CHARACTERIZATION OF A HEAVY OIL RESERVOIR USING

COMPUTER ASSISTED TOMOGRAPHY OF CORE MATERIAL

A. KANTZAS D.F. MARENTETTE B. ERNO S. HARDING

this article begins on the next page F

PETROLEUM SOCIETY OF CIM PAPER NO 16 SS89-16 THIS IS A PREPRINT - SUBJECT TO CORRECTION CHARACTERIZATION OF A HEAVY OIL RESERVOIR USING COMPUTER ASSISTED TOMOGRAPHY OF CORE MATERIAL Apostolos Kantzas Daniel f. Marentette Brian Erno Nova Husky Research Corporation Steve Harding Husky Oil Operations PUBLICATION RIGHTS RESERVED THIS PAPER IS TO BE PRESENTED AT THE THIRD TECHNICAL MEETING OF THE SOUTH SASKATCHEWAN SECTION, THE PETROLEUM SOCIETY OF CIM, HELD IN REGINA, SEPTEMBER 25-27, 1989. DISCUSSION OF THIS PAPER IS INVITED. SUCH DISCUSSION MAY BE PRESENTED AT THE TECHNICAL
MEETING AND WILL BE CONSIDERED FOR PUBLICATION IN CIM JOURNALS IF FILED IN @VRITING WITH THE TECHNICAL PROGRAM CHAIRMAN PRIOR TO THE CONCLUSION OF THE MEETING. ABSTRACT The characterization of reservoir rock, using Computer Assisted Tomography (CAT) with x-rays has been an objective of increasing popularity in the petroleum industry. In the present study, a medical x-ray CAT scanner was used for the study of core material from the "B" member of the Clearwater Formation of the Lower Cretaceous Mannville Group in the Cold Lake area. Two pieces of core, each
approximately 0.3 m long, were scanned at O.Olm intervals at a resolution of 0.00075m x 0.00075m x 0.005m. The raw CAT scan data were analysed using a post- processing computer package developed in our laboratories at NOVA Husky Research. This analysis gives a detailed picture of the various heterogeneities within each core section including cracks, shale barriers and clay clasts, and also allows determination of core properties such as bulk density and porosity. Bulk densities determined from the CAT scan data are in good
agreement with direct measurements obtained from several plugs taken from the core. The CAT scanning of the core successfully identified the locations for good plug selection. In addition,
various types of heterogeneities were mapped in three dimensions. Significant core damage was observed which is attributed to the coring and freezing processes. This paper presents the methodology of the scanning and the subsequent data analysis. INTRODUCTION Computer Assisted Tomography (CAT) x-ray scanners have become a useful tool for the petroleum scientist and engineer. A detailed literature
survey has recently been prepared by Kantzas (1989) while various applications are described in a variety of papers (e.g. Wellington and Vinegar (1987), Kantzas etal. (1989)). of particular relevance to this work are -various papers describing applications to heavy oil recovery (Fransham and Jelen (1986), Sedgwick and Miles-Dixon (1987),
Cuthiell and Sedgwick (1988)). The quantitative representation of core properties has been the Eocus of a lot of our work recently. The effort focused on consolidated media and particularly iii ttie Alberta dolomitic reservoirs. In this paper we address the potential of using the same technique for unconsolidated media. The CAT scanner can measure bulk density and effective atomic number directly by doing the so-called dual scanning i.e. one scan at an energy which is predominantly producing Compton scattering phenomena and one energy where
photoelectric absorption is dominant. Fluid saturations can I)e measured in a core if the core is scanned while saturated and while dry at different (,nerg, levels, and if the fluids are scanned separi5Ltely (Kantzas, 1988). This is very import:ant when flow processes are monitored (e.g. Fransham and Jelen (1986), Cuthiell and
 55 '19- /4
PETROLEUM SOCIETY OF CIM PAPER NO 16

THIS IS A PREPRINT - SUBJECT TO CORRECTION

CHARACTERIZATION OF A HEAVY OIL RESERVOIR

USING COMPUTER ASSISTED TOMOGRAPHY
OF CORE MATERIAL

Apos(olos KanlzBs
Daniel f. Marentette
Brian Erno
Nova Husky Research Corporation
Stave Harding
Husky 011 OperatIons

