquisition, or loss of production.
4. In order to better identify injection WC1lresponses and imp-
rove the analysis of long-term injection data, we need to
develop type curve techniques similar to those currentJy used
in dcclinc curve and pressure transient analysis.
5. At the NRU, we sce the same problems that arc associated
with the majority of hctcrogcncous, IOWpermeability c~-
bonatc reservoirs--a lack of reservoir continuity, low water-
flood sweep efficiency, early water breakthrough, and water
channeling.
6. Surface pressure acquisition during pressure falloff tests
yields data of sufficient quality for interpretation even when
low precision pressure gauges arc utilized. This is a cheap
..,i A.
-s
.rfxt
-b..
;wnwatmflond
. . . . . . . --
sllr$fCfl]aCC
-
100!.
7. The preferential fracture direction at the NRU appears to be
cast-west. Several of the injection wells are in communica-
f:fi .,;. hwdm resu!ting from long-
L,r,,,- 1.J
IIliral
AM..-..
IV induced
J .... . ...-
fr~c~u~M
tcrm injection at pressures WC1labove the fracture pressure of
the reservoir.
0
0.
t~r..-.
W awl
-.. -!;s., uluil..
quauty
---:1 cl.s.m. @h I)lrl k a mainr
~ -..0. - .-.-J-- ~
n~. ~f~ Cff~-
tivc waterflood surveillanceplan.
9. The results of the previous hydraulicfracture treatments at the
NRU have been extremely poor, resulting in extremely shofi
low conductivity fractures.
16 An Integrated Geologic and Engineering Reservoir Characterizationof the Notth Robertson (Cltiork) Unit SPE 29594
A Case Study, Part 1

