INTERNATIONAL SYMPOSIUM
ON SUBSURFACE INJECTION
OF OILFIELD BRINES
Proceedings
Sponsored By
UNDERGROUND INJECTION
PRACTICES COUNCIL, INC.
LU
o
RESEARCH FOUNDATION
Royal Sonesta Hotel
New Orleans, Louisiana
May 4 through 6t 1987
-------
-------
Proceedings o£ a International symposium
on
SUBSURFACE INJECTION OF OILFIELD BRINES
May 4 through 6, 1987
Sponsored by the
U.S. Environmental Protection Agency
and the
Underground Injection Practices Council
Research Foundation
President UIPCRF - Paul Roberts, Director, Nebraska Oil &
Gas Conservation Commission
Executive Secretary UIPCRF - Michel J. Pague, Director,
UIPC
Chairman UIPCRF - Science Advisory Committee, Dr. Wayne
Pettyjohn, Sun Professsor, Oklahoma State Univ.
Symposium Coordinator - Rosemary A. Marmen
Symposium Registrar - Betty J. Robins i
Published by the
Underground Injection Practices Council
525 Central Park Drive, Suite 304
Oklahoma City, OK 73105
(405) 525-6146
Additional copies available at $75
-------
UNDERGROUND INJECTION PRACTICES COUNCIL
RESEARCH FOUNDATION
Approximately two years ago, five state members of the UIPC
took an action authorizing the formation of the Underground
Injection Practices Council Research Foundation. The UIPC
Research Foundation exists with an independent Board of Directors
and functions as a separate entity. The purpose of the Research
Foundation is strictly to promote research in underground
injection which it feels are necessary and to provide a means for
the funding of those projects. It takes recommendations for its
research program from the UIPC Board of Directors and its own
members as well. To date, the Research Foundation has conducted
the following projects:
1) Hydrogeological and Hydrochernical Assessment of the Basal
Sandstone and Overlying Paleozoic Age Units for Wastewater
Injection and Confinement in the North Central Region.
£) A Pilot Survey of State Mechanical Integrity Testing
(MIT) - New Mexico.
3) Conducted a major national symposium on Subsurface
Injection of Oilfield Brines.
4) Conducted a Well Construction Seminar in Washington, DC.
5) Conducted two Mechanical Integrity Seminars ana will
conduct a third in Long Beach, California, July 14-16th.
6) Will conduct an International Symposium Class V injection
Well Technology on September 2£-£4 in Washington, DC.
The Foundation has also established the UIPC Library, funded
the first UIC Bibliography, and has as one of its ongoing
committments the further development of what will hopefully be
the largest collection of UIC texts and articles in the country.
The officers of the UIPC Research Foundation are:
1) Paul Roberts - President
£) Jim Watkins - Treasurer
3) Jarnes Welsh
4) Al Rarick
5) Jim Collins
6) Manual Sirgo
7) Michel Paque - Secretary
-------
TABLB OF CONTENTS
PAGE I
OPERATIONAL EXPERIENCE
1. Brine Disposal at Sour Lake Field, Texas: The
Interplay of Area of Review, Mechnical Integrity
and Geology in Evaluating Returns to the Surface 1
2. Application of the Temperature Survey in
Demonstrating the Mechanical Integrity of Injection
Wells 22
3. Injection Monitoring and Control; Dollarhide
Clearkfork "AB" Unit 63
4. Subsurface Injection of Fluids for the Recovery
of Petroleum 79
5. Oilfield Brine Disposal into the Wilcox Aquifers
in S.E. Mississippi - A Case History 132
6. Mechanical Considerations of the Disposal of
Fluids into Poorly Consolidated Sandstone Reservoirs 134
7. Leaking Abandoned Wells Caused by Class II
Injection Operations - Case Histories from the
Texas Railroad Commission files 166
8. Sources of Ground-Water Salinization in Parts of
West Texas, USA 224
9. Field Results of Tracer Tests Conducted in Oil
Field Steam and Non-condensible Gas Injection
Projects 254
10. Environmental Protection Agency's Pennsylvania
Compliance Initiative for Blow Box Operations 256
11. Mathematical Evaluation of Operating Parameters
Identified in a Class II Brine Disposal Well Permit
Application 262
12. The Use of Controlled Source Audio Magnetotellurics
(CSAMT) to Delineate Zones of Ground Water
Contamination - A Case History 286
13. Convective Circulation During Subsurface Injection
of Liquid Waste 318
-------
PAQI ft
14. Monitoring, Troubleshooting and Repairing Wellbore
Communication o£ Waterflood Injection Wells in
the Ville Platte Field - A Case History 342
15. Some Aspects of Monitoring a Waterflood, Ventura
Avenue Field Waterfloods 368
16. Status of Mechanical Integrity Testing in
Mississippi 423
WELL TECHNOLOGY
17. Well Integrity Maintenance Using Pumpable Sealants 438
18. Measuring Behind Casing Water Flow 468
19. A Pilot Survey of State Mechanical Integrity
Testing Programs in New Mexico 485
20. Planning Successful Temperature Surveys 512
21. Mobil's Attempt to Obtain a Waiver from the Surface
Cementing Requirements for Rule Authorized Class II
Enhanced Recovery Wells in the Springfield North
Unit 535
22. A Method to Convert Multiple-Shop Section Openhole
Completions into Cased-Hole Completions with Zonal
Isolation 556
23. How to Locate Abandoned Wells 578
24. "Ada" Pressure Test 598
-------
BRINE DISPOSAL AT SOUR LAKE FIELD, TEXAS:
THE INTERPLAY OF AREA OF REVIEW, MECHANICAL INTEGRITY AND GEOLOGY IN
EVALUATING RETURNS TO THE SURFACE
BY
T. LAWRENCE HINELINE
KEN E. DAVIS ASSOCIATES
3121 SAN JACINTO, SUITE 102
HOUSTON, TEXAS 77004
ABSTRACT
Sour Lake Field in Hardin County Texas, approximately 20 miles west of
Beaumont, Texas is one of the oldest producing fields in the country, having been
discovered a commercially productive field in 1903. The field continues to produce
about one million barrels of oil a year with approximately seven million gallons of
water being produced which must be disposed of. Throughout the recent operating
history of the field, the produced brines have been returned to the subsurface
through the use of injection wells. The injection of produced brine was either
purely disposal or in some cases, into producing zones, for the purpose of
secondary recovery.
A unique feature at Sour Lake is a twelve acre lake, commonly referred to in
the area as the "sink hole". This lake formed in the late 1920s as a result of
subsidence due to the oil and water withdrawal.
In 1980, the Texas Railroad Commission, which has regulatory authority over
oil and gas operations in the state, held a hearing to review all of the existing
disposal permits for possible cancellation which would have the ultimate effect of
virtually shutting in the field. Investigations and incidents relevant to the
hearing included a wellbore flowing saltwater to the surface, injection wells
without tubing or with mechanical integrity infractions, a reported two foot
increase in water level in the sink hole coincident with the injection of 288,000
-1-
-------
barrels of brine into a disposal well and an increase in chlorides concentrations
in the sink hole from 2000 ppm to over 25,000 ppm. These events and their
subsequent resolutions are discussed.
INTRODUCTION
Sour Lake Field in Hardin County, Texas, approximately 20 miles west of
Beaumont, Texas is one of the oldest producing oil fields in the country. The
field has produced over 90 million barrels of oil and continues to produce in the
vicinity of a million barrels of oil per year. In the early days of production
produced brines were discharged to the surface and were carried off in drainage
ditches. Later in the history of the field, produced brines were disposed of down
wells. For the most part, brine disposal in an old salt dome field is fairly
routine procedure. In the early 1980s however, there was a series of events that
could have virtually shut down the field.
SITE GEOHYDROLOGY
As mentioned above, Sour Lake Oil Field is located on the north side of the
town of Sour Lake about 20 miles west of Beaumont. This places the field in the
Gulf Coast Salt Dome Basin Province. Other than around salt domes, oil and gas
production is from Frio and Yegua sands along this trend. The Sour Lake Salt Dome
is a piercement feature, cutting through the Yegua, Jackson and Frio sections.
Miocene sands thin considerably over the dome, having a thickness of over 4500
feet less than two miles off of the flanks of the dome and less than 1000 feet at
the crest. A schematic cross section of the west flank of the feature is
illustrated in Figure 1. Oil and gas production in the immediate vicinity of the
dome is from a series of Miocene sands. On the flanks of the dome, the deeper
-2-
-------
Yegua sands are productive. Oddly, despite ideal structural and stratigraphic
trapping, there is virtually no Frio production from the flanks of the dome.
Salt dome areas are prolific oil and gas producers because of the
stratigraphic and structural traps formed in the surrounding lithology. The
upwarping of sediments and extremely complex faulting in the sand-shale sequences
results in a multitude of individual reservoirs. Unfortunately this complexity
also makes precise geohydrological analysis nearly impossible.
Aquifers in the Sour Lake Area include the Miocene Oakville Sand, the Pliocene
Willis Sand and Goliad Sand, the Pleistocene Lissie Sand and recent alluvium.
Because of the similar character of all of these formations, because it is
difficult to distinguish them in the subsurface with drillers logs or electric logs
and because it is assumed they all are hydrologically connected, these formations
are generally referred to collectively as the Gulf Coast Aquifer.
On the flanks of the dome, the depth to the base of fresh water reaches 2000
feet. At the crest of the dome fresh water is found to an approximate depth of 100
feet. This configuration is illustrated in Figure 2. The transition from fresh to
salt water is easily detected by resistivity logs as shown in that same figure.
There are numerous water wells in the Sour Lake area (Figure 3) from which
water quality information can be obtained. The nearest shallow well to the crest
of the dome that there is water quality information for is about 7000 feet to the
southeast. It is drilled to a depth of 60 feet and in 1962 had a total dissolved
solids (TDS) concentration of 1,025 ppm. The City of Sour Lake operated two
municipal supply wells about one mile south of the crest of the dome which drew
water from 177 feet. In the years from 1941 to 1949 the TDS concentration in these
wells rose from 520 ppm to over 1500 ppm. These wells were replaced by two new
wells another two miles to the south. These wells were drilled to a depth of 812
feet and 224 feet. The deeper well initially had a TDS concentration of 548 ppm
-3-
-------
and a chlorides concentration of 188 ppm. After ten years of operation the IDS
concentration rose to 1460 ppm and chlorides to 645 ppm. The quality of water in
the shallower well remained fairly constant over the same period of time with 500
to 600 ppm IDS and approximately 200 ppm chlorides. This change in water quality
is typical around a salt dome, especially in wells on the down gradient side of the
dome. As the wells are pumped over time, salt water encroachment is to be
expected.
HISTORICAL REVIEW
Production History
The poor quality of water in the Sour Lake area can not be attributed to oil
and gas operations. This can be assumed from the name given to the town of Sour
Lake which was founded in 1835. In fact, there is a legend that the original Sour
Lake, now gone, caught fire, inciting the rather superstitious Indians in the area
never to return. Seeps of oil and sulfur to the surface first brought those
seeking medicinal treatment to the area and as early as 1893 brought oil
prospectors into the area. In the later years of the 1890s there was minimal oil
activity. On January 6, 1903 the first significant well was drilled by the Texas
Co., coming in as a 15,000 barrel a day gusher. Well over a thousand wells have
since been drilled at Sour Lake. Texaco, alone has drilled in excess of 800 wells
on the major 815 acre lease of the field as well as on some smaller surrounding
leases, and to date over ninety million barrels of oil have been produced.
Sink Hole Development
A relatively unique feature at Sour Lake Field is a depression that developed
in the ground in 1929. The twelve acre by 40 to 50 foot deep feature did not
evolve over time, but in two brief incidents on October 9 and 12 of that year.
-4-
-------
Several oil wells up to 2000 and 3000 feet from the sink went entirely to water.
All of the affected wells were apparently producing from the caprock of the dome
rather than the overlying sands.
The formation of the sink is attributed to the dissolution of the salt and cap
rock, the production of over 73 million barrels of oil and the production of an
undetermined quantity of saltwater, sand and dissolved solids. The volume of
displaced earth at the surface was estimated to be 98,000 cubic yards.
In the years from the early 1930s until the late seventies, drilling,
production, brine disposal and the sink hole apparently coexisted with little
adverse consequence.
EVENTS LEADING TO THE SOUR LAKE ORDER
The interplay of area of review, mechanical integrity and geology became
apparent at Sour Lake when the Texas Railroad Commission took action to investigate
brine disposal at that field. The investigation was triggered by one incident, and
in the several months that followed new circumstances either developed or were
uncovered.
In January 1980, Texaco reported to the Texas Railroad Commission that an oily
accumulation was collecting on the surface of the sink hole. The material which
was collecting at an estimated rate of ten to fifteen gallons per day was described
as a "fibrous, oily, muddy looking material." At the time of the report, the
material covered about two acres of the twelve acre lake. Texaco offered to
make every effort to contain the material but felt that they were not responsible
and would seek assistance from other operators in the field. The Railroad
Commission made periodic inspections of the sink hole and surrounding area in the
following months. One such inspection followed a report of contaminated water in
Clemmons Gully into which the sink occasionally drains. The inspection revealed no
-5-
-------
problems, although later testimony alleged that some cattle died as a result of
drinking this water. In May, Texaco received permission to skim 3000 barrels of
fluid off of the sinkhole for transport through a pipeline.
In May and June a new commercial brine disposal well was permitted, drilled,
and completed 1,500 feet north of the sink hole. The well was drilled to 1912 feet
and reached total depth in the caprock. The well was completed with 10 3/4-inch
surface pipe to 113 feet, 7-inch casing to 1740 feet, a 4 1/2-inch liner from 1700
feet to 1904 feet and 3 1/2-inch tubing set on a packer at 1638 feet.
Texaco maintains storage caverns that were dissolved in the salt dome for the
storage of hydrocarbon products. To control the movement of product in or out of
the caverns, Texaco had two lined pits at the surface to hold brine which was
pumped into or out of the caverns. Around the time that Luther Hendon completed
his disposal well, one of the Texaco pits developed a leak and needed to be drained
of the several hundred thousand barrels of brine it contained so that repairs could
be made. Luther Hendon was given permission to dispose of this brine by the
Railroad Commission and had disposed of approximately 250,000 barrels of the brine
before he suspended injection in late June.
In the middle of June, an operator in the field reported a rise in the fluid
level in the sink hole and at the end of June a program was begun to monitor the
water level in sink hole. Precipitation and evaporation were taken into account.
The area of the water surface was surveyed to be 12.24 acres and the calculation
made that a one inch rise represented 7,854 barrels.
In consideration of these occurrences, the Railroad Commission District 3
office requested that a hearing be held in Austin to show cause why
1. All disposal wells into charged zones should not have permits cancelled
and be plugged in accordance with Railroad Commission Regulations.
-6-
-------
2. All disposal wells should not require tubing to be set on a packer with
annual pressure tests.
3. Any present or future disposal wells should not have permits cancelled if
testing reveals that the injection zone is charged or under pressure.
In July there was an official call for this hearing which was to be held on
September 25, 1980. In the mean time, the Railroad Commission requested that Dome
Holding Company and Luther Hendon shut in their disposal wells. These two
operators were singled out because they were injecting water not produced at Sour
Lake.
At the same time all of this was transpiring, the Luther Hendon application
was pending, despite there having been emergency authority to dispose of the Texaco
brine. Apparently other small operators in the field felt that the occurrences at
the sink hole which jeopardized their operations could be attributed to the Hendon
operation. They therefore joined together and called for a hearing to protest the
Hendon permit. This hearing was called by the Railroad Commission and held on
August 21, 1980.
In the course of the two hearings a great deal of information as well as some
speculation was brought forth on the events surrounding brine disposal in the
field.
The opposition to the Hendon application provided testimony which they believe
connected the rise in the sink hole water level to the Hendon operation. Although
at the time records had not been kept, photographs indicated a two foot rise in the
water level between June 18 and June 28 during which time Hendon had injected
approximately 250,000 barrels of brine. The two foot increase represents about
190,000 barrels of additional water in the sink hole. The opposition alleged that
injection of the heavy brine from the Texaco pits (10.5 pound per gallon) at 825
-7-
-------
psi would fracture the receiving formations. Records of level began being kept
after this incident but also after the time Hendon shut in his well and were kept
from July 10 to August 18. In this period, minus the effects of rain, the level
increased approximately three inches. During August, excluding effects of rain,
there was a net loss in the level. A note was made that during that period, Texaco
disposal wells were shut in for testing or repairs. A great deal of speculation
arose as to how injected brine could end up it the sink hole. The complexity of
the geology on the crest of the dome makes any specific analysis virtually
impossible despite the dense well control. Whether or not the sands at
approximately 1700 feet at the Hendon well actually contact the sink hole or
whether or not fractures, faults or abandoned wellbores may have allowed
communication was only theorized.
The point was also made that any effects on the sink hole were most likely the
cumulative effects of the 20 or so disposal wells operating in the vicinity.
In consideration of the facts presented at the Hendon hearing which included
that the well was properly completed and that the operation was given approval by
the Texas Water Commission and based apparently on the fact that the opponents had
not proven connection between that operation and detriment to the field, the
hearing examiner made his recommendation. This included injection be allowed at
1675 feet to 1730 feet, injection pressure be limited to 400 psi, injection be
through tubing and packer, only brines produced in Sour Lake Field be injected and
that annual mechanical integrity tests be performed.
In view of the fact that the September 25, 1980 hearing required all operators
of disposal wells in the field to defend their operations, a great deal of
information was produced. Photographs were presented which pictured flow to the
surface of the sink as indicated by an area of disturbed water and some bubbling.
In April and into May, operators made an effort to stop the flow into the sink
-8-
-------
hole. Reportedly, divers were able to locate a submerged wellhead. A pipe was run
into this wellbore to a depth of 310 feet and 1,377 sacks of cement were pumped
into the well. The bubbling to the surface reportedly stopped.
The most significant rise in water level however, was reported to have
occurred in June. This would indicate that if brine disposal was responsible for
the level rise, there were other avenues than the abandoned wellbore. Dome Holding
and the Hendon well were shut in upon Railroad Commission request yet the water
level continued to rise. Between the time of the call for the hearing and the
hearing, ten of the twenty or so disposal wells at the field were tested and found
to be injecting into overpressured zones. Additional testing also indicated that
some of the active disposal wells failed mechanical integrity tests.
Texaco, the major operator in the field, operated two disposal wells into the
caprock near the crest of the dome. Texaco produced approximately 61 percent of
the field's million barrel a year production and disposed of approximately 11,000
barrels a day of salt water of the field's 21,000 barrels disposed of daily.
Approximately seven million barrels of water are injected annually. Neither of the
Texaco wells met the standards set forth for the hearing, so prior to the hearing
Texaco repaired both of the wells. The number one well was fitted with tubing and
packer in servicing that also found a leak in the casing. Because of a restriction
in the casing of the number two well, a packer could not be set, so tubing was
cemented into the entire length of the casing. Both wells passed subsequent
radioactive tracer tests. Sun Oil Company is another major company that operates
in the field, though of considerably less consequence than Texaco. Sun only
injected about 200 barrels per day into two wells, both of which were about two
miles from the sink hole. Also, both Sun wells met the mechanical standards
required.
-9-
-------
Considerable testimony was provided by a group of small operators called the
Sour Lake Operators Group. The group's testimony had two fundamental themes. The
first was that any difficulties with the sink hole could be attributed to one or
two disposal operations that were injecting brine not associated with oil
production in Sour Lake Field. The second was a verification that all of the wells
utilized by the group were completed with tubing and packer and therefore met the
standards set forth by the Railroad Commission or else they were shut-in.
There were other conditions or incidents that led to the conclusion that there
was brine migration at the field. There was a report, although not documented,
that there had been a drilling rig active in the vicinity of the sink hole run by
an unknown operator. Following this operation, an abandoned pipe was found to be
flowing salt water to the surface at that location. A second similar incident
which is documented, occurred after the 1980 hearings. This incident involved a
well drilled to 902 feet that reached total depth in the caprock. When an attempt
was made to log the well, it began flowing salt water in an eight inch stream that
rose four feet into the air and continued to do so for 24 hours.
Another factor indicating flow into the sink hole was the quality of the
water. Testimony was given that the sink hole water had always been relatively
fresh, derived primarily from runoff. Reportedly, as late as January 1980, the
chlorides content of the water was around 2000 ppm. The water from the sink or
from Clemmons Gully into which it drains had been reportedly used for irrigation,
cattle watering and mixing drilling mud. In September of 1980, water quality was
analyzed at different depths. The chlorides content was reported as follows:
Depth (ft.) Chlorides (ppm)
7 28,700
10 28,910
19 28,595
-10-
-------
Depth (ft.) Chlorides (ppm)
20 28,830
30 28,520
40 25,995
THE SOUR LAKE ORDER
Following the consideration of all data and testimony, on October 5, 1980 the
Railroad Commission issued the following order:
All existing disposal or injection permits currently in effect in the
Sour Lake Field will terminate 90 days after the effective date of this order.
However, an existing permit may be renewed by the refiling of Railroad
Commission Forms W-14 or H-l and other Commission required supporting data. A
renewed permit or any future disposal or injection permit for the Sour Lake
Field will be subject to the following conditions as well as any other
limitation that may be required by the Commission.
1. Injection must be through tubing set on a packer located immediately
above the disposal zone.
2. The injection fluid will be limited to saltwater produced in
association with oil and gas production in the Sour Lake Fields.
3. Prior to injection, and annually thereafter, a surface monitored
downhole survey must be conducted under the supervision of the
District Director to insure that the injected material can enter no
other strata than that approved in the permit; and provided further
that should it be determined by the Commission that such injected
material is not confined to the approved strata, the authorization
given hereby shall be suspended and the injection stopped until all
migration from such strata is eliminated.
-11-
-------
In April 1981, the Railroad Commission investigated operators' compliance with
the order. There were 38 injection or disposal wells in existence at the time of
the order. Nineteen disposal wells and two injection wells were found to be in
compliance and were reissued permits. Thirteen disposal or injection wells were
not in compliance and issued letters cancelling permits. Three of those wells were
rejected because they had been recompleted into zones too shallow by Water
Commission standards. Other reasons for rejection included no tubing, holes in
tubing, a wellhead leak or other incidents of mechanical integrity test failure.
Other rejections were due to the fact that operators failed to perform tests or
submit the results of the tests.
SUMMARY AND CONCLUSIONS
In 1980 and 1981 the water filled sinkhole at Sour Lake experienced increases
in levels that were not attributed to rainfall or run off. The fact that
overpressurization of saltwater disposal zones was leading to flow into the
sinkhole was substantiated by disposal wells with shut-in pressure, observed
indication of submerged flow into the sink hole, wellbores flowing brine and an
increased chlorides concentration in the sink hole. Although there were
approximately 20 disposal wells in the field, Texaco and Luther Hendon were the
most significant operators. Luther Hendon had been injecting 15,000 barrels of
brine a day when the most significant level increase occurred. Texaco had been
injecting 11,000 to 14,000 barrels a day into wells with no tubing. Putting tubing
in these wells restricted their volume to the point of having to shut in producing
wells.
Although it was implied that flow to the surface was the result of
overpressured disposal zones, losses of mechanical integrity, complex geology and
-12-
-------
old abandoned wellbores, none of this could really be verified due to the
complexity of the area.
Luther Hendon never operated his well again despite being issued a permit to
do so. Texaco converted three wells on the west flank of the dome to disposal
wells. Disposal would be into non-productive Frio sands which pinch out at a safe
distance from the sink hole and caprock so that there should never be any
complications. All other smaller operators in the field either shut in their wells
or verified that they met the standards of the order.
-13-
-------
REFERENCES
Baker, E. T., (1964), Geology and Groundwater Resources of Hardin County, Texas,
Texas Water Commission Bulletin 6406.
Sellards, E. H., (1930), Subsidence in Gulf Coastal Plains Salt Domes, University
of Texas Bulletin, 3001, pp. 9-36.
-14-
-------
FIGURES
-15-
-------
FIGURE 1
SCHEMATIC CROSS SECTION OF WEST
FLANK OF SOUR LAKE DOME
-16-
-------
SINK HOLE
1000
2000
4000
5000
UNDIFFERENTIATED MIOCENE
AND YOUNGER SANDS
1000'
FIGURE 1 SCHEMATIC CROSS-SECTION OF WEST FLANK OF SQUR LAKE DOME
-------
FIGURE 2
CROSS SECTION SHOWING DEPTH OF
FRESH WATER OVER SOUR LAKE DOME
-18-
-------
DV
Approximate
surface
FIGURE 2 CROSS-SECTION SHOWING DEPTH OF FRESH WATER OVER SOUR LAKE DOME
(FROM BAKER, 1964)
-------
FIGURE 3
MAP OF SOUR LAKE AREA
-20-
-------
TEXACO FLANK WELLS
TEXACO CAPROCK WELLS
CREST OF DOM
FIGURE 3 MAP OF SOUR LAKE AREA
-21-
-------
APPLICATION OF THE TEMPERATURE SURVEY IN DEMONSTRATING THE MECHANICAL
INTEGRITY OF INJECTION WELLS
MALCOLM D. JARRELL AND RICHARD LYLE
KEN E. DAVIS ASSOCIATES, 300 N. MICHIGAN
SUITE 409, SOUTH BEND, INDIANA 46601
ABSTRACT
The temperature log has an important role in demonstrating the absence of
fluid migration behind casing in injection wells. At the present time, many
regulatory agencies require temperature logs to satisfy the mechanical integrity
test requirements of the Underground Injection Control (UIC) program. However,
these agencies have not established guidelines for conducting a temperature survey.
Two methods have been successfully applied and approved by regulatory agencies in
specific cases. The first method involves an injecting temperature log and a
series of shut-in logs run immediately after normal injection is ceased. The
second method requires a stabilized static base log followed by a period of
injection and a subsequent suite of post injection temperature logs. As these logs
have become more prevalent in mechanical integrity evaluations, experience shows
that the three (3) critical survey parameters are the shut-in time prior to static
base log, the volume of water injected, and the temperature differential between
the injected water and the formation water in the zone of interest. Recommended
procedures for running and presenting temperature logs have been developed based on
case histories of both Class I and Class II injection wells in the Midwest and
Nevada. These cases include logs conducted in wells with and without tubing, and
utilize both traditional differential temperature tools and the newer radial
differential temperature tool.
-22-
-------
INTRODUCTION
Ascertaining the mechanical integrity of injection wells is a major objective
of the Underground Injection Control (UIC) program. Under the program, various
state agencies and the United States Environmental Protection Agency require
injection well operators to demonstrate that the fluids they inject are staying
within the permitted disposal intervals and not contaminating underground sources
of drinking water. Also, the UIC program is concerned about any other flow between
zones penetrated by a well through channels behind the casing. The high
sensitivity of temperature logging tools to minute thermal disturbances has made it
a valuable tool in evaluating flow anomalies in injection wells. Properly run, the
temperature log can detect where injected fluids are being stored; whether injected
fluids are remaining in the receiving zone or channeling; and whether or not there
is other interzonal flow which may affect the quality of potentially useable
water.
PRINCIPALS OF THE TEMPERATURE SURVEY
An injection well is part of a complex heat transfer system where heat energy
is exchanged with formations surrounding the well. This heat transfer is dependent
on whether the well is acting as a heat source or sink. By analyzing the system,
information on the disposition of fluids into the well, and more importantly,
outside the wellbore in the formations can be obtained. The effect of fluid
movement will have a measurable influence on the heat flow.
Temperature logs enable the heat transfer that exists in a well to be
recorded. Also, by using various logging techniques, the heat transfer system can
be altered to investigate fluid migration problems. Using proper techniques, a
temperature log can provide information on the flow distribution taking place
-23-
-------
inside or near the wellbore. On most injection wells, the temperature log can
show:
1) where fluids are being stored within a disposal zone,
2) the point of entry or exit at the wellbore,
3) the source and path of flow behind casing, and
4) locations of interzonal flow not necessarily related to injection
activity.
To understand the temperature log application to injection wells requires a
consideration of heat transfer mechanisms within the earth.
GEOTHERMAL TEMPERATURE GRADIENT
The temperature within the earth increases with depth. A constant heat flow,
with its source at the molten core, is carried through the rock up to the surface
by conduction. The temperature decreases toward the surface which acts as a
radi ator.
Geothermal temperature gradient is a measurement of heat dissipation as it
rises through the earth to the surface. Generally, it is defined as the change in
temperature per 100 feet of depth. Geothermal gradients vary widely in the United
States and are dependent on the geology of a particular area and the ability of the
specific rock sequence to conduct heat. In the Midwestern United States the
geothermal gradient could be as low as 0.6 °F/100 feet, whereas in the Gulf Coastal
area it is about 2.3 "F/100 feet. In central Nevada where a case history is
presented, the geothermal gradient is 1.3 "F/100. Figure I shows the variance in
the temperature gradient in different regions across the United States.
Figure 2 illustrates how the temperature gradient response varies with type of
formation. These changes are due to the differences in the thermal heat transfer
-24-
-------
coefficient of the individual formations. These lithological differences must be
taken into account when investigating fluid channeling in a well. A lithological
log should be evaluated during both the planning and interpretation stages of a
temperature survey.
The effect that injected fluids have on the natural temperature gradient of a
well is best shown by case history presentations. First, however a description of
the logging tools is necessary.
LOGGING SYSTEMS
Two logging systems used to investigate the mechanical integrity of injection
wells include the conventional temperature system and the new radial differential
temperature system.
The basic configuration of the conventional system is described in Figure 3.
Three sections are common to this system. These are:
1) A tool which consists of a single temperature sensing element. This is
usually a high resolution platinum thermistor sensitive to temperature
changes of 0.1 °F.
2) A temperature section, which processes the line signal into a gradient
curve. This is the absolute temperature recorded by the tool.
3) A differential section which provides a calculated curve. This curve
responds to differences in the rate of temperature change.
The differential temperature curve increases the sensitivity of the
temperature gradient data. Although it does not furnish any new information not
included on the gradient curve, it presents the data in terms of relatively small
temperature changes which may not appear significant on the gradient curve. The
sensitivity of the differential curve can be varied by the logging engineer over a
-25-
-------
wide range. Figure 4 is an idealized presentation illustrating the differential
temperature response of a conventional single element temperature tool.
The second type of logging system used to evaluate injection wells is the
radial differential temperature (RDT) logging system. The ROT is a specialty tool
used to detect flowing channels behind casing. It is normally run in conjunction
with other investigative logs to confirm channeling where channeling is suspected.
Its primary function is to pinpoint the orientation of flowing channels once a
temperature anomaly is detected utilizing a conventional differential temperature
logging system. The use of the RDT tool as the primary source of demonstrating
mechanical integrity is not recommended.
A typical RDT tool is shown in Figure 5. The tool has two arms equipped with
temperature sensors positioned 180° apart that extend to contact the casing walls.
The contact diameter of the arms are adjusted to exert optimal pressure to maintain
contact between the temperature sensors and the interior of the casing. A motor
rotates the tool at a speed which is recorded on the left hand margin of the log.
The RDT tool is typically run into the injection well after the well is
shut-in. The tool is positioned adjacent to a point where channeling outside the
casing is suspected. The logging operator then extends the mechanical arms against
the casing and activates tool rotation. Where there is no flowing channel the
temperature sensors should measure a uniform temperature at all contact surfaces as
shown in Figure 6. If a flowing channel exists however, a sinusoidal wave will be
recorded as shown in Figure 7, indicating unequal heating or cooling of the
casing.
PLANNING AND EXECUTING A TEMPERATURE SURVEY
The planning and preparation of a production logging survey requires the
definition of the type of flow condition that may be encountered. This will enable
-26-
-------
the prediction of the expected temperature responses. The expected well conditions
will influence the log scale, injection procedures, intervals over which the
temperature logs will be run and whether or not the tubing and packer should be
removed.
The first step in temperature log planning and interpretation is to determine
whether the bore is acting as a heat sink or source. This will depend on whether
the fluids injected into the well are greater or less than the normal gradient.
For proper interpretation their must be a sufficient temperature change taking
place at the zone of interest.
The most important objective is to determine which portions of a zone are
accepting fluids and whether any migration out of that zone is taking place.
Shut-in temperature logs are the most effective means to detect whether an injected
fluid is remaining in the zone or channeling behind casing.
The path and storage of injection fluids are associated with the heat sink
effect of the earth. Generally, injected fluids are close to the surface
temperature which is usually less than the natural bottom hole temperature. The
following examples illustrate temperature surveys conducted in the Midwest and in
Nevada to demonstrate mechanical integrity. The first two case histories show
containment of injected fluids within the receiving interval with no upward
migration behind casing. The third example shows suspected channeling above the
fluid entry point into the formation. The channeling could be confirmed by
specialty temperature logging techniques.
Case History - 1
The first example is a temperature log performed to show the absence of fluid
channeling behind the long string casing of a Class I industrial disposal well in
Illinois. Non-hazardous wastewater had been injected into this well for the past
-27-
-------
17 years. The wastewater is injected at ambient temperatures. Since this logging
was conducted during the winter, the cold fluids injected had a cooling effect on
the well. The total depth of this well is 5524 feet with a disposal interval
consisting of 565 feet of porous dolomite.
The well was undisturbed for a period of 48 hours prior to running the base
temperature log. The base log showed the static geothermal gradient to be 0.6
°F/100 feet. The fluid level in this well was discernable from the base log and
recorded at a depth of 170 feet. Changes in the recorded conductivity are also
noted on the base log. The most significant occurs at a depth of 4060 feet as
shown in Figure 8. At 4060 feet there is a transition from the St. Peter, a
predominantly sandstone formation to the Prairie du Chien, a predominantly dolomite
formation.
These temperature shifts occur naturally due to the different thermal
conductivity of the changing rock matrix.
The heat transfer between the Eminence-Potosi injection zone and the Prairie
du Chien upper confining zone is evident by the cooling effect noted on the base
log below 4850 feet. This response above the disposal zone is resulting from the
vertical conductive cooling due to the injection of cool fluids below 4968 feet.
After the base log was completed approximately 163,800 gallons of cool (46°F)
fresh water was injected into the well down the 7" diameter casing. Three post
injection temperature logs were performed sequentially at 15 minutes, 2 hours and 4
hours following cessation of injection.
The post injection logs show that the majority of the injected fluid entered
the zone from 4970 feet to 5110 feet where the largest cooling effect is seen.
Each sequential post injection log pass shows the heat flow recovery to gradient
taking place. The recovery in various sections of the wellbore will be directly
proportional to the amount of cooling that has occurred under injection.
-28-
-------
The first post injection log shows the entire well remaining at a fairly
constant temperature. The second and third post injection logs show a definite
warming effect above 4970 feet. The logs are approaching the base log gradient
with similar slope which is characteristic of natural warming. This signifies
little or no fluid migration above 4970 feet. This depth corresponds to the base
of the confinement system indicating proper fluid isolation.
At the disposal zone, the rate of thermal recovery is reduced where the cold
water is stored. The post injection logs are showing this as a cooling effect due
to the mass of cold water in the disposal zone absorbing the heat flow.
Case History - 2
The next example is a case where produced brine was injected into a disposal
well for two years prior to conducting a temperature log. This well was a Class II
injection well located in Nevada. A section of the composite log showing the base
temperature log and three post injection runs is presented in Figure 9. This log
demonstrates that a good temperature log can be recorded with the tubing and packer
installed if the temperature differences between the injected fluids and the
formation are sufficient.
The base temperature log was run 92 hours after the well was shut-in. The
amount of shut-in time was due to an obstruction in the injection tubing which had
to be removed to allow the logging tool to go below the packer.
The cooling effect apparent on the base log between 8105 feet and 8155 feet is
the result of a temperature sink between the extremely cool formation below 8155
feet and the normal gradient at about 8105 feet. From 8155 feet to 8325 feet the
extreme cooling on the base log is due to the large volume of cooler water which
had been injected into the formation for approximately two years. Although the
geothermal injection water at the surface is 210 °F, the cooling effect of the
-29-
-------
formation temperature, (being less than 210 °F to a depth of 6100 feet), cools the
water to less than 190 °F from the surface to the injection point because the water
traveled 6100 feet being affected by cooler formations. Between 6100 feet and 8114
feet the temperature of the formation tries to raise the temperature of the
injected fluid, but due to the rate of pumping and the fluid traveling only 2014
feet at this increased temperature, the heating effect does not bring the water
back to any temperature above 190 °F before going into the disposal zone. The
temperature differential at the disposal zone then is 50 °F resulting in extreme
cooling of the formation. The base log shows that the largest quantity of
injection fluid is going into the formation between 8155 feet to 8300 feet.
The post injection temperature logs were run after pumping 375 barrels (15,750
gallons) of 52 °F surface water at a rate of three barrels per minute (126 gallons
per minute). The passes were made at approximately 30 minute intervals from 7100
feet to total depth at 8400 feet.
In post injection pass number one the fluid in the tubing just above the
packer is about 55 °F cooler than the base log. From 8137 feet to 8277 feet the
formation is being cooled by the 52 °F surface water being injected into the
formation. From 8277 feet to 8400 feet there is a heating back to a bottom hole
temperature of 235 °F. Injection occurred between 8137 feet and 8277 feet with
very little water being injected between 8277 feet and 8382 feet. No injection
occurred below 8382 feet.
Post injection log pass number two shows the fluid in the tubing just above
the packer to be about 45 °F cooler than the base log. This increase in
temperature means that the wellbore fluid is trying to reheat to normal gradient.
Between 8137 feet and 8277 feet the formation cooling is still evident due to the
375 barrels of cooler fluid injected into the formation. The heating effect still
evident from 8277 feet to 8400 feet further confirms that the largest quantity of
-30-
-------
fluid was injected between 8137 feet and 8277 feet, with very little water being
injected below that point and none below 8382 feet.
Post injection pass number three demonstrates the same effects as the two
previous runs except for the gradient heating another five degrees above the
packer. All three of the post injection logs come back to the same temperature
below the perforated intervals indicating good log quality control.
There is no channeling evident in this well.
Case History - 3
The previous examples showed temperature logging techniques applied using the
suggested procedures included in this paper. The results are exactly as expected.
In the next example, shown in Figure 10, a channel is suspected using conventional
gradient temperature and differential temperature logging techniques. Since the
injection well had previously been shut-in, a base temperature log was run before
injecting the cold test fluid. Two post injection logs were run to verify the
fluid entry point and demonstrate external mechanical integrity.
The majority of the injected fluid appears to be entering the upper perforated
interval from 5040 feet to 5060 feet as indicated by the cooling effect. There
also appears to be some injection into the upper ten feet of the lower perforated
interval from 5140 feet to 5160 feet. A rapid return to normal gradient indicates
that there is little or no injection below 5150 feet.
The gradient log shows possible inadequate cementing of the long string casing
above the perforation and possible channeling. The static base pass and the post
injection passes have opposing gradients between 4670 feet and 4990 feet indicating
a cooling effect from fluid moving outside the casing. The cooling effect extends
upwards to approximately 4750 feet before returning to normal gradient. This depth
correlates to a lithological change in the open hole logs and represents the top of
-31-
-------
the injection formation. There is no evidence of migration of fluid above the top
of the disposal zone.
Another log was run, using the ROT tool and a flowing channel was again
detected as previously depicted in Figure 7. The sinusoidal presentation confirmed
the presence of the channel at 4800 feet along with its orientation and vertical
extent. The tool was run above and below the suspected channeling interval. The
wave presentation above and below the zone of interest indicated no channeling as
previously illustrated in Figure 6. This example highlights the use of the RDT
tool as a secondary source of flowing channel identification.
TEMPERATURE LOGGING PROCEDURES
The following recommended procedures for running temperature logs on injection
wells is a compilation of recommended practices from various logging service
companies and the authors personal experiences. Injection wells and injection
practices are extremely varied and there are certain to be exceptions to these
rules.
The general approach will depend on the normal differential temperature
between the injected fluids and the receiving formation and the chemical properties
of the injected fluids. Obviously if the injected fluids are highly corrosive or
toxic it is recommended to run the logs after flushing the wellbore with fresh
water or brine. If the temperature of the injected fluid is near the same
temperature as the receiving zone special injection procedures may be required.
The two general cases are described as follows and shown graphically on Figure 11.
Case I
For wells in which the injected fluid temperatures are at least 35 °F greater
than or less than the temperature of the receiving zone the general approach is to
-32-
-------
run a log with the well in its stabilized normal condition prior to shut-in. Then
run a series of logs after the well is shut-in. This would mean a stabilized
injecting log and a series of post injection shut-in logs.
Case II
If the temperature of the injected fluids are similar to that of the disposal
zone or if the well has already been shut-in for a period of time the procedure is
more complex. In this case an artificially high or low temperature fluid may have
to be injected to impart a thermal change as was done in the Case Histories
presented. The well is generally shut-in for a time and allowed to stabilize
before the heated or cooled fluid is injected. Three critical survey parameters
must be determined. These are:
1) The time period that the well is shut-in prior to running a base log,
2) The temperature of the test fluid, and
3) The volume of the test fluid.
The logic diagram presented in Figure 12 may help determine whether Case I or
Case II above may be employed.
Shut-In Time
It is not necessary for the well to be shut-in until the temperature reaches
static conditions. This could take days or weeks in some cases. The pertinent
information is how the temperature is changing with time at all depths in the well
after the well's condition has changed. The tools available today are capable of
detecting small temperature changes accurately without having to wait a long time.
Also the longer the well is shut-in the longer it is unavailable for normal
-33-
-------
injection activities. A shut-in period of 24 hours is generally satisfactory for
mechanical integrity demonstration.
Temperature of Test Fluid
Although todays temperature tools are capable of extreme sensitivity, the
recorded logs are much easier to interpret when the changes in the wellbore are
relatively large. This is especially true if the survey is conducted without
removing the injection tubing. The best results are obtained when the difference
between the injected fluid temperature and the well bore temperature at the zone of
interest is at least 35 °F. The maximum heat flow occurs in the early part of the
post injection period. The maximum temperature difference between the borehole and
the surrounding formation exists at this time. This horizontal heat flow decreases
rapidly with shut-in time. Also the effects of vertical heat flow are less during
the early part of the shut-in period. For this reason it is recommended that a
post injection log be run immediately after the cessation of injection and one hour
after the cessation of injection. The timing of any subsequent post injection logs
can be determined based on the response of the initial post injection logs.
Volume of Test Fluids
The amount of temperature change induced in the wellbore is a function of the
volume of the injected fluid and rate of displacement as well as its temperature.
Heat transfer starts immediately as the injected fluid enters the well and the
thermal exchange takes place across the entire depth of the well above the
receiving interval. A sufficient volume must be injected so that there is enough
differential left to uniformly heat or cool the receiving zone. Injection should
also take place long enough to build up an injection pressure near to that of
normal injection operations. As a general rule of thumb the injection volume
-34-
-------
should be the greater of either three well volumes or one barrel of fluid per each
foot of disposal interval. For example if a 7" well is considered with 1500 feet
of disposal zone and a total depth of 4500 feet, a volume of at least 1500 barrels
or 63,000 gallons of fluid would be desirable. If a 7" well (2 gallons/ft) with
100 feet of disposal zone and a total depth of 4500 feet is considered, then a
volume of 4500 x 2 x 3 = 27,000 gallons would be desired. The higher the injection
rate the greater will be the differential temperature imparted at the zone of
interest. In general, the rate should be near that of the maximum permitted
injection rates or should be limited by the maximum permitted injection pressure.
Logging Speed and Direction
Most temperature logs are designed to give the best results when run at a
logging speed of 25-35 feet per minute. Running at a faster speed will tend to
spread out temperature anomalies or entirely miss small changes.
It is important to keep the logging speed constant throughout the survey.
Stopping the tool during a log run should be avoided. The logging speed should be
kept constant for all sequential passes.
The direction in which the well is logged is also an important factor.
Ideally, temperature surveys should be run only through undisturbed fluid. Since
the logging tool and electric line will disturb the fluid in the wellbore, the
temperature log should always be run while going into the hole.
Interval of Investigation
The temperature log should be started at least 300 feet above the area of
interest. In most injection wells which are undergoing routine mechanical
integrity testing, the objective is to determine if there is any channeling above
-35-
-------
the permitted disposal interval. Therefore in these cases the temperature log
should be started a minimum of 300 feet above the top of the receiving zone.
Calibration Scales
The calibration scale selected will depend on the differential between the
post injection logs and the base or injecting log. Frequent shifts in the log will
be required if the scales selected are too small. This makes the log difficult to
interpret. A scale range of 4 °F/inch to 10 °F/inch is generally best for
injection well logs conducted according to the preceding guidelines.
The actual scale determination may have to be made at the time that the log is
conducted.
Data To Include With Log
A temperature survey is meaningless when it cannot be correlated to the well
construction or conditions under which it was run. Data that should be included
on, or accompanied with, the log include:
1) Well pressure,
2) Time log was run,
3) Well conditions, shut-in or injecting,
4) Scales,
5) Injection rates if injection is taking place,
6) Construction features, and
7) Logging Speed.
To correlate the log back to other well logs, it is desirable to run the
temperature log in tandem with a casing collar locator and/or a gamma-ray log.
This is especially important if the log is being conducted with the tubing and
packer installed or if there are lithological changes at the zone of interest.
-36-
-------
REFERENCES
Cooke, Claude E., 1973, Radial differential temperature (RDT) logging - a new tool
for detecting and treating flow behind casing; Paper SPE 7558 presented at the
53rd Annual Fall Technical Conference of SPE-AIME, October 1978, 8 pp.
Dresser Atlas, Dresser Ind. Inc., Home Office, 1982, Interpretive methods for
production well logs.
N. L. Ind. Inc., N. L. McCullough, 1984, Systems approach to production logging, a
training manual for logging engineers.
Wei lex, no date, Temperature log interpretation, Document No. CL-2002, a training
document for logging personnel, Wei lex, a Halliburton Company.
-37-
-------
FIGURES
-38-
-------
FIGURE 1
TEMPERATURE GRADIENT VARIANCE IN THE UNITED STATES
-39-
-------
S31V1S
(2861 'd3SS3UCJ WOHd
3Hi NI aoNviavA iN3iavaD
tn to f-
DEPTHS IN THOUSANDS
-------
FIGURE 2
TEMPERATURE GRADIENT VARIANCES WITH TYPE OF FORMATION
-41-
-------
TEMPERATURE INCREASES
GYPSUM
ANHYDRITE
Thermal Conductivity in 10"^ Calories/Sec./Cm./°C
Shale 2.8 - 5.6
Sand 3.5 - 7.7
Por. Lm. 4-7
Dense Lm. 6-8
Dolomite 9—13
Quartzite 13
Gypsum 3.1
Anhydrite 13
Salt 12.75
Sulphur .6
Steel 110
Cement .7
Water 1.2-1.4
Air .06
Gas .065
Oil .35
FIGURE 2
TEMPERATURE GRADIENT VARIANCES WITH TYPE OF FORMATION
(MODIFIED FROM WELLEX)
-42-
-------
FIGURE 3
CONVENTIONAL SINGLE ELEMENT
TEMPERATURE LOGGING SYSTEM
-43-
-------
WIRELINE
TEMPERATURE
TOOL
TEMPERATURE
PROBE
GRADIENT
TEMPERATURE
PANEL
DIFFERENTIAL
TEMPERATURE
PANEL
FIGURE 3
CONVENTIONAL SINGLE ELEMENT
TEMPERATURE LOGGING SYSTEM
(FROM N-L-McCULLOUGH)
-44-
-------
FIGURE 4
DIFFERENTIAL TEMPERATURE RESPONSE
-45-
-------
TEMPERATURE
GRADIENT
NATURAL GRADIENT
x \
\
\
Q.
UJ
Q
DIFFERENTIAL
FIGURE 4
DIFFERENTIAL TEMPERATURE RESPONSE
(FROM N-L-McCULLOUGH)
-46-
-------
FIGURE 5
RADIAL DIFFERENTIAL TEMPERATURE TOOL
-47-
-------
(
-ANCHOR SPRING
•ROTATION MOTOR
ELECTRONICS
-CONVENTIONAL
TEMP. SENSOR
RDT ARM WITH
SENSOR PROBE
-CENTRALIZER
FIGURE 5
RADIAL DIFFERENTIAL TEMPERATURE TOOL
(MODIFIED FROM COOKE, 1973)
-48-
-------
FIGURE 6
RDT RESPONSE SHOWING ABSENCE OF CHANNELING
-49-
-------
-
CCL A
-20 MV 20
4600
DEPTH
1
4600
DEPTH
t—
\
\
\
\
\
\
}
(
}
(
\
\
1
S
5
^
\
i
<
\
I
\
t
j
(
\
i
r
)
'
*
\
RAD. DIFFERENTIAL
4 DES F 6
FIGURE 6
RDT RESPONSE SHOWING ABSENCE OF CHANNELING
-50-
-------
FIGURE 7
RDT RESPONSE SHOWING FLOWING CHANNEL
-51-
-------
TIME DATE ROT SERIAL # PROGRAM MODE JOB # FILE
STAT
CCL A
-20 MV 20
.
CCL A
-20 MV 20
4800
DEPTH
1
\
1
1
4SOO
DEPTH
RAD. DIFFERENTIAL
A DES F 6
•!
/
•^
<"
t,
^
'^
X""
^
^
\
1
y
f
•**,
/
^.
s
\
•^
s
>*
)
"N^
^
^\
r
^*^
x1
s^
X
V
,
RAD. DIFFERENTIAL
4 DEC F 6
FIGURE 7
RDT RESPONSE SHOWING FLOWING CHANNEL
-52-
-------
FIGURE 8
CASE HISTORY NO. 1 -
BASE AND POST INJECTION GRADIENT LOGS
-53-
-------
50°
•• • 1
— ' 1
— — t
— 1
— 1
!—
i — i
-
—
j*
—
1— j
«•
"f~
qp
1
T —
4-
f-
4 —
p^
*
I'M
H
'
t=
1
i
/
T-
-H-
-1—
-r-
_ L
— i
V"
I
1
^
— r
—
r
-i
-1-
-+
-i-
\
T~i
,
J—
[-4
.'.-1
)
-i- *
Li
tt-
-v
=±
-—*-
2=3
—
1
4-
T*
p<
-?»
33
S
4900
RUN
RUM 3
LOGS
5000
5100
s
BASE LOG
NH
EST
CASING
COLLAR:
DV TOOL
CASING:
SHOE
V)
O
eo
"
i
<
O
100°
-54-
-------
FIGURE 9
CASE HISTORY NO. 2 -
BASE AND POST INJECTION GRADIENT LOGS
-55-
-------
"' iiRUN 3
PERFORATIONS^
--DEGREES FAHRENHEIT
u.
O
-56-
-------
FIGURE 10
CASE HISTORY NO. 3 -
BASE AND POST INJECTION GRADIENT AND DIFFERENTIAL
LOGS SHOWING SUSPECTED CHANNELING
-57-
-------
TEMRDEG. F)
90.00 110.00
DIFFERENTIAL TEMP
i:.
X
FIGURE 10
CASE HISTORY NO. 3 - ,
BASE AND POST INJECTION
GRADIENT AND DIFFERENTIAL
LOGS SHOWING SUSPECTED
CHANNELING
-58-
-------
FIGURE 11
GENERAL TEMPERATURE LOGGING PROGRAMS
-59-
-------
CASE - 1
NORMAL INJECTION ^-SHUT WELL IN ^--RETURN TO NORMAL OPERATION
INJECTING TEMP. LOG^ VPOST INJECTION LOGS
o
i
CASE - 2
NORMAL INJECTION SHUT-IN /BASE TEMP. LOG x POST INJECTION LOGS
/ ^- /
{ I L I
t
^SHUT-IN PERIOD MNJECT TEST FLUID ^RETURN TO NORMAL
OPERATION
TIME
FIGURE 11
GENERAL TEMPERATURE LOGGING PROGRAMS
-------
FIGURE 12
LOGIC DIAGRAM FOR TEMPERATURE LOGGING
TO DEMONSTRATE MECHANICAL INTEGRITY OF INJECTION WELLS
-61-
-------
M TO MOHMAt OPtHATIOH [ ! (JO TO HOT. HOMf OH RAT LOO
FIGURE 12
LOGIC DIAGRAM FOR TEMPERATURE LOGGING
TO DEMONSTRATE MECHAN5CAL INTEGRITY
OF INJECTION WELLS
-62-
-------
INJECTION MONITORING AND CONTROL
DOLLARHIDE CLEARFORK "AB" UNIT
T. S. Collier
Unocal
Midland, Texas
ABSTRACT
The Dollarhide Clearfork "AB" Unit, a West Texas waterflood,
currently produces 1600 BOPD and is expected to recover 37
percent of original oil in place. Of this 37 percent, more
than half is attributable to effective waterflood operations.
In order to effectively waterflood this field, control of
injection water plays a critical role.
This paper describes the benefits of injection monitoring and
control both from a standpoint of protection of ground water
and increased oil recovery. It describes how injection
performance, production performance, radio-active tracer
surveys, and temperature surveys were used to quantify and
identify injection that was not entering the target interval
in the Dollarhide Clearfork "AB" Unit. Discussions are
presented on the various causes of "out-of-zone" injection as
well as several remedies for this problem. Finally,
-63-
-------
additional oil recovery is shown to be directly related to
the monitoring and control of injection water-
Background Information
The Dollarhide Clearfork "AB" Unit is located in Andrews
County, Texas near the Texas-New Mexico border (Fig. 1). Oil
production averages 1600 barrels per day from 60 producing
wells. Average daily water injection is 7000 barrels per day
into 30 water injection wells.
The Clearfork "AB" Unit has three productive zones. They are
the Upper "A", the "A", arid the "B" zone. The Clearfork
formation is encountered at an average depth of 6500 ft.
(1980 m). As shown in Figure 2, the Clearfork formation is a
North-South trending anticline, and although it is not shown
on the figure, there is closure to the south. The lithology
is predominantly limestone.
On June 1, 1959, the various leases in the Dollarhide Field,
excepting one operator, were unitized for the purpose of
establishing a waterflood. A small scale pilot waterflood
comprising two water injection wells was initiated. This
pilot waterflood was expanded in November, 1961 to include
six injectors. Increased water production in offsetting
wells was detected indicating a possible problem with
-64-
-------
injection control. This problem was corrected and the
waterflood was expanded to full scale in May, 1964.
In 1959, prior to unitization and waterflood operations, the
life of the Dollarhide Clearfork Field was estimated to be
twelve years, based on production decline data. By the
implementation of a well designed, closely monitored
waterflood, the life of this field has been extended into
the next century and will, in all probability, allow it to be
produced using CO2 for enhanced oil recovery.
Benefits of Controlled Injection
Ultimate Recovery.
The chief benefit of controlled injection to the operating
company is reduced operating costs. This is accomplished
through several mechanisms, allowing the operating company to
recover more oil economically from any given project.
The cost reduction takes many forms. The most obvious of
these is associated with injecting less water to achieve the
desired waterflood performance. The procurement and
pressurization of water in a waterflood is often a costly
process. Since the cost of injection is the same for water
that enters the target interval and enhances oil production
as for water which does not, it is important that injection
-65-
-------
water is confined to the target interval.
Sometimes, injection water will exit the wellbore, into to a
high permeability lens (or "thief" zone), and then proceed to
an offset producing well. For example, if the well was
completed "open hole" (the casing is set just above the
target interval leaving the target interval uncased)
injection water may preferentially exit the open hole into a
few fairly thin intervals. Another example involves wells
which are completed with casing cemented through the
producing zone. Occasionally, the bond between the formation
and the cement used to secure the casing has insufficient
strength to isolate these "thief" zones. Water entering thin
intervals having high permeability contributes little, if
any, to additional oil production, but requires additional
expense to produce. The costs associated with producing a
barrel of water are the same as producing a barrel of oil.
Since oil recovery is predicated on continuing favorable
economics, each increase in operating costs is associated
with a decrease in ultimate oil recovery-
Leak detection.
Injection monitoring and control can assure that injected
water does not enter the ground water aquifer- This aspect
-66-
-------
of injection control is especially important in the
Dollar-hide Clearfork "AB" Unit as the climate is semi-arid
and water wells are the sole source of water for livestock.
Fortunately, the signals that a leak has occurred into the
annulus which could further escape into the ground water
aquifer is readily detected.
Figure 3 shows a schematic cross section of a typical water
injection well including (1) surface casing which is solidly
cemented from the base of the aquifer to surface, (2)
production casing which penetrates the production - injection
zone and is cemented in place, and (3) tubing string with a
packer set immediately above the zone into which water is
injected. Deviations from the type of completion shown in
Figure 3 are often necessary, or desirable. For example,
unusual drilling problems which are encountered at Dollarhide
make it necessary to run an additional or "intermediate"
casing string at 3100 ft. (940 m), which is placed between
the surface and production casing string. In contrast, other
shallow oil reservoirs require only a production string, thus
eliminating the need for surface casing.
Water to be injected is introduced into the tubing at the
surface and enters the oil zone through perforations in the
casing. The packer prevents water from contacting the
-67-
-------
production casing opposite the aquifer. Surface pressure of
the annular space between tubing casing is monitored and if
communication occurs, a pressure increase will be observed at
the surface. Corrective action can then be taken to repair
or replace the tubing or packer, as necessary, which is the
first line of defense. In this schematic, second and third
lines of defense are provided by the production and surface
casing strings, respectively.
Monitoring Methods
Tubing-Casing Annulus Pressure.
As noted above, the most effective way to verify that fresh
water aquifers are not being impacted is by the diligent
monitoring of tubing-casing annulus pressures. If no
pressure exists, then communication with the ground water
aquifer is not taking place.
Well Performance.
A valuable tool available to the petroleum engineer in
evaluating the effectiveness of underground injection is the
analysis of the production performance of the wells which
offset an injection well. Early water breakthrough into the
producing well indicates that injected water is most likely
exiting the injection wellbore into a high permeability lens
of limited size. Further corrective action may be warranted,
-68-
-------
as it was in the Dollar-hide Clearfork "AB" Unit.
Radio-Active Tracer Surveys.
One very useful tool for tracking and quantifying water exit
from an injecting well is the use of radio-active tracers.
By injecting a small amount of radio-active material and
measuring the length of time it takes to travel a certain
distance within the wellbore, it is possible to determine the
amount of injection water exiting the wellbore over a given
interval. This information is very useful in designing any
corrective action which may be required.
Temperature Surveys.
Another useful tool in determining injection water exit from
the wellbore is the temperature survey. By recording the
wellbore temperature vs. depth, an analysis may be made of
intervals where injection water is leaving the wellbore. The
temperature survey yields interpretations which are more
qualitative than the radio-active tracer, but offer a
slightly better idea of what happens to the injectant
after it leaves the wellbore.
-69-
-------
Mechanisms of Out-of-Zone In.lection.
Mechanical Integrity.
One cause of injection outside of the target zone may be a
lack of mechanical integrity. This is evidenced by an
increase in pressure on the tubing-casing annulus. Prompt
attention to the situation and timely repair should ensure
that injection water does not enter the tubing-casing
annulus.
Primary Cementing Procedures.
Occasionally, the primary cementing procedures used in older
wells did not achieve sufficient bonding to the pipe or the
formation to prevent the flow of injection water behind
casing. This situation can be corrected by squeezing cement
into the formation and behind the primary cement.
Dollarhide Clearfork "AB" Pilot Flood Performance
Description of Pilot Flood.
As shown in Figure 4, six producing wells were converted to
water injection service in November, 1961. Of the six, five
were completed open hole. The remaining well was a dual
completion and was perforated in the Lower "A". Water was
injected into the six wells and production performance was
-70-
-------
monitored in the offsetting producing wells.
Early Injection-Water Breakthrough.
After only six weeks of water injection, water production was
observed in well number 15-72-C. By July, 1962, just eight
months after injection was initiated, water breakthrough had
been observed in eight offset producing wells. Water
production steadily increased while oil production
diminished. Finally, in October, 1962, injection was
discontinued.
In the case of the Dollarhide Clearfork "AB" Unit, the key
indicator had been early water breakthrough. Further
investigation using tracer surveys indicated that a thin,
highly permeable zone at the top of the Upper "A" interval
was acting as a "thief" zone within the oil reservoir. Water
injected into the wells was not reaching the target interval,
but instead was virtually all exiting from the top 150 feet
of open hole.
During the period while injection was discontinued between
October, 1962, and March, 1963, oil production from two
offsetting wells decreased by a total of 20-30 barrels per
day. This indicated that some oil response had been achieved
and that a solution of the breakthrough problem would result
-71-
-------
in a significant increase in oil production.
The Solution.
Several alternatives were considered, such as re-cementing
and recompleting, attempting to cement off selected
intervals, and cementing inner liners which are placed
opposite the injection interval. The latter alternative was
selected as the most effective solution for several reasons.
A steel liner has the best mechanical integrity of the three
methods considered. In addition, liners offer the best
wellbore stability and are superior to cement repairs for
isolation of the injection intervals. The drawback of this
alternative was the reduction in internal diameter which
would make future workover operations more difficult. Also,
it was the most expensive of the three solutions considered.
In designing the liner installations, care was taken to
properly balance all considerations. In this type of design
there is a trade-off between the size of the pipe (the larger
the pipe the fewer the problems during future injection and
workover operations), and the likelihood of obtaining a good
primary cement bond. After considering the size and
condition of the open hole section, liner sizes were were
determined on a well-by-basis. During the five month period
from October, 1962, to March, 1963, five liners were
-72-
-------
installed and cemented, after which time water injection was
resumed.
Results.
Following the installation of liners in the Dollarhide
Clearfork "AB" Unit Pilot Waterflood, water production was
stabilized and oil production was increased, (Fig. 5).
Water production was virtually eliminated in three of the
offsetting production wells . Based on the results of the
pilot waterflood, full scale water injection was initiated in
the Dollarhide Clearfork "AB" Unit in May, 1964. The
techniques used to monitor and control the water injection in
the pilot water flood have been extended throughout the field
and have enabled Unocal to double total oil recovery over
primary depletion.
Conclusion
Close control and monitoring of injected fluids in a
secondary recovery project can improve economics and
reserves. This objective may be achieved by close
surveillance of tubing-casing annulus pressures, by carefully
monitoring the production performance of offsetting producing
wells, and by using various wellbore surveys.
-73-
-------
FIGURE
't
INDEX MAP
SHOWING
DOLLARHIDE FIELD
ANDREWS COUNTY,TEXAS
-74-
-------
FIGURE 2
""
~K.lt' UNIT
CLEAR FOHK WELLS ONLY
DOLLARHIDE CLEAR FORK AB UNIT
STRUCTURE - TOP OF B ZONE
1 MILE
-75-
-------
FIGURE 3
///^V/A^
\
X
CASING
PERFORATIONS
///sy/A\y7/
AQUIFER
-SURFACE CASING
-INJECTION TUBING
-PRODUCTION CASING
INJECTION PACKER
TARGET INTERVAL
TYPICAL WELLBORE SCHEMATIC
-76-
-------
FIGURE 4
H
ol
o
Gp-
2 j >." ,
JL
UJ
4]
>-
z
13
0
o
_ _
Tlnio, HS
.51 ^iO •"
1 a
..o. X-l X .I
1 /o | / X"
x'"
/
/
^Its
V
v
-^
^
91 j
II
II
i
jd
1* •"
1
-
1
— r
i
i
/
1 * 1 ^ *
/ 1
i PI LOT AREA
, X " i • •* X"
* 1 /4
f />. Co.tfr/i
\ Unlor. j.
^."1 x" -1*
\, 'A" /A"
l^^l ^iw)
L L
1 3Z
• rf 10 •
1 •
/«•' «!'
f 4 Ce*din
Mob.l j
i
•
JLL .
I6| ||
f .
1 ' *
1
1 -
vttu
/•*•*• •" i /»
7 1 5
•*' X"| •'*
\ L-' L
jfn ^ •• 1 >'••
17
.«
J
X"
4
•I
/" i /ff
/5 / 1
/ | I
* i
/ C P. Comiiit \
' Union
X' •""{" x"
18 X*o
24
«
. 109 X'°*| • I0t X'0'
I
JJ |(pj J4 |() 1 J5
1 '
j
" I X' 1
IC'.P Cavttn
."
. Mobil } Humbll 4
\ •"'
1 i
1
— LEOCNO —
UAL COMPLETION! DESIGNATED IV
OS (>•*«". a»- »il»f lan
.
:. " . .
^' \
1 |
'I
Mobil
P
)LK.
I*
SL
A - 52
1 "*K£ltf Dtvllopmtnt
/
19
20
Uobil
Mobil |
• * j
UU.U.^J
— -1
,
^' !
j]
X1 a
"M
l~* — * ""MO
1
1
1
1
c,,ain
bil
23
n
t.P Cowc
Mobil
3
AREA MAP
AREA OF
PARTICIPATION Foil
CLCAKFORK '»* RCSCHVOIR
DOLLARHIOE
FIELD
ANORCWl COUNT*
TIXAI
"
-77-
-------
400 .
300 .
Q
O.
CD
Q
O
cc
CL
200 .
100 -
Runni ng Li ners
FIGURE 5
PRODUCTION PERFORMANCE
PILOT FLOOD AREA
J
1963
-78-
-------
NIPER Paper No. EPR/OP-87/10
SUBSURFACE INJECTION OF FLUIDS FOR THE RECOVERY OF PETROLEUM
By A. Gene Collins and Herbert B. Carroll, Jr.
IIT Research Institute
National Institute for Petroleum and Energy Research
P. 0. Box 2128
Bartlesville, OK 74005
To be presented at the
UNDERGROUND INJECTION PRACTICES COUNCIL
INTERNATIONAL SYMPOSIUM ON SUBSURFACE
INJECTION OF OILFIELD BRINES
New Orleans, Louisiana, May 4-7, 1987
COPYRIGHT WAIVER
By acceptance of this article for publication, the publisher recognizes the
Government's (license) rights in any copyright and the government and Its authorized
representatives have unrestricted rights to reproduce In whole or in part said
article under any copyright secured by the publisher.
DISCLAIMER
This report was prepared as an account of work sponsored by an agency of the United
States Government. Neither the United States Government nor any agency thereof, nor
any of their employees, makes any warranty, express or Implied, or assumes any legal
liability or responsibility for the accuracy, completeness, or usefulness of any
Information, apparatus, product, or process disclosed, or represents that its use
would not infringe privately owned rights. Reference herein to any specific
commercial product, process, or service by trade name, trademark, manufacturer, or
otherwise, does not necessarily constitute or imply its endorsement, recommendation,
or favoring by the United States Government or any agency thereof. The views and
opinions of authors expressed herein do not necessarily state or reflect those of
the United States Government or any agency thereof.
-79-
-------
TABLE OF CONTENTS
Page
Acknowl edgments v
Abstract vi
Introduction 1
011 Recovery Mechanisms 3
Primary Recovery 4
Secondary Recovery 4
Tertiary Recovery 6
EOR Selection Methodology. 7
Laboratory Tests 7
Water and Rock in Secondary and Tertiary Recovery Operations 8
Injection Water 8
Water Sources 8
Formation Water „ 9
Fresh Water 10
Seawater 11
Water Compatibility 12
Core Flow Tests 12
Corrosion 13
Bacteria 14
Formation Rock Minerals 14
Fluid Injection Treatment Systems 16
-80-
-------
TABLE OF CONTENTS (Continued)
Page
Types of EOR Operations 16
Micellar-Polymer EOR Operation 16
Polymer 18
Alkaline 19
Carbon Dioxide 20
Steam 21
In Situ Combustion 21
Miscible Hydrocarbon 22
Inert Gas Injection 22
Microbial Flooding 23
Cyclic Microbial Flooding 23
Quantity of Chemicals Used in EOR 24
Mobility Control Agents (Polymers) 24
Cosurf actants 25
Alkaline Flooding Agents, Preflush Agents, Thermal Enhancers... 25
Surfactants 25
Biocides, Chelating Agents, Oxygen Scavengers 26
Transport and Fate 26
Conclusions 26
References 28
TABLES
1. Geochemical Water Analyses 31
2. Tertiary System 32
3. Toxicological Data 33
-81-
-------
TABLE OF CONTENTS (Continued)
Page
ILLUSTRATIONS
1. Oil Production 35
2. Crude Oil and Water Produced (Including Alaska) 36
3. Crude Oil and Water Produced (Excluding Alaska) 37
4. Chemical Flooding (Micellar-Polymer) 38
5. Chemical Flooding (Polymer) 39
6. Chemical Flooding (Alkaline) 40
7. Carbon Dioxide Flooding 41
8. Steam Flooding 42
9. In-Situ Combustion 43
10. Nitrogen — C02 Flooding... 44
11. Microbial Flooding 45
12. Cyclic Microbial Flooding 46
-82-
-------
ACKNOWLEDGMENTS
The authors appreciate the support of this work by the U. S. Environmental
Protection Agency (EPA) through Contract/IAG DW89931947-01-0 and the U. S.
Department of Energy (DOE) through an interagency agreement with the EPA. The
authors also thank Bill Linville for his encouragement and for editing the
manuscript and Joe R. Lindley who prepared the drawings of enhanced oil
recovery processes.
-83-
-------
SUBSURFACE INJECTION OF FLUIDS FOR THE RECOVERY OF PETROLEUM
By A. Gene Collins and Herbert B. Carroll, Jr.
National Institute for Petroleum and Energy Research
Bartlesville, OK 74005
ABSTRACT
This report addresses the major methods used to recover petroleum which
are classified as (1) primary, (2) secondary, and (3) tertiary or enhanced oil
recovery (EOR). Further, EOR methods which include miscible, thermal, and
chemical are described. Subsurface injection of fluids is used in secondary
and tertiary petroleum recovery operations.
The report notes that one of the most important criteria relevant to an
injection operation is adequate geologic and engineering characterization of
the subsurface reservoir. Reservoir screenings and detailed characterizations
of reservoirs are made by use of appropriate computer models.
Laboratory studies are conducted using core samples taken from the target
injection zone in conjunction with appropriate dynamic flowthrough core
apparatus, whereby porosity, permeability, ion exchange, clay sensitivities,
rock wettability, miscibility, etc. are determined. The laboratory data and
the characterization data are used in an appropriate computer model to predict
the probable hydrologic transport and flow of the injected fluids and the
targeted petroleum. If these studies indicate a high probability of success
for economic petroleum recovery, the next step is a pilot field test. If the
pilot test indicates that an economic amount of petroleum can be recovered,
then a full-scale field operation is designed and properly sited, wells are
drilled, injection and production equipment is installed, and the petroleum
recovery operation begins.