PUBLICATION RIGHTS RESERVED

THIS PAPER IS TO BE PRESENTED AT THE THIRD TECHNICAL MEETING OF THE SOUTH SASKATCHEWAN
SECTION, THE PETROLEUM SOCIETY OF CIM, HELD IN REGINA, SEPTEMBER 25·27, 1989. DISCUSSION OF
THIS PAPER IS INVITED. SUCH DISCUSSION MAY BE PRESENTED AT THE TECHNICAL MEETING AND WILL
BE CONSIDERED FOR PUBLICATION IN CIM JOURNALS IF FILED IN WRITING WITH THE TECHNICAL PROGRAM
CHAIRMAN PRIOR TO THE CONCLUSION OF THE MEETING.

ABSTRACT INTRODUCTION

The characterization of reservoir rock, using Computer Assisted Tomography (CAT) x-ray
Computer Assist.ed Tomography (CAT) with x-rays scanners have become a useful tool for ehe
has been an objective of increasing popularity petroleum scientist and engineer. A detailed
in the petroleum industry. In the present literature survey has recently been prepared by
study a medical x-ray CAT scanner was used for
I Kantzas (1989) while various applications are
the study of core material from the "BIT member described in a variety of papers (e.g.
of the Clearwater Formation of the Lower Wellington and Vinegar (1987), Kantzas et.aL
Cretaceous Mannville Group in the Cold Lake (1989»). of particular relevance to this work
area. Two pieces of core, each approximately are various papers describing appl~cations to
0.3 m long, were scanned at a.Dlm intervals at a heavy oil recovery (Fransham and Jelen (1986),
resolution of O.00075m x O.00075m x O.OOSm. The Sedgwick and Miles-Dixon (1987) J Cuthiell and
raw CAT scan data were analysed using a post- Sedgwick (1988). The quanticative
processing computer package developed in our representation of core properties has been the
laboratories at NOVA Husky Research. This focus of a lot of our work rece.ncly. The effort
analysis gives a detailed picture of the various focused on consolidated media and particularly
heterogeneities within each core section in the Alberta dolomitic reservoirs. In this
including cracks, sha.le barriers and clay paper we address the potential of using the same
clasts, and also allows determination of core technique for unconsolidaeed media. The CAT
properties such as bulk density and porosity. scanner can measure bulk density and effective
Bulk densities determined from the CAT scan data atomic number directly by doing the so-called
are in good agreement wich direct measure.ments dual scanning i. e. one scan at an energy which
obtained from several plugs taken from the core. is predominantly producing Compton scattering
The CAT scanning of the core successfully phenomena and one energy where photoelectric
identified the locations for good plug absorption is dominant. Fluid saturations can
selection. In addition, various types of be measured in a core if the core is scanned
heterogeneities were mapped in three dimensions. while saturated and While dry at different
Significant core damage was observed which is energy levels I and, if the fluids are scanned
atcributed to the coring and freezing processes. separately (Kantzas, 19B8). This is very
This paper presents the methodology of the important when flow processes are monitored
scanning and the subsequent data analysis. (e.g. Fransham and Jelen (1986), Cuthiell and
Sedgwick (1988), ~ellington and Vinegar (1987), 8. Core Analysis
Kanezas (198B))_
The core analysis from the cored intervals
In the. case of heavy oil unconsolida.ted sands, immediately next: to the pieces used for
the core is drilled and frozen immediately in an CATs canning gave the results listed in Table l.
effort to preserye it at conditions as close to The piece No. I in each COre is the piece
the reservoir as possible_ When ehe core directly above the scanned sample, while the
reaches the core lab, plugs are taken at piece No _ 2 is the piece direccly below the
different inter,Tals and core analysis is scanned sample. Piece No. 1 wichin COL·e A
commencing. consists of an unconsolidated fine to medium
grained sandseone, rarely interrupted by chin
The. recovery of unconsolidated sand during argillaceous bands; Glauconite was identified as
coring involves using a rw)ber or plastic sleeve an accessory _ Piece No. 2 within Gore A was a
core barrel. Rubber sleeve cores may be fine-grained sands cone Iolith nrgillaceous
stabilized before sampling by freezing the rock laminaeion; glauconite was ideneified as an
and/or by immobiliz.ing th(~ core by surrounding accessory once again. 80th pieces encompassing
it wi th waK, p las ter of I'aris, foam, or other the Core 8 sample were described simply as a
suitable materials. Test samples are often clean, fine-grained sandstone. The general
