ENHANCE OIL RECOVERY

ASSIGNMENT # 01

STEAM FLOODING
Duri Steam Flood (DSF) Project in Indonesia
The DSF project was the largest steam flood project in the world. The field was located in the
Riau Province, on the island of Sumatra, in Indonesia. Duri was the second largest field in the
country, producing about 200,000 BOPD by steam flood. The DSF project was planned to
develop over 15,000 ac of reservoir utilizing over 4000 producing wells.
The reservoir depth was 600 ft with a net pay of 109 ft. The porosity and permeability were
36% and 1550 mD, respectively. The reservoir temperature was 100_F. The oil viscosity at
the reservoir temperature was 157 cP.
The API gravity of oil was 23. The field began production in 1958. Cyclic steam stimulation
began in 1967. By 1977, 339 steam-stimulation jobs had been performed. A steam flood
pilot was initiated in 1975. With the success of the pilot, steam flood was expanded in Area
1. Area 1 consisted of 95 inverted seven-spot patterns, each 11.625 ac in size, which was
the main pattern used in the field. There were more than 420 producers.
The earlier design was a concentric injection string in which one string of tubing was inside
another. The problem was that heat transfer between the flow streams resulted in highquality steam entering one interval, while low quality stream or even hot water entering into
another. That caused one sand being preferentially heated over the other.
The solutions employed were:
(1) To utilize a downhole choke configuration to inject steam down a single string and use
the principle of critical flow to achieve a proper split between sands; and
(2) To use twin existing injectors for patterns in which the total rate per sand was greater
than that achievable with the choke design. It was discovered that hybrid development
scheme with a combination of 15.5-acre five- and nine-spot patterns maximized oil recovery
and improved economics. Additionally, by changing to nine-spot patterns in the thickest net
pay areas, the producer-to-injector ratio improved from 2.4 for seven-spot patterns to 3.5 for
nine-spot patterns.
The DSF project is unique in that it simultaneously involved the management of existing
steam flood areas, the development of new steam flood areas, and the design of future
areas to maximize both oil recovery and production efficiency.

Kern River in California

Kern River was a large heavy oil field 5 miles northeast of Bakersfield, CA. The original oil in
place (OOIP) was higher than 4 billion bbls. The reservoir was shallow and consisted of an
alternating sequence of unconsolidated sands with inter bedded silts and clays. The porosity
and permeability ranged 28-35% and 1-5 D (4 D on average), respectively. The oil viscosities
were on the order of 4000 mPa-s at the reservoir temperature of 70F. The reservoir pressure
was low (100 psig). At 250F, the viscosity was reduced to15 mPa-s. The Kern River
formation represented a continental-alluvial fan deposit derived largely from the westwardflowing Kern River. The heat had been used in Kern River since 1950s. First downhole heaters
were used, and then hot water was injected. In 1964, the hot-water project was converted to
a steam-drive project and expanded to 47 injectors. This project is known as the Kern
project. After 1964, four other projects were started: San Joaquin, Kern A, G & W A, and
Reed. Five-spot pattern with 21/2 spacing were used. Below Figure is the map of Kern River
field with steam flood pilots and 1970-71 expansions. More detailed summary report of the

steam injection projects in different parts of the Kern River field was provided by Five basic
types of completions were used in these steam projects. They were punched liners, slotted
liners, selectively perforated cemented casing, inner liner completions and gravel-packed
liners.
Since 1966, producing wells had been completed by cementing casing through the oil zone
and selectively jet-perforating 50-60 ft of interval near the bottom of the zone. Where sand
production became a problem, inner liners were run. Although this helped to limit the sand
production, in many cases it caused plugging.
Steam flood performance showed poor vertical sweep efficiency. Two methods were used to
improve steam profile: mechanical restriction of the production interval and adding foam
diverters in the injected system.
Kern River sands were separated by silt and clay inter beds. Individual sands were believed
to be isolated. Steam injection was started from the deepest sand and moved upward to the
next zone after the current zone was depleted. This was done by selectively perforating
injection wells in the desired zone. Production wells were generally completed with all zones
open. Because steam was injected for several years, some of the heat lost to the overburden
preheated the reservoirs above the injection layer.

