Mapping Steam and Water Flow in

Petroleum Reservoirs
Michael Wilt* and Clifford Schenkel, Lawrence Livermore Natl. Laboratory; Tom Daley, John Peterson, and Ernest
Majer, Lawrence Berkeley Natl. Laboratory; A.S. Murer,** Mobil E&P U.S.; R.M. Johnston,** SPE, CalResources LLC;
and Louis Klonsky, Chevron U.S.A. Production Co.

Summary data on fundamental reservoir properties. Seismic measurements

During the past 5 years, we have applied high-resolution geophysical often reveal the reservoir structure and the location of fracture zones
methods [crosswell seismic and electromagnetics (EM) and passive and bedding planes. EM data define the resistivity of the rock, which
seismic] to map and characterize petroleum reservoirs in the San Joa- in turn is sensitive to the distribution of pore fluids. We apply these
quin Valley and to monitor changes during secondary-recovery opera- techniques to small- to moderate-scale field experiments so we can
tions. The two techniques provide complementary information. Seis- improve our understanding of fluid flow during recovery. Data are
mic data reveal the reservoir structure, whereas EM measurements are collected in observation and production wells at depths encompass-
more sensitive to the pore-fluid distribution. Seismic surveys at the ing the injection and production zones. We use core and log data to
South Belridge field were used to map fracture generation and monitor help interpret results. In addition to these active source techniques,
formation changes caused by the onset of steamflooding. Early results we also apply passive seismic monitoring to image fracture creation
show possible sensitivity to changes in gas saturation caused by the and growth during hydrofracturing and steam injection.
steamflooding. Crosswell EM surveys were applied at a shallow pilot Crosswell EM System. The LBNL/LLNL crosshole EM system
at Lost Hills for reservoir characterization and steamflood monitoring. consists of separate transmitter and receiver stations that deploy
Images made from baseline data clearly show the distribution of the tar- tools in wells separated by up to several hundreds of meters (Fig.
get oil sands; repeated surveys during the steamflood allowed us to 1).1 The transmitter uses a vertical-axis magnetic core wrapped with
identify the boundaries of the steam chest and to predict breakthrough 100 to 300 turns of wire and tuned to broadcast a low-frequency si-
accurately. Applications of the EM techniques in steel-cased wells are nusoidal signal at frequencies from 100 to 20 kHz.
at an early stage, but preliminary results at Lost Hills show sensitivity The transmitter signal induces electrical currents to flow in the
to formation resistivity in a waterflood pilot. formation between the wells. These currents, in turn, generate a sec-
ondary magnetic field in proportion to the electrical resistivity of the
Introduction rock where they flow. By measuring the fields from the receiver bore-
Although large quantities of petroleum are produced through water- hole, we can infer the resistivity distribution between wells. At the re-
and steamflooding, the process is typically poorly understood. This ceiver borehole, a custom-designed coil detects the vertical magnetic
leads to inefficient recovery and associated production problems, such field and sends it up the logging cable to a lock-in amplifier located
as premature water/steam breakthrough, fracturing of reservoir rock, at the surface. The lock-in amplifier operates like a radio by measur-
and well failures. In a new effort to understand these fluid-displacement ing only those signals that are coherent with the transmitted signal
processes and associated reservoir changes, we are applied crosswell while rejecting incoherent background noise. An in-field computer is
geophysical methods to monitor secondary-recovery processes. used to keep track of the transmitted signal, the detected magnetic
The goal of this project is to use high-resolution geophysical field, and the depths of both source and receiver coils.
methods jointly to map and characterize petroleum reservoirs dur- By positioning both the transmitter and receiver tools at various
ing secondary-recovery operations. We view the introduction of levels above, below, and within the zone of interest, we can create an
steam- and waterfloods in petroleum reservoirs as natural tracers to image of the resistivity distribution between the wells. The EM data
map fluid flow and to define the reservoir structure. Efficient use of are interpreted by computer modeling in which the rock between the
such tools can help determine flow mechanisms, map creation and wells is divided into two-dimensional square blocks 1 to 5 m on a side.
destruction of fracture porosity, and track injected flow through nat- Each block is assigned an electrical resistivity value estimated from
ural channels that connect (and isolate) petroleum deposits. It is an the borehole resistivity log (if available). The computer then modifies
ideal mechanism for detailed reservoir characterization; the reser- the resistivity of these blocks until the calculated and measured EM
voir is defined (and redefined) as it is produced. data agree to within the measurement error, usually 1 to 2%. This pro-
In 1994, a partnership was formed between Lawrence Berkeley cess usually requires 10 to 12 hours per data set on a 50-MHz work-
Natl. and Lawrence Livermore Natl. laboratories (LBNL/LLNL) station to produce a detailed image of the underground strata.
and a group of San Joaquin Valley oil producers, including majors
CalResources, Chevron, and Mobil and the independent Bakers- Crosswell Seismic System. The crosshole seismic system consists
field Energy. The partners agreed to apply geophysical techniques of a high-frequency piezoelectric source and borehole sensors. The
to characterize several pilot areas and to monitor existing and incipi- receivers may be deployed in combination with the source for active
experiments, such as crosshole tomography, or separately for pas-
ent steam- and waterflood projects within the valley. The multiyear
sive listening during hydrofractures or reservoir injection. The seis-
project consists of a series of geophysical surveys in several pilot
mic source is a high-voltage piezoelectric transducer, which can be
areas and joint efforts to interpret the data in terms of reservoir prop-
operated in a pulse or controlled-signal-sweep mode. It operates at
erties and reservoir changes during oil production. In this paper, we frequencies up to 15 000 Hz and is effective at well separations of
briefly describe the technologies applied and examine some early 200 m or more, depending on the rock type. The high-frequency
field results from several of the pilot field projects. characteristics of this source make it ideal for resolving small-scale
features, such as fractures and injection fronts.
Geophysical Methods We deploy both hydrophones and three-component clamped geo-
The two main geophysical tools used are crosswell seismics and phones for signal detection. The hydrophone string is often used
crosswell EM. These tools provide complementary high-resolution with the piezoelectric source in water-filled boreholes; this tool de-
ploys rapidly. The clamped geophones are typically used for passive
**Now with ElectroMagnetic Instruments Inc. monitoring where it is necessary to detect lower-level signals and
**Now with Aera Energy LLC
obtain multicomponent data.
Copyright 1995 Society of Petroleum Engineers