10. As with all oil field operations, wc recommend that quality REFERENCES
data be taken early and often to ensure more accurate analyses 1. Ghauri, W.K., et ak Changing Concepts in Carbonate
and intcrprctalions. Watcrflooding - West Texas Denver Unit Project - An
NOMENCLATURE
IIlustrativc Example, JPT (June 1974) 595-606.
2. Stiles, L.H.: OptimizingWaterflood Rccovcty in a Mature
Field Vatiabks Waterflood, The Fullerton Clearfork UniL paper SPE 6198
Formation and Fluid Parameters:
presented at the 1976 SPE Annual Fall Technical
A= draina~carea, fL2 ConEcrcnccand Exhibition, Ncw Orleans, LA, Octobtx 3-6.
B= oil formation volume factor, RB/STB
cl = total system compressibility, psi-l 3. George, C.J. and Stiles, L.H.: Improved Techniques for
cti = initial total systcm compressibility,psi-l Evaluating Carbonate Watcrfloods in West Texas, JPT
= porosity, fraction (November 1978) 1547-1554.
t= formation thickness, ft 4. Ghauri, W.K.: Production Technology Experience in a
S~irr = irreducible water saturation, fraction Large Carbonate Waterflood, Denver UniL Wasson San
k= forma[ionpermeability, md Andrea Field, JPT(Scptcmber 1980) 1493-1502.
re = reservoir drainage radius, ft 5. Barber, A.H., et ah: Infill Drilling to Incrcasc Reserves -
rw = wellbore radius, ft Actual Experience in Nine Fields in Texas, Oklahoma, and
rwl = apparent wellbore radius (includcaformation damage Illinois, JPT, (August 1983) 1530-1538.
or stimulation effects), ft 6. Barbc, J.A. and Schnocbclen, D.J.: Quantitative Analysis
P= fluid viscosity, cp. of Infill Performance: Robertson Clearfork UniL JPT,
Pressur~ate/Time Par-ters: (Dcccmber 1987) 1593-1601.
FctkovichfArpslSdcclinc curve exponent 7. Calhoun, J.C.: Fundamental of Reservoir Engineering,W
constant in the pscudostcady-statequation for liquid University of Oklahoma Press, Norman, (1982).
t-low 8. Ehrlich, R. and Davies, D.K.: image Anaiysis of Fore
oil flow rate, STWD Geometry: Relationship to Reservoir Engineering and
% original oil in place, STB Modeling, paper SPE 19054 presented at the 1989 SPE
NP cumulative oil production, STB Gas Technology Symposium, Dallas, TX, June 7-9.
N p,mov movable oil, STB 9. Clelkmd, W.D. and Fens, T.W.: Automated Rock Chamc-
P pressure, psia tcrization with SEM/Image Analysis Tcdmiqucs,n SPEFE,
Pi initial reservoir pressure, psia (December 1991),437-443.
Pw/ flowing bottomhole pressure, psia 10. Wardlaw, N.C. and Cassan, J.P.: Estimation of Recovery
Ap PCPUII, pressure drop, psi Efficiency by Visual Observation of Pore Systems in
r radial distance, ft Reservoir Rocks, Bulletin Canadian Petroleum Geology,
t time, days 26 (1978).
i N~q, material balance time, days 11. Thomccr, J.M.: Introduction of a Pore Geometrical Factor
s Vari @ Real Domain Dctincd by the Capillary Pressure Curve, JPT, (March
1960), 73-77.
911i = dimcnsiordcssdcclinc rate function as dcfiicd by
Fctkovich18 12. Hutt, P.R., et ak: Applications of Cross-Borehole seismic
Q-M = dimensionlessdcclinc rate integral as dcfmcd by Tomography to Monitor EOR DisplacementFronts in Heavy
McCraylg Oil Fields of Kern County, California, paper SPE 19855
q~ = dimcnsionkss decline rate integral derivative presentedat the 1989 SPE Annual Technical Conference and
function as defined by McCraylg Exhibition, San Antonio, TX, October 8-11.
rD = rlrw, dimensionless radius 13. Harris, J.M.: Cross-Well Seismic Measurements in
r~ = dimcnsionks.sdrainage radius of reservoir Sedimentary Rocks, paper SPE 19871 presented at the
s = skin factor for near WCIIdamage or stimulation 1989 SPE Annual Technical Conference and =hiiiition,
Im = dimensionlessdcclinc time as defined by San Antonio, TX, October 8-11.
Fctkovich18 14. Baker, L.J. and Harris, J.M.: Cross-Borehole Seismic
Imaging, presented at the 1984 Annual Meeting of the
Dd = dimcnsionlcsadcclinc variable SEG, Atlanta, GA, December 2-6.
MP = match point 15. Bregman, N.D., Hurley, P.A., and West., G.F.: Seismic
pss = pscudostcady-state Tomography at a Fire-Flood Site, Geophysics, (1989) 54,
i = intcgrdl 1082-1090.
id = integralderivative 16. Chon, Yu-Taik: Crosswcll Bcd connectivity Analysis,
o = initialvalue prcscntcd at lhc 1993 SEG International Exposition,
ACKNOWLEDGMENTS Washington, D.C., Scptcmbcr 26-30,
Wc acknowledge the permission to publish field data and the i?. -----
I%,,h]et-.. -.-.,
r. 1? -. . ... Dc~iinc CIMVe Ana!ysis Using TYPC
et al.:

cmmputingsuppor~scrviccs provided bv Fins Oil and Chemical,
Co. (Western Division, USA). In particular, we would like to
thank Debbie Taylor, Manjou Banthia, and Susan King for their
help in indexing large amounts of data.
We also acknowledge the technical and computing support scr-
viccs provided by the Department of Petroleum Engineering at
Texas A&M University and the University of Tulsa, as well as the
financial support of the United States Department of Energy
(DOE) for funding provided through the DOE Class II Oil
Program.
Curves--Analysis of Oil Well ProductIon Data Using
MaicriaiBaiancc Time: Appikatiim io Fkid C2iSes, papiX
SPE 28688 presented at lhc 1994 Petroleum Conference and
Exhibition of Mexico, Vcracruz, Mexico, October 10-13.
18. Fctkovich, M.J.: Dcclinc Curve Analysis Using Type
Curves, JPT (June 1980) 1065-1077.
19. McCray, T.L.: Reservoir Analysis Using Production De-
cline Data and Adjusted Time, M.S. Thesis, Texas A&M
University, College Station, TX (1990).
20. Palacio, J.C. and Blasingame, T.A.: Decline Curve Anrdy-