Qi-40 Block in Laohe, China

An SF pilot was conducted in the Qi-40 block in the Huang Xi Ling field, Laohe Field, China.
The pilot was in the Lian (Hua) II layer. The average porosity was 25% and the permeability
was 1.49 D. The average oil column was 60.5 m and the net-to-gross ratio was 0.484. The
reservoir depth was 910-1045 m. The original reservoir pressure was 8-10 MPa, and the
original reservoir temperature was 36.8_C at 850 m. The dead oil viscosity at
50C was 2639 mPa-s (the in situ oil viscosity was 3750 mPa-s). The block had been under
steam soak since June 1987. The initial development plan was to use 200 m square patterns
in the areas of greater than 15 m reservoir thickness. Two development patterns were used
in the areas of greater than 20 m reservoir thickness. Before SF, oil saturation was 0.57, and
the recovery factor was 24%.A simulation study was conducted to compare continuous

steam soak, WF and SF in 1997. The results for the Lian II are shown in Below Table. The
results showed that if steam soak was continued, the recovery factor would be low. If
converted to hot WF, the recovery factor was not significantly increased. The recovery factor
was the highest if converted to SF. The conversion of steam soak to hot WF was tried
starting in April 1996. Three inverted five-spot patterns with well distance of 141 m were
used. After conversion, the water cut increased from 31% before the conversion to 85%
within 2 months. The oil rate decreased. The WF was stopped in May 1997. And it was
decided to convert steam soak to SF.
Four inverted nine-spot patterns of a well distance of 70 m were initially steam soaked in
January 1998, and converted to SF in October 1998. There were 4 injectors, 21 producers
and 2 observation wells. The performance may be divided into three phases Phase I, from
the start of conversion to March 1999, the group liquid rate increased from 154 to 330 t/d,
and water cut increased from 63.3% to 90%. The oil rate decreased from 56 to 30 t/d
following the trend from steam-soak period.
Phase II, March-July 1999, six low-liquid-rate wells were steam soaked. The group liquid rate
increased to 440 t/d, the water cut decreased slightly, and the oil rate increased to 70 t/d.
Phase III, from July 1999 on, the group liquid rate increased to 578 t/d by March 2000, but
decreased to 470 t/d during August-September 2000 owing to well sand production and low
pumping efficiency. The oil rate fluctuated within 35-125 t/d. Until the end of 2003, the
cumulative OSR was 0.21, the SF recovery factor was 40.3%, and the total recovery factor
including steam soak was 64.3%. The pilot was extended to include seven injectors in July
2003. The SF was continued until the end of 2004. By that time, 815 tons/d of steam was
injected, and the average production rate was 7.8 tons/d per well in the expanded pilot area.
The water cut was 83.3% and the OSR was 0.2. Compared with the steam-soak recovery
factor, 17.65% incremental oil recovery was obtained. It was predicted that the incremental
oil recovery factor would be 24.66%, the production time would be 6 years, the sweep
efficiency by steam was 45%, and the heat efficiency would be 38.8%
The measured steam profiles in injection wells showed 23.4-64.9%, indicating overall low
profiles. High-temperature profile modification was conducted. Before modification, steam
injection was stopped to convert to hot water injection for 10-15 days. Then hightemperature profile-modification agent was injected. After that steam injection was
resumed. Downhole gauges were installed below production pumps to monitor pressure and
temperature. The success rate was 100%. The longest monitoring time was 6 months. The
data from 5 wells showed 230C during the initial steam-soak period, 80-90C in the middle
of production period, and 60-70C in the late production period. The data showed a decline
trend in temperature. The flow pressure was 5-6 MPa in the initial production period, but 0.20.3 MPa at the lowest during the late production period. The monitoring data helped to
adjust well production parameters. For low-rate wells, more perforations were added or
some chemical stimulation measures were taken. For high-potential wells, larger pumps
were installed. For steam channeling wells, profile-modification measures were taken. The
measure steam qualities from downhole samples at 7 injection wells for 30 times were 6573% at 30 m depth, 56-66% in a middle depth, and 50-56% at a lower depth. The steam
quality from
the boiler was 75-76%.