Important operations may include reservoir preflush for the removal of the
connate brine; injection fluid treatment to mitigate clay sensitivities or to
-84-
-------
prevent corrosion and incompatible reactions. The waters used in injection
operations consist of formation water, fresh water, or seawater, and
consideration must be given to fluid-fluid interactions and fluid-rock
interactions.
Micellar-polymer, polymer, alkaline, carbon dioxide, steam, in situ
combustion, miscible hydrocarbon, inert gas, and microbial EOR processes are
briefly described. The types and amounts of some of the injected chemicals
also are addressed.
-85-
-------
INTRODUCTION
Subsurface petroleum reservoirs possess natural pressure, and when a
producing well is drilled into the reservoir, the pressure is reduced,
creating a pressure differential which moves the oil and gas from the
reservoir into the well and to the surface. This pressure is caused by water
pressing upward from beneath the petroleum (water drive); a gas pressing
downward (gas cap drive); by gas in solution (solution gas drive); or by all
of these working together. In most reservoirs, initial pressure is strong
enough to lift the oil to the surface of producing wells; however, as
reservoir pressure declines with cumulative oil withdrawals, "artificial lift"
is required to raise petroleum to the surface. This is accomplished with
downhole pumps lifting the oil to the surface or by injecting gas deep into
the fluid column to lighten the weight of the fluid (gas lift).
Even when reservoir pressure is depleted and no longer lifts oil to the
surface, the reservoir pressure may be adequate to move petroleum through the
formation into the well bore. Primary recovery, or production relying
entirely on natural forces, often recovers a substantial portion of a field's
total petroleum reserve.
Natural forces are wastefully dissipated when inefficient production
procedures are used. In the oil booms of yesterday, when "boomers" rushed to
drill as many wells as possible and produce oil as fast as they could, total
recovery was far less than that of today's methods. Oil reservoirs must be
carefully managed to conserve pressure and optimize recovery.
Today, the number, location, and producing rates of oil wells are planned
to maximize recovery and to maintain production as long as possible. Natural
forces are augmented by injecting replacement fluids like water and/or gas,
and these efforts are known as secondary recovery operations.
-86-
-------
The methods described as "primary" or "secondary" operations move only
part of the oil, often leaving as much as 40 to 80 percent unrecovered. Even
a we11-engineered waterflood leaves more than one-third of the original oil as
unrecovered residual oil. The national average for oil recovery by both
primary and secondary methods is only about 34 percent.
Enhanced or "tertiary" methods recover residual oil by increasing the
volume of the reservoir contacted and by reducing interfacial tension. These
enhanced methods are classified as follows:
• Thermal recovery. Heated oil flows more easily through the reservoir
rock. It may be heated by injecting high-pressure steam into the
reservoir or by actually burning some of the crude oil in the
reservoir rock (fireflooding).
• Miscible recovery. Miscibility is the ability of fluids to mix with
each other to form a single phase. Normally, oil and water separate
into layers and are not miscible. Some fluids that mix with oil are
effective in displacing oil from reservoirs; for example, light
liquid hydrocarbons, such as propane and ethane, which are extracted
from natural gas. Carbon dioxide is also miscible with oil.
• Chemical recovery. Chemicals with large molecules, such as polymers
which "thicken" water when added in low concentrations to water, are
used to enhance recovery by improving the ability of water to "wash"
or "sweep" oil from the rock pores. Surfactant flooding calls for a
combination of surfactants (special detergents) and polymers used to
recover residual oil that remains trapped after secondary recovery.
A "bank" or "slug" of fluid (mostly water) containing surfactant is
injected to reduce the interfacial forces trapping the residual oil
allowing it to flow to the producing wells. The surfactant bank is
-87-
-------
followed by water usually thickened with polymer to maximize the
volume of reservoir contacted.
OIL RECOVERY MECHANISMS
There are three major types (or mechanisms) of recovery of oil from
subsurface reservoirs: primary, secondary, and enhanced. Each type of
recovery is associated with the original-oil-in-place, the remaining oil-in-
place (subsequent to recovery or production operations), and the pressures
within the reservoir. For example, when a well is drilled into a subsurface
reservoir containing oil, tests are conducted to determine the amounts of oil,
water, and gas that are present. This information plus knowledge of the
depth, reservoir thickness, reservoir pressure, reservoir lithology, and
results from specific production tests permits accurate calculations of the
amount of oil in the reservoir. Further, calculations can indicate how much
oil should be produced by primary recovery when primary recovery is defined as
oil produced from a well as a result of oil flowing and finally pumping the
reservoir until it is depleted or no longer economical to operate. Secondary
recovery usually involves repressuring by gas injection or water injection,
i.e., simple waterflcoding. The third or tertiary phase employs more
sophisticated technology such as altering one or more properties of the crude
oil to reduce surface tension. This technology is known as enhanced oil
recovery. Tertiary recovery often is accomplished by injecting water mixed
with specific chemicals that "free" the oil adhering to the porous rock so
that it is taken into the solution and pumped out of the well.
Figure 1 illustrates the three major oil recovery operations where, during
primary recovery, 12 to 15 percent of the original oil-in-place is produced.
Secondary recovery can produce an additional 15 to 20 percent of the oil
reserve, and enhanced oil recovery (EOR), another 20 percent.
-88-
-------
Primary Recovery
As noted in figure 1, primary recovery refers to oil that can be recovered
from the subsurface reservoir through the natural energy of the reservoir.
Artificial lift such as pumping may be used, but injection of water is not
used in primary recovery.
Secondary Recovery
The widespread application of waterflooding (Craig, 1971) to boost
production after initial decline in primary production led to this process
being called secondary recovery. For regulatory and pricing purposes
waterflooding has been set apart from other forms of EOR. In a typical
waterflood, the "watercut" in the produced fluid continually increases, and
the expenses of pumping, separation, and disposal of the floodwater eventually
exceed the income from the oil recovered. Then secondary recovery efforts are
halted even though oil may remain in the reservoir.
The effectiveness of secondary recovery is dependent on the volume of the
reservoir contacted by the injected fluid, which is dependent on the
horizontal and vertical sweep efficiency of the process. Factors which
control the sweep efficiency are (1) pattern of injector wells, (2) off-
pattern wells, (3) unconfined patterns, (4) fractures, (5) reservoir
heterogeneity, (6) continued injection after breakthrough, (7) mobility ratio,
and (8) position of gas-oil and oil-water contacts. (Langnes et al., 1985)
Selection of an injection pattern is one of the first steps in the design
of a secondary recovery project. In making the choice, it is necessary to
consider all available information about the reservoir. The adverse effects
of the factors listed above can be offset if they are considered during the
pattern selection. Other factors to consider in pattern selection are
(1) flood life, (2) well spacing, (3) injectivity, (4) response time, and
-89-
-------
(5) productivity.
Waterflood life depends on the availability of injection water, the rate
at which it can be injected, well spacing, and proration policies. The
performance and economics for various well spacings and pattern sizes should
be analyzed in order to pick the economically optimum choice. These analyses,
however, cannot be made without considering injectivity, which is best
determined using pilot operations, and a well designed and applied pilot
operation is essential to understanding all the pattern selection factors.
An ordinary waterflood, operated at practical rates with ordinary water or
brine, is physically incapable of displacing all of the oil from reservoir
rock. Capillary forces acting during the waterflood may cause part of the oil
to be retained in water-wet rock as disconnected structures which do not flow
under the pressure gradient from the flow of water. The detail of these
structures is directly related to the microscopic mechanism of oil
entrapment. Thus, even in those regions of the reservoir which are relatively
well-swept, i.e., regions through which relatively large quantities of water
flowed, a residual oil saturation can range from 15 to 40% of pore space. The
residual oil saturation in well-swept regions of proven accessibility with
respect to injected fluids is an important target, though a difficult one, for
EOR.
Ordinary waterflooding is a less expensive process than most EOR
operations. However, the economics of waterflooding becomes uneconomical when
the revenue produced by the amount of oil recovered is less than the cost of
waterflood injection, which may occur when the residual oil saturation is as
high as 40% of pore space to as low as 15% of pore space.
-90-
-------
Tertiary Recovery
The target oil for recovery is the residual oil in the reservoir that is
left after primary and secondary recovery operations. Tertiary recovery by
EOR methods usually is a more expensive operation and is not usually applied
unless the price of oil is sufficient to pay the costs of producing the oil
from the subsurface reservoir. In this report, we shall refer to tertiary
recovery as enhanced oil recovery or EOR.
Petroleum production from reservoirs under primary, secondary, or EOR
processes involves the simultaneous flow of two or more fluids. Multiphase
flow, particularly three-phase flow, is not well understood or adequately
described analytically, even for pipeline flow. With natural porous media
with complex geometry, a microscopic description of the multiphase fluid flow
process is not possible. Empirical macroscopic descriptions based on Darcy's
work, relating fluid velocity to pressure gradient and viscosity through a
constant called permeability, permits the needed fluid-flow calculations.
Multiphase flow of fluids through porous media is related to a relative
permeability of each phase, fluid viscosities, pressure drop, capillary
pressure, and permeability; however, the relative permeabilities are the least
understood and the most difficult quantities to measure.
The effectiveness of EOR is dependent upon the same variables as secondary
recovery with regard to sweep efficiency, injection patterns, etc. Since EOR
usually is more expensive to implement per barrel of oil recovered, the
preliminary work before implementation often is more detailed and exacting
than for primary and secondary recovery operations. The studies often involve
geological reservoir characterization, laboratory studies, computer simulation
studies, and field pilot studies.
-91-
-------
EOR SELECTION METHODOLOGY
Since the oil targeted for EOR is difficult and expensive to obtain, the
oil producer wishes to apply only the most cost-effective technology to
extract the oil. Selection of the most cost-effective technology requires
several studies, as noted by Goodlett, et al. (1986). Detailed information
concerning geological, chemical, physical, and engineering characteristics of
the target reservoir rocks and fluids (oil/gas/water) is used along with
screening parameters to make a preliminary EOR selection. Subsequent to
selection of a candidate method, basic laboratory tests are performed
including dynamic fluid flowthrough core experiments using simulated
subsurface pressures and temperatures.
Information gathered from these tests, plus other relevant knowledge, is
used as input variables for numeric computer models which helps decide the
viability of the selected EOR process. Other relevant knowledge includes
reservoir characterization in as much detail as possible. The presence of
certain minerals and/or reservoir heterogeneities adversely affect EOR.
Knowledge of micro-scale reservoir heterogeneities such as dead-end pores,
pore throat size, and tortuosity also is important.
Laboratory Tests
Goodlett, et al. (1986) described some of the numerous experiments and/or
tests that should be conducted before implementation of even a pilot EOR
operation. For example, scaling should be determined by application of linear
scaling principles to better reproduce the basic operative physical and
chemical mechanisms which will occur in the reservoir. Scaling experiments
are accomplished through the use of laboratory core floods. Cores used in
laboratory core floods range from sandpacks to native-state reservoir samples
which are obtained and retained at subsurface conditions of temperature,
-92-
-------
pressure, and fluid saturations. Native-state cores are the most expensive
and most useful porous-media system for EOR evaluation.
Core wettability is a critical factor in evaluation, and alteration of the
wettability can occur during the operations of obtaining a core. Other
important tests include injectivity, plugging, mobility control, relative
permeability, oil saturation, rock-fluid and fluid-fluid interaction, etc.
WATER AND ROCK IN SECONDARY AND TERTIARY RECOVERY OPERATIONS
INJECTION WATER
Items that should be considered before implementation of a fluid injection
project involving any type of injection water include the following:
(1) formation type; (2) formation quality such as clay content; (3) formation
porosity and permeability; (4) depth of formation; (5) fracture-opening
pressure of formation; (6) fracture-breakdown pressure of overlying and
underlying formations and; (7) compatibility of injection solutions with
fluids already in the formation and with the formation rock material.
Petroleum reservoir rock formations are filters and are susceptible to
plugging by any type of solid material which may be suspended in or
precipitated from an injection fluid. Even materials such as oil and grease
from the pumps, corrosion inhibitors, and bactericides can cause plugging
problems.
Table 1 lists the items typically requested in analyses of a produced
oilfield water; water used in injection for pressure maintenance for secondary
recovery for EOR; water used to generate steam for steam injection; and water
injected into a disposal well.
Water Sources
Three major types of water are used for injection: formation water,
seawater, and fresh water.
-93-
-------
Formation Hater
Formation water is subsurface brackish or brine water usually produced
from a petroleum producing formation. Table 2 illustrates the composition of
some formation waters taken from some Tertiary Age formations. The table
gives the highest value found in milligrams per liter for a given constituent,
the average values, and the number of samples used to estimate the average
value, Collins (1975).
An estimate of the amounts of water that are in various reservoirs was
made for the State of Oklahoma. The estimate indicated that Oklahoma has
about 3.4 trillion gallons of surface water possessing a quality of 100 to
1,000 ppm dissolved solids (OS); about 5.0 trillion gallons of ground water
with a quality of 280 to 4,000 ppm OS; about 23.6 trillion gallons of
formation water down to 5,500 feet deep with a quality of 15,000 to 110,000
ppm DS; and 35.8 trillion gallons of formation water down from 5,500 to 8,500
feet deep with a quality of 15,000 to 110,000 DS. Further, it was> determined
that the State of Oklahoma has no exact information on the quantity or quality
of water injected or produced in petroleum operations involving primary,
secondary, and EOR. Related information for other states was not determined
(Collins and Wright, 1982).
An analysis was made of the approximate amount of water produced with
crude oil in 14 states. The states and their percent of total U.S. crude oil
production are: Alabama, 0.3%; Alaska, 19.9%; California, 11.7%; Colorado,
1.0%; Florida, 1.4%; Louisiana, 13.4%; Montana, 1.0%; Mississippi, 1.2%;
Nebraska, 0.2%' New Mexico, 2.3%; North Dakota, 1.4%; Texas, 31.2%; Utah, 0.8;
and Wyoming, 4.2%.
Figure 2 indicates the crude oil and water production from wells in the 14
states. The figure indicates that about 4.3 barrels of water is produced per
-94-
-------
barrel of oil. Figure 3 is a similar graph for 13 states excluding Alaska.
This figure indicates that about 5.2 barrels of water is produced per barrel
of oil. Further it can be shown that oil wells produce more water as
cumulative oil production increases. In other words, the older the well, the
higher the water-to-oil ratio.
Fresh Water
Fresh water primarily is water that can be made potable by flocculation,
filtration, and/or chlorination; contains less than 2,000 ppm dissolved solids
(DS); and can come from surface sources such as lakes, rivers, or underground
sources. In any EOR project, a first consideration must be given to the water
source. In some projects where a fresh water preflush is necessary, it is
obvious what the water source must be. Usually some sodium chloride is added
to the fresh water to inhibit clay swelling. Some EOR chemicals can tolerate
a more salty water. In such cases formation water, a mixture of formation
waters, a mixture of formation water and fresh water, or even seawater might
be feasible. When surfactants, polymers, and caustics are used with these
waters, precipitates caused by reactions with multivalent cations pose major
problems. The two most problematic cations are calcium and magnesium,
primarily because they are so highly concentrated in some waters.
The first step in determining the suitability of any water is to analyze
the water for physical properties and for chemical and biological
constituents. Next, the composition of the formation into which it is to be
injected should be determined. Clays such as smectites, kaolinites,
chlorites, and illites are sensitive to fresh water. Permeability reduction
may occur because of clay dispersion and clay swelling, Mangan (1965).
Increasing the salinity of the water usually minimizes the effect.
-95-
-------
Smectites and illites are the more common clays sensitive to fresh
water. They can absorb water on their edges and surfaces. Fresh water can
penetrate between the layers of a smectite to cause the plates to separate and
disperse. Therefore, formation damage caused by fresh water usually is most
severe in a formation that contains smectite.
Seawater
Several companies use seawater for water injection as a pressure
maintenance technique or for secondary recovery in some giant oil reservoirs,
Davis (1974); Mitchell (1978); and Carlberg (1979). It is injected into both
sandstone and carbonate reservoirs. Some of the negative aspects of seawater
injection are described by Ogletree and Overly (1978).
Eventually seawater will be used as an injection fluid in EOR technology,
Jerque (1984). The use of seawater presents the same problems associated with
any open system; that is, where air-water contact exists. Seawater presents
some additional problems; one of the most notable is the biomass; for example,
organisms such as copepods, diatoms, and dinoflagellates.
Mitchell and Finch (1978) outlined some of the necessary water quality
tests including: membrane filter test, examination of the filtered
particulates with light and scanning electron microscopy, on-site core
injectivity tests, particle size distribution in the injection water with
respect to the pore size distribution in the reservoir, amount and type of
biomass (other than bacteria) in the raw seawater, and bacterial levels
(aerobic and anaerobic). They found that cores are superficially plugged by
lipids derived from copepods plus inorganic debris. They also emphasized the
plugging of cores by bacterial debris, which was documented by Fekete (1959).
-96-
-------
Water Compatibility
Waters that can be mixed without the formation of precipitates are
considered to be compatible. Henkel (1953,1955) reported testing brine and
wastewater compatibility by allowing a mixture of the two liquids to stand
from 8 to 24 hours at the approximate aquifer temperature. The mixture is
considered compatible if it remains free of precipitates. Others have
suggested that this criterion may not always be entirely satisfactory, since
reactions may require considerable time for completion and because gaseous
reaction products may also cause reduction in permeability (White and Delany,
1982).
If the planned project is EOR using chemicals such as surfactants,
polymers, or caustics, the compatibility tests become even more complex. For
example, various studies indicate that sulfonates and polymers react with the
multivalent cations in formation water, Meister, et al. (1980). The tolerance
of petroleum sulfonates to the multivalent cations depends upon the average
equivalent weight (AEW) of the sulfonate. In general, the amount of cation
tolerated increases as the AEW of the sulfonate decreases.
Ostroff (1979) presents two methods of determining water compatibilities
and information on how to predict scale formation. Collins (1975) presents
some information on brine stabilization and methods for calculating over and
under saturation of some relatively insoluble compounds. A method approved by
the American Society for Testing and Materials (ASTM) Subcommittee D-19.09
appears in section 11.02 of the ASTM 1985 Annual Book of Standards.
Core Flow Tests
Core flow testing is the only good method of determining the effects of
the proposed injection fluid upon the permeability of the formation
reservoir. McCune (1977) describes some core test equipment for a flow
-97-
-------
test. An ASTM standard practice using core flow testing is in press and will
appear in the 1987 ASTM Annual Book of Standards on Water, Section 11, Volume
11.02.
Near-well filtration is the filtration of small particles on the face of
the formation from injected solutions which causes injection rates to lower.
Eventually, the permeability of the interior of the formation will decrease.
For example, it is not unusual for water injection rates to decline by 50% in
12 months. The only way to circumvent this is to inject water that contains
no suspended solids and is compatible with the formation water and formation
rocks, especially the clays. Workovers can improve the injection rates after
a decline but are expensive and time-consuming.
Corrosion
Ostroff (1979) lucidly defines corrosion and the forms of corrosion found
in oilfield operations. As he points out, electro-chemical corrosion of steel
is the usual type found in the oilfield. He further notes that "it is
necessary to have an (1) anode; (2) cathode; (3) electrolyte, and (4) external
connection. Remove any one of these and corrosion will cease." Obviously the
electrolyte is the water, and it is impossible to remove it in an oilfield
water system. Also it usually is impossible to remove the anode, cathode or
the external connection in most oilfield systems. Complete coating of the
steel lines and vessels or use of non-conducting lines and vessels (cathodic
protection) would solve the problem, but this is not yet feasible for all
systems.
The gases in some EOR injection waters, which are deleterious because of
potential corrosion problems, are 02, H2S, and C02. The presence of these
gases in salt water presents severe corrosion problems because salt water is
-98-
-------
an electrical conductor and is corrosive, and the corrosivity increases as the
water becomes saltier and as the concentration of 62, ^S, or CC^ increases.
These dissolved gases drastically increase the corrosiveness of salt
water. Fewer corrosion problems exist if they are removed and if the
injection water is maintained at a neutral or slightly higher pH; however,
because of the effect of high pH on clay swelling, a pH above 7 may be
undesirable.
Bacteria
Injection waters must be free of bacteria because they can cause corrosion
as well as plugging of the equipment and the face of the injection well.
Bacteria can reproduce rapidly, and they populate in extremely diverse
conditions such as low and high pH, temperature, pressure, and even in the
absence of oxygen. Patton (1975) and Collins and Wright (1982) describe tests
and problems bacteria cause in oilfield water injection operations.
FORMATION ROCK MINERALS
As noted by Collins and Kayser (1985), a small number of minerals comprise
the mass of most sandstone aquifers, and the average sandstone consists of
66.8% Si02 (mostly quartz), 11.5% feldspars, 11.1% carbonate minerals, 6.6
percent micas and clays, 1.8% iron oxides, and 2.2% other minerals. Limestone
and dolomite aquifers are primarily CaC03 and CaMg(C03)2, respectively, but
some contain 50% noncarbonate constituents such as Si02 and clay minerals.
Quartz, the main constituent of sandstones, is the least reactive of the
common minerals and generally can be considered nonreactive except in highly
alkaline solutions. Clays can react with highly basic or highly acidic
solutions; however, an injected fluid need not be highly acidic to attack
-99-
-------
certain clay minerals. The degree of reaction of feldspars and micas with
injected solutions is uncertain, but some reaction is likely to occur.
Sandstone aquifers often are cemented with carbonate minerals, which react
with acid solutions. Reaction of acid wastes with the carbonate cement in
sandstone causes an evolution of C02 that increases the pressure and reduces
the permeability. In the special case of acid aluminum nitrate wastes, it was
determined that the reaction of the waste with CaC03 creates a gelatinous
precipitate that plugs sandstone pores. Many sandstones are composed of
gypsum and limonite cementing material. These two minerals can dissolve,
reprecipitate, and block pores. Deep limestone, dolomite, or calcareous
sandstone aquifers usually contain brines which are in chemical equilibrium
with the aquifer, and dissolution and/or reprecipitation are not as likely to
occur.
If injected EOR fluids are at a lower pH than formation waters, solution
of the carbonate reservoir material can occur. This reaction is beneficial if
gelatinous precipitation does not occur. If alkaline injected fluids mix with
formation water and raise its pH, dissolved salts can precipitate and plug
pores.
Clay minerals are present in sedimentary rocks, and sandstones containing
less than 0.1% clay minerals probably do not exist anywhere except in small
deposits of almost pure glass sand. Clay minerals reduce the permeability of
sandstone to water versus its permeability to air, and the degree of
permeability reduction to water versus air is the water sensitivity of a
sandstone. Collins and Kayser (1985) address phenomena associated with
injection of oilfield waters into formation rocks; for example, anhydrite
versus gypsum, clay sensitivities, ion exchange, and adsorption.
-100-
-------
FLUID INJECTION TREATMENT SYSTEMS
Water and/or EOR fluid injection systems are divided into two categories;
(1) closed systems and (2) open systems. A closed system is designed to
exclude air or oxygen, whereas an open system makes no attempt to exclude
oxygen. Ostroff (1979) and Patton (1981) present detailed information
concerning injection water chemistry; chemicals used in scale and corrosion
prevention; chemicals used to control microorganisms; and methods used in
coagulation, sedimentation, filtration, degasification, etc. Modifications
and/or extensions of these methods are used in EOR injection fluid
pretreatment.
TYPES OF EOR OPERATIONS
MICELLAR-POLYMER
Figure 4 shows a single 5-spot injection-production pattern for a
micellar-polymer EOR operation. In this particular operation, a reservoir
preflush was first used to condition the reservoir followed by the micellar
fluid for releasing oil, polymer solution for mobility control, a fresh water
buffer to protect the polymer, and the final drive injection water.
Surfactant-polymer floods are chemical EOR processes. Surfactants are
micellar or surface-active agents including soaps and soap-like substances.
To be useful in enhanced oil recovery, they must reduce the interfacial
tension between water and oil. They have an amphiphilic molecule that is
attracted, at one end, to water (the hydrophilic or water-loving end), and the
other end is attracted to oil (the oleophilic or oil-loving end).
Alcohol improves the quality of some micellar solutions and, when used, is
a cosurfactant. The cosurfactant also aids the micelle in solubilizing oil or
water, stabilizes the solution, and reduces adsorption.
-101-
-------
The water-soluble polymers used in EOR consist of chain-like molecules
with molecular weights up to or exceeding 20 million. Polymers such as
polyacrylamides and polysaccharides often are used as mobility-control buffers
for permeability reduction and/or increased viscosity. Polysaccharides
sometimes are called biopolymers. Polymers increase the viscosity of the
waterflood and prevent it from running ahead of the oil. Increased resistance
to flow, particularly in high permeability zones, improves the volumetric
reservoir sweep efficiency resulting in increased oil recovery.
Water-soluble synthetic polyacrylamides consist of high-molecular-weight,
chain-like molecules with CONH2, COOH, and COONa groups attached to every
other carbon atom on a carbon chain. Naturally occurring polysaccharides
consist of cyclic carbohydrate monomers alternating in the polymer
structure. These additives aid oil recovery by decreasing the floodwater's
mobility. The polyacrylamides, for example, are most susceptible to breakdown
because of mechanical shear degradation and are more likely to adsorb on clay
or silicate surfaces than the polysaccharides. However, the fact that the
polysaccharides react with low concentrations of polyvalent cations, react
with bacteria, and in general plug filters or well sand faces because of
numerous reactions gives polyacrylamides a wider acceptance in oil recovery
operations.
In many of the surfactant-polymer EOR operations, a preflush is used.
This preflush often consists of fresh water to which sodium chloride is
added. More specifically, it probably will consist of fresh water, plus
sodium chloride, plus a bactericide, plus a corrosion inhibitor. A preflush
may continue for a year or until 80% of the rock pore volume (PV) is
flushed. The purpose of the preflush is to remove the connate brine from the
area of the reservoir where the operator wants to form an oil bank. After
-102-
-------
completion of the preflush, the sulfonate solution is injected. The preflush
theoretically removes most of the divalent ion cations (calcium and magnesium)
that were in the connate brine. These divalent cations react with many
sulfonates causing them to precipitate or become inactive or useless in the
entrainment or entrapment of the oil phase.
Other constituents in this surfactant or micelle phase may be sodium
hydroxide, sodium chloride, polymer, crude oil, and, of course, fresh water.
The polymer is added to increase the viscosity of the solution. Sodium
hydroxide, if used, may aid in forming a multiphase microemulsion system. The
microemulsion has at least three components: oil, water, and surfactant,
Collins and Kayser (1985).
Much of the preliminary work on an EOR operation is conducted to determine
possible interactions and compatibilities of injected fluids with the
indigenous reservoir fluids and rocks. This work is performed to minimize
losses of the injected solutions because of incompatible reactions with the
reservoir fluids and rocks and to ensure maximum oil recovery per dollar value
of injected chemical.
POLYMER
Figure 5 illustrates a single 5-spot injection-production pattern for a
polymer EOR operation. As shown, a preflush was performed to condition the
reservoir. This was followed by an injection of polymer solution primarily
for improved mobility control and an improved volumetric sweeping of oil
through the reservoir. Next, a fresh water buffer was injected to protect the
polymer followed by injected drive water.
A polymer operation is similar to a surfactant-polymer operation. The
notable exception is that the surfactant phase is not injected. The polymer
phase only is used; therefore, it might be called a thickened or polymer-
-103-
-------
augmented waterflood. The polymer increases the mobility ratio of the flood
and tends to move more oil without allowing the flood to finger through the
oil.
A preflush usually is used. Fresh water is used in many of the preflushes
— in the polymer phase and in the first drive water phase. Brine-tolerant
polymers will decrease the necessity of using fresh water. Many polymers
react with divalent cations such as calcium and magnesium.
ALKALINE
Figure 6 illustrates a single 5-spot injection-production pattern for an
alkaline EOR operation. As shown, a preflush of the reservoir is used to
condition the reservoir followed by an injection of an alkaline or
alkaline/polymer solution to form surfactants in situ to release oil from the
reservoir rock. Next, a solution of polymer is injected for mobility
control. Then injection of fresh water buffer to protect the polymer is
followed by injection of the driving fluid (water).
In general, an alkaline (caustic) flood is performed only in a sandstone
reservoir because of the abundance of calcium in a carbonate reservoir
brine. The most common chemical used in caustic flooding is sodium
hydroxide. Sodium orthosilicate and sodium carbonate are also used. Other
chemicals that have been used include ammonium hydroxide, potassium hydroxide,
sodium silicate, trisodium phosphate, and polyethylenimine. Since cost is
important, sodium hydroxide is more likely to be used than potassium
hydroxide.
Divalent cations such as calcium and magnesium in the connate water can
deplete a caustic slug by precipitation of hydroxides. Also, if anhydrite or
gypsum are in the rock, calcium will react with the slug to precipitate
calcium hydroxide. High ion-exchange-capacity clays will exchange hydrogen
-104-
-------
for sodium rendering the caustic slug ineffective by producing water and tying
up the sodium. Caustic usually reacts with the silica in sandstone too slowly
to cause problems. Most dolomites and limestones will not react with the
caustic to cause deleterious effects.
Krumrine, et al. (1982) reported on the effects that alkaline additives
have on dilute surfactant systems for low-tension waterflooding and how
interfacial tension, hardness removal, and surfactant retention affect oil
recovery in high-hardness core systems. They also examined the effects of
alkaline additives on dilute surfactant systems for improved oil recovery.
CARBON DIOXIDE
Figure 7 illustrates the carbon dioxide oil flooding process, a miscible
displacement process applicable to many reservoirs. A slug or a prescribed
amount of carbon dioxide is injected into the reservoir followed by an
injection of water and a subsequent injection of carbon dioxide.
Most C02 floods uses a water-injection phase as a preflush and as a water-
alternating-gas injection (WAG). For example, the preflush may be a fresh
water to which salt is added or it may be a softened salt water. In some
areas softened seawater is used.
At least four methods of carbon dioxide and water injection have been
studied or used: (1) continuous injection of carbon dioxide for the life of
the flood, (2) injection of carbon dioxide followed by water, (3) injection of
alternate slugs of carbon dioxide and water, and (4) simultaneous injection of
carbon dioxide and water. The water in some field applications consists of
polymer-thickened water. Carbon dioxide floods are useful in both carbonate
and sandstone reservoirs.
The depth of the reservoir should be 2,500 ft or more. If it is not, the
overlying rock may be fractured. If the pressure in the reservoir containing
-105-
-------
an oil of 30° API gravity or greater has been depleted to less than 1,200 psi,
the pressure must be built up by injection of water before the C02 injection
begins. Of course, C02 could be injected to build up the pressure— but this
would be very expensive at the current prices for C02«
STEAM
Figure 8 illustrates a steamflcoding operation. Heat from steam injected
into a heavy-oil reservoir thins the oil making it easier to push through the
formation toward production wells. Steam and hot water flooding account for
most of the oil recovered by all EOR operations. There are two steam recovery
processes: (1) steam stimulation, sometimes called cyclic steam injection,
steam soak, or huff and puff and (2) steamflooding which is a process similar
to waterflcoding. Water used in a steamflood usually is a high quality water
and usually is softened before it goes into the steam generator to prevent
scale problems in the boiler. Steamflooding accounts for the most oil
recovered by any EOR technology.
IN SITU COMBUSTION
Figure 9 illustrates an in situ combustion operation where heat is used to
thin the oil and thereby permit it to flow to the production well. In this
operation, the oil in the formation is ignited, and by continued injection of
air the fireflood front advances through the reservoir.
There are two fundamental processes of in situ combustion -- forward
combustion and reverse combustion. Water is used in variations of the forward
combustion process. When water is injected with air, it forms superheated
steam near the injection well. At the combustion front, it mixes with
nitrogen, carbon monoxide, carbon dioxide, and other gases. This hot gas
mixture displaces the oil. Heat reduces the viscosity of the oil allowing the
-106-
-------
oil to flow toward the production well. The benefit of the wet method is that
it allows a threefold reduction in air to produce a barrel of oil.
MISCIBLE HYDROCARBON
As the name of the flood implies, the injected gas or liquid hydrocarbon
becomes miscible with the hydrocarbons in the reservoir. This miscibility
usually is accomplished at elevated temperatures and pressures; therefore,
depth of the reservoir is important because of the need to maintain a high
pressure.
Three different techniques are commonly used: (1) miscible slug process,
whereby a slug of liquid hydrocarbon about 0.05 PV is injected followed by gas
and water as the drivers; (2) enriched gas process, whereby a slug of enriched
gas is injected followed by lean gas and water as the driver; and (3) high-
pressure, lean-gas process, whereby lean gas is injected at high pressure to
cause evaporation of the crude oil and formation of a miscible phase.
INERT GAS INJECTION
Increased costs of natural gas and carbon dioxide have prompted operators
to look at other methods to maintain the pressure in petroleum reservoirs.
With natural gas, miscibility could be achieved in some reservoirs. The
miscibility state allows almost 100% displacement efficiency in the swept
zone; however, this is not always the goal. Often pressure maintenance is the
goal.
Figure 10 illustrates the use of nitrogen in a carbon dioxide flood
operation where the nitrogen is used for economic reasons. Inert gases such
as nitrogen are not miscible with many oils at low pressures. Also, the API
gravity of the oil should be 35° or higher for application of this process.
-107-
-------
MICROBIAL FLOODING
Microbial flooding is performed by injecting a solution of microorganisms
and a nutrient such as industrial molasses down injection wells drilled into
an oil-bearing reservoir. As the microorganisms feed on the nutrient, they
metabolically produce products ranging from acids and surfactants to certain
gases such as hydrogen and carbon dixoide. These products act upon the oil in
place in a variety of ways, making it easier to move the oil through the
reservoir to production wells.
The microbial and nutrient solution and the resulting bank of oil and
products are moved through the reservoir by means of drive water injected
behind them, as illustrated in Figure 11.
CYCLIC MICROBIAL RECOVERY
This well-stimulation method is one of the newest EOR methods and requires
the injection of a solution of microorganisms and nutrients down a well into
an oil reservoir. This injection can usually be performed in a matter of
hours, depending on the depth and permeability of the oil-bearing formation.
Once injection is accomplished, the injection well is shut in for days to
weeks. During this time, known as an incubation or soak period, the
microorganisms feed on the nutrients provided and multiply in number. These
microorganisms produce products metabolically that affect the oil in place in
ways that make it easier to produce. Depending on the microorganisms used,
these products may be acids, surfactants, and certain gases, most notably
hydrogen and carbon dioxide.
At the end of this period, the well is opened, and the oil and products
resulting from this process are produced.
This method eliminates the need for continual injection, but after the
-108-
-------
production phase is completed a new supply of microorganisms and nutrients
must be injected if the process is to be repeated. Figure 12 illustrates the
process.
QUANTITY OF CHEMICALS USED IN EOR
In general, a micellar injection is in the range of 5-20% pore volume
(PV), with 5-20% of the injection slug containing sulfonate and 1-20% of the
injection containing alcohol. Polymer injections vary greatly, ranging in the
area of 25-75% PV, depending on polymer concentration, and tapered to lower
concentrations as injection progresses. For better economics, efforts are
being made to lower the amounts of chemicals used, and, in fact, no new
micellar-polymer field operations were started in the past 2 years.
Alkaline operations inject 15-40% PV caustic slugs composed of less than
2% caustic compounds, such as NaOH and sodium orthosilicate. Biocides are
added to the surfactant slug if it is biochemically unstable; however, they
normally are injected with the polymer injection. Concentrations used are in
the order of 10-150 ppm, and the volume injected is less than 1% of the
injection.
Silvestro and Desmarais (1980) divided EOR chemicals into five functional
groups as shown in table 3 and below:
1. Mobility Control Agents (Polymers)
In general, these are considered to be low in toxicity; many of them are
used in small amounts as food and drug additives or constituents of food
packaging. The main hazards from polymers are associated with on-site
handling and dispensing.
Degradation products of polyacrylamide and polysaccharide polymers are
generally smaller fragments of the respective polymer. In the cases of
-109-
-------
polysaccharides, the degradation products are ultimately the monomers often
used in synthesizing the polymer. Polysaccharides are hydrolyzed at the ring-
ester linkages to form simpler sugars, while polyacrylamides tend to be
hydrolyzed at the amide linkage and form a low-viscosity polymer with reduced
mobility control properties. The alkylcellulose ethers degrade to simpler
starches, sometimes hydrolyzing at available linkages under higher-pH
conditions. It is unlikely that toxic hazards should be expected from any of
these degradation products.
2. Cosurfactants
Cosurfactants are generally used in relatively small amounts. They are
composed primarily of longer chain aliphatic alcohols whose hazards have been
well documented through industrial usage and are not expected to cause
environmental problems in EOR projects.
3. Alkaline Flooding Agents, Preflush Agents, Thermal Enhancers
Some of the compounds in this group are quite caustic and require
conscientious handling (sodium hydroxide, sodium orthosilicate); others are
organics with relatively high toxicity levels or carcinogenic potential
(hydrazine, quinoline). The sodium compounds are generally considered safe in
the diluted amounts used in EOR; little is known about the safety of on-site
disposition of the organic compounds used.
4. Surfactants
Recent standards established within the United States consider up to 0.5
mg surfactant per liter of water as being safe for human purposes. Although
alkylaryl and petroleum sulfonates are minor irritants to eyes and skin,
systemic chronic effects and toxicological data are not generally known. The
high toxicity of sulfonates to aquatic life may be an indicator of toxic
potential. Incomplete degradation of alkylbenzene sulfonates does occur in
-HO-
-------
the environment, possibly introducing free benzene rings into the formation or
a surface disposal site.
5. Biocldes, Chelatlng Agents, Oxygen Scavengers
The biocides are moderate to severe irritants, particularly to eyes, skin,
and upon inhalation. Certain ones, such as acrolein, glutaraldehyde,
formaldehyde, pentachlorophenol (PCP), and 2,4,5-trichlorophenol, are
extremely toxic over short exposure periods. Bioaccumulation is high, and all
five are implicated as carcinogens. Pentachlorophenol and 2,4,5-
trichlorophenol contain contaminants (dioxin, chloroquinone,
tetrachlorobenzene) which may be more toxic than the pure compound.
TRANSPORT AND FATE
Physical, chemical, and microbiological processes affect the transport and
fate of fluids injected into subsurface reservoirs. Geohydrology provides a
quantitative understanding of the flow of fluids through the subsurface, and
as a discipline it includes the mathematical, chemical, geological, and
physical sciences. Although many methods are available to aid in solving
mathematical problems associated with flow, transport, and fate of injectants
into subsurface reservoirs, many of the problems require further study, and
new methods need to be developed and tested.
CONCLUSIONS
Water is important in petroleum recovery operations. Adequate
considerations should be given to the type, quality, and quantity of water
available. Necessary tests should be made to ensure that the water used is
compatible with the recovery technology planned and the reservoir rock and
associated indigenous fluids. After the recovery operation is begun,
necessary tests should be conducted on a routine basis to ensure that the
system is maintained at optimum conditions.
-ill-
-------
Knowledge of abiotic and biodegradation transformations and mobility
pathways in soils, surface waters, and groundwaters for many chemicals used in
petroleum recovery is nonexistent. Better information concerning abiotic and
biodegradation transformations, transport, and ultimate fate of EOR chemicals
and their by-products in soils and waters should be obtained for (1) mobility
control agents, (2) cosurfactants, (3) surfactants, (4) alkaline flooding
agents, (5) preflush agents, (6) thermal enhancers, (7) biocides, (8) chelat-
ing agents, (9) oxygen scavengers, (10) solid wastes from steamfloods, and
(11) potentially dangerous chemicals used in any EOR operation.
-112-
-------
REFERENCES
American Society for Testing and Materials (ASTM), 1985, Philadelphia, PA.
Standard Practice for Calculation of Supersaturation of Barium Sulfate,
Strontium Sulfate. Dihydrate (Gypsum) in Brackish Water, Sea Water, and
Brines, Section 11, volume 11.02, pp. 551-556.
Carl berg, B. L. 1979, How to Treat Seawater for Injection Projects.
World Oil, v. 189, No. 1, pp. 78-81.
Collins, A. 6. 1975, Geochemistry of Oilfield Waters. Elsevier
Scientific Publishing Co. New York, 496 pp..
Collins, A. G. and M. B. Kayser. 1985, Interaction, Compatibilities, and
Long-Term Environmental Fate of Deep-Well-Injected EOR Fluids and/or Waste
Fluids with Reservoir Fluids and Rocks - State-of-the-Art, Oept. of Energy
Report No. NIPER-70, NTIS Order No. DE85000146, 103 p.
Collins, A. G. and C. C. Wright. 1982, Enhanced Oil Recovery Injection
Waters. Dept. of Energy Report. No. DOE/BETC/RI-82/5, Apr., 82 pp.
Craig, F. F., Jr. 1971, The Reservoir Engineering Aspects of
Waterflcoding, Society of Petroleum Engineers, Morgraphy Series 3: 134 pp.
Davis, J. 1974, Big Waterflood Begins Off Abu Dhabi. Oil and Gas
Journal, v. 73, No. 33, pp. 49-51.
Fekete, T. 1959, The Plugging Effect of Bacteria in Sandstone Systems.
M.S. Thesis, University of Alberta Canada, 1959.
Goodlett, G. 0., M. M. Honarpour, H. B. Carroll, P. S. Sarathi. 1986,
Screening for EOR - 4 Parts, Oil and Gas Journal, June 23, 1986 ending July
28, 1986.
Henkel, H. 0. 1953, Surface and Underground Disposal of Chemical Wastes
at Victoria, Texas. Sewage and Industrial Wastes. Chemical Engineering
Progress, v. 25, No. 9., pp. 1044-1049.
-113-
-------
Henkel, H. 0. 1955, Deep-Well Disposal of Chemical Wastes. Chemical
Engineering Progress, v. 51, No. 12, pp. 551-554.
Jorque, M. A. 1984, How to Treat Seawater for Water Injection, Petroleum
Engineer, Nov. 28-34.
Krumrine, P. H., J. Falcone, and T. Campbell. 1982, Surfactant Flooding
2: The Effect of Alkaline Additives on Permeability and Sweep Efficiency.
Society of Petroleum Engineers Journal, v. 22, No. 6, pp. 983-992.
Langnes, G. L. Robertson, J. 0. Jr., Mehdizadeh, A., Torabzadeh, J., Yen,
T. F., Donaldson, E. C., and Chilingarian, G. V. 1985, Waterflcoding, Ch. 8
in Enhanced Oil Recovery, 1. Fundamentals and Analysis, Elsevier, p. 251-334.
McCune, C. C. 1977, On-Site Technology to Define Injection Water Quality
Requirements. Journal of Petroleum Technology, v. 1, pp. 17-24.
Meister, M. J., C. A. Wilson, and A. G. Collins 1980, Tolerance of
Petroleum Sulfonates to the Presence of Calcium Ions, Chapter in Solution
Chemistry of Surfactants, Plenum Press, pp 927-940.
Mitchell, R. W. 1978, The Forties Field Sea-Water Injection System.
Journal of Petroleum Technology, v. 30, pp. 877-884.
Mitchell, R. W. and T. M. Finch 1978, Water Quality Aspects of North Sea
Injection Water, Society of Petroleum Engineers, (UK) LTD Europe Offshore
Petroleum Conference, Proceedings, v. 1, pp. 263-276.
Mungan, N. 1965, Permeability Reduction Through Changes in pH and
Salinity. Journal of Petroleum Technology, v. 12, pp. 1449.
Ogletree, J. 0. and R. J. Overly 1973, Sea-Water and Subsurface Water
Injection in West Block 73 Waterflood Operation. Journal of Petroleum
Technology, v. 25, pp. 623-628.
Ostroff, A. G. 1979, Introduction to Oilfield Water Technology, National
Association of Corrosion Engineers, 394 p.
-114-
-------
Patton, C. C. 1975, Oilfield Water Systems. Campbell Petroleum Series,
Norman, OK, 65 p.
White, A. F. and J. M. Delany 1982, Investigation of Surface Interactions
Between Silicate Rocks, Minerals, and Groundwater. Annual Report, Earth
Sciences Division, Lawrence Berkeley Laboratory, LBL-15500, pp. 112-115.
-115-
-------
TABLE 1. - Geochemical water analyses
Property
or
Constituent
PH
Eh
Specific resistivity
Specific gravity
Bacteria
Barium
Bicarbonate
Boron
Bromide
Calcium
Carbonate
Carbon dioxide
Chloride
Hydrogen sulfide
Iodide
Iron
Magnesium
Manganese
Oxygen
Potassium
Residual hydrocarbons
Sodium
Silica
Strontium
Sulfate
Suspended solids
Total dissolved solids
Produced
Water
X
0
X
X
0
X
X
0
0
X
X
0
X
0
0
X
X
0
0
0
X
0
0
X
X
Injection
Water
X
X
X
X
X
X
X
X
X
X
X
X
X
0
X
X
0
X
X
X
X
X
Steam
Generation
Water
X
X
X
X
X
X
X
X
X
0
0
0
X
0
X
X
Disposal
Water
X
0
X
0
X
X
X
X
0
X
0
0
X
0
0
0
0
0
0
X
X
X
X usually requested
o sometimes requested
-116-
-------
TABLE 2. - Tertiary system - highest concentration of a
constituent found, average concentration, and
number of samples analyzed - Collins (1975)
Constituent
Lithium
Sodium
Potassium
Rubidium
Cesium
Calcium
Magnesium
Strontium
Barium
Boron
Copper
Chloride
Bromide
Iodide
Bicarbonate
Carbonate
Sulfate
Organic Acid as acetic
Ammonium
Concentration
highest
27
103,000
1,200
0.6
0.4
38,800
5,800
420
240
450
1
201,300
1,300
35
3,600
300
8,400
1,900
2,700
(mg/1)
average
4
39,000
220
0.24
0.20
2,530
530
130
60
36
0.63
64,600
85
28
560
75
320
140
230
Number of samples
169
379
176
11
9
376
368
142
140
170
3
380
323
322
364
8
139
53
64
-117-
-------
TABLE 3. - Toxicological Data
Survey Chemicals Arranged by General Use In EOR
(Reference Silvestro and Desmarais, 1980)
Group I Mobility Control Agents
Polyacrylamides
Xanthan gums
Carboxymethylcellulose
Hydroxyethy1 eel 1ulose
Polyethylene glycol monobutyl ether
Polyethylene oxide
Group II Cosurfactants
1-hexanol 2-hexanol
1-octanol
2-octanol
n-butanol (and tert-, sec-, iso-isomers)
Cyclohexanol
Polyethoxyalkylphenol
Group III Blocides, Chelating AGents, Oxygen Scavengers
Quaternary ammonium chloride
2,4,5-trichorophenol
Pentachlorophenol
Phenol
2,2-dibromo-3-nitrilopropionamide
Copper sulfate
Glutaraldehyde
Formaldehyde
Sodium hypochlorite
Acrolein
EDTA
1,6-hexanediamine
Group IV Surfactants
Alky aryl sulfonates
e.g., Alkyl benzene sulfonate
Octadecyltoluene sulfonate
Tridecyl benzyl sulfonate
Decyl benzyl sulfonate
Alkyl naphthenic sulfonates
Petroleum sulfonates (toxicity as groups)
-118-
-------
Group V Alkaline Flooding Agents. Preflush Agents. Thermal Enhancers
Sodium nitrate
Sodium hydroxide
Sodium orthosilicate
Sodium carbonate
Sodium borate
Sodium hydrosulflte
Sodium bisulfite
Sodium sulfate
Hydrazlne
Qu1nol1ne
Toluene
Xyl1d1ne
Aniline* 2,2-d1bromo-3-nitr1loprop1onam1de
Copper sulfate
Glutaraldehyde
Formaldehyde
Sodium hypochlorite
Acrolein
EDTA
1,6-hexanedlamine
-119-
-------
FIGURE 1
K!
O
I
OIL PRODUCTION
Improved technology through research is
enhancing oil recovery.
PRIMARY RECOVERY
Produces 12-15% of the
original oil-in-place*
SECONDARY RECOVERY
Another 15-20% of the
original oil-in-place* may be
produced by waterflooding
ENHANCED OIL RECOVERY (EOR)
An additional 4-11% of the original oil-in-place* may be
produced using current and advanced technology
ADVANCED PROCESSES
• Improved Mobility
Control
• Deep Steam
• Microbial
y Mining
y^&ZKvmK*!
Approximately 65% (300 billion bbls)
of original oil in place* still locked
in earth after secondary recovery
Approximately 460 billion bbl of
oil estimated to be in place
before any production
-------
31,000
30,000 -
7,000 r
6,000
1975 76 77 78 79 80
YEAR
FIGURE 2. - Crude oil and water produced (x 1,000 barrels
per day) from wells in 14 states including
Alaska (Collins and Wright, 1982).
-121-
-------
3 1,000
30,000
29,000 -
Q 28,000 -
CD
CD
; 27,000 -
Q
O
OC.
26,000 -
25,000 -
7,000 -
6,000 -
5,000
- (EXCLUDING ALASKA)
FIGURE 3. - Crude oil and water produced (X 1,000 barrels
per day) from wells in 13 states excluding
Alaska (Collins and Wright, 1982).
-122-
-------
FIGURE 4
CHEMICAL FLOODING
(Micellar-Polymer)
The method shown requires a preflush to condition the reservoir, the injection of a micellar
fluid for releasing oil, followed by a polymer solution for mobility control to minimize channeling,
and a driving fluid (water) to move the chemicals and resulting oil bank to production wells.
(Single 5-Spot Pattern Shown)
ro
t_o
I
Additional
Oil
Recovery
(Oil Bank)
Preflush
to Condition
Reservoir
-------
FIGURE 5
CHEMICAL FLOODING
(Polymer)
The method shown requires a preflush to condition the reservoir, the injection of a polymer
solution for mobility control to minimize channeling, and a driving fluid (water) to move
the polymer solution and resulting oil bank to production wells.
Mobility ratio is improved and flow through more permeable
channels is reduced, resulting in increased volumetric sweep.
(Single 5-Spot Pattern Shown)
Production Well
Polymer
Solution
For
Mobility
Control
Fresh
Water
Buffer
to Protect
Polymer
Additional
Oil
Recovery
(Oil Bank)
Preflush
to Condition
Reservoir
Driving
Fluid
(Water)
-------
FIGURE 6
CHEMICAL FLOODING
(Alkaline)
The method shown requires a preflush to condition the reservoir and injection of an alkaline
or alkaline/polymer solution that forms surfactants in situ for releasing oil. This is followed
by a polymer solution for mobility control and a driving fluid (water) to move the chemicals
and resulting oil bank to production wells.
Mobility ratio is improved, and the flow of iiquids through
more permeable channels is reduced by the polymer
solution resulting in increased volumetric sweeo.
(Single 5-Spot Pattern Shown)
Production Well
Alkaline
Solution
Forms
Surfactants
In Situ For
Releasing
Oil
Additional
Oil
Recovery
(Oil Bank)
Preflush
to Condition
Reservior
-------
FIGURE 7
CARBON DIOXIDE FLOODING
This method is a miscible displacement process applicable to many reservoirs. A CO2 slug followed
by alternate water and CO2 injections (WAG) is usually the most feasible method.
Viscosity of oil is reduced providing more efficient miscible displacement.
I
I—»
hO
I
Produced Fluids (Oil. Gas and Water)
Separation and Storage Facilities
-------
I
t—»
N5
I
FIGURE 8
STEAM FLOODING
Heat, from steam injected into a heavy-oil reservoir, thins the oil making it easier
for the steam to push the oil through the formation toward production wells.
Heat reduces viscosity of oil and increases its mobility.
Production Fluids (Oil, Gas and Water)
Separation and Storage Facilities
^W&^W^-^c®
•^MMc^'^P^^^df
aaaiafcQ^y-»y-csfa»;Yvffq-B?-^
Oil and
Water Zone
Near Original
Reservoir
Temperature
Steam and
Condensed Water
-------
OO
I
FIGURE 9
IN-SITU COMBUSTION
Heat is used to thin the oil and permit it to flow more easily toward production wells. In a fireflood,
the formation is ignited, and by continued injection of air, a fire front is advanced through the reservoir.
Mobility of oil is increased by reduced viscosity caused by heat and solution of combustion gases.
J-^J
-r-1-
^
^F
L-r-L .-'I , 1 |
1 ' 1 '"I
T - 1
L-rr-^r^
1 ^~1 — '
E^3
r-^-T-i-
i '
1
i — '
1
r-1-
r-J-r-J
| ' 1 ' 1 ' 1 1 ' 1 ' T^1 1 ' 1
. 1. Injected Air and Water Zone (Burned Out)
2. Air and Vaporized Water Zone
3. Burning Front and Combustion Zone (600
^— 4. Steam or Vaporizing Zone (Approx. 400CF
I
i
1
,° .
:)
[ — — c~
i — i — i-
1200°F)
-i — '
5,
6
7
i ' i '", •' i H
r^-r-
L-i
L 1 1 , 1 1 1 1 1 1 ' 1 •
. Condensing or Hot Water Zone
(50° - 200°F Above Initial Temperat
. Oil Bank (Near Initial Temperature
. Cold Combustion Gases
1 l ' I i-T-^
i
ure) J
) -
~r — i — ' — j — "~
L-r-1 ! ' 1 ' '
L^~l l ' ! '
H^ . ' . '
[ • -r 1 , 1 ,1-
-------
FIGURE 10
NITROGEN — CO2 FLOODING
In a CO2 flood, the use of nitrogen to displace the C02 slug and its miscible oil bank
might be desirable due to the lower cost of the nitrogen.
Viscosity of oil is reduced providing more
efficient miscible displacement.
Produced Fluids (Oil, Gas and Water)
Separation and Storage Facilities
' \ 1 1 1 1 * 1 ^
-------
u>
o
I
FIGURE 11
MICROBIAL FLOODING
Recovery by this method utilizes the effect of microbial solutions on a reservoir. The reservoir is usually
conditioned by a water preflush, then a solution of microorganisms and nutrients is injected. As this
solution is pushed through the reservoir by drive water, it forms gases and surfactants that help to
mobilize the oil. The resulting oil and product solution is then pumped out through production wells.
(Single 5-Spot Pattern Shown)
Microorganisms
and Nutrients
-------
FIGURE 12
CYCLIC MICROBIAL RECOVERY
A solution of microorganisms and nutrients is introduced into an oil reservoir during injection.
The injection well is then shut in for an incubation period allowing the microorganisms to produce
carbon dioxide gas and surfactants that help to mobilize the oil. The well is then opened and oil
and products resulting from the treatment are produced. This process may be repeated.
Schematic portrays one well during the 3 phases of this
process. Flow pattern is stylized for clarity.
INCUBATION (Shut-in Phase) PRODUCTION
Days to Weeks Weeks to Months
INJECTION
Hours
LO
i—'
I
Microorganisms
and Nutrients
K
Produced Oil\
and Products/
^ I I I
Depleted
Oil Sand
Injected Microorganisms
and Nutrients
Metabolic Activity Produces
CO2 and Surfactants
-------
PAPER UNAVAILABLE. WILL BE PRINTED IN UIPC JOURNAL.
ABSTRACT
Oilfield Brine Disposal into the Wilcox Aquifers
in S.E. Mississippi - A Case History
Author
Lee Thomas
U.S. EPA Region IV, Atlanta, Georgia
Since the 1940's oilfield brines have been disposed of by
injection into the Wilcox Aquifers in Clarke, Jones, Jasper,
Wayne and Smith counties in Mississippi. When the
regulations for the Safe Drinking Water Act were promulgated
it was required that any aquifer with less than 10,000 m/g
per litre total dissolved solids be protected. Regional
studies for these counties subsequent to the promulgation of
the UIC regulations indicated that the Wilcox Aquifers in
these counties contained less than 10,000 m/g per litre total
dissolved solids. In order to insure that injection did not
endanger any protected aquifer, the Environmental Protection
Agency requested that all owners of Wilcox disposal wells in
this area submit permit applications. To evaluate these
permit applications it was necessary to give careful
consideration to the hydrogeology in this area. Each permit;
application was evaluated with respect to the injection
aquifer using geophysical logs since water samples showing
ambiant conditions were generally not available. In many
permit applications the actual injection sand was shown to be
less than 10,000 m/g per litre total dissolved solids. In
other permit applications a Wilcox aquifer contained
protected waters in its upper section and injection was into
a lower sand with greater than 10,000 m/g per litre total
dissolved solids water. The issue in these permit
applications became whether adequate confining layers existed
within a specific Wilcox aquifer. In order to provide
confinement a zone must extend continuously ton enough in all
directions so that it is beyond the zone of endangering
influence, it must be between the injection zone and the
lowest protected water, it must have sufficiently low
hydraulic conductivity to preclude migration of injection
fluid or formation fluid into protected sands. Determining
whether Wilcox injection might cause endangerment of
protected water required an understanding of the complex
ge;ology and hydrogeology of the Wilcox aquifers in this area.
-132-
-------
BIOGRAPHICAL SKETCH
Lee Thomas
Lee Thomas has 7 years professional experience as a
Geologist. He is presently with the Underground Injection
Control Section, Ground Water Protection Branch, U.S. EPA
tiegion IV in Atlanta, Georgia. He has a Bachelor of Arts in
Geology from the University of Tennessee at Chattanooga. A
Master of Science in Geology from Memphis State University.
He is presently attending Georgia State University doing
graduate study in Hydrogeology-
-133-
-------
MECHANICAL CONSIDERATIONS OF THE DISPOSAL OF FLUIDS INTO POORLY
CONSOLIDATED SANDSTONE RESERVOIRS
by J.G. Roberts and R.F. Stiles, Completion Services, Inc
ABSTRACT
This paper discusses the mechanical aspects of disposing fluids into a
poorly consolidated sandstone formation. In certain areas of the U.S.
unconsolidated or poorly consolidated formations are prevalent and the
injection of fluids into these strata require specific mechanical
considerations in the injection well design. Among the most of
important of these considerations are the perforating program, gravel
pack design and maintenance of the wellbore environment. Each of these
items are discussed in detail and a recommended procedure is
presented.
In order to see the impact of these items on the pressure/rate
relationship, a theoretical model is used to calculate the effect of
changing these design considerations. This work shows the importance
of proper well design in minimizing the pressure drop across the final
wellbore completion.
INTRODUCTION
The disposal of fluids into an underground strata has been used
extensively in the waste industry for several years. The technology
for these applications originally came from the oil and gas industry
however have been modified to meet the specific requirements of
disposal projects. The systems and procedures necessary to implement
these projects are complex and can involve a variety of different
disciplines.
-134-
-------
In certain areas of the country the"construction of these injection
systems are further complicated by the type of strata underlying the
region, in these areas, the formations that are available for waste
injection consist of unconsolidated or poorly consolidated sandstone
reservoirs. Injecting fluids into these formations theoretically does
not present a problem because of the direction of flow. In practice
however the actual operation of the entire injection system creates
situations in which the unconsolidated formation sand may fall into
the wellbore area. A pressure surge caused by an emergency shut-down
system, fluctuation of injection rates and pressures, operator error
and so forth all may allow the introduction of formation sand into the
wellbore. This formation sand can lead to an increase in injection
pressures and ultimately total failure of the injection well.
l
As a method of controlling this problem a variety of different sand
control techniques have been tried. The sand control system that has
produced the best results in both injection well technology as well as
in producing wells is gravel packing. Gravel packing involves the
placement of a wire-wrapped screen or slotted liner across from the
injection interval and packing the screen-casing annulus with high
quality gravel pack sand. This technique, while developed for use in
the oil and gas industry to prevent the production of formation sand,
provides many of the same benefits for injection well applications.
The basic difference between an injection well gravel pack and a
producing well gravel pack is the final "direction" of production. The
desired results for both are the same:
1. High volumes of production or injection
2. Low pressure drop across the completion
Gravel packing has been proven to be a viable method of controlling
the movement of formation sand from the reservoir into the wellbore
while minimizing the pressure drop across the completion. This process
has steadily improved over the past 50 years to the point that in
certain areas, (Gulf Coast, Florida, California) a majority of the
wells are completed initially using gravel packing as a sand control
measure.
-135-
-------
The initial success and overall life of the gravel pack is greatly
effected by several factors. These factors include wellbore
maintenance, perforating, and completion design. This paper will
address the design considerations in relation to these factors.
THEORY
Sand problems are most common in younger Tertiary sediments,
particulary of the Miocene epoch. Notable examples are the extensive,
troublesome sand production areas in such sediments in the U.S. Gulf
Coast, the Los Angeles basin of California, and the Florida panhandle.
In other areas, however sand or rock failure can occur in other
formations when local earth stress states and rock strength are
affected by certain completion practices and production operations
that create an unstable condition.
Gravel pack technology has evolved over the past 40-50 years and
developed into a highly specialized service. A gravel pack completion
consists of packing high quality gravel pack sand around a screen or
slotted liner. In addition to packing this screen-casing annulus,
gravel pack sand is also pumped into the perforation tunnels. The
function of the gravel pack sand is to prevent the formation sand from
flowing into the wellbore while at the same time allowing the produced
fluids into the production screen.
The gravel pack sand in the annulus requires the produced or injected
fluid to flow through this sand pack and therefore will increase the
pressure drop across the completion. For this reason high quality sand
is used in gravel packing because its permeability is significantly
higher than the formation sand's permeability. For example, smaller
gravels such as 40-60 U.S. mesh have about 69 darcies of permeability
while others sands such as 20-40 have about 170 darcies. In
comparison, a good Gulf Coast formation sand will have only 500
milli-darcies or 1/2 darcy permeability.
The amount of pressure drop created by the produced or injected fluids
-136-
-------
flowing across the gravel pack sand can be calculated using Darcy's
law for fluid flow through a porous media.
kh (Pe - Pw)
Q - 7.082
(u ln(re/rw))
As can be seen in the above equation, the permeability of the porous
media has a large effect on the calculated pressure drop. In order to
keep the pressure drop at a minimum over the gravel packed section,
the permeability of the gravel pack sand must be kept at a maximum.
Any reduction of pack sand permeability will cause the pressure drop
at any specific flow rate to increase.
The gravel pack sand must therefore be sized in order to restrain or
bridge the formation sand while maintaining the highest permeability
possible. When the gravel pack sand has been correctly sized, the
formation sand will bridge exactly at the gravel pack sand, formation
sand interface. Since there is no migration of formation sand into
the gravel pack, maximum permeability will be retained. If any mixing
of the pack sand and formation sand occurs then the permeability of
the resultant mix is less than either the formation sand or the gravel
pack sand alone. This in turn will increase the pressure drop across
the completion because of the reasons described above.
To further investigate the pressure drop across a gravel packed
completion a model of the flow environment must be used. To describe
this flow, three separate regions are assumed: Flow through the sand
packed screen-casing annulus, flow through the sand packed perforation
tunnels and flow through the surrounding reservoir. Each of these
areas can be modeled using various flow equations. In general flow
through the screen-casing annulus and flow through the surrounding
formation is described by the radial form of Darcy's law. The flow
through the perforation tunnels, however is described by the linear
form of Darcy's law. This modification of Darcy's general equation
predicts the pressure drops through a porous media where flow is
confined to a uniform cross-sectional area.
When these equations are applied to the gravel pack model it can
-137-
-------
readily be seen that by far the largest pressure drop occurs in the
perforation tunnel. This indicates that the permeability of the sand
in the perforation tunnel is probably the most important aspect of the
gravel pack. This is why it is so important that gravel pack sand be
packed into both the screen-casing annulus as well as the perforation
tunnels themselves.
The importance of placing high quality gravel pack sand into the
perforations can be illustrated by calculating pressure drops due to
one darcy formation sand filling the perforations. One darcy of
permeability is about twice as much as normally occurs in Gulf Coast
formation sand. Assuming a flow rate of 1 BPD/perf, the pressure drop
across a 3/8" diameter perf would be 450 psi. Across a 1/2" diameter
perf the pressure drop would be 190 psi and across the larger 3/4"
perf the pressure drop would be 64 psi.
If the flow rate is increased, the pressure drop in the perforations
will become quite large. Increasing the rate to 10 BPD/perf, the
pressure drop across a 3/8" diameter perforation will be 27,760 psi.
There are no wells in the Gulf of Mexico which are capable of flowing
at this pressure drop. Even in a 1/2" diameter perforation the
pressure drop will be over 9000 psi.
If these same perforations are filled with a medium permeability 20-40
U.S. mesh gravel, the pressure drops are significantly less. Flowing
at 1 BPD/perf the pressure drop in a 3/8" diameter perf is 2 psi; in a
1/2" diameter perf the pressure drop is 1 psi; and in a 3/4" diameter
perf the pressure drop is 0.4 psi. Increasing the rate to 10 BPD/perf
produces only a 6 psi pressure drop in a 3/4" diameter perforation. As
can be seen, placing high quality sand into the perforation tunnel is
very important in limiting the pressure drop across the completion. A
well that is gravel packed with a high quality gravel can easily flow
at a rate between 50-100 BPD/perf with a relatively small pressure
drop.
The selection of the gravel size needed to restrain sand production
but maximize the permeability of the pack has been an area of debate
for some time. Prior to 1966 up to 30% of the gravel packs performed
-138-
-------
resulted in failure. Early gravel pack design was based on the works
of Coberly, Wagner and Hill and suggested using a gravel pack grain
size with a diameter equal to 10 times the formations 10% coarse point
on a cumulative sieve analysis. Formation sand is made up of several
different particle sizes and some method must be used to describe the
overall characteristics of the sample. A cumulative sieve analysis is
a standard method of describing these various particle sizes that make
up the formation sand.
In additional work done by Winterburn it was suggested that the
gravel size determination be based on the fines end of the cumulative
sieve analysis. A finer gravel will naturally impede the movement of
formation sand into the wellbore, however, lower productivity may
occur as a result of the decreased permeability of the pack sand.
In 1974 Saucier performed a series of experiments to simulate a
perforation tunnel . Part of the simulated tunnel was packed with
gravel pack sand and the other part was packed with formation sand.
By flowing liquid through this model and measuring the flow rate,
cross sectional area and pressure drop, the permeability of the sand
in the tunnel could be calculated using Darcy's linear flow equation.
Saucier's work used the median gravel pack sand size to the median
formation sand size as the design criteria. The median sand size was
defined as the 50% point on a cumulative sieve analysis plot. When
this ratio between the formation sand size and the pack sand size was
6 and the flow rate was 8.2 BPD/perforation the pressure drop was 16
psi. When the flow rate was increased to 14 BPD/perforation the
pressure drop increased to 30 psi. Since both the flow rate and the
pressure drop were increasing proportionately, the permeability
remained constant. The flow rate was then lowered back to the
original 8.2 BPD/perforation and the pressure drop came back to the
original 16 psi. This indicated that the permeability remained
constant throughout the experiment and no damage was incurred at the
higher flow rate.
When the gravel size was increased to 8.5 times the formation sand
size, a higher pressure drop of 54 psi was obtained with a lower flow
-139-
-------
rate of 7.7 BPD/perforation. This demonstrates that the permeability
was already being damaged due to formation sand migration into the
gravel pack sand. When the flow rate was doubled to 13.0
BPD/perforation the pressure drop went up by more than double
indicating that the permeability had been further damaged. Finally
when the flow rate was lowered back to the original 7.7
BPD/perforation the pressure drop was now 94 psi indicating that the
permeability of the gravel pack had been permanently damaged.
The results of Saucier's experiments show that as long as the diameter
ratio between the median pack sand and the median formation sand is
less than 6, none of the formation sand will migrate into the gravel
pack section. This will maintain the permeability at a maximum and
provide the minimum pressure drop. Whenever this diameter ratio is
exceeded, the permeability will begin to drop as formation sand
migrates into the gravel pack. It will finally reach a point at about
14 times the diameter where there will be unrestrained formation sand
production through the gravel pack.
One of the major problems of gravel packing technology prior to the
early 70's was that the permeability of the final pack was extremely
low. This would create a large pressure drop across the completion and
only highly prolific wells could produce in this manner. The reasons
for this low pack sand permeability have been identified in the last
several years and can be summarized into three major areas:
1) Poor quality gravel pack sand
2) Fluid systems during placement
3) Placement technique itself
Each of these areas will be covered in the following sections with
steps and recommendations to minimize any of the above problems. The
importance of quality gravel packing techniques cannot be over
emphasized. In many cases the results of the gravel packing operation
will determine the success or failure of the well. While it is
understood that the gravel packing of a well may represent a major
portion of the total project cost, the incremental costs associated
with performing quality gravel packs are minimal when compared to the
-140-
-------
cost of well failure and re-completions
Wellbore Preparation And Maintenance
Preparation and maintenance of the wellbore environment is one of the
most critical steps in the implementation of a sand control
completion. This step begins with the choice of completion or workover
fluid. The majority of wells today are completed using clear water
brines. These brines are used because of their non-formation damaging
characteristics and low solids content. The use of a low solid fluid
is extremely important in a gravel packed completion because of the
reduction in permeability that would result in mixing solids from the
completion fluid with the gravel pack sand itself.
Clear water brines are available in a weights of 8.4 to 19.2 Ibs/gal
For those well not requiring high density fluids for hydrostatic
control, brines made of 2% potassium chloride or 3% ammonium chloride
are frequently used.
Frequently the fluids left in the casing prior to the completion
process are of a different weight and viscosity than the fluids to be
used for the completion or workover. For this reason the casing fluid
must be displaced with the completion fluid prior to operations
commencing. It is important that this operation be carried out
efficiently in order to insure that no fluid or debris will
contaminate the clear brine completion fluid.
The procedure to changeover from drilling mud to completion fluids
depends on mud type, storage facilities, logistics, and environmental
conditions. Before displacement of the drilling fluids begins, all
surface equipment should be thoroughly cleaned. All circulation during
this procedure should be in the reverse circulation mode where fluid
is pumped into the annulus and returns taken up the workstring or
drillpipe. This technique insures maximum turbulence during fluid
circulation and more efficient transport of debris to the surface.
-141-
-------
The following is a general procedure for changeover from drilling
fluids to completion fluids. This process, used by several operators
&11, achieves a clean, closed system in an efficient, and economical
manner:
1. Run bit and scraper on DP to TVD.
2. Reverse circulate drilling fluids through surface
cleaning equipment to remove solids while diverting
cleaned mud to storage.
3. Clean all surface lines and tanks with high pressure
hose.
4. Make up and store in portable tanks required volume of
completion fluid.