480
SPE 29594 L.E. Doublet, P.K. Pande, M.B. Clark, J.W. Nevans, and T.A, Blasingame 17

I sis Using Type Curves: Analysis of Gas Well Production

Data , paper SPE 25909 prcscntcd at tie 1993 SPE Rocky
Mounlain Regional/Low Permeability Reservoirs
Symposium, Denver, CO, April 12-14.
21. The Production Analyst Program (Version 2.3), OGCI
Software, Inc., Tulsa, OK, USA, 1993.
22. Jordan, J.K.: Reliable Interpretation of Watcrfioori
Production Data, JPT (August 1958) 18-24.
23. Hall, H.N.: How to analyze waterflood injection WC1lpcr-
formancc: World Oif (October 1963) 128-130.
24. Thakur, G.C.: Watcriiood Survciiiancc Tcchniqttes - A
Reservoir Management Approach, JPT (October 1991)
1180-1188.
25. Robertson, D.C. and Kclm, C.H.: Injection-Well Testing
to Optimize Walcrflood Pcrformancc, JPT (November
1975) 1337-1342.
26. Ncvans, J.W., Pandc, P.K., and Clark, M.B.: Improved
Reservoir Management With Water Quality Enhancement at
the North Robertson Unit, paper SPE 27668 prcscntcd at
the 1994 SPE Permian Basin Oil and Gas Rccovcry
Confcrcncc, Midland, TX, March 16-18.
27. Banthia, B.S., Meyer, B.J., and Blasingamc, T.A.: Use
of Surface-Derived Pressure Mcasurcmcnts for Cost-
~rrnt-f;.,n
b,-.. DmwlwoirSu.Wciiizn~eof Waterflood Operations
.* *i-
n the Pcrrnian Basin, paper SPE 27685 prcscntcd at the
1994 SPE Permian Basin Oil and Gas Recovery
Confcrcncc, Midland, TX, March 16-18.
28. Hazcbrock, P., Rainbow, H., and Matthews, C.S.: Prca-
surc Fall-Off in Water Injection Wells, Trans., AIME
(i958)5 2 i2S0-260.
3:

29. Mathcron, G.: i%e Theory of Regionalised Variables and

its Applications, Cahicr No. 5, Ccmre de Morphoiogie
Mathcmatiqucdc Fountainblcau(1971).
30. Kclkar, M.: Introduction to Geostatistics, Tutorial Paper
prcscntcd at the 1991 Intcmational Reservoir Characteriza-
tion Confcrcncc, Tulsa, OK.
31. Kclkar, M.: App[ied Geostatistics For Reservoir Character-
ization, draft, University of Tulsa, Tulsa, OK, (1993).
32. Igor-Graphing and Data Analysis Program (Version 2.7),
WavcMctrics, Lake Oswcgo, OR, USA, 1992.
33. panSystcmTM-Well Test Analysis Program (Version 1.8),
Edinburgh Petroleum Scrviccs, Ltd., Edinburgh, Scotland,
UK, April 1991.
34. Tcrrascicnccs Reservoir Analysis Package (Version 5.4),
Tcrrascicnccs, Inc., Littlcton, CO, USA, 1994.
35. SAS (Version 6.08), SAS Institute, Inc., Cary, NC, USA,
i 993:

481
 An Integrated Geologic and Engineering Reservoir Characterizationof the Norlh Robertson (Clearfork)Uniti SPE 29594
A Case Study, Part 1

TAIII.11 1 - Reservoirand Fluid Property Dala for T/\13LE 4 - QuantitativeAnalysis of Mereury-Air Capllhry
the North Rohcrtsnn(Clearfork) Unit. Pressure Data, alter Thomeer (Ref. 11).

R k ronl fksdiuz I)isplmccmcnl Iner[cclivc

Reservoir Prnpertlrs T;PC (#m) Pressure rorudly al Soa
(psi.) p&$:Jcut~m
WcllhmC rdiw. r. 0.3I fed
in P
EUkMcd avcrakx gnm pay interval 1200- Imft-et

Avzncc net Fly thickncm.h unknown I 1 I 76-S30I 2-10 I 2-296

Avemgc Pormily. e 0.07s

AVUUZ irmkibfe water SWALIII, S=.n 0.30

f%maiim fmmdlilily. i O.1-lonul

Rcrerwir IMIPCCnUrC 1l@P

C2tiginal nrmiml well SPzcing 40 ame:

Currm mmiml -11 Sp2ring 20 -m

Fluid Properlirr I 5 I ,.07-,.78 I 60.,s0 I 21.7 -57.2

Jnitid saruratimpmsum 17CK3
psia (UCF)
1S40 psia(fIF)
Initial brrm!ion volume factor. B 1.285 RBKTO (UCFl
1.382 Ro/sTf3 (LCP)
I 610331mlm
lnitid oil viscosity.II I .05 q (Ucrf 7 ..
0.81 Cp(La=)
Initial I.JUIccwnfncs$itility.cti 12.OXl@ P.@
8 .- .-

Initial API oil grwuy al 60#/tC+ 34.40 (UCF3

39.W WFi
Average fomu!ion volume kclor, B 1.30 R0/STfS

Averageod viscosity.II I ..30Cfl

Avmsc owl compms<ihili!y.rti TA23LE S - S20tlomhole Premore Dsts For Reservoir Simulation History Rfstchimg.