Original SPE manuscript received for review 15 January 1997. Revised manuscript received Field Experiments
14 August 1997. Paper peer approved 2 September 1997. Paper (SPE 37532) first presented
at the 1997 SPE International Thermal Operations and Heavy Oil Symposium held in Bakers-
Seismic and EM field projects were initiated as early as 1994 and
field, Californina, 10–12 February. continue in four separate field areas. Most projects are in early to in-

284 SPE Reservoir Engineering, November 1997

 Fig. 1—Schematic of crosshole EM system.

termediate stages of completion, but several have just begun sec- Three crosshole seismic surveys were completed: one immediately
ondary operations. The following sections provide selected high- after the onset of steam in June 1995, the second in August 1995, and
lights from some of these. the third in February 1996. The initial survey included passive seis-
mic monitoring during hydrofracturing. The results of this monitor-
Seismic Measurements at the CalResources Pilot, South Belridge ing showed no observable activity. The lack of microseismicity was
Field. CalResources currently is applying a high-energy steamdrive surprising in light of the high seismicity observed earlier in a different
for oil recovery in diatomite reservoirs at the South Belridge field. part of the field during steam injection. Seismic monitoring at two
This pilot is the third in a series to study the steam-flow mechanism subsequent hydrofractures with Chevron at Lost Hills (August 1996
and May 1997), however, did show seismicity associated with hydro-
in this very tight rock. Previous results were generally encouraging
fracturing in the Opal A section of the reservoir.
for enhanced oil production, but there was evidence of steam bypass
Initial crosshole measurements showed strong attenuation of the P-
and breakthrough into intervals outside the targeted injection zone. wave phase despite good energy in later arriving S waves. The weak
Well drilling for the third pilot began in April 1995. The new wells P-wave energy was quite surprising in view of strong P-wave energy
included two fiberglass and two steel-cased observation wells in addi- during crosshole surveys at Mobil leases within the same field. A fol-
tion to the injection and production wells in the 5/8-acre pilot. Because low-up survey 2 months later showed similar results [that is, weak P-
of unexpectedly encountering a hot zone within the overlying Tulare wave energy but strong S waves (Fig. 2a)]. The third survey, made
sands, both fiberglass wells were lost almost immediately after drill- 6 months after the steam injection had begun, had dramatically differ-
ing. We were therefore unable to collect EM data for this site. ent results. In this case, the P waves are noticeably stronger over