drilled from fro:;cen core using liquid nitrogen descripcion for the entire cored intervals may
as ehe bit lubr:'cant. In some studies these be generalized as follows: the reservoir is a
samples have been transfen"ed immediately Lnto a sandstone with shale partings. The ~andstone is
core holder and allowed co thaw in place. The medium brown in colour, rarely argillaceous,
samples are then. cleaned and tested without moderately well sorted, fine-grained, with
removal from the holder (Swanson and Thomas, framework constituents predominantly subangular
1980). The con~s must be ~;ubjected to simulated to sub rounded . Sands tend to be thinly bedded
downhole (net oVE'rhurden) stress to arrive at a with stratification varying from horizontal to
representative value of porosity and low angle. Shale beds are typically 1-3cm thick
permeability (Keelan, 1982). and are evenly laminated. Angular and rounded
shale and carbonaceous intraclasts are scattered
The normal procedure for selecting samples from throughout.
unconsolidaeed reservoir core involves first
cuttLng the core sections lengthwise, and then b, CAT Scanning
selecting locations for cutting plugs or other
samples by visu.3.l inspection of the exposed Both core pieces represent 0.30 m of reservoir
surface. Wieh typical core diameter of under depth and will be reported in two sections
O.lm, plugs longl~r than about O.05m have to be Part A being core 451.3-451.6 m and Part B being
chosen from full diameeer core. Since the core core 451.7-452_05 m. An EMI GT5005 full body x-
is not visible through the plastic coring pipe, ray scanner equipped with a movable core-
selection of plugs from full diameter core is mounting bed was used in this ,""ork. The
done with no visual guidance, This may make it customized bed allows precise positioning of an
difficult co obtain the type of sample desired. object in the scanner. The scans were taken at
The x-ray CAT scan techniques assessed in this l40kV-28rnA wieh a 0.005m pre-patienL collimator
work offer the p',ssibility of non-destructively on the x-ray tube and full D.D15m pose-paeient
previe~ing of full diameter cores, and thus make collimator on the detectors. Calibration
the selection 0:: core plugs ea!!:ier and more procedures and software development used are
reliable. based on work by Kantzas (1989) and Kantz as
et.s!. (1989). The cores were mounred inside a
The main advantage of CAT scanning is that we core holder assembly consisting of a hollow
can extract infarmation about the core non- 0.125m ID tube with a Q.250m 00 plexiglass
destructively and prior to any destructive annulus. This is referred co as the "donut",
sampling. This paper identifies some of the and is filled with water. This "donut" assists
benefits of the method and discusses associated in hardening of the x-ray beam and increases the
problems. quality of the image produced by reducing
EXPERIMENTAL areifacts in the perimeter of the sample. The
piece of core was scanned at a resolution of
The two pieces considered in this study were .00075m by Q.00075m by D.OQ5m beginning O.Dlm in
from two wells in the Clean.'ater B Formation of from one end. Additional scans were made every
ehe Lower Cretaceous Mannville Group in the Cold O.Olm along the length of ehe core, ellch SCBn
Lake area. A dec ailed geological description of giving data on ehe full cross section of ehe
[his area is given in Kendall (1977), where five core. Output from the scanner computer consists
se lec ted reservoir [ypes are discussed and of an array of 320 by 320 numbers. Each number
compared on the basis of rock properties, fluid is normalized linear attenuation coefficient for
saturation and reservoir geometry" The core was one volume element and is called a CT number.
taken in JanUal")' 1987 and pieces were cut and The raw CT data ....as trans ferred via magnetic
kept frozen for future analysis. Two of these cape to a SUN4 workstation and processed using
pieces were taken for x-ray scanning" The scans software developed at NHRC. The resules are
were initially l·aken to see whether we could presented as slice views and corresponding GT
pre-select areas for plugs without having to cut number frequencies, and as sagittal vie ....s, every
the core. lmm,~diacely after scanning the core ten pixels, and their corresponding frequencies.
was cut and plug5 were taken, thus not allowing We define as sagiteal the core views that resule
_ _ _ _ _ _ _ _---1.
any additional work. after stacking all the two dimensional slices ----'