5. Prepare and pump the following pills*:
a. Spacer containing 2% surfactant
b. Caustic wash with 3-5 Ibs caustic per barrel
c. Scavenger slurry containing 1 Ib/barrel HEC and
25-50 Ibs per barrel sand blast sand
d. High viscosity pill containing 3-5 Ibs/bbl HEC
6. Follow pills with completion fluids
7. After pills have been reverse circulated to surface
change to closed system
* Pills should be designed for a 10 min. contact time and
turbulent flow. Do not stop pumping during changeover as
the different weights and viscosities of the fluids will
cause contamination of the pill stages and they will have
to be discarded. The workstring should be reciprocated
and/or rotated to enhance the displacement process.
Once the changeover has been completed, the use of pipe dope should be
held to a minimum. It has been found that pipe dope is a major cause
of plugging not only in the formation but also on the gravel packing
screens. This plugging of the screens may cause poor gravel packs due
to the inability to effectively dehydrate the slurry through the
screen. In recent field tests pipe dope was applied to only the pin
end of the workstring using a 1 in. (2.54 cm) paint brush and then
only when required. It was found that as many as 5 round trips were
-142-
-------
made without galling or leaks developing
10
By using the changeover procedure outlined above, acceptable clarity
levels may be obtained in as little as 1-1/2 circulations. In the case
of severe hole contamination, this process may need to be repeated 2
or 3 times, in general time and money spent on the proper execution of
the changeover procedure will be less than that of additional
filtering and loss of production if improper displacement methods are
used.
Once the system has been closed it is imperative that all completion
fluids be filtered before being pumped into the well. There are two
methods generally used to filter completion fluids; cartridge filter
systems and Diatomeacous Earth (DE) systems.
Cartridge systems utilize filter elements for the removal of solids
from the completion fluids. These elements are constructed by wrapping
a perforated tube with a woven material made of cotton or
polypropylene fibers. The tightness of the weave determines the size
particle the elements are capable of filtering. Cartridge elements are
rated by the nominal size in microns of the smallest particle to be
filtered out. A 10 micron element is constructed to filter those
solids with a diameter 10 microns or larger. However, differential
pressure across the filter element may distort the weave allowing
larger particles to pass through. For this reason it is necessary for
the filter system to be monitored closely and the filter elements
changed often. Absolute cartridges are available that will correct
this problem but are in most cases cost prohibitive. Due to the
frequency of element changes required to maintain the fluid clarity,
it is recommended that systems having volumes greater than 100 barrels
use DE filtering systems.
DE filter systems are proven to be an effective and economical method
of cleaning completion fluid systems . The DE system, however must be
sized properly in order to achieve optimum filtering capabilities. An
acceptable rate of filtering is .25-.50 gallons per minute per square
foot10. When fluids of higher weights are used, care must be taken
when sizing DE systems as the increase in weight and viscosity of the
-143-
-------
fluid greatly reduces the efficiency of the filter system in terms of
rate and solids removal.
Additional problems that may be encountered with improperly sized DE
systems are solids blow-by and DE bleed through. These two problems
are caused by operating the system at a higher pressures than
recommended in order to maintain an acceptable rate. These problems
were addressed by Glaze and Echols and found increasing the square
footage and reducing the flowrate increased filter efficiency. Due to
the plugging ability and the fact that DE is virtually insoluble in
acid, it is recommended that a cartridge filter unit be placed down
stream of the DE unit as a guard against contamination of the
completion fluid with DE material.
In order to prevent excessive fluid loss to the formation, completion
fluids should be of the lowest density possible while maintaining a
safe hydrostatic over-pressure (100-150 psi overbalance). In many
cases however fluid loss materials will be required in order to
maintain hole stability. These materials may be used for the control
of fluid loss to the formation and/or as a method of preventing
formation sands from "sloughing" into the wellbore. In either case
these materials must be non damaging and 100% removable. The size and
use of wellbore stabilizing slurries should be held to a minimum, that
is use only what is required to continue with normal completion
procedures.
There are three major fluid loss systems in use today; calcium
carbonate, saturated salt systems, and HEC gel slurries. Calcium
carbonate (CaCO.,) is often used as a well stabilizer because of it's
ease of handling and 100% solubility in hydrochloric acid. Calcium
carbonate is normally mixed in a HEC pill at a load of 50 Ibs/barrel
and spotted across the perforations. If calcium carbonate is used
during the completion or workover process it is necessary that all of
the material be removed with hydrochloric acid prior the start of the
gravel packing operation. If not the calcium carbonate forms a thick
filter cake that will not allow for good injectivity into the
perforations.
-144-
-------
Another form of wellbore stabilizer is the super-saturated salt
systems. This involves the building a viscosified salt pill to a point
that exceeds the saturation limits of the fluid. Additional granulated
salt in then added to the pill, however will not dissolve because of
the fully saturated state of the fluid. The pill is then spotted
across the formation and the salt crystals act as a plugging agent.
The salt is removed by circulating or injecting a fluid with a low
salt concentration and therefore dissolving the granulated material.
As in the case of calcium carbonate, the salt pill must be removed
prior to the actual gravel packing operation.
Finally HEC (hydroxyethyl cellulose) is the most widely used method of
lost circulation control. HEC can be used to viscosify both treated
fresh water and brines. Through many test and field applications it
has proven to be the least damaging gelling agent. HEC pills will
degenerate as a function of time and temperature without additional
treatment. Pills are normally mixed with a loading of 3.5-4 Ibs of HEC
per barrel of brine and then spotted across the formation. The
viscosity of the gel will then impede the loss of fluid to the
formation.
PERFORATING
Perforating is that part of the completion procedure that allows for
communication between the wellbore and the formation. The main
objective of a perforating program is to achieve channels which allow
for efficient flow of fluids from the reservoir into the wellbore.
Early perforating systems involved the use of mechanical, hydraulic or
bullet perforators. These systems often were a source of high
formation damage and excessive rig time and have been largely replaced
by explosive shaped charges. Shaped charges were developed originally
for use in anti-tank guns by the military in the late 1930's. The
typical shaped charge consists of a steel case, charge and a liner.
Upon detonation of the charge, the force generated by the rapidly
burning material is focused by the construction of the case and
-145-
-------
creates a high speed jet of metal which penetrates the casing walls.
Explosive shaped charges are available in many forms and
configurations. The two basic groups however are classified as
expendable guns and hollow carrier guns.
Expendable guns are shaped charges run in the hole on an explosive
train to the proper depth. The charges are then detonated and as the
name implies, all firing mechanisms and charge housings are destroyed
and left in the hole. The advantage of expendable guns is that a
larger charge may be used for any given gun diameter. The
disadvantages are a.) debris left in the hole may interfere with
further completion operations, and b.) potential casing and cement
damage caused by improperly positioning the guns.
Hollow carrier guns consist of a shaped charge confined in a pressure
housing and may be run on wireline or tubing. When detonated, this
hollow carrier retains the debris from the charge housings and firing
mechanisms. The additional strength and protection of firing
mechanisms make hollow carrier guns the more reliable choice of
perforating guns.
In order to design an effective perforating program many factors
should be considered:
1. Perforating Fluids - This is the fluid that will be across
from the perforations when the guns are actually fired.
2. Perforating Debris -Debris consisting of copper, lead, and
copper are injected into the formation every time a
perforating gun is fired.
3. Perforation Compacted or Crushed Zone -In test performed by
Saucier and Lands on Berea cores showed that an area of
severe damage consisting of crushed or compacted zone extended
radially from the perforation to a distance of up to .7
in.(1.7 cm.)
-146-
-------
4. Perforation Cleaning - The perforating process has inherent
damaging characteristics that cannot be completely eliminated.
In order to correct these problems an effective method of
cleaning the perforations needs to be considered.
5. Depth of Penetration vs. Perforation Diameter - In general, as
shot penetration increases the shot diameter diameter
decreases. The depth of penetration should always be
sufficient to extend past the damaged area caused by drilling
and cementing.
6. Shot Density -The number of shots per foot to achieve maximum
production.
The perforating program required in a gravel pack completion is
significantly different than perforating in harder formations. One of
the decisions that must be made in this area is the type of
perforation cleaning system. Two general methods exist; 1) the well is
perforated with wireline casing guns and the perforations washed with
a mechanical wash tool assembly and 2) the well is perforated
underbalanced with tubing conveyed peforating equipment and the
underbalanced condition allowed to clean the perforation tunnels.
Perforating underbalanced (the formation having a higher pressure than
the wellbore at the moment of perforation) will help overcome much of
the damage caused by perforating, drilling, and cementing operations.
In soft sand formations a 500 to 1000 psi underbalance is used to
perforate the formation. While underbalanced perforating is an
excellent method of removing damage from the perforations it is
doubtful and should not be expected that all perforations will be
affected the same.
Formation damage may also be removed by washing the perforations. For
this method a mechanical wash tool is lowered to the perforated
interval on tubing and fluids forced into the formation at 6"-!'
increments. After being injected into the formation the fluid washes
the areas directly adjacent to the casing with returns being taken
through the perforation immediately above the section being washed.
-147-
-------
In test performed by Penberthy perforations were tested under
several different conditions and the following conclusions were made
in regards to washing:
a. Voids can be washed between perforations and then filled
with gravel pack sand.
b. The amount of formation sand removed by washing increases
with increasing pump rates.
c. Low viscosity rather than high viscosity fluids are more
effective in washing perforations
d. Perforation wash volume geometry is dependent on the
permeability of the formation sand. High permeability sand
is more easily removed than low permeability sands
e. Perforation washing can precipitate pressure parting if the
pressure gradient is exceeded
It must be noted that washing perforations runs the risk of
intermixing formation clays and shales with productive sand causing
possible formation damage. Either perforating systems may be used
successfully in the completion process as long as a viable cleaning
process is utilized.
The next major item in the design of the perforating program is the
selection of shot size, shot diameter and shot density. With the
perforating equipment available today shot diameter and shot
penetration are mutually exclusive, as shot diameter increases, shot
penetration decreases. In sand control applications this trade off is
decided in favor of shot diameter for the following reasons.
As detailed in the Theory section of this paper, the pressure drop in
the perforation tunnels for a gravel packed completion is quite large.
For this reason it is important to have as many open perforations as
possible to pack gravel pack sand into.
Furthermore in soft formations requiring sand control, penetration is
not generally a problem. The damage caused by the shaped charge in the
compacted or crushed zone is also not as severe due to the formation's
-148-
-------
ability to absorb the force of the perforating charge. For these
reasons it is much more productive to increase shot diameter in
relation to shot penetration. A normal perforating program for a
gravel packed completion will consist of a shot diameter of the
largest size capable, usually greater than .75", and a shot density of
12-16 shots/foot.
GRAVEL PACK DESIGN
The purpose of a gravel pack is to place a high quality sand in the
perforation tunnels and around a screen which has been positioned
across from the productive interval. The sand acts as a filter and
keeps the formation sand from being produced with the well fluids. As
with any phase of the completion operation, there are several design
parameters which need to be addressed to assure a quality gravel pack.
Most gravel packed completions are performed under cased hole
conditions and this discussion of gravel pack design will be aimed
more towards these types of completions. Open hole gravel packs differ
primarily in the perforating program, however many of the items
covered in this paper will also apply to these types of completions.
Gravel Pack Sand Sizing
The first step in designing a gravel packed completion is to obtain a
sample of the formation material in order to be able to size the
required gravel pack sand. Rubber sleeve and conventional cores are
excellent methods of accomplishing this because they obtain a large
volume of sample which is representative of the true formation sand
size- The difference between these two coring techniques is that the
rubber sleeve core has a rubber sleeve which lines the core barrel.
This sleeve contracts to hold the formation material in place while it
is being tripped out of the hole. The conventional core barrel does
not have this feature therefore allowing unconsolidated formation sand
to fall out of the core barrel while being tripped to the surface-
-149-
-------
The most common and readily available formation sample is the side
wall core. Side wall cores are easily obtained prior to setting casing
by the use of a core gun. The gun is run into the hole and shot at the
intervals of interest. The samples can then be individually tested and
studied. Many companies go to the additional expense of obtaining two
sets of side wall cores. One set is given to the geologist and the
other set used to design the sand control completion.
Bailed and produced samples are the worst method of collecting
material, although they are sometimes the only sampling technique
available. Bailed samples are poor because of the inability to
determine where the bailer caught the material. Finer particles will
settle out of the fluid first, resulting in coarser material resting
on bottom and the finer material on top. Because we cannot determine
where the sample is being taken in this gradated sand column, a true
representation of the formation is not known.
The next step in the design process is to analyze the formation sand
sample to determine the median sand grain size. The sample is sieved
on a series of sieves to obtain the weight percentage retained on each
screen. The cummulative percentage on each consecutively smaller
screen is plotted against the sand grain size. When plotted this graph
looks like an S-curve. The sand size which is the most representative
of this particular formation sample is chosen to be the 50 percentile
point on the S-curve. This design point determines the median sand
grain size of the formation. As discussed in the Theory section of
this paper the size of the gravel pack sand can now be determined.
The selection of the size and quality of the gravel pack sand is of
utmost importance. The size of the gravel pack sand will determine
whether the formation sand is restrained while the use of poor quality
gravel pack sand may cause a reduction in the permeability of the
final pack. Fines can be generated by erosion of the sand grains
during transportation or during the placement process. These problems
are almost always associated with angular type gravel. When forces are
applied to angular gravels by handling, trucking, shipping, or pumping
operations, the gravel tends to be eroded to a more spherical shape.
-150-
-------
The particles that are broken off will plug the pores therefore
causing a reduction in permeability.
i
Fines can also be present in the gravel source due to improper quality
control by the gravel supplier. Poor quality control can be seen in a
gravel that has a large percentage of fines or oversize particles. A
good quality gravel pack sand should be within 96% of specifications
and should not have any grain size varying by more than 2%.
The perfect sand grain will have a sphericity and a roundness factor
higher than 0.6, with 1.0 being a perfect circle. It will also have a
very rough crater-like surface which gives the gravel enough
frictional resistance to form a stabilized pack that will not be
fluidized by the production or injection process. Glass beads have a
slick surface and a very low friction factor and if used in a gravel
pack can be very easily fluidized.
•
Attention must be given to avoid using inferior gravel pack sand. The
use of a gravel with a guaranteed low quantity of fines and oversized
grains will result in a better gravel pack. For the 20 mesh cuts (i.e.
20-40, 40-60, 50-70) there should be no more than 2% by weight of
oversized or undersized particles. For the 10 mesh cuts (i.e. 20-30,
30-40, 40-50) 1% oversized and undersized can be tolerated.
Rounded gravel will greatly reduce erosion during the placement
operation and limit the amount of fines generated. The use of a gravel
with a high percentage of quartz is also beneficial because the high
quartz content increases the strength of the sand grain. This results
in the sand being very resistant to crushing and erosion as well as
being very resistant to acid exposure.
Gravel Pack Fluids
The next step in designing a gravel pack is determining the type of
gravel packing fluid to be used. There are a variety of gravel pack
fluids available to the industry today and the correct system for a
-151-
-------
specific well depends on various parameters that must be examined
before a final design can be determined. The two most widely used
systems are the slurry pack and the water pack.
The water pack systems was the original method for the placement of
sand in which the un-viscosified workover fluid is used to carry the
gravel pack sand to bottom. Because of the poor carrying capabilities
of water, the concentration of sand must be kept low. This operation
is carried out by the placement of a sand injector in line with the
gravel pack pump. The gravel pack sand is then injected into the well
with completion fluid at a rate of approximately 50-100 Ibs/barrel.
Gravel packing with a sand injector tends to co-mingle gravel and
formation sand in the perforation tunnels therefore causing a severe
reduction in permeability. In addition water packs require large
amounts of fluids and time to execute. Due to the disadvantages
encountered with this system, water pack are discouraged.
A slurry pack is performed by loading a viscosified fluid with gravel
pack sand and pumping this sand ladened fluid into the screen-casing
annulus. The viscosity of the carrier fluid is such that the
concentration of gravel pack sand can be greatly increased, normally
around 300 Ibs/bbl. This highly concentrated slurry moves into the
perforation tunnel as one mass and allows for little inter-mixing of
formation sand and gravel pack sand. This maintains the permeability
in the perforation tunnels at a maximum and therefore minimizes the
18
pressure drop across the completion.
The base fluid used for the slurry is usually fresh water treated with
3% Ammonium Chloride or 2% Potassium Chloride. The completion brine
may also be used with an HEC loading at a ratio of 2.5-3.0 Ibs
HEC/barrel. With this high sand/fluid ratio, much less fluid is
required to perform the gravel pack therefore reducing the required
placement time. The slurry pack method is by far the most popular
fluid method in gravel packing and is highly recommended for the
completion design.
151a
-------
Gravel Pack Techniques
The next major decision concerns the method of placing the slurry
across the perforated interval. Although it will not be covered in
this paper it is recommended that a matrix acid treatment be performed
prior to gravel packing regardless of the technique used. This
procedure assures that the perforations are open and taking fluid and
therefore can be packed with the gravel pack sand. The actual method
of placement will depend on factors such as well deviation, length of
interval, and tool spacing. There are three techniques available for
slurry placement. These are:
1. Squeeze pack
2. Conventional circulating pack
3. Bottom-up circulating pack
Due to limited service tool manipulation, squeeze packs are generally
less complicated to perform than circulating packs. When a slurry is
squeezed into place, the slurry is circulated to the gravel pack
packer and then forced into the formation. Because there is no way to
insure the slurry has been introduced to the entire interval and
cannot be squeezed through the production screen there is a
possibility that voids will exist in the gravel pack. These void may
result in movement of formation sand into the perforations and
wellbore, greatly reducing productivity. For this reason it is
recommended that a squeeze pack be performed on those intervals no
greater than 20 feet.
A circulating pack, as the name implies, utilizes a circulation path
to position the slurry across the production interval. This process is
accomplished with the use of washpipe in a four position service tool.
A lower screen section (Tattle-Tale) is placed below the production
interval and separated from the production screen with a seal bore
assembly. Wash pipe is then placed in the seal bore to direct the
circulation path through the tattle-tale (lower circulation)
Once the tattle-tale is covered the circulation path is blocked and
-152-
-------
the slurry squeezed into the formation. After sand-out the wash pipe
is pulled from the seal bore and circulation is re-established with
the washpipe through the production screen (upper circulation). This
circulation path allows the slurry to be de-hydrated through the
screen. This placement method is much more successful at placing sand
both in the perforations and around the screen.
The technique described above has worked well in straight holes and
produces satisfactory results. In a deviated well however, the gravel
fails to pack uniformly and voids are developed in the packed annulus.
In recent years, several studies ' ' have been performed to
investigate this problem.
All of these studies indicate that as the degree of deviation
increases, the percent pack in the annulus decreases. The main reason
for this correlation is that gravitational forces tend to cause the
gravel to prematurely settle out near the upper part of the zone to be
gravel packed. As a result, a small dune begins to form at the upper
end of the zone. As the dune enlarges and desends down the annulus,
more and more of the carrier fluid is diverted through the screen by
the fluid flowing over the dune. This causes the velocity of the
slurry to decline therefore resulting in additional sand settling.
This process continues until the dune completely blocks flow to the
lower portion of the annulus. When this shut off occurs, the slurry
fluid is diverted through the top section of the screen and no further
slurry can be paced in the annulus. This will adequately pack the
upper section but leave a void in the lower section.
19
In field and laboratory studies conducted by Stiles a new technique
for placement of sand slurry was developed. This technique has been
named bottom-up after the manner in which the sand is placed in the
screen annulus. In conventional gravel packs, the sand is pumped from
the top of the interval to the bottom. The bottom-up method reverses
this flow and the sand is circulated from the bottom of the interval
to the top. This concept is similar to the method used in cementing
casing strings in a well.
As identified in the gravel packing studies referenced above, the
-153-
-------
settling rate of the gravel pack sand is what causes the "duning"
phenomena to occur. The bottom-up system reduces this settling rate
and therefore reduces or eliminates the formation of these dunes.
Another way of describing this effect is that the sand is being
constantly "bumped" up the hole by the force of the fluid and
therefore not allowed to settle out. This will keep the sand in the
fluid instead of settling out and being deposited on the low side of
the casing. The sand will then move completely to the top of the
screen in plug flow and not accumulate at the bottom of the zone.
Downhole Equipment
The final step in the design of a gravel pack completion is the
selection of the downhole equipment. Selection of packers and screens
will depend on the type of fluid to be injected, the type of gravel
pack to be performed and the production requirements.
In many injection wells the fluid being disposed of is corrosive to
the standard metal and rubber products used in conventional downhole
equipment. For this reason premium metals, such as stainless steel or
other exotic metals are used in the construction of the tubulars,
packers, screens and other downhole accessories. Information
concerning the application of such materials is available from the
equipment suppliers.
In conventional oilfield equipment, seals and packer elements are
normally made of butal-nitrile. Under corrosive conditions however
premium rubbers such as vyton, ryton, and teflon must be used. These
premium rubber products are resistant to many corrosive fluids but
require certain design modifications to the wellbore hokk-up. For
example these materials are often quite brittle, and therefore once
located in a seal bore they should not be allowed to move. This
requires certain modifications to the downhole equipment and must be
designed for in the completion setup.
In gravel pack completions, there are two basic types of production
-154-
-------
packers: Permanent-retrievable packers and mechanical hook-wall
packers. The type of packer used is dependent upon depth, system
pressures, production or injection rates, access to well, and
regulatory guidelines. Hook-wall packers are run on tubing and
normally set with a combination of torque and set down weight. Upward
pull of the tubing unsets these packers at which time they can be
reset without redressing the tool.
Permanent retrievable packers are set with a "setting tool" that
imparts a force to the packer that sets the slips and expands the
packing elements. The setting tool is then pulled from the hole
leaving the packer in a set position. A seal assembly is then run into
the hole and sealed into the polished bore on the packer.
In general the hook-wall packers are less expensive than the permanent
type packers and are used in shallow to medium depth land
applications. The major disadvantage to this type of packer is the
possibility of it becoming unset during production or injection
operations. As stated above these packers will unset with upward pull
of the tubing. This same upward pull however can be generated by
tubing shrinkage caused by normal well operations. From tubing
movement analysis it can be seen that this problems becomes more acute
when the depth of the well increases, the system pressure increases,
high volumes of fluid are being produced or injected, and bottom hole
temperature increases. For these cases a more permanent installation
is required and a permanent type packer should be used.
Another problem encountered with hook-wall packers is the difficulty
in the setting process. It is sometimes difficult to determine when
enough torque has been applied to correctly position the setting
mechanism on the packer. This can result in over-torquing the
workstring. Furthermore when the tubing is lowered to apply set down
weight on the packer, the tubing buckles in the form of a helix and a
significant amount of the applied weight is supported by the friction
between the tubing and casing. These problems are amplified by well
deviation and depth and can cause difficulty in setting the packer.
A permanent-retrievable packer is more costly than a hook-wall packer
-155-
-------
but has the advantage of being both extremely stable and tubing
retrievable. This type packer cannot be unset by tubing movements and
normally has working pressures which greatly exceed other production
packers. The permanent-retrievable packer was developed for offshore
use, however, because of the ability to support a variety of gravel
pack configurations it is frequently used in land operations. These
types of system also require that a seal assembly be run on the end of
the tubing in order to seal into the bore of the already set packer.
This seal assembly isolates the injection fluid from the production
tubing annulus.
The seal assembly may be used in a floating seal or fixed seal «
configuration. A floating seal assembly allows the seal to move within
the seal bore during production. A tubing movement analysis should be
performed to insure that sufficient seals are run to prevent
communication between the production tubing and the tubing annulus.
Furthermore in some high volume injection wells, forces encountered
during injection may cause sufficient stress in the form of a bending
moment that permanently distorts the tubing. In these cases the seals
are not allowed to float or move in the packer bore but are held in
place with some type of anchoring system. The forces generated by
tubing movement in the well can be substantial and should always be
analyzed and incorporated into the final well design.
Gravel Pack Systems
The selection of the production packer is not only dependent upon well
conditions but also dependent upon the gravel packing method to be
used. In order to select the type of production packer to be used an
understanding of the different gravel pack systems is first required.
The gravel pack process itself requires a packer to be set during the
pumping operation. The packer is needed in order to be able to squeeze
the sand slurry into the perforations without applying the squeeze
pressure to the entire casing string. After completion of the pack, a
production packer must also be set on top of the production screen in
order to control the flow of fluids into the tubulars. In general the
-156-
-------
following two systems are available to the operator to achieve these
results:
1) Two trip system
2) One trip system
The two trip system was the original method for completing gravel
packed wells and utilizes a mechanical hook-wall packer as the gravel
pack packer. The hook-wall packer is used during the gravel packing
operation and then pulled from the well after completion. A production
packer and an overshot is then run to bottom and sealed over a hook-up
nipple that is on top of the screen and liner assembly. The production
tubing is connected directly to the packer and therefore no seal
assembly is required. Almost any type of production packer can be used
with this system although most often another hook-wall packer is
selected. As the name implies this method requires two trips of the
pipe in order to finish the gravel pack.
Due to the design of the hook-wall packer used in the two trip method,
this technique is primarily used to perform squeeze packs. Although
some modifications may be employed to allow a circulating pack, it is
not capable of providing an upper and lower circulating position. Due
to the lack of a true lower circulating position the slurry may not be
"forced" to the bottom of the interval resulting in a premature sand
out. For these reasons it is not recommended that this type of gravel
pack be used for zones greater than 20' in length.
The one trip system was developed in the mid 70's for offshore
completions. This system utilizes the same packer for both the gravel
packing operation as well as the production operation. A permanent-
retrievable packer is set prior to the gravel pack and then left
behind to serve as the production packer. This concept requires one
less trip and is used exclusively in offshore completions.
The one trip system allows for the most flexibility of all of the
gravel pack systems. Both a squeeze pack and a circulating pack can be
performed with the system depending upon how the screen assembly is
configured. Many service companies offer a four position one trip
-157-
-------
system and these should be used when possible. This is especially true
for gravel packing zones in excess of 20' or on wells deviated greater
than 45 degrees.
In general these one trip systems are run as follows: The packer,
crossover and setting tool, and screen assembly are run in the hole
and positioned across the production interval. The packer is then set
and the crossover tool released with mechanical and/or hydraulic
force. Once the packer has been set and the gravel pack has been
performed the crossover and setting tool are pulled out of the hole.
The seal assembly can then be run into the well and the remainder of
the completion process continued. Appendix A contains a sample
completion procedure using a one-trip type gravel pack system.
The actual type of gravel pack system to be used is dependent upon the
well conditions. In general the one-trip type systems produce superior
packs because of the four positions which are available during the
packing operation. In addition these systems utilize a stronger more
stable packer for the final production packer. The two trip systems
however can be used successfully on short zones, shallow depth, low
pressure and other field applications.
CONCLUSIONS
1) Gravel packing injection wells is a viable technique that can
control unconsolidated formations while maintaining high
injectivity.
2) The placement of high quality gravel pack sand in the perforation
tunnels is the most important aspect of the gravel pack procedure.
3) The use of clean low solids fluids is a requirement for high
quality gravel packs.
4) The perforating program for gravel packed completions should allow
for some type of perforation cleaning, high shot density and large
-158-
-------
shot diameter.
5) The gravel pack system to be used should be matched with the
wellbore conditions and the equipment requirements.
REFERENCES
1. Liebach, R.E. and J. Cirigliano: Gravel packing in Venezuela,
Seventh World Petroleum Cong. Proceedings, Mexico City (1967),
Sec. Ill, p. 407-418
2. Williams, B.B., L.S. Elliott, and R. H. Weaver: Productivity of
inside casing gravel pack completions, J. Petroleum Technology
(April 1972), p. 419-425
3. Saucier, R.J.: Conciderations in gravel pack design, J. Petroleum
Technology (Feb 1972) p. 205-212
4. Holman, G.B.: Evaluation of control techniques for unconsolidated
silty sands, J. Petroleum Technology (Sept 1976) p. 979-984
5. Monroe, S.A. and W.L. Penberthy, Jr.: Gravel packing high volume
water supply wells, J. Petroleum Technology (Dec 1980), p.
2097-2102
6. Manthooth, M.A.: "Statistical Analysis of Recent Sand Control
Work", API Committee on Sand Control, API Paper 926-13-G (1968)
7. Coberly, C.J. and Wagner, E.M.: "Some Considerations in Selection
and Installation of Gravel Pack Oil Wells", Pet. Tech. (Aug.
1938) TP 960.
8. Hill K.E.: "Factors Affecting the Use of Gravel in Oil Wells,"
Oil Weekly, (may 26, 1941 p. 13-20
9. Winterburn, Read: "Control of Unconsolidated Sands in Wilmington
-159-
-------
Oil Field," Drill, and Prod. Prac., API (1947) p. 63-79
10. Ledlow, L.B. and Sauer, C.W.: "Recent Design, Placement, and
Evaluation Techniques Lead to Improved Gravel Pack Preformance,"
SPE 14162, 1985
11. Sollee, S.S., Elson, T.D. and Lerma, M.K.: "Field Applications of
v Clean Completion Fluids," SPE 14318, 1985
12. Barren, C.W., J.A. Young, and R.E. Munson: "New Concept-High
Density Brine Filtration Utilizaing Diatomaceous a Earth
Filtration System," SPE 10648, 1982
13. Glaze, O.K. and J. B. Echols: "Filtering oil field brines is not
that simple," World Oil, (Oct. 1984), p. 85-90
14. Penberthy, W.L.: "Gravel Placement and Perforation Cleaning for
Gravel Packing", SPE 14161 1985
15. Maly, G.P., Robinson, J.P- and Laurie, A.M.: "New Gravel Pack
Tool for Improving Pack Placement," J. Pet. Tech. (January, 1974)
19-24
16. Gruesbeck, C., Salathiel, W.M. and Echols, E.E.: "Design of
Gravel Packs in Deviated Wells," paper SPE 6805 presented at SPE
52nd annual Fall Technical Conference, Denver, Oct. 9-12, 1977.
17. Shryock, S.G.: Gravel Packing Studies in a Full-Scale, Deviated
Model Wellbore," paper SPE 9421 presented at SPE 55th Annual Fall
Technical Conference, Dallas, Sept.21-24, 1980.
18. Sparlin, Derry D., "Pressure Packing with Concentrated Gravel
Slurry," paper SPE 4033 presented at the 47th Annual Fall
Meeting, San Antonio, Oct. 8-11 1972.
19. Stiles, R.F, Colomb, G.T., and Farley, D.L.:"Development of a
Gravity-Assisted Gravel Pack System, " SPE 15409 presented at
61rst Annual Technical Conference, New Orleans, LA, October 1986.
-160-
-------
20. Saucier, R.H. and Lands, J.F.: "A laboratory Study of
Perforations in Stressed Formation Rocks," JPT, February 1978
-161-
-------
Appendix A
Operating Procedure
Perform any necessary squeeze work. Make a bit and scraper run
to 4;50' below the desired interval to clean the casing wall of
any debris (such as mud cake or scale) which might obstruct the
running of any tools.
After the casing has been scraped, reverse circulate 2 hole
volumes with clean, filtered workover fluid to wash out any
debris which has been scraped from the casing and POOH.
RU the wireline unit and make a gauge ring and junk basket run to
assure a constant casing ID from surface to the desired plugback
and POOH.
GIH with wireline set sump packer and set it to 10-12 feet below
the lowest perforation of the desired interval and POOH
GIH with the perforating equipment and perforate the desired
interval. The recommended perforating shot density is 12-16 spf
with the largest hole diameter allowable. POOH
GIH with the centralized screen and liner assembly. Caliper and
check all tool connections, screen gauges, and record all
lengths. Lubricate all connections with available lube oil. DO
NOT USE PIPE DOPE. For a single zone circulating gravel pack the
screen and liner assembly will consist of:
a. 1/2 muleshoe, collet latch, 2' of seals and locator sub
b. Crossover to tattle tale screen
c. Tattle tale screen
d. "0" ring seal sub
e. All-weld production screen wrapped from bottom up and
-162-
-------
centralized every 15' for casing size used
f. Blank tubing
g. Safety shear out sub
Hang off the screen and liner assembly with clamps. Pick up the
sized washpipe and GIH. Sting into "0" ring seal sub and space
out to the rotary with the proper sized pup joints.
Make up Gravel Pack assembly consisting of:
a. Crossover and setting tool
b. Gravel Pack Packer
c. Slotted extension
d. Lower seal bore
e. Lower seal bore extension
f. Interference collar
g. Interference collar extension
Connect the washpipe to the washpipe adapter on crossover tool.
Next connect the gravel pack assembly with the screen and liner
assembly. Finally connect the workstring to the gravel pack
assembly and GIH with the entire assembly.
Immediately prior to tagging the sump packer, PU on the
workstring and note actual pick up weight.
Gently sting into the sump packer with collet latch and seals.
In order to verify the position of the assembly, pull 30001 over
the pipe pick up weight. Once a positive indication of the latch
in is observed, slack off to neutral weight and close the hydril.
Slowly reverse circulate to fill the tubing.
With the gravel pack assembly properly positioned and the tubing
full, drop the ball and wait for it to gravitate to the ball
seat. Pressure up on the tubing slowly in 500 psi increments to
3000 psi. Three shears will be observed during the packer
setting procedure. The first shear (at +1000 psi) indicates that
the packer has been set. The second (at +2000 psi) shear
indicates that the setting sleeve on the setting tool has been
-163-
-------
sheared which deactivated the setting tool. The final shear (at
+3000 psi) indicates that the releasing sleeve has shifted to the
lower position allowing the threaded lugs to disengage, releasing
the setting tool and crossover assembly from packer assembly and
indicates that the ball seat has been blown to below the
crossover ports.
One the three shears have been observed place 10,000#'s on the
packer and mark the tubing to indicate the squeeze position. PU
8" and mark the pipe to indicate the lower circulating position.
Establish a circulating rate and note the pressure required to
break circulation. PU 2' while bumping against the interference
collar and mark the pipe to indicate the upper circulating
position. Slack off to the squeeze position and set 10,000#'s on
the packer. Establish an injection rate with filtered workover
fluid at less than the calculated fracture pressure.
After the packer has been set and the tool positions established,
a mutual solvent acid job should be performed prior to the
introduction of the sand slurry to the perforation tunnels. A
mutual solvent dcid job will enhance the injection profile,
alleviate damage near the wellbore and provide for rapid clean up
of the well. Position the crossover tool in the reverse
circulating position and spot the acid 2-3 barrels above the
tool. Slack off to the squeeze position and set 10,000t's on the
packer and squeeze the acid into the formation at matrix rates.
Directly behind the afterflush for the acid treatment, pump a
viscous spacer of gelled fluid to prevent any sand from entering
the neat fluid while being pumped downhole. Next, pump a sand
slurry consisting of a gelled fluid containing quality controlled
sands whose sizes are determined by a sieve analysis of
representative samples obtained from the zone of interest.
The use of a high density slurry allows gravel placement into the
formation and perforations tunnels with minimum fluid loss and
prevents the mixing of the pack sand and formation sand. When
two sizes of sand are mixed, the resulting permeability is less
-164-
-------
than either.
When the afterflush has been displaced to the formation, position
the crossover in the lower circulating position and continue
pumping until a sand out of the lower telltale screen is
obtained. The sand out is indicated by an increase in pressure
and a decrease of the returns at the surface.
At this sand out the pressure should be allowed to increase
1000-1200 psi over the circulating pressure. Slack off to the
squeeze position and set 10,000#'s on the packer. Squeeze the
slurry into the formation until a pre-determined squeeze pressure
is attained. Allow the pressure to bleed off into the formation.
PU to the upper circulating position and circulate through the
production screen. Continue circulating until a sand out on the
production screen is obtained. Again, an increase in pressure
and diminishing returns at the surface serve as indicators that a
sand out on the production screen has been obtained. Allow the
pressure to increase o 1000-1200 psi over the circulation
pressure.
Bleed off the pressure and pull the collet through the
interference collar. Pick up 3-4' to get into the reversing
position. Reverse out the excess slurry plus a minimum of 2
tubing volumes. POOH with the crossover tool.
Go in the hole with the production seals with a locator sub,
landing nipples and other required equipment. Continue the
procedure to place the well on production.
-165-
-------
FACTORS THAT CAN CAUSE ABANDONED WELLS TO LEAK
AS VERIFIED BY CASE HISTORIES FROM CLASS II INJECTION,
TEXAS RAILROAD COMMISSION FILES
J. E. CLARK, M. R. HOWARD, D. K. SPARKS
E. I. DU PONT DE NEMOURS & CO., INC.
P. O. BOX 3269
BEAUMONT, TEXAS 77704
ABSTRACT
An abandoned well is a well where use has been permanently
discontinued or is in disrepair such that it cannot be used for its
intended purpose nor for observation purposes. A properly plugged
well is a well where upward migration of fluids does not occur as a
result of increased reservoir pressures.
Abandoned wells are possible sources of pollution to water supplies
if fluids are allowed to migrate into Underground Sources of Drink-
ing Water (USDW) from the over-pressured injection zone. Federal
Underground Injection Control (UIC) regulations require the critical
identification and evaluation of all abandoned wells in the Area of
Review (AOR) during the permitting process.
-166-
-------
Case histories from the Texas Railroad Commission files on leaking
abandoned wells reportedly caused by Class II injection wells (salt
water and enhanced recovery) were studied. Important factors have
been identified from these case histories that can cause an improp-
erly plugged abandoned well to leak due to overpressuring the
injection zone. The factors include: 1) depth of the injection
zone, 2) casing left in the borehole which is open to the injection
zone, providing a direct path for upward fluid migration, 3) reser-
voir properties and flow rates, 4) drilling method, and 5) bore-
holes in "hard" rock which tend to remain open indefinitely, as
opposed to boreholes in "soft" rock where expandable clays or
sloughing shales close the borehole.
An important finding of this study was that wells drilled in
unconsolidated (soft) rock, such as the Texas Gulf Coastal Plain
experience natural borehole closure, which drastically reduces the
potential for leakage from these abandoned wells. This study showed
that the most likely pathway for leakage is a production well
improperly abandoned with the production casing left open to the
injection zone.
All abandoned wells in the AOR must be identified to satisfy Federal
UIC regulations. Abandoned wells that are satisfactorily plugged
are dismissed from further review, and remaining wells are consid-
ered for plugging or modeling to determine the maximum permissible
injection pressure. The maximum injection pressure is set to
prevent the hydraulic lift of the injected fluid or other non-native
-16-7-
-------
fluids into an overlying USDW from improperly plugged abandoned
wells. During modeling it is important to consider the entire well
field of surrounding injection or production wells which may affect
the injection zone. From case studies of several Class II injection
wells suspected of causing leakage through abandoned wells in Texas,
we believe that operators can achieve responsible compliance through
the use of historical records and available modeling techniques.
INTRODUCTION
Since 1859, when the first petroleum well was drilled in the United
States, approximately three million oil and gas wells have been
drilled and over two million have been abandoned (Anzzolin and
Graham, 1984). According to 40 CFR 146, a well is considered
abandoned if its use has been permanently discountinued or is in a
state of disrepair such that it cannot be used for its intended
purpose nor for observation purposes. Of particular concern to the
Class II UIC program are improperly plugged wells that penetrate the
injection zone or within 300 feet of the injection zone, because
they have the potential for conveying fluid from the injection zone
to an overlying Underground Source of Drinking Water (USDW).
Of the approximately 150,000 Class II (brine injection) wells
operating in the United States (Fryberger and Tinlin, 1984), approx-
imately 54,000 are in Texas (Roth, 1987). The State of Texas has
recognized the need for proper plugging of abandoned wells since
1899 when the first regulations were passed. In 1919 the Texas
-168-
-------
Railroad Commission (TRC) was given regulatory responsibility for
proper well plugging. The TRC is also responsible for a program to
remedy improperly abandoned wells where the operator is unknown or
financially insolvent. Through this program approximately 1400
wells have been plugged since 1965 with state funds (Ross and Steed,
1984).
AREA OF REVIEW CONCEPT
The AOR is the main UIC requirement to protect an USDW against
potential upward migration of fluid from boreholes that penetrate
protective confining layers. Abandoned wells come under the current
review process for a UIC permit. In Texas, the AOR encompasses the
area within a 1/4-mile radius of the injection well. If unplugged
wells are known to exist nearby, but outside the AOR, they may
require reservoir simulations to determine the adequacy of the
1/4-mile radius (Engineering Enterprises, 1985).
This State UIC program requires that records of all artificial
penetrations (boreholes that penetrate the confining/injection zone)
be examined during the AOR to locate wells that are improperly aban-
doned. A properly completed or abandoned well is one where inter-
formational movement of fluids will not occur as a result of an
increased reservoir pressure.
-169-
-------
We developed a protocol to identify and evaluate artificial
penetrations in the AOR (Figure 1). All wells identified as being
inadequately plugged must be modeled to verify that no upward
migration will occur. If upward migration is possible, then one of
the following steps must be taken before the injection well is
allowed to operate:
1) Reenter and properly plug the potential problem well.
2) Lower the proposed injection rate to reduce the pressure (head)
driving force.
3) Complete the injection well in a lower zone so that the aban-
doned well can tolerate a higher pressure without fluid
migration.
4) Complete the injection well in a lower zone which the abandoned
well does not penetrate.
5) Increase the density of the injection fluid to prevent upward
migration.
6) Drill a monitor well next to the potential problem well to
monitor possible upward fluid movement.
-170-
-------
FACTORS RELATED TO LEAKAGE THROUGH
IMPROPERLY ABANDONED BOREHOLES
Class II wells are generally constructed with surface casing
cemented below freshwater aquifers, long-string casing perforated
through the injection zone, and injection tubing to deliver brine to
the subsurface. Figure 2 shows the construction of a Class II
injection well and three improperly abandoned wells that provide
potential fluid migration pathways. A leaking abandoned well can
mean a leak at the surface or interformational flow of fluids which
does not reach the surface (Figure 2). Injected fluids will move
laterally through the injection zone and can migrate into an impro-
perly plugged well. A discussion of important factors that relate
to leaking abandoned wells follows.
For the purposes of the study, two rock types were identified:
consolidated ("hard") rock and unconsolidated ("soft") rock. These
two types are geologically distinct and their characteristics
greatly influence the behavior of abandoned wells.
ROCK TYPES
Unconsolidated formations such as the geologically young Tertiary
shales in the Texas Gulf Coastal Plain have hydration (expanding
clays-smectities) and plastic properties which result in the natural
closure of man-made boreholes (Johnston and Greene, 1979; Davis,
1986). Smectite exhibits a high amount of swelling when hydrated.
-171-
-------
Non-expanding clays or illite swell much less on being wetted than
expanding clays. Collins (1986) reported that shales penetrated by
drilling fluids experience a significant water exchange due to an
osmotic process which is dependent upon ionic activity of the mud
and the brine in the shale. This water exchange can lead to
swelling of the shale and sloughing into the borehole.
A change in mineralogy from smectite to illite occurs with
increasing depth and temperature and is associated with squeezing
water out of the clay lattice (Grim, 1968). This alteration is
called clay diagenesis (See Figure 3). Powers (1967) found that
when montmorillonite (smectite) is buried to a depth of approxi-
mately 3000 feet, most of the water is expelled from it, except for
the last few bound layers that are along the basal layers between
the unit layers of clay. At this depth, the effective porosity and
permeability are essentially zero because all space is occupied by
the solid layers of clay and the rigid water layers bound to the
clay. In a laboratory experiment by Darley (1969) most of the free
water in clays was squeezed out of the expanding clay members at a
pressure of 2500 psi, approximately equivalent to 5000 feet of
overburden.
Borehole closure by hydration occurs at depths less than 10,000 feet
in the Gulf Coast. Alteration of smectite to illite (mixed-layer
clay) begins at a depth of 6000 feet (Figure 3) and continues until
a near total transition has occurred by a burial depth of
approximately 10,000 feet in the Gulf Coastal Plain (Powers, 1967).
-172-
-------
Below 14,000 feet in the Gulf Coast, there is no swelling component
remaining in the illite (Burst, 1959).
Borehole closure by plastic flow is associated with high pore
pressure shales being relieved of the overburden stress by penetra-
tion of the drill bit. This geopressured zone (plastic flow) occurs
at approximately 10,000 feet in the Gulf Coastal Plain (Figure 3).
Because the pore pressure and shale plasticity is abnormally high
relative to the overburden strata, the shale is extruded into the
borehole by plastic flow if the drilling fluid pressure (mud column
weight) is less than the fluid pressure in the rock pores being
drilled.
Drilling muds are generally conditioned to prevent borehole closure.
If the mud breaks down or settles out, the borehole will seal itself
by natural closure (Ammons, 1987). Johnston and Knape (1986)
reported after interviewing several experienced drilling engineers
that the geologically young and unconsolidated sediments of the Gulf
Coast tend to slough and swell, and an uncased borehole will com-
monly squeeze shut within hours, resulting in natural borehole
closure. According to Cheatham (1984), shale hydration has been one
of the more significant causes of borehole instabilities in the
past; however, improved drilling fluids in the last 20 years have
provided better control of swelling shales. Therefore, old
abandoned wells which typically did not have good drilling muds
would have exhibited natural closure even more rapidly.
-173-
-------
Reentering and plugging abandoned wells near Du Pont injection
facilities in the Texas Gulf Coastal Plain has confirmed that the
boreholes are closed by natural processes (Klotzman, 1986; Meers,
1987). Old abandoned boreholes have healed across shale sections to
the extent that the reentering is like drilling a new hole. Natural
borehole closure is also verified by day-to-day experience of field
engineers who encounter difficulty in keeping boreholes open while
drilling, running casing, and logging. Our experience in this area
indicates that borehole closure while running casing can result in
being stuck ("wall stuck") in the well and not able to bring circu-
lation of fluids ("break circulation") to the surface. Generally a
wiper trip is made (drill bit is run in the hole and the borehole is
conditioned with mud) to keep the borehole open for logging if it
needs to be left open for more than 24 hours.
Typically, dry holes drilled in the Gulf Coastal Plain have been
abandoned with surface casing set and plugged, but without long
string casing, thus providing ready opportunity for natural closure
below surface casing.
Consolidated formations, such as in west Texas, are generally rigid
("hard rock") and lack the shale mineralogical properties that help
the borehole to close by caving or sloughing (see Figure 4).
Abandoned wells may remain open here indefinitely because the
factors for natural closer are limited. Lost circulation zones are
more common in consolidated rock areas where drilling fluids and
cement may have been displaced from the borehole. Johnston and
-174-
-------
Knape (1986) stated that abandoned wells in this region may remain
open for many years, and reentering the boreholes for plugging may
be done by merely washing down with a drill bit. Most reports of
leaking abandoned wells or groundwater contamination have been
reported as occurring in consolidated rocks (Johnston and Greene,
1979).
A major exception to the normal stability of the West Texas
boreholes is exhibited in uncased sections of wells penetrating
shale formations of the Triassic "red beds". These beds consist of
water-sensitive clays which swell and slough in the borehole,
causing well construction problems and total hole closure during and
after well abandonment. This is typically below the base of the
surface casing in a well where the long-string casing is absent or
has been pulled for salvage prior to abandonment (Johnston and
Knape, 1986).
DRILLING METHODS
The method used to drill a well can influence the potential for
leakage after it is abandoned. Three dominant drilling methods
examined were rotary mud, rotary air, and cable tool.
Rotary drilling with mud as the drilling fluid has been the
preferred method, especially in the Gulf Coastal Plain, since its
invention in 1901. It is almost impossible to drill shale with
-175-
-------
other techniques in coastal plain areas and keep the borehole open
to advance the bit and casing.
The rotary mud rig uses a water-based drilling fluid (mainly a
suspension of bentonite, a swelling clay), weighting material, and
chemical additives as a medium to carry drill cuttings to the
surface, control pressures encountered in underground formations,
and lubricate the bit.
In most wells drilled prior to the 1930's, rotary drilling fluid was
a mixture of water and the drill cuttings. This was called "native
mud", derived from the clay formations penetrated by the drill bit.
Water was continually added to thin native muds, and the minimum
weight for these drilling fluids was probably not less than 9
Ibs/gal (Johnston and Knape, 1986).
When a well reaches logging depth, the mud is conditioned to keep
the borehole open prior to running geophysical logs (a practice
since the 1930's). The density of mud left in the borehole can be
determined from plugging records or from the geophysical log header.
Rotary drilled dry holes can be assumed to have been left full of
mud as a minimum condition because there is no economic reason to
recover the drilling mud prior to abandonment (Johnston and Knape,
1986). However, if the mud were recovered for another project, the
borehole would be filled with a bentonite type mud. Totally
removing the mud system from the borehole with the drill pipe on
-176-
-------
bottom of the well is taking an unnecessary risk of getting the
drill pipe (salvagable material) stuck in the hole, because removing
the mud can cause hole instability and caving.
Mud density, primarily used for well control while drilling, can
also be used to prevent interformational fluid flow. Permeability
of the mud left in the borehole is less than the surrounding produc-
tive formations and the pressure maintained by the mud column in the
hole is high enough to prevent the displacement of the plugging
material. Drilling fluid that is suitably conditioned after
drilling can satisfy these requirements (Polk and Gray, 1984).
In plugging mineral exploration holes, Polk and Gray (1984) found
that by increasing mud viscosity to 20 sec/quart, the exploration
holes that were drilled were sealed with permeabilities less than
10"8 cm/sec. The sealing effectiveness of the mud conditioner
treatment was confirmed by observations of surface hole intercepts
made during the mining operations. This fact minimizes the chance
of encountering a truly open conduit in an abandoned dry well which
was rotary drilled using mud.
Cable tool drilling is sometimes used in consolidated rock forma-
tions, but it has not been used very much in unconsolidated rock
regions for the past 50 years because caving sands and sloughing
shales caused operating problems. If a well were drilled by cable
tool or rotary air drilling methods, then the fluid in the hole is
probably native water or brine. Generally, cable tool holes are
-177-
-------
hard to locate because the surface casing was never cemented and was
removed after drilling.
HUD HEIGHT
The mud column provides a downward force, or higher hydrostatic
head, than the fluids in formations encountered by the drill bit to
maintain well control (keep the well from "blowing out"). This same
mud column can keep the abandoned well bore from "breaking out" due
to injection in other wells, if the formation pressure is not
increased above the hydrostatic head of the mud column. Figure 5 is
an example of pressure resistance of a static mud column exerted at
different depths and mud weights. Figure 6 represents normal for-
mation pressure at depth for two pressure gradients. Figure 7
represents pressure resistance differential based on the hydrostatic
pressure resistance of the mud column minus the formation pressure,
for several different cases. Formation pressure must be greater
than the pressure resistance of the mud column to cause movement of
fluids in the improperly plugged borehole. This is a conservative
calculation because it assumes no credit for borehole closure, gel
strength, or pressure required to break the mud cake gel at the
borehole face.
High-density muds undergo density changes due to gravitional
settling. In a field experiment, Cooke, et al (1983, 1984) made
direct determinations of change in the density of bentonite mud left
standing in the annular space where pressure transducers at various
-178-
-------
levels along the outer casing were located. The water-based mud
weighted with barite to 11.0 Ibs/gal was reduced to 9.1 Ibs/gal in
eleven months. The weight of natural and modern muds left in the
borehole have a reported low range of 9 to 9.5 Ibs/gal (Price, 1971;
Johnston and Knape, 1986; Collins, 1986; Davis, 1986; and Alford,
1987). A 9 Ibs/gal mud would be a conservative value to use in
modeling calculations to predict upward migration in abandoned
wells. This value of 9 Ibs/gal would be valid for rotary mud-
drilled dry holes and for cased holes with long string or production
casing only if records indicate mud/cement left in the boreholes.
Of course, if the records indicate lost circulation zones, or if
casing is pulled from the borehole, the mud column cannot be assumed
to fill the borehole.
GEL STRENGTH
A second mud parameter, gel strength (Gs), helps prevent upward
fluid movement in a mud-filled borehole. Gel strength is the prop-
erty which acts to suspend the drill cuttings in the static mud
column when circulation stops. Drilling mud gels under static
conditions as a function of the amount and type of clays in sus-
pension, time, temperature, pressure, pH, and chemical agents in the
mud system. The pressure required to displace the gelled mud can be
significantly large.
Gel strength may be the main factor in preventing brine from
migrating up abandoned wells from a fluid flow injection well driven
-179-
-------
by pressure build-up (Collins, 1986; Johnston and Knape, 1986).
Collins (1986), in simple laboratory experiments (pipe with collars
or shoulders to simulate different hole sizes and filled with
bentonite mud) to test gel strength, demonstrated that mud gel and
hole irregularities interacted to yield a large contribution (five-
fold or more increase in gel strength) to sealing pressure and help
prevent upward migration.
Gel strength is increased by flocculation which enhances clay
particle contact. Several studies were conducted which showed that
gel strength increases with time (Garrison, 1939; and Gray, et al.,
1980) at borehole conditions. An increase in pH (Garrison, 1939)
increases gel strength. High pressures in thousands of psi (Killer,
1963) pressures generally much greater than those encountered in
Class II injection wells, decrease gel strength. The gelling nature
of mud has been observed and reported in replugging abandoned wells
(Johnston and Knape, 1986).
Minimum gel strength for drilling muds has been reported as 20 to 25
lbs/100 ft2 (Barker, 1981; Johnston and Knape, 1986; Davis, 1986;
Collins, 1986; and Gurke, 1987) and would provide a considerable
safety factor in modeling most situations. Figure 8 is a plot of
gel strength and pressure resistance to prevent upward migration.
The added pressure resistance for a well 5000 feet deep with a gel
strength of 20 lbs/100 ft2 and a 6-inch borehole would equal 50 psi.
-180-
-------
DEPTH OF INJECTION ZONE
Injection zone depth is important because a shallower borehole will
have a lower hydrostatic head (downward force) due to the shorter
fluid column weight in the abandoned well. A longer column of fluid
(deeper injection zone) can counterbalance more formational pressure
buildup in the injection zone. Table 1 shows the hydrostatic mud
pressure for 9.0 Ibs/gal mud at depths from 1000 to 5000 feet. The
mud column has a pressure differential resistance to initiate upward
flow (hydrostatic mud pressure minus formation pressure) of 18 psi
at 1000 feet, and 90 psi at 5000 feet.
CASING LEFT IN BOREHOLE
Special attention should be placed on abandoned wells with long-
string or production casing remaining in the borehole and left open
to the production/injection zone. Generally, if production casing
is intact, then a mud-filled hole cannot be safely assumed, unless
records indicate the presence of mud or cement at abandonment to
counterbalance higher injection pressures.
If an operator abandons a depleted well or dry hole without proper
plugging, then injected fluid from a Class II well (Figure 2, Well
A) could enter the improperly abandoned well from the same pro-
duction zone (Figure 2, Well D). Another potential avenue for upward
migration exists if the well is cemented across only part of the
well bore, and drilling mud was displaced ahead of the cement from
-181-
-------
the annular space between the casing and the open hole (Figure 2,
Well B). If cement was not circulated to the surface, the annular
space above the cemented portion would be filled with drilling mud.
If driving pressures are high enough, fluids can enter the
uncemented or mud annulus and migrate upward if not cemented above
the injection/production zone.
The annular mud space provides resistance as in the mud-filled
borehole to upward migration because of the increased hydrostatic
head of the mud column and gel stength of the mud (Davis, 1986). In
addition, in the Gulf Coastal Plain, shale can close around the
casing and seal off the borehole.
RESERVOIR PROPERTIES
Transmissivity and injection rates are the main variables that
control formation pressure buildup in an injection zone. Trans-
missivity is equal to permeability of the injection zone multiplied
by the pay thickness (injection zone height). Figure 9 shows the
relationship between pressure buildup and distance from the injec-
tion well for various transmissivities and injection rates. Higher
disposal injection pressure buildups are related to zones of low
transmissivity and higher flow rates. Because flow rates are
important to formation pressure buildup, it is imperative to
consider other nearby disposal and production operations utilizing
the same injection zone when determining the potential for leakage
through abandoned wells.
-182-
-------
MODELING UPWARD MIGRATION
Well-established, conservative, engineering models are available for
computing the pressure at which upward migration will begin. The
formation pressure necessary to initiate upward flow (Pf) through an
abandoned well is determined first by calculating the pressure
exerted by the well's mud column and then adding the pressure for
gel strength (note that no additional credit is taken for borehole
closure resulting from shale hydration or the plastic nature of
abnormal pressured shales). Second, the formation pressure prior to
injection (Po) is subtracted from Pf. This difference (Pf-Po)
represents the injection formation pressure buildup which must occur
at an abandoned well to initiate upward flow. This difference is
the key for limiting the maximum permissible pressure increase in an
injection formation at the location of an improperly plugged
abandoned well. An equation developed by Barker (1981) to calculate
the pressure resistance in an improperly abandoned well is as
follows:
Pf = Pt + 0.052*p*H + (0.00333*Gs*H/ Dw) (1)
where : Pf = pressure required in the formation to initiate
upward flow in an abandoned borehole (psi)
Pt = surface well pressure (psi)
p = density of mud (Ibs/gal)
-183-
-------
H = height of mud column (feet)
2
Gs = Gel strength of mud (lbs/100 ft
Dw = maximum diameter of well bore (inches)
Davis (1986) reported an equation to calculate the opposing forces
(mud hydrostatic head and gel strength) that act in resistance to
upward fluid migration along a uncemented/mud casing annulus if not
cemented above an injection or production zone:
Pf = Pt + 0.052*p*H + (0.00333*Gs*H/ Dw-Dc) (2)
where : Pf, Pt, p, H, Gs, and Dw are defined as in equation 1 and
DC = outside diameter of casing (inches)
The AOR for an injection well is dependent upon the following
variables:
1. unit weight of mud plug, gel strength, and borehole diameter,
2. reservoir properties: permeability (k) and pay zone (effective
injection zone) thickness(H),
3. injection rates (Q),
-184-
-------
4. injection or production operations utilizing the same injection
zone,
5. initial reservoir pressure and surface pressure,
6. depth of injection zone,
7. injection and formation fluid properties.
When pressure modeling calculations indicate that injection well
operations are sufficient to cause fluid migration in an abandoned
well, one of the alternatives previously discussed under AOR must be
pursued.
Figure 10 shows cross-section modeling calculations for a reservoir
and indicates that with a 9 Ibs/gal mud at 5000 feet, the area of
review for abandoned rotary drilled dry wells would be less than
1000 feet from the injection well. Figure 11 is a plan view for the
above modeling calculations.
CASE HISTORIES FOR LEAKING ABANDONED WELLS
IN TEXAS
Agency Information Consultants, Inc. (AIC) of Austin, Texas has
examined records on file with the Texas Railroad Commission (TRC)
for pollution problems associated with abandoned wells in the
-185-
-------
following cases: 1) significant problem leaking abandoned wells in
Texas cited by EPA (1975), (AIC, 1987a), 2) proper plugging
hearings from selected counties along the Texas Gulf Coast (uncon-
solidated rock areas) to determine pollution problems in connection
with the upward migration of fluids in improperly abandoned wells
(AIC, 1987b), and 3) proper plugging hearings for fluid migration
from improperly plugged wells in unconsolidated (TRC Districts 2, 3,
and 4) and consolidated rock areas (TRC Districts 7-B, 7-C and 9)
(AIC, 1987c).
CASE 1
The TRC gained authority and funds in 1967 to plug those wells
causing a problem or presenting a potential pollution threat. EPA
(1975) found approximately 830 wells that were plugged from 1967 to
1974 and identified approximately twenty-eight leaking, abandoned
wells that were significant problems and reportedly caused by Class
II injection wells (Figure 12, location map). These wells were
found in a review of the TRC files on unplugged or improperly
plugged wells that have been plugged by State authority. AIC
(1987a) studied these 28 problem wells.
The AIC study identified the following as important factors that
contribute to the potential for upward migration due to injection
operations in the unconsolidated rock areas: 1) long-string casing
left in the borehole and left open to the production or injection
-186-
-------
zone, and 2) significantly overpressured injection zones because of
the low reservoir transmissivity.
Out of 28 problem wells, only 4 leaking abandoned wells were from
the unconsolidated rock area (Figure 12). Three improperly
abandoned wells in the unconsolidated rock region had production
casing set and left open to the injection zone, providing a direct
pathway to the surface and eliminating possibilites for borehole
closure. In one of these wells, a cause-and-effect relationship was
shown when a suspect injection well reduced its flow by two-thirds
and another suspect well was shut in, the problem well stopped
leaking.
The fourth well cited in the unconsolidated rock area was drilled to
a total depth of 1395 feet, abandoned with 21 feet of surface pipe
in the borehole and filled with heavy mud. The well suspected of
causing the problem injected between 1810 to 1900 feet, or 400 feet
below the depth of the leaking well. Thus, this suspect well is not
likely to have been the cause of the leaking well. The most likely
source of salt water for the abandoned well is the fact that fresh
groundwater at this location is very shallow (less than 100 feet).
When the leaking well was entered to stop the leak, "A partial
obstruction was encountered at approximately 20 to 25 feet and it
was found that a solid obstruction of clay and shale was encountered
at approximately 50 feet. It is obvious that this obstruction will
have to be drilled out rather than washed out in order to properly
plug the well" (Eikel, 1969). This record on the attempt at
-187-
-------
reentering the abandoned well confirms that borehole closure can
occur in unconsolidated formations.
In summary, improperly plugged abandoned wells in unconsolidated
formations with long-string casing left open to the injection
interval may have only mud and mud gel strength or formation brine
to withstand pressure buildup. Thus, depth of injection is critical
in these cases. It is important to review the records of all
production wells within the AOR because they are commonly abandoned
with casing intact and they have the greatest potential for upward
migration.
In 21 of 24 cases in the consolidated rock area, leaking abandoned
wells were again due primarily to injection by the suspect wells
into the same interval to which the leaking wells had been open;
but, it was through the production casing or the open borehole.
In the other three cases, AIC (1987a) could not find an injection
well after searching a radius of 1.5 miles for well No. 25. In
addition, the abandoned well was not leaking salt water but was
identified as a well that was not properly plugged. A second
leaking well was drilled to a depth of 4156 feet in consolidated
formations and abandoned with 112 feet of surface casing in the hole
with 75 sacks of cement and heavy mud. An injection well
approximately 3/4 mile away (injection zone 518 to 535 feet) was
suspected of causing the leak; however, when the injection well was
shut down for a week, there was no change in the leaking well.
-188-
-------
Thus, the suspect well was probably not the cause of the leakage.
Additionally, the sand used for injection pinches out in the
direction of the leaking well (Krusekopy, 1970). Lack of sand
continuity prohibits lateral fluid migration. Thus, the suspect
well was probably not the cause of this leakage. The third leaking
well that did not fit the same zone as the suspect well was drilled
to a total depth of 4,050 feet in consolidated formations and
abandoned with 101 feet of surface casing in the hole and filled
with mud. An injection well approximately 1700 feet away was sus-
pected of causing the problem. This injection well was disposing of
salt water through the annulus between 354 and 2302 feet. Modeling
the suspect well based on the following limited reservoir parameters
and sensitivity analysis:
where, Q (flow rate) = 110 bpd
H (pay zone) = 35 feet
p (injection pressure wellhead) = 175 psi
r (radius from well) = 1700 feet
indicated that pressure buildup due to injection was approximately
50 psi at the 530 foot depth injection zone. Assuming 9 Ibs/gal mud
in the abandoned borehole, the borehole can only support 10 psi
buildup before fluid migrates upward (Figure 13, Case No. 3).
In all cases where there was sufficent reservoir data available to
model pressure buildup at the leaking abandoned well, the reservoir
-189-
-------
pressure buildup exceeded the calculated pressure resistance for 9
and 10 Ibs/gal mud systems (Figure 13).
In nearly all 28 cases cited by EPA (1975), AIC (1987a) found that
•*
records pertaining to cement and/or mud plugs in the leaking wells
were inadequate, incomplete, or non-existent. Plugging with mud was
more common than plugging with cement, but in either case, details
on the mud weight ("heavy") and cement (amount and location of
plugs) are usually not given. If this information is unavailable,
then conservative values should be used in modeling (9 Ibs/gal mud
and no cement).
Two other important mechanisms that are related to reservoir
modeling include well depth and distance from leaking well to
suspect injection well. Figure 14 shows that the average depth for
a leaking well in this case study is less than 2500 feet. Figure 15
shows that the maximum reported distance from a leaking well to a
suspect Class II injection well is less than 6000 feet and the
average is less than 2000 feet. This is consistent with reservoir
modeling where greater formation pressure buildup is associated
closer to the injection well.
CASE 2
A second study also conducted by AIC (1987b) involved the
examination of proper plug hearing files in selected Gulf Coast
counties. Proper plug hearings are called by the TRC "when it comes
-190-
-------
to their attention that a well has been abandoned or is not being
operated and is causing or likely to cause pollution to freshwater
above or below the below the ground or if gas or oil is escaping
from the well, the commission shall determine at a hearing, after
due notice, whether or not the well was properly plugged." These
hearings are called under Statewide Rule 14 (b) (2) of the "Texas
Statewide Rules For Oil, Gas, and Geothermal Operations."
This study was undertaken to determine the magnitude and mechanisms
of pollution problems associated with improperly abandoned wells in
unconsolidated sediments. From six selected counties along the Gulf
Coastal Plain (Figure 16), 2531 oil and gas fields were examined.
From these fields, 171 proper plug hearing orders were identified,
only three involved actual leakage incidents of which only two were
directly related to an injection well (Figure 17). These three
pollution incidents were examined to verify the factors that caused
the abandoned wells to leak.
Pollution incident No. 1 consisted of three wells on one lease that
were in violation of proper plugging. Subsequent field investiga-
tions by the TRC revealed that surface pollution existed but was not
the result of upward migrating fluids. Oil found in a pit near one
well was leaking from a 250-barrel tank. Operator negligence was
cited.
Pollution incidents Nos. 2 and 3 were the result of upward migration
of fluids due to subsurface injection of Class II wells in San
-191-
-------
Patricio County. Incident No. 2 involved an improperly abandoned
production well leaking oil to the surface. This well had been
drilled to 2590 feet. The well was abandoned with 885 feet of 8-5/8
inch surface casing, 2444 feet of 5-1/2 inch casing, and 2316 feet
of 2-inch production tubing in the hole. The 5-1/2 inch casing was
plugged back to 2345 feet and perforations were noted from 2446 to
2590 feet. The 2-inch tubing was cemented to the surface and
mud-laden fluid was pumped into the well along with a 25 sack-cement
plug (set at an unknown depth).
A suspect injection well was located approximately 2550 feet from
the leaking well. This suspect well was probably not a likely cause
of the pollution because its injection interval (5128 to 5132 feet)
is far below the producing interval (2446 to 2590 feet). In addi-
tion, the leaking well never penetrated the injection interval. Oil
migration has probably been the result of natural fluid migration
from the production zone through the improperly abandoned production
well.
Pollution incident No. 3 involved another improperly abandoned
production well, cited for leaking oil and water to the surface from
the thread of a "home-made" cap on the 5-1/2 inch casing. The well
was abandoned with 210 feet of 8-5/8 inch surface casing, 1358 feet
of 5-1/2 inch production casing, and 1355 feet of 2-inch tubing in
the hole. No records of cement were found on this well indicating
that it was ever plugged. The well was completed from 1331 to 1337
feet. A suspect injection well was located approximately 1300 feet
-192-
-------
away and the injection interval was from 1110 to 1155 feet. The
suspect well was permitted to operate at an average of 300 bbl/day
with maximum surface pressure of 30 psi.
Both pollution incidents Nos. 2 and 3 involved actual upward
migration of fluids and had protection/production strings left in
the hole, eliminating any possibility of borehole closure.
It is important to note that out of 2531 fields examined (the number
of abandoned wells may exceed the number of fields by a factor of
ten) along the Gulf Coast, only two leakage incidents were found.
This case study confirmed that the number of pollution problems in
the unconsolidated rock areas is small and indicates that natural
borehole closure is an important mechanism in eliminating upward
fluid migration.
CASE 3
To enhance our understanding and defend the conclusions of the
second study, a third study was conducted of proper plug hearings
for pollution incidents in "hard" and "soft" regions in Texas (AIC,
1987c). TRC Districts 7-B, 7-C, and 9 were selected as the "hard
rock" area and Districts 2, 3, and 4 comprised the "soft rock" area
(Figure 18). Districts were chosen primarily for their rock
environment and large number of oil and gas fields (i.e., production
wells).
-193-
-------
According to Anzzolin and Graham (1984, citing A. D. Little), 95% of
all production wells and 78% of all abandoned wells (Anzzolin and
Graham, 1984) fall within the AOR of Class II injection operations.
Accordingly, because each district contains a substantial number of
oil and gas fields, we can assume that a significant number of Class
II wells exist in each region studied. The study concluded that
pollution incidents resulting from Class II injection operations in
"hard rock" areas outnumber those cited in "soft rock" areas by a
factor of 10. Our conclusions are explained in the following
paragraphs.
Proper plug hearing files for 12,461 oil and gas fields in the
"consolidated rock" area were studied for pollution incidents (AIC,
1987c). Seven hundred and ninety (790) hearing files were located,
and further examination of these files found that 112 hearings were
called as the result of fluid migration from improperly abandoned
wells (Figure 19).
On the other hand, hearing files for 34,512 oil and gas fields in
the unconsolidated area were studied for leakage incidents. Six
hundred, seventy-four (674) hearings were found and only 16
indicated fluid migration. Nearly three times as many fields were
examined in unconsolidated rock areas as compared to consolidated
rock areas, but only 13% (16) of the 128 proper plug hearings from
both areas resulted from upward fluid migration in unconsolidated
rock.
-194-
-------
The 16 unconsolidated rock pollution incidents were studied to
determine the factors which caused the abandoned wells to leak.
Fourteen of the pollution incidents involved wells abandoned with
production casing left in the hole; two pollution incidents had
incomplete or nonexistent records.
It is important to note that all sixteen unconsolidated rock
incidents (leaking wells) were once production wells, and most, if
not all, were completed or abandoned with production casing intact.
In turn, by improper cementing across production intervals, improper
abandonment, or both, these wells were left open to upward migrating
fluids. Thus, natural borehole closure, common in the Gulf Coastal
Plain or unconsolidated rock areas, was restricted because of
production casing left open to the injection zone.