(a) (b)

Fig. 2—Crosshole seismic recordings after initiation of a steamflood at the South Belridge field in July 1995; common offset gather from
Wells L06 (source) and L07 (sensor). (a) Data from the August 1995 survey showing highly attenuated P-wave phase and strong S-wave
energy. (b) Data from February 1996 showing large increases in P-wave energy. Change is believed to be related to a decrease in gas
saturation caused by reservoir pressurization.

SPE Reservoir Engineering, November 1997 285

 At frequencies greater than 50 Hz, the effect of the formation causes
the total response curve to deviate from the curve with the casing alone.
The observed data seem to correspond much more closely to the model
with the formation present. This promising result, coupled with other
research, strongly suggests that the formation resistivity may be ob-
tained with crosshole EM even in steel-cased wells. In future surveys,
our goal is obtain a complete tomographic resistivity section through
one steel well casing. In addition, we expect that the ongoing EM moni-
toring will yield information within 1 or 2 years on the subsurface resis-
tivity changes caused by waterflooding.

Discussion and Conclusions

We have shown geophysical data that relate to reservoir changes
during steam and waterflooding and during hydrofracturing opera-
tions. Although the projects are clearly at an early stage, the results
indicate that the geophysical measurements may provide a diagnos-
tic tool for monitoring reservoir behavior during secondary and ter-
tiary operations. The next stage is to attempt to match these behav-
iors with known or suspected reservoir behavior during production.
Fig. 4—Crosshole EM sounding measurements through a steel-
cased borehole at the Lost Hills field. Plot indicates that col- Acknowledgments
lected data are consistent with a model that accounts for both This work was supported by Chevron U.S.A. and the U.S. Dept. of
the steel casing and formation resistivity. Energy, Assistant Secretary for Fossil Energy, Office of Oil and Gas
Technology through the Natural Gas and Oil Technology Partnership
steam. The image also indicates that almost no steam has gone into the Program at LLNL and LBNL under Contract DE-AC03-76SF00098.
upper oil sand. In addition, Fig. 3c indicates that the steam preferential-
ly flows to the west in the middle sand but to the east in the lower unit. References
The EM surveys showed the steamflood to be much less uniform 1. Wilt, M. et al.: “Crosshole Electromagnetic Tomography-A New Technol-
than the operator anticipated, providing valuable information on the ogy for Oil Field Characterization,” The Leading Edge of Exploration, Soc.
progress of the flood and the parts of the reservoir affected by the of Exploration Geophysicists, Tulsa, Oklahoma (1995) 173–77.
steaming. In fact, our predictions have been confirmed by recent 2. Wilt, M. et al.: “Crosshole EM in Steel-Cased Boreholes,” Expanded Ab-
temperature and induction resistivity logs in fiberglass-cased ob- stracts, 1996 Soc. of Exploration Geophysicists Annual Meeting, Den-
servation wells. We continue to monitor the progress of this steam- ver, Colorado.
flood and are concentrating efforts on the upper oil sand.
SI Metric Conversion Factors
Waterflood Monitoring With Crosshole EM in Steel-Cased acre 4.046 873 E*01 +ha
Wells. Crosswell EM surveys are difficult to do because most cycles/sec 1.0* E)00 +Hz
available boreholes in operating oil fields use steel casing, which ft 3.048* E*01 +m
causes a high attenuation and phase delay of the transmitted signal. *Conversion factor is exact. SPERE
Recent research suggests that, although difficult, making effective
crosswell measurements is possible, even in steel wells.2 Extend-
Michael Wilt is Director of Borehole Geophysics at ElectroMagnetic
ing the crosshole EM technique to steel-cased wells is an important Instruments Inc. in Richmond, California. Previously, he was with LawĆ
objective of our research because it dramatically broadens the use rence Livermore Natl. Laboratory in Berkeley, California. He has more
of this technique. than 15 years' experience in the use of electrical and electromagnetic