Prior 1* Unllllslkn rruydnnyP#sd ed ~,~-~g, ,nu.p

Production Parameters (3/s7)
NO. IDIIP Wrl-L No. SIBNP W5LL rut SIBHP
hlual IXXrvuIr pressure.p, 2(125psiu(UL19 tplis) (+ I
tfnia)
2%Uf psi. (LCFI I(JI 201 Iv>)
F6nwinghti!,mlmle pressure.p,.f unknown 2M 651 m 119
s 20! 452 207 11s1 .W am
Ian 9s5
7 41M !320
# 2WI X6 4.X)2 207 4N1 ilso
9 lKII E64
T.\Dl,E 2 . k%rr Types Idcntilied FromCoreThin-srclion% XRD. and SEhl Anmhsis. 291 1 I(I2 lt.a 21UI 7(J6 1I(N I.w
110s 1247 I
1%,, Size hp. C..rdin.d.a Aped R.trn A178wIIC.1 G*.1.#1. 1010
M* (Pm) Is lna.lplc+
Number
3106 >.s0- Iw.1
Orscrtpll.
1
2203
2ZOP
>24
mo7
176
2217
2224
222S
IJXI
1074
1221
-Tn7
2224
222s
2600
I MO
A I wmlm Omwadck)
I (MtakMc) (h-) lNUrMnR@
I
LG3zk
2211
2216 719
WJ 1Icil
.P*=I Skll hwdl I IM

F
Olu,h) Ildmti DilmMaa 326 nil 953 2701 8431 lrnl
v.,, 2a01 1296
plolwl -
(High) .&l 172 ml I 991
.!:7 .3003 Imt .3D13 lsya
<501 11s9
*

+!-++=+
:
(k-) 157
- I
..
,- .. . .VJ,=
...- .-
II106
,.!,.$
716
ii.;
I I I
i
(-) i8tmmm@ .72s mlrl 47>
3WI Ins 3510 W.2
<2), I 7:- 3517
. .. . . ..
,n, J
>:=
F 3105 rClrsr.hl
-6
tl,gh
(Low) ,MfKmnmd I ,.lmauilw .r2*
3$16
ton
661 3!11
V.w
I 199

I ...
c <3 .Sklm19 (Ink.lc.yudim (:lNm-p4trp4t
I* nm,l.)

TAIILE 6- Itesulls of PVT Analysis on 1991 Surhce Fluid

Samples (Upper & Lower Clearfork)
TAf21.E 3 . Rnck Types Dclintd for the Clrwfork snd Gloricla Fc.rmations.

I
Ffuid Pro~rlirs NRU 3522 NRU 3013
(Scclion 329) (.SectIan 327)
I I
I WOumR A B.C.D 5.4I 5.$2

I l B.C
I
D.E 6.10 I .71
I S2n@i.gDYe I FetItu2rY7,1991 I FebruaIYll,WPl I
I I
3 cmlmlm= c D.E 4.19 0.20
I Rccm+ination Duhhk
Foinl
1.330 pzis 1.30J ptia 1

d ImlmMt 0 F 6.W 002 ,, l*lMmlc

I I I I
L 1 I
. . .5 [-alcld
011Vlsmnily M Lruklhk
Pain!
2.67 Cp 1.32 Cp
I . 1>- c A.D.E.F 7..12 0.14 (W8CI (w-

m
P- oil FMnYii Volum 1.132 RkVSTB 1.280 RWSTB

I *
Fata u Buhhfr ibim

oil 12Mg: frubbfr o.64a gnlkx 0.762 cmfcc

IP6WICIG.91
17
1 0
1-1- Stak Tmk Oil API
Gmvity at MPF
SO.Y 33.P

482
 ,.
19
SPE 29594 L.E. Doublet, P.K. Pande, M.B. Clark, J.W. Nevans, and T.A. Blasingame