Regarding the 112 pollution incidents in "hard rock" regions, AIC
(1987c) noted that the producing zones were much shallower than in
"soft rock" areas. Abandoned wells in "hard rock" areas would tend
to have smaller hydrostatic heads due to the shorter static mud
column. Thus, pressure differentials between injection or
production intervals and static mud columns are small and more
likely to allow upward fluid migration than deeper injection or
production zones in "soft rock". "Hard rock" areas accounted for
87% of the total 128 leakage incidents resulting from upward fluid
migration.
-195-
-------
CONCLUSIONS
Case studies of Class II injection wells from the Texas Railroad
Commission files showed that only a small number of pollution
problems from leaking abandoned wells are associated with the
Texas Gulf Coastal Plain. These studies also documented natural
borehole closure as an important mechanism in preventing upward
fluid migration in the unconsolidated rock of the Texas Gulf Coastal
Plain.
The most important factors providing potential for upward
fluid migration due to injection operations in the unconsolidated
rock regions are: 1) production wells which had protection or pro-
duction casings left in the hole left open to the injection zone,
eliminating any possibility of borehole closure; and 2) signi-
ficantly overpressurized injection zones because of low reservoir
t ransmi ss ivi ty.
The case studies for west Texas (consolidated rock) indicate
a higher percentage of pollution incidents resulting from improperly
abandoned wells. The important factors relating to upward migration
are: 1) boreholes abandoned with or without casing remaining open
to the injection zone, 2) significantly overpressurized injection
zones because of low transmissivity, and 3) shallower production or
injection zones resulting in shorter static mud columns to counter-
balance increased formation pressure.
-196-
-------
This study of case histories has shown that all of the leaking
abandoned wells could have been identified as potential problem
wells. Preventive measures could have been taken prior to injection
operation. We believe operators can achieve responsible compliance
through the use of historical records and reservoir modeling to
conduct injection operations in a manner that protects the
environment.
-197-
-------
TABLE 1
MUD WEIGHT PRESSURE RESISTANCE
Assuming 9.0 Ibs/gal mud and formation pressure gradient of 0.45
psi/ft:
Depth
(ft)
1000
2000
3000
4000
5000
Hydrostatic
mud pressure
(psi)
468
936
1404
1872
2340
Formation
pressure
(psi)
450
900
1350
1800
2250
Pr
di
(P
18
36
54
72
90
differential
-198-
-------
PROTOCOL FOR IDENTIFYING
RBflNDONED WELLS IN FM flRER OF REVIEW
SERRCH DHTfl
SOURCES
STRTEX
FEDERRL
RECORDS
IWtRNFL
OOCUKMT5
corn.
LOG.
CO. 'S
OIL
COMPPNY
RECORDS
IDENTIFY WELLS IN RRER OF REVIEW
_L
PENETRRTE CONFINING -' INJECTION ZONE
FORMflTION TYPE
UNCONSOLIDHTEDXINDURRTED
DRILLING METHOD
ROTRRY/CRBLE
PLUGGING RECOfiDS
RVRILHBLE
SERRCH CE^E^^•,
INJECTION RNIM3R
PRODUCTION RECORDS
SEflRCH CEJCNT.
INJECTION HND^OR
PRODUCTION RECORDS
_L
PLUGGED
rwjttKLT
Y
OK
MR»ET<*ETER
POTENTIRL UPWRRD MIGRHTION
LOCRTE HBRNDONED
1
MFP WELL LOG
COORDINHTES COORDINHTES
WELL
1
NO UPWRRD MIGRHTION
METHL
DETECTOR
1
OK
SURVEYOR
OTHER
WELL LOCRTED
N N
LOWER INJECTION
ZONE
LOWER INJECTION
RHTE
Figure 1
-199-
-------
POTENTIRL PRTHS OF FLUID MIGRRTION
FROM CLRSS II INJECTION WELLS
B
EXPLRNRTION
fl - CLRSS II INJECTION WELL
B - PRODUCTION WELL - COMPLETED IN
DEEPER ZONE, RNNULUS PRRTIRLLY
UNCEMENTED TO SURFRCE, FORMRTION
PRESSURE » STRTIC RNNULRR MUD
COLUMN
C - IMPROPERLY PLUGGED RND RBRNDONED
DRY HOLE - PENETRRTING THE
INJECTION / PRODUCTION ZONE
FORMRTION PRESSURE » STRTIC
FLUID COLUMN
D - IMPROPERLY RBRNDONED PRODUCTION
WELL - DEPLETED WELL WITH
PRODUCTION STRING LEFT OPEN TO
INJECTION ZONE, NO CEMENT OR
MUD PLUGS
i
o
o
CN
SURFRCE COSING
CEMENT
ION
PERFORATIONS
CEMENT
DIRECTION OF
FLUID MOVEMENT
Figure 2
-------
Stepwise Dehydration of Clay
Gulf Coast Well, Chambers County, Texas
in Mixed Layer Components
From Burst, 1969)
w «.v -,0 60 80 100
o
I—'
I
10,000
15,000
©
i
• Zone « .
^*
Closure
Top Anahuac
* Top Frlo
• Top
Geopressure
10,000
Restricted
Dehydration
Zone
Plastic
Borehole
Closure
±
15,000
Figure 3
-------
CONSOLIDATED AND UNCONSOLIDATED
BOCK TYPES IN TEXAS
P*.5"S;!«fl5?*^;.T
•&#^^#:£;;#3$
• SA!
:«!?£:-ff7J
•&::':$&£:V'
£$££$•$!
^•Syv-Stt*
M
;•?(?:%
^fe^iiwffiwiih.^v
*!
8
1
IT
'sa^
LEGEND
Railroad Commission
District Numbers
Consolidated Sediments
("Hard Rock")
Unconsolldated Sediments
("Soft Rock")
Figure 4
-202-
-------
5000T-
4500--
4000--
MUD COLUMN PRESSURE VS. DEPTH
3500--
P
r
e 3000-
s
s
u
r
e
2500--
2000--
1500--
1000--
500--
500 1000 1500 2000 2500 3000 3500 4000 4500 5000
Depth, feet
9 Ibs/gal mud
10 Ibs/gal mud
11 Ibs/gal mud
Figure 5
-203-
-------
FORMATION PRESSURE VS. DEPTH
5000T
4500--
4000--
3500-
P
r
e 3000-
s
s
u
r
e
2500--
2000--
1500-
1000-
500--
500 1000 1500 2000 2500 3000 3500 4000 4500 5000
Depth, feet
0.45 psi/ft
0.47 psi/ft
Figure 6
-204-
-------
PRESSURE DIFFERENTIAL BASED ON MUD WEIGHT
AND FORMATION GRADIENT
400-r
D
i
f
f
e
r
e
n
t
i
a
1
P
r
e
s
s
u
r
e
P
s
i
300--
200--
100--
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
Depth, feet
10 Ibs/gal mud & 0.45 gradient
10 Ibs/gal mud & 0.46 gradient
9 Ibs/gal mud & 0.45 gradient
9 Ibs/gal mud & 0.46 gradient
Figure 7
-205-
-------
GEL STRENGTH VS. PRESSURE RESISTANCE
Pressure
resistance (psi)
per 500 ft depth
35T-
30-
25-
20--
15--
10--
20 40 60 80 100 120
Gel Strength (Gs), lbs/100 sq. ft.
6 inch hole diameter
8 inch hole diameter
10 inch hole diameter
• 13 inch hole diameter
Figure 8
-206-
-------
700-r
P
S 600--
D 500-
I
F 4QO--
F
I 300--
E
N 200--
T
I
A
L
0-
100
1000
DISTANCE IN FEET FROM INJECTION WELL
10000
Q » 10 gpm, T
Q - 100 gpm, T
100 gpd/ft
100 gpd/ft
250-r
1000
DISTANCE IN FEET FROM INJECTION WELL
10000
Q =. 600 gpm, T
Q = 1200 gpm, T
3700 gpd/ft
3700 gpd/ft
Injection Zone Pressure Buildup After 30 Years vs. Distance
and Relationship Between Transmissivity and Injection Rates.
Figure 9
-207-
-------
AREA OF REVIEW CALCULATION
o
00
p
R 400-p
E
S
S
U
R
E
I
F
F
E
R
N
T
I
A
L
P
S
350- =
300--
250+
200-
150 +
100--
50 +
Pressure Buildup From Injection
100
1000
10000
DISTANCE IN FEET FROM INJECTION WELL
Mud pressure resistance, 10 Ibs/gal mud @ 5000 ft
Mud pressure resistance, 10 Ibs/gal mud @ 2000 ft
Mud pressure resistance, ° IK,-/~,I -..^ a inAn **-
Mud pressure resistance,
9 Ibs/gal mud @ 2000 ft
9 Ibs/gal mud @ 2000 ft
Area of Review Calculation Based on Formation Gradient = 0.45 psi/ft,
9 Ibs/gal Mud at 5000 ft Depth, and an Injection Zone where Q = 600 gpm
and T = 3700 gpd/ft. Other Weight Muds at Various Depths Shown.
Figure 10
-------
CRLCULRTED PRESSURE INCRERSE DISTRIBUTION
RRER OF REVIEW CRLCULRTION
O
^>
I
EXPLRNHTIQN
CRLCULRTED RRER OF REVIEW
•60 PSI
Figure 11
Q - 600 gpm
T • 3700 gpd/-ft
(30 yr. Injection)
9 Ibs/gal MUD ® 5000 FT
-------
Cast 1
CASE HISTORIES RESEARCHED
FROM TEXAS RAILROAD COMMISSION FILES
r
•
YOUNQ
COOKE
JACK / STEPHENS
COLEMAN
COMANCHE
HARRIS
DUVAL
LEGEND
1 • C«»» Hlitorl** R«**>rch*d
Figure 12
-210-
-------
CASE 1
RESERVOIR PRESSURE BUILDUP VS. MUD PRESSURE RESISTANCE
I
ho
P
R 800
E
S
g 700-
u
600-
E
D 500
I
F
F 400
E
R
£ 300-
N
200-
A
L 100-
P
Sn
0_j — ,
I
170
15
A^jP
1
70
^
50
250
37
16
s\s*
3* 7
63
%
500
50
$
8
19C
I
800
600
224
I
60 V>
\Q< //
13
15 16
500
50
42
cv
vv
I'Si^
17
165
/y
400
250
140
/; 90
36 // ,2 /V
^
-------
CASE 1
TOTAL DEPTH OF LEAKING WELLS VS. NUMBER OF OCCURRENCES
I
I—'
N3
Occurrences
Figure 14
-------
CASE 1
DISTANCE FROM LEAKING WELLS TO SUSPECT CLASS II WELLS
i
NJ
00
I
4501-5000
5001-5500 >5
£•£
5501-6000 $
10
12
14
16
18
20
22
Occurrences
Figure 15
-------
Case 2
PROPER PLUGGING HEARING SURVEY
SELECTED GULF COAST COUNTIES
Flgure 16
-214-
-------
CASE 2
NUMBER OF FIELDS EXAMINED, PROPER PLUG HEARINGS, AND
POLLUTION INCIDENTS REPORTEDLY CAUSED BY CLASS II INJECTION
1000-r
800--
600 —
t_n
I
400--
200--
887
260
222
47
19
554
405
40
Harris Jefferson
Nueces
San Patricio Victoria
County
Figure 17
fields
examined
(2,531)
Proper Plug
Hearings
(171)
leakage
incidents
(2)
-------
CaM 3
PROPER PLUGGING HEARING SURVEY
SELECTED RAILROAD COMMISSION DISTRICTS
LEGEND
Railroad Commission
District Numbers
Consolidated Sediments
("Hard Rock")
Unconsolldated Sediments
("Soft Rock")
Figure 18
-216-
-------
CASE 3
CONSOLIDATED VS. UNCONSOLIDATED FORMATIONS
i
to
35000-
30000-
25000-
0
c
c
u 20000-
r
r '.
e
n 15000-
c !
e
s
10000-
34.512
12,461
5000-
112
16
Consolidated Unconsolidated
formations
Oil & gas fields
Proper Plug Hearings
(PPH)
PPH's with
well bore leaks
Figure 19
-------
REFERENCES
AIC (Agency Information Consultants, Inc.), 1987a, Survey of Cited
EPA Problem Leaking Wells in Texas: Prepared for E. I.
Du Pont.
AIC (Agency Information Consultants, Inc.), 1987b, Survey of
Pollution Abatement Hearings for Selected Counties Along the
Texas Gulf Coast: Prepared for E. I. Du Pont.
AIC (Agency Information Consultants, Inc.), 1987c, Survey of Proper
Plugging Hearings for Fluid Migration from Unplugged or
Improperly Plugged Wells in Texas Railroad Commission
Districts 02, 03, 04, 07B, 07C, and 09: Prepared for E. I.
Du Pont.
Alford, S. E., 1987, Conoco, Senior Drilling Engineer (drilling mud
specialist), Houston, TX; personal communication.
Ammons, C. T., 1987, Conoco, Drilling Engineer, Lafayette, LA;
personal communication.
Anzzolin, A. R., and Graham, L. L., 1984, Abandoned Wells—A
Regulatory Perspective, in D. M. Fairchild, ed., Proceedings
of the First National Conference on Abandoned Wells: Problems
and Solutions: Environmental and Ground Water Institute,
University of Oklahoma, Norman, OK, p. 17-36.
Barker, S. E., 1981, Determining the Area of Review for Industrial
Waste Disposal Wells: Master's Thesis, The University of
Texas at Austin, Austin, TX, 146 p.
Burst, J. F., 1959, Postdiagenetic Clay Mineral Environmental
Relationships in the Gulf Coast Eocene, in A. Swineford, ed.,
-218-
-------
Clays and Clay Minerals: 6th National Clays and Clay Mineral
Conference Proceedings, Pergamon Press, 411 p.
Burst, J. F., 1969, Diagenesis of Gulf Coast Clayey Sediments and
Its Possible Relation to Petroleum Migration: American
Association of Petroleum Geologists Bulletin, v. 53, p. 73-93.
Cheatham, Jr., J. B., 1984, Wellbore Stability: Journal of Petroleum
Technology, V. 36, p. 889-896.
Collins, R. E., 1986, Technical Basis for Area of Review: Prepared
for Chemical Manufacturers Association, 112 p.
Cooke, Jr., C. E., Kluck, M. P., and Medrano, R., 1983, Field
Measurement of Annular Pressure and Temperature During Primary
Cementing: Journal of Petroleum Technology, V. 35, p. 1429-
1438.
Cooke, Jr., C. E., Kluck, M. P., and Medrano, R., 1984, Annular
Pressure and Temperature Measurements Diagnose Cementing
Operations: Journal of Petroleum Technology, v. 36, p.
2181-2186.
Darley, H. C. H., 1969, A Laboratory Investigation of Borehole
Stability: Journal of Petroleum Technology, v. 21, p.
883-892.
Davis, K. E., 1986, Factors Effecting the Area of Review for
Hazardous Waste Disposal Wells: Proceedings of the
International Symposium on Subsurface Injection of Liquid
Wastes, National Water Well Association, Dublin, OH, p. 148-
194.
Eikel, B. C., 1969, Assistant District Director, Railroad Commission
of Texas, letter of August, 20, 1969 to R. D. Payne, Director
-219-
-------
of Field Operations, Railroad Commission of Texas: Railroad
Commission of Texas file 00000101834.
Engineering Enterprises, Inc., 1985, Guidance Document for the Area
of Review Requirement: Norman, OK, prepared for EPA.
EPA, 1975, Proposed Injection Well Regulations for Brine Produced
with Oil or Gas: US EPA Document from J. T. Thornhill to E.
Hockman, 24 p.
Fryberger, J. S., and Tinlin, R. M., 1984, Pollution Potential from
Injection Wells via Abandoned Wells, in D. M. Fairchild, ed.,
Proceedings of the First National Conference on Abandoned
Wells: Problems and Solutions: Environmental and Ground
Water Institute, University of Oklahoma, Norman, OK, p.
84-117.
Garrison, A. D., 1939, Surface Chemistry of Clays and Shales:
Petroleum Transactions of AIME, v. 132, p. 191-203.
Gray, G. D., Darley, H. C., and Rogers, W. F., 1980, Composition and
Properties of Oil Well Drilling Fluids: Houston, Gulf
Publishing.
Grim, R. E., 1968, Clay Mineralogy (2nd ed.): New York, McGraw-Hill,
596 p.
Gurke, R., 1987, Halliburton Service Training Course, Duncan, OK,
personal communication.
Hiller, K. H., 1963, Rheological Measurements on Clay Suspensions
and Drilling Fluids at High Temperatures and Pressures:
Journal of Petroleum Technology, v. 15, p. 779-789.
Johnston, 0., and Green, C. J., 1979, Investigation of Artificial
Penetrations in the Vicinity of Subsurface Disposal Wells:
Texas Department of Water Resources.
-220-
-------
Johnston, 0. C., and Knape, B. K., 1986, Pressure Effects of the
Static Mud Column in Abandoned Wells: Texas Water Commission
LP86-06, 99 p.
Klotzman, 1986, Consulting Geologist; Concerning Plugging Abandoned
Wells Near Victoria, TX; personal communication.
Krusekopy, Jr., H. H., 1970, Geologist, Railroad Commission of Texas
letter of January 22, 1970 to R. D. Payne, Director of Field
Operations, Railroad Commission of Texas: Texas Railroad
Commission file 00000300113.
Meers, R. J., 1987, Petroleum Consultant; Concerning Plugging
Abandoned Wells Near Orange, TX; personal communication.
Polk, G., and Gray, G. R., 1984, Plugging Mineral Exploration Holes
with a Drilling Fluid Conditioner, in D. M. Fairchild, ed.,
Proceedings of the First National Conference on Abandoned
Wells: Problems and Solutions: Environmental and Ground Water
Institute, University of Oklahoma, Norman, OK, p. 295-302.
Powers, M. C., 1967, Fluid-release Mechanisms in Compacting Marine
Mudrocks and Their Importance in Oil Exploration: American
Association of Petroleum Geologists Bulletin, v. 51, p.
1240-1254.
Price, W. H., 1971, The Determination of Maximum Injection Pressure
for Effluent Disposal Wells, Houston, Texas area: Master's
Thesis, The University of Texas at Austin, Austin, TX, 84 p.
Ross, C. C., and Steed, W. C., 1984, Well Plugging in Texas, in
D. M. Fairchild, ed., Proceedings of the First National
Conference on Abandoned Wells: Problems and Solutions:
-221-
-------
Environmental and Ground Water Institute, University of
Oklahoma, Norman, OK, p. 251-270.
Roth, T., 1987, Head of UIC Program (Class I) for State of Texas;
Concerning Number of Class II Injection Wells; personal
communication.
-222-
-------
Biographical Sketches
James E. Clark holds a B.S. in geology (1972) from Auburn University and
an M.S. in geophysical sciences (1977) from Georgia Institute of Technology.
As a geohydrologist with Law Engineering Testing Co., he worked on suitability
studies of salt domes as repositories for nuclear waste. He is a consultant
with Du Font's (E. I. du Pont de Nemours & Co., Inc., Engineering Department,
P. 0. Box 3269, Beaumont, TX 77704) solid waste and geological engineering
group and is active in permitting and evaluation of disposal wells.
Milton R. Howard received his B.S. degree in geology from Texas A&M
University (1985). He served as a petroleum geologist for SOHIO and Albaine,
active in on-shore database evaluation and oil and gas exploration. In 1985
he joined the waste and geological engineering group of Du Pont as a contract
consulting environmental geologist responsible for permitting and evaluation
of the Federal UIC Class I disposal wells.
Diane K. Sparks received her B.S. degree (1977) in geology and her M.S.
degree (1978) in geology from Bowling Green State University. She was a
petroleum geologist with Amoco Production Company and Helmerich and Payne,
Inc. Sparks is now a consulting geologist and currently works as a contract
geologist for the Engineering Service Division of Du Pont, in evaluation of
Class I disposal wells and fluid migration studies.
-223-
-------
Sources of Ground-Water Sal1n1zat1on in Parts of West Texas, U.S.
Bernd C. Richter and Charles W. Kreitler
Bureau of Economic Geology, The University of Texas at Austin,
University Station, Box X, Austin, Texas, USA 78713-7508
Acknowledgments
Funding for this project was provided by the Railroad Commission of Texas under
contract no. IAC(84-85)-2122. Appreciation is expressed to Railroad Commission of
Texas personnel at District 7-C in San Angelo, Texas, and to many individuals in
Tom Green, Runnels, and Concho Counties for assistance during data collection.
Tonia J. Clement assisted in data preparation. The manuscript was reviewed by
Jules R. DuBar and Alan R. Dutton, Bureau of Economic Geology, The University of
Texas at Austin. Figures were drafted under the supervision of Richard L. Dillon,
Bureau of Economic Geology.
Abstract
Determination of chemical constituent ratios allows distinction between two
salinization mechanisms responsible for shallow saline ground water and vegetative-
kill areas in parts of West Texas. Mixing of deep-basin salt water and shallow
fresh ground water results in saline waters with relatively low Ca/Cl,
*Publication authorized by the director, Bureau of Economic Geology, The University
of Texas at Austin.
-224-
-------
Mg/Cl, S04/C1, Br/CI, and N03/C1 ratios. In scattergrams of major chemical
constituents versus chloride, plots of these waters indicate trends having brine
values as high-Cl end members. Evaporation of ground water from a shallow water
table, in contrast, results in saline water that has relatively high Ca/Cl, Mg/Cl,
S04/C1, and Br/CI ratios. Trends indicated by plots of this water type do not
coincide with trends indicated by plots of sampled brines. Leaching of cultivation
nitrate in areas with a shallow water table accounts for high N03 concentrations in
shallow ground water.
Introduction
Salinization of soil and shallow ground water and the appearance of vegetative-
kill areas are major concerns of farmers in parts of Texas and in other
agriculturally important areas in the United States. In many parts of the country
natural and agricultural factors are responsible for salinization. In Texas,
pollution hazards associated with the exploration and production of oil are
additional possible sources of salt water. These hazards complicate the problem of
determining the sources of soil and ground-water contamination.
Residents of Tom Green, Runnels, and Concho Counties in West Texas (Figure 1)
blame oil-field-related activities for widespread contamination. They point out
that (1) water was of better quality before drilling for oil began and (2) locally,
formerly productive land has become so salty that plant growth is limited or has
ceased. Many cases of oil-field-related water and soil pollution, caused by brine
flow from abandoned holes and leaky injection wells, are known in the area.
Unknown, however, is the areal extent of contamination that has occurred or is
occurring from thousands of oil wells, core holes, shot holes, and injection wells
and from the use of open pits for brine disposal, a practice which was abandoned in
the late 1960's. The area is underlain by an artesian brine aquifer (Coleman
-225-
-------
Junction aquifer, Table 1) that flows to land surface where pathways are provided
and that stratigraphically overlies most of the major oil-producing horizons in the
area. Therefore, most holes drilled for oil penetrate this aquifer and thus create
a potential, artificial pathway for brine movement into shallow fresh ground water
or to land surface.
Researchers (for example, Reed [1962] and Marshall [1976]) claim that a
combination of natural conditions and inappropriate agricultural and water-well
drilling techniques is responsible for salinization of soils and ground water in
the area. During severe droughts in the 1950's many water wells that had run dry
were deepened until saline water was encountered (Marshall, 1976). Many of these
wells have not been plugged (Marshall, 1976) and therefore constitute a possible
source of ground-water pollution. At about the same time, extensive closed-contour
terracing of land and destruction of former drainage networks began in the area in
an attempt to reduce surface runoff. Unusual heavy rainfalls in the early 1960's
following the droughts of the 1950's and the practice of land terracing have had
the combined effect of gradually raising the water table closer to land surface
during the last 30 years. Today, ground water stands at or within a few feet of
land surface in many topographically low localities in the eastern part of the
area, causing waterlogging and subsequent salinization of vadose and ground waters
owing to evaporation. Salts that precipitate in the soil during this process
inhibit growth of non-salt-resistant plants and are dissolved and flushed into
ground water after rainfall, thus spreading the pollution hazards to other areas.
These processes occur in the absence of any oil field activity or artesian brine
aquifers, as evidenced by hundreds of thousands of acres affected throughout the
Great Plains from Texas to Montana (Miller et al., 1981).
-226-
-------
Purpose
This study was designed to discover whether certain hydrochemical methods, such
as determination of Na/Cl, Br/Cl, and I/Cl ratios, could differentiate surface
salinization caused by evaporation from a shallow water table from surface
salinization caused by discharge (natural or man-made) of deep-subsurface brine
aquifers. The study was conducted from January 1 through April 30, 1985. Water
samples were collected from water-supply, oil, and injection wells for chemical and
isotopic analyses designed to establish the chemical characteristics of ground
water in the area.
Geologic Setting
The study area is underlain by Permian to Quaternary sediments (Figure 1).
Cretaceous rocks, which consist of argillaceous limestone, form topographic highs
that border the study area in southern, western, and northern Tom Green County.
southern Concho County, and northeastern Runnels County. Pleistocene and Recent
alluvial deposits of variable thickness directly overlie Permian strata in central
and eastern Tom Green County and parts of Runnels and Concho Counties. Permian
strata crop out in north-south-trending belts in central Tom Green and northern
Concho Counties and are scattered throughout Runnels County. Permian strata dip to
the west and northwest at approximately 50 ft/mi (10 m/km) and include sandstone,
limestone, shale, gypsum, and dolomite beds (Willis, 1954).
Thousands of oil wells have been drilled in the area since oil exploration
started at the end of the last century. Most oil production is from Pennsylvania*!
strata at depths greater than 3,000 ft (915 m) in the western part of the area and
greater than 2,000 ft (610 m) in the eastern part of the area. Some production is
from shallow depths from the San Angelo Formation (Table 1), approximately 1,000 ft
(305 m) below land surface, in southwestern Tom Green County. Oil has been
-227-
-------
encountered in wells within 50 to 300 ft (15 to 90 m) of land surface in the San
Angelo area in western Tom Green County (Udden and Phillips, 1911).
Hydrogeologlc Setting
Cretaceous limestones and Quaternary alluvial deposits form principal aquifer
units bordering the area (Table 1). In the remainder of the three counties, no
extensive, major fresh-water aquifers are present at shallow depths. Local supplies
of potable water are found in outcrops of Permian limestone and gypsum. However,
the quality and quantity of ground water is very erratic in these units. Many dry
holes have been drilled in the immediate vicinity of high-capacity wells. At one
location in northern Concho County a 100-ft (30-m) deep dry hole was drilled just
20 inches (50 cm) from a flowing well of the same depth, which indicates that
ground water flows through solution channels or fractures in that area.
Saline water is encountered downdip of potable water supplies in outcrops of
Permian strata. Highly mineralized water occurs under artesian pressure and at
shallow depths in the Permian San Angelo and Blaine Formations (Table 1) of west-
central Tom Green County (Udden and Phillips, 1911; Willis, 1954). The brine
aquifer in the Permian Coleman Junction underlies the area at depths between
3,000 ft (915 m) in the southwest and 800 ft (245 m) in the east. Brine has the
potential to flow to land surface from this aquifer via natural and artificial
pathways, with surface-casing pressures exceeding 100 psi in individual wells in
Runnels County (Raschke and Seaman, 1976).
Water levels in eastern Tom Green County have generally increased during the
last 30 years but remain 50 ft (15 m) or more below land surface. In contrast, in
southern Runnels County water levels approach land surface in many wells, causing
seepage of ground water at topographically low areas.
-228-
-------
Marshall (1976) reported that probably hundreds of water wells were drilled
down to depths of 500 ft (150 m) west of the city of San Angelo during the drought
in the 1950's, and, although these wells encountered highly mineralized water, many
of them were not plugged. These water wells create a pollution hazard by allowing
saline water to mix with potable water resources (Marshall, 1976).
Geochemical Approach
In a study of salt-water sources in north-central Texas, Richter and Kreitler
(1986) showed that differences in ratios of Na/Cl, Br/Cl, I/C1, Mg/Cl, K/C1, and
(Ca+Mg)/S04 indicate two salt-water types. (1) Salt water derived from dissolution
of halite by fresh water relatively close to land surface is characterized by Na/Cl
and (Ca+Mg)/S04 molar ratios of approximately 1, and by low Mg/Cl, K/C1, Br/Cl, and
I/C1 ratios. (2) Salt-water derived from deep-basin brines is characterized by
Na/Cl ratios of less than 1, (Ca+Mg)/S04 molar ratios of greater than 1, and high
Mg/Cl, K/C1, Br/Cl, and I/C1 ratios. This differentiation worked especially well at
concentrations of greater than 10,000 mg/L of total dissolved solids. In addition,
stable isotopes of oxygen and hydrogen characterized halite-dissolution brine as
local, meteoric ground water. Deep-basin brine proved to be of nonlocal origin.
Two principal sources of saline water exist in Tom Green, Runnels, and Concho
Counties: deep-basin brines and agricultural salinization. Goals of the present
study were to obtain clear definitions of deep-basin brine characteristics and of
seep-water characteristics using the parameters previously mentioned. However, in
contrast to the study by Richter and Kreitler (1986), most of the polluted ground
waters in the area are of relatively low salinity (less than 5,000 mg/L) and
halite-dissolution brine is not present. Therefore, it was unknown how well these
ratios could be applied in this case.
-229-
-------
In the present study, nitrate was chosen as an additional possible tracer of
pollution sources. Shallow ground water in the Runnels County area typically
exhibits high concentrations of nitrate owing to dissolution of cultivation nitrate
by water recharging through the vadose zone {Kreitler, 1975). High nitrate
concentrations in shallow ground water are caused by changes in agricultural
practices in the area. Dryland farming prior to the 1950's had caused oxidation of
organic nitrogen to nitrate in the soil zone. Nitrate was leached below the root
zone by percolating ground water but was out of contact with the water table until
the late 1950's and early 1960's, when extensive terracing raised the water table
to within a few feet to land surface. The latter caused leaching of nitrate into
shallow ground water (Kreitler, 1975). Ground water at or slightly below land
surface in seep areas, therefore, could contain elevated nitrate concentrations.
Deep-basin brines, in contrast, normally do not contain appreciable amounts of
nitrate.
Brines in the area were expected to be isotopically enriched in oxygen and
deuterium with respect to fresh ground water. Evaporation of ground water from a
shallow water table also may result in an isotopic shift toward higher values.
Therefore, seep waters too were expected to be isotopically heavier than local
precipitation. The magnitude of the shift and the difference between brines and
seep water, however, were not known.
In addition to water sample data obtained from published and unpublished
sources, 46 samples were collected during this study: 39 from shallow water wells
and 7 from oil field wells and holes (Figure 1, Table 2). Five of the 39 samples
were obtained from shallow wells drilled in seep areas. Three of these were from
water wells and two from shallow holes drilled for this investigation.
To establish the characteristics of water types, sampling included (1) oil
wells, (2) a Coleman Junction well (3) allegedly polluted wells, (4) stock wells,
-230-
-------
(5) house wells, and (6) seep wells. Samples for chemical analyses were stored in
500 ml polyethylene bottles, and samples for isotope analyses were stored in 250 ml
glass bottles. During collection, samples were filtered using 0.45-micron membrane
filters and nitrogen gas to remove particulate matter. The filter bottle was
cleaned thoroughly between samples and checked for residual ion content, using
distilled water and silver nitrate, to prevent cross-contamination of water
samples.
Results
Data from previous investigations, when plotted on Piper diagrams, show that
ground water in Tom Green County is characterized by four chemical facies. At
chloride concentrations of less than 250 mg/L, Ca-Mg-HC03 water is the major facies
type (Figure 2). This type occurs predominantly in Cretaceous (limestone)
formations. Another facies type is Ca-Mg-S04 water, the result of dissolution of
gypsum or anhydrite in Permian strata. At chloride concentrations of greater than
250 mg/L, an increase in sodium and especially chloride percentages results in Ca-
Mg-Cl and Na-Cl waters (Figure 2). In Runnels County, cation percentages in ground
water are evenly distributed without a shift toward sodium dominance at chloride
concentrations of greater than 250 mg/L (Figure 2). Anions too are distributed
relatively evenly at chloride concentrations of less than 250 mg/L, but show a
shift toward the chloride apex at chloride concentrations of greater than 250 mg/L.
Therefore, at chloride concentrations of greater than 250 mg/L, Piper diagrams of
ground water in Tom Green and Runnels Counties indicate that different mechanisms
control the distribution of cations in ground water in the two counties.
Only 6 of 39 water samples collected during this study contain chloride
concentrations of less than 250 mg/L because emphasis was put on collection of
allegedly contaminated ground water. The configuration of data points from these
-231-
-------
samples within a Piper plot (Figure 3) is similar to the distribution of data
points in the plot of ground water in Tom Green County for chloride concentrations
of greater than 250 mg/L (Figure 2). Most of the samples collected during this
study are of the Ca-Mg-Cl or Na-Cl types. Within the cation triangle, a linear
trend between Ca-Mg-dominated ground water and Na-dominated brine is indicated.
On bivariate plots of Ca, Mg, $04, and Br/Cl versus Cl, evolution or mixing
trends are indicated that contain fresh water and brine as end members (Figure 4).
At high chloride concentrations, the plots of Mg and $04 versus Cl suggest that
possibly two trends exist, where one trend points toward brine values and is
relatively low in Mg and $04 and the other points away from brine values and is
relatively high in Mg and $04. Bromine and nitrate were the only minor chemical
constituents that were above detection limits both in the brine and in the ground-
water samples and that showed some differences between water samples. Ratios of
Br/Cl in brines underlying the area are lower than ratios in, for example, shallow
subsurface brines in the southern Rolling Plains of North-Central Texas and in oil
field brines of Kansas (Figure 4). In contrast, Br/Cl ratios in fresh water are
typical of this water type. With increasing chlorinity, Br/Cl ratios in ground
water in the area decrease to values similar to Br/Cl ratios in brines underlying
the area, possibly indicating a mixing trend between fresh ground water having high
Br/Cl ratios and brines having low Br/Cl ratios.
Nitrate concentrations range from less than 1 mg/L to more than 200 mg/L
(Table 2). Lowest concentrations were measured in ground water in western Tom Green
County and in central Runnels County, as well as in brines underlying the area.
Concentrations in excess of 100 mg/L prevail in northeastern Tom Green County,
southern Runnels County, and northern Concho County (Figure 5). Four of five seep
samples have nitrate concentrations of between 121 mg/L and 158 mg/L.
-232-
-------
A differentiation of salinization sources by use of stable isotopes of oxygen
and hydrogen was not possible from analyses obtained during this study. All
samples, regardless of chlorinity and geographic or stratigraphic origin, plot
within one cluster and indicate no apparent trends (Richter and Kreitler, 1985).
Mixing and dilution of waters from different sources may account for this
relationship between water types and isotopic composition.
Discussion
Water-table elevation is close to land surface in many topographically low
areas in Runnels, Concho, and eastern Tom Green Counties, whereas it is well below
land surface throughout western Tom Green County. Associated with a high water
table, saline seeps and vegetative-kill areas are widespread phenomena in the
eastern part of the area but are less frequent in the western part. Therefore,
salinization by evaporation should be more prevalent in the eastern part of the
area than in western Tom Green County.
Mixing between brine and fresh water seems to be indicated in 16 of the 39
water samples, as suggested by ratios of major chemical constituents in ground-
water samples when compared with ratios typical of sampled brines in the area
(Table 2). Of these 16 samples only 4 were obtained from Runnels County, Concho
County, and eastern Tom Green County, whereas 12 were obtained from western Tom
Green County. The remaining 23 water samples, which include only 3 from western Tom
Green County, do not indicate any similarity with brines underlying the area.
Grouping of the data according to sample location (east versus west) breaks up
the cluster and the tentatively suggested trends of Figure 4 into two, fairly well-
defined trends (Figure 6). Trend 1, characterized by high Ca, Mg, and $04
concentrations, is made up mainly of samples from Runnels, Concho, and eastern Tom
Green Counties. This trend does not include values typical of brines in the area,
-233-
-------
which indicates that a salinization mechanism other than mixing of fresh ground
water and deep-basin brine is responsible for increases in salinity in ground water
of the area. Possible mechanisms are (1) evolution of ground water through mineral
dissolution, (2) mixing of different ground-water types, and (3) evaporation from a
shallow water table. Trend 1 approaches a slope of one in the bivariate plots of
molar concentrations, which eliminates the possibility of mineral dissolution as
the most dominant salinization mechanism. During evaporation, the molar ratios of
chemical constituents stay constant in absence of precipitation or dissolution
reactions. Also, the relative position of cation percentages in a Piper diagram
does not change during evaporation, which is suggested for water samples from the
east (Figure 7). In contrast, two water types would be expected to plot within two
discrete clusters in a Piper plot, where mixing would be indicated by a trend that
connects the two clusters. Although the possibility cannot be dismissed that two
nonrelated waters fall within the same cluster, it seems most likely that
evaporation is the mechanism that accounts for the trends observed in the Piper
plot and in the bivariate plots for waters from the eastern part of the area.
Trend 2 is made up of samples low in Ca, Mg, and $04 and is represented by samples
obtained mainly from the western part of Tom Green County. This trend includes
values of Coleman Junction and oil field brines as high-chloride end members,
suggesting mixing of fresh ground water and brine rather than evolution of fresh
ground water to a brine through water-mineral reactions. At low concentrations of
dissolved chemical constituents, the two trends overlap and do not allow
differentiation of salinity sources. As chloride increases, the trends increasingly
deviate from each other, making it possible to determine salt-water sources.
Seep samples, although not indicating mixing of fresh ground water and brine in
any of their chemical constituents (Table 2), do not plot clearly within Trend 1
but within the zone of overlap between Trend 1 and Trend 2. Because sample
-234-
-------
collection was during early February, when the effects of evaporation are at their
lowest, seep samples are similar to other samples from the area. It can be expected
that seep samples collected during summer months will plot as high-chloride end
members of Trend 1. Four of the five seep samples (6, 7, 8, and 10) were obtained
from wells in topographically low areas where the water table was within a few feet
of land surface, indicating stagnant water- These samples have nitrate
concentrations in excess of 100 mg/L owing to dissolution of nitrate in the shallow
soil zone. Seep sample 11 was obtained from a flowing well that is used to drain
the seep area in an attempt by the owner to reclaim waterlogged land. According to
the owner, this well stops flowing whenever irrigation from nearby wells is
activated. Therefore, the sample from this well is part of an active ground-water
flow system (activated by the well), in contrast to a sluggish or stagnant ground-
water system at the other seeps. Continuous flushing of this particular flow path
may explain the low nitrate concentration of sample 11 when compared with other
seep samples.
Samples 2 and 4 are high-chloride waters that were obtained from wells in
central Runnels County. These samples consistently fall within Trend 2, which is
the trend of samples from western Tom Green County. This suggests that two sources
of salinity exist in the eastern part of the area. Most samples follow Trend 1, and
therefore evaporation seems to be the most dominant salinization mechanism. The
distribution of cations from samples in the east form one big cluster in a Piper
plot (Figure 7), similar to the cluster typical of Runnels County at chloride
concentrations of greater than 250 mg/L (Figure 3). This indicates that the samples
obtained during this study are representative of the area and that salinization
through evaporation is of widespread nature. In contrast, few samples from the east
follow Trend 2, suggesting that mixing between fresh ground water and brine is a
local phenomenon in the area. Samples 2 and 4 were obtained from abandoned water
-235-
-------
wells close to producing oil wells. Ratios of Br/Cl and concentrations of N03 in
both samples are very low and similar to brine values, which is atypical of ground
water in the Runnels County area. Therefore, the location and the atypical chemical
composition of the two samples indicate that mixing of fresh water and brine
accounts for the salinity of the samples. In western Tom Green County, mixing of
fresh ground water and brine seems to be an area! phenomenon. All but three samples
indicate mixing in at least one of the major chemical constituents, Ca, Mg, and $04
(Table 2), and more than half of the samples indicate mixing in two or all of these
constituents. Mixing is also indicated by the cation percentages of ground water in
western Tom Green County, as shown by a linear trend from Ca-Mg-dominated water to
Na-dominated water (Figure 7). This trend could also be interpreted as an evolution
trend. However, considering the position of brine and ground-water values of
Trend 2 in the bivariate plots (Figure 4), mixing rather than evolution through
mineral reactions appears to be the most likely explanation for this cation trend.
Some Br/Cl ratios seem to be additional tracers of salinization sources, with
ratios of less than 30 X 10~4 being indicative of possible mixing of brine and
fresh ground water. However, absolute bromide concentrations, the range of bromide
concentrations, and the range of Br/Cl ratios in all samples are relatively small
in this study, which makes Br a less favorable tracer. Ratios in seep samples fall
within the range of ratios in fresh ground water and are only twice as much as
ratios of Br/Cl ratios in brines underlying the area. In comparison, differences in
Br/Cl ratios of approximately 1:10 were used by Whittemore and Pollock (1979) and
by Richter and Kreitler (1986) to distinguish brine sources. Even more important,
at concentrations of approximately a few mg/L of Br, analytical errors will greatly
affect Br/Cl ratios. For example, a bromide concentration of 1.5 mg/L places
sample 24 within the field of possible mixing of brine and fresh water (Figure 6,
Table 2). In contrast, a concentration of 2 mg/L would place this sample within the
-236-
-------
range of fresh water and seep water in the area. Similarly, nitrate concentrations
may or may not serve as additional tracers of salinization sources. In Runnels
County, where extremely high nitrate concentrations in shallow ground water have
been measured for the past 15 years, low nitrate concentrations in combination with
high chloride concentrations may indicate mixing of fresh ground water and brine,
the latter being high in Cl and low in N03. In contrast, high chloride
concentrations combined with high nitrate concentrations may suggest a common
source of Cl and N03, such as animal waste. However, mixing of Cl-rich brine and
N03~rich ground water would result in a similar relationship between chloride and
nitrate. In western Tom Green Counties, where nitrate concentrations in shallow
ground water are much lower than in Runnels and eastern Tom Green Counties, nitrate
is a less favorable tracer of salinization sources. In general, N03/C1 and
especially Br/Cl ratios are not good tracers of salinization sources in this study
because of their relatively narrow ranges and overlapping trends. At best, these
ratios can be used as supportive arguments for salinization sources, but within a
suite of diagnostic ratios and plots rather than by themselves.
There are four possible mechanisms for the mixing of fresh water and deep-basin
brine in the area. (1) Western Tom Green County includes an outcrop of the Permian
San Angelo and Blaine Formations. These formations contain salt water under
artesian conditions downdip, which indicates the potential for natural discharge of
saline water at formation outcrops and by movement across confining layers. (2)
Discharge of salt water from the San Angelo and Blaine Formations is possible
through unplugged, exploratory water wells that were drilled into saline parts of
these aquifers. The locations of these numerous wells are poorly known. (3) Tom
Green County and Runnels County have been sites of extensive exploration for and
production of oil. Most oil reservoirs in the area underlie artesian brine
aquifers, such as in the Coleman Junction, and thus pathways for upward flow of
-237-
-------
brine from the artesian aquifer or from oil reservoirs along poorly cemented wells
may have been created by exploration and production of oil. Also, shallow seismic
holes may connect saline parts of the San Angelo and Blaine Formations with
overlying fresh ground water. (4) Open-surface pits for brine disposal were used in
the area until the late 1960's. This practice of brine disposal was abandoned in
Texas after numerous cases of ground-water contamination by brine had been
documented. However, brine may still be migrating from below these former disposal
areas into shallow ground water. The amount of salt still present in the subsurface
at those sites and the rate of migration are unknown. There are indications that
all of these potential mechanisms of brine pollution were or are active in the
area. At this time we do not have enough data to chemically characterize these
contamination sources and to explain particular mechanisms for mixing between deep-
basin brine and fresh ground water in this part of West Texas.
Conclusion
In this study, determination of Ca/Cl, Mg/Cl, and SO/j/Cl ratios, and to a
smaller degree Br/Cl and N03/C1 ratios, allowed differentiation between salt-water
pollution derived from evaporation of shallow ground water and pollution derived
from mixing with Na-Cl brine. All these ratios should be considered, rather than
only chloride concentrations or the sole ratio of one constituent over chloride,
because chemical characteristics of these two sources of contamination overlap.
Overlaps are most pronounced at low ionic concentrations because dilution by fresh
water masks chemical characteristics of salt-water sources. Therefore,
differentiation of contamination sources is most successful where concentrations of
dissolved solids are high.
In western Tom Green County, the chemical composition of ground water appears
to result from mixing of fresh ground water and Na-Cl deep-basin brine. This is
-238-
-------
indicated (1) in Piper plots by a mixing trend between Ca-Mg-dominated ground water
and Na-dominated ground water and (2) in bivariate plots by low Ca/Cl, Mg/Cl, and
S04/C1 ratios that indicate trends with deep-basin brine values as high-chloride
end members. Mixing of fresh ground water and deep-basin brine appears to be an
areal phenomenon, but the mechanism of mixing and the source of salt water are
unknown.
In Runnels, Concho, and eastern Tom Green Counties, there appear to be two
causes of deterioration of water quality. Most poor-quality waters result from the
evaporation of shallow ground water. These waters typically have Ca/Cl, Mg/Cl,
S04/C1, and Br/Cl ratios that are higher than those observed in sampled deep-basin
brines. On bivariate plots, these waters suggest trends indicative of evaporation,
that is, ratios are constant with increases in salinity. The potential for ground-
water evaporation and subsequent salinization increases as the water table becomes
shallower. Therefore, salinization by evaporation should be more prevalent in
Runnels County and eastern Tom Green County, where the water table is generally
shallower than in western Tom Green County. In combination with a shallow water
table, nitrate concentrations in most samples from the east are very high owing to
leaching of nitrate in the shallow subsurface. Other poor-quality waters collected
in the area during this study result from mixing between Na-Cl brine and fresh
ground water, which occurs on a local basis. These waters, which were obtained from
shallow water wells close to producing oil wells, have low Ca/Cl, Mg/Cl, SO/j/Cl,
N03/C1, and Br/Cl ratios. The latter water type is similar to brines underlying the
area and to ground water in western Tom Green County, suggesting mixing of fresh
water and brine.
-239-
-------
References
Barnes, V. E., 1975, San Angelo Sheet: The University of Texas at Austin, Bureau of
Economic Geology, Geologic Atlas of Texas, Scale 1:250,000.
Barnes, V. E., 1976, Brownwood Sheet: The University of Texas at Austin, Bureau of
Economic Geology, Geologic Atlas of Texas, Scale 1:250,000.
Kreitler, C. W., 1975, Determining the Source of Nitrate in Ground Water by
Nitrogen Isotope Studies: The University of Texas at Austin, Bureau of Economic
Geology Report of Investigations No. 83, 57 pp.
Marshall, M. W., 1976, City of San Angelo Pollution Abatement Program, Water
Department: Memorandum to T. L. Koederitz, P. E., Water Pollution Control and
Abatement Program Director.
Miller, M. R., Donovan, J. J., Bergatino, R. N., Sonderegger, J. L., Schmidt,
F. A., and Brown, P. L., 1981, Saline Seep Development and Control in the North
American Great Plains—Hydrogeological Aspects, ^n Holmes, J. W., and Talsma,
T., (eds.), Land and Stream Salinity: Elsevier Development in Agricultural
Engineering, v. 2, 391 pp.
Raschke, A. J., and Seaman, W. H., 1976, Leaking Core Hole Problem Review, Hatchel
Area, Runnels County, Texas: Railroad Commission of Texas, Oil and Gas Division
District 7-C, San Angelo, Texas, 14 pp.
Reed, E. L., 1962, Letter to Mr. James K. Anderson: Midland, Texas, April 2.
Richter, B. C., and Kreitler, C. W., 1985, Sources of Shallow Saline Ground Water
in Concho, Runnels, and Tom Green Counties: The University of Texas at Austin,
Bureau of Economic Geology, Report prepared for Railroad Commission of Texas
under Contract No. IAC(84-85)-2122, 31 pp.
Richter, B. C., and Kreitler, C. W., 1986, Geochemistry of Salt Water Beneath the
Rolling Plains, North-Central Texas: Ground Water, v. 24, no. 6, pp. 735-742.
-240-
-------
Udden, J. A., and Phillips, W. B., 1911, Report on Oil, Gas, Coal and Water
Prospects near San Angelo, Tom Green County, Texas: Report to the Chamber of
Commerce, San Angelo, Texas, 36 pp.
Whittemore, D. 0., and Pollock, L. M., 1979, Determination of Salinity Sources in
Water Resources of Kansas by Minor Alkali Metal and Halide Chemistry:
Manhattan, Kansas, Kansas Water Resources Research Institute Contribution
No. 208, 28 pp.
Willis, G. W., 1954, Ground-water Resources of Tom Green County, Texas: Texas Board
of Water Engineers Bulletin 5411, 60 pp.
Work Projects Administration, 1941, Tom Green County—Records of Wells and Springs,
Drillers' Logs, Water Analyses, and Map Showing Locations of Wells and Springs:
Texas Board of Water Engineers Work Projects Administration, Project 17279,
80 pp.
-241-
-------
List of Figures
Figure 1 Area location map showing outcrop areas of major geologic units
(from Barnes, 1975, 1976) and location of sample sites.
Figure 2 Piper diagrams of ground-water chemistry in Tom Green and Runnels
Counties (data from Work Projects Administration [1941], Willis
[1954], and Texas Natural Resources Information System).
Figure 3 Piper diagram of ground-water chemistry in Tom Green, Runnels, and
Concho Counties (data from this study).
Figure 4 Bivariate plots of ground water and deep-basin brine chemistry in
Tom Green, Runnels, and Concho Counties (data from this study).
Figure 5 Nitrate concentrations in ground water from Tom Green, Runnels, and
Concho Counties (data from this study).
Figure 6 Bivariate plots of ground water and deep-basin brine chemistry in
Tom Green, Runnels, and Concho Counties, with data sorted according
to sample location (data from this study).
Figure 7 Cation diagrams of ground-water chemistry in (a) Runnels, Concho,
and eastern Tom Green Counties and (b) western Tom Green County
(data from this study).
-242-
-------
List of Tables
Table 1 Generalized relationship between strati graphic and hydrogeologic
units found in study area; see text for discussion (modified after
Willis, 1954).
Table 2 Chemical analyses of ground water and brines in Tom Green, Runnels,
and Concho Counties.
-243-
-------
I
ho
STRATIGRAPHIC UNIT
System Formation
Quaternary Alluvium
Cretaceous
Blaine Gypsum
San Angelo
Sandstone
Coleman Junction
HYDROGEOLOGIC UNIT
Yields small quantities of
potable water
Yields potable water from
two aquifer units that are
separated by confining beds
of massive limestone
Yields small amounts of
highly mineralized water
Yields small amounts of
moderately to highly
mineralized water
Highly overpressured brine
aquifer
Pennsylvanian
-------
Table 2. Chemical analyses (in mg/L) of ground water and brines
in Tom Green, Runnels, and Concho Counties.
ID Ca
No.
Mg
Na
Coleman
B1+ 2310
B2@ 1940
B30 2500
B4+ 4530
B5$ 1605
B6$ 2400
B7& 931
1120
1059
1122
5
1110
881
696
Ground Water:
1 113
2 255
3 335
4 1172
5 731
6** 350
7** 319
8** 299
9 414
10** 202
11** 585
12 129
13 369
14 273
15 525
16 252
17 359
18 229
19 189
20 185
21 188
22 212
23 157
24 669
35
216
138
524
198
115
154
137
339
82
192
108
50
764
123
82
128
96
62
118
115
111
64
242
25700
22500
22900
31600
7440
26100
15600
S04
Cl
Br
N03
Junction and Oil Field
4080
2310
4170
3750
3390
3930
9
41900
38000
38300
51600
15500
41200
27200
Runnels, Concho,
173
1140
269
1790
249
295
305
289
512
245
633
218
271
952
178
169
334
143
114
91
192
233
156
369
108
378
940
1092
1815
591
567
501
1485
249
2115
251
223
2415
1008
270
174
167
156
474
465
258
261
2040
166
2330
452
5130
595
699
723
685
983
454
735
343
720
1460
516
461
980
454
236
205
367
482
184
639
70.8
70.2
70.9
93.5
37.2
83.4
56.7
and
0.9
5.5
1.8
3.8
2.4
2.5
2.5
2.4
3.2
1.6
3.1
1.6
2.5
8.5
0.2
1.9
3.3
2.0
1.0
0.8
1.4
2.0
0.8
1.5
< 1
< 1
< 1
< 1
< 1
< 1
< 1
Eastern
155
< 1
5
1
32
149
128
158
57
169
< 1
121
165
30
147
115
229
115
98
35
28
131
20
< 1
Ca*
TTT
Mg*
rr
S04*
CT
IT3
(BrxlO4)
Cl
Brines
0.05
0.04
0.06
0.08
0.09
0.05
0.05
Tom
0.60
0.09
0.66
0.20
1.09
0.50
0.39
0.39
0.37
0.40
0.71
0.34
0.46
0.16
0.90
0.49
0.33
0.50
0.71
0.80
0.46
0.39
0.76
0.93
.04
.04
.04
.00
.11
.03
.03
Green
.31
.14
.45
.15
.49
.24
.31
.29
.51
.27
.38
.46
.10
.77
.35
.26
.19
.31
.39
.84
.46
.34
.51
.55
0.03
0.02
0.04
0.03
0.08
0.04
0.00
.00
.00
.00
.00
.00
.00
.00
16.9
18.5
18.5
18.1
24.0
20.2
20.2
Counties
0.24
0.06
0.76
0.08
1.11
0.31
0.28
0.27
0.55
0.20
1.05
0.27
0.10
0.61
0.71
0.21
0.07
0.14
0.24
0.84
0.46
0.20
0.52
1.16
.93
.00
.01
.00
.05
.21
.18
.23
.06
.37
.00
.35
.23
.02
.28
.25
.25
.25
.41
.17
.07
.27
.11
.00
48.2
23.6
39.8
7.2
38.7
34.3
34.6
35.0
31.5
33.0
42.2
46.6
33.3
57.5
3.9
39.0
33.7
41.8
42.4
39.0
38.1
39.4
43.5
23.5
-245-
-------
Table 2 (continued)
ID Ca Mg Na S04 Cl Br N03 Ca* Mg* $04* N03 (BrxlO4)
NO. FT n~ rr TTT ~n
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
Ground Water: Western Tom Green County
268
452
181
448
536
385
188
73
90
212
498
560
519
280
921
97
152
50
139
177
124
69
30
41
89
185
263
223
192
491
243
363
391
731
744
386
232
259
113
422
1770
978
220
284
7185
161
192
284
402
386
131
113
180
128
318
432
462
753
225
2070
735
1310
573
1622
1970
1230
479
211
161
712
3380
2650
1060
976
11630
2.7
4.4
1.9
4.7
5.6
4.3
1.3
0.7
0.6
2.2
6.9
6.2
4.5
3.2
9.9
125
87
29
173
43
63
8
2
2
29
43
41
46
13
13
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
.32
.31
.28
.25
.24
.28
.35
.31
.50
.27
.13
.18
.44
.26
.07
.19
.17
.13
.13
.13
.15
.21
.21
.37
.18
.08
.15
.31
.29
.06
0.08
0.06
0.18
0.09
0.07
0.04
0.09
0.31
0.29
0.17
0.05
0.07
0.26
0.09
0.07
.17
.07
.05
.11
.02
.05
.02
.01
.01
.04
.01
.01
.04
.01
.00
36.7
33.6
33.2
28.4
28.4
35.0
27.1
33.2
37.3
30.9
20.1
23.4
42.4
32.8
8.4
EXPLANATION
$
&
**
0.09
Mol ratios
Leaky injection well with flow from bradenhead
Flowing well completed and abandoned in the Coleman Junction
aquifer; sample B2 was obtained after 10 minutes of flow, sample
B3 was obtained after 90 minutes of flow
Flowing core hole, approximately 100 ft deep
Producing oil well, possibly affected by waterflooding
Seep sample
Ratio similar to ratios of Coleman Junction and oil field brines
-246-
-------
ion"
I
K>
-p-
^J
I
EXPLANATION
o Water well • Brine well
Cretaceous
;;:; ;j ;; CONCHQ i
QA5896
-------
00
I
CK250 mg/L
TOM GREEN COUNTY
Cl>250 mg/L
RUNNELS COUNTY
-------
EXPLANATION
o Ground water sample A Brine sample
QA 5900
-------
4000 H
0
o
400 H
Ln
O
I
2400H
240H
Log Cl (mmol/L)
I 2
J i
350
3500
Cl (mg/L)
35,000
Log Cl (mmol/L)
l 2
J i
350
I
3500
Cl (mg/L)
35,000
o'
E
9600H
O
C/5
960H
60 H
40H
20H
Log Cl (mmol/L)
I 2
350
3500
Cl (mg/L)
35,000
, Southern
I Roljing
. Plains of
| North-
. Central
| Texas
100 1000 10,000
Cl (mg/L)
100,000
QA 5901
-------
EXPLANATION
20 Nitrate concentration as N03 (mg/L)
• Good-quality ground water (Cl < 250 mg/L)
x Mixing between fresh water and deep-basin brine
I
N}
Ol
-32°
I N
]28go!58_j2l RUNNELS |
_ CONCHQ |
QA 5902
-------
4000-
400-
Log Cl (mmol/L)
I 2
350 3500
Cl (mg/L)
I
35,000
-2'
9600-
960-
Log Cl (mmol/L)
I 2
I I
I
350
3500
Cl (mg/L)
35,000
-2
I
M
Ul
ro
I
2400-
240-
Log Cl (mmol/L)
2
I
350
3500
Cl (mg/L)
35,000
-I CT
O
-
60-
40-
-
20-
/"^\
. Southern
[ Rolling
Plains of
1 North -
u 0 Central
O o Q, D • 1 Texas /
• aSfffa i~^~~ ^/
« • (Kansas /
Brines^
4° *
1 1 1
100
1000 10,000
Cl (mg/L)
- EXPLANATION
4 Sample numbers (see table 2)
O Water sample from Runnels, Concho, and eastern Tom Green Counties
n Water sample from seep area
• Water sample from western Tom Green County
100,000
OA 5903
-------
-253-
-------
ABSTRACT
COFRC-GREG PIETRUSZKA
CHEVRON U.S.A. - T. R. BEVINS
Fie 1 ct Results of Tracer Teats Conducted i _n_ 0i 1 Field
Steam and Non-Condensible Gas Inject ion Projects
chevron has more than twenty years of experience in using
rhemi r:.\ I and radioactive tracers to determine flow patterns
of injected steam and non-condensible gases in oil field
reservoirs. Chemical tracers used include sodium salt ions
(btumide, chloride and nitrate), sulfur hexaflouride (SF6),
and £1 ourocarbons (Freons 11 and 113). Radioactive tracers
1.1 ;o".:d include krypton 05, tritiated water and tritiated
methane, Tracer test design considerations wj 11 be discussed
including reservoir characteristics, amounts of tracers,
Irriect. ion and monitoring techniques and impact on
env ironment.
Results of tracer testing of several different Injection
fluids will bp presented. The Kern River Ten-Pattern Steam
Flood tracer program demonstrated the use of tracery in steam
nnoil Lug; the SACROC tracer test program applied tracer
t.eM.ing to C02 flooding; and the current Painter Reservoir
Unit tracer test program applies tracer testing to a nitrogen
infection project. Uses of tracer testing results will be
ii iscussetj including as an aid in: geologic modeling,
:<:.'Hjuct i on/reservoir engineering, and reservoir modeling for
• . ruir ] a t i on .
-254-
-------
BIOGRAPHICAL SKETCH
Greg Petruszka graduated from the University of Tulsa in l':)80
with a Bachelor of Schience Degree in Petroleum Engineering.
He worked for Chevron USA> Denver between June 198d dud July
1984 on assignment in drilling production and reservoir
engineering. He hds worked for Chevron Oil Research Company
since July 1984 and his current assignment as Research
Engineer is in the Production Research Department. He
implements new research technology in oilfield app 1. {<"* t \ on.i:;
NB - A FULL TEXT IS NOT AVAILABLE FOR INCLUSION IN THESE
PROCEEDINGS
-255-
-------
Identification and Closure
of Shallow Brine Disposal Wells in
Pennsylvania
By Jon M. Capacasa, P.E., Chief
Drinking Water/Ground Water Protection Branch
U.S. EPA, Region III
Presented at UIPC International Symposium
New Orleans, LA
May 5, 1987
-256-
-------
Identification and Closure of Shallow
Brine Disposal Wells (Blow Boxes) in
Pennsylvania
By Jon M. Capacasa
US EPA Region III, Philadelphia, PA
The shallow gas producing fields in the Commonwealth of Pennsylvania
are "wet gas" formations which produce hrine in association with the
natural gas. The brine is a highly saline formation fluid which contains
a variety of chemical elements. The EPA has established Primary or Second-
ary Drinking Water Standards under Parts 141 and 142 of the Code of
Federal Regulations for a number of these elements, including: arsenic,
barium, cadmium, chromium, chlorides, iron, sodium, sulfate and strontium.
Produced brines consistently exceed established MCL's and secondary
standards and can be a threat to public and individual water supplies if
not disposed of in an environmentally prudent manner.
Gas producers use a variety of methods to dispose of their waste
brines including injection wells, annular disposal, pretreatraent to a
municipal system, stream discharges, road spreading, etc. A number of
these are approvable methods for disposal of brines and drilling fluids.
However, it has been the common practice of producers in the
Southwestern PA area to dispose of produced brines on site into a dry
well known as a blow box. Blow box is a generic terra used to describe a
bottomless wooden or concrete structure of varying construction type;
most common types seen within the study area were rectangular concrete,
circular concrete, and square wooden cribs extending, in depth, approx-
imately 1.8 - 3.6 meters (6 to!2 feet). Many were of similar construction
to septic system tanks.
At gas well sites employing blow boxes, the brine is typically directed
from the well head, sometimes via a brine storage tank, to the blox box, from
which it percolates into the subsurface potentially contaminating under-
ground sources of drinking water. After a period of time, the blow boxes
tend to fill with silt, resulting in reduced fluid capacity. As a result
of this reduced fluid capacity and seasonally high water tables, surface
releases of brine may occur, causing extensive vegetation damage and
discharges of contaminants into local streams. Sand or gravel is often
placed in a portion of the box to promote percolation.
Blow boxes are in essence, shallow brine disposal wells, discharging
brine directly into the zone of aeration and surficial aquifers. Such
injection is prohibited under EPA's Underground Injection Control
(UIC) Program, specifically in such sections as: the definition of a
Class II well §144.6(b); §144.12 prohibits movement of fluids into under-
ground sources of drinking water; §144.21 requires authorization by rule
or permit to inject; §144.27 requires inventory submittals by owners or
operators authorized to inject; and §144.28 specifies casing and cementing
requirements to prevent fluid movement. Discharges to ground water are
also prohibited under the Pennsylvania's Clean Streams Law (P.L. 1987
No. 394) and Section 207 of the Oil and Gas Regulations (P.L. 223).
-257-
-------
The environmental risk associated with the blow box practice
was generally assessed based upon interviews of state field inspectors,
by EPA review of drinking water quality data in the area, analytical
results of brine samples and ancedotal information from news reports of
individual water well contaminantion. Although contributing on average
small volumes of brine to the ground water (ranging from .25 to 1
barrel per day), the widespread use of the practice in a 13 county area
of Southwestern PA brought EPA's original estimate of 3000 or more gas
well sites using boxes. Data on total volumes of brine are as yet, incomplete,
however one large company provides an illulstrative example in documenting
that their yearly brine production of 98,000 barrels is now directed to
a treatment plant. An early drinking water survey of sodium levels of
public supplies in the State of PA left the overall observation that the
highest levels of sodium could be found in the SW PA Region. In fact
the levels ranged as high as 250 ppm in Indiana County. Coupled with
this observation were several documented barium MCL violations in the
area under the SDWA. On-site visits to blow box locations also provided
frequent ancedotal accounts about individual water wells or agricultural
use wells fouled by high sodium or barium levels.
Given the regulatory mandate for protection of USDW's, the chemical
composition of the brines, the estimate of in excess of 3000 active blow
boxes in this small region of PA, and the other conerns for localized
drinking water quality impacts based on data reviews, EPA Region III
developed in April, 1985 a strategy for the identication, notification
and enforced closure of blow boxes in PA under the UIC Program.
The Blow Box Compliance Strategy was developed with a full appreciation
of the 40 or more year history of blow box use with little previous
interference by regulators, the marginal economics of the gas industry
In this area characterized by many independent owners and operators and a
depressed gas price, and the large number (3000) of small sites which
were involved. The decision was made to set up a strategy which sought
the cooperation of the industry through early and frequent notification
of the problem and requirements, provided sufficient lead time for
conscientious operators to close the wells without undue economic burden, and
establish a series of progressively more severe enforcement actions for those
owners who denied operations or resisted closure efforts. A reasonable
goal of a 2 year closure project was developed. The strategy was divided
into 3 phases some of which proceeded on concurrent paths:
I - EPA Identification of Blow Boxes Owners and Locations.
II - Outreach and Notification to Owners of Record and Verfication.
Ill - Closure Plans/Methods.
-258-
-------
The first phase of the project was critical to its success. The
public outreach and notification sought to advise all potentially
impacted owners of the problem, seek their feedback and cooperation
in the strategy, and provide them an opportunity to comply with DIG
inventory deadline of June 25, 1985. Initial briefings/meetings were
held with PA Natural Gas Association (PNGA) and PA Oil and Gas Associa-
tion (POGAM) before the strategy was finalized or other outreach occurred.
The Associations responded in a positive way to the upfront communication
of EPA's goals and the reasonable compliance deadlines. An aggressive
series of press releases, paid news ads, trade journal articles, and
direct mail notices followed to all gas well owners of state record to
promote inventory identification by the June 25, 1985 deadline. As a
result of these efforts, over 1200 well sites were inventoried in a 2-3
month period. For these individuals and companies EPA negotiated up to
18-month closure schedules based on the number of boxes owned. Five
bilateral compliance agreements were executed to confirm the closure
schedules and methods for closure. There were no penalty assessments
for operation during this period.
To maintain the enforcement aspect of the project and as an incentive
to self-identification, EPA moved on a separate track to independently
identify blow box locations and then ownership using several innovative
techniques. The techniques used were gas well ownership records of the
state, interviews of state field inspectors, field inspections by EPA,
and aerial surveillance. No official inventory of blow box locations
existed at the outset of this project.
A pilot study was conducted to assess the feasibiity of identifying
wells using blow boxes in a cost efficient manner using aerial photography.
Three types of film and three different imagery scales were considered
in determining the best combination for identification accuracy and
cost effectiveness. The best combination of film type and scale
was found to be 9" color infrared film at a scale of 1:12,000. Ground
truthing by EPA personnel was done to establish features and signatures
associated with blow boxes. Based on the pilot study findings, the
entire study area was flown, in early 1985.
The coverage area was approximately 10,748 square kilometers (4150
square miles). The EPA Environmental Photographic Interpretation Center
performed this sutdy at the request of EPA Region III.
Stereoscopic viewing of the backlit transparencies provided a three-
dimensional effect which, when viewed at various magnifications,
enabled the identification of signatures or features associated
with gas wells using blow boxes. The "signature" refers to a combination
of visible characteristics (such as color, tone, shadow, texture, size,
shape, pattern and association) which permit a specific object or condition
to be recognized.
-259-
-------
It is not prudent for EPA to provide the specific features
associated with the presence of blow boxes here. Suffice it to say
that no one feature provided conclusive identifications, only through a
combination of well construction and surrounding environmental features
could they be identified as "possible," "probable" or "definite" blow
boxes from the imagery.
From this work, over 1526 sites were identified for followup of
which only 118 or 7.7% were determined to be invalid identifications
through field inspections.
The followup to this information could have been extremely resource
intensive with field visits to each. However, Phase II of the strategy
targetted notices to owners/operators of these well sites to solicit
voluntary compliance. EPIC "calls" were crosschecked versus state records
to obtain well ownership. Due to the potential ramifications of improperly
accusing persons of blow box operation, initial letters were sent out
educating the addressees about EPA's program and the probable blow box
ownership. Persons in this group fell into two categories: those denying
ownership and those cooperating to close the boxes. Field inspections
were used only to supplement ownership identification where state permit
files were not complete.
Owners were asked to voluntarily provide:
- inventory forms for all blow boxes operated; specific site location
on maps; total numbers of facilities owned.
Following receipt of the data, the owners were placed on a compliance
schedule for closure of each identified blow box. Those denying ownership
received priority attention for field inspection and followup in terms
of potential enforcement and penalties.
Phase III or Closure Phase of the project involved followup on the
compliance schedules and field verification of closures. As of April, 1987,
voluntary compliance efforts have identified roughly 2154 blow boxes. Of
these, only 20 remain active and the rest are temporarily or permanently
abandoned, subject to field verification. The 2154 closures were achieved
without one legally enforceable order being issued or civil action taken.
EPA is just now in process of issuing administrative orders to the very
small number of blox box owners which came to our attention at this
stage of the project. All of the EPIC leads have been addressed.
As blow boxes were closed, EPA personnel verified a representative
number of closures relative to accepted methods and assessed whether or
not alternate means of equally illegal disposal of produced brines were
replacing the blow boxes. A listing of state accepted alternate brine
disposal methods was provided to each blow box opeator identified by
this project.
-260-
-------
Summary and Results
The blow box compliance project was a highly successful,
efficiently managed project for EPA Region III. The practice of illegal
brine disposal into surficial aquifers via blow box injection wells
has for all intents and purposes been brought to a close by this effort.
The elimination of many minor surface discharges to streams has also been
avoided from these sites. The project has seen the placement of many
above ground storage tanks for brine hauling to approved disposal sites
and facilities. In this regard, over the period from 1985 to the present,
three new permitted brine treatment or pretreatment facilities have
opened in PA for commercial use. In addition, three new brine injection
wells were permitted for private use or commercial use. This compliance
effort, in redirecting brine to such alternate facilities, has helped to
make proper brine disposal practices more economically viable.
More importantly, this initiative by EPA's UIC program helped establish
a momentum for proper brine disposal practices in PA and EPA/State
enforcement of the same.
The key to the success of this project was the early public outreach
and notifications by EPA to the gas industry to seek their voluntary efforts
to comply in lieu of enforced efforts by EPA. This single factor resulted
in 2154 closures of illegal wells in less than two years.
Acknowledgements
The author wishes to acknowledge the innovative efforts and dedication
of the following who were critical to this project's success and of
assistance in this paper: Karen DeWald, Gary Naumick, George Hoessel,
Anthony Spano, Alfred Sturniolo and EPA's Environmental Photographic
Interpretation Center (EPIC).