In 1994, we began a collaborative project with Mobil to test the methods in applied geophysics and application of these methods in
petroleum engineering, contaminant waste migration, and enhanced
crosshole EM technique for mapping the resistivity structure be- oil recover, with his ongoing research focused primarily on those apĆ
tween wells and to track resistivity changes caused by water injec- plications for singleĆborehole and crosshole environments. Wilt holds a
tion. Fiberglass Monitoring Well 003 is 10 m from a water injector PhD degree in engineering (geophysics) from the U. of California, BerkeĆ
and 90 m from a steel-cased production well. Induction resistivity ley. Ernest Majer heads the Subsurface Geosciences Dept. at LawĆ
rence Berkeley Natl. Laboratory (LBNL) in Berkeley, California, where he
logs in Well 003 show that the average resistivity of the diatomite is involved in analysis of seismic waves generated by an impulsive
is 2.5 to 3 W@m, varying between 1.5 and 4 W@m in the production source embedded in a medium, with the goal of specifying the dynamĆ
interval from 550 to 770 m. In addition, repeat induction logs show ic source properties and parameters of the material. Majer holds a PhD
that the water injection has reduced the resistivity from 20 to 40%. degree in geophysics from the U. of California, Berkeley. Tom Daley is a
staff research associate in the Earth Sciences Div. of LBNL. His research
For the initial test, we positioned both source and receiver tools interests are borehole seismology, seismic anisotropy, and seismic
at the same vertical level and adjusted the frequency of the transmit- detection of fractures. Before joining LBNL in 1988, he worked for 5 years
ter beginning at the lowest frequency detectable with our receiver with Seismograph Service Corp. performing borehole seismic surveys.
(20 Hz). Then, we increased the frequency until the signal was atte- Daley holds a BS degree in geophysics and an MS degree in engineerĆ
ing geosciences, both from the U. of California, Berkeley. Anthony MurĆ
nuated by the casing to below the section threshold (approximately er is a geologist with Aera Energy LLC's Diatomite New Process Group
500 Hz). These frequency soundings were made at 10 different lev- in Bakersfield, California, currently working on designing and testing of
els, corresponding to different casing segments and small differ- new recovery processes in diatomite. Previously, he was involved in a
ences in formation resistivity. The sounding data were found to re- variety of E&P assignments in both unconventional and conventional
reservoirs for Aera and Mobil Oil. Murer holds an MS degree in geology
peat over time to approximately 3%. from the U. of Arizona. Mike Johnston works for Aera Energy LLC in BaĆ
Fig. 4 shows a crosshole frequency sounding from a depth of 550 kersfield. His current work focuses on new recovery processes for diatoĆ
m. The data are plotted with numerical calculations for steel-casing mite. Since 1984, he worked in various engineering capacities on
models with and without a 3-W@m formation. The casing parame- California oil fields. Johnston holds BS and MS degrees in geological enĆ
gineering from the U. of Missouri, Rolla. Louis Klonsky is Geophysical
ters were obtained with the Schlumberger METT casing evaluation Coordinator of Chevron U.S.A. Production Co.'s Western Business Unit in
log for the thickness and by trial-and-error fitting of the lower-fre- Bakersfield, California. Since joining Chevron in 1980, he has held posiĆ
quency section (t50 Hz) of the crosshole soundings to obtain the tions in exploration, development geology, and development geoĆ
conductivity and magnetic permeability of the casing. At these low- physics. Previously, he worked for Amoco and Unocal. Klonsky holds a
BS degree in physics from Columbia U. and an MS degree in geology
er frequencies, the calculated models with and without conducting from Rutgers U. Biographies of Clifford Schenkel and John Peterson are
formation effects agree to within a few percent. unavailable. Photographs of all authors are unavailable.

SPE Reservoir Engineering, November 1997 287