PiM4JClt0N/WJECIDN(1956- 1994)
D41LY
TA131.I? 7 - Rcmdts O( PVT Analysis on Original IJotlomhnle
Fluid Snmplcs (Upper & Lnwer Clearfork) 105

I FluidI-rnpcrlim Wdl Fe* 11-21. Wdl I.-m E.4U

I
?Alllc Upper LMuhk Lnwcr Ckulark

10
!!!H!!!!!!!!!!!!!!!!
SumplingO.uc July 16.1958 January 17.1947
r I 1 1 1 1 1 I i
1 1 1 1 8 1 I r 1 1 1 1 1 1 1 1 1 1 [ 1 1 1

Illllllllllljllll I I [ I I 1 I
Wngitwl- Nuhhk Pnml i .7ti~ p~ia ,, ,.W
..,I ..,.
,rw. I I II

I>rmsurc

2,625 psia 29$4)psia

lmtid I@.motr Pressure

tu

1 StockTank Od API
Gnwity at M?T
34.40 39.6
I
10

Tut k.ofs

I Fqure 3. Production@ I@clion Hklory for b Nwlh Robe- (Clsarfwk) Urut

I . .....
I ,,.., ,.
I .... ..
I
;/ ...C*
,++.:(_.L. _._. T. L._. _.r.L
~fl ,+$
I i
-,s I
rl- .- .- A-1.-l1

I
[

C.C.50 & R.c.tlc.,

Figure 1- LosatoOnof the Norlh RObe!tsOn(Clearlork) Unil, Permian Basin, West Texas.
, r--
-\
/f 1.
IL. .
I PUBLICBS:,fMxL MND
-- ,:- ~~:
x< 8
FWro 4. I.oulion d bad Web w+lhiilb North.%berlson (Chatfork) Uml.

-
--------------- J

..<s :
&---- ..-.
IlOanl nOumrawUm
----.xr.,----- -, -.. ..n
Fgure 5- OapOsilknal SnvirOnmn IS - upper wdde ckeaddr, UPW cMW* ~ G1-.
---

w ! le.-

483
20 An Integrated Geologic and Engineering Reservoir Characterizationof the North Robertson (Clearfork) Unit: SPE 29594
A Case Study, Pm 1

F~ra 9. Slmclum M6p - lop 01 Gbrisls Seclion d NRU.

=POREWEA UFORETWEB tZPOmmEc

POREI-ITED PORETY7EE J70REIWEF

PORSTWEO

Figure 10- Rdaliw Vdums Propdons 01%s Typss (A-G) in Esch Rod Type (l-S).

Figure 7- Stfuclure Map. Top d Tubb Seclii al NRU.

+ NRUZ07.7142.S II
NRU~717tll
ONRuazsfl
o tmuwzzmt.ssn
m )MIJZ07.7014R

2Q 40 60 m
\\,c!linf h= *twr.lrn. (%}
Figure 8- Slruclurc Map - Top of Uppsr C1.adti SscOon d NRU.

Figuro 11. Mmwymif Cspissry Pr.ssuro Cunf.s. Rack Typs 1.

484
 L.E. Doublet, P.K. Pande, M.B. Clark, J.W. Nevans, and T.A. B1a.singame 21
SPE 29594

NRU :7wan
ObIiu

1- Cbf Fork

Icku%k

II Chat F&k
I
I

0 :0 S4 1o11
Bm. Satw.nion (e. We u me)
FQWO 15. ClsJ8riOIWG101MaType LW . NRU Well 207.

Figure 12. Waler/Oil Relalive Permeability CuIVes. Flock Type

3,
ANHYDRITE

&J
..,.,. .. ..*C

19
t

10

N
#27 o Qo

Y
CALCITE
QUARTZ
. Rock Type 14
16 0- Psck TYP~ 5
. . Reck Type 6 (anhydritc]
O. Rock Type 7
. . Rock Type 8

4 6 8 10 12 14 16
Usess (@masks)
O.ml 1 1
0 5 15 Zo
CORE POROS% (%)

Figure 13- Porosity/Permeability Relationship - Entire ClearIorWGlorieta Section.