-261-
-------
MATHEMATICAL EVALUATION OF OPERATING PARAMETERS
IDENTIFIED IN A CLASS II BRINE DISPOSAL WELL PERMIT APPLICATION
MARC EDWARD HERMAN
U.S. ENVIRONMENTAL PROTECTION AGENCY
REGION 8 (8WM-DW)
999 18TH STREET, SUITE 500
DENVER, COLORADO 80202
ACKNOWLEDGEMENTS
The author gratefully acknowledges Victoria Parker Christensen, Lester
Sprenger, Gustav Stolz, Jr., Eric Koglin, Joseph J. D'Lugosz, and Debra G.
Ehlert, for their invaluable comments and thorough reviews of the manuscript.
As chairman of the internal EPA review committee, Mr. Sprenger ensured that
peer review proceeded smoothly and efficiently. Daily technical discussions
between the author, Mr. Stolz, Ms. Ehlert, and Ms. Parker Christensen serve to
continually improve EPA Region 8's UIC program implementation. Thanks are
also extended to Ms. Kay Stortz for her careful proofreading of the paper.
ABSTRACT
The purpose of the Underground Injection Control (UIC) program is to
prevent contamination, caused by improper injection operations, of underground
sources of drinking water (USDW's). In Montana, there are approximately 150
Class II brine disposal wells that must be regulated under the UIC program,
which is administered by U.S. Environmental Protection Agency (EPA) Region 8
offices.
Any person who proposes to operate a new Class II brine disposal well is
required to submit a permit application to the EPA. Permit applications for
rule-authorized (existing) brine disposal wells must be submitted within 4
years of the program promulgation date (June 25, 1988 for Montana).
-262-
-------
The following operating data are reviewed in the course of evaluating a
permit application: (a) specific gravity and viscosity of the injection
fluid; (b) injection zone rock type, thickness, porosity, depth, and
permeability; (c) depth to top of perforations and extent of perforated
interval; (d) fracture pressure data and pore pressure of the injection zone;
(e) proposed average/maximum injection rate and pressure; and (f) expected
operating life of the well.
Some potential injection operation impacts are: (1) the fracturing of
either the injection or confining zones; (2) the amount of injection zone
pore space available for fill-up; (3) the extent of the fluid plume; (4) the
length of time the well should operate, based on volume fill-up calculations;
and (5) the feasibility of disposing of proposed fluid volumes at proposed
injection pressures.
Numerical approximations are obtained through the use of analytical
equations that take into account injection pressure, volume, and rate.
Estimation of formation fracture pressure values may be accomplished by
evaluating the results of a step-rate test.
Comparisons are made between: fracture pressure and proposed maximum
injection pressure; projected total volume of fluid to be injected and
available formation pore volume; theoretical injection rate and proposed
injection pressure; and formation pressure build-up and proposed injection
pressure.
INTRODUCTION
The purpose of the Underground Injection Control (UIC) program is to
protect underground sources of drinking water (USDW's) from the improper
operation of injection facilities. The UIC program for the State of Montana
-263-
-------
is administered by the U.S. Environmental Protection Agency (EPA) Region 8
offices, and became effective on June 25, 1984, under the authority of Part C
of the Safe Drinking Water Act (SDWA).
Many of the underground injection facilities in Montana are Class II
injection wells. These are wells in which brine and salt water, brought to
the surface in association with oil production, are injected into the
subsurface. The Class II well category can be further divided by defining the
purpose of a given injection operation.
Two basic reasons for salt water disposal are: (a) injection into
water-bearing formations to dispose of the salt water; and (b) injection into
oil-bearing formations to enhance the recovery of hydrocarbons. The former
are termed "salt water disposal" wells, and are the subject of this report.
According to the UIC regulations for EPA administered programs (40 CFR
Subpart D Section 144.31(c)(l)), Class II brine disposal wells operating prior
to a program promulgation date are authorized by rule until 5 years after the
date of promulgation. However, a permit application for each and every
existing brine disposal well must be submitted no later than 4 years from the
promulgation date of the UIC program. At the end of the 5-year period, all
existiny brine disposal wells shall have been issued permits or permit
denials, thereby replacing any rule-authorized status. The 4-year deadline
for applications is designed to provide a transition period from
rule-authorized to permitted injection status.
Except for rule-authorized wells, all other brine disposal wells are
prohibited unless authorized by permit. Generally speaking then, any company
that proposes or performs a Class II brine disposal operation is, or will be,
required to submit a permit application to the EPA. All applicants for UIC
permits must provide the EPA with a completed application form (40 CFR Subpart
-264-
-------
D Section 144.31 (e)) and supplemental information unique to the specific
facility.
OBJECTIVE
Proper implementation of the permit review process is based on the
establishment of consistent guidelines for evaluating criteria associated with
proposed operating parameters. The purpose of this paper is to elaborate on
several techniques for analyzing permit application operating data. This
paper does not discuss all aspects of the UIC program, nor does it discuss all
Class II injection well technical issues. Rather, an attempt is made to
present quantitative methods for evaluating certain injection well operating
parameters.
These methods are used as a means for gaining a more objective
understanding of the impacts due to injection operations. The equations and
results serve as tools to supplement additional information obtained during
the course of an application evaluation. Figure 1 highlights the topics and
parameters to be addressed.
DATA REQUIREMENTS
Prior to issuance of a permit or permit denial for the construction or
conversion of a new Class II brine disposal well, the following information is
considered by the EPA:
(a) information required on EPA form 4;
(b) a map locating the disposal well and other wells within the
applicable area of review;
(c) a tabulation of data on all wells within the area of review that
penetrate the injection zone;
-265-
-------
(d) proposed average and maximum rate/volume of fluid to be injected;
(e) proposed average and maximum injection pressure;
(f) source and analysis of the injection fluid;
(g) hydrogeologic data on the injection and confining zones;
(h) hydrogeologic data on all USDW's present;
(i) construction schematic of the disposal well;
(j) a demonstration of financial responsibility;
(k) available logging and formation testing data;
(1) a demonstration of mechanical integrity;
(m) injection procedures; and
(n) status of defective wells within the area of review.
Except for items (b), (c), and (n), the information above must also
accompany any application for an existing Class II brine disposal well.
Applicants for both existing and new brine disposal wells may be required to
submit additional numerical data identifying facility operating parameters not
previously mentioned.
Numerical data that are used to analyze the operational aspects of a
Class II brine disposal well include, but are not limited to:
(a) specific gravity of injection fluid;
(b) viscosity of injection fluid;
(c) fracture pressure data for the injection zone;
(d) injection zone pore pressure;
(e) proposed maximum injection pressure;
(f) porosity of injection zone;
(9) permeability of injection zone;
(h) injection zone thickness;
(i) depth to top of perforations;
-266-
-------
(j) description of injection zone rock type;
(k) number of years the well has operated and is expected to operate;
(1) proposed maximum injection rate and cumulative injected volume; and
(m) theoretical radial limitation of the injection fluid plume, measured
from the injection well.
This information will serve as input for mathematical equations that
approximate the relationships between injection pressure, injection rate, and
ground-water flow. Environmental effects that are of particular concern are:
(1) the potential extent of the fluid plume; (2) the potential for fracturing
either the injection or confining zones; and (3) the feasibility of disposing
of proposed fluid volumes at proposed injection pressures.
GOVERNING EQUATIONS
Analytical ground-water models have proven to be useful tools for
evaluating many ground-water problems. By combining the results of these
models with a qualitative analysis of accompanying hydrogeologic information,
the EPA permit writer can conduct a comprehensive review of any brine disposal
well permit application.
Five analytical equations are employed to evaluate the impact of the
operating parameters on the injection formation, and the reasonability of the
brine disposal operations:
pressure due to hydrostatic head:
Pd = 12Sh (1)
-267-
-------
injection/fracture pressure equivalent at specific depth:
Pi=Pw
injection zone pore volume:
V = Crrr2bn)/5.6 (3)
length of time to fill pore volume:
T = V/365q (4)
steady flow from a well in a confined aquifer:
Q = (7.07kb[Pi - Pf])/m(ln re/rw) (5)
The following assumptions are made so that the above analytical equations
may be used with some degree of confidence (Bear, 1979):
(a) ground-water flow obeys Darcy's law;
(b) ambient ground-water flow is negligible;
(c) ground-water flow is radially symmetric, steady-, and horizontal;
(d) the injection zone is a homogeneous, isotropic, confined aquifer;
(e) injection zone hydraulic conductivity and thickness are constant;
(f) the base of the injection zone is horizontal ;
(g) the injection zone has an infinite areal extent;
(h) the injection rate is constant; and
(i) the injection zone is fully penetrated by the well.
-268-
-------
For many brine disposal wells, this last assumption is not valid.
However, for a well that partially penetrates an aquifer, at a distance of
"2b" from the well, the effects of partial penetration become negligible and
ground-water flow is essentially horizontal (Bear, 1979).
PROBLEM FORMULATION
The hypothetical example discussed in this paper is based on an actual
permit application, and although the numerical values for the parameters are
not identical to the original problem, they represent realistic estimates for
operating and hydrogeologic conditions in Montana.
The brine disposal operation to be reviewed is a rule-authorized well
that has been operating for 5 years. Figure 2 is a well schematic that
illustrates several important parameters. Much of the technical data needed
to perform an analytical evaluation of a given permit application are basic
operating parameters that are easily obtainable (i.e. injection pressure,
rate, formation thickness, depth to perforations, etc.).
Values for the specific gravity and dynamic viscosity of the injection
fluid can be found or derived from the chemical analysis submitted with the
permit application. The applicant is also often able to provide acceptable
estimates for the porosity, permeability, and bottom hole pore pressure of the
injection zone. Table 1 is a compilation of parameters and equivalent
numerical values to be used for this sample problem.
In the event that certain hydrogeologic data are not available, the
permit writer has several options. First, textbooks by Davis and DeWiest
(1966), Freeze and Cherry (1979), and Mott (1979) can be used to provide
approximations for fluid and formation characteristics.
-269-
-------
Perhaps of more value is the practical experience the permit writer
accumulates from each application that is processed. As each application is
reviewed, the permit writer gains a better understanding of the geologic
characteristics that are unique to the individual oil-producinq areas of
Montana. By cross-referencing the data, the permit writer can assess the
validity of submitted information.
FRACTURE PRESSURE ESTIMATION
According to UIC regulations (40 CFR Part 146 Subpart C Section
146.23(a)(l)),
"Injection pressure at the wellhead shall not
exceed a maximum which shall be calculated so as
to assure that the pressure during injection does
not initiate new fractures or propagate existing
fractures in the confining zone adjacent to the
USDW's."
Essentially, each permit must establish a maximum injection pressure to
ensure that fractures are not initiated in a confining zone and that injected
fluids do not migrate into USDW's. Realistically, fracture pressure data for
a confining zone is rarely available. On the other hand, the injection zone
is almost always tested.
It has been observed that injection formations usually possess lower
fracture pressure values than the confining zones overlying and underlying
them. Therefore, it has been concluded that fracture data obtained for the
injection formation will represent conservative estimates that can be
confidently applied to the requirements set by the regulations.
For existing or converted disposal wells, the applicant usually submits
the results of a fracture treatment that was conducted shortly after the well
was constructed. Experience indicates that the average value for fracture
-270-
-------
pressure can be expected to increase during the operational life of a well.
This implies that an old fracture test will most likely be an underestimate of
current formation conditions. With this in mind, it is in the applicant's
best interest to perform an up-to-date step-rate test to determine a more
appropriate fracture pressure value.
For the purposes of this paper, fracture (or breakdown) pressure will be
defined as the instantaneous shut-in pressure (ISIP) plus the additional
pressure needed to overcome fractional losses in the well. JSIP is the
pressure needed to maintain an open fracture. In a step-rate test, the
formation is intentionally fractured in order to obtain values for the
breakdown, ISIP, and frictional loss pressures.
The results of a step-rate test can be graphed (Figure 3) to determine an
approximate value for fracture pressure. The break in slope between the two
lines is taken as the point at which a fracture is initiated. For this
example, the fracture pressure appears to be approximately 875 psig. This
value will be compared to the proposed maximum injection pressure. A more
conservative estimate of fracture pressure will be provided if the ISIP value
is used.
If an applicant is unable to supply the EPA with current results from a
step-rate test, the following equation is applied (40 CFR Part 147 Subpart BB
Section 147-1353(a)):
Pw = (0.733 - 0.433Sjh (6)
w g
where S is the specific gravity of the injection or fracturing fluid. This
equation assumes that the fracture gradient for any given formation is 0.733
psig/ft. Fortunately, a sufficient number of permit applications has provided
-271-
-------
the EPA Region 8 staff with a clearer picture regarding fracture
characteristics for many of the formations in Montana. This knowledge allows
permit writers to assess the appropriateness of equation 6.
INJECTION PRESSURE CALCULATIONS
To determine an acceptable value for maximum allowable injection
pressure, the permit writer can compare proposed operating pressures with
fracture pressure data derived from field-determined step-rate tests. Values
for fracture and proposed maximum injection pressures at specific depths
(usually the top of the perforated interval) can be calculated in the
following manner.
Pressure at the top of the perforations, induced by applying a surface
pressure equivalent to the fracture pressure, can be calculated as follows.
First, the pressure due to hydrostatic head is determined with equation 1.
Pd = 12Sh (1)
NOTE: the fluid used in the step-rate test is the same as the injection
fluid; the specific gravity of the fluid is 1.107.
S = [(1.107M62.4 Ib/ft3)]/(12 in/ft)3
= 0.03997 = 0.040 pel.
Substituting 0.040 pci for "S" and 1400 feet for "h",
Pd = (12)(0.040)(1400) = 672 psig.
-272-
-------
Equation 2,
pi = pw + Pd
combines fracture pressure measured at the surface (Pwfrac = 875 psig) with
the pressure due to the hydrostatic column within the tubing (Pd = 672 psig).
Pifrac = 875 psig + 672 psig = 1547 psi9
P.f is the total pressure, at the perforations, associated with the
fracture pressure applied at the surface.
Pressure at the perforations, induced by applying the maximum proposed
surface injection pressure (P^,.,., = 700 psig) is calculated in a similar
WIHoA
manner.
Pd = (12)(0.040){1400) = 672 psig
and with P = 700 psig = maximum proposed injection pressure,
WlMClA
Pimax = 70° ps1g + 672 psig = 1372 psl'9'
Pimax rePresents pressure, at the perforations, caused by a surface
injection pressure of 700 psig and is the maximum proposed injection pressure
at the same depth as the calculated fracture pressure. Comparing the proposed
injection pressure (P.jmax = 1372 psig) to the fracture pressure (Pifrac =
1547 psig), it can be seen that the company will be operating below the
pressure necessary to fracture the injection formation.
-273-
-------
INJECTION VOLUME LIMITATION
The proposed maximum injection rate is 2000 BWPD, and the well has been
in operation for approximately 5 years. Results from equations 3, 4, and 5
will quantitatively indicate:
(a) the injection zone pore volume available for fill-up;
(b) the amount of time it will theoretically take for the plume to
extend to the designated limit; and
(c) whether the proposed injection rate is consistent with the proposed
injection pressure.
Equation 3 is used to estimate the volume of fluid that is theoretically
necessary to fill up a subsurface cylinder, centered around the disposal well,
with a radius of 1/4 mile and height equal to the injection zone thickness.
This equation is similar to the equation used to calculate the volume of a
cylinder (volume = TTr h, where h = height of cylinder). The porosity term
represents the pore space of the cylinder.
Before equation 3
V = (TTr2bn)/5.6 (3)
can be used, however, several intermediate calculations must be performed, in
order to take into account the fact that the well has been operating for 5
years. If the disposal well was a newly constructed or converted well, these
calculations would not be necessary. According to the applicant, the average
injection rate over the 5-year period was 1300 BWPD.
Equation 4,
T = V/365q (4)
-274-
-------
which describes the length of time it will take to fill the pore volume, can
be rearranged to solve for pore volume filled during a known time period.
q, = average injection rate during specified time period = 1300 BWPD
a
V = subsurface volume filled during time period (barrels)
a
T = specified time period = 5 years
V = 365qaTa = (365)(1300)(5) = 2.3725x1 O6 barrels.
a a a
Naturally, if records for total volume of injected fluid are available,
they should be used for V,.
a
As a rule-authorized well, the facility was originally limited to an
injection plume extent of 1/4 mile from the wellbore. However, the equations
used in this analysis neglect the effects of salt-water dispersion and ambient
ground-water flow within the formation. Therefore, a safety factor is
incorporated into the analysis, in an attempt to acknowledge the phenomena of
molecular diffusion, mechanical dispersion, and regional flow.
5% of 1/4 mile is 66 feet or 0.0125 miles,
(0.25 - 0.0125) miles = 0.2375 miles = 1254 feet.
A value of 1254 feet will be used as the radial distance (r), instead of
the 1/4-mile value (1320 feet). Although for this situation a 10% safety
margin is considered adequate, it should not be viewed as necessarily
acceptable for all disposal operations.
Total subsurface volume available (Vt) is determined by substituting
the following values into equation 3: 1254 feet for "r", 50 feet for "b", and
0.30 for "n".
Vt = [(Tr)(1254)2(50)(0.30)]/5.6 = 1.323269xl07 barrels
-275-
-------
The subsurface volume, currently available (Vb) for fill-up, takes into
consideration the reduction of available pore volume caused by previous
injection.
Vt - Va = Vb = 1.323269x1O7 - 2.3725x1O6 = 1.08602x1O7 barrels
The proposed maximum injection rate (q.) is 2000 BWPD. Using the
calculated value for Vb, the length of time remaining for injection
operations (Tb) is calculated through the use of equation 4,
Tb = V365% = H. 08602x107)/(365) (2000) = 14.8 = 14 years.
Tb is rounded down to incorporate an additional margin of safety. The
value of 14 years is the theoretical length of time the company may continue
to operate the disposal well. In the permit application, the company proposed
an operating life of approximately 20 years; beginning from the date of
application. Unless additional technical data is submitted to prove
otherwise, the permit would be written such that it would expire no later than
14 years after issuance.
PRESSURE & RATE COMPATIBILITY
Equation 5 is used to assess the feasibility of injecting at the proposed
maximum rate and pressure, and is derived from a standard equation that
describes the drawdown curve for steady flow to a well in a confined aquifer
(Bear, 1979).
Q = (7.07kb[Pi - Pf])/m(ln re/rw) (5)
-276-
-------
This equation can be used to estimate the theoretical, maximum allowable
injection rate that would be operationally consistent with the proposed
injection pressures, and is an objective method for comparing theoretical rate
with proposed rate. In addition, equation 5 makes a comparison between
proposed injection pressure and the formation pressure.
If an applicant is unable to provide estimates of aquifer permeability,
the permit writer has several options. Referring to inhouse hydrogeologic
records, maintained for the same or similar formations, has proven to be
helpful. EPA UIC staff discussions further serve to guarantee that there will
be an ongoing exchange of new or pertinent information.
An operator will often submit a bottom hole pressure measured shortly
after the well was drilled. If the well has been in operation for some time,
this value will be an underestimate of current formation pressure. One
approach for estimating pore pressure is to assume that formation pressure is
equivalent to a hydrostatic column measured from the top of the perforations
and extending just to the land surface. For disposal zones that are not
highly pressured, this approach may overestimate pore pressure, but will
actually reduce the value of "Q" and provide a conservative limit for the
maximum allowable injection rate.
From the previous pressure calculations,
Pj = 1547 psig, and Pf = Pd = 672 psig.
Substituting 0.05 darcys for "k", 50 feet for "b", 0.4 cp for "m", 1254
feet for the adjusted "r ", and 0.333 feet for "r ",
C W
Q = ((7.07)(0.05)(50)[1547 - 672])/((0.4)[ln (1254/0.333)]) = 4695 BWPD.
-277-
-------
The proposed maximum injection rate (q = 2000 BWPD) is much less than the
calculated, theoretical maximum allowable injection rate (Q = 4695 BWPD).
However, this does not mean that the company would necessarily be allowed io
arbitrarily increase the proposed rate.
SUWIARY
This paper presents a mathematical approach for evaluating Class II brine
disposal permit applications. Certain physical processes associated with well
hydraulics and ground-water flow can be approximated through the use of
analytical models. Once numerical estimates are assigned to specific
variables, operating conditions can be evaluated in terms of compliance with
the UIC program. It should be remembered that mathematical equations are
tools to be used in conjunction with a qualitative review of all available
hydrogeologic information pertinent to the injection operation.
FUTURE WORK
Technical reviews of permit applications would be greatly enhanced
through the use of appropriate ground-water computer models. However, access
to documentable hydrogeologic data is often limited. In many cases, numerical
values for specific parameters must be approximated. Under these
circumstances, it is not appropriate to make use of data-intensive models,
particularly when the data base itself is based on generalized assumptions.
Time constraints and computer hardware capabilities limit programming choices,
further complicating the matter.
Fortunately, valid computer codes have been developed for almost any
hardware setup. It is hoped that in the forseeable future, a FORTRAN program
written by Hsieh (1986) will be incorporated into the permit application
-278-
-------
review procedures. Hsieh's program evaluates the analytical solution of the
radial dispersion problem by analyzing dispersive transport in radial flow
from a recharge/injection well. Most of the input items required for the
model are data that are regularly reviewed during an application evaluation.
In addition, the analytical solution is predicted to be computationally more
efficient than previous solutions.
-279-
-------
SCIENTIFIC TERMS
2
psig = pounds per square inch gauge (Ib/in )
3
pci = pounds per cubic inch (Ib/in )
BWPD = barrels of water per day
Pd = hydrostatic pressure at a specific depth (psig)
S = specific weight of injection or fracturing fluid (pci)
= (specific gravity of fluid)(specific weight of fresh water)
h = height of fluid column (feet)
12 = conversion factor for feet to inches (1 foot = 12 inches)
PW = fracture or injection pressure at the surface, or wellhead (psig)
PI- = pressure, due to injection/fracturing, at a specific depth (psig)
V = subsurface injection zone pore volume (barrels)
r = radial distance of injection plume limitation (feet)
b = thickness of injection zone (feet)
n = porosity of injection formation (dimensionless)
3 3
5.6 = conversion factor for ft to barrels (5.6 ft =1 barrel)
q = injection rate (BWPD)
T = time period to fill the injection zone pore volume (years)
365 = conversion factor for days to years (365 days = 1 year)
Q = theoretical injection rate (BWPD)
k = injection zone permeability (darcys)
Pf = injection zone pore pressure (psig)
m = viscosity of water (centipoise)
r = distance of theoretical plume limitation (feet)
r = well bore radius (feet)
7.07 = conversion factor
-280-
-------
METRIC CONVERSIONS
(Ib/in3)*(2.767990x104) = kg/m3
(psi)*(6.894757x103) = Pa
(centipoise)*(1.000000xlO~3) = Pa-second
(barrel )*( 1.58987 3x10"1) = m3
(ft3)*(2.831685xlO~2) = m3
(feet)*(0.3048) = meter
(darcy)*(9.870x10"13) = m2
(jiffy)*(3.3602x10"12) = sec/m
REFERENCES
Bear, Jacob. 1979. Hydraulics of Groundwater. McGraw-Hill Inc., New York,
569 pp.
Davis, Stanley N. and Roger J.M. DeWiest. 1966. Hydrogeology. John Wiley
& Sons, Inc., New York, 463 pp.
Freeze, R. Allan and John A. Cherry. 1979. Groundwater. Prentice-Hall, Inc.,
Englewood Cliffs, 604 pp.
Hsieh, Paul A. 1986. A New Formula for the Analytical Solution of the Radial
Dispersion Problem. Water Resources Research, volume 22, number 11,
October, pp. 1597-1605.
Mott, Robert L. 1979. Applied Fluid Mechanics. Charles E. Merrill Publishing
Co., Columbus, 2nd Edition, 405 pp.
Nielsen, David M. and Linda Aller. 1984. Methods for Determining the
Mechanical Integrity of Class II Injection Wells. National Water Well
Association, Worthington, OH, July, Report No. EPA-600/2-84-121, 263 pp.
-281-
-------
Table 1. Hydrogeologic & Operational Parameters
parameter
injection zone permeability
viscosity of fluid
maximum proposed surface injection pressure
fracture pressure measured at the surface
maximum proposed injection rate
arbitrary radial plume limitation (1/4 mile)
specific gravity of fluid
previous average injection rate
injection zone thickness
porosity of injection zone
depth of perforations
well bore radius
proposed operating life of the well
numerical value
0.05 darcys
0.4 centipoise
700 psig
875 psig
2000 BWPD
1320 feet
1.107
1300 BWPD
50 feet
0.30
1400 feet
0.333 feet
20 years
-282-
-------
UIC PROGRAM
I
Class I
CLASS II
Class III
Class IV
Class V
BRINE DISPOSAL
I
enhanced recovery
PERMIT APPLICATION INFORMATION
well
construction
details
p&a
plan
OPERATIONAL DATA
financial
demonstration
EXISTING
new
area
of review
corrective
action
INJECTION RATE-
INJECTION PRESSURE-
OPERATING TIME •
RICAL
NG DATA
OPERATING'PARAMETERS HYDROGEOLOGY
MAXIMUM INJECTION PRESSURE
MAXIMUM INJECTION RATE.
OPERATING TIME •
RADIUS OF INFLUENCE
POROSITY'
PERMEABILITY'
THICKNESS•
DEPTH-
GEOLOGIC DESCRIPTION-
FRACTURE PRESSURE -
PORE PRESSURE-
FLUID DATA
I
INJECTION
FLUID
injection
zone
SPECIFIC
GRAVITY
I
VISCOSITY
INJECTION ZONE
confining zones-
total
dissolved
solids
(tds)
content
- geologic
description
• fracture
pressure
— depth
— thickness
Figure 1. Generalized Flow Chart of Information Requirements (capitalized items are
discussed in paper).
-283-
-------
'////S
INJECTION ZONE
Figure 2. Injection well schematic (after Nielsen and Aller, 1984, p. 18)
-284-
-------
1200-
i
800-
600-
I
400-
200-
0
8 75 PSIG
INJECTIOH RATE, IH
BARRELS PER MINUTE fSPMJ
8PM
0.59
1.80
3.35
4.02
4.76
5.10
PSIG
240
560
930
1015
1160
1187
in
oo
oj
\
\ \
0 1
Figure 3. Step-rate test results.
\
5
-------
THE USE OF CONTROLLED SOURCE AUDIO
MAGNETOTELLURICS (CSAMT) TO DELINEATE ZONES
OF GROUND WATER CONTAMINATION - A CASE HISTORY
By Richard M. Tinlina, Talib Syedb, Steve Figginsc,
and A. Roger Anzzolin*1
aVice President
Geraghty & Miller, Inc.
3322 E. Fort Lowell Road, Tucson, Arizona 85716
°Geophy s i c i s t
Zonge Engineering and Research Organization
3322 E. Fort Lowell Road, Tucson, ARizona 85716
Project Officer
Office of Drinking Water
U.S. Environmental Protection Agency
Washington, D.C.
ABSTRACT
A significant potential for the pollution of fresh water
aquifers exists due to oil-field water-flood operations. The
sources of potential pollution are surface spills, a lack of
mechanical integrity of injection wells, and improperly plugged
wells which are in communication with the injection zone.
Surface spills are relatively easy to detect and control. Pro-
cedures for checking the mechanical integrity of a properly
constructed injection well are available. Making a determination
in the absence of good records as to whether or not a well is
improperly plugged, providing a conduit for the vertical migra-
tion of formation brines from the production zone to shallower
fresh water aquifers, is very difficult.
Electrical surface geophysical methods offer considerable
promise in detecting the movement of formation brines into
-286-
-------
fresh water aquifers, through improperly abandoned or plugged
wells.
An electrical surface geophysical technique. Controlled
Source Audio-Frequency Magnetotellurics (CSAMT) has been applied
to locate the presence of anomalies resulting from the upward
movement of formation brines through improperly plugged wells.
The primary objective in a CSAMT Survey is to provide apparent
resistivity and the phase angle between the electric and magnetic
fields over a prospect area. CSAMT has the advantages of
excellent lateral resolution, good depth penetration (a kilometer
or more) and is relatively inexpensive. The frequency and
resistivity of the subsurface controls the depth of penetration.
The lower the frequency, the deeper the penetration.
A CSAMT Survey was run in an oil producing field in east
central Oklahoma which is currently on waterflood and has many
abandoned and apparently improperly plugged wells. The water in
the Vamoosa aquifer underlying the study area has a high chloride
content. The objective of running the CSAMT Survey was to locate
suspected low resistivity anomalies due to formation brines in
the vicinity of improperly plugged wells and to attempt to map
their extent.
Introduction
The study area includes approximately 324 hectares (800
acres) in the Sac and Fox Reservation located in Lincoln County,
east central Oklahoma (Figure 1) . The area is underlain by the
Vamoosa Formation, which consists of alternating thin to massive
-287-
-------
OO
00
I
SAC & FOX
RESERVATION
R6E
SAC & FOX UNIT
STUDY AREA LOCATION
Lincoln County, Oklahoma
Figure 1
-------
sandstones and sandy-silty shales. The sandstone layers are fine
to coarse-grained and provide a reservoir source for one of the
major fresh water aquifers in Oklahoma.
Oil production in the study area began in the 1930's. The
unit's cumulative production through August 1982 is approximately
28 million barrels, with an estimated current monthly production
of 4500 barrels. Water injection for secondary recovery and/or
salt water disposal purposes began in the 1950's and the
cumulative water injected through August 1982 is approximately 75
million barrels, with an estimated current monthly injection of
48,000 barrels.
The major objectives of the ground water contamination study
included the determination of the cause of the high chloride
concentrations in the Vamoosa aquifer underlying the study area
and whether this resulted directly from oilfield activities on
and around the study area. A large number of well logs and
plugging and abandonment records were evaluated and several tests
conducted to determine the source of the high chloride waters in
the Vamoosa aquifer underlying the study area. Test holes were
drilled and logged to obtain ground water samples and to
determine the ground water quality profile at the test sites. In
addition, an electrical surface geophysical survey (Controlled
Source Audio Magnetotellurics - CSAMT) was run in order to locate
the presence of anomalies that might result from the upward
movement of formation brines high in salt content through
-289-
-------
improperly plugged wells into the overlying Vamoosa fresh water
aquifer.
Background Studies
Hydrogeology
The surface geology of the area is part of the Ada Group.
Underlying the Ada Group is the Vamoosa Formation. The Vamoosa
aquifer includes the Vamoosa Formation and underlying and
overlying Pennsylvanian formations that are 1ithologically
similar and hydrologically interconnected. The Vamoosa aquifer
consists of a complex sequence of fine to very fine grained
sandstone, siltstone, shale, and conglomerate, with interbedded
very thin limestones. The water-yielding capabilities of the
aquifer are largely controlled by the lateral and vertical
distribution of the sandstone beds and their physical
characteristics (D'Lugosz et al, 1978). Figure 2 illustrates the
hydrogeology underlying the area and was prepared from driller's
logs and plugging records of five abandoned wells. The
orientation of the geologic cross-section is approximately
southwest to northeast (Tinlin et al, 1984).
Earlier investigations (Hart, 1966) reported that the base
of the fresh water beneath the area could range from 50 meters to
more than 150 meters below ground level. An evaluation of
electrical resistivity logs run on oil wells in and around the
area shows the base of the fresh water to be in the range of 40
to 90 meters below ground level. A base of fresh water contour
map (Figure 3) was drawn utilizing data from resistivity logs and
-290-
-------
SAC & FOX GROUNDWATER CONTAMINATION STUDY
• Approx. 3000'
I
NJ
0-
500-
1000-
1500-
2000-
2500-
3000-
3500-
Sac. 10 Sae. 16 Sac. 16 Sac. 10 Sac. 18 Sao. 19
No. 9 No. 4 No. 3 No. 2 No. 1 No. 18
a>
8
CO
3
0
c5
a
to
3
'-' '"i •-'~~~^.
^-Ttj~Li7j^j^rurij^_r^rir;
Sao.
L No.
-
HI-vvir^HHHjiHH
^ l ^ ', l ' ' • ^
19
5
V Prua Sandatona
lonzonlal Scale: 1' - 500'
/ertical Scale 1' - 500'
HOT6: Qaologlc Formatlona ara Pannaylvanlan In Aga. «.UM., „=« «Brt, ««,^ ™«oo o=^,.«..
GENERALIZED GEOLOGIC CROSS SECTION
Lincoln County, Oklahoma TUN - R6E
-500
-1000
-1500
-2000
-2500
-3000
L3500
Figure 2
-------
SAC & FOX GROUNDWATER CONTAMINATION STUDY
R6E
UEQEND
A^»A' Crou Section Location
—250—Contour truetvrt
• ••••• Boundary at A<*a
• Ten Well Drilled In 1M3
EEI Tett Well OrlDwl In 1970
©
O
I
f-O
Ni
I
S»c 1 Fox T««t WWK
Drlltod In 1879
EEI T*M Vttol 1963
Othar W»ll Control
4
N
Seal* I - 1/4 ml*
Conlout Inlirvat -SO'
Contour* and Control Polnli
a/« in Fe»l Below Ground Level
10 It. - Ground Uvel Datum)
BASE OF FRESH WATER CONTOUR MAP
Lincoln County. Oklahoma
Figure 3
-------
from test wells drilled earlier in the study area (Tinlin et al,
1984). The contour map shows the base of fresh water to be
relatively shallow over a large portion of the area.
Ground Water Quality
A study (Bingham and Moore, 1975) showed that the quality of
the ground water adjacent to the area to be generally good with a
IDS of 500 mg/L or less.
In the spring of 1979, four test holes were drilled by
Engineering Enterprises, Inc. in order to determine the
ground water quality. The test well data showed an anomalous
occurrence of salt water in a portion of the Vamoosa aquifer
that was expected to contain only fresh water. Three
possibilities as to the cause of the high chlorides in the
Vamoosa were postulated; (1) natural occurrences of salt
(halite) in the Vamoosa, (2) upconing of the fresh water - salt
water interface due to overpumping of the Vamoosa aquifer, and
(3) accidental introduction of salt water into the aquifer due to
various phases of oil field activities. The Vamoosa and
associated rock units do not contain salt beds nor does the
sedimentary environment in which the Vamoosa was deposited allow
for the development of bedded salt. Overpumping of the Vamoosa
aquifer was also ruled out as the recharge rate of the Vamoosa is
much higher than the pumping rates of the existing supply wells.
-293-
-------
Determination of Salt Water Contamination Source in the Vamoosa
by Water Sampling
Water samples from a test well (located in Section 22}
drilled under the supervision of Engineering Enterprises, Inc. in
July 1983 were analyzed by Dr. Donald 0. Whittamore of the Kansas
Geological Survey. Whittamore's (1983) procedure is very
effective in distinguishing oilfield brines from halite (rock
salt) solution brines and thus can determine which may be the
source of ground water contamination.
Bromide concentrations were determined by Dr. Whittamore and
the chloride concentrations were determined by Environmental
Control Laboratory in Norman, Oklahoma. The results of the water
analyses are presented in Table 1.
The bromide/chloride ratio is the key in determining the
brine contamination source. The ratios in the saline waters are
what could be expected if oilfield brine pollution had occurred.
Bromide/chloride ratios for most Oklahoma oilfield brines range
from 0.003 to 0.01. The bromide/chloride ratio expected for
waters with chloride concentrations of 10,000 mg/L for a halite-
solution source of salinity is 0.0002 + 0.0002. The
bromide/chloride values are much higher than the expected values
for a halite-solution source and fall within the values for most
Oklahoma oilfield brines. Since this situation fits the special
case where the injected fluid is the same or very similar to the
Prue formation fluid, the source of salinity in the Vamoosa
ground waters beneath the area was concluded to be Prue formation
brine.
-294-
-------
TABLE 1: CONSTITUENT CONCENTRATIONS AND RATIOS
FOR VAMOOSA WATER SAMPLES
Sample No.
1
2
3
Depth, meters
42
61
79
Cl, mg/1
200
7,520
10,800
Br, mg/1
0.78
33
50
Br/Cl
0.0039
0.0044
0.0050
-295-
-------
In December of 1983, two samples of Prue formation brine
were obtained from a surface separator on site, and the bromide/
chloride ratios determined by Dr. Whittamore. The purpose of
this testing was to obtain a comparison of the Prue brine
bromide/chloride ratios with the previous Vamoosa ground water
bromide/chloride ratios of July, 1983. The bromide and chloride
concentration and bromide/chloride ratios are listed in Table 2.
The bromide/chloride ratios from the Vamoosa aquifer closely
match the bromide/chloride ratios of the brine in the Prue
Formation. Due to this close match of geochemical ratios, the
Prue oilfield brine is considered to be the most probable source
of brine polluting the Vamoosa aquifer.
Location and Evaluation of Improperly Plugged Wells
An evaluation of all plugged and abandoned wells in the
study and surrounding areas was made from Oklahoma Corporation
Commission records. A location map of these wells, including
identification of the properly and improperly plugged wells, is
shown in Figure 4.
A number of the wells were determined to be improperly
plugged. For a proper plugging operation in an area where
protection of the fresh water aquifer is important, at least four
downhole cement plugs should be installed in every well that is
to be plugged and abandoned. These four plugs are a bottom plug
opposite the injection interval, an isolation plug across the top
of cut casing, a surface casing protection plug, and a surface
plug.
-296-
-------
TABLE 2: CONSTITUENT CONCENTRATIONS AND RATIOS
FOR PRUE OILFIELD BRINES
Sample No. Cl, mg/1 Br, mg/1 Br/Cl
Prue 1 74,200 377 0.0051
Prue 2 74,600 372 0.0050
-297-
-------
© Producing Oil Well
0 Oil Well - Properly Plugged
0^ Oil Well - Improperly Plugged
A Infection Well
{& Injection Well - Properly Plugged
Surface Rights Only
-298-
SAC & FOX UNIT
Lincoln County, Oklahoma
Figure 4
-------
Many of the older wells were plugged by loading the hole
with mud. With time mud can settle out and allow channeling of
salt water through the borehole. The cement plugs in most of the
older wells were determined to be inadequate. Many of the wells
had only the top surface cement plug and no additional downhole
cement plugs. In addition, in most of the plugged wells the
surface casing was set too shallow. Surface casing should be set
below the base of fresh water and cemented all the way to the
surface to effectively seal fresh water zones from deeper
injection fluids. Although these plugging methods satisfied
regulations at the time, they are inadequate as they provide
potential flow paths for upward migration of reservoir and/or
injection fluids to shallower fresh water zones.
Controlled Source Audio-Frequency Magnetotelluric Survey
Controlled source audio-frequency magnetotellurics (CSAMT)
is a relatively new technique first used on a consistent
commercial basis in 1978 by Zonge Engineering of Tucson, Arizona.
The CSAMT technique is similar to the conventional audio-
frequency magnetotelluric (AMT) method, with the exception that a
fixed current source is substituted for a natural earth-telluric
source resulting in a fixed, dependable signal.
The primary objective in a CSAMT Survey is to provide
apparent resistivity and phase angle soundings over a prospect
area. The technique is particularly effective at identifying
buried, conductive features. It has been successfully applied in
hydrocarbon exploration, mineral exploration, and geothermal
-299-
-------
exploration, and mapping of EOR fronts (steam front). In this
particular case the application of interest was to use the CSAMT
to locate suspected low resistivity anomalies in the vicinity of
wells thought to be improperly plugged and abandoned.
CSAMT has several advantages over the other geophysical
methods. It has good lateral resolution, good depth penetration,
is fast and relatively inexpensive, $1000 to $2000 per line
kilometer (or approximately $200 per station), depending on the
receiver dipole spacing used. It is also relatively insensitive
to "cultural" features such as pipelines, power lines, fences,
well casings, etc. Disadvantages associated with CSAMT include
difficulty in data interpretation due to near-field effects and
difficulty in estimating depths to anomalous two and three-
dimensional features without extensive computer modeling and some
geologic input. It is important to consider the specific
requirements of a field project before deciding whether or not to
use CSAMT.
CSAMT Layout in Study Area
A typical layout for a CSAMT Survey is shown in Figure 5.
The large transmitter dipole is located as far away from the
receiver dipole as is practical - usually three skin depths at
the lowest frequency being used. Skin depth (6) is related to the
signal penetration into the ground, and is defined as:
-300-
-------
Controlled source AMI
400 Cycle Engine
co
o
NOTE: Not to Scale
Current Electrodes
Potential Electrodes
AMTCoi
LAYOUT FOR
CONTROLLED SOURCE AMT SURVEY
Figure 5
-------
Skin depth (
-------
EXPLANATION
f> Water Injecnan Well
• Oil Well
-T T- Powerlme
Pipeline
Fence
Sac 8 Fo« Boundary
NOTE a =200 , Deormg N60E
BASE MAP Bl A/Tenneco niap supplied uy
Engineering Enterprises
T.UW.II o Line 783
Figure 6 - Layout of Lines for CSAMT Survey.
-------
5 W W
18.0 17.0 16.0 15.0 1M.0
13.0
11.0 ie.0
9.0
B.0
7.0
—I N 6« E
COQNIRRD RESISTIVITY
values in
13.8
1C2U Hz
512 Hz 4
250 Hz 4
12* Hz 4
I
OJ
o
*-
i
W Hz J-
24U8 Hz
1«2U Hz
512 Hz
256 Hz
4 ize HZ
5.S
6U Hz
22 Hz
CQ
C
CD
UW»ITM1IC CONTOURS ( Intcrvah «.!«)
i:?'
2.51
3.16
3.96
S.fl
6.31
T.SK
It.t
12.6
2ong* « 347
Plot bu CPLOT 3F
Plotted BPR 16 IB
-------
resistivity of the ground underlying line 3 (in ohm meters)
plotted against the frequency (in Hz) at each station on line 3.
The same data is also shown in Figure 8 except that the apparent
resistivity is plotted against depth. A pseudo-section such as
Figure 8 can be viewed as showing relative depths in an
approximation to a vertical slice through the ground.
The depth of penetration in a layered environment can be
estimated from the following MT equation:
Depth of penetration = 356 /_£_ meters
V f
p = ground resistivity in ohm meters
f = frequency in Hz
Figure 7 shows the vertical low resistivity anomaly
extending laterally between stations 10 to 16 and extending to
depth below. The low resistivity anomalies are more significant
at approximate depths of 14, 40, and 80 meters at station 13 and
at an approximate depth of 13 meters at station 11. The low
resistivity (high conductivity) anomalies can be identified more
clearly in Figure 8. The low anomaly extends to a depth of at
least 82 meters below station 13. These significant low
resistivity anomalies are most likely caused by improperly
abandoned and plugged wells. However, this will need to be
verified and confirmed in future studies through detailed test
drilling and water quality sampling.
Figure 9 is a horizontal view at a depth corresponding to 64
Hz. At station 13 this frequency corresponds to a depth of
approximately 63 meters.
-305-
-------
17.0 16.6
iz.t 11.1 t«.a
6.*
S £• H I
!.» >-i~ ...&.,
• ^r*l"5^ '•*^*^** X LI \ • « « ».•
J'* \ J.V^Tt? /J \'»*f X"' /'! !»'•' !\ M UOMIIIH
^i i»:r ••?,«(•»;« )vi s »•• / i'/ 7 v^ ,«
"'• /.;-,w v- ,y.. .- cf -^ s
•: V , l,s, ....
> al ' _.. . I . . .
DESISTtVITY
in
uewirwic canouu i
.H
•;»
t.l
1.3*
I.H
v A
M 4.3
r 1 i-s
I——I J.I
t.t
luat » M>
f loTow CfUJI 3H
riciiM MC M it
Figure 8 - Vertical Pseudo-Section Line 3 Resistivity versus Depth.
-306-
-------
CflGNIflRD RESISTIVITY at 64 Hz ior lines 1 to 5
o
1
7.8
in
FEET
< = = = = increcTs'ing station numbers
.8.2
o. u
-------
In Area of Current Injection Activity
In the area of current injection activity, line 8 was
selected for presentation. Figure 10 is a vertical pseudo-
section showing apparent resistivity versus frequency while
Figure 11 plots apparent resistivity versus depth. Note the low
resistivity anomaly extending to depth below station -4.0, Figure
10, and a depth of 114 meters below station -4.0, Figure 11. It
is particularly interesting because it spreads out near the
surface and the apparent resistivity has the lowest value near
the surface, indicating possible surface or near surface spills
or leaks.
Figure 12 is a horizontal view at a depth corresponding to
a frequency of 64 Hz. This corresponds to a depth of 63 meters
at station -4.0.
Test Drilling of CSAMT Anomalies
Two test well sites in line 8 were selected on the basis of
the CSAMT results. Test well #1 was located at station -2.0,
line 8 because of the contrasting resistivites encountered with
depth. Test well #2 was located at station -4.0, line 8 where
the surface resistivity is the lowest and the conductive plume
appears to be deep seated.
The stratigraphic sequence of rock penetrated in both wells
drilled in July, 1984 is shown in Table 3. Rock cuttings were
logged continuously during the drilling of each well. The first
good water bearing zone encountered was in the Vamoosa, in the
interval 120-150 feet in test well #1, and in the interval 113-
-308-
-------
Ml 63
1.0
U)
o
(Q
c
—»
CD
S 60 N
0.0 -i.e
-2.0
—i—
-3.0
i2« HZ
-1.0
—t—
-5.0
1 N S0 E
-r utas HI
812 Hi
25fl Hi
129 Hi
eu HI
32 Ui
COCNIflRD RESISTIVITY
values In ohn-nelers
LOCflRITHMIC CONTOURS I Interval, «.!« I
2.M
2.51
3. IE
3.08
S.tl
6 31
7.8M
Zonoe » 347
Plot by CPLOT 3F
Plotted APR 16 li
-------
U«ll S3
L.9 9.9 -1.8 -Z.9 -3.8 -4.9 -5.9
3 69 H I 1 1 1 1 1 1 N 69 E
CflGNIfKD RESISTIVITY
values in ohm-meters
UWWITHMIC CONTOURS ( Interval, 0.14)
2.W
2.51
3.16
3.08
5.41
S.31
r.w
19.9
12.6
SWING LEGEND
^^ I SO
2-51
3.16
3.98
9.5
r
5.1
Zonge » W7
Plot by CaOT 3H
Plotted DEC 31 I98H
-310-
Figure 11
-------
ZONGE ENGINEERING & RESEflRCH ORGflNIZflTIGN
COGNIflRD RESISTIVITY qt SU- Hz for Linas 7.8. and 9
SAC 8. FOX PROJECT
-------
Table 3 Stratigraphic Sequence of Rock Penetrated in July, 1984 Test Wells
Alluvium Ma Group Vamoosa Formation
Test Well #1 0-9 m. 9-23 m. 23-91 m.
Test Well #2 0-10 m. 10-20 m. 20-91 m.
-312-
-------
137 feet in test well #2. Water samples were collected from a
shallow, intermediate, and deep zone in each well and the
composition and bromide/chloride ratios determined. Results of
the analyses are shown in Table 4.
The bromide/chloride ratios are similar to the ones obtained
earlier and are within the range of typical Oklahoma oilfield
brine which is 0.003 to 0.01. This indicates that the probable
source of polluted Vamoosa ground water is Prue oilfield brine
regardless if it came from surface spills, improperly plugged
wells, or mechanical integrity failure of injection wells. After
logging and testing operations, both wells were completed as
monitoring wells to provide future monitoring data of the ground
water.
Conclusions
The CSAMT technique has been successful in locating low
resistivity anomalies which appear to result from oilfield brines
in the vicinity of improperly plugged wells and near active
injection wells. Two deep conductive plumes and one shallow
conductive feature were detected and traced in the area of active
injection wells in Section 15, while on the grid in Section 16 in
the area of improperly plugged wells, one localized deep feature
and two deep plumes were detected and traced. Several of these
plumes ran out of the edge of the survey grids and their extent
is not known.
Test drilling confirmed the low resistivity anomalies
detected by the CSAMT Survey. Future studies of this type should
-313-
-------
Table 4 Constituent Concentrations and Ratios for Vamoos Water Samples
(July. 1984 Test Wells)
Well
No.
I
2
Sample
No.
1
2
3
1
2
3
Depth,
m.
40
61
87
38
56
90
Cl,
mg/L
950
1,180
1,200
650
664
643
Br,
mg/L
4.9
5.9
6.1
3.5
3.6
3.4
Br/Cl
0.0052
0.0050
0.0051
0.0053
0.0054
0.0053
-314-
-------
include sufficient test drilling and logging to gather data to
permit comparison and correlation of known salinity contrasts
with resistivity contrasts. Once the baseline contrasts for a
given area or oilfield are established, detecting and tracing of
conductive plumes might be combined with estimates of water
salinity to give even more useful results.
ACKNOWLEDGEMENTS
Partial funding for the Vamoosa ground water contamination
study (including the CSAMT Surveys) was obtained from the U.S.
Environmental Protection Agency as part of Project 68-01-6389 of
the Underground Injection Control Program and is gratefully
acknowledged. The cooperation of the Sac and Fox officials
including Mr. Truman Carter is also acknowledged with thanks.
REFERENCES
Bingham, R.H. and Moore, R.L., 1975. Reconnaissance of the Water
Resources of the Oklahoma City Quadrangle, Central Oklahoma:
Oklahoma Geological Survey, HA-4.
D-Lugosz, J. , and McClaflin, 1978. Geohydrology of the Vamoosa
Aquifer, East-Central Oklahoma: U.S. Geological Survey
Open-File Report 78-781.
Fryberger, J.S., and Tinlin, R.M., 1984. Pollution Potential
from Injection Wells via Abandoned Wells, presented at First
National Conference on Abandoned Wells: Problems and
Solutions, Norman, Oklahoma, May 1984.
Hart, D.L., 1966. Base of Fresh Water in Southern Oklahoma, U.S.
Geological Survey, Hydrologic Investigations Atlas HA-223.
Whittamore, D.O., 1983. Geochemical Identification of Salinity
Sources in Proceedings of International Symposium on State-
of-the-Art Control of Salinity; Ann Arbor, Michigan.
-315-
-------
BIOGRAPHICAL SKETCHES
Dr. Richard Tinlin is a Vice President with Geraghty &
Miller, Inc. located in Tucson, Arizona. He has over 20 years
experience in hydrogeology and geophysics. He holds a Ph.D. from
the University of Arizona in Watershed Hydrology with a minor in
mineral exploration, and a B.S. in Geology from Arizona State
University.
Talib Syed graduated with a degree in Chemical Engineering
from the University of Madras, India in 1971 and earned a Masters
or Petroleum Engineering from the University of Oklahoma, Norman,
in 1983. He previously worked as a petroleum engineer for the
Arabian-American Oil Company in Saudi Arabia from 1974 to 1979
primarily in production, reservoir, completions and workovers,
both onshore and offshore. He currently is Program Manager - UIC
Projects for a nationwide contract providing technical assistance
to the EPA in the implementation of the UIC Program in the non-
primacy states.
Steve Figgins is a geophysicist with Zonge Engineering in
Tucson, Arizona. He obtained a B.S. in geophysics from the
University of Arizona in 1982 and has experience in CSAMT and CR
methods in Australia and the United States. He is present
working as General Manager of Zonge Engineering while pursuing an
MBA degree at the University of Arizona.
Roger Anzzolin graduated with a degree in geology from
Wichita State University, Kansas (1970) and a Master's degree in
-316-
-------
Environmental Engineering from Florida Institute of Technology,
Melbourne, Florida (1979). He worked with the U.S. Army for four
years on pollution abatement and installation restoration (IR)
projects. Since joining the Environmental Protection Agency in
1979. Roger has held numerous varied assignments. He also
manages the UIC Implementation Contract for regional program
support in Direct Implementation states, and chairs the
mechanical integrity workgroup in the branch for the UIC program.
Mr. Anzzolin is a member of the Groundwater Technology Division
of NWWA.
-317-
-------
CONVECTIVE CIRCULATION DURING SUBSURFACE INJECTION OF LIQUID WASTE
John J. Rickey
U.S. Geological Survey, Tampa, Florida 33634
ABSTRACT
Injection of liquid waste into a highly transmissive, saltwater-bearing,
fractured dolomite provided an opportunity to study density-dependent flow
associated with two miscible and density-different liquids. Mean chloride
concentration of the injectant during two tests of 91 and 366 days duration
was 180 and 170 milligrams per liter, respectively; whereas chloride concen-
tration of native saltwater ranged from 19,000 to 20,000 milligrams per liter.
During the 366-day test, chloride concentration in water from a well open to
the upper part of the injection zone 223 meters from the injection well
approximately stabilized at about 3,900 milligrams per liter. Approximate
constancy of chloride concentration in water from this observation well at a
level significantly greater than the injectant concentration suggested the
hypothesis that convective circulation with saltwater flow added chloride ions
to the injection-zone flow sampled at the observation well.
In order to assess the acceptability of the convective circulation
hypothesis, information was required about the velocity field during injec-
tion. Mass-transport model simulations were used to provide this information
-318-
-------
after determining that the fractured injection zone could be treated as an
equivalent porous medium with a single porosity. The mass-transport model was
calibrated using the 91-day test data from the observation well 223 meters
from the injection well. The model was then run without parameter changes to
simulate the 366-day test. Mass fractions of injectant computed for observa-
tion wells during the 366-day test compared favorably with observed mass
fractions. Observed mass fractions were calculated as a function of chloride
concentration and density. Comparisons between model-computed mass fraction
and velocity fields in a radial section showed convective circulation with
saltwater flowing toward the injection well in the lower part of the injection
zone, then mixing with the injectant, and the mixture flowing away from the
injection well in the upper part of the injection zone. Based upon the model
results and the assumed reasonableness of treating the injection zone as an
equivalent porous medium with a single porosity, the hypothesis of convective
circulation during subsurface injection of liquid waste into a highly trans-
missive, saltwater-bearing, fractured dolomite was judged acceptable.
INTRODUCTION
Subsurface injection of liquid waste into a highly transmissive,
saltwater-bearing, fractured dolomite provided an opportunity to study
density-dependent flow associated with two miscible and density-different
liquids. Two injection tests were run at a site in the city of
St. Petersburg, Florida (Figure 1). The first test was run for 91 days and
-319-
-------
86°
i
84°
i
ST. PETERSBURG
TEST SITE
0 10 20 MILES
82°
i
80°
i
-30°
-28C
-26°
0
32 KILOMETERS
Fig. 1. Location of test site.
-320-
-------
the second test for 366 days. During the second injection test, chloride
concentration in water from an observation well open to the top of the
injection zone at 223 m from the injection well changed slowly after passage
of the injectant front and, toward the end of the test, became approximately
stable at concentrations (3,900 mg/L) significantly above the concentration of
the liquid-waste injectant (170 mg/L). The liquid waste was treated municipal
sewage similar in composition and density to freshwater. Native saltwater in
the injection zone was similar in composition and density to seawater. The
observed chloride-concentration data raised questions about the transport
processes that occurred in the injection zone during injection. Approximate
constancy of chloride concentration in water from the observation well at a
level significantly above the injectant concentration suggested the hypothesis
that convective circulation with saltwater flow, related to a convection cell,
added chloride ions to the injection-zone flow sampled at the observation well
(Hickey and Ehrlich, 1984).
Convective circulation in variable-salinity ground-water flow has been
discussed as a possibility by several other authors. Flow of saline water
opposite to the flow of overlying and less dense fresher water in an isotherm-
al, variable-salinity ground-water environment was discussed by Cooper (1959)
and Cooper et al. (1964). Cooper theorized that saltwater would flow landward
in a coastal aquifer in response to dispersion of saltwater into seaward flow-
ing freshwater. A similar idea was proposed by Carrier (1958). Hubbert
(1957) and de Josselin de Jong (1969) concluded that circulation would occur
in a variable-density ground-water environment as a result of density gradi
ents that were related to salinity variations.
-321-
-------
Even though the mechanisms and resulting flow patterns postulated by some
of the above mentioned authors differ, they all are in agreement that some
form of convective circulation, related to a convection cell, may occur during
variable-salinity ground-water flow. This article evaluates the hypothesis of
convective circulation as an explanation for the approximate constancy of
chloride concentration at a level significantly above the injectant concentra-
tion in water from an observation well during subsurface injection. To
achieve this purpose, the 91-day and 366-day injection tests are described.
Then, the hydrogeologic characteristics of the test site are described with an
emphasis on treating the fractured injection zone as an equivalent porous
medium with a single porosity. This is followed by a discussion wherein a
numerical, mass-transport model is calibrated with data from the 91-day test
and run without parameter changes to simulate the 366-day test. Finally,
model-computed mass fraction and velocity fields are compared and interpreted
with regard to the acceptability of the convective-circulation hypothesis.
INJECTION TESTS
Injection of treated sewage into a saltwater-bearing aquifer was tested
for 91 days in 1977 and for 366 days in 1979 and 1980 at a test site in the
city of St. Petersburg, Florida (Rickey, 1982; 1984b; Rickey and Ehrlich,
1984). During the 91-day test, the injection well became partially plugged by
algae in the liquid waste (Rickey, 1982). Mean injection rate for the 91-day
43 33
test was 1.54x10 m /d with a standard deviation of 1.28x10 m /d, and the
4 3
mean injection rate during the 366-day test was 1.33x10 m /d with a standard
deviation of 2.31x10 m /d. The mean injection rate for the 366-day test was
about 14% less than the 91-day test.
-322-
-------
473 m-
223 m
O
Cl
O
B2, B3
B6 O'
.INJECTION WELL
M.
OJ
3
100 METERS
O-
C3
Fig. 2. Areal configuration of wells open to the injection zone.
-323-
-------
Chemical composition of injected sewage was similar during both tests.
The mean concentration of dissolved solids during the 91-day test was 508 mg/L
and during the 366-day test was 466 mg/L. The mean density of the treated
3
sewage injectant was 999 kg/m for both tests, and mean chloride concentra-
tions were 180 mg/L and 170 mg/L during the 91-day and 366-day tests, respec-
tively. Chloride concentration of native water from the injection zone before
injection occurred was similar to seawater and ranged from 19,000 mg/L in the
upper part of the zone to 20,000 mg/L in the lower part. Density of the
3 3
native saltwater ranged from 1,025 kg/m to 1,026 kg/m .
Areal configuration of wells open to the injection zone at the test site
is shown in Figure 2. The injection-zone interval and water-producing inter-
val open to each observation well are shown in Figure 3. Wells B3, B6, Cl,
and C3 are open to the upper part of the injection zone, and the injection
well and well B2 are open to the lower part of the zone. Wells Cl and C3 were
constructed after the 91-day test to monitor the 366-day test for evidence of
anisotropy in the plane of the injection zone.
One month before the start of the 366-day test, chloride concentration in
water from well B3 was 14,000 mg/L; from wells B6, Cl, and C3, it was
18,000 mg/L; and from well B2, it was 20,000 mg/L. These data suggest that
chloride concentrations are stratified in the injection zone at the start of
the 366-day test.
Chloride concentrations in water from well B3 during the 91-day and 366-
day injection tests are shown in Figure 4. For the period common to both
tests, chloride concentrations are similar and show little or no influence
from plugging of the injection well. Similar concentrations during both tests
would be expected because the mean chloride concentration of injectant for
each test was very similar and the mean injection rate for each test differed
-324-
-------
ro
t_n
CO
ct:
GJ
e-<
UJ
ft 1
*--' '
2
^^
O
i — i
O
QJ
—l
~
O
CJ
in
tr.
en
Cd
"^>
O
en
(T-
ao -
84-
70-
56-
42-
28-
14-
n -
J B3 L INTERVAL THAT |
RR ' -~~-^^ PRODUCES WATER '
1 Bt> ! ^^SAMPLES ~*
-------
(jj
_J
UJ
Q_
CO
n
cc
ce:
o
I—I
cc
UJ
o
z
o
o
QJ
Q
i — i
ct:
o
o
£UUUU
16000
12000
8000
4000
n
V
V V V
* * 47 V* *
91-DAY-
B2
^^366-DAY
4T
^ V
t7
H^v
B3 '^
*
-------
by only 14%. During the 366-day test, chloride concentration slowly
decreased after passage of the injectant front and became approximately stable
at about 3,900 mg/L after 190 days from the start of the test. Because this
concentration was significantly larger than the injectant concentration (170
mg/L) , it appeared that chloride ions were somehow being added to the
injection-zone flow sampled at well B3. This consideration played an import-
ant role in developing the hypothesis of convective circulation with saltwater
flow (Hickey and Ehrlich, 1984). Convective circulation was thought to be
related to a convection cell similar to the type proposed either by Cooper
(1959) or by de Josselin de Jong (1969). Well B3 is emphasized in mass-
transport model calibration, discussed below, because it showed no apparent
influence from injection-well plugging, as did well B6.
Chloride concentrations in water from well B2 during the 91-day and 366-
day injection tests are also shown in Figure 4. During both tests, chloride
concentrations did not change and remained at the native chloride concentra-
tion of 20,000 mg/L.
HYDROGEOLOGIC CHARACTERISTICS OF THE INJECTION ZONE
The injection zone at the test site is composed of a dolomite that is in
the middle Eocene series. The injection zone is in the Upper Floridan aquifer
and is overlain and underlain by semiconfining beds (Figure 5) . The top of
the injection zone is at 234 m below sea level. Thickness of the zone is
about 98 m. Darcian flow occurs in the injection zone, at least beyond 11 m
from the injection well (Hickey, 1984a). Transmissivity of the zone at the
2
test site is about 75,000 m /d, whereas transmissivity of the zone beyond
-327-
-------
DEPTH BELOW LAND SURFACE, IN METERS
J) Ul A OJ PO —
D 0 0 000
D O 0 O O O O
-
HYDROGEOLOGIC
UNIT
Surficial aquifer
system
Intermediate
\ confining /
\ unit /
Upper
Floridan
aquifer
Middle
confining unit
of the
Floridan
aquifer system
semiconfining
bed
INJECTION
ZONE
semiconfining
bed
Fig. 5. Hydrogeologic section of the injection site.
-328-
-------
24 km and east of the site may be less (Hickey, 1981). The vertical component
of hydraulic conductivity of the overlying and underlying semiconf ining beds
is estimated to lie between about 0.03 m/d and 0.3 m/d (Hickey. 1982).
Regional saltwater flow toward the southeast likely occurs in the injec-
tion zone (Hickey, 1982) . However, the apparent stratification of chloride
concentrations prior to the 366-day test suggests that flow of native salt-
water in the neighborhood of the site was of little importance before and
during injection.
A total porosity of 14% for the rocks comprising the injection zone was
estimated in one borehole at the test site from geophysical logs (Hickey,
1982). Transport-model simulation of the arrival time of the injectant front
at well B3, 223 m from the injection well, discussed later, required an effec-
tive porosity of 10%. Both estimates are remarkably consistent considering
that the borehole geophysics estimate is based upon measurement of rock prop-
erties in the immediate vicinity of the borehole. Even though this comparison
of porosities may be fortuitous, it does suggest that porosity may be distrib-
uted more or less uniformly throughout the injection zone.
The injection zone generally can be characterized as a fractured crypto-
crystalline to microcrystalline dolomite with minor solution enlargement along
some of the fractures. The fracture pattern in the dolomite is very complex
and cannot be described by a simple repeating pattern of similarly spaced and
oriented fractures. A borehole television survey and a caliper log of a test
hole drilled into the injection zone showed the wall of the hole to be very
blocky with fractures between the blocks generally occurring about every 0.3
to 1.0 m. Cores taken from holes within 24 km of the site showed oblique
fractures oriented between 30 to 60 degrees from the axis of the cores. Some
of the cores also showed horizontal fractures. Added to this fracture pattern
-329-
-------
are shattered intervals that collapsed during drilling. A shattered interval
was cored at another site in the area (Hickey, 1977) and showed fracture
.3
spacing on the order of 5x10 m and less. Shattered intervals appeared at
different depths in holes drilled into the injection zone, not only at this
site but at other sites in west-central Florida. When encountered in a bore-
hole, shattered intervals were always water-producing intervals, although not
all water-producing intervals identified in boreholes were shattered (Hickey,
1982). Conceptually, the varying positions of shattered intervals and other
water-producing intervals, along with the blocky character of the injection
zone, suggest a branching network of numerous intersecting fractures. As
noted by Long et al. (1982), when fracture density is increased, when fracture
orientation is distributed (as would be the case with numerous intersecting
fractures), and when larger sample sizes of fractures are tested (as would be
the case with 6-m to 33-m thick water-producing intervals), fractured systems
behave like a porous medium.
Supporting the perception of a high fracture density are the observations
2
that the injection zone has a very high transmissivity (about 75,000 m /d) ,
yet the injection well became plugged during the first injection test at the
site (Hickey, 1982). In order to explain both of these occurrences, the
injection zone at the well would have to be composed of numerous fractures
with relatively small aperture rather than a few fractures with relatively
large aperture. Other injection wells within 24 km of this site also became
plugged during injection of secondarily treated sewage indicating that the
zone has similar physical characteristics throughout the area.
The injection zone also can be described as a collection of variably
sized crystalline dolomite blocks with bounding fractures. Visual examination
of the crystalline dolomite indicates that it generally has no visible
-330-
-------
porosity and thus should have very small, if measurable, hydraulic conductiv-
ity and effective porosity. Five representative cores of the crystalline
dolomite blocks taken from test holes within 24 km of the site have laboratory
measurements of hydraulic conductivity that did not exceed 2.0x10 m/d with
four of the five cores at or below the detection limit of the permeameter
(Hickey, 1977; Mickey and Barr, 1979). An additional 10 representative cores
of the crystalline dolomite blocks, also taken from holes within 24 km of the
site, have laboratory measurements of effective porosity that have a mean
value of 0.9% (Hickey. 1977; Rickey, 1979; Hickey and Barr, 1979). The
branching network of numerous intersecting fractures and the very small
hydraulic conductivity and effective porosity of the crystalline dolomite
blocks strongly suggest that the injection zone can be treated as an equiva-
lent porous medium with a single porosity.
CALIBRATION OF THE SWIP MASS-TRANSPORT MODEL USING
THE FIRST INJECTION TEST
The SWIP finite-difference mass-transport model (Intercomp, 1976; Intera,
1979) was calibrated using data from the 91-day injection test. The model, as
used in this article, solves for two dependent variables --pressure and mass
fraction of injectant--in two dimensional, cylindrical (r-z) coordinates under
isothermal conditions. Central-in-space and central-in-time finite-difference
equations were used in the numerical model. The reduced band-width direct-
solution procedure was used to solve the equations.
In addition to equivalent porous medium and single porosity assumptions,
other major assumptions in the mass - transport model application are:
-331-
-------
hydrostatic conditions prevail in the injection zone at the start of
injection; the injection zone is confined; hydraulic characteristics of the
injection zone are radially extensive; and flow during injection is
isothermal. The first assumption ignores background flow of the resident
saltwater and restricts model computations to effects caused solely by
injection. The second and third assumptions restrict model computations to
effects caused by injection within a confined, radially symmetric cylinder
that has uniform characteristics. The fourth assumption restricts model
computations to effects caused by constant temperature ground-water flow.
Because the injection zone, as discussed in the hydrogeology section, only
approximately satisfies these assumptions, it was hoped that the hydraulic
characteristics assumed for the model, although likely nonunique, would be
close enough to the actual characteristics of the injection zone such that the
major flow processes during subsurface injection would be approximately
simulated.
The original plan for calibrating the mass - transport model envisioned
using mass - fraction data observed in wells B2 and B3 and holding all
injection-zone hydraulic characteristics constant except for dispersivity,
which was to be varied. During the trial and error calibration process, a
minor alteration to the plan was necessary. In addition to varying dispersiv-
ity, porosity was changed, as mentioned above, from 0.14 to 0.1 to improve
comparisons between the observed and computed arrival time of the injectant
front at well B3. During the 91-day injection test, the front arrived at well
B3 sometime between 9 and 15 days from start of injection. Model runs using a
porosity of 0.14 consistently computed the front arriving between 15 and 20
days. After changing porosity to 0.10, the arrival time computed by the model
was consistent with the observed data.
-332-
-------
en
M
o
_
o
z
o
O
cc
ct:
1.0
0.3
0.6
0.4
0.2
CD
£ o.o
Q
(COMPUTED CL
OO
o o o
RVE IS COINCIDEN
n n o n nn
JP**
^
-------
Mass fractions of injectant computed by the calibrated model and observed
mass fractions of injectant at wells B2 and B3 are shown in Figure 6. The
model calibration, in addition to what has already been mentioned, mainly en-
tailed varying the longitudinal and transverse dispersivities in a trial and
error fashion until what was considered an acceptable fit between computed and
observed mass fractions occurred. Mass fractions computed at well B3 differed
from observed mass fractions by no more than 0.06 and were generally much less
than this, as can be seen in Figure 6. Also, mass fractions computed at well
B2 showed no changes and, as such, agreed with the observed data. Longitudi
nal and transverse dispersivities of the calibrated model were 2.85 m and 0.85
m, respectively. These values compare reasonably well with the longitudinal
(6 66 m) and transverse (0.66 m) dispersivities found by Segol and Finder
(1976) in their model analysis of saltwater intrusion into a highly transmis-
sive aquifer in southeast Florida.
SIMULATION OF THE SECOND INJECTION TEST
After calibration, the model was run without changing any of the param-
eters for the purpose of simulating the 366-day injection test. Figure 7
shows the distribution in a radial section of observed and computed mass
fractions of injectant at the end of the 366-day test. The leading edge of
the injectant front in the upper part of the injection zone is approximately
at the position of the 0.30 mass-fraction contour as it was for the 91-day
test. In general, the observed mass fractions of injectant shown in Figure 7
compare very favorably with the computed mass fractions. This is true even in
the immediate vicinity of the injection well where observed and computed mass
-334-
-------
Ol
I
CD
ct:
QJ
E-H
LJ
CJ
o
z
o
t—(
C-H
O
CD
OC
CD
CiJ
>
O
CD
d
QJ
93 -
m
JO
LINE OF EQUAL MASS
FRACTION OF INJECTANT
0\ 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600
\ DISTflNCE PROM INJECTION WELL,IN METERS
INJECTION WELL
Fig. 7. Observed and computed mass fractions of injectant in a radial section of the injection
zone at the end of the 366-day test.
-------
fractions for well B6 were, respectively, 0.91 and 0.98. Also shown in Figure
7 is the initial mass-fraction distribution that was used in the model at the
beginning of the injection-test simulation.