CYCLIC STEAM STIMULATION

Cold Lake in Alberta, Canada
This was the largest CSS project in oil sands. Cold Lake was one of the four major Alberta oil
sands deposits. It contained an estimated 160 billion barrels of low gravity (10.2 API) and
highly viscous oil (100,000 mPa-s at 13C reservoir temperature). The reservoir depth was
from 300 to 600 m. Therefore, the oil was too deep to be produced by surface mining or too
viscous to be pumped at a reasonable rate at original conditions. The formation porosity was
37% and the permeability was 3000 mD. The reservoir thicknesses were on the order of 33
m. Esso Resources Canada began laboratory and engineering studies in the early 1960s,
progressing to small-scale field pilots in 1964. In order to provide a sound planning base for
future operations, an evaluation program of drilling and coring was started in 1973.
Because Cold Lake oil (bitumen) at reservoir conditions (450 psi and 13 C) was practically
immobile, it was necessary to stress the formation to the point of yielding for steam
injection. It was found that the Clearwater formation of main interest would yield to a
downhole pressure of 1300 psi. The initial breakdown pressure might be 30-50% higher. With
high injection pressure, vertical and horizontal fractures were generated to accommodate
large volumes of hot fluids.
The Ethel pilots were initiated in late 1964 and operated until 1970. The stimulation wells
were completed in the Clearwater bitumen zone and were stimulated through eight cycles.
The size of steam treatments ranged from 3000 to 5000 bbl. Gas was injected with steam in
seven of the cycles, and air and water with steam were injected in two cycles. These
additives were not convinced to be beneficial.
A soak period of about 5 days was unusually allowed for heat dissipation in the reservoir.
The well was then opened for production for a few weeks which might continue for 5-8
months, depending on the fluid temperature and observed decline in oil rate.
In October 1969, a bottom water 5-spot steam flood was initiated. The flood contained one
central producer, four steam injectors, and four confining producers, all of which were open
to the bottom water. The objective was to determine whether heating conformance in the oil
zone could be improved by injecting steam into the more mobile lower water zone. The rate
of vertical heating was found to be slow and the experiment was terminated in April 1970.
Steam generation and fluid handling facilities were upscaled to the commercial scale, based

on the pilot tests and engineering studies: 20% bitumen recovery, well production rate of 80
bbl/day over an average of 6 year life, and 0.4 OSR.

Gaosheng Field, China

This field in Liaohe, China, had gas cap. Although it also had a bottom water, there was a
barrier so that water coning was not observed. The gas-oil level was 1510 m, the oil-water
level was 1690 m, and the reservoir depth was
1500-1800 m. The developed area was 14.5 km2. In the horizontal direction, there were
seven blocks among which the blocks 3, 246, and 3618 were the mainly oil-bearing blocks.
In the vertical direction, there were eight layers. Among these layers, Layers L1_L4 were
gas-bearing layers, L5, L6, and L7 were the main oil layers (88% oil in place), and L8 was the
aquifer layer. The reservoir thickness was 67.7 m on average. The porosity was 22-26% and
the air permeability was 1000-2300 mD. The reservoir temperature at 1600 m was 60 C
and the initial reservoir pressure was 16.1 mPa. The oil viscosity in situ was 74-605 mPa-s.
The oil viscosity decreased to 6 mPa-s when the temperature was raised to 200-220C.
Initially the field was produced by mixing light oil and heating rod pumps. Starting in
September 1982, CSS was tested and found successful. In 1984, a development plan was
designed which included:
1. Five-spot patterns of 210 m later in filled to 150 m.
2. Separately developing L5, L6, and L7 because of existence of gas cap and bottom water.
3. Four phases: initial depletion by mixing light oil and heating rod pumps, CSS, steam
flooding, and cold water flooding.
4. Completion included gravel packing, wiring wrapped screen, and perforated pre stressed
casing.
5. Wells were drilled along the gas-oil ring in L5 to make use of gas cap energy and control
pressure.
Because the reservoir was deep, it was important to reduce heat loss through wellbores.
Measures to reduce heat loss included tubing insulation, high-temperature metal packer,
and filling nitrogen in the annulus.
The heat loss was controlled to be less than 12%. It was observed that the back-produced
water was only 7.8% of the injected. Such low flow back was caused by high content of clay
(7-10%), especially montmorillonite (90%). Clay swelling adsorbed a lot of water and
reduced permeability. The cumulated water slowed down the heat dissipation into the
reservoir during injection. To solve this problem, surfactants and chemicals to prevent clay
swelling were added in the steam. Adding nitrogen in the steam also helped water
production. Adding thin film spreading agents also helped. To stop gas cap breakthrough,
several wells were drilled to produce gas under a controlled mode. The pressure of gas cap
was controlled not lower than 8 mPa, and the pressure difference between gas cap and oil
layer was controlled.