Figure 16- Rock Type Delineation Plot.

lm
m
r- ,.?Hi!M4
8
10



e
1
==
==
==
0.1
==
=:
0.01

0001
0 5 10 15 20 k%-., ,.-1 .- l-l
J
---
r
..
------

-:=2::
~=i=-=-.
Figure 14- Porosily/Permeability
CORE
--.. POROS~ 1%),
. . . . . . . . ,..

Relationship by Rosk Type - Rock Type 1.

1 . .-. .-1

485
22 An Integrated Geologic and Engineering ReservoirCharacterizationof the North Robertson (Clearfork) UniL
A Case Study, Pan 1

cro~~ Bor~ho!~ r~rn.~manhv Data.-

~- r... __. Acquisition
T..-. ..
Im
NW ,Cin.4c,b) ?,, COlhUy btcdan 4&m WM h
PDSA l-survey CO@CtC(! July, 1994
I -s

as
.-.
. .-
...: NRU 40J Injaclor

(\ $#
@ourco)
i A

~ /

20-
~ -: -.. .

-.
. ..-

%--------
E-Haad
.-\.
.

------.:.
.!
.$

* . .-.! ------------

; . .
:
.

:
, NRU 207 Producer /
[Raceivw)
Corad Well
.8$
/

E- 1*-k PDSA III-Survey Planned for February, 199S

I

? z a !
NRU 3a22 Producer

k
07 Corad W*II
6 K&
--fnl
X& & aim .

%
%
Fgure 18- Continuity Study From 40-acre WeU Dala - NRU Saclii$ 5,325,329. NRU 3S1$ Injaeeor
~

B
NRU 3S24 Producer b

*
,\ c?

Figure 2! - Complalad nd Planwd Tomogrvhy Work (Saclbm 5 h 329)

Connectivity Mapping
Processing Flow Chart
A
I ln~ulDaln1
J

Calculate
Transmissivity
Functions

u u i

I
Spectral
COrrdalion
I I
J

NUU 403
]hy Paths for NRU 403 to 207 NRU 207
.50..,., w.]! Tomography Survey R...i... W*N F,gure 22- Processiq Fbw Chatl 101ConnaclMly Mapping.
n n
wnr
(U1rl+.wwcw Connccwity Map I*
WdlI
SP .

I u u
\
Fw*20-Rv PtibIMU~7T_Wb SWWY. Tnmdwau r.urar Iamb

**as01814181a
, # 1
aan
F~ro 23. NRU 40&7 ~ Map tih Q- Ray nd VOlOdty Lcqs.

486
 10 W+ d d WJ lti
l#vJ.zOql

FrIWNe24- FelkO@VMC&Sy qm q= d qm Type CurveS.

Figure29- Match01ProductionOstaforNRUWetl3510- Clesrlodr(RsdmlFlowTypeCuwe).

1 I 1 1
1401
m{ .
p?miq
,-----,
..
. t
tm
4 .
*

Figure25- NRU Well 3S10 - Seclion 329 of NorthRobartacm(Cleariork) Unil, & eo-

1 1
Id

9-U . =ww
0
s. . 0 loomo Smom Smmo 4oomJ
1$ . . . h.., -t N,, L3TB

s
G1
., .:-= Figure 30- Movable 011Estimation from Ra!e History.

10 I
.

1
o 2&l uioo emo mm moo lsooo
low

Figure 26- Semilog Production Plot for NRU Well 3510- (Clearfork).

. .

:

.

nlw -

Fgure 27- Log-log Producticm Plot for NRU Well 3510- (Clearlork). F~rc 31. Malefiel Bdensa Oadine Type Cuwa Analysis - Map or Odginal Oil-in--- MM3TB.

487
24 An Integrated Geologic and Engineering Reservoir Characterization of the North Robertson (Clearfork) UniL SPE 29594
A Case Study, Part 1

m
Td?-
, Joooo
oo*. .g@
0000..:

-. .- . .-
u
~igure 32- Maiwisi =aianca k-m Type CIJIV9&Idysis - Mep wi%msfy tuH, MS I b.

Figure
35- Bubble Map d NRU Primary CurmJ181A.eOd Pmduolbn (1956-198T).

Ei%15Ynfl(
000..0
o* @o.
? .**.. ? m

F,gur& 33- MaIeriel Balsnce Oedine TYIM Curve hdyeis - Map of Flow Capacity (kb), d-low

figure 36- BubbleMap of NRU Prirnery Cumulative Water Production (195G1 9S7).