Because the observed and computed mass fractions of injectant compare
favorably with each other during the 366-day test, particularly at wells Cl
and C3 , this not only strongly supports treatment of the injection zone as an
equivalent porous medium with a single porosity, but also, that the model is
likely simulating the major flow processes that occur during subsurface injec-
tion. Thus, it appears that model computed velocity fields may be interpreted
with some degree of confidence.
MODEL-COMPUTED VELOCITY FIELD
Figure 8 shows the model-computed velocity field at the end of the 366-
day injection test. The velocity field shows convective circulation in a
radial section of the injection zone related to a convection cell. Flow
within the region of convective circulation is generally away from the
injection well in the upper part of the zone, whereas flow is generally toward
the injection well in the lower part of the zone. Separating these outward
and inward flows is a shear zone wherein velocity vectors are about oppositely
directed.
Figure 8 shows that buoyant or free convection occurs at radial distances
of less than 100 m from the injection well. This is consistent with the
interpretation of buoyant convection based upon the mass-fraction distri
butions shown in Figure 7. Beyond the region of convective circulation,
Figure 8 shows flow directed away from the injection well throughout the
injection zone.
-336-
-------
OJ
OJ
to
Q±
CO
f-n
Ld
E-
_J
GZ
U— 100 m/d-»|
EXPLANATION
PORE VELOCITY VECTOR. MAGNITUDE OF VECTOR IS PROPORTIONAL TO SHAFT
LENGTH. FOR LOW VELOCITIES, VECTORS HAVE NO MEASUREABLE SHAFT. IN
THESE CASES, TIP OF ARROWHEAD IS LOCATED AT THE NODE POINT. THE FIRST
COLUMN OF VECTORS IS SHOWN AT 6 METERS FROM THE INJECTION WELL.
THEREAFTER, EVERY FIFTH COLUMN OF COMPUTED VECTORS IS SHOWN
- •=* •=>- •>•
->:=- ;=-==-=-
^ZT-T^II
j(F^~* -=
^>^s«=-*=:-=:-=c -=^ -=.
^r_
-TV— oc«aKa*=s-=c-
-------
Comparison of Figure 7 with Figure 8 shows that downward flow in the
outermost band of convective circulation is in the vicinity but beyond the
leading edge of the injectant front. For the test, the leading edge of the
front is approximated by the intersection of the 0.3 mass-fraction contour
with the top of the injection zone. Comparison of these figures, in addition,
shows that counter flow directed toward the injection well in the lower part
of the zone is mostly saltwater.
Between 100 and 800 m from the injection well, model results for the 366-
day test also show that some of the counter flow of saltwater mixed with flow
away from the well at a rate of about 4,100 m /d. Also, saltwater with a rate
3
of about 5,600 m /d flowed past 100 m to subsequently mix with flow in the
neighborhood of the injection well. Because mostly saltwater was added to
flow occurring away from the injection well in the upper part of the zone, the
magnitude of mass fraction of injectant and the temporal rate of change in
mass fraction of injectant at points within the upper part of the injection
zone should be markedly influenced.
CONCLUSIONS
This article represents an effort to test the hypothesis of convective
circulation used to explain the magnitude and approximate constancy of chlo-
ride concentration in water from an observation well during subsurface injec-
tion of liquid waste into a saltwater-bearing, fractured dolomite. A numeri
cal model was constructed as a simplified representation of the very complex
fractured injection zone by assuming that it could be treated as an equivalent
porous medium with a single porosity. Observed mass fractions from two
-338-
-------
observation wells during a 91-day injection test were used to calibrate the
mass-transport model. After calibration, the model was run to simulate a 366-
day injection test. Comparisons between observed and computed mass fractions
for the 366-day test at five observation wells suggest that the model is not
only conceptually appropriate, but also does likely simulate the major flow
processes during subsurface injection. Convective circulation with counter
flow of saltwater in the injection zone was portrayed by the model as a major
process related to a convection cell that influenced the distribution of mass
fraction and, thus, chloride concentration in space and time. Because of
these model results and the assumed reasonableness of treating the injection
zone as an equivalent porous medium with a single porosity, the hypothesis of
convective circulation is judged acceptable for explaining the magnitude and
approximate constancy of chloride concentration in water from an observation
well during subsurface injection.
REFERENCES
Carrier, G. F. , The mixing of ground water and sea water in permeable
subsoils, Jour, of Fluid Mechanics, 4, 479-488, 1958.
Cooper, H. H. , Jr., A hypothesis concerning the dynamic balance of freshwater
and saltwater in a coastal aquifer, Jour. Geophys. Res., 64(4), 461-467,
1959.
Cooper, H. H. , Jr., F. A. Kohout, H. R. Henry, and R. E. Glover, Seawater in
coastal aquifers, U.S. Geol. Surv. Water-Supply Pap. 1613-C, 84 pp.,
1964.
-339-
-------
de Josselin de Jong, G. , Generating functions in the theory of flow through
porous media, Flow Through Porous Media, edited by R. J. M. De Weist,
pp. 377-400, Academic Press, New York, 1969.
Hickey, J. J., Hydrogeologic data for the McKay Creek subsurface waste-
injection test site, Pinellas County, Florida, U.S. Geol. Surv. Open-File
Rep. 77-802, 94 pp., Tallahassee, Fla. 1977.
Hickey, J. J. , Hydrogeologic data for the South Cross Bayou subsurface waste-
injection test site, Pinellas County, Florida, U.S. Geol. Surv. Open-File
Rep. 78-575, 87 pp., Tallahassee, Fla., 1978.
Hickey, J. J., Hydrogeology. estimated impact, and regional monitoring of
effects of subsurface wastewater injection, Tampa Bay area, Florida, U.S.
Geol. Surv. Water-Resour. Inv. 80-118, 40 pp., Tallahassee, Fla., 1981.
Hickey. J. J. , Hydrogeology and results of injection tests at waste-injection
test sites in Pinellas County, Florida, U.S. Geol. Surv. Water-Supply
Pap. 2183, 42 pp., Reston, Va., 1982.
Hickey, J. J., Field testing the hypothesis of Darcian flow through a carbon-
ate aquifer, Ground Water, 22(5), 544-547, 1984a.
Hickey, J. J., Subsurface injection of treated sewage into a saline-water
aquifer at St. Petersburg, Florida-- aquifer pressure buildup, Ground
Water, 22(1), 48-55, 1984b.
Hickey, J. J. , and G. L. Barr, Hydrogeologic data for the Bear Creek subsur-
face waste-injection test site, St. Petersburg, Florida, U.S. Geol. Surv.
Open-File Rep. 78-853, 53 pp., Tallahassee, Fla., 1979.
Hickey, J. J., and G. G. Ehrlich, Subsurface injection of treated sewage into
a saline-water aquifer at St. Petersburg, Florida--water-quality changes
and potential for recovery of injected sewage, Ground Water, 22(4), 397-
405, 1984.
-340-
-------
Rickey, J. J., and R. M. Spechler, Hydrologic data for the Southwest subsur-
face injection test site, St. Petersburg, Florida, U.S. Geol. Surv. Open-
File Rep. 78-852, 104pp., Tallahassee, Fla., 1979.
Hubbert, M. K., Darcy's law and the field equations of the flow of underground
fluids, Bulletin de Association Internationale d' Hydrologic
Scientifique, 5, 24-59, 1957.
Intercomp Resource Development and Engineering, Inc., Development of model for
calculating effects of liquid waste disposal in deep saline aquifers,
parts I and II, Rep. USGS/WRI-76-61, PB 256 903, 236 pp., Reston, Va.,
1976.
Intera Environmental Consultants, Inc., Revision of the documentation for a
model for calculating effects of liquid waste disposal in deep saline
aquifers, U.S. Geol. Surv. Water-Resour. Inv. 79-96, 72 pp., Reston, Va.,
1979.
Long, J. D. S., J. S. Remer, C. R. Wilson, and P. A. Witherspoon, Porous media
equivalents for networks of discontinuous fractures, Water Resources
Res., 18(3), 645-658, 1982.
Segol, G. and G. F. Pinder, Transient simulation of saltwater intrusion in
southeastern Florida, Water Resources Res., 12(1), 65-70, 1976.
-341-
-------
Monitoring, Troubleshooting and Repairing Wellbore
Communication of Waterflood Injection Wells in the
Ville Platte Field - A Case History
Steven K. Whiteside, Conoco Inc.
SUMMARY
Three waterflood projects are currently in operation in the Ville Platte
Field, Evangeline Parish, Louisiana. In mid-1986, casing pressure began to
develop on each of the three injection wells that serve these waterflood
projects. This paper outlines the history of these injection wells, the
troubleshooting techniques employed in an attempt to identify the source of
casing pressure, and the workover procedures which led to the successful
elimination of wellbore communication.
INTRODUCTION
The three waterflood projects that are presently in operation in the Ville
Platte Field include the Cook Mountain "B" Sand (VP CM B RA SU WF), the
Basal Cockfield Sand (VP BSL CF RD SU WF) and the Middle Cockfield Sand (VP
MDL CF RA SU WF) waterfloods.
The Cook Mountain "B" Sand is a channel sand of Middle Eocene age located
at a subsea depth of 8,050'. Initially, the Cook Mountain "B" Sand was a
normally pressured reservoir of approximately 3,770 psig (9.0 ppg pore
pressure). However, by the time the waterflood was initiated in March
1985, the reservoir pressure had declined to approximately 870 psig (2.1
ppg). Average reservoir porosity and sidewall core permeability are 30%
and 900 md., respectively. The approximate reservoir area is 180 acres
with an average net effective pay of 12 feet.
-342-
SKW1/011
-------
At the outset of the waterflood project, there were three producing wells
which had a combined production rate of 120 BOPD. Waterflood response has
been quite favorable in the Cook Mountain "B" Sand with the peak production
rate reaching 862 BOPD in September 1986.
The Basal Cockfield Sand is a Middle Eocene age sandstone located at a
subsea depth of 7,900'. The initial reservoir pressure in the Basal
Cockfield Sand was 3,720 psig (9.0 ppg). Upon initiation of the waterflood
in April 1986, the reservoir pressure had declined to approximately 2,900
psig (7.1 ppg). The average reservoir porosity of the Basal Cockfield Sand
is 25.5% with a sidewall core permeability of 66 md. The estimated
drainage area of the reservoir is 210 acres with an average net effective
pay of five feet. There are currently three producing wells in this
waterflood unit with a combined average production rate of 120 BOPD. To
date, waterflood response has not yet been detected in the Basal Cockfield
Sand.
The Middle Cockfield Sand is a Middle Eocene age sandstone located at a
subsea depth of 7,730'. The waterflood encompasses a drainage area of
approximately 260 acres with an average net effective pay of six feet. The
Middle Cockfield Sand has an average porosity and sidewall core
permeability of 25% and 33 md., respectively. The original reservoir
pressure was approximately 3,620 psig (9.0 ppg). At the beginning of the
waterflood project in April 1986, the reservoir pressure was approximately
2,520 psig (6.3 ppg), with a combined production rate from the two
producing wells of 72 BOPD. In recent weeks, the first indications of
waterflood response have been seen in the nearest offset producer, as
-343-
SKW1/011
-------
evidenced by a significant increase in the producing fluid level as well as
a 20 BOPD increase in production.
There are presently three injection wells that serve the aforementioned
waterfloods. The Ludeau-Haas No. 14 Well is completed as a single
injection well in the Cook Mountain "B" Sand waterflood from perforations
at 8,147'-82' (see Figure 1). This is the sole injection well for the Cook
Mountain "B" waterflood. Since initiation of the waterflood in March 1985,
cumulative injection into the Cook Mountain "B" Sand has been approximately
985,000 barrels of water.
The August Attales No. 3 and the Opelousas St. Landry Securities Co. No. 11
Wells are dually completed injection wells serving both the Basal and the
Middle Cockfield Sand Waterfloods (see Figures 2 and 3). Since initiating
these floods in April 1986, cumulative injection into the Basal and Middle
Cockfield Sands has been approximately 65,000 and 102,000 barrels of water,
respectively.
In mid-1986, casing pressure began to develop on each of the waterflood
injection wells. This paper outlines the history of the injection wells,
describes the troubleshooting techniques employed to identify the source of
the casing pressure, and details the quality control measures utilized
during workover operations which led to the successful elimination of
casing pressure.
-344-
SKW1/011
-------
LUDEAU-HAAS NO. 14 WELL
The Ludeau-Haas No. 14 Well was completed as a single injection well in the
Cook Mountain "B" Sand in March 1985 (see Figure 1). During the first 15
months of the waterflood, injection rates averaged approximately 1,100 BWPD
with a corresponding average surface injection pressure of less than 500
psig (see Figure 4). However, in June 1986, injection pressures began to
increase significantly as a result of the increases in injection rates.
During this time, injection reached a peak rate of 2050 BWPD with injection
pressures in excess of 2,100 psig. These injection increases were
necessary in order to match withdrawal rates from the reservoir.
Corresponding to the sudden increase in injection rates and pressures,
pressure began to develop on the casing. Initially, the casing pressure
built-up rather slowly, approximately 50-100 psig per day. However, within
a matter of weeks, the casing pressure began to build up by more than 1,000
psig per day. During this time, the casing pressure was being very closely
monitored and bled off daily. Conoco immediately informed the Underground
Injection Control Division of the State Office of Conservation of the
development of casing pressure and requested permission to continue
injection while attempting to identify the source of casing pressure.
After review by the State to ensure that there was no risk of contamination
to fresh water sands, permission was granted to continue operation. This
permission was granted provided that the subject well was closely monitored
and efforts were made to identify and repair the wellbore communication.
-345-
SKW1/011
-------
The initial investigation into a possible cause of the casing pressure
indicated that the bypass valve on the existing compression-set retrievable
packer may be opening, as a result of the increased injection rates and
pressures. The bypass valve is located on the top of the packer and is
designed to provide a means of circulating fluid in the wellbore without
having to release the packer. The packer is set by rotating and slacking
off weight on the tubing string. Set down weight of approximately 8,000 to
10,000 Ibs. is required to close the bypass valve and set the packer.
A tubing stress analysis indicated that under the existing injection
conditions the tensile forces acting on the tubing string were great enough
to cause the bypass valve on the retrievable packer to open, thus providing
communication between the tubing and the casing annulus. In an attempt to
offset the tensile forces acting on the tubing string, an additional 6,000
Ibs. of set down weight was applied to the packer. Also, in conjunction
with this work the tubing hanger was pressure tested to 2,250 psig to
ensure that no leaks were present in the hanger. Two days after restoring
injection into the Ludeau-Haas No. 14 Well, casing pressure had built back
up to 1,025 psig.
At this point, it was still suspected that casing pressure was associated
with the bypass valve on the retrievable packer. On September 24, 1986,
workover operations were begun to pull tubing and replace the existing
packer with a retrievable packer more suitably designed for the injection
conditions. Unlike its predecessor, the new packer did not have to be set
in compression and was not equipped with a bypass valve. Upon retrieving
the packer, a close inspection of the bypass valve did not show any signs
of wear or erosional effects to support our theory that the bypass valve
-346-
SKW1/011
-------
had been prematurely opening downhole. While rerunning the tubing and the
new packer, each joint of tubing was internally hydrotested to 3,000 psig
for approximately 10 seconds. No tubing was discarded during pressure
testing. After setting the packer, the casing annulus was pressure tested
to 1,000 psig for 10 minutes with no bleed off. Injection was then
restored at 1,700 BWPD and 1,600 psig injection pressure. Within two days,
1,140 psig of pressure had developed on the casing annulus.
During the first week of October 1986, the Ludeau-Haas No. 14 Well was
shut-in in order to conduct extensive pressure testing to establish the
source of the casing pressure. This work involved setting a plug in the
landing nipple located immediately above the retrievable packer and
pressuring up on the tubing to approximately 3,000 psig while monitoring
the casing pressure. Testing also included pressuring up on the casing
annulus while monitoring the tubing.
There were two primary reasons for conducting this testing. First, in
order to comply with applicable State regulations, every effort was being
made to identify the source of casing pressure so that the appropriate
steps could be taken to effectively eliminate communication. Secondly,
since the first workover proved unsuccessful in repairing the
communication, it was critical that every attempt be made to identify the
source of casing pressure before initiating further costly wellwork.
However, after four days of testing no conclusive evidence was obtained to
substantiate the source of the casing pressure. Thus, another workover was
necessary in order to identify and repair the wellbore communication.
-347-
SKW1/011
-------
Workover
On October 27, 1986 the second workover operation was begun on the
Ludeau-Haas No. 14 Well. During this workover, the following work was
performed:
1. The tubing and retrievable packer were pulled from the well.
Each joint of tubing was laid down on pipe racks.
2. While the tubing was on the pipe racks, the threads on the pin
and box ends were cleaned and visually inspected for signs of
wear or galling. Thread protectors were then placed on the pin
ends. Each joint of tubing was inspected for any obvious
corrosion or defects.
3. The retrievable packer was inspected and redressed before being
rerun in the well.
4. The tubing was rerun into well in the following manner:
a. Each joint of tubing was made up to the API recommended
optimum make-up torque for 2-3/8", 4.7#/ft., J-55, EUE 8rd
tubing of 1,290 ft.-lbs.
b. Each connection was internally pressure tested with helium
gas to 5,000 psig for 30 seconds.
-348-
SKW1/011
-------
c. Each joint of tubing was internally hydrotested to 5,000
psig for 15 seconds.
5. The wellhead and Christmas tree were gas tested with helium and
hydrotested to 5,000 psig.
In performing this work, every effort was made to implement effective
quality control measures in order to ensure that the wellbore communication
was identified and repaired. The following quality control measures were
taken while rerunning the tubing.
Tubing Make-Up
In making up each joint of tubing, a torque gauge was installed on the
hydraulic power tongs to ensure that the tubing was made-up to the API
recommended optimum make-up torque for 2-3/8", 4.7#/ft., J-55 tubing of
1,290 ft.-lbs. Special attention was given to orienting the load cell at
right angles to the lever arm on the power tongs and horizontal to the rig
floor. Otherwise, significant error in the torque gauge reading can result
from improper orientation of the load cell. Each connection was made-up
with the power tongs operated in low gear. Experience has indicated that
it is physically impossible to maintain control of the tongs and
achieve the optimum make-up torque of the tubing while operating in high
gear. While the tubing was being made-up, the pin and coupling were
visually inspected to make sure that the last round of threads on the pin
-349-
SKW1/011
-------
shouldered up to the coupling. If several rounds of threads remain exposed
after optimum make-up torque is achieved, this indicates the possibility of
crossthreading. Conversely, if the optimum make-up torque is not achieved
by the time that the threads are completely buried, this is an indication
of a pin or coupling that is out of tolerance or the threads are stripped.
Gas Testing
The fundamental operation of a helium gas test involves placing an internal
test tool across from the coupling area of the tubing connection. After
packing off above and below the coupling with the test tool, the coupling
area is filled with a helium test-gas mixture to a prescribed test
pressure. Once the test pressure has stabilized, a gas containment sleeve
is placed around the exterior of the coupling where helium will accumulate
if a leak exists in the coupling. After a designated accumulation time, a
probe is inserted into a sampling port on the containment sleeve to detect
the presence of helium. The helium concentration is determined by
utilizing a portable thermal conductivity meter, which compares the thermal
conductive properties of the atmosphere with those of the sleeve gas (see
Figure 5).
Several quality control measures were taken to ensure accurate gas test
results. Before testing began, the helium concentration of the test gas
was checked to confirm that sufficient levels existed that could be
detected by the thermal conductivity meter. During testing, the meter was
recalibrated regularly in order to make sure that the meter was reading
accurately. This calibration simply involved "sniffing" a helium test
bottle with the meter probe to check the response of the meter. After each
-350-
SKW1/011
-------
joint of tubing was made up, excess pipe dope would accumulate around the
coupling. Since this pipe dope could mask a small leak, each coupling was
wiped clean with a rag before placing the containment sleeve around the
coupling.
Hydrotesting
Internal hydrotesting involved simultaneously pressure testing the
connection, which had previously been tested with gas, and the tubing body.
This testing was performed above the rotary table so that a visual
inspection of the joint of tubing could be made during testing to check for
any signs of leakage.
In addition to the quality control measures described above, several other
steps were taken to ensure accurate test results. Company personnel were
located on the rig floor and the pump truck to make sure that strict
quality control measures were constantly adhered to during testing. All
pressure testing was conducted with the tubing string hung in tension
above the rotary table, in order to simulate the downhole tensile loading
conditions that would exist on the tubing. Finally, the gas coupling test
was conducted before the hydrotest. If the hydrotest is performed first,
water can enter into a potential helical leak path in the tubing threads,
resulting in masking a leak which would otherwise be detectable with gas.
-351-
SKW1/011
-------
Workover Results
When the tubing was pulled from the Ludeau-Haas No. 14 Well, it was obvious
that some of the tubing had not previously been made-up adequately, as
evidenced by the fact that several rounds of threads were exposed on the
pin ends. While cleaning and inspecting the tubing on the pipe racks, a
total of nine joints were discarded due to obviously galled or worn
threads. In two instances, a majority of the threads were worn completely
smooth, indicating that severe crossthreading had occurred during make-up.
While rerunning the tubing into the subject well, seven couplings and one
landing nipple failed either the gas test or the hydrotest and had to be
replaced. Also, one collar became distorted as a result of making up the
connection and consequently was discarded.
After running the tubing into the well and setting the retrievable packer,
the casing annulus was pressure tested to 1,000 psig for one hour with no
bleed-off occurring. On October 31, 1986, injection was restored to the
Ludeau-Haas No. 14 Well at a rate of approximately 2,000 BWPD with a
corresponding injection pressure of 2,100 psig. To date, no casing
pressure has developed on the subject well.
Overall, the workover operations went as planned and the wellbore
communication was successfully eliminated. One drawback to this testing is
the considerable running time associated with performing a combination gas
and water pressure test. The average running time was approximately 10 to
-352-
SKW1/011
-------
11 joints per hour; which includes the downtime associated with tool
failures and mechanical problems. Generally speaking, this author does not
advocate gas testing for this type of application, considering that the
subject well is a waterflood injection well operating at relatively low
pressures. However, due to the importance of minimizing the downtime of
this well and the previous workover attempt which proved unsuccessful in
identifying the wellbore communication, the additional precautions taken
were economically justifiable. Nonetheless, if more stringent quality
control measures had been implemented during the initial workover on the
Ludeau-Haas No 14 Well, a second costly workover would most likely not have
been required.
AUGUST ATTALES NO. 3 WELL
The August Attales No. 3 Well was completed as a dual waterflood injection
well in the Basal and Middle Cockfield Sands in April 1986 (see Figure 2).
Due to the low permeability and the high reservoir pressure of these sands,
the initial injection rates were relatively low (100 to 200 BWPD) with
correspondingly high injection pressures (in excess of 2,000 psig).
Immediately following the completion of the August Attales No. 3 Well,
pressure began to develop on the casing. Initially, the casing pressure
built up very gradually (approximately 100-300 psig per day) which
suggested that any leak(s) were small. However, within two months the
casing pressure was building up by as much as 1,200 psig in one day. In
an attempt to identify the source of casing pressure, several
trouble-shooting techniques were employed.
-353-
SKW1/011
-------
Troubleshooting
The first method utilized in an attempt to identify the cause of the casing
communication simply involved shutting in one side of the dual completion
while maintaining injection into the other. Prior to doing so, the casing
annulus pressure was bled off to 0 psig. During injection into only one
string of tubing, the casing pressure was closely monitored for any
build-up. This method proved unsuccessful in conclusively establishing the
source of casing pressure. Regardless of which string of tubing was being
injected into, casing pressure would develop.
There were two primary factors which contributed to the inconclusive
results. First, thermodynamic effects were present due to the injectivity
of a relatively cool fluid into the wellbore. Significant fluctuations in
the casing pressure were apparent when changes in the injection rates would
occur. By injecting the cool saltwater down the tubing, the temperature of
the wellbore would begin to decrease resulting in a corresponding decrease
in the casing pressure. Conversely, as the injection rates were reduced or
the well was shut-in, the wellbore would begin to warm up, causing pressure
to increase. Due to the thermodynamic effects, it was impossible to
establish how much of the casing pressure was actually attributable to
wellbore communication.
A second factor which contributed to the inconclusive test results was the
slow bleed-off of the tubing pressure after one of the completions was
shut-in. Due to the low reservoir permeability, it took a considerable
length of time for the pressure to dissipate in the reservoir and the
-354-
SKW1/011
-------
tubing pressure to bleed to zero. Therefore, until the tubing pressure had
bled off entirely, it was impossible to determine which completion
accounted for the increase in casing pressure. Since the duration of the
tests were relatively short, the wellbore never had time to stabilize in
order to establish the source of casing pressure.
The second method employed to identify the cause of the casing pressure
involved setting plugs in the landing nipples in the shortstring and the
longstring at approximately 7,712' and 7,944' respectively. Once the plugs
were set, the entire wellbore was bled off to zero. Then, each string of
tubing and the casing was individually pressure tested while the remainder
of the wellbore was closely monitored for any pressure build-up. The
primary advantage of this method over the previously described alternative
was the ability to isolate the wellbore from the reservoir pressure. This
provided more flexibility in testing in addition to having the ability to
quickly reach a stabilized wellbore condition. The results of this testing
clearly established that the longstring was the source of casing pressure
(see Figure 6).
After identifying the longstring as the source of casing pressure,
additional testing was conducted in order to determine if the tubing hanger
was leaking. If the hanger was the cause of the casing pressure, the
communication could be repaired without requiring a costly workover.
A tubing bridge plug was set on a collar stop one joint below the surface.
Once the plug was set, red dye was poured into the longstring before
pressure testing began. Only subtle changes in the casing pressure were
-355-
SKW1/011
-------
apparent during the previous testing, indicating that a relatively small
leak existed. Therefore, it was decided that the only way to conclusively
establish if the tubing hanger was leaking was to actually detect colored
dye returns on the casing during testing.
Initially, the longstring was pressured up to 3,000 psig. Within four
minutes, the longstring had bled down to approximately 1,000 psig while the
casing pressure remained unchanged. At this point, it was suspected that
the bridge plug was not holding. Testing then began on the shortstring.
After pressuring up to approximately 3,000 psig, the pressure on the
shortstring held steady while pressure on the casing increased only
slightly. The increase in casing pressure was caused by ballooning effects
of the tubing. After the wellbore remained stable for several minutes the
casing and tubing pressure was bled off. While bleeding off the casing,
red dye was recovered from the annulus, indicating that a leak in the
hanger did exist. Further testing confirmed that a leak was present in the
tubing hanger. The most likely explanation for not detecting the leak
initially with the pressure gauges is that the casing annulus was not
completely full of fluid when testing began.
Subsequently, a workover rig was brought in to repair the leak in the
tubing hanger. After nippling down the Christmas tree and picking up the
longstring tubing and the dual split hanger, the tubing threads that screw
into the bottom of the tubing hanger were hydrotested again in order to
visually witness the leakage (see Figure 7). It was obvious that the leak
in the hanger was caused by inadequate make-up of the tubing into the
bottom of the hanger, thus creating a helical leak path in the threads.
-356-
SKW1/011
-------
The leak was then repaired by taping the tubing threads with teflon and
completely making up the tubing into the hanger. Additional pressure
testing confirmed that the tubing hanger leak had been eliminated.
Three days after restoring injection into both completions, casing pressure
again developed. At this point, the longstring was shut-in and the casing
pressure ceased. For nearly three months, injection was maintained in the
shortstring, with no sign of casing pressure developing.
Working over the subject well had intentionally been delayed until wellbore
communication was eliminated on the Ludeau-Haas No. 14 single completion.
By delaying this work, valuable experience could be gained from
successfully repairing communication in the No. 14 Well before initiating
workover efforts on the dual completion.
Workover
In November 1986, workover operations began on the August Attales No. 3
Well in order to repair the longstring-to-casing communication. The
workover procedure that was carried out on the longstring was identical to
the procedure previously implemented on the Ludeau-Haas No. 14. However,
cleaning and visually inspecting the threads as well as gas testing was
excluded for the shortstring since it was not the cause of the casing
pressure.
-357-
SKW1/011
-------
After cleaning and visually inspecting the longstring on the pipe racks,
two joints of tubing were discarded; one due to a slight scar on the
threads and the other because the tubing body was partially crimped.
Although these joints were thrown out as a precautionary measure, they were
not suspected to be the cause of the communication.
It was obvious when the longstring was pulled out of the well, that the
entire tubing string had been inadequately made-up during the initial
completion. Three or four rounds of threads were exposed on each
connection. It was immediately suspected that a helical leak path existed
in the 8 round connections, thus causing communication between the
longstring and the casing.
While rerunning the longstring, effective quality control measures were
taken to ensure that the optimum make-up torque was applied to each
connection and that proper testing procedures were followed. During
pressure testing, not a single joint of tubing failed to test. Since
performing this workover, the August Attales No. 3 Well has been on
injection for nearly five months at injection pressures in excess of 2,000
psig, with no sign of casing pressure.
-358-
SKW1/011
-------
OPELOUSAS ST. LANDRY SECURITIES CO. NO. 11 WELL
The Opelousas St. Landry Securities Co. No. 11 Well, like the August
Attales No. 3, is completed as a dual waterflood injection well in the
Basal and Middle Cockfield Sands (see Figure 3). Shortly after completing
the subject well in April 1986, pressure began to develop on the casing.
Of the three injection wells discussed within this paper, the development
of casing pressure was least pronounced in the Opelousas St. Landry
Securities Co. No. 11. Despite realizing surface injection pressures in
excess of 2,000 psig into both completions, the casing pressure never built
up by more than 200 psig per day, indicating that only a very minute
leak(s) existed in this wellbore. Extensive pressure testing utilizing the
various troubleshooting techniques described previously, was conducted on
the No. 11 Well.
While preparing to pressure test the tubing head, water was discovered in
the test port indicating that a leak existed in the wellhead or the tubing
hanger (see Figure 7). To seal off the leak, plastic packing was injected
into the tubing head test port. Afterward, the tubing head was pressure
tested to 5,000 psig for one hour with no bleed off. Since performing this
work, casing communication has been eliminated. Presently, injection
pressures into both completions are greater than 2,200 psig, with the
casing pressure remaining constant at approximately 200 psig.
-359-
SKW1/011
-------
CONCLUSIONS
Several valuable lessons have been learned as a result of the work
performed on these injection wells. First, by closely monitoring and
troubleshooting a problem well, in many instances the source of
communication can be identified without requiring costly workover
operations. Secondly, implementing effective quality control measures
during remedial work can ensure that tubing leaks are successfully
identified and repaired. Thirdly, exercising rigorous completion practices
initially will help ensure the mechanical integrity of the completion
configuration and minimize costly workover operations.
Also, timely and thorough communication with the State Office of
Conservation provided them with an understanding of the testing objectives
and an assurance that fresh water sands were not being endangered during
this lengthy program.
ACKNOWLEDGEMENTS
I thank Conoco Inc. for permission to publish the material presented in
this paper. Thanks are also due Mark McClelland and Chuck Spisak for their
critical review of the manuscript and Ty Maxey for his field assistance in
making this work a success.
-360-
SKW1/011
-------
Figure 1: Ludeau Haas No. 14 Wellbore Schematic
LUDEAU HAAS #14 WELL
16* DRIVE PIPE SET__v-J
AT 93'
10 3/4* CASING SET ^
AT 1818'
8.7 PPG FSW IN ANNULUS
RETRIEVABLE PACKER /
AT 8,072'
5 1/2" CASING SET AT v-
//
X^
-— -J
X
X
$&£:0:£&i£CJj'ti$fi
II
T r» o
-------
Figure 2: August Attales No. 3 Wellbore Schematic
AUGUST ATTALES ^3 WELL
16" DRIVE PIPE AT
10-3/4" CASING SET AT
1822'
8.8 PPG IN ANNULUS
2-3/8", 4.7*/FT.,N-80,
EUE, 8RD. TUBING
W/BEVELED COLLARS.
7-5/8" DUAL PACKER SET
AT 7,744'
LANDING NIPPLE AT 7,944'
7-5/8" PERMANENT —
PACKER SET AT 7,977'
7-5"/8" CASING SET AT
8494'
X
I
~>
X
X
I
VILLE PLATTE FIELD
EVANGELINE PARISH, LA.
LANDING NIPPLE AT 7,712'
LANDING NIPPLE AT 7,759'
MIDDLE COCKFIELD SD.
: PERFS. AT 7,868'-950'
BASAL COCKFIELD SAND
PERFS. AT 8,036'-64'
FLOAT COLLAR AT 8,405'
T.D. = 8,500'
-362-
-------
Figure 3: Opelousas St. Landry Securities Co., No. 11 Wellbore Schematic
OPELOUSAS ST. LANDRY SEC. CO.
WELL
16' DRIVE PIPE AT
130'
10-3/4* CASING SET
AT 1804*
8.7 PPG FSW IN ANNULUS
7-5/8" DUAL PACKER AT
7708'
LANDING NIPPLE AT 7,9271
7-5/8" PERMANENT -
PACKER AT 7960'
FLOAT COLLAR AT 8403'
7-5/8' CASING SET AT
8497'
J
•>-*
s
»
)
!7'
\
/
i
\
/
\*+
,'
\ '
'//
///
X
^
>>
V
\
/
<
II
/
\
//
'/
''/
X
^
\
/
-J
>~
II
/
\
*s
'?'
''/'
",
S
\
X
s
\
<
LVILLE PLATTE FIELD
EVANGELINE PARISH, L
2-3/8", 4.7*/FT, N-80,
EUE, 8RD TUBING
W/BEVELED COLLARS.
LANDINft NIPDI P AT 7 R7*
LANDING NIPPLE AT 7 72%
x MIDDLE COCKFIELD SO.
== PERFS. AT 7,840'-928'
BASAL COCKFIELD SD.
'==^ PERFS AT 8 O18'-50*
T.D. = 8,500'
-363-
-------
i'lgure 4: Ludeau-Haas No. 14 Water Injection History
LUDERU HflflS « 14 NELL
INJECTION RflTE VS. TIME
10'
CD
d.
CO
10' l-f
WflTER INJ
L.
JFMflMJJflSONDJFMRMJJFlSONpJFMRMJ
1985 1986 1987
3500
3000
• 2500
CS
JC 2000
LJ
f, '50Q
LJ
Qi
Q_
; 1000
500
LUDERU HRRS
INJ. TUBING PRESSURE VS. TIME
WflTER INJ
n
r
J i
_J
JFnflMJJflSONDJFMnMJJnSONDJFnflMJ
1985 1986 1987
-364-
-------
Figure 5:
INTERNAL GAS TESTING CONFIGURATION
WIRELINE
PACKER
PACKER
PRESSURE SUPPLY LINE
ANNULAR
PRESSURE
N2/He SUPPLY
PRESSURE
GAUGE
TEST PROBE
COUPLING
TEST AREAS
GAS
CONTAINMENT
SLEEVE
INTERNAL
TEST TOOL
THERMAL CONDUCTIVITY
METER
-365-
-------
FIGURE 6:
AUGUST ATTALES NO. 3
WELLBORE PRESSURE TESTING RESULTS
L.S.
CSG.
20
40 60 80 100
TIME, MINUTES
a)
3000
2500
2000--
100--
0
S-S- 1000--
CSG.
L.S.
20 40 60 80 100
TIME, MINUTES
b)
950-
300-
200-
100
20 40 60 80 100 120
TIME, MINUTES
C)
a) PRESSURE TESTING THE LONGSTRING
b) PRESSURE TESTING THE SHORTSTRING
c) PRESSURE TESTING THE CASING
-------
Figure 7:
DUAL SPLIT TUBING HANGER
TUBING HEAD
TEST PORT
TUBING HANGER
THREADS
-367-
-------
SOME ASPECTS OF MONITORING A WATERFLOOD
VENTURA AVENUE FIELD WATERFLOODS
R. A. Deans and E. Jean Hill
Texaco USA, P.O. Box 811, Ventura, California 93002
ACKNOWLEDGMENTS
The authors wish to thank the management of Texaco USA for permission to publish this
paper. Special thanks are also extended to Ms. M. O. Sorensen and the local drafting personnel
for their assistance in the preparation of this manuscript. The authors are indebted to those
many engineers who have spent countless hours analyzing the information and formulating the
recommendation which have led to the current operating strategy for Texaco's Ventura Avenue
Field Waterfloods.
-368-
-------
TABLE OF CONTENTS
Page No.
List of Tables . iii
List of Figures iv
I. INTRODUCTION 1
II. WATERFLOOD REVIEW 2
A. Reservoir Description 2
B. Historical Background 3
III. INJECTION WATER QUALITY
A. Mechanical Treatment 5
B. Chemical Treatment 7
IV. MONITORING TECHNIQUES
A. Injection Well Surveys 9
B. Spinne r Surveys 9
C. Temperature Surveys 10
D. Waterflood Tracer Surveys 10
V. CONTROLLING INJECTION WATER 12
A. Mechanical Methods 12
B. Chemical Methods 14
VI. SUMMARY 16
VII. APPENDIX 18
VIII. REFERENCES 19
-369-
-------
LIST OF TABLES
Table No. Title
1 Fluid and Rock Properties (C-Block Average), Ventura Avenue Field
2 Optimum Requirements - Injection Water, C-Block Unit Waterflood
3 Staged Acid Program
<4 Gross Pore Volume Injection, Phases 1-10, C-Block Unit Waterflood
5 Water Entry Surveys, Producing Wells, C-Block Unit Waterflood
-370-
-------
LIST OF FIGURES
FiRure No. Title
1 Areal Conformance
2 Vertical Performance Saturation Fronts
3 Field Location Map
4 Ventura Avenue Field North-South Cross Section
5 Type Electric Log Ventura Avenue Field (Lloyd No. 244)
6 Ventura Avenue Field - C-Block and D-Block Unit Boundaries
7 C-Block Unit Waterflood
8 Project Production History; Phases 1-1 n, C-Block Unit Waterflood (1961-1984)
9 Ventura Avenue Field, VL&W East D-6, 7U Waterflood, D-Block Unit
10 Ventura Avenue Field; D-Block Unit Waterflood Status
11 Ventura Avenue Field; Performance of D-Block Waterfloods
12 Ventura Avenue Water Cleaning System
13 Typical Radioactive Tracer Detector Tool Configuration
14 Lloyd //234 Injection Profile Surveys
15 Spinner Survey Tool Configuration
16 Typical Temperature Survey Response, C-Block Unit Waterflood, Well Lloyd
17 External Casing Packer Schematic
18 Injection Well Flow Regulation Assembly
19 Conceptualized Polymer Treatment
20 C-Block Unit Polymer Treatment Results - Treated Sands
21 C-Block Unit Polymer Treatment Results - Untreated Sands
22 Gross Pore Volume Injection, C-Block Unit Waterflood, Phases 1-10
23 Water Entry Surveys, Producing Wells, C-Block Unit Waterflood
24 Injection Profile Status C-Block Unit Waterflood
25 C-Block Unit Oil Production
-371-
-------
SOME ASPECTS OF MONITORING A WATERFLOOD
VENTURA AVENUE FIELD WATERFLOODS
R. A. Deans and E. Jean Hill
Texaco USA, P.O. Box 811, Ventura, California 93002
ABSTRACT
Effective waterflooding relies, in part, on the efficient use of the injected water to
displace movable oil toward a producing well. Because of this requirement, steps must be
taken to direct water to the zones containing the oil reserves and data must be obtained to
reflect the true path of the water. Every barrel of water that does not go where it is intended
reduces the recovery and, consequently, negatively impacts the economics of the operation
Texaco operates two waterflood units in the Ventura Avenue Field. This field is massive
and, to waterflood it properly, careful attention to the quality and placement of the injected
water is required. Many techniques are used to help direct the water to the desired location.
They include mechanical means (external casing packers, cement, mechanical flow regulation,
selective perforations) and chemical means (acid treatments and crosslinked polymer). In
addition, downhole data are collected (from temperature, injection profile surveys and
chemical waterflood tracers) to identify the water path.
Through these efforts, the integrity of the waterflood is maintained, the condition of the
injection wells is understood and the injected water is used effectively.
-372-
-------
I. INTRODUCTION
As administered by the Division of Oil and Gas, the State of California laws governing the
conservation of oil and gas deposits include two fundamental goals:
(1) the protection of fresh water supplies, and
(2) the conservation of oil and natural gas resources.
Texaco USA and the Division of Oil and Gas have a mutual interest in accomplishing these
goals. Texaco (and its predecessor companies), by implementing specific engineering concepts,
has increased waterflood reserves in the Ventura Avenue Field and has simultaneously reduced
operating costs without impugning the integrity of the environment. This presentation will
discuss Texaco's experience in monitoring Ventura Avenue Field waterflood projects utilizing
operations designed specifically to not only comply with state regulations, but to do so
consistent with good oil field practice.
Waterflooding operations have been "without question responsible for the current high
level of producing rate and reserves with the U.S. (sic) and Canada." 1 Using the concepts
forwarded by individuals like F. F. Craig, Jr.l, proper engineering of any waterflood requires an
ever increasing understanding of the nature of the reservoir involved and how it reacts to the
injection of water. Improper placement of the water, known or unknown, results in, at best, an
inefficient flood and, at worst, unwanted damage.
Unbalanced areal coverage will not provide the displacement of available oil toward a
producing well. As depicted in Figure 1 the circles on the right and their size represent an
idealized, proportional volume of injected water. To the left, the oil between the ineffective
injection wells and the producers will eventually be moved away by the disproportionate
injection. The injection of uniform volumes in each well would prohibit such adverse
consequences. An equally important but less frequently considered view of inequitable water
injection is depicted in Figure 2. The oil that, in fact, should be produced can actually be
moved away resulting in a corresponding loss of reserves, if equal injection into each zone is not
-373-
-------
present.
Efforts in the C-Block and D-Block Unit Waterfloods of the Ventura Avenue Field have
been designed to address these important issues through thoughtful study and careful operation.
The success is verified by the results.
II. WATERFLOOD REVIEW
A. Reservoir Description
The Ventura Avenue Field is on the Ventura Anticline in the northwestern onshore area of
the Ventura Basin, about two miles north of the city of San Buenaventura ("Ventura")
California (Figure 3). In this field, the Pliocene Age Sands of the Pico Formation are composed
of the Upper Pico and Lower Repetto members and consist of a sequence of sands, silts and
shales more than 10,000 feet thick. The Ventura Anticline, a major structural feature of the
Ventura Basin, is tightly folded and oriented in a generally east-west direction. It is broken into
major areas by two large longitudinal thrust faults known as the "Taylor" and the "Barnard".
These faults divide the Ventura Avenue Field into two major oil producing sections, the "C-
Block" and the "D-Block". The C-Block section is that portion of the field which lies between
these two faults and the D-Block section is located below the Barnard Fault (Figure 4).
Typically, these major producing blocks have been divided into many sand sequences which
are interrupted by shale laminations. The C-Block, for example, consists of twelve major sand
bodies, some of which are two hundred feet thick (Figure 5). The producing Blocks, then, are a
very complex series of reservoirs to which general waterflood principles must be judiciously
applied. Monitoring procedures which are useful for less complex operations are not always
applicable for the Ventura Avenue Field because of its large gross sand interval, the large
number of sands and the permeability variations. Table 1 is compendium of fluid and rock
properties for the Ventura Avenue Field, representative of the C-Block. The general
orientation of the "C"- and "D-Block" waterflood units is shown on Figure 6.
-374-
-------
B. Historical Background
The first commercial oil production in the Ventura Avenue Field was established in 1916
by the State Consolidated Oil Company with the completion of their well, Lloyd #2, which
produced 100 barrels per day of 50° gravity oil from a light oil zone at a depth of approximately
2,500 feet. By 1921, Associated Oil Company had acquired State Consolidated and in January
(1921) completed the first C-Block producer (129 barrels of oil per day, 3,778 feet total depth).
The following year, Lloyd //5 was completed at 1900 barrels of oil per day from a depth of 4,051
feet. C-Block total production peaked in 1929 at 30,000 barrels of oil per day.
By 1956, the C-Block producing rate had declined to 4,900 barrels of oil per day and a
waterflood feasibility study was begun. Completed in 1958, this study defined a waterflood
development plan and recommended a waterflood pilot to test injectivity and waterflood
response of the C-Block sands. The first water injectivity test was initiated in August 1961
with the "Lloyd Lease (C-Block) Pilot Waterflood", consisting of one producer-conversion to salt
water injection, one producer-redrill completed as an injection well and five observation wells.
Two years later, neither injection well had attained planned injection rates (965 and 770 barrels
of water per day) nor had waterflood response been noted in any observation well. A second
pilot waterflood (East VL&W) was commenced in 1964 and the Lloyd Waterflood Pilot was
subsequently discontinued (September 1964).
Although performance of the second pilot was poor, the final waterflood development plan
was designed in 1967 and unitization agreements were signed in 1968 (final unitization July 1,
1970). The C-Block Unit Waterflood was divided from east to west into ten "Phases" denoting
individual sections of a staggered line-drive flood pattern (Figure 7) which included the C-2, 3
and 4 Sands from the "S" Sand marker to the "AT" Sand marker (shown earlier on Figure 5). By
January 1972, Phases 1 through 8 had been developed. [Development of Phases 9 and 10 was
hampered by the lack of an operator's agreement with Shell Oil Company, the offset operator
to the west of the C-Block Unit. Although negotiations were vigorously pursued, final
-375-
-------
agreements were not signed until August 1979 which finally enabled Phases 9 and 10 to be
expanded to full-scale injection and the C-Block Unit to become fully developed.
Figure 8 is a composite graph of waterflood production histories for all ten C-Block Unit
Waterflood Phases. To date, Phase 8 has provided the highest waterflood response while
reflecting a relatively shallow decline. Phases 9 and 10 have been slow to respond because of
injection delays in that particular area, but are currently responding well. Phases 1, 2 and 3
responded only slightly to the C-Block Unit Waterflood because of water influx and reservoir
heterogeneity prevalent in those areas. These phases also recorded particularly severe declines
following peak response. Phase 5 responded well initially, but production declined rapidly as a
result of early water breakthrough. Ultimate waterflood recoveries, when adjusted to an acre-
foot basis, also indicate superior Phase 8 performance. Phases 4, 5, 6, and 7 ultimate
waterflood recoveries are substantially less and recoveries from Phases 1, 2 and 3 are very
poor.
Tidewater Oil Company began development of the Ventura Avenue Field "D-Block" Zones
in April 1931 with the completion of Lloyd #57. At that time, this well's total depth of 8,823
feet made it the world's deepest producing oil well. There were many technological limitations
associated with drilling at these depths and development of the D-Block was, by necessity,
rather slow. However, by 1938, technology had advanced sufficiently to support additional
drilling and, while oil production peaked in 1949 at 23,600 barrels/day, active development of
the D-Block Sands actually continued through the early 1960's.
A D-Block waterflood plan was designed in 1970 and unitization was finalized in October
1978. Following a successful water injectivity test in 1979, the first D-Block Unit Waterflood
was initiated in January 1980 ("VL&W East D-6,7 Upper"). Figure 9 indicates the original
pattern and location of this waterflood relative to the D-Block Unit. Fourteen months later,
this waterflood was expanded. (Initially, an inferred fault was expected to form the west
boundary. By February 1981, it was apparent that the inferred fault either did not exhibit
-376-
-------
enough displacement to seal against injection or that the injection water traveled around the
fault.) Because of the numerous fault blocks and the massive zone thickness associated with the
D-Block, many floods will be required to properly develop this unit's waterflood potential. A
total of fourteen waterfloods have been d signed for initiation by the year 2008 (Figure 10).
Oil production in the waterflood area was averaging approximately 300 barrels/day when
full-scale water injection was initiated and, as shown on Figure 11, performance of this
waterflood has been excellent.
HI. INJECTION WATER QUALITY
Waterflooding porous media requires excellent water quality to aid in effective secondary
recovery. In the C-Block Unit Waterflood alone, poor water quality could account for
reductions in the proved reserves approaching 9.6 million barrels of oil. Because the sands in
the D-Block waterfloods generally have lower permeability, the impact on them could be
equally dramatic. Potential losses of reserves of this magnitude provide the basis for the
extensive water treating efforts in both waterfiooding projects and economically substantiate
the capital expenditures and assigned manpower required to maintain and improve the water
quality.
"Excellent" water quality is often a relative term which may constitute a wide range of
water standards. Frequently, a level of five to fifteen ppm total suspended solids is considered
as "excellent" quality. However, experts in oilfield water systems have established certain
criteria as listed in Table 2 which quantify more restrictive standards for high quality injection
water. Mechanical and chemical means, the effects of each on the other cannot be separated,
are used to achieve the desired quality.
A. Mechanical Treatment
Because the surface locations of the C-Block and the D-Block Unit Waterfloods physically
overlap, processing the water for injection is performed as one major "facility" and then
apportioned to each waterflood. A simplified schematic representation of the processing is
-377-
-------
shown in Figure 12. Approximately 95,000 barrels of water are processed in the facilities for
injection each day. Water produced with and then separated from the oil amounts to about
70,000 barrels, 10,000 barrels come from saltwater source wells, and the remaining 15,000
barrels come from a nearby freshwater lake.
The water produced along with the oil is first separated and then transferred to four
vessels ("WEMCOS") for removal of solids and oil remaining in the water after the initial
separation. Unfortunately, the efficiency of the "WEMCOS" does not meet our rigid standards
and further processing is necessary. Water from several source wells is added to the freshly
treated water stream and then it is all routed to two storage tanks. This water is then blended
with fresh water and transferred to tanks that supply water to five downflow multimedia "sand"
filters.
After being filtered, the water is held in more tankage that supplies both diatomaceous
earth (DE) filters and centrifugal pumps. This second filtration is to "polish" the water for
injection into the D-Block. About 30,000 barrels of water are injected into the D-Block daily.
The remaining 65,000 barrels processed on a daily basis are distributed to the C-Block injection
wells. Because of the remote location of many of the injection wells in the waterfloods, several
other "plant sites" are located throughout the field. Primarily, however, these are basically
stations containing limited tankage and pumps to boost the injection pressure.
Mechanically, with the help of some of the changes in the water imposed by chemical
treatment, the total suspended solids (TSS) levels of the water consistently and continuously are
reduced as a given body of water moves through the facilities. (Tankage is not sufficient to
hold the entire daily volume required by the flooding operations.) Figure 12 also indicates the
TSS levels at various points in the system. Overall, the facilities lower the TSS from an
average inlet level of 90 ppm to an average outlet level of less than 1 ppm.
The testing that provides this information uses filters which will collect particles larger
than 0.45 microns. Thus, not only is the TSS level down, the size of the material is quite small.
-378-
-------
These data verify that the technology being used in the water treatment facilities is capable of
approaching the desired levels for "excellent" water quality.
B. Chemical Treatment
Major problems affecting Ventura Avenue Field water quality attacked by chemicals are
primarily related to scale and bacteria. Although a corrosion inhibitor program is in place,
corrosion problems in the field are not considered significant; however, corrosion does affect
the scaling tendencies.
1. Scale; Scale deposition in the Ventura Avenue Field is a continual problem.
Calcium carbonate, barium sulfate, calcium sulfate and iron sulfide precipitation is frequently
observed. Because all of the water in the C-Block and D-Block have high bicarbonate and
sulfate levels, chemically treating the water is necessary. Although many of the injection wells
are treated with acid periodically, appropriate treatment for scale has decreased the frequency
of the remedial work.
Carbonate and sulfate scale is generally controlled by organic phosphonates and
polyelectrolytes. Both types of chemicals aid in the removal of the precipitate rather than
allow deposition or continued crystal growth. Because of the complex nature of oilfield brines,
the myriad of chemical equilibria and the chemical kinetics, the precipitation mechanisms are
not well understood; scale inhibition is still closer to an art than a science.
Control of iron sulfide precipitation and the subsequent fouling of equipment is attempted
by reducing or removing the reagents used to form iron sulfide. Restricting the corrosion rate
reduces the soluble iron. Eliminating the presence of hydrogen sulfide removes a significant
source of sulfide ion. Corrosion is controlled by a combination of the following:
...using alloy steels and plastics in construction,
...removing the dissolved gases,
...using plastic, epoxy or cement coatings on steels,
...organic film forming chemicals, and
-379-
-------
...cathodic protection.
2. Bacteria; Bacteria have associated with them the attendant problems affecting
corrosion and scale. The three strains commonly attacked in oil field operations are sulfate
reducing bacteria, iron bacteria and slime forming bacteria. In each case, detrimental effects
on the faciliites and, in turn, the solids content of the water, can be very deleterious.
The first rule for the successful application of any bactericide is to generate and then
maintain a "sterile" system. This means that all surface facilities should be purged of
biomasses in the tanks, along the walls of the pipes and in the filters. Because bacteria usually
prefer to grow in the nonturbulent zones of water systems and even under scale or debris, the
effectiveness of a biological control scheme will depend on the manner in which the scheme
overcomes the obstacles. Bacteria will remain very difficult to kill when they are shielded by
scale, debris or even their own secretions (biopolymers and iron hydroxide, for example).
In the Ventura Avenue Field, these obstacles are tackled by "pigging" the injection lines.
This means physically removing scale, debris and even biomass from the tubular goods by
forcing a scraping device, a "pig", through the lines. Additionally, the filters and the tanks are
emptied and cleaned when evidence suggests that a problem exists.
A new biocide program using chlorine dioxide has been implemented in the Ventura
Avenue Field. Many months of optimization (which included the monitoring of biocide levels,
introduction of hydrogen sulfide scavenging and ferrous iron oxidation chemicals, changes in the
frequency, duration and location of the chemical addition and cleaning of the surface lines,
tanks and filter vessels) have yielded the values indicated on Figure 12.
When compared to the requirements for excellent water quality listed in Table 2, the
water treatment efforts for the C-Block and D-Block injection fluid approach ideal. In general,
suspended solids are "om or less, oxygen levels are less than 10 ppb, hydrogen sulfide levels
are less than 0.1 ppm, corrosion rates exceed the guidelines and the soluble iron is essentially
zero.
-380-
-------
IV. MONITORING TECHNIQUES
A. Injection Well Surveys
Injection profile surveys are the primary means by which the entry of the injected water
is monitored. The method commonly used is fluid velocity profiling which involves timing the
movement of an injected slug of radioactive material in the flowing stream. This procedure is
effective for determining the zones receiving fluid injection. In addition, it is diagnostic for
fluid movement behind casing, leaking packers and evaluation of well stimulation procedures
such as acidization.
Shown in Figure 13 is the typical configuration of a radioactive tracer tool. The casing
collar locator (CCL) is a sensor that responds to the increased metal density at the casing
coupling. This information is used to correlate the tracer survey data with the correct depth.
The ejector port is the point from which the radioactive chemical (usually either Iodine-131 or
lridium-192) is injected. Finally, two gamma ray detectors with known spacing lie below the
ejector.
An example of injection profile survey results can be seen in Figure 14. These data
reflect the sensitivity of profile on the rate within the C-Block but also show the resolution of
the information available. To explain further, at an injection pressure of 1100 psi, five
identifiable zones are receiving water injection and, with the exception of the uppermost
interval, the distribution is more or less equitable for each interval. With an injection pressure
of 1800 psi, additional zones are taking water, although the water is more confined. Finally, at
higher pressures, still a different distribution is observed. These types of data provide the
reservoir engineer and operations personnel with the basis upon which to make informed
decisions on the proper operations for the waterflood.
B. Spinner Surveys
Another type of device used to follow the water as it leaves the wellbore is the "spinner"
tool. This tool, conceptually shown in Figure 15 is nothing more than a flow meter located on
-381-
-------
the bottom of a cable. The "spinner" is simply a propeller that spins in response to the water
flowing past. The rate of spinning is detected by a receiving coil which surrounds a magnet
attached to the propeller.
The simplicity of the tool makes its operation easy to understand, but, it is also limited by
flow rate and orientation in the hole. However, these limitations are well known and can be
overcome for many situations. The spinner surveys are often conducted along with the
radioactive injection profile survey to verify results.
C. Temperature Surveys
Reservoir temperatures in the C-Block and D-Block are higher than the temperature of
the injection water and this difference in temperature allows identification of the zones that
have received significant amounts of injection water. As the cooler water is injected, the rock
and fluid temperatures are lowered. By recording the downhole temperature with respect to
depth, any cooling observed can be distinctly attributed to the action of the injection water.
Figure 16 shows the results of a temperature survey that was conducted in a C-Block well
(Lloyd //246). Several areas of cooling can be observed. Substantial cooling has occurred in the
"AC", the "AE" and the "AK" Sands with some cooling also seen in the "AA" Sand. Had no
influence by the injected water occurred, the temperature survey would have shown the normal,
gradual, steady increase in temperature as the depth increased.
D. Water flood Tracer Surveys
Various chemicals have been used to follow the movement of the injected water through
the reservoir. Although the information from injection profile surveys and spinner surveys is
extremely useful, these data only indicate the depths at which the water is exiting the wellbore.
They do not provide insight on the movement beyond the wellbore. By adding a "tag" to the
water that can be analyzed in the subsequently produced fluid, communication between
injection wells and producing wells can be defined. To accomplish this task, however, the
chemical "tags" must (1) not react in the formation chemically, (2) must move with the injected
-382-
-------
fluids and not suffer adsorption, and (3) must be detectable in the produced fluids. To date,
fluorescent dyes, ions and radioisotopes have been used in the C-Block Unit Waterflood with
good success.
For tests in which very rapid communication between injection wells and producing wells
is suspected, flourescent dyes are recommended. Their large losses through adsorption and
reactivity with reservoir constituents prevent a prolonged life. Analytically, the presence of
the dyes can be detected colorimetrically and even visually.
Specific ions (especially thiocyanate) are recommended for situations anticipated to be of
longer duration. Although nitrate and sulfate may be used as tracers, thiocyanate has been
proved more successful. Thiocyanate can be detected colorimetrically after complexation with
ferric iron; however, ion chromatographic techniques provide more reliable information with
much less sample preparation.
Other ionic chemical tracers should not be used in the C-Block Reservoir. Chloride,
bromide, iodide and phosphate are all present with inconsistent levels in the produced water.
Lithium is expected to exhibit a significant exchange problem with the cations weakly bonded in
the reservoir clay material and has a relatively high cost. Finally, fluoride has a very limited
solubility in the C-Block water (30 ppm maximum).
Radioisotopes are the best tracers for the long term, complex, flow-studies. Although the
major disadvantages are the costs and inconvenience of the analytical services, they have
better characteristics than either of the other types of tracers for the following reasons:
...They have few compatibility problems.
...Naturally occurring background levels of the radioisotopes are usually zero.
...Introduction of radioisotopes into the injection stream is very fast.
Recent chemical tracer work has produced very good results. In an extensive test to
define the nature of the interaction of a C-Block injection well, Hartman #58, and a producing
well, Hartman //9, a combination of fluorescent dye, tritiated water and thiocyanate was used.
-383-
-------
Because either the "AK" Sand or the "AH" Sand was suspected of contributing excessive water
to Hartman #9, several tracers were needed for primary and confirming information. The final
analysis identified the "AK" Sand as the offending zone when the injection pressure was above
1000 psi. Subsequent reduction in the injection pressure at the "AK" Sand has resulted in lower
water production with no loss of oil production; a more effective use of the water.
Sulfur-35, introduced as a sulfate, was tested with inconclusive results. Although it
should have been detected, it was not. This observation has fostered speculation that the
sulfate moiety may have been consumed by the activity of the resident bacteria.
V. CONTROLLING INJECTION WATER
Because the Ventura Avenue Field waterfloods are in structurally thick reservoirs with
multiple layers of sandstone and shale, the equitable vertical distribution of the injected water
across the waterflood interval has been a severe, vexing, continual problem for our engineers.
Since waterflood inception, many procedures have been tried, and, unfortunately, many have
failed to materially improve injectivity profiles. The C-Block injection interval includes as
much as 1,000 feet of net oil sand and several hundred noncommunicating individual sand layers.
The designed injection rate is 5 BW/day/foot of sand; however, some sands will take no water
while some "thief" zones take over 100 BW/day/foot of sand. The majority of the injection
wells have 7-inch cemented casing, perforated at intervals with two 1/2-inch holes per foot.
Many older producing wells have been converted to injection wells and their slotted-liner
completions make injection profile improvement very difficult. Left uncontrolled, water will
preferentially enter the zone which yields the least amount of resistance. For these reasons,
care is exercised to specifically direct the injected water, again, by both mechanical and
chemical means.
A. Mechanical Methods
1. External Casing Packers; The primary cementing of the casing within the wellbore
is very important to the integrity of the waterflood. External casing packers have been added
-384-
-------
to the methods of segregating vertical sections of a given wellbore in an effort to improve the
primary cement effectiveness and to provide a reasonable guarantee for success
In brief, the external casing packer is an elastomeric sleeve surrounding a mandrel on
which an inflation valve system is mounted. This design, illustrated in Figure 17, allows the
sleeve to expand upward from the bottom as it is being filled with cement only after the
primary cementing operation has been completed. Placement of this type of packer in the new
wells being drilled (or redrilled because of failure) eliminates fluid migration behind the casing
between the sands.
2. Mechanical Flow Regulation; Generally, a single wellbore is used to inject water
into more than one of the sands in the C-Block and, as has been mentioned, the variety of rock
characteristics will not allow the desired distribution of injection water over a large interval.
Therefore, regulation of the flow between sand bodies having diverse qualities is required to
improve the effectiveness of water.
Figure 18 illustrates the type of flow regulation currently being used in the C-Block and,
to a much more limited extent, in the D-Block. The assemblies consist of packers to isolate
zones intended for injection and a side-pocket mandrel containing a flow regulating orifice to
limit the flow rate by generating a backpressure. Usually, no more than five packers with four
mandrels have been successful because of the difficulties with their operation. Once installed,
the flow rates can be adjusted by changing the size of the orifice within each mandrel. This
operation can be performed remotely (by "wireline" recovery) so that the entire assembly does
not require removal.
This method of controlling the injected water has proved to be the most effective
technique when several zones of significantly different permeabilities are open to injection in a
given well at the same time.
3. Cemented Isolations; As a general practice, isolations between one or more zones
may be needed in existing wells. External casing packers have not been used on many wells and
-385-
-------
were not used on any wells prior to 1985. Routine procedures for these situations require the
operator to establish a segregation, most often in an identifiable shale zone.
Usually, the casing is perforated with six holes, sixty degrees apart along a 1.5 foot casing
section. This orientation and low-angle phasing of the perforating holes provide a very good
chance of intersecting existing channels in the primary cement sheath. Following a "spearhead"
of hydroflouric acid to dissolve the drilling mud filter cake, large volumes of cement slurries
are generally beneficial to permit casing wellbore annulus fill up with cement.
After an attempt is believed successful, the casing is reperforated above and below the
segregation to pressure test the zones, ensuring the integrity of the procedure.
*. Selective Perforations: Because a great wealth of information is available to the
Reservoir and Operations Engineers and Development Geologists, the maintanence and plans for
the waterflood are carefully thought through. Each new project (involving new wellbores and
the maintenance to replace failed wells) is designed to selectively perforate specific major
sands within the C-Block.
For example, flooding selective sands is a viable option (assuming the recoverable
reserves assigned to the project can make the project economic). This technique certainly will
mechanically restrict the water only to those zones thought to contain moveable oil and
improve the effectiveness of the injected water. However, not many areas are available in the
C-Block in which to define this type of project.
The nature of the deep waterflooding operations in the D-Block are essentially selective
sand floods. The difference in reservoir pressures requires special measures and the difficulty
of achieving good mechanical isolation at depths greater than 10,000 feet make the flooding of
multiple sands less attractive. The history of the C-Block waterflood suggests that initial
selective flooding of the C-Block may have resulted in a much more manageable project today.
B. Chemical Methods
1. Acidization; Regardless of attempts to maintain exceptionally high quality injected
fluids, injectivity unfortunately decreases in the water injection wells. This loss of injectivity
-386-
-------
is usually caused by damage from scale deposition and particulate matter. The loss is
monitored by a gradual but continual increase in the injection pressure and, eventually, when
the maximum available pressure is reached, a decrease in the injection rate occurs. To rectify
this damage, wells are usually treated with acid in an attempt to restore the well closer to its
former condition. Unfortunately, the original conditions can never be attained again.
Although the details of the entire procedure are quite involved, the usual procedure for an
i
acid treatment consists of three "stages" of injection using 20 gallons of an acid mixture per
foot of open interval. The entire perforated interval is "washed" in two-foot increments using
one stage at a time. The composition of typical acid stages in a program is given in Table 3.
The acid provides significant improvements to the injectivity. As the quality of the
injected water continues to improve, less frequent acidization programs will be required.
Because of the reduced frequency of acidization, water will be entering the zones preferred by
the engineering staff for a much longer time and the adverse consequences of reduced
injectivity will not be as severe.
2. Crosslinked Polymer; A program using crosslinked water-soluble polymers to aid the
mechanical methods in the redistribution of injected water was implemented in the C-Block
Unit Waterflood. This project was certified as a qualified tertiary recovery method under the
Windfall Profit Tax Act of 1980 and consisted of a series of injection well treatments designed
to curtail the ability of certain zones to accept the injected water while not impairing the
injectivity of other sands. The goal of each treatment was to provide a means to more evenly
distribute the injection water when no other method would be available. Again, the result of
the treatments would be a more effective placement of the injected water and would yield a
more efficient recovery of incremental oil. The results of a well treatment are shown in the
conceptualized drawing of Figure 19. As shown, the polymer enters a zone previously open to
water injection and effectively restricts further flow. The water is then forced into areas less
affected, or even previously unaffected, by the waterflood.
-387-
-------
Each well selected as a treatment candidate was extensively reviewed to determine how
better to redistribute the injected water. Usually, a sequential injection of a polymer and
crosslinking agent followed by injection of polymer fluid containing a crosslinking agent formed
the basis for the treatment. Each treatment was specifically tailored to the conditions of the
individual well. As the complex nature of the process became better understood by the project
engineers, the ability of the process to intentionally reduce the water injection in selected
zones improved and, as a consequence, the redistribution of the injected water occurred as
planned.
Data from the injection profile surveys taken following the treatments conducted in 1985
indicate consistent improvements. As shown diagramatically in Figure 20, the average
reduction in daily water injection into the sands receiving the polymer treatment was about 36
percent (from 876 to 562 barrels of water per day). For the individual treatments, the best
reduction of injectivity was 76 percent while the lowest was actually an increase of about 12
percent.
However, each well treatment resulted in an increase in injection into the sands targeted
for an increase, as depicted in Figure 21. On the average, the eight treatments resulted in a 38
percent increase of injection. As a result of these treatments, in total, about 2,371 barrels of
injected water have been redirected to zones having more potential.
Although the project was planned to continue for several years, the dramatic drop in the
price of oil made these expensive, individual well treatments uneconomic. Consequently, the
program has been discontinued; the manpower has been reassigned.
VI. SUMMARY
Historically, water entry surveys performed in the C-Block Unit wells indicated that,
generally, the highest water production occurred in those zones of greatest water injection,
confirming zone isolation with no apparent crossflow (Figures 22 and 23, Tables 4 and 5). Static
temperature surveys, conducted at very regular intervals, substantiate that none of the injected
-388-
-------
water is moving out of the targeted waterflood zones. Figure 24 clearly reflects that the "AH"
and "AK" Sands have been successfully flooded while the "AM" through the "AQ" intervals
remain basically unflooded. Utilizing procedures discussed in this paper, the art of "selective
waterflooding" is now being applied to these less permeable sands with promising results.
Much has been said here about "technology" and the "proper application" of that
technology. However, we must also address the assimilation and the interpretation of these
data. As with any new project, hindsight has fostered the creation of invaluable information
systems for the C-Block Waterflood which can (if properly used) immediately denote specific
problem areas. Conversely, these same data banks can often affirm the successful pursuit of
increased reserves recovery by waterflooding. (These monitoring systems are already in place
as a monitoring device of the newer D-Block Waterfloods).
Records must be accurately and diligently maintained. Engineers must periodically
review fluid levels, production and injection rates, as well as injectivity profiles. Complex
waterflood projects are best reviewed on an individual well basis; composite data per flood have
a tendency to hide or distort problematic areas resulting in counterproductive effects.
Having arrested the waterflood base decline rate (as shown on Figure 25) the C-Block Unit
Waterflood is considered extremely successful. However, this pinnacle will not be sustained
without future extensive waterflood maintenance. Well replacements (new or redrilled
wellbores), improved injection well profiles and the continuing stimulation of producing wells
will be favorably reflected in reserves recovery. Our waterflood achievement in the Ventura
Avenue Field is also a direct and positive manifestation of the skillful application of proper
oilfield technology blended with the vital synergism of the reservoir engineer and operating or
field personnel. Through this fundamental collaboration, we have learned to emphasize correct
operating strategy, rather than merely relying on proper procedures. As a result, Texaco's
success is twofold: ...increased reserves recovery with its associated economic rewards and
...the secure knowledge that the integrity of the surface and subsurface ecological systems has
been preserved.
-389-
-------
English Unit
Acre
°AP1
bbl
cp
ft
oF
gal
mil
psi
scf/STB
in
VIL APPENDIX
English to Metric Conversion Factors
x 4.04687 x 10+3
x 1.58987 x 10"1
x 1 x 10-3*
x 3.048x10-1*
(OF - 32)/1.8
x 3.78411
x 2.54 x ID'3*
x 6.89476
x 1.80118 X 10-1
x 2.54*
Metric Unit
g/cm3
m3
P2 • s
m
°C
cm
kPa
std m3/stock-tank
cm
*Conversion factor is exact.
-390-
-------
Vffl. REFERENCES
1. "The Reservoir Engineering Aspects of Waterflooding", Craig, F. F.: Monograph Volume
3, Society of Petroleum Engineers of American Institute of Mechanical Engineers, Ne-
"ork, 1971, p. 9.
Much of the information in this paper has been summarized from the following internal
reservoir engineering reports and papers:
2. "Reservoir Engineering Analysis. C-Block Unit Waterflood (1970-198ft)". Goble, P.G.:
Texaco USA, October 1984.
3. "Ventura District 1987 Capital Budget", Texaco USA, June 24, 1986.
4. "Water Injection Well Monitoring Ventura Avenue Field", Goble, P.O., Reis, T.A., and
Davis, A.K.: Getty Oil Co., September 1983.
5. "Here's How Getty Controls Inactivity Profiles in Ventura", Froning, S. P., Birdwell, R.F.:
Oil and Gas Journal, February 1975.
6. Hartman 58 Tracer Tests Results and Recommendations; C-Block Unit Waterflood",
McHenry, J.: Texaco, USA, July 9, 1986.
-391-
-------
TABLE 1
FLUID AND ROCK PROPERTIES (C-BLOCK AVERAGE)
VENTURA AVENUE FIELD
PRESSURE (datum 5500 feet subsea)
Initial psia
Bubble point psia
Reservoir pressure psia (1956)
TEMPERATURE (datum 5500 feet subsea)
Reservoir °F
ROCK
Anticline plunge
Formation dip degrees (South Flank)
(North Flank)
Average formaton depth (subsea) ft.
Porosity (range %)
(with overburden)
Permeability
(air-absolute) md (range %)
(effective brine) md
Permeability variation
Authigenic clay content, thin section analysis, Lloyd #235
(total)
(expandables, fraction of total)
Authigenic clay content, pipette analysis, Lloyd
(total)
(expandables, fraction of total)
SATURATION
Soi (fraction)
Swi (fraction)
Sgi (fraction)
OIL
Gravity (range o API)
Viscosity (initial, cp)
Bo (initial)
(1956)
2800
2800 (Assumed)
550
160
3°-8° Eastward
up to 60°
5500
22.0 (18.8-24.8)
20.0
160 (70-350)
70
0.7
1-17% avg. 5%
6-79% avg.
avg. 3%
avg. 50%
.70
.30
0
30.50 (29°-32°)
1.3
1.315
1.141
-392-
-------
TABLE 1 (concluded)
PRODUCED WATER (C-3 SANDS)*
Na (ppm)
Ca (ppm)
Mg (ppm)
Cl (ppm)
HCO3 (ppm)
SO^ (ppm)
B (ppm)
ph
* Average of eight analyses - primarily Phases 1 and 2
GAS
Original gas formation volume factor = Bgi (RB/scf)
(1956) = Bg
Original gas compressibility factor = Zi
(1956) = Z
Initial solution GOR scf/STB
(1956)
Original gas viscosity (Cp)
Gas specific gravity 7g
= Rsi
= Rs
10,467
205
17,097
750
7.5
56
7.7
0.00517
0.02941
0.835
0.911
570
135
0.0186
0.8
-392-
-------
TABLE 2
OPTIMUM REQUIREMENTS - INJECTION WATER
C-BLOCK UNIT WATERFLOOD
Our experience indicates the following general requirements are necessary for excellent water
quality:
Item
Suspended solids
Scale
Iron (total)
Corrosion rate
Pit depth (30 days)
Pit frequency (30 days)
Oxygen
Bacteria count
H2S
Level at Inj. Wellhead
Less than 0.5 ppm
No decrease in calcium, sulfate or bicarbonate
levels through the system
Less than 1 ppm
Less than 0.1 mils/year
1 mil
1 pit/sq.in.
Less than 10 ppb
Less than 100 colonies/ml
Zero
-394-
-------
TABLE 3
STAGED ACID PROGRAM
ACID COMPOSITION
STAGE DESCRIPTION
I A solution of 15 percent hydrochloric acid containing 0.5 percent
corrosion inhibitor, 0.3 percent water wetting surfactant, 15 pounds per
1000 gallons of iron chelating agent, 1.0 percent acetic acid for buffer,
20 pounds per 1000 gallons of a reducing agent and 3 percent of a mutual
solvent.
II A solution of 13.5 percent hydrochloric acid and 1.5 percent hydrofluoric
acid containing 0.5 percent corrosion inhibitor, 0.3 percent water
wetting surfactant, 15 pounds per 1000 gallons of iron chelating agent,
1.0 percent acetic acid for buffer, 5 pounds per 1000 gallons of a
reducing agent.
Ill A solution of 7.5 percent hydrochloric acid containing 0.5 percent
corrosion inhibitor, 0.3 percent water wetting surfactant, 15 pounds per
1000 gallons of iron chelating agent, 1.0 percent acetic acid for buffer, 5
pounds per 1000 gallons of a reducing agent and 3 percent of a mutual
solvent.
-395-
-------
TABLE 4
GROSS PORE VOLUME INJECTION
PHASES 1-10. C-BLOCK UNIT WATERFLOOD
Za
AA
AC
AE
AGa
AH
AK
AM
AO
AP
AQi
AS
Floodable
Sand Vol.
(AC.FT.)
24,940
19,361
39,295
64,474
57,569
42,075
90,508
46,083
16,675
16,760
21,787
22,932
0
.20
.20
.20
.20
.18
.20
.20
.20
.18
.18
.18
.16
Gross PV
(MBbls.)
38,697
30,041
60,970
100,038
80,392
65,284
140,432
71,502
23,286
23,404
30,424
28,465
Inj. To 3/1/8*
(MBbls.)
26,505
16,457
27,462
30,057
36,179
58,634
84,397
17,237
4,305
987
841
7,393
%
PV
68
55
45
30
45
90
60
24
18
4
3
26
-396-
-------
TABLE 5
WATER ENTRY SURVEYS
PRODUCING WELLS. C-BLOCK UNIT WATERFLOOD
BWPD
Lloyd #56 Lloyd #99 Lloyd #10* Lloyd #1*2
Markers 8/7/78 10/21/7* 12/30/82 */28/78
Za MINOR 280
AA
AC
AE 200
AGa
AH 175
AK MAJOR 515 60
g AM 120
^ AN
AO 25 50
AP
AQi
AR
AS 20
Lloyd #226 VL&W #2 VL&W #160 McGonigle #*3 Marker
12/16/77 9/5/78 12/26/79 5/20/76 Totals
300 580
220 220
100 170 270
200
0
130 290 595
80 210 865
120
320 320
75
0
0
0
20
-------
Figure 1
AREAL CONFORMANCE
POOR OIL RESPONSE
VD
00
POOR INJECTION BALANCE AND COVERAGE
GOOD OIL RESPONSE
GOOD INJECTION BALANCE AND COVERAGE
-------
Figure 2
VERTICAL PERFORMANCE
SATURATION FRONTS
POOR OIL RESPONSE
POOR VERTICAL CONFORMANT
GOOD OIL RESPONSE
BALANCED INJECTION
-399-
-------
Figure 3
FIELD LOCATION MAP
R 24 W
R 23W
R 22 W
R 21 W
R 20 W
R 19 W
R 18 W
R 17 W
HOPPER .
CANYON^
MARILLO
CONEJO (Abd)
PIRU CREEK
AMONA
T
5
N
T
4
N
EUREKA CANYON
CANYON GAS * _ _—NORTH TAPO
(Abd)-~^5~v»^*. ^-TAPO RIDGE
, TORREY ^f VBANTA SUSANA
SOUTH |H;N^CTKOR%GE^ %™™> ]
MOUNTAIN^0AANKYOpNARKN I -BIG MTN J CANYOXN
\
1
1
1
1
1
°SIMI
SANTA SUSANA
THOUSAND OAKS \
T
2
N
T
I
N
I
S
-------
Figure 4
VENTURA AVENUE FIELD
NORTH-SOUTH CROSS SECTION
C-BLOCK
UNIT
BARNARD FAULT
TAYLOR FAULT
D-BLOCKUNir
-401-
-------
Figure 5
TYPE ELECTRIC LOG - VENTURA AVENUE FIELD
(LLOYD NO. 244)
C-5
-402-
-------
Figure 6
VENTURA AVENUE FIELD
C-BLOCK AND D-BLOCK UNIT BOUNDARIES
o
CO
I
LEGEND
— — — — C-BLOCK UNIT
—^— D-BLOCK UNIT
-------
Figure 7
C-BLOCK UNIT WATERFLOOD
i
-O
o
UNIT BOUNDARY ^
-------
Figure 8
PROJECT PRODUCTION HISTORY
1-10, C-BLOCK UNIT WATERFLOOD
(1961 - 1984)
i
-p-
o
Ijl
I
q
£
1961
1966
1971
1976
1981
-------
Figure 9
VENTURA AVENUE FIELD
V.L.&W. EAST D-6, 7U WATERFLOOD
D-BLOCK UNIT
D-Block
Unit Boundary
Waterftood
Area
N
LEGEND
WATER NJECTOR
PRODUCER
WATERFLOOD AREA
-406-
-------
I
->
o
~J
I
Figure 10
VENTURA AVENUE FIELD
D-BLOCK UNIT WATERFLOOD STATUS
D-5 ZONE
WATERFLOOD
AREA
Lloyd CD
Hartman
V.L.4W.
Weat
V.L.4W.
Central
V.L.4W.
East
-------
Figure 11
VENTURA AVENUE FIELD
PERFORMANCE OF D-BLOCK WATERFLOODS
I- 100M
WATER INJECTION
- 10M
- 1M
PRIMARY OIL
I 1078 I 1979 I 1980 | 1961 I 1982 I 1983 I 1984 I 1985 I 1986
100
-408-
-------
Figure 12
VENTURA AVENUE WATER CLEANING SYSTEM
o
VO
I
BIOCIDE
INJECTION TO -
D-BLOCK 0.4 ppm TSS
(30 MBBLS/DAY) 100 COLONIES/ML SRB
OXIDIZER/BIOCIDE
^ 1
PRODUCED WATER
(70 MBBLS/DAY)
SOURCE WELLS
(10 MBBLS/DAY)'
SOURCE TANKS
WEMCOS
1.0 ppm TSS
100 COLONIES/ML SRB
-^ 1—&—oooo
D.E. FILTERS
MULTIMEDIA
SAND FILTERS
>. INJECTION TO
C-BLOCK
(65 MBBLS/DAY)
POST-FILTER TANKS
10 ppm TSS
106 COLONIES/ML SRB
PRE-FILTER TANKS
r^ ^-r^ » l
/ FRESH
^ 1 \A/ATi-r>
\ TANK
LAKE CASITAS
WATER
(15 MBBLS/DAY)
-------
Figure 13
TYPICAL RADIOACTIVE TRACER DETECTOR
TOOL CONFIGURATION
CCL
EJECTOR PORT
TOP OF
GAMMA DECTOR
TOP OF
GAMMA DECTOR
-410-
-------
Figure 14
7100
8EQ.i .!
7160
7200
7250
7300
17
42
73
1$
37
68
as
21
7360
7400
•29
7460
7600
LLOYD 234
INJECTION PROFILE SURVEYS
6/29/83
460 BWPD
1100 P8i
lilJ
B
6/30/83
760 BWPD
1800 psi
2222223
7/1/83
860 BWPD
2200 psi
V.VWiYMWYM
0 10 20 30 40 60 0 10 20 80 0 10 20 30
INJECTION RATE
B/D/FT
-411-
-------
Figure 15
SPINNER SURVEY TOOL CONFIGURATION
PIPE
WIRELINE
CABLE
INSTRUMENT
IMPELLER
—I °
£
o
CASING CABLE-
MAGNET
PICKUP COIL
SPINNER
WELL CASING
•H
5pj)
t
N\
-412-
-------
Figure 16
TYPICAL TEMPERATURE SURVEY RESPONSE
C-BLOCK UNIT WATERFLOOD, WELL LLOYD #246
TEMPERATURE »F
4000'-s
5000'-
6000'- —
STATIC
TEMPERATURE
°F
-413-
-------
Figure 17
EXTERNAL CASING PACKER SCHEMATIC
MUD DISPLACED
FROM ANNULUS
HIGH SEAL LOAD AT
ROCK INTERFACE
FILTER CAKE COMPRESSED
AND DEHYDRATED
-414-
-------
INJECTION WELL FLOW REGULATION ASSEMBLY
X
\
-415-
-------
Figure 19
CONCEPTUALIZED POLYMER TREATMENT
.".'•••.'•'•••. '• '•:'.'.•%'••'TV "••e.'.•»*.*>• •*•" ••- '.c '•.:.< '.•;.•:,•«'''• •'
•?•';?• ..";•..*.••,.'•;"•• i••'« ••*'••• .••••,•"."•••••
FLUID FLOW
1. FLUID ENTERS STREAM TUBE - A
2. INJECT POLYMER
*. FLUID DIVERTED TO STREAM TUBE - B
LEGEND
WATER
OIL
WATER,
SOME OIL
[POLYMER [ POLYMER
FORMATION
-416-
-------
Figure 20
C-BLOCK UNIT
POLYMER TREATMENT RESULTS
TREATED SANDS
INJECTION RATE. BWPD
2000
1500-1
1000-
500-
BEFORE TREATMENT
AFTER TREATMENT
AVG. L-49 H-64 L-270 L-216 V-14 L-67 H-12a L-62
-------
2000'
Figure 21
C-BLOCK UNIT
POLYMER TREATMENT RESULTS
UNTREATED SANDS
INJECTION RATE, BWPD
BEFORE TREATMENT
AFTER TREATMENT
1600'-
00
1000'-
500'-
AVG. L-49 H-64 L-270 L-216 V-14 L-67 H-12a L-62
-------
Figure 22
GROSS PORE VOLUME INJECTION
C-BLOCK UNIT WATERFLOOD, PHASES 1
- 10
1.00
.80-
O
P
o
UJ
UJ
O
UJ
ffi
o
£L
.60-
.40-
.20—
o <
<2
uj O
< o
3*
si
-I UJ
c co
O uj
a.
-------
Figure 23
WATER ENTRY SURVEYS
PRODUCING WELLS, C-BLOCK UNIT WATERFLOOD
900
NJ
O
I
q
a.'
•
£
•
to
800-
700
600-
600 —
400-
300-
200-
100 -J
i I I r
Za AA AC
AE AQa
T
AH AX AM
MARKER
•T" I 'T'
AN AO
AP AQ1 AR AS
-------
Figure 24
INJECTION PROFILE STATUS
C-BLOCK UNIT WATERFLOOD
j>
t—>
i
i
Zs
AA
gs
AC
1-
®
AE
A3a
a 5
a »
i i
sJiJL
AH
a
a
SIS
a=
AK
i
..,
AP
L
-40
-30
3
•20 O
iu
o
-10
y s
AQ1
AS
SANDS
uugiiigiiiiiiig
LEGEND
FLOOOABLE AC. FT. - % UN8T TOTAL
CUM. INJECTION - % UNIT TOTAL
12/1 sea INJ. RATE - % UNIT TOTAL
8/1984 INJ. RATE - % UNIT TOTAL
-------
Figure 25
C-BLOCK UNIT OIL PRODUCTION
8000-
7500-
5% DECLINE RATE
o
Q.
O
OQ
«•*
111
z
o
I-
o
D
O
O
a-
a.
1979 '1980 ' 1981
5500-
5000
1982' 1983
YEAR
1 1984 I 1985 ' 1986 '
-------
STATUS OF MECHANICAL INTEGRITY TESTING
IN MISSISSIPPI
by
Lynnette A. Gandl and Desiree A. Landry
KEN E. DAVIS ASSOCIATES, INC.
11805 Sun Belt Court
Baton Rouge, Louisiana 70809
(504)293-2561
I. ABSTRACT
Pressure testing of Class II wells in Mississippi for mechanical
integrity has been witnessed for the last two years by Ken E. Davis
Associates (KEDA) personnel under contract to Engineering Enterprises
Incorporated and the Region IV United States EPA office in Atlanta,
Georgia. Initially, only those wells used for secondary recovery were
tested; however, all Class II disposal wells are currently being
tested. The testing conducted from September 1985 through March 1987
resulted in a failure rate of approximately 20* for all first time
tests. However, some of these failures were due to pressure .increases
caused by insufficient temperature stabilization prior to testing,
rather than pressure increases or decreases due to leaks. Evidence of
well failures has been observed at some sites, and some unusual well
completions have resulted in variable testing procedures. Suspected
groundwater contamination due to oilfield operations has also been
reported and preliminarily investigated in some areas of the state.
-423-
-------
II. INTRODUCTION-DEFINITIONS
The primary function of KEDA personnel in Mississippi has been
to act as witness to a variety of procedures required by the U.S. EPA
of the Class II operators in the state. The procedures are intended
to assure that all Class II wells in operation are operating in a
manner which assures that all Underground Sources of Drinking Water
(USDW's) are protected from contamination by the injected saltwater.
The principle concern has been to determine if the operating wells
are properly constructed and have mechanical integrity.
The procedure which is being used to assure internal mechanical
integrity is a pressure test of the annular space between the
injection tubing and the protection casing. A file review of all well
completion records is also conducted to determine if each well is
properly constructed to protect the USDW. Proper construction
includes surface or production casing set below the base of the USDW,
and sufficient cement behind the casing to prevent migration from
saltwater bearing zones into USDW's or freshwater bearing zones.
Other functions carried out by KEDA for the EPA have included
witnessing plugging and abandonment procedures to assure that the
plans approved by EPA are followed. Wells scheduled for plugging and
abandonment have included formerly operating Class II wells which do
not meet construction or MIT requirements, and unplugged or
improperly plugged abandoned oil and gas exploration wells within the
area of review of the permitted Class II wells.
-424-
-------
Unannounced inspections have also been performed on wells which
have been ordered shut down due to failure to meet permitting
requirements. The operational status and condition of these wells are
noted when inspected, and each is photographed by the field
inspector.
Other tasks requested specifically by EPA which have also been
performed are described in the following sections.
III. EESDLTS OF MECHANICAL INTEGRITY TESTS
A. Pressure Test Requirements to Prove Internal Mechanical
Integrity
The requirements for mechanical integrity which have been used
in Mississippi were designated by the U.S. EPA Region IV office in
Atlanta, Georgia. The pressure test to confirm mechanical integrity
of the well casing requires that the annulus be pressured to a
minimum of 300 psig and that this pressure must hold for a minimum of
30 minutes with no more than a 3% change in pressure. However, most
tests have been run at a minimum of 500 psig in order to meet the
requirements of the state of Mississippi. Tests which exceed 30
minutes may be allowed one additional percentage point of change for
each ten minutes, to a maximum of 6%. Failure is considered as either
an increase or decrease in pressure in excess of the allowed
percentage.
1. Pressure Decreases
A decrease in annulus pressure during testing can be caused
by leaks at a variety of locations in the system as follows:
The downhole packer on the tubing can leak,
-425-
-------
The casing can leak through a corroded spot or a parted joint
of casing,
The injection tubing can leak from the annulus into the
tubing during a test, or
The seals and valves at the surface can leak.
Additionally, leaks in the surface pipes leading to the
injection tubing can also cause surface and shallow subsurface
contamination.
A decrease in annulus pressure can also be the result of
injecting cold fluid down a well that is geothermally stable. As
the cold fluid is injected, the temperature of the annulus fluid
drops due to contact with the cold injection tubing, and the
corresponding pressure in the annulus drops. The temperature
will eventually stabilize if injection of cold fluid is
continued over a sufficient period for the well to reach thermal
equilibrium. At that time the well can be successfully tested
with meaningful results.
2. Pressure Increases
Pressure increases due to lack of internal mechanical
integrity can occur if the injection pressure exceeds the
annulus pressure and a leak in the injection tubing or packer
allows injection fluid to bleed into the annulus during testing.
However, a common reason for failure due to an increase in
pressure has been temperature increases caused when cold fluid
is placed into the annulus on the day of the test, and the
temperature is not allowed to stabilize before the annulus is
-426-
-------
pressured up. When cold fluid is placed in a deep borehole, the
increasing temperature of the earth with depth will heat the
annular fluid and cause the pressure in a sealed annulus to
rise. This rise in pressure as a result of thermal expansion
could mask a small leak and therefore such a test can not be
considered valid.
3. Results of Pressure Tests
The MIT's conducted during the period from startup in
August 1985 through March 1987 have included both secondary
recovery wells and saltwater disposal wells. Most of those wells
which failed their first MIT have been retested or will be
retested. The figures below reflect the results of MIT's through
the end of March, 1987.
Large Operators
Five oil companies operating in the state have each
had 19 or more of their wells tested. The total tested for
each company ranged from 19 to 47, and the failure rate
ranged from 5* for one operator to 32% for two operators.
Interestingly, the operator with the most wells had the
highest success rate. For the group as a whole, 22% or 37
of the total 171 wells tested failed their first MIT. Of
the failed wells which have been retested almost all passed
the second MIT, although not all of the failed wells have
been retested. Some companies have chosen to temporarily or
permanently abandon the wells which failed.
-427-
-------
Svall Operators
Fourteen companies have each had fewer than 10 of
their wells tested; ten of these companies have had only
one to three wells tested. A total of 37 wells were tested
for these fourteen companies with an overall failure rate
of 30* resulting from 11 failures. These small operators
are not necessarily independent oil companies, some are
majors with few wells in the state and/or few wells which
have been selected to be tested.
Well Locations
Of the 208 wells tested through March 1987 the
majority were located in Wayne and Yazoo counties, with 67
and 47 tests, respectively (Figure 1). Jasper and Lincoln
Counties had 28 each tested, and Pike county had 19. The
following counties had fewer than ten wells each tested:
Jefferson, Adams, Wilkinson, Lamar, and Amite.
As mentioned previously, wells were selected by EPA first
on the basis of use as secondary recovery wells, and then on the
basis of concentration of wells in an area and potential for
USDW contamination.
4. Unusual Well Completions and Tests
One type of completion for which an annular pressure test
is inappropriate is those wells in which the injection tubing
has been cemented into the hole. Because a minimal amount of
-428-
-------
annular space is present above the top of the cement a pressure
test of the annulus can not be used to test the mechanical
integrity of such a well. For wells such as these a radioactive
tracer survey is a more appropriate test.
Several wells are injecting into more than one disposal
zone, and due to this unique construction have been difficult to
test. Failures due to pressure drops occurred frequently and
successful tests were accomplished only after the packer was
reset or replaced. Three different methods of accomplishing
injection into multiple zones have been observed and are
described below.
a. Two tubings are run side by side to two different
depths (Figure 2B). A dual packer is in place which seals
off both tubings above an upper disposal zone, and a single
packer is set on the lower tubing, sealing off the lower
disposal zone. Two sets of perforations are usually
present: one at the base of the upper tubing, between the
upper and lower packers, and one below the lower packer
into which the lowest tubing injects. In one case, a zone
above the upper packer is also perforated, and injection in
the past has been down the annulus of the well.
b. Two tubing strings are run, one inside the other,
with the smaller, inner tubing at a greater depth than the
outer tubing (Figure 2C). Packers are set at the base of
each tubing string and injection is into perforations below
-429-
-------
each packer. In one case a third set of plugged
perforations is present above the upper packer and
injection was previously into this zone.
c. A tubing string is run with one or two sliding
sleeves set across one or two different disposal intervals
(Figure 2D). Packers separate the sleeve openings to
provide for controlled injection into either of the zones.
In some of these completions, the wells were originally
used to simultaneously produce oil from one tubing and inject
saltwater down the other tubing or down the annulus. However, we
are not aware of any wells currently being used in this manner-
B. Factors to Demonstrate External Mechanical Integrity
Casing and cementing records for each well tested are also
reviewed to determine the location of each string of casing, and the
calculated or measured height of the cement behind the production
casing. The location and thickness of the injection and confining
intervals are also reviewed, if known, in order to determine whether
the USDW's are being protected by sufficient confinement. The results
of the reviews are forwarded to EPA for followup as required. The
significance of USDW protection is discussed in the following
section.
IV. INJECTION ZONES AND USDW PROTECTION
In Wayne County in the eastern part of the state, the Lower
Wilcox Aquifer is the deepest USDW in most of the county. This
-430-
-------
aquifer consists of the lower portion of the Lower Wilcox Group and
the upper portion of the Naheola Formation. Under present laws,
injection of oilfield brines or drilling fluids into a USDW via a
Class II well is prohibited; however, numerous wells in Wayne County
in some adjacent counties have been injecting into the Lower Wilcox
for 30 to 40 years under previous authorization. The permit
applications for these Lower Wilcox wells in Wayne County have been
denied when application is made, and the wells have been ordered shut
down within a specific period of time. Some operators have requested
that an aquifer exemption be granted, on the basis of the historical
injection into the zone, and the water quality degradation which has
already occurred.
Insufficient confinement of the Lower Wilcox is suspected in
some of the areas near and within Wayne County. Therefore, EPA is now
requiring that supporting information such as geophysical logs, cross
sections and hydraulic conductivity values from core analyses be
submitted with each application to operate.
*
V. INSPECTIONS AND OTHER TASKS
KEDA has also been asked to perform additional tasks related to
the operation of Class II wells, which do not deal specifically with
Mechanical Integrity Testing. These tasks generally are performed in
order to verify whether or not wells which have not been permitted
have ceased operation and/or been properly plugged and abandoned.
Some additional tasks specifically requested by EPA have also been
performed.
-431-
-------
1. Unannounced Inspections
Since August 1986, wells which have been denied permits and
ordered shut down by a certain date have been inspected
following the ordered shut down date to determine if they are
operating. Additional wells for which permit applications have
been requested by EPA but have not been received have also been
inspected following a specific deadline. In some cases the wells
are operating when inspected; in many cases they appear operable
but are not operating when inspected. In some cases they are
obviously inoperable due to disconnected injection lines. In
other cases the well has apparently been plugged and abandoned
because no evidence of a wellhead is present at a site.
A total of 138 wells have been inspected unannounced since
August 1986. Of those wells, 41 (30*), were operating at the
time of the inspection. Four wells could not be located, either
due to plugging and abandonment or to incorrect location
coordinates. The majority of the 93 wells which were not
operating appeared to be operable as indicated by fresh paint,
connected injection lines, and new gauges and valves, but were
reported to EPA as not operating. The presence at some wells of
automatic timers in an "off" cycle when checked, indicates that
these wells may be use, although they were reported to EPA as
not operating.
Several of the wells which were inspected were operating
gas wells rather than disposal wells, due to apparent mis-
identification or filing errors.
-432-
-------
2. Witness Plugging and Abandonment Procedures
KEDA personnel have witnessed P & A procedures for fewer
than one dozen different wells, after approval by EPA of the
proposed procedure. Former Class II wells and former producing
wells and dry holes are both required to be plugged if they are
in the area of review of an operating Class II well, and have
not been properly plugged. The approved plugging and abandonment
procedure has generally consisted of the placement of 3 or 4
cement plugs as follows:
1) One opposite the perforated zone,
2) One at the base of the USDW (10,000 mg/1 TDS), to a
height of at least 100 feet above the USDW,
3) One at the base of the freshwater zone (1,000 mg/1 TDS),
to a height of at least 100 feet above the freshwater
limit, and
4) A 50 foot plug at the surface.
The casing has been pulled when possible, and has generally been
perforated at the plugging interval when it could not be pulled.
In some cases the base of freshwater and the base of the
USDW are essentially the same, and one continuous plug has been
set from 100 feet below the USDW to 100 feet above the base of
freshwater. In some cases additional plugs have been required at
the base of casing left in the hole due to insufficient cement
behind the casing. In other instances wells with multiple
-433-
-------
perforated zones or badly deteriorated casing have been plugged
from the base of the perforations to the surface. The
requirements for plugging are currently undergoing modification
by the EPA.
3. Suspected Surface Water and Groundwater Contamination
Chloride contamination of surface water and domestic water
supply wells has been reported in certain areas of south
Mississippi. The Water Resources Division of the U.S. Geological
Survey has conducted several studies of the situation in recent
years, (see the list of references), however, no conclusions
have been drawn in these reports regarding the specific source
of the chlorides: they could have originated from the former
storage of drilling mud and brine in unlined pits, from direct
discharge of brine into surface water, from disposal of brine
into Class II wells lacking mechanical integrity, or from leaks
through the unplugged well bores of abandoned oil and gas wells
and exploratory holes.
At EPA's request, KEDA field inspectors have collected
conductivity data from water supply wells in Pike and Lincoln
Counties; and have noted any reports of water quality problems,
such as cloudy water in drinking water wells coinciding with
injection into nearby Class II wells. The data collected is
currently being evaluated by EPA's Region IV office.
-434-
-------
V. SUMMARY
Through March 1987, a total of 23 % of the Class II injection
wells in Mississippi which were pressure tested for mechanical
integrity had failed their first test. However, the successful
retesting of the majority of the wells which failed indicates that
many operators have a desire to operate their wells correctly. By
contrast, 30 * of the unpermitted Class II wells which have been
inspected after being ordered shut down were still operating,
indicating a reluctance on the part of these operators to comply with
the permitting requirements.
Pressure testing is expected to continue until all of the
authorized or permitted wells in the state have been tested, and
those that fail are repaired or plugged. This testing is expected to
require several more years but is necessary in order to protect the
drinking water resources in the state of Mississippi.
VI. REFERENCES
Gandl, L.A., 1981, Characterization of Aquifers Designated as
Potential Drinking Water Sources in Mississippi; U.S. Geological
Survey Open-File Report 81-550, 90 p.
Hem, John D., 1970, Study and Interpretation of the Chemical
Characteristics of Natural Water; U.S. Geological Survey,
Water Supply Paper 1473, 383 p.
Kalkhoff, S.J., 1982, Specific Conductance and Dissolved Chloride
Concentrations of Freshwater Aquifers and Streams in Petroleum
Producing Areas in Mississippi; U.S. Geological Survey Open-File
Report 82-353, 33 p.
Kalkhoff, S.J., 1985, Brine Contamination of Freshwater Aquifers and
Streams in Petroleum Producing Areas in Mississippi; U.S.
Geological Survey Water Resources Investigation 85-4117, 116 p.
Kalkhoff, S.J., 1986, Brine Contamination of Shallow Groundwater and
Streams in the Brookhaven Oilfield, Lincoln County, Mississippi;
U.S. Geological Survey Water Resources Investigation 86-4087,
57 p.
-435-
-------
FIGURE 1.MAJOR AREAS OF OIL AND/OR GAS PRODUCTION
AND NUMBER OF WELLS TESTED IN MISSISSIPPI
THROUGH MARCH 1987.
-436-
-------
A. TYPICAL CLASS ff WELL; ONE TUBING,
ONE PACKER.
=
s
MM?
m
s s
"*
^^>-^^
\
{riST'
B. TWO TUBINGS; SIDE BY SIDE,
TWO PACKERS,
C. TWO TUBINGS; ONE INSIDE
THE OTHER. TWO PACKERS.
D. ONE TUBING WITH TWO SLIDING
SLEEVES, TWO PACKERS
FIGURE 2. TYPICAL AND UNUSUAL CLASS H WELL COMPLETIONS.
-437-
-------
WELL INTEGRITY MAINTENANCE USING PUMPABLE SEALANTS
R. Clay Cole and Kurt Lindstrom
Halliburton Services
ABSTRACT
Integrity of Class II disposal wells can be restored and maintained by
applications of a variety of pumpable sealants.
In this paper many of the diverse causes for failure of Class II disposal
wells to pass state and federal integrity tests are discussed, as well as
methods for identifying these causes. Experience with many well histories has
shown that, because of their diverse nature, not all of these problems can be
remedied with a single product or technique. Therefore, thorough problem
diagnosis is emphasized in addition to a description of the application of
several sealant systems to overcome the problems.
Considerations are given to overall well conditions, age of the casing
string, magnitude of the failure, associated well temperature and hydraulic
pressures, and the nature of the fluids against which a seal is required. Port-
land cement slurries (used alone, or in conjunction with secondary sealants),
true solution type sealants capable of entering the formation matrix, micro-
annuli, or pinholes in pipe are discussed.
INTRODUCTION
In the oil and gas production industry, wells used to reinject brine, to
aid in enhanced oil recovery, and for storage of hydrocarbons are referred to as
Class II wells.1 2 Environmental considerations require that these wells meet
certain standards to prevent leakage into ground water aquifer zones and/or to
other adjacent zones. Complete isolation of the wellbore is required, and
-438-
-------
mechanical integrity tests (MIT) are performed to establish the condition of the
wells.
Failure of MIT may be caused by a variety of mechanical failures in the
hardware installed to provide protection required.
1. Corrosion from extended exposure of casing to fresh water or brines. A
continuous and competent cement sheath around the casing helps prevent the
exposure, however if the sheath was not proper when installed or has failed
from subsequent damage, casing may corrode. Early completion practices did
not provide cement all the way to the surface, so casing was exposed to
corrosive fluids for long intervals.
2. Tubing leaks inside the casing can cause corrosion from inside out. Where
this occurs, the corroded interval may be extensive.
3. Fractured confining zones may allow fluid migration even when the well
itself is completed properly.
Restoration of well integrity is most commonly accomplished by squeeze
cementing, a technique that is also used to stop fluid migration through frac-
tured confining zones behind sound pipe strings. Externally catalyzed silicates
(ECSS) were developed specifically to control subsurface brine flow in producing
or injection wells. Also, an epoxy based system has been successfully used to
make a high strength bond between pipe and cement, thus plugging microannuli and
collar leaks.
This paper discusses the above methods briefly and provides an extensive
bibliography of references that give complete details. A complete description
and detailed operating procedure is presented on a more recently developed
family of sealants known as internally catalyzed silicate sealants (SS-I.
SS-II). Since there is little information on this method to be found in current
literature, more detail is offered. Sections on sealant selection and
-439-
-------
diagnostic methods apply to all sealing methods discussed.
DIAGNOSTIC METHODS
Forming a good definition of the problem is a vital early step in treating
a damaged disposal well. The original cement bond log might establish the
condition of the cement sheath and locate top of cement (TOC). If squeeze
cementing has been done, logs conducted after the squeeze are needed.
Temperature logs may help locate TOC. These logs are recordings of incre-
mental temperature changes occurring while the logging tool is lowered into the
well. From these records a temperature curve can be made; the curve may
indicate perforations, casing leaks and fluid channelling.
Direction and rate of travel taken by leaked fluids may be established by
injecting short half-life radioactive isotopes into fluids being pumped into the
well and monitoring their route with a gamma ray logging tool. The route and
rate of fluid travel relative to the physical geometry of the system may
indicate casing leaks, channels, packer and bridge plug leaks, etc.3
A spinner survey may be used to determine the location of casing leaks.
Fluid movement turns a propeller in the tool which directly measures the fluid
rate flowing past the instrument.1*
By isolating the hole with packers, the magnitude of the casing leak may be
determined from the injection rate. A straddle packer used in precision
perforation breakdown as described by Hushbeck can be used to isolate intervals
as short as 3 in.5
Although the foregoing tests may be expensive and time consuming, the
information they provide is critical to selection of the correct sealant and
execution of the treatment on the first try.
-440-
-------
PORTLAND CEMENTS
The method most commonly used to repair casing leaks has been, and is now,
to squeeze portland cement slurries into the spaces and voids around the casing.
When performed successfully, a squeeze cement job can plug leaks by forming a
permanent, high-compressive strength seal at a reasonable cost. Fluid loss
control, thickening time, and cement density are designed for each squeeze job
according to downhole conditions .
Best results are usually achieved by preceding the squeeze with a thin
fluid such as water to open and clean the zone of interest. The squeeze slurry
itself should have a sufficient thickening time and proper low fluid loss
characteristics to allow it to be pumped all the way to the repair area. If
these properties are not correctly selected, (1) the cement may dehydrate too
soon and leave the pipe plugged, or (2) a squeeze may enter in the wrong loca-
tion which would prevent slurry from penetrating the intended zone.
Displacement pressure should be limited to the minimum needed to accomplish
the squeeze as excess pressure can break down a weak formation. Foam cements
have provided significantly higher success ratios in controlling water zones and
in sealing corroded casings where low fracture gradients exist.7
EXTERNALLY CATALYZED SILICATES
An externally catalyzed silicate system (ECSS) has been developed specifi-
cally to control brine flow in producing or injecting wells. Generally. ECSS is
applied to the most severe channels behind pipe where brine flow is severe
enough to dilute conventional squeeze cement slurries. This system has been
used extensively in flood operations to improve oil/water ratio, and for repair
of casing leaks.® ^
ECSS may consist of two or three fluids, applied in either two or three
-441-
-------
stages. The process is designed so that chemicals in the second stage react
instantly with chemicals in the first stage to yield a plugging material. If
this plug is sufficient to complete the job, then the process is two fluid, two
stage. In the three stage process, portland cement is pumped behind the plug.
The following paragraphs describe each stage.
Stage 1 consists of pumping a special brine preflush into the formation to
cause a gelling reaction in the second stage. Although the material used in
Stage 2 will react with most formation brines, use of the preflush (Stage 2)
helps achieve the rapid formation of a plugging gel when chemicals of the two
stages meet.
In Stage 2, solids-free, non-Newtonian, inorganic silicate fluid intermixes
with the brine after being displaced from the tubing to form a gel which inhi-
bits flow through previously open channels. This chemical has a viscosity of
200 cp, and can carry up to 10 Ib (4.5 kg) of inert filler per gallon. To
obtain bridging in severe cases, silica sand and other special materials may be
added.
The third stage, which is normally used, is a low-water-loss portland
cement slurry. Since the first two stages drastically reduce flow within
channels, the leading edge of the Stage 3 slurry is able to combine with the
silicate of Stage 2. Both stages thicken at the point of interface.
Stages 1 and 2 should be pumped at pressure lower than the fracture gra-
dient. By the time Stage 3 enters the flow channel, a pressure buildup should
occur but the fracture gradient should not be exceeded. If pressure does not
build up, sequential injection of Stage 2 and 3 components are repeated until a
five-minute standing pressure can be maintained on the perforated interval.9
-442-
-------
EPOXY-BASED SEALANT
Persistent collar leaks and/or microannuli that do not respond to lower
strength treatments may occasionally be repaired with epoxy sealants. These
high-strength, true solution sealants bond tenaciously to both pipe and cement,
and success ratios in treating small leaks are high. If the leak to be repaired
is more than 5 bbl (0.79 m3)/day, inert fillers such as silica should be added
to fill voids and reduce fluid loss.10
The epoxy resin is strengthened by the addition of a chemical hardener.
Hardening of the mixture is accelerated by high temperature and a chemical
accelerator speeds the reaction by (1) reacting along with the hardener compound
to accelerate its reaction with resin and (2) reacting independently with the
epoxy resin, which further hardens the resin.10 Although this system is
successful, it is expensive and usually considered for unique or severe problems.
INTERNALLY CATALYZED SILICATES
General
A family of internally catalyzed silicate sealants (ICSS) was developed
specifically to offer an alternative to portland cement for the purpose of
squeezing off casing leaks and re-establishing zone isolation. ICSS sealants
offer flexibility of job design, competent sealing, casing protection, and ease
of removal.
Depending on the level of silicate content, these sealants are referred to
as SS-I (low level of silicate content), and SS-II (high level of silicate).
SS-I is used to provide a moderate seal that can be removed later. SS-II is
formulated to enter the matrix and form a permanent seal against moderate to
high pressures.
-443-
-------
SS-I Sealants
Properties
SS-I sealants are composed of water and inorganic silicates routinely used
as soil stabilization/solidification agents.11 Ungelled SS-I without fillers
(neat SS-I) has a viscosity of 1.2 cp, enabling it to penetrate tight forma-
tions. The co-reactant gel initiators, dissolved in fresh water, react with
silicates, become part of the gel network, and do not leach out with time.
Resultant gel is pH 10 to 11 and resists fluid penetration until broken up
physically.12
With a pH 11 gel surrounding a pipe or inside the pipe, corrosion due to
brines or fresh water is resisted. Figure 1 shows results of a 90-day study in
which the corrosion rate of J-55 grade tubing surrounded by SS-I gel was com-
pared to that of J-55 tubing immersed in tap water.12
Inert fillers
SS-I properties are enhanced by addition of a composite of inert fillers
such as diatomaceous earth, which is the optimum filler for SS-I, yielding
slurry density of 9.2 Ib/gal (1.1 kg/L). The nature of this inert filler is
such that it allows some slow fluid loss through the structure of the diatoms
(Fig. 2 shows structure). When establishing a fluid seal against a porous
formation loss of some fluid into the formation is desirable, since that fluid
also has some gelling qualities. Conversely, fluid loss must be limited so the
slurry will not dehydrate. Qualities of diatomaceous earth inert filler provide
both of these desirable characteristics.
Another filler in the composite slurry helps keep the other fillers in
suspension so that pipe removal is easier. Settled and compacted fillers can
prevent pipe removal.
-444-
-------
A third filler in the composite slurry is a bridging agent that helps
correct severe lost circulation. In situations where there is a brine column
behind the pipe and a 9.2 Ib/gal (1.1 kg/L) slurry inside the pipe, too much
fluid loss could cause a continuing "U-tube effect", using a high volume of
sealant. The bridging agent allows complete pipe fill-up if desired.
All the inert fillers in the composite slurry provide added strength to the
gelled product.
Gel Strength and Sealing
The texture of SS-I gel is best described as friable (easily crumbled) yet
rigid. When gelled in a container, the gel retains the shape of the container
when removed, thus the term "rigid gel" is used. The unsupported gel can be
broken up by very light physical disturbances, after which it does not congeal.
Broken pieces of the gel can act as check valves across pin holes and split pipe
openings.
SS~I gel strength is measured by using a penetrometer to gauge resistance
to penetration by a sharply pointed cone weighing 200 gm (Fig. 3). Neat SS-I
gel allowed 23.5 mm penetration; an SS-I slurry containing inert fillers as
described above allowed 17.4 mm penetration, or 25% improvement in strength.
Two tests were conducted to determine the level of hydraulic pressure at
which SS-I gels could provide an adequate seal.12
1. Two 10 in. (254.0 mm) x 2 in. (50.8 mm) stainless steel reservoirs were
plumbed together to allow series flow in the vertical direction (Fig. 4).
The bottom portion of the lower reservoir contained approximately 200 grams
of No. 70-170 U.S. sieve sand to provide a porous matrix media having a
permeability of around 4 darcies. This reservoir was filled with fresh
water containing fluoricein dye.
-445-
-------
Once the SS-I had gelled, a pressure test was performed on the system by
applying pressure on the fluorescein dye water in 50 psi (344.7 kPa) incre-
ments up to 300 psi (2068.4 kPa), with flow rates measured at each incre-
ment. The pressure test data are shown in Table 1. Fluid flow was checked
every 1/2 hour. No dye flow was detected and no damage to the gel was
observed during the test period.
2. A section of 2 3/8 in. (60.325 mm) tubing was placed inside a 4^ in. (114.3
mm) casing to simulate placement of SS-I sealant in the annulus (Fig. 5).
Twelve 3/4 in. (19.05 mm) holes were drilled on a 3 in. (76.2 mm) spacing
in a 12 in.
(304.8 mm) section of the casing to simulate casing damage. Three sets of
four holes were oriented 90° apart around the casing. The holes were packed
with resin consolidated 40-60 mesh Ottawa sand to provide a permeable
medium to simulate the leakage of fluid from the hole into the formation
matrix. The average permeability of the consolidated 40-60 mesh Ottawa
sand was about 40 darcies. Enough SS-I solution was placed in the annular
space to cover the entire 12 in. (304.8 mm) section. The annulus space was
then filled with dyed water. After the SS-I was squeezed with 100 psi
(689.5 kPa) for an hour, the test was shut-in to allow the SS-I gel to
form. After 48 hours of shut-in, the pressure test was performed by slowly
increasing the pressure from 0 to 500 psi (0-3447.4 kPa) with nitrogen on
top of water. The leakoff rate was measured at each pressure increment.
Results are shown in Data Table 2. This test presented an extreme condi-
tion, wherein only enough SS-II was used to just cover the target leaks.
It was concluded that even with 12 high permeability holes present in a 12
-446-
-------
in. long (304.8 mm) section of pipe that they can be sealed off to a
sufficient degree to pass an MIT. Only 1.0 psi (6.9 kPa) was lost in 30
minutes at 500 psi (3447.4 kPa) test pressure. Usually, several hundred
feet of SS-I is run above the shallowest known point of leakage to be sure
complete coverage is attained.
Allowance for Pipe Retrieval
The friable consistency of SS-I allows partial or complete filling of an
annular space with sealant, leaving it there, and subsequent retrieval of the
workstring. If cement is used, it is necessary to either reverse out liquid
cement after a squeeze pressure was attained, or allow the cement to attain a
soft set and drill out what remains in the casing. Drilling out cement can
cause severe damage to casing strings and liners.
From laboratory tests, it was soon recognized that SS-I type gels might
have the properties to allow a permanent annulus application that would permit
easy inexpensive removal. It has been determined that tubing, with or without a
packer, can be pulled through SS-I gel. Eight full-scale tests were conducted
using a test rig (Fig. 6) to determine the pull required to lift 180 ft (54.9 m)
of tubing string through a casing filled with SS-I gels, both neat and slurry.
In four of the eight tests an unseated retrievable type packer of the size
corresponding to the casing size used was attached to the bottom of the tubing
string. The various pipe specifications used are given in Table 3 along with
the pull data from all eight tests.
Tests were conducted by preparing SS-I formulations under field conditions,
then pumping the liquid sealant into the annulus between the pipes being used in
each test. The SS-I was allowed to gel and age overnight. Samples of the SS-I
material were saved to verify that a gel had formed in each case, and that its
-447-
-------
strength was normal. Then, the tubing was pulled without rotation, reciproca-
tion, or fluid circulation to break the initial bond.
Job Design and Performance
SS-I Jobs are designed to place sealant over the entire corroded casing
zone (Fig 7). Although placing sealant just to cover the leak has been shown to
seal sufficiently to withstand 500 psi (3447.4 kPa), it is advisable for two
reasons to place at least 300 to 500 ft (91.4 m - 152.4 m) of SS-I above the
shallowest known point of corroded casing.
1. Inaccuracies in determining the location of the leak may lead to use of
insufficient volume. If an interval of several hundred feet of leaking
exists, some small section may be overlooked. Some sections of the leak
may be temporarily plugged at the time of testing, also resulting in the
use of insufficient sealant.
2. If the leaks are fairly large, or the temporarily plugged sections of the
leak become open, more SS-I is lost to the voids outside the casing than
anticipated. This could result in some of the upper holes being left
untreated.
If the precise location of the leaks is not known, it is best to run SS-I
up to where the bottom of the surface casing is located, or to surface.
Two typical placement procedures are used. One is to preflush with fresh
water or light sodium or potassium chloride brines, then pump the sealant into
the annulus with the production packer seated. In effect this can be considered
a "bullhead squeeze" technique. Although many good results have been achieved
with this procedure, it can allow sealant contamination. With this procedure,
pump rates are restricted by the leak size, not the annulus size. Therefore,
with very low placement rates SS-I could fall through the brine in the annulus
-448-
-------
and become contaminated, resulting in no gel and no seal.
A better approach is to unseat the packer, pump a preflush and then spot
(pump rapidly with circulation to get the SS-I in the proper position) the SS-I
down to the packer. The packer is then set. The required pressure is then
applied to the annulus to accomplish a squeeze. When a squeeze is achieved, the
well is shut-in for a minimum of 24 hours after which the treated wellbore is
ready to be pressure tested. It is recommended that any pressure test be
carried on in a step-wise fashion so that if a premature failure occurs, at
least the extent of improvement will be known.
Example Treatment 1: The Oklahoma Corporation Commission (OCC) required a
customer in northeastern Oklahoma to pressure test the annulus of a disposal
well. The annulus between 7 5/8 in. (193.7 mm) casing and 5% in. (139.7 mm)
casing would take fluid at 3 bbl/min (0.48 bbl/min) at 150 psi (1034.2 kPa). A
retrievable packer was set inside the 5^ in. (139.7 mm) casing and pressure
applied to the 2 7/8 in. (73.0 mm) - 5^ in. (139.7 mm) annulus. It held pres-
sure with no leakage. At the same time fluid was flowing to surface from the 7
5/8 in. (193.7 mm) - 5% in. (139.7 mm) annulus. This indicated a leak inside
the 7 5/8 in. (193.7 mm) casing. A 3000 gallon batch of SS-I was prepared to
repair this leak.
Time (Minutes) __ Operation
0000 SS-I Solution mixed
0020 Solid fillers added to the neat SS-I
0065 Began to pump the 71.5 bbl (11.4 m3) of composite SS-I down
the 7 5/8 in. (193.7 mm) - 5^ in. (139.7 mm) annulus at 2
bbl/min (0.32 m3/min) at 0.0 psi
-449-
-------
0085 Pump rate slowed to 1 bbl/min (0.16 m3/min) [52 bbl (8.3 m3)
pumped]
0095 Stopped pumping (annulus dead)
0097 Resumed pumping SS-I at less than 1 bbl/min (0.16 m3/min) [61
bbl (9.7 m3)] pumped)
0115 Stopped to reprime pump
0120 Rate slowed to 1/4 bbl/min (.04 m3/min) as pressure began to
build
0125 All 71.5 bbl (11.4 m3) of SS-I in place in the annulus; Pressure
reached 300 psi (2068.4 kPa)
0130 Well shut in
Results: at 68 hours after the well was shut in, it held 1200 psi (8273.7 kPa)
for 10 minutes. In the final pressure test the annulus held 800 psi (5515.8
kPa) for 30 minutes, and passed the OCC test. Estimated savings in terms of
manpower, rig time, down time on the well, and the cost difference between the
SS-I job and other means of repairing the leak was about $25,000.
Example Treatment 2: In Kansas, a disposal well completed into the Arbuckle
formation had developed a casing leak. Initially, when 300 psi (2068.4 kPa)
pressure was applied to the annulus, bleedoff to 75 psi (517.1 kPa) occurred in
10 minutes.
Step No. Operation
1 Packer bypass opened
2 70 bbl (11.1 m3) of fresh water pumped down the annulus
3 17 1/2 bbl (2.8 m3) of a SS-I type treatment pumped down the
-450-
-------
annulus and displaced to packer
4 Packer bypass closed
5 Tubing volume of light sodium chloride brine pumped down the
tubing to displace any SS-I below the packer
6 Squeeze pressure applied to the annulus
7 Well shut in for 48 hours
Results: the well was pressure tested in three steps. In the final pressure
buildup the annulus held 285 psi (1965.0 kPa) for 30 minutes. The state
accepted the test and, one year later, the well is being used for disposal.
Table 4 lists results of mechanical integrity tests performed on 12 wells
that have received SS-I treatments to repair casing leaks. All of these wells
were given approval by the OCC and were put into service as injector or brine
disposal wells shortly thereafter.
SS-II Sealants
Properties
SS-II is placed as a low viscosity (1.7 cp) solution which contains no
undissolved solids, although fillers may be added if desired. The internal
catalyst allows a controllable pump time before the system sets to a stiff gel.
The material does not have significant strength in its neat form [15 psi (103.4
kPa)]. Its virtue lies in the matrix sealing quality of the system.8 13
In laboratory extrusion tests, SS-II gel was placed in matrices of various
types of unconsolidated sand packs. In 40-60 U.S. Mesh sand (40 to 50 Darcys),
gelled SS-II withstood 1500 psi (10342.1 kPa) before the seal failed. In 70-170
U.S. Mesh sand (9 Darcys) 2000 psi (13789.6 kPa) broke the seal. Sand packs
-451-
-------
were inside 1 in. (25.4 mm) ID by 3 in. (76.2 mm) long pipes, at room tempera-
ture. 11
Penetrometer tests, as described earlier, showed 16.6 mm average penetra-
tion for neat SS-II gels, showing a tougher structure than SS-I gels.
Applications
SS-II gel sealant provides higher strength and longer pump times (up to 600
minutes) than SS-I gels.8 n 13 SS-II is especially suited for repair of
pinhole leaks and the long placement times help achieve entry of SS-II into
formation pore throats and microannuli. The key to casing repair appears to be
in the ability of the operator to place sealant outside the casing, regardless
of the volume of sealant inside the casing.
SS-II may be used alone or as part of a combination of ICSS sealant/cement
squeeze. Primary use of SS-II material has been to help prevent bottom water
coning and help to seal selected zones. Qualities of SS-II allow radial injec-
tion deep into the formation. A cement tail-in provides the following synergis-
tic benefits on the treatment.8 ^ 13
1. Cement provides high compressive strength near the wellbore where differen-
tial pressure is greatest.
2. Cement give a positive indication that proper displacement has occurred.
3. SS-II reacts with cement to flash set near the wellbore and the cement
begins hydration almost immediately.
After setting, SS-II forms a firm, permanent gel inert to most chemicals.
It may be used to help form a barrier to lower brine zones or to help prevent
acid from subsequent treatments from contacting water zones below the treatment.
Application steps for SS-II may be the same as for SS-I, or the optional
procedure presented below may be used.
-452-
-------
Step No. Procedure
1 Locate damaged area
2 Pull tubing and packer
3 Perforate if needed to gain access to channel or leak area
4 Set a bridge plug below damage
5 Set a retrievable packer above zone of repair
6 Pump a fresh water preflush, followed by SS-II, spacer, and
cement (optional)
7 Squeeze to displace SS-II into formation or annulus; maintain
slight pressure on SS-II as it sets; avoid rapid injection of
SS-II at this point
8 Shut well in to allow strength to build
9 Wash out SS-II left inside casing
10 Retrieve packer and bridge plug
11 Test seal
SEALANT SELECTION
Selection of the proper sealant for the application requires consideration
of several factors. Shown below in tabular form are descriptions of some
commonly found conditions and a recommended sealing method for that condition.
These recommendations can not be considered "ironclad" since there may be other
variables that influence the selection.
-453-
-------
CEMENT BOND LOG INDICATIONS
Magnitude of Leak
Flow rate 5 bbl/
min at low pres-
sure. Most likely
well will not
stand full
above packer.
Medium flow
behind pipe.
(1 to 3 bbl/min)
Leak less than
1.0 bbl/min
Competent cement
with good bond to
pipe and formation
Evidence of frac-
tured formation be-
hind cement. Per-
forate to intersect
channel. Use ECSS
with low water loss
cement slurry as
tail-in. Repeat
Stages 2 and 3 un-
til well holds
pressure.
Method 1
If not accessible,
perforate to inter-
sect, try ECSS
squeeze with re-
peated steps of
brine and ECSS.
Method 2
If interval is less
than 50 ft, inject
neat SS-II formula-
ted to gel as it is
injected. If in-
terval is over 50
ft, tail-in SS-II
slurry.
Method 1
Squeeze with neat
SS-II followed by
SS-II slurry. Re-
move excess from
casing.
Method 2
Squeeze neat SS-I
followed by SS-I
slurry. Leave in
casing.
Method 3
If well has high
pressure, or if
well is Class I,
consider using
epoxy based sealant.
Some cement present
but poor bond to
pipe and formation
Try ECSS squeeze
first. Perforate if
necessary to inter-
sect channel. Re-
peat Stages 2 and 3
until well holds
pressure.
Method 1
Perforate to inter-
sect flow. Squeeze
with ECSS and
cement slurry.
Method 2
Treat with SS-I or
SS-II slurry.
Method 3
Use low pressure
cement squeeze if
drillout is no
problem.
Method 1
Squeeze with neat
SS-II followed by
SS-II slurry. Re-
move excess from
casing.
Method 2
Squeeze neat SS-I
followed by SS-I
slurry. Leave in
casing.
Very poor or no
cement present
behind pipe
Apply combination
of ECSS and ce-
ment. May require
repeated applica-
tions of both.
Consider adding
bridging agents
to ECSS and cement
slurry.
Method 1
Perforate to
intersect flow.
Apply ECSS squeeze,
repeating Stages 2
and 3.
Method 2
Apply Hesitation
squeeze using Port-
land cement alone.
Method 1
Apply ECSS squeeze
with one or more
cement stages.
Method 2
Apply large volume
squeeze with SS-I
or SS-II slurry.
Consider preflush-
ing with calcium
brine.
-454-
-------
Pinhole leak, Method 1 Consider perfor- Method 1
well stands Use neat SS-I or ation to open ac- Perforate to inter-
full but will neat SS-II. cess to fractures sect void area. Run
not hold an
-------
and Gas Conservation Division," 1986 Edition.
3. Murphey, Bill, "Squeeze Cementing Requires Careful Execution for Proper
Remedial Work" Presented to SPE Squeeze Cementing Symposium, Dallas, Tx.,
1985.
4. University of Texas at Austin, "A Dictionary of Petroleum Terms", Second
Edition, 1979.
5. Hushbeck, D.F., "Precision Perforation Breakdown for More Effective Simula-
tion Jobs," Presented to the International Meeting on Petroleum
Engineering, March 17-20, Beijing, China.
6. Smith, Dwight, "Cementing" SPE Monograph Series, Volume 4.
7. Bour, D.B., Creel, P.G., "Foam Cement for Low-Pressure Squeeze Applica-
tions," presented at the 1987 Southwestern Petroleum Short Course, Lubbock,
Tx, April 22-23, 1987.
8. Cole, R. Clay, Mody, Bharat, Pace, James, OE81 SPE10396.1 "Water Control
For Enhanced Oil Recovery", Presented to the Offshore Europe Technology
Conference, Aberdeen, Scotland, 1981.
9. Smith, C.W.; Pugh, T.D.; Mody. B.: "A Special Sealant Process for
Subsurface Water", Presented at the Southwest Petroleum Short Course,
Lubbock, Tx., Aug., 1978.
10. Cole, Robert C. , SPE 71874, "Epoxy Sealant for Combatting Well Corrosion"
Presented to the International Symposium on Oilfield and Geothermal Chemis-
try, Houston, Tx., January, 1979.
11. Dalrymple, Dwyann; Sutton, David; Creel, Prentice: "Conformance Control in
Oil Recovery", Presented at the Southwest Petroleum Short Course, Lubbock,
Tx., April, 1985.
12. Cole, R. Clay, Dalrymple, D., McDuff, C.H., Jones, Mark, "Chemical Process
Seals Leaks in Injection Wells", Presented at the Southwest Petroleum Short
-456-
-------
Course, Lubbock, Tx, April, 1987-
13. Koch, R.R., Mclaughlin, H.C., "Field Performance of New Technique for
Control of Water Production or Injection in Oil Recovery," paper SPE 2847,
presented at SPE Improved Recovery Techniques meeting, Fort Worth, Tx,
March 1970.
Before Treatment
During Treatment
After Treatment
Table 1
SS-I Pressure Test
Fluid Loss
in 30 min
(cc)
107.0
1.2
4.0
16.0
19.0
20.5
21.0
21.0
Flow Rate
Through Sand
(cc/min)
960.0
3.56
0.04
0.1333
0.5333
0.6333
0.6833
0.7000
0.7000
Pressure
(psi)
10
10
50
100
150
200
250
300
-457-
-------
Table 2
SS-I Sealing Test
Holes per foot = 12
Pack sand = Consolidated Ottawa 40-60 mesh sand
SS-I Volume = 3 liters (13 in. - 14 in. from bottom)
Temperature = 75°F
Before Treatment
Flow Pressure
(psi)
2.5
2.5
4.5
4.5
Flow During Treatment
Squeeze Pressure
(psi)
10
20
30
40
50
60
70
80
90
100
Flow Rate
(cc/min)
4500
4420
5240
5220
Flow Rate
(cc/min)
0
0
0
0
0
0
0.02
0.02
0.02
0.0366
Leak Off
Rate
(cc/min/psi)
1800
1768
1164
1160
Avg. = 1473
Leak Off
Rate
(cc/min/psi)
0
0
0
0
0
0
0.0003
0.0003
0.0002
0.0004
-458-
-------
After Treatment
Test
Pressure
(psi)
100
200
300
400
500
Table 2 (Con't)
SS-I Sealing Test
Flow
Rate
(cc/min)
0.56
1.10
2.10
2.9
1.42
Leak Off
Rate
(cc/min/psi)
0.0056
0.0055
0.0070
0.0073
0.0028
Leak Off
Pressure
Per 30 Min
(psi)
1 psi
Table 3
Pull test data obtained from full scale testing using 180 feet of tubing inside
casing.
Pull Test Data for SS-I Neat Formulations (No Filler)
Test
No.
1
3
5
7
Test
No.
2
4
6
8
Casing Size
(in.)
7.0
7.0
7.0
4.5
Pull Test Data
Casing Size
(in.)
7.0
7.0
7.0
4.5
Tubing
(in.)
4.50
2.875
2.875
2.875
for SS-I Slurry
Tubing
(in.)
4.50
2.875
2.875
2.875
Packer
no
no
yes
yes
Formulations
Packer
no
no
yes
yes
Final Pull
lb/ft2
19.47
21.7
34.37
43.72
Final Pull
lb/ft2
28.42
40.40
57.00
56.63
-459-
-------
Table 4
Mechanical Integrity Test Results
Location
Kansas
Oklahoma
Oklahoma
Oklahoma
Oklahoma
Oklahoma
Kansas
Kansas
Oklahoma
Oklahoma
Oklahoma
Oklahoma
Depth
(ft)
1,539
1,290
712
2,205
2,200
2,800
2,370
2,717
1,270
2,600
3,375
4,000
Treatment
Volume
(gal)
1,000
600
500
2,000
1,500
3,000
500
1,000
800
250
1,000
1,000
Required
Test Pressure
(psi)
300
200
200
200
200
800
300
300
1,000
250
300
300
Achieved
Test Pressure
(psi)
300
200
100
300
200
880
300
300
1,000
350
300
300
-460-
-------
0.1
g* 0.04
ff
0.01
0.001
0.0001
Fig. 1
CORROSION OF J-55 CASING SAMPLES
LIQUIDS IN WHICH THE J-55 CASING WAS TESTED
m TAP WATER • SS-I SLURRY
DATA FOR 75 DEGREE F EXPOSURE
20
40 60
TIME (days)
80
100
The maximum acceptable corrosion rate is 0.04 Ib/sq ft.
This occurred with water after 60 days exposure.
-461-
-------
FIGURE 2
1000X Photomicrograph showing the SS-I filler structure through which some very
limited fluid loss into adjacent formations is achieved.
-462-
-------
FIGURE 3
A Precision Scientific Penetrometer
This instrument is used to determine the relative strength of SS-I gels by
measuring the distance of penetration attained by the cone point into the gel
structure. ASTM (D-217) test procedures and specifications were used in these
determinations.
Y./
-463-
-------
Fig. 4
Sealing test 1 apparatus used to evaluate SS-I gels
Gauge
N, Inlet
Water Reservoir
SS-I Gel Inside
a 10" x 2" Reservoir
70-170 U.S. Mesh Sand
200 gpm
-464-
-------
.5
Process To Seal Leak
injection Well
IP
N, Inlet
Water
Tank
Reservoir
Pressure Bleed Off
Valve
4Vz" Casing
23/a" Tubing
Dyed Water
SS-I Gel
3/4" Pipe Nipples
Packed with Consolidated
40-60 Mesh Sand
To Drain
-465-
-------
Fig. 6
SS-I Service
Pull Test Well Configuration
/si
V\V\V\>yS
fch
•^
Casing
QQ_I in 1Qn' nt anniilne
rioai uoiiar on ena OT luoing
or Retrievable Type Packer
— - Cap on end of casing
•Cemented Wellbore in Test Well
-466-
-------
.7
SS-S Job Placeman! Schematic
Tubing
Holes in Casing
Packer
Casing Damage Area
-467-
-------
MEASURING BEHIND CASING WATER FLOW
by
T. M. Williams
Texaco, Inc., E & P Technology Division
Box 425, 5901 S. Rice Avenue, Bellaire, Texas 77401
ABSTRACT
A common problem encountered in water injection operations is
locating and stopping undesired water channeling. The Texaco
E & P Technology Division has developed a Behind Casing Water Flow
(BCWF) measurement system to measure vertical water flow in or
behind multiple casings. This nuclear logging system can measure:
• the direction of flow
• the linear flow velocity
• the volume flow rate
• the radial distance of the flow from the sonde.
The system uses a neutron generator tube to provide a source
of high energy neutrons to activate oxygen in the flowing water.
The resulting high energy gamma rays are detected with two crystal
detectors. By using the counts in different energy ranges in the
two detectors, the system computer can calculate the water flow
velocity, volume flow rate, and radial distance from the flow to
the sonde. Velocities of between 0.75 and 10 in/sec (19 and 254
mm/sec) can be measured. This 3-5/8 inch (92 mm) diameter logging
sonde is reversible so either upward or downward flowing water can
be detected. This logging system has been used in several Texaco
fields and has proved its value in detecting undesired water flow.
-468-
-------
INTRODUCTION
A common problem encountered in production and water injection
operations is locating and stopping undesired water channeling.
The Texaco E & P Technology Division has developed a Behind Casing
Water Flow (BCWF) measurement system to measure vertical water flow
in or behind multiple casings. This nuclear logging technique can
measure:
• the direction of flow
« the linear flow velocity
• the volume flow rate
• the radial distance of the flow from the sonde.
PRINCIPT.K OF BCWF LOG
The BCWF log is based on a nuclear activation technique in
which flowing water is irradiated with high energy (14 MeV)
neutrons emitted by a neutron generator within the logging sonde.
These neutrons interact with the oxygen nuclei in the water to
produce the radioactive isotope nitrogen-16 through the
016(n,p)N16 reaction. Nitrogen-16 decays exponentially in time
with a halflife of 7.13 seconds, emitting 6.13 and 7.12 MeV gamma
radiation. An oxygen activation gamma ray spectrum is shown in
Figure 1.
The characteristic activation gamma rays are identified in
this plot of gamma ray intensity per energy versus gamma ray
energy. The 2.615 MeV thorium peak, which is used for energy
calibration, is also marked. The water flow parameters of interest
are computed from the energy and intensity response of two gamma
ray detectors mounted within the logging sonde.
-469-
-------
Figure 2 shows schematically a two-detector BCWF sonde in a
well bore where water channeling occurs within the cement annulus
behind the well casing. The channeling water is activated as it
flows past the neutron source. The gamma rays from the activated
flowing water are first measured when the water passes the first
detector and then again when it passes the second detector. During
its travel from detector 1 to detector 2, the gamma intensity
decays by an amount determined solely by the travel time; that is,
by the distance between detectors and the linear velocity of the
water. Consequently, the ratio of the detector 1 to detector 2
count rates is an exponential function of the water velocity only.
The linear water velocity is determined from this function
regardless of the radial position of the water channel from the
BCWF sonde. The direction of the water flow to be measured is
determined by the relative position of the neutron generator and
detectors. That is, so that flow can be measured in both
directions, the sonde has been made reversible.
Figure 3 graphically shows the laboratory apparatus used to
calibrate a BCWF sonde. Water of various metered rates can be
pumped through different PVC pipes simulating flow channels. The
PVC pipes are positioned at various radial distances from the sonde
and one or several casings can be inserted between the pipes and
sonde.
Figure 4 shows the linear velocity calibration results for a
BCWF sonde containing two 2-inch (51mm) diameter x 6-inch (152mm)
long Nal(Tl) detectors spaced 18 inches (457mm) apart. As
predicted, the logarithm of the ratio of the detector count rates
-470-
-------
is a linear function of I/velocity, and velocity is given by:
velocity = A^/ClnfcJ/cJ) - AQ] (1)
where
AO and A^ are calibration constants,
C? = net counts in detector 1 from 3.7 to 7.2 MeV, and
G£ = net counts in detector 2 from 3.7 to 7.2 MeV.
For this detector spacing, linear velocities of 0.75 to 10 in/sec
(19 to 254 mm/sec) can be measured with reasonable accuracy.
The material between the water channel and the detectors
degrades the primary gamma ray energy; that is, the number of
primary (6.13 and 7-12 MeV) gamma rays is reduced and the number of
low energy gamma rays is increased. This effect is used to
determine the radial distance from the sonde to the flow. The
H
ratio of: (1) the counts in the energy window (C?) from 4.9 to 7.2
MeV, to (2) the counts in the energy window (C,) from 3.25 to 4.0
MeV, is related to the total number of electrons per unit area
between the water channel and detector 1. The relation for a
detector spaced about 40 inches (1 m) from the source is given by
where aQ, a^, and a2 are calibration constants and pem is the
total number of electrons per unit area between the sonde and flow
channel .
The electron density for different materials is known. Thus
with the borehole fluid, casing, and cement information, the radius
from the sonde to the water flow can be calculated from /9em.
Theory suggests and experiments have shown that the gamma
ray activity divided by the volume flow rate is a function of:
-471-
-------
(1) water velocity, (2) the radial distance of flow channel to the
sonde, (3) the type and amount of material between the flow and
sonde, and (4) the output of the neutron tube.
Data analysis reveals that the logarithm of gamma count rate
per unit volume flow rate can be expressed by a second order
polynomial in ln(v) as given in equation 3.
where
ln[Cn/qw] = b0(R,pem) + b1ln(v) + b2[ln(v)]2 (3)
c" = the net detector 1 count rate from 3.25 to 7.2 MeV,
gw = the water volume flow rate,
v = the linear water velocity,
bO = C0 + C1R + C2R2 + C3'em' and
b,, b-/ CQ, c,, c-, and c- are calibration constants.
This relationship is demonstrated graphically in Figure 5,
where the detector 1 net count rate/water volume flow rate is
plotted versus linear velocity for various radii of the flow
channel to the sonde and for various casing sizes and combinations.
The volume flow rate can be determined from this plot
regardless of the flow channel cross-sectional area, which is not
accessible to measurement. Velocity and gamma ray count rate are
measured by the BCWF sonde; then, knowing radial distance and type
of intervening material, the volume rate can be obtained as shown
in Figure 5.
In short, the water velocity and volume flow rates can be
determined from the gamma ray spectra measured by the BCWF sonde
without knowledge of the location and cross-sectional area of the
flow channel and the intervening material.
-472-
-------
Additional information on the theory of operation of this
logging system may be found on page 121 of the January 1979 issue
of the Journal of Petroleum Technology.
FIELD TESTS
Field tests with the BCWF system have been successfully
performed in several wells, including the four in Texas chosen as
examples for this presentation.
Two (No. 123 and 124) were logged in June 1985 to determine
the source of the salt water channeling behind casing to the
surface. These wells had been recently drilled to about 2000 ft
(610 m). 5-1/2 inch, 17 Ib. (140 mm, 25.3 kg/m) production casings
were set to TD and 13-3/8 inch, 72 Ib. (340 mm, 107 kg/m) surface
casings were set to 40 ft (12 m). Both surface and production
casings were cemented with cement circulated to the surface.
Prior to the BCWF water flow measurements, a long spaced
neutron-gamma ray log or a natural gamma log was run in each well
to select the depths for the BCWF measurements and to correlate
BCWF depths with those of available commercial logs.