I
I
I 1
h\ \ \ .

b sacs ..-
.

. .
h ,.. \

000.0...
.0 ,000000 000

/
.wll//ll/ l-{l\ \ 1 II
@@**@@~
oo:6000@oo:o
a~~e~

MOmLm. z20um
gwe 34- Malarial Balarm O@chneTyps Curve AIIS@iS - Msp cd Eslbneted Primery Drdrse90 Area

F-37 - Bubble -p d NRU Sewn&y Cumddiw Oil Proshdon (19S7.1604).

488
 .
SPE 29594 L.E. Doublet, P.K. Pande, M.B. Clark, J.W. Nevans, and T.A. Blasingame 25

F o
0
0000


Q

Icf hplsd,

7 I
8WS (1987- 1994)

I
J
1
#
,*
. o ::00
00000 00

f 00000 00
o@. 0000

oQ &
t
000,0
.*Q.
co 000

~~ --lmm
~

CumWalerInj (Mbbls)

Figure 41. Waterflood Parlormansa Dlagnoetk PM. drw Jordan (f4ef. 22).

II NORTHROBERTSON
UNIT

Cumddk WolsrPrcdussd.
m (19s7-1994)

c.w&mdlww8csm
\

.-
/-

.
---
o- Peuwmr.a@w
/

.
- --
P0aun2k91a

.
----
I

----
I
a
:OC

J
/
00

.**e
f

. . * ..
.0 .0. 0 :

x
/
000 :
,0. / ----
B
-- I / ---
coo=
* /
9. 0.ooem ----
/ -z-
*****
/
*O-O
**.*
/
/
t .000 C

%: ****

b-=


*
r

o ~tlm~
.
u
Clmd8swwmmw

Figure 39- Bubble Map of NRU Cumrslaltie Waler Injaclion (1987.1994).

F~re 42- Example of Hal Pbl, flar Thakur (Rd. 24).

I 3 - b!dlwd W (Cdm4Y Od (tw)

_ Wd.hjtdm Rde (CdEI@ b) (*)

1 1 J I

10~1 . . . ..
o 4ccoo
Cum Wolcr Inj (Mbbk) Cum.bliie Woterhjesled (Mbbls)

FtguIo 40. NRU Daily Told ln@cd and Pmdusad Fluii Vdurnas (1 987-1 WW4). FWIC 43- Hal PW Fkld Exam#a.

489
 SPE 29594
26 An Integrated Geologic and Engineering ReservoirCharacterizationof the North Robertson (Clearfork)UniL
A Case Study, Part 1

w
I
I
v I

lml-
90

.:
.
I
. t ,8
. : ,i
.
I I
! I
v

08 mo 400 Xo
10
,., I
0 Im ?00 18

CiurIIJlolii Woler Injecled (Ubbls) 3k4-n Ti411w

F$gure44- Hall Plot Ios NRU Well 301 Figbw 47- NRU Wdl 301 Proscum Fdbll &sdysi8 Using D41iv6tiv4 usd Inlogrd Andy8is.

%iH

,o.r_l_
.
,0* ! 0
shuhrwtik
31n4+Tmaath

Figure 45- NRU Well 1000 Pressure Buldup Analysis Using Oarivdivc and Intagrd Analysis. Figure
46-NRU WCU3510 Pmswra Falloff Analysis Udng O.rWdivs nd InWgml AMt@c.

,TBU

I
l,,*-
.P~yTs

+00.10 +-
II ,.a a! n
.. . . .

I I 1 d
!0 L lIY 10 d 1(
Jo
-hr~ab SWM rm.u *
F-46 - NRu wd 3s10 I+OSSW aIJMUP hdpk using 04mmow and IIsbgrd hdph FiWM
40. Cu16eian Pbt 01 P- VWU8 Skl-kI Tsw fos NRU W63 301 F%8ssumFdlD41Toct
 27
SPE 29594 L.E. Doublet, P.K. Pande, M.B. Clark, J.W. Newms, and T.A. Blasingame

, 1 L
5000 ,IOlm !31Q M TOSM
NMI . F- P*

I I
4800tHrrrm

:!gure 50- Compari=X 01 Average Parting Pressures from Step-Rate Tesla (19SS-1 993).

.-. -. u
->=z. _
-----~ ---------~
- n

Figure 51- Map of Partial-Unit Model Areas to be Used for Reservoir Simulalior

491