Well No. 123
A cement-bond/gamma ray/CCL log and a temperature log were
available on Well No. 123. The temperature log indicated a
potential source of the salt water channeling at about 750 feet
(229 m). The BCWF log was run to confirm the temperature log
results and to locate any additional sources of salt water
channeling. Stationary flow measurements, each of 15 minute
duration, were made opposite shales at 12 locations. These 12 were
between 13 and 900 feet (4 and 275 m) and were above formations
-473-
-------
which could contribute and/or be the source of the salt water
breaking out at the surface. The results are given in Table I.
Based on a preliminary wellsite interpretation, the casing was
perforated below 470 feet (143 m) and the well was squeezed with
100 sacks of cement, which stopped the breakout of salt water at
the surface.
The BCWF log also revealed what appeared to be flow from a
zone near 700 feet (213 m) to a zone near 600 feet (183 m). This
was reported to the field for corrective action.
Well No. 124
Well No. 124 was logged with the BCWF log about 24 hours after
casing was set. The well was about 600 feet (183 m) east of No.
123 and it was suspected that the source of the channeling salt
water was at about the same depth interval as in well 123.
Consequently, stationary flow measurements of 15 minute duration
each were made at 12 locations opposite shales and above formations
which could contribute to the salt water breaking out at the
surface. The results of the log is given in Table II.
Based on a wellsite interpretation, the casing was perforated
and a cement squeeze made. These measures stopped the water
breakout at the surface.
Well Nos. 3521 and 5334
Tests were also performed in two other Texas wells in
September 1985. This field had a history of casing problems
between 3200 and 4200 feet (973 and 1281 m). These two wells were
logged primarily to determine if behind casing water flow was a
cause of casing erosion/corrosion in this field. A second reason
-474-
-------
for running the BCWF log in these wells was to determine if
cementing of the casing in the problem zone had stopped the
suspected water flow.
Well No. 5334 produced 20 BOPD and 22 BWPD before the pump and
tubing were pulled for logging operations. This well had 7-5/8
inch (194 mm) casing set to 3166 feet (965 m) and cemented to the
surface. The 5-1/2 inch, 15-1/2 Ib. (140 mm, 23 kg/m) production
casing was set to 7949 feet (2423 m) and cemented. A temperature
survey indicated the top of the cement was at 5390 feet (1643 m).
Stationary measurements were made at nine locations between 4485
and 3250 feet (1367 and 990 m). The measurements at 4485 and 4040
feet (1367 and 1231 m) indicated no water flow. Each station
between 3726 and 3250 feet (1136 and 990 m) indicated an upward
water flow behind the casing of approximately 4 BWPD.
Well No. 3521 was shut in because it produced only water. The
well had 8-5/8 inch (219 mm) intermediate casing set to 3287 feet
(1002 m) and cement circulated to the surface. The 5-1/2 inch,
15-1/2 Ib. (140 mm, 23 kg/m) production casing was set to 7300 feet
(2225 m) and cemented. A temperature log indicated the top of the
cement was at 2650 feet (808 m). Logging runs to measure both up
and down flow were made at ten locations between 3020 and 4675 feet
(920 and 1425 m). No water movement was detected. This indicates
that cementing the problem interval did stop the water flow behind
the casing.
PLANNED DEVELOPMENT
This logging technique has been effective in locating behind
casing water flow in many locations. We are currently improving
-475-
-------
the electronics to double the data acquisition rate. This will
reduce the time needed at each station to measure the water flow.
We are currently studying the feasibility of constructing a
1-11/16 inch (43 mm) diameter model of the BCWF sonde. This small
diameter would increase the number of wells in which the BCWF
logging system could be effectively used. To use the standard
3-5/8 inch (92 mm) diameter sonde, the tubing must be pulled in
most wells. In many wells, water flows behind the casing only when
the well is being produced or during water injection. Thus, with
the tubing pulled, no flow is observed with the BCWF logging system
unless the casing can be pressurized to induce flow.
SUMMARY
The system uses 14 MeV neutrons to create radioactive nitrogen
from the oxygen in the water. The decay of this radioactive
nitrogen can be measured to locate water flow.
The capabilities of the BCWF system are:
• the direction of water flow (up or down)
• the linear flow velocity from
0.75 to 10 in/sec (19 to 254 mm/sec)
• volume flow rates > 10 BWPD
• flow radially displaced up to
10 inches (254 mm) from sonde
• flow behind up to 3 casings
This logging technique has been licensed to Dresser-Atlas.
They are developing a field system and plan to offer this system as
part of their regular service.
-476-
-------
FIGURE CAPTIONS
Fig. 1 A typical BCWF gamma ray spectrum when water is flowing.
Fig. 2 A BCWF sonde in a well bore with water channeling within
the cement annulus behind the well casing.
Fig. 3 BCWF calibration facility. Different pipe sizes can be
placed at selected radial distances from the sonde. Different
casing combinations can be used during calibration.
Fig. 4 This is the velocity calibration results for a typical BCWF
sonde. V is the velocity and C^ and C^ are respectively, the
counts in detectors 1 and 2 between 3.7 and 7.2 MeV.
Fig. 5 This plot shows the volume flow rate can be determined
without knowing the flow channel cross-sectional area.
-477-
-------
CD
I
ID
O
O
1400
1200
1000
800
600
400
200
0
0
BEHIND CASING WATER FLOW
TEST PIT DATA
ENERGY (MeV)
6 7
Figure 1
-------
DUAL DETECTOR BCWF SONDE
14 MeV
SOURCE
-479-
Figure 2
-------
CALIBRATION FACILITY
CONTROL
VALVE METER
3-5/81
nfciznr^
PUMP
-480-
Figure 3
-------
0.7
CN/CN VS
I 2
L - L = 18 INCHES
52 5i
_| L
0.2
I L I I
0.6
'/v
0.8
-481-
Figure A
-------
DETECTOR I TO SOURCE = 39.4 INCHES
50
30
20
10
O
o
3 2
ZD
_J
O
.07
R CASING
2.60 NONE
3.00 NONE
4.30
5.65
5 10 15 20
VELOCITY (In/sec)
NONE
71
7'+4'/21
NONE
9%'
9 % '+7'
9 s/8'+7i+4l/2>
-482-
Figure 5
-------
TABLE I
BCWF LOG RESULTS
WELL No. 123
RECORD
NUMBER
1
2
3
4
5
6
7
8
9
10
11
12
DEPTH
feet meters
822.7
801.7
771.4
751.4
721.4
690.7
613.7
562.8
233.2
79.8
29.2
13.3
250.8
244.4
235.1
229.0
219.9
210.5
187.1
171.5
71.1
24.3
8.9
4.1
VOLUME
BWPD
0
0
0
2.8
2.8
50
38
9
286
86
-483-
-------
TABLE II
BCWF LOG RESULTS
WELL No. 124
RECORD
NUMBER
1
2
3
4
5
6
7
8
9
10
11
12
DEPTH
feet meters
782.6
766.4
740.6
694.0
635.7
594.7
511.7
482.0
442.1
82.0
35.1
24.1
238.5
233.6
225.7
211.5
193.8
181.3
156.0
146.9
134.8
25.0
10.7
7.3
VOLUME
BWPD
0
0
79
107
83
13
15
3.2
>107
3.1
-484-
-------
A PILOT SURVEY OF STATE MECHANICAL
INTEGRITY TESTING (MIT) PROGRAMS IN NEW MEXICO
Kelly Nash, ERT, 12655 North Central Expressway,
Dallas, Texas 75243
Raleigh Kreuz, ERT, 3000 Richmond Avenue,
Houston, Texas 77098
Jack Marr, ERT, 3000 Richmond Avenue,
Houston, Texas 77098
ACKNOWLEDGEMENTS;
The authors gratefully acknowledge the assistance and the
patience of the New Mexico Oil Conservation Division staff, in
particular, Prentiss Childs, Jerry Sexton, Frank Chavez, Evelyn
Downs, Bonnie Prichard, and Charles Gholson. Special thanks
also to Judith Anderson of ERT for compiling the bulk of the
data used in this study* Finally, the UIPC Research Foundation
provided critical review and funding for the project.
ABSTRACT;
A pilot survey of State MIT Programs for Class II wells
under the EPA UIC Program, was conducted in New Mexico.
Records of 217 annulus pressure test failures (of 1309 MITs) ,
witnessed and recorded by the New Mexico Oil Conservation
Division in 1984 and 1985,, were reviewed. File data on test
conditions, well construction, and subsequent workovers were
listed on an automated database. The database allows an
evaluation of the diagnostic abilities of positive pressure
testing of the tubing - casing annulus as compared to
monitoring annulus pressure, for which records are also
available. Information listed in the database includes:
-485-
-------
1.) Pressure conditions in the tubing and casing strings
prior to and at the beginning and ending of testing.
2.) Well construction data - including initial completion
dates, casing and packer set depths, and injection
intervals.
3.) Failure type indicated - primarily casing, packer and
tubing failures were indicated by the survey.
4.) Well repair data - the survey included information on
the type of repair, estimated cost of the repair, and
details such as the casing hole interval (where
identified).
The spreadsheet feature of the database allows the
calculation of frequency distributions of such features as
casing hole interval, age of well, and repair types. Reviewing
the information in the database allows an evaluation of the
various factors which may lead to pressure test failures.
The study indicates that the annulus pressure test will
detect holes in casing. 64% of the test failures were
associated with casing holes. Most holes were in uncemented
sections adjacent to saline zones below underground sources of
drinking water.
The age distribution of the injection wells which failed
the annulus pressure test was a function of general historical
drilling activity if they were completed prior to the early
1970's. Wells completed later were not significantly
represented in the database.
The average cost of repairs necessitated by conditions
leading to an annulus pressure test failure was estimated to be
$11,000.
The New Mexico MIT program was reviewed and evaluated with
the test data. The New Mexico program is more stringent than
the EPA program in that annulus monitoring is conducted
annually on all wells. Annulus monitoring in positive pressure
-486-
-------
injection wells identifies leaks of injection fluid through
tubing or packer and casing leaks opposite pressurized zones.
The first level of USDW protection in the New Mexico program is
centered around monitoring of the tubing and casing annulus.
The pressure test looks at the second level of protection - the
production casing string. There is a third level of protection
- surface casing. Usually, this third level would have to be
breached before a USDW would be endangered by a failure of the
second level.
The annulus pressure tests did not identify any evidence
of injection wells which caused groundwater contamination.
-487-
-------
INTRODUCTION;
ERT, A Resource Engineering Company, was retained by the
Underground Injection Practices Council (UIPC) Research
Foundation, to conduct a Pilot Survey of the Mechanical
Integrity Testing (MIT) program for Class II injection wells in
the State of New Mexico.
The survey focused on Mechanical Integrity Tests of the
casing, tubing or packer. Potential fluid movement into an
Underground Source of Drinking Water through vertical channels
adjacent to the injection well bore was not considered as part
of the survey.
The survey included a review of the New Mexico MIT
program and comparison with EPA MIT requirements and the
compilation of MIT and well workover data to indicate the
diagnostic abilities of annulus pressure tests.
ENVIRONMENTAL PROTECTION AGENCY (EPA) MECHANICAL
INTEGRITY TESTING (MIT) REQUIREMENTS;
EPA Mechanical Integrity Testing requirements,
promulgated in 40CFR146.08, defined Mechanical Integrity as
having two parts: 1) the absence of a significant leak in the
casing, tubing or packer; and 2) the absence of significant
fluid movement into an Underground Source of Drinking Water
(USDW) through vertical channels adjacent to the injection well
bore. The regulations specified that one of the following
tests must be used to evaluate the absence of significant leaks
as defined in 146.08: 1) monitoring of annulus pressure; or 2)
pressure test of the annulus with liquid or gas. The
regulations also provided an avenue for the use of alternate
Mechanical Integrity Tests, when approved by the EPA
Administrator. These testing requirements were prepared for
-488-
-------
use by EPA in direct implementation programs, and EPA expected
states to use these tests in the State UIC programs implemented
under the authority of the Safe Drinking Water Act.
On December 5, 1980, the Safe Drinking Water Act was
amended and, among other changes, the amendments added a new
Section 1425 to the Act. Section 1425 established an
alternative method for a state to obtain primary enforcement
responsibility for those portions of its UIC program related to
the recovery and production of oil and gas (i.e., Class II
Injection Wells). The Amendments specified that if a State
program meets the requirements of Sub-paragraphs a-d of Section
1421(b)(1) of the Safe Drinking Water Act, and represents an
effective program to prevent underground injection which
endangers drinking water sources, EPA shall approve the
program. On May 19, 1981, EPA published guidance on the
implementation of the alternative demonstration provided for in
the new Section 1425. The guidance included the criteria EPA
would use in approving or disapproving applications under
Section 1425. The guidance established the following tests as
adequate to demonstrate the absence of significant leaks: 1) a
pressure test of the annulus with liquid or gas; 2) the
monitoring of annulus pressure in those wells injecting at a
positive pressure, following an initial pressure test; or 3)
all other tests or combinations of tests considered effective
by the State Director.
NEW MEXICO MIT REQUIREMENTS;
The UIC program for Class II (enchanced recovery or salt
water disposal) injection wells in New Mexico is administered
by the Oil Conservation Division (OCD) of the New Mexico Energy
and Minerals Department. The OCD was granted primary
enforcement authority for the UIC program under the federal
Safe Drinking Water Act on February 5, 1982.
-489-
-------
The MIT requirements for Class II wells are found in Rule
704, pursuant to the Oil and Gas Act, as follows:
1. Prior to commencement of injection, initial
integrity testing of the casing, tubing, and packer
(if used) including pressure testing of the
casing-tubing annulus.
2. At least every five years thereafter, testing to
assure continued mechanical integrity, including:
a.) measurement of annular pressures in wells
injecting at positive pressures under a packer
or balanced-fluid seal;
b.) pressure testing of the casing-tubing
annulus for wells injecting under vacuum
conditions; and
c.) other tests which are demonstrably
effective and approved by the OCD.
3. The OCD can require additional testing when deemed
advisable, including the use of tracer surveys,
noise logs, temperature logs, or other test
procedures or devices.
4. The OCD may order tests to be conducted prior to the
expiration of five years if conditions warrant.
5. The operator must notify the OCD of the scheduling
of MITs so that a Division representative may
witness the test.
-490-
-------
6. Rule 704 was amended in 1986 to require casing
pressure tests whenever the tubing is pulled or the
packer is unseated.
In addition to MIT requirements, Rule 704 also specifies
monitoring requirements which serve as a continuous
demonstration of mechanical integrity:
1. Injection wells must be equipped such that injection
pressure and all casing annular pressures could be
measured at the well head and the injected volume
may be determined at least monthly.
2. Injection wells used for storage must be equipped
such that both injected and produced volumes may be
determined at any time.
When tests indicate that wells are defective, the
operators are required to take corrective action. Operators
must submit for approval a description of proposed repairs on
Division Form C-103, "Sundry Notices and Reports on Wells",
which is also used to report on completed work. District
inspectors schedule and witness follow-up pressure tests.
Rule 116 requires operators to report mechanical failures
or downhole problems which might endanger fresh water, such as
"spills, leaks or blowouts". The Rule requires appropriate
corrective action for injection well failures.
Rule 705 requires notification to the Division of
commencement, discontinuance and abandonment of injection
operations (e.g., plugging and abandonment).
Rule 706 requires monthly or annual reporting of
injection volumes or pressures on Form C-115 (enchanced
recovery and pressure maintenance) or Form 120-A (salt water
disposal).
-491-
-------
In addition to the testing, monitoring, and reporting
requirements specified in the OCD rules, when easily corrected
problems such as small surface leaks or excessive injection
pressures are noted by Field Inspectors, the problems are
brought to the attention of the operator for immediate
corrective action, under the general authority of Rule 1303,
"Duties and Authority of Field Personnel".
NEW MEXICO TESTING PROGRAMS;
At least twenty-five percent of the pressure tests
carried out on wells injecting on a vacuum are witnessed by
District Field Inspectors (NMOCD, 1981). When tests are not
witnessed, the operators are required to file the test results
with the District offices.
The District offices place a strong reliance on
observations and records of pressures and other pertinent data
for each string of casing and tubing during injection. The
procedure for field observations of annulus pressure conditions
(referred to in New Mexico as a "bradenhead test") is as
follows:
1. The operator closes in the valves on the bradenhead
24 hours prior to testing.
2. The operator opens each valve during the test and
the NMOCD inspector records pressures and other
pertinent information. Often, a short puff of air
or a short flow of water will result from the
expansion of the tubing caused by injection
pressures and temperature differentials. A
continuous pressure or fluid flow at the surface
indicates either a tubing or packer leak (for wells
injecting at a positive pressure) or casing leaks
adjacent to pressurized water-bearing formations.
-492-
-------
NMOCD observations of annulus pressure have been carried
out in District 1 since 1974, District 2 since 1979 and
District 3 since 1981 (NMOCD, 1981). The testing program
includes producing wells in addition to injection wells.
In some injection wells, the weight of the fluid column
in the well is sufficient to push water into the injection zone
without applying pressure at the surface. If a well takes
water at a particular rate faster than it can be filled to the
surface, the surface injection pressure will be less than zero
and the well is said to operate under vacuum conditions. For
these wells, bradenhead tests may not indicate if tubing or
packer leaks are occurring. Therefore, the State requires
periodic annulus pressure tests, under Rule 704, for wells
injecting under a vacuum.
Aside from the annulus monitoring program and the
periodic pressure tests, bond logs, radioactive tracer surveys,
temperature logs and other special tests are carried out in
areas where problems are suspected (NMOCD, 1986).
Most of the New Mexico pressure test and well records are
located in three District offices in Hobbs (District 1) ,
Artesia (District 2) and Aztec (District 3) which serve the
eight oil and gas producing counties in New Mexico. There are
presently approximately 4400 active Class II injection wells
which are distributed roughly as follows;
District 1 (Hobbs) - 2,600
District 2 (Artesia) - 1,300
District 3 (Aztec) - 500
Since 1983, New Mexico has reported to EPA on the number
of MITs conducted and the number of MIT failures, as follows:
-493-
-------
Year MITs # Failures % Failing
1983 3502 75 2.1%
1984 3713 148 4.0%
1985 3199 430 13.4%
1986 (Jan-Sept.) 2519 98 3.9%
TOTAL 12,933 751 5.8%
Higher reported failure rates in 1985 are partially due
to a special study for which approximately 2,000 wells
injecting at positive pressures were pressure-tested between
March, 1985 and March, 1986 (NMOCD, 1986).
SPECIAL STUDIES OF MIT IN NEW MEXICO;
The New Mexico OCD conducted a study for EPA (completed
in June, 1986), which discusses the results of comparison
testing between annulus monitoring (referred to in the study as
"bradenhead testing") and pressure tests of 416 injection wells
in 1984, and summary data on mechanical integrity testing of
2,091 injection wells in 1985 and 1986 (NMOCD, 1986). The OCD
study and the accompanying background data provided the
foundation for the design and implementation of this survey.
The conclusions from the Phase 1 testing of 416 wells in
1984 included the following (NMOCD, 1986):
"1. The bradenhead test is adequate to find tubing and
packer leaks or casing leaks with pressure on the
zone where the leak is located.
2. The problems found during this study represent no
significant threat to fresh water since most of the
casing leaks were found below the surface casing in
the salt sections where pump-in pressure is known to
be betweern 400 psi and 1000 psi. Such pressures
resulting from tubing or packer leaks would be
detected by the regular bradenhead testing.
-494-
-------
3. The data collected supports the conclusion that
without a significant tubing or packer leak the
majority of casing leaks cannot be found with the
bradenhead test. However, it should be pointed out
that where the bradenhead test did not show casing
leaks, the tubing and packer were mechanically sound
so no movement of fluid was occurring in the casing
annulus.
4. The bradenhead test is not adequate for finding
tubing or packer leaks on vacuum injection wells."
The Phase 2 testing in 1985 and 1986 resulted in
additional conclusions;
"Problem Areas - Wells in Districts I and II have fairly
comparable geologic conditions and, as would be expected, found
fairly comparable problem situations; viz., in the salt
sections or at the base of the red beds (tertiary). The salt
section, approximately 1000 feet in thickness, provides a
corrosive environment to external surfaces of intermediate or
production casing strings when left uncemented through the
salt. However, cemented surface casing above the salt protects
fresh water. The red beds tend to swell in the presence of
fresh water to the extent of closing off the annular space
between pipe and formation, setting up a zone of intense
corrosivity. Caliche beds near the surface are also highly
corrosive.
In District III, the most corrosive zone appears to be
the Menefee which is the middle productive zone of the
Mesaverde formation. This is an electrolysis problem in the
entire area for all types of wells. It is not related
specifically to injection."
The final recommendations of the OCD study were:
1. Continue annual observations of annulus pressure
during well operation bradenhead tests.
-495-
-------
2. As resources permit, supplement annulus monitoring
with positive pressure tests.
3. Require casing pressure tests during workovers.
4. Require reports of tubing repairs and changes in
packer set depths, and require that the packer be
set not more than 100 feet above the perforations.
STUDY AREAS;
A subset of records from the NMOCD special study,
consisting of wells which failed the positive pressure tests
during the special study discussed in Section 4 were reviewed
for this project. Specifically, records for pressure test
failures in 1984 and 1985 from the District 1 (Hobbs) and
District 3 (Aztec) offices were reviewed. This consists of
approximately 220 test and well construction records. These
are the most complete records available on well integrity and
provide representative comparisons of annulus pressure tests
and bradenhead tests in New Mexico. District 1 wells, in the
southeast part of New Mexico, are completed in formations which
are primarily structurally controlled by the Delaware and
Permian Basins. District 3 wells, in the northeast part of the
State, are completed in formations which are structurally
controlled by the San Juan Basin. These are the major oil and
gas producing regions in New Mexico.
DISTRICT 1 GEOLOGY;
District 1, in southeast New Mexico, lies generally
within the Great Plains Physiographic Province. The plains are
a remnant of an alluvial plain built up by eastward-flowing
streams from the Rocky Mountains. There are essentially no
surface streams in District 1 and the Tertiary sands and
gravels of the Ogallala formation provide the only extensive
source of fresh water for the area (Boyer, 1986). The
-496-
-------
saturated thickness of the Ogallala formation is approximately
200 feet in the eastern part of District 1 and the formation is
removed by erosion to the south and west (USGS, 1984). Minor
local sources of fresh groundwater are also present in
sandstone layers in the Triassic "Red Beds" and the Permian
Rustler formation (Nicholson and Clebsch, 1981). All known
USDWs lie above the Permian salt section (NMOCD, 1981).
Subsurface structure in District 1 is controlled by the
Permian and Delaware basins and the associated shelf-reefs
which formed during the Paleozoic era. The hydrocarbon
reservoir rocks are entirely Paleozoic and the production is
from the fields on the Central Basin Platform and the Northwest
shelf. Ninety percent of the state's oil production has come
from southeast New Mexico, with commercial production since
1924 (NMBMER, 1981). The reservoir rocks are predominantly
limestone, largely Permian in age. Over half the oil
production is from the Grayburg and San Andres formations.
Other important hydrocarbon reservoirs include the Yates, Seven
Rivers, Queen and Abo Reef formations. Most of the oil and gas
produced in southeast New Mexico is structurally trapped in
anticlines (Landes, 1970).
Standard casing practice in District 1 is to set casing
to the top of the Permian salt section (Salado formation) and
cement to the surface. The salt section itself acts as a
confining layer to prevent out-of-zone water from hydrocarbon
reservoirs from entering shallower USDWs (NMOCD, 1981).
DISTRICT 3 GEOLOGY;
District 3, in northwest New Mexico, lies within the
Colorado Plateau Physiographic Province. The topography is
much more rugged than in the southeast. The San Juan, Rio
Puerco and Chaco Rivers provide important surface water
supplies and the associated alluvium is a source of fresh
groundwater.
-497-
-------
The subsurface geology of northwest New Mexico is
structurally dominated by the San Juan Basin. The major fresh
groundwater sources in the San Juan Basin are Cretaceous
sandstones, most importantly the Ojo Alamo sandstone (which
immediately overlies the Kirtland shale (Brimhall, 1973).
Freshwater also occurs in the Menefee member of the Mesa Verde
Formation and locally in the Morrison formation. Fresh water
may be found at depths to 3000 feet in District 3. Since these
aquifers are often artesian, cemented production casing is used
to isolate these aquifers. Artesian conditions also contribute
to the effectiveness of annulus monitoring in indicating casing
leaks (pers. comm., E. Busch, OCD).
In northwest New Mexico, most of the hydrocarbon
production has been from Cretaceous rocks, mostly sandbar-type
stratigraphic traps, and also from fractures in the Mancos
shale (NMBMER, 1981). Gas has been commercially produced in
District 3 since 1921 and oil since 1922.
DATA ELEMENTS;
After reviewing the file material available at the New
Mexico Oil Conservation Division (NMOCD), a survey form was
developed, an example of which is shown in Figure 5-4. The
main sources of information available in the OCD files were the
inspector worksheet for the pressure tests, the workover forms
filed after the pressure tests, and the initial well completion
report, which contained construction details of the wells.
After the records were reviewed, the key data to conduct
the evaluation was determined and set up on a spreadsheet
program. The data elements included:
1) Record Number - unique for each well test.
2) Pressure drop-rate in psi/minute. In general, the
tubing-production casing annulus was pressured to
-498-
-------
approximately 300 psi. In many cases, it was not
possible to reach the initial pressure of 300 psi,
or even to fill the annulus with packer fluid.
3) Completion date - the year that the well was
initially completed.
4) Bradenhead Tests Pass/Fail - whether the well was
considered to have passed or failed an initial
observation of annulus pressures.
5) Test Data - the initial, start test and end of test
pressures inside the surface, intermediate, and
production casing and the injection tubing.
6) Type of Failure - a review of the test records and
workover forms indicated, in many cases, the type of
failure which was indicated by the pressure test
failure.
7) Squeeze/Leak Interval - when the interval of the
casing hole was identified by subsequent workovers,
including plugging and abandonment, the depth
interval was recorded.
8) Approximate Cost - cost estimates were made for each
repair.
9) Surface Casing Depth - as indicated by the
construction records.
10) Surface Casing Cement Circulated - Yes or No.
11) Packer Set Depth,
12) Injection Interval.
-499-
-------
13) Type of Packer - (if known).
14) Injection Pressure.
15) Additional Remarks.
COST ESTIMATES OF REPAIRS;
Costs were derived by reviewing each record, summarizing
each job, estimating the total time required to complete, and
then totaling the cost based upon the work performed. For
example, if a packer was re-set only and the well re-tested and
passed, this would require one day, as follows:
Rig - 1 day $1,250
Pump Truck - 4 hours 600
Estimated Cost $1,850
The cost applied to each phase of a job from quoted daily
costs obtained from various oilfield service companies and
operators, including workover contractors, cement companies,
tool rental companies, logging companies, etc.
The total cost of the repairs for which sufficient
information was available (143) and for which wells
subsequently passed the annulus pressure test is $1,564,000 or
an average of $11,000 per job.
DATA ANALYSIS;
Summary Statistics.
A total of 217 records of annulus pressure test failures
were reviewed. These represented all the records which could
be located for the District 1 and 3 annulus pressure tests
-50CK
-------
conducted in 1984 and 1985 for the (NMOCD, 1986) study
previously discussed. A test failure in the study was
generally defined as the decrease of more than 10% in annulus
pressure over 15 minutes. The annulus pressure test failure
records identified in this study are from a total of 1301
mechanical integrity tests, of which 263 failed. The procedure
for testing was that pressure tests followed bradenhead tests.
Not all wells which failed bradenhead tests were subjected to
annulus pressure tests. District 3 did not conduct any annulus
pressure tests for wells which failed the bradenhead test
(pers. comm.( E. Busch, OCD).
The average depth of the top of the injection zone in the
database is 3,866 feet.
The age distribution of the wells in the database is
shown in Figure 1. In general, the age distribution shows good
correlation with historical drilling, activity until the early
1970gs. Figure 2 shows a comparison of the distribution of
ages (initial well completions) of wells in District 1, which
failed the annulus pressure test with the total well
completions for each year in eastern New Mexico as reported by
the International Oil Scouts Association (I.O.S.A.,
1930-1983). The figure suggests that improvements in casing
materials and cementing technologies in the 1970's will be
evaluated by future pressure tests. In the 1990's, wells
completed during the "boom" years of the late 1970's will reach
the age of those which showed casing problems in this survey.
In general, the findings of this survey supported the
conclusions discussed in OCD, 1986 regarding the problems
indentified by the annulus pressure test. The primary problem
type identified was holes in the production casing. The
distribution of problem and repair types is illustrated in
Figures 3 and 4.
Casing problems were identified in 88 wells in one of
three ways:
1. Holes found during subsequent workovers,
-501-
-------
2. Problems reported by operators, or
3. Flow of fluid or pressure increase in the pipe
string outside the string being tested.
The locations of casing leaks were determined in 79
wells. The distribution of the top of the leaks is shown in
Figure 5. The average depth of the top of the casing holes
found was 1,595 feet. Most of the holes were in uncemented
sections of the production casing below the surface casing, as
discussed in (NMOCD, 1986).
Records of ten wells with old casing perforations above
the packer were found. In one case, the Bureau of Land
Management was requiring the operator to monitor annulus fluid
levels. Twenty-two wells with packer problems were
identified. Five of these showed evidence of packer leaks
during the bradenhead tests, four were shut-in and two were
injecting on a vacuum. Tension-set packers may have been
unseated by the pressure test in some instances. For example,
the packer in one well was unseated by the operator during the
well workover by applying 500 psi annulus pressure. OCD, 1986
also discusses packer failures, and states that some reported
packer failures were repaired by moving the packer slightly
up-hole to cover a casing leak, but remaining set in the
injection zone.
Tubing leaks were identified in nine instances. Two of
these failures were in wells that passed the bradenhead test.
The remaining test failure types were nine miscellaneous
valve, line and wellhead packing leaks which were identified
after the initial testing and do not indicate mechanical
integrity problems.
The type of repair was determined for 175 wells. The
most common repair was a cement squeeze for casing leaks (61 of
175 wells or 35%). Twenty-eight wells (16%) were simply
plugged and abandoned. Forty wells (23%) were reported as
shut-in, with seven of these having a cast iron bridge plug
-502-
-------
set. The packer was reset on 15 wells (9%). In some cases, it
was not necessary to pull the tubing to accomplish this. On
nine wells (5%), it was necessary to repair or replace the
packer. Twelve wells (7%) were tested again and passed the
annulus pressure test the second time.
GROUNDWATER IMPACTS;
An actual determination of groundwater impacts would
require a site-specific investigation at each suspected problem
well. While the actual determination of groundwater impacts
from injection wells which failed the annulus pressure test is
beyond the scope of this survey, a general assessment of
contamination potential can be made.
OCD, 1986, which formed the basis for this study, noted
that no evidence of contamination of a USDW was found at any of
the 2,507 well locations where tests were conducted. Likewise,
this survey did not identify any evidence of contamination of a
USDW.
From OCD, 1986 and the previous discussion, it is obvious
that annulus monitoring can detect tubing or packer problems in
wells injecting at a positive pressure. It can also detect
casing problems opposite pressurized zones. Where such
conditions exist, the potential for contamination is greatest.
These conditions are readily detectable by annulus pressure
observations.
Injection under a vacuum poses minimal threat to USDWs
since it is unlikely, under these conditions, that the
hydraulics exist that could cause injected fluids to enter and
contaminate a USDW. Before injection fluids could enter the
formation, fluid levels would rise to a level above the USDW
sufficient to overcome the natural pressure gradient of the
USDW formation.
The increase in pressure down-hole in an injection well
is due to the weight of fluid in the hole (approximately 0.5
psi/ft.). If the pressure at the surface is negative (i.e.,
-503-
-------
vacuum conditions), positive pressure conditions will exist at
some point down-hole. It can be intuitively reasoned that if
this point is below a USDW, no contamination of the USDW from
injection fluid can occur. Even though the annulus monitoring
is less definitive for vacuum wells, these wells are less
likely to contaminate groundwater.
SUMMARY;
o ERT reviewed the New Mexico Oil Conservation
Division's program of mechanical integrity tests
(MITs) for Class II injection wells.
o Records of 217 pressure test (PT) failures were
reviewed. Records represented 1309 MITs in New
Mexico during 1984 - 1985.
o A database was developed listing test and well data.
o Failure types, repairs and repair costs were
analyzed.
o The PT failures were mostly due to casing holes.
Most holes were in uncemented saline zones.
o The age distribution of well failures tracks
historical drilling trends up to the 1970«s.
o The average cost per well associated with a PT
failure was estimated at $11,000.
CONCLUSIONS;
o The pressure test will detect holes in casing.
-504-
-------
The analysis of pressure test failures did not
identify evidence of USDW contamination.
The New Mexico annulus monitoring program is more
stringent than the EPA UIC requirements and provides
a high level of USDW protection.
The first level of protection for underground
sources of drinking water in the New Mexico Class II
UIC program is centered around monitoring of the
tubing and casing annulus. The pressure test looks
at the second level of protection, the production
casing string. There is a third level of protection
surface casing. Usually, the third level would
have to be breached before a USDW would be
endangered by a failure of the second level. The
$1,564,000 spent to repair wells failing pressure
tests was necessary to meet New Mexico well
construction requirements, but was not required
because of an imminent threat to a USDW.
-505-
-------
REFERENCES:
Boyer, David G., 1986, Differences in Produced Water
Contaminants from Oil and Gas Operations in New Mexico
Implications for Regulatory Action, Presented at National Water
Well Association Conference on Southwestern Groundwater Issues,
Tempe, Arizona, October 20-23, 1986.
Brimhall, Ronald M., 1973, Ground Water Hydrology of Tertiary
Rocks of the San Juan Basin, New Mexico, in: "Cretaceous and
Tertiary rocks of the Southern Colorado Plateau, A Memoir of
the Four Corners Geological Society, p. 197-207.
Gutentag, E., et. al., 1984, Geohydrology of the High Plains
Aquifer in Parts of Colorado, Kansas, Nebraska, New Mexico,
Oklahoma, South Dakota, Texas and Wyoming, U.S.G.S. Prof. Paper
1400-B.
International Oil Scouts Association (1939-1980), International
Oil and Gas Development Yearbooks, I.O.S.A., Austin, Texas.
Landes, Kenneth (1970), Petroleum Geology of the U.S.,
Wiley-Interscience, New York, NY.
New Mexico Bureau of Mines and Energy Resources (NMBMER), 1981,
New Mexico's Energy Resources '80', Circular 181, Santa Fe, NM.
Oil Conservation Division (OCD), 1981, Primacy Application for
Class II Injection Wells, EPA Region 6, Dallas, Texas.
Oil Conservation Division (OCD), 1986, Comparison Test Between
a Bradenhead Test and Pressure Test, Final Report, EPA Grant
No. X811232-01-3, OCD, New Mexico Energy and Minerals
Department, Santa Fe, New Mexico.
-506-
-------
AGE DISTRIBUTION OF ALL RECORDS
i
Ul
o
10 -
17 -
16 -
15 -
U-
13 -
12 -
11 -
10 -
0 -
8 -
7_
6 -
5 -
-
3 -
2_
1 -
0_
R
I * t i . I
—
r
/
?
/
X
^
•j
'
j
'
'
/
/
X
/
J
/
y
y
X
X
^
/
j
r
/
R
r
/
r
j
X
J
/
J
f
/
n
fl
p
/
7
/
^
r
>
r
/
n
/
r
/
/
•
^
/
^
^
V
^
^
^
/
/
*
^
/
/
X
4
>
>
>
7
/
/
>
r
>
f
s
t
/
'
7
/
X
/
— ,
^
llp^ ^ ^^
y v v ^ 1^
r r r r ^r
/> I ^
r ^r
' ' P1F1 *" ' Plr PI P1P1 PI
LIT , , rjr j JT j
"HP"* lR 1 RfflR
i • • • • i • • • . i . . . . , . . . , ,r
M
O
193O 1935 1940 1945 195O 1955 196O 1965 1970 1975 1980 1985
DATE OF iNmAL WELL COMPLETION
# WELLS COMPLETED
-------
AGE DISTRIBUTION - DISTRICT ONE
O
00
I
O
o
m
o
j»
I
I
5*
o
1945 1950 1955 1960 1965
DATE OF INITIAL WELL COMPLETION
1975
1980 1985
-------
Type of Failure
Other (6.5%)
Tubing (6.5%)
Casing" Perf. (7.2%)
i
£>
I
Packer (15.9%)
H
O
a
Casing (63.8%)
-------
Type of Repair
Shut In (22.9%)
f Other (1.1%)
Rep. Casing (1.7%)
Repair Tubing (2.9%)
Repair Packer (5.1%)
Re-test OK (6.9%)
Cement Squeeze (34.9%
H
a
a
Reset Packer (8.6%)
Rug and Abandon (16.0%)
-------
b.
O
*
tu
O
2
D
Z
DISTRIBUTION OF CASING HOLE TOPS
24
22
20 -
18 -
16 -
14 -
12 -
10 -
8 -
6 -
4 -
2 -
AS IDENTIFIED DURING REPAIRS
z
8
500 1000 1500 2000 25OO 3000 3500 4000 4500 5OOO 55OO 6OOO 6500 7000
TOP OF HOLE INTERVAL (DEPTH IN FEET)
No. OF WELLS
-------
PLANNING SUCCESSFUL TEMPERATURE SURVEYS
John G. Berner, Conoco Inc., Houston, Texas
ACKNOWLEDGEMENT
The author appreciates the critical review of this paper
provided by fellow employees, Rob Bedard, Ed Dew, and Paul
Pilkington. Permission to present this paper was provided
by Conoco Inc.
ABSTRACT
Temperature surveys are often run in injection wells as soon
as a problem is suspected. Very little thought goes into
analyzing the possible problems and thinking about how they
will affect the temperature survey, if at all. Often the
survey is simply run and then an attempt is made to analyze
the results.
Some pre-testing, data gathering and considerable thought
should precede the temperature survey. Pressure communica-
tion does not necessarily mean fluid flow is occurring.
Temperature anomalies do not exist if fluid flow is confined
to its normal injection path.
-512-
-------
Injection should be maintained at its normal state unless it
is known that contamination of the USDW is occurring and can
be stopped by shutting in the well. A survey should be
planned, and if it does not show anomalies, then some can be
created according to the test procedure.
A differential temperature survey will accentuate an anomaly
from the continuous temperature survey. The most definitive
survey, however, may be several successive surveys known as
a temperature decay log. A decay log is normally successful
at distinguishing zones of fluid entry into the formation.
TEMPERATURE SURVEYS FOR PROBLEM INJECTION WELLS
Temperature surveys are run for a variety of reasons. They
are run in some areas on a schedule, perhaps annually, to
monitor injection profiles. They are also run as the first
choice when a problem is discovered. The temperature survey
is quick and relatively cheap. The problem is that the
planning stage is also "quick". Very little thought has
been given concerning what the temperature log should or
might show. The result is that the log is ultimately sent
to the company "expert", with inadequate accompanying data,
for interpretation.
-513-
-------
A problem in an injection well could be discovered by a
sudden change in the injection rate-pressure relationship.
The well could suddenly develop pressure on the tubing-
casing annulus or a surface seepage could be spotted. A
surface seepage would require closing in injection because
the ground waters are being contaminated. In the other
cases it would be preferable to not disturb the injection
until after the first temperature survey is run.
Fast action is required. Shutting in the injection will not
stop pollution of the ground waters if it is occurring,
because the injection zone will probably back flow for a
long period of time. Shutting off injection and doing
nothing creates a worse day of reckoning later. The problem
needs to be reported and a remedial program planned.
WELLBORE DIAGRAMS
Temperature surveys should be planned for success. One of
the first tasks is to draw a diagram of the downhole equip-
ment. The well file should be scanned thoroughly in order
to determine the equipment currently installed in the well.
The minimum I.D. of all of the pieces should be determined
along with any obstruction that might exist in the injection
string. Circulating valves should be noted and all perfora-
-514-
-------
tions, open or squeezed, below the packer or above, should
be included.
Figure No. 1 shows a simple injection well with casing,
tubing, packer and one set of perforations in the injection
zone. This is the simplest situation that is possible and
should lead to a simple analysis of a temperature log.
Hopefully, several feet of rathole will exist below the
bottom perforation. This will allow the temperature survey
to read the normal temperature at that depth. Complications
occur when multiple injection zones, uphole perforations,
circulating valves, liners and twin well injection are
added, as in Figure No. 2.
Another item of interest would be the type and density of
fluid in the annulus. In a lot of old wells, this value is
probably not known with any degree of accuracy. A sudden
indication of pressure on the casing-tubing annulus could
mean a lot of things. An analysis of the situation could
point out possibilities to plan for. The injection pressure
(Pi) can be compared to the annulus pressure (Pa). If the
leak is at the bottom of the injection string, Pa = Pi minus
tubing friction pressure losses minus difference in densi-
ties of fluid in annulus and tubing.
-515-
-------
WELL DIVGNOSIS
In order to remove the unknown value of friction pressure,
it may be desirable to shut off injection for a few minutes.
A comparison between stabilized Pi and Pa can then be made
to see if they are directly related. If the annulus fluid
is considerably heavier than the injection fluid, then a
large difference in Pa and Pi would be expected. On the
other hand, there could also be a hole in the casing. If
Pa = Pi, a shallow tubing leak could be expected with no
leak in the casing.
If a well has multiple injection strings, it would be normal
to suspect the upper injection string. By comparing Pa to
the Pi's with various strings shut in, or injecting, it
should be possible to determine which string has a leak. If
the pressures observed during injection indicate a shallow
tubing leak, the shutin pressures should indicate which
tubing string is at fault. If Pi reduces considerably when
injection ceases and Pa mirrors that pressure drop, then the
shut-in tubing string is at fault. A multi-pen chart re-
corder could give a good record of these test pressures.
Figure No. 3 indicates some pressure relationships in a
simple injection well.
-516-
-------
In order to think of all the possibilities for the particu-
lar well design, it may be desirable to construct a decision
tree for the wellhead test. The decision tree would force
more thought into the situation, provide a logical sequence
to the testing, and allow written results at each test
point. See Figure No. 4 for an example.
Another factor to be considered is a casing leak due to
outside sources of injection. In a new injection project
this could be likely if twin injection wells exist. In
older injection projects it could come from any pressured
zone uphole of the problem well's injection zone. This is
one instance where failure of uphole perforations could be a
strong possibility, even if they have been squeezed.
NORMAL EARTH TEMPERATURE
It is important to know what the normal temperature-depth
profile is for the area. The well files, log files, and
lease files, by now have hopefully been searched for tem-
perature information. Continuous temperature surveys are
not usually run during normal primary operations and none
may be available. In planning an injection project, the
additional cost to run two or three con-tinuous temperature
profiles in the field would be very small. If temperature
-517-
-------
profiles are available, make sure they were run under static
conditions with all temperatures stablized.
The alternative to a continuous temperature log is a multi-
point curve created from maximum recording thermometer
readings at various open hole log depths. Open hole logs
are often run at surface pipe depths and at total depth. In
some areas, other logs may be run if an intermediate casing
string and/or liner are set. More than one log run is
generally used at each logging depth. Each succeeding tool
run will usually have a higher temperature than the pre-
ceding tool run. The temperature from the last tool run in
the hole is still probably low but can be considered as
representative at that depth.
The resulting temperature graph will usually be a straight
line as shown in Figure No. 5. The temperature gradient
will probably be about 1° F. per 100 feet but can be much
higher or quite a bit lower. Extrapolation of the downhole
temperatures to the surface can safely be made. Surface
temperatures will vary with area considerably but will range
from 80° F. offshore Gulf of Mexico to 50° F. at the
Canadian border. Surface temperature is a constant tempera-
ture just below the earth's surface which is unaffected by
atmospheric temperature changes.
-518-
-------
The accuracy of the plotted log points can be improved by
creating a temperature buildup plot at each depth. This
plot, see Figure 6, is similar to a "Homer Plot" for pres-
sure buildup.
In deep wells and in areas having over pressures, abrupt
gradient changes can occur. See Figure No. 7- This knowl-
edge can greatly improve interpretation accuracy. Data from
several wells will offset possible bad data from a single
well.
INJECTION TEMPERATURE PROFILE
Water injection at surface temperatures will change the con-
tinuous temperature profile considerably. The cool water
temperatures absorb heat from the tubing, annular fluids,
casing, cement and formation until the gradient in the well
becomes very low. The amount of cooling varies with rate
and volume injected. The temperature decrease does not
occur very far out into the formation from the wellbore.
Injection into the reservoirs will have a much greater
effect, however, because the "cool" water actually pene-
trates the formation. Figure No. 8 shows the cooling
effects.
-519-
-------
Actual continuous temperature surveys run during normal
injection are rare. Very few companies run them on a regu-
lar basis for monitoring purposes. Each subsequent run
would be a little different if they were available, but the
difference should only be in the slope of the line and the
ultimate injection zone temperature. Any anomaly in the
curve is probably due to equipment changes or a rare forma-
tion change. Papers have been written on temperature
profile modeling. A lot of data on thermal conductivity of
wellbore equipment and fluids and all formations from
surface to TD would be necessary to obtain a good model.
This could provide a good base model if it is built
properly.
SEARCHING FOR ANOMALIES
Now that we know what to expect from the data collected at
this point, it is time to run a continuous temperature
survey and see what anomalous indications occur.
Very high rates of injection may prevent the temperature
tool from "seeing" anything but the temperature of the water
in the injection string. Small leaks in the tubing or
channels behind the pipe or up past the packer may go un-
noticed or provide only small anomalies. Most service
companies can run a differential temperature curve which
-520-
-------
accentuates the anomaly. It may be either a one or two
detector tool, but it measures the difference in temperature
between two depth points (2' to 10' apart) in the tubing.
It is not a different temperature survey but merely a dif-
ferent presentation displayed on an exagerated scale.
Figure No. 9 is a combination continuous temperature log and
a differential temperature survey.
If anomalies occur, it is time for interpretation. Many
papers have been written on this subject and it will not be
discussed here. If anomalies are not present, then why not?
just because we have pressure on the annulus, it doesn't
mean we have flow outside the injection string. Fluid flow
outside the intended fluid flow path is required to create
an anomalous temperature change.
CHEATING AN ANOMALY
Now is the time to create an anomaly or increase what looks
like an insignificant one. A significant increase in injec-
tion rate may create a prominent anomaly on the next temper^
ature survey. Another method of attack would be to open the
annulus and flow it to a tank for a while and run the
temperature survey again. A leak in the injection tubing, a
packer not set, communication behind the casing back into
the annulus, a faulty circulating valve; all of these will
-521-
-------
now show a temperature anomaly. The amount of temperature
change will depend on the amount of flow coming out of the
annulus. A large flow volume will require a shorter flow
period to obtain a significant temperature change. A small
flow volume may require several hours of flow period to
obtain the same amount of change.
Very shallow wells with low injection rates will have little
temperature change from surface to TD. The temperature
profile may change more from night to day with injection
water temperatures changes than it does from top to bottom
of the well. The best solution here may be to get a hot
oiler on .pa location and heat a tank of injection water to
120° F. A significant difference will now occur from top to
bottom of the well.
The biggest change of all after running the first tempera-
ture log during injection may be to shut the well in. A
series of temperature surveys can now be run on some
schedule such as 1, 3, 5 - - - hours after shutin. This is
called a temperature decay log. See Figure No. 10. It
illustrates the earth's return to temperature equilibrium.
The areas with the shallowest cooling (all areas where water
did not enter the rock) will return to the earth's normal
temperature gradient first. This is the best type of tem-
perature log to find leaks in the casing.
-522-
-------
A lot of papers have been written on mathematical interpre-
tation of temperature surveys, but most people consider
their interpretation a work of art. The radioactive tracer
survey can be run with the temperature survey and is a good
tool for finding small leaks. It is also a good tool to
confirm confinement of injection fluid to the injection
zone. The temperature survey may solve the problem, but it
would be advisable to have radioactivity tracer logging
equipment on location.
Development of a good data base and proper planning of the
temperature survey will go a long way in assuring successful
interpretation of the well's problem.
REFERENCES
1. Cocanower, R. D., Morris, B. P., and Dillingham, M.:
"Computerized Temperature Decay - An Asset to Tempera-
ture Logging," J^ Pet^ Tech^ (Aug. 1969) 933-9^1.
2. Loeb, J., and Poupon, A.,: "Temperature Logs in Produc-
tion and Injection Wells." Twenty-Seventh Meeting of
the European Association of Exploration Geophysicists
in Madrid - May 5-6-7, 1965.
-523-
-------
3- Fagley, John, Folger, H. Scott, Davenport, C. Brent,
and Millhone, Ralph S.: "An Improved Simulation for
Interpreting Temperature Logs in Injection Wells."
Paper SPE-AIME 10081, 56th Annual Fall Meeting, San
Antonio, Texas, Oct. 5-7, 1981.
4. Western Company Technical Leaflet. No date.
5. Joslyn, C. D., and Chilton, L. F.: "Analysis of Well
Problems Through the Use of Differential Temperature
Logs." API Preprint Paper No. 875-24-H, for presenta-
tion at Spring Meeting, Rocky Mountain District Divi-
sion of Production, Denver, Colorado, April 27-29,
1970.
-524-
-------
WELLBORE DIAGRAM
ZERO
GRD. ELEV.
7" CASING
2 3/8" TUBING
to
Ul
I
15' ALF
936'
TENSION PACKER @ 6357-59' _E
TOP/MILO SAND
BASE/MILO SAND
PBTD
7" CASING SHOE
T.D.
E-6250'
^6350'
PRESENT WELL CONDITION:
INJECTING 430 BWPD
1250 PSI
CONDUCTOR PIPE 70'
SURFACE CASING 10 3/4" H-40 @ 2900'
CEMENT CIRCULATED
CASING 7" J-55 23#/FT. #/FT.
TOP OF CEMENT 3600'
ANNULAR FLUID - 80,000 PPM NaCI
MINIMUM ID - PACKER 1.9"
VBOTTOM OF TUBING
L_OPEN ENDED @ 6369'
i?-PERFS
l~6400'
^6423'
~^_6429'
- 6433'
FIGURE NO. 1
-------
ZERO
GRD. ELEV.
15' ALF
936'
ON
I
TOP/CATOOSA SAND —
BASE/CATOOSA SAND-
SLIDING SLEEVE
DUAL HYDRAULIC PACKER
TOP/JONES —
BASE/JONES—
PERMANENT PACKER
TOP/MILO SAND —
PERFS
BASE/MILO SAND
PBTD
7" CASING SHOE
TOTAL DEPTH
^-6300'
^6400'
WELL BORE DIAGRAM
PRESENT WELL CONDITION:
INJECTION
MILO 430 BWPD
1250 PSI
JONES 300
1200 PSI
CATOOSA TWIN WELL INJECTION
200 BWPD
1400 PSI
PERFS SQUEEZED
NORRIS FORMATION
5310-31 IS BEING
FLOODED IN THIS
PART OF THE FIELD.
CASING 7" J-55 23#/FT.
TOP OF CEMENT 3600'
ANNULAR FLUID - 80,000 PPM NaCI
MINIMUM ID
MILO TUBING 1.94 IN.
JONES TUBING 1.875 IN.
FIGURE NO. 2
-------
PRESSURE BALANCE
INJECTION
PW = Pi -
x
Pf -
h p
>
x
H
a
FLUID LEVEL
Y* Prr —
r ' I u
PR = Pj - 1
Pi =
Pa =
PTC =
PW =
Pf =
PH -
PR =
Pn =
INJECTION PRESSURE
ANNULUS PRESSURE
TUBING CASING
BOTTOM HOLE PRESSURE
BOTTOM HOLE PRESSURE
FRICTION PRESSURE
HYDROSTATIC PRESSURE
RESERVOIR PRESSURE
PRESSURE LOSS
ACROSS PERFORATIONS
Pa + PH
FIGURE NO. 3
-------
DECISION TREE CLOSE ANNU.
TAKE WTR. ENDTEST
SAMPLES NO CHANGE IN
CK. INJ. RATE INJ. RATE & PRESS.
RECORD TIME./ CLOSE-IN ANNU.
& PRESS. ^ P END TEST
CONT. RATE INCR. AND/OR ANNU.
RECORD L°W PRESS. DECR. FLOW DIES
PRESS
FLOW X \NJ.
S EXISTS ""•"•"• \ ANNU. FLOW
t° m MONITOR ^/ \ CONT.
°ANNU PRESS.X N,0 , V CLOSE ANNU.
NO FLOW n END D RESUME INJ.
PRESS. \ r TEST END TEST
V. RECORD & \ NO PRESS.
END TEST \ BUILD-UP
SHUT-IN y ^END TEST
ANNU.
PRESS TUBING PRESS. DROP &
RETURNS ANNU. PRESS. DROP
TBG. PRESS DROP
& ANNU. PRESS. DROP
END TEST
FIGURE NO. 4
-------
i
Ul
VD
I
ESTIMATION OF FORMATION TEMPERATURE GRADIENT
FROM BHT DATA
240
18,000
FIGURE WO. 5
-------
220
210
o
t 200
190
180
0.4
TEMPERATURE BUILDUP
0.5
0.6
0.7
0.8
At
t + At
0.9
1.0
FIGURE NO. 6
-------
01
Co
BHT DATA FROM RESISTIVITY LOGS
CUSTER CO., OKLAHOMA
o
o
a.
g
6
8 -
10
12
16 •
18 •
20
SNIDER NO. 1-A
X
TOP
OVERPRESSURE
WOLFCAMPIAN
BROWN DOLOMITE
U. TONKAWA-
•
COTTAGE GROVE •
CHECKER BOARD LS.'
U. RED FORK
COUNTY LINE LS.
L. TONKAWA
HOGSHOOTER
DEESE
L. RED FORK
FORMATION TOPS AND
OVERPRESSURE DATA
TAKEN FROM SNIDER NO. 1-A
"A"
A __
ni-> i • • r><*\O c ^^^
PRIMROSE —
v
V
CUNN|NGHAM
100° 120° 140° 160° 180° 200° 220° 240° 260° 280° 300° 320°
TEMPERATURE °F
FIGURE WO. 7
-------
INJECTION PROFILES
ARE RATE AND VOLUME
DEPENDENT
(M
CO
U.
o
o
o
0.
UJ
Q
90 100 120
TEMPERATURE
140
160
FIGURE NO. 8
-------
CONTINUOUS AND DIFFERENTIAL
TEMPERATURE CURVES
TUBING LEAK
I
Ul
COLLAR LOG
TUBING
10,000
10,050
10,100
10,150
10,200
GRADIENT ONE DEG.
PER INCH
DIFFERENTIAL
TUBING PRESSURE
7200 PSI
FLUID LEVEL
IN ANNULUS
ABSOLUTE TEMP.
DIFFERENTIAL CURVE
TUBING LEAK (10,118')
T.D. @ 21,700'
FIGURE NO. 9
-------
90° 91° 92° 93°
U)
TEMPERATURE DECAY LOG
DETERMINES FLUID MOVEMENT
IN THE FORMATION
R-1 2:00 P.M. -IN J. RATE -400 BPD
JV2 __ 4jO_0_P_J\/L. ~ 30 MINUTE SHUT IN
R3 6:00 P.M/
R-52:00 A.M
R-7
94° 95° 96° 97° 98° 99°
4200-
FIGURE 10
-------
Mobil's Experience in Applying for a Waiver from the Surface Cementing
Requirements for Rule Authorized Class II Enhanced Recovery Wells in the
Springfield North Unit.
N. H. Ginest, Sr. Regulatory Engineer
Mobil Oil Corporation
J. V. lerubino, Operations Engineer
Mobil Oil Corporation
ABSTRACT
On July 20, 1986 the Region V office of the EPA issued a casing and cementing
policy for all Class II injection wells to provide guidance to its UIC permit
writers. Mobil's rule authorized Class II enhanced recovery wells in the
Springfield North Unit (SNU) do not meet the surface cementing requirements
set out in the subsequently promulgated 40 CFR § 146.22 (b) and § 147.754 (b)
and, under the aforementioned Region V casing and cementing policy, Mobil would
be required to squeeze cement to isolate USDW's. The high costs and risks
associated with squeeze cementing the 35 to 40 year old injection wells in the
SNU would force Mobil to abandon this waterflood project. In order to
demonstrate that USDW's are being adequately protected under existing operating
conditions, evidence was collected which included cement bond logs, radioactive
tracer surveys, cyclic activation logs and pressure tests. This evidence was
submitted in a waiver request to satisfy the burden of proof for protection of
USDW's which is placed on the operator. It should be noted that the costs of
gathering evidence to illustrate protection of USDW's can easily reach an
amount which could make a mature waterflood uneconomic. Should the economics
of compliance dictate that Mobil abandon its SNU, approximately 32,000 barrels
-535-
-------
of recoverable oil will be left in place. Not only will foregone production
in cases such as this be detrimental, but rising compliance costs will make
many secondary recovery projects much less economically attractive and fewer
operators will be willing to make the increased investment necessary to
recover secondary reserves.
-536-
-------
INTRODUCTION
Examination of a typical injection wellbore configuration for existing or rule
authorized Class II wells, as shown in Figure No. 1, reveals two areas where
the cement and/or casing may not meet the EPA's construction requirements for
Class II injection wells. The areas in question are:
1. the area through and above the injection zone; there may not be
sufficient cement to fill the casing/wellbore annulus to a point 250
feet above the injection zone, and
2. the area from the surface to the base of the USDW's; surface casing
and cementation may not be sufficient to isolate all USDW's.
In existing fields which were developed prior to enactment of the Safe
Drinking Water Act, these conditions will generally exist on a field or
area-wide basis since, during development, drilling and completion practices
would have remained relatively uniform.
By monitoring and utilizing available technology, operators can show on a
case-by-case basis that the USDW's are being adequately protected from
contamination in a specific area under existing operating conditions. This
paper summarizes Mobil's attempt to provide the EPA with the data and
information required to illustrate that USDW's are being adequately protected
in the Springfield North Unit (SNU).
-537-
-------
FIELD HISTORY
The Springfield North Unit (SNU) is located in Posey County, Indiana, and the
field was discovered with the drilling of the Highman Heirs No. 1 (renamed as
the SNU No. 30) on June 4, 1946. The unit contains approximately 970 acres
upon which there are currently 47 wellbores capable of production or
injection. The producing reservoir is the Palestine Sandstone.
SNU waterflooding operations began in February of 1963 and there are currently
13 wells permitted by rule for injection. A field map is shown in Figure
No. 2.
INJECTION WELLBORE CONFIGURATION IN THE SNU
Figure No. 3 illustrates the typical configuration of an injection well in the
SNU. The stratigraphic succession of geologic formations with their
approximate thicknesses is set out on the left side of the wellbore. The
dashed line represents surface casing which was run to various depths up to
125' on the injection wells in the SNU. In all cases where surface casing was
run, it was cemented to the surface. Although there is little or no
information available as to how deep sands which could be defined as USDW's
occur, it is generally accepted that the USDW's run to at least the top of the
West Franklin limestone. The injection wells in the SNU do not meet the EPA's
surface cementing requirements because surface casing setting depths are not
sufficient to cover all the sands classified as USDW's. It is important to
note that when these wells were permitted by rule and converted to injection
-538-
-------
service the definition of a USDW was quite different from that contained in
the SDWA.
During initial completion of the injection wells, cement was circulated up
from the base of the long string casing into the annulus between the wall of
the hole and the outside of the casing. Cement tops in this casing/hole
annulus range from approximately 366 feet to 1271 feet above the top of the
injection interval. Cement tops were calculated using 80% of the volume
circulated in the casing/hole annulus.
UIC REGULATIONS PERTAINING TO THE SITUATION IN THE SNU
Pursuant to §144.22(b), existing Class II enhanced recovery wells in Indiana
must comply with casing and cementing requirements by June 25, 1987. These
casing and cementing requirements are set out specifically for the State of
Indiana in §147.754(b). Mobil is requesting a waiver for the 13 existing Class
II enhanced recovery wells in the Springfield North Unit from the requirements
in §147.754(b)(l)(i) and (ii) which state that the USDW's must be protected by:
"Cementing surface casing by recirculating the cement to the surface from a
point 50 feet below the lowermost USDW; or isolating all USDW's by placing
cement between the outermost casing and the wellbore."
Under provisions in §144.16(a), the Director may authorize less stringent
requirements for construction than those required in the previously mentioned
sections. In Mobil's SNU cementing waiver request to the EPA, data and
evidence were compiled and submitted to justify variances from required
surface cementing standards.
-539-
-------
EVIDENCE GATHERED TO ILLUSTRATE THAT USDW'S ARE BEING ADEQUATELY PROTECTED
FROM INJECTED FLUIDS
A. Adequate cementation above the injection interval in the casing/
hole annulus. All thirteen (13) injection wells have approximately
366 feet to 1271 feet of cement coverage above the top of the
injection interval. All injection wells have more than 250 feet of
cement coverage which is considered sufficient as required in
§147.754(b)(2).
Cement bond logs were run on SNU Wells No. 2 and 14 and confirmed the
calculated values for cement tops.
B. Radioactive tracer surveys. Radioactive tracer surveys provide an
effective means for locating and evaluating leaks in the casing,
tubing and/or packer and channeling behind the casing. The primary
advantage of a radioactive tracer survey is that it is run during
injection operations and can therefore provide a clear picture of
what is taking place in the well under actual operating conditions.
Radioactive tracer surveys were run in Wells No. 2 and 14. Both logs
indicated injected water to be entering the Palestine zone with no
evidence of upward channeling behind the casing.
-540-
-------
Noise and temperature logs can also be run separately or in
combination to detect tubing and/or casing leaks and also fluid
channeling in the cement sheath behind the casing. Neither of these
logs were run in the SNU injection wells because they were all
cemented adequately above the injection interval.
C. Mechanical integrity pressure testing. Pressure tests, which were
witnessed by EPA field inspectors, were conducted on all the
injection wells in the SNU. This testing demonstrated the tubing,
casing, and packer in the injection wells were mechanically sound.
D. Wellhead injection pressures below the formation fracture pressure.
All injection wells are operated with wellhead injection pressures
below the calculated fracturing pressure. These calculated fractur-
ing pressures are based on a frac gradient of 0.8 psi/ft and an
injection fluid specific gravity of 1.015.
It should be noted that the formation fracture gradient is actually
approximately 0.98 psi/ft as is evidenced by the instantaneous
shut-in pressure (ISIP) data from a stimulation treatment on the No.
14 well. Mobil has recently submitted a request for increased
allowable injection pressure in this field supported by copies of
actual treatment reports which show treating pressure and ISIP data.
-541-
-------
E. Overlying formations prevent upward migration of injected water.
The Palestine Sandstone is overlain by several limestone and shale
beds which consist of several impermeable layers and serve to prevent
upward migration of the water injected into the Palestine zone.
These beds are illustrated in the stratigraphic section on Figure
No. 3.
EVIDENCE GATHERED TO ILLUSTRATE THAT INTERMINGLING OF FLUIDS
IN USDW'S IS NOT OCCURRING
A. Monitoring of long string casing/surface casing annulus shows no
pressure changes or fluid movement. Mobil currently monitors the
aforementioned casing/casing annulus on a weekly basis and has found
no indication of a pressure change or fluid flow. This is a critical
monitoring point which allows an operator to sense any changes or
disturbances in the freshwater zones not covered by surface casing.
There are two likely situations which could exist in the long string
casing/hole annulus which are as follows:
-542-
-------
1. Since the casing was run in the open hole while it was filled
with drilling mud, the casing/hole annulus will be filled with
the fresh water, solid-based drilling fluid. The drilling fluid
is more dense than fresh water and will exert a hydrostatic
pressure due to the column of drilling fluid that will tend to
keep formation fluids out of the casing/hole annulus and
discourage intermingling of formation fluids.
2. If after standing in the casing/hole annulus for several years
the drilling fluids were to dehydrate, it is unlikely that a
void would exist adjacent to the production casing. The
overburden stresses present in the geologic column will tend to
force the formation to compact around the pipe. This would also
tend to prevent any fluid movement from occurring adjacent to
the long string casing.
B. Pulsed Neutron Logs show no evidence of fluid movement behind the
casing. Arnold and Paap described a water-flow monitoring system
(referred to as the Cyclic Activation or CA log) based on a nuclear
activation technique in which water is irradiated with neutrons emitted
by a source in the logging sonde. These neutrons interact with
Arnold, D. M. and Paap, H. J., Quantitative Monitoring of Water Flow
Behind and in Wellbore Casing, JPT, January 1979.
-543-
-------
oxygen nuclei in the water to produce the radioactive isotope
nitrogen-16. N decays with a half-life of 7.13 seconds and
emits gamma radiation during decay. If water flow is occurring
outside the casing, its velocity can then be computed from the
energy and intensity response of the two gamma ray detectors
mounted in the logging sonde. Basically, the difference in
gamma ray count rates (above the normal background gamma
emission) of the two detectors is used to calculate a linear
fluid velocity.
This water flow detection system is similar to various radio-
active tracer techniques, but is unique in the sense that the
tracer, N is "manufactured" in the water. This eliminates the
need to perforate the casing and to inject tracer material from
an external source.
The CA log was run on SNU Well Nos. 14 and 17. The CA log
analysis obtained from these two logging runs indicated that no
fluid flow was occurring outside the casing.
ECONOMIC IMPACT OF COMPLIANCE COSTS
There are several different cases or scenarios which could be used to attain
compliance for the injection wells in the SNU. For the following cases, costs
and the resulting unit economics are listed which illustrate the associated
-544-
-------
economic repercussions. The economic indicator referred to in each of the
following cases is payout. Payout is defined as the time required to recoup
all investment costs. Economic assumptions include an initial unit production
rate of 35 BOPD and an annual decline rate of 11%.
CASE I
Case I is an example of the work completed to date in the SNU. It entails
running two (2) cyclic activation (CA) logs on a representative sampling basis
at a cost of $7,000 ($3,500/well). Economic runs indicate a payout on this
$7,000 investment would occur in approximately 0.4 years. The SNU can support
the compliance-related investment associated with Case I.
CASE II
Case II assumes that the EPA would not accept Ipgs run in sample wells for an
area-wide waiver and would require Mobil to run CA logs in all existing
injection wells to obtain a waiver for the entire unit. Running CA logs on
all thirteen (13) injection wells would require a total investment of $45,500.
Economic runs indicate a payout on this investment never occurs. Mobil could
not afford to run CA logs on all SNU injection wells and still maintain a
profitable operation.
-545-
-------
CASE III
Case III is the scenario most likely to occur should Mobil be required to
squeeze cement to isolate the USDW's. In order to realistically estimate
costs involved in squeeze cementing the mature injection wells in the SNU,
costs were gathered to perform the following work:
1) Four (4) of the thirteen (13) squeeze jobs would proceed with no
problems and cement returns to the surface would be obtained after
the first squeeze attempt,
2) Five (5) of the injection wells would require a cement bond log after
the initial cement squeeze and have to be reperforated and resqueezed
two (2) more times before isolation of USDW's could be obtained and
3) Four (4) of the injectors would require not only the work and expense
incurred in 1) and 2) above, but also would require new production
casing strings (liners) due to casing failures during workover/
squeezing operations. The new liners would have to be run and
cemented if the original long string casing collapsed. The risk of
casing collapse is very high when exerting the high pump pressures
required to break circulation to the surface.
-546-
-------
The investment required to complete the work described in the Case III scenario
is $286,000. Economic runs indicate that a payout of this investment never
occurs.
The three economic cases are summarized in Table 1. Table 1 illustrates that
the only case under which Mobil can economically justify the costs associated
with compliance is Case I. A waiver, based on a showing of adequate
protection of USDW's under current operating conditions, would have to
include running CA logs on a representative sampling basis to illustrate
unit-wide compliance. It should also be npted that Mobil or any other
company, for that matter, will not operate a project at a loss with income to
be made up in another producing area. Each project must stand on its own and
be economical or it will be abandoned.
-547-
-------
SUMMARY
Commercial oil production in the SNU is dependent entirely upon water
injection. Should the economics of compliance dictate that Mobil abandon the
waterflood in the SNU, approximately 32,000 barrels of recoverable oil will be
left in place. Not only would foregone production in cases like Mobil's SNU be
detrimental, but the rising costs of compliance could have a major effect on
oil production nationwide. As compliance costs rise, secondary recovery
projects will become significantly less economically attractive. Fewer
operators will be willing to make the increased investment necessary to recover
the additional reserves left in place after primary recovery has been
completed.
-548-
-------
CONCLUSIONS
1. It is important that the EPA and state regulatory agencies approve the
use of available technology to gather data which can show injection
operations are not contaminating USDW's. This available technology
includes logging techniques.
2. It is equally important that the EPA respond to Mobil's waiver request in
a timely manner. All enhanced recovery wells in Indiana must comply with
Class II casing and cementing requirements by June 25, 1987. Not only
would a timely response allow Mobil sufficient time to make plans for the
future operation of the SNU, but it would also provide guidance to other
operators that will be required to attain compliance with construction
requirements.
3. As is the case in the SNU, squeeze cementing to isolate USDW's will not
be economically viable in many of the existing mature waterfloods in the
United States.
-549-
-------
TABLE 1
SUMMARY OF THREE COMPLIANCE CASES
COMPLIANCE
RELATED
WORK
INVESTMENT
REQUIRED
PAYOUT
(YRS.)
CASE I
2 CA LOGS RAN ON
REPRESENTATIVE
BASIS
$ 7,000
0.4
CASE II
CA LOGS RAN ON
ALL INJECTORS
IN UNIT
$ 45,500
DOES NOT
PAYOUT
CASE III
SQUEEZE USDW'S
IN ALL INJECTORS
IN UNIT
$286,000
DOES NOT
PAYOUT
-550-
-------
Noel Ginest is a Senior Regulatory Engineer with Mobil Oil Corporation in
Denver, Colorado. He received a BS degree (1981) in Petroleum Engineering
from the Colorado School of Mines. Mr. Ginest was employed by Mobil in Lake
Charles, Louisiana, in 1981 as an Operations Engineer and worked in Mobil's
Gulf Coast Operations until being transferred into the Environmental and
Regulatory Affairs Department in 1985. He is a member of the Society of
Petroleum Engineers.
James lerubino is an Operations Engineer with Mobil Oil Corporation in
Crossville, Illinois. He received a BS degree (1982) in Geology from Rider
College, where he was published by the GSA and various other journals
following his research on sedimentation patterns on the continental shelf.
Mr. lerubino also holds an MS degree (1985) in Petroleum Engineering from the
Colorado School of Mines. He is a member of the Society of Petroleum
Engineers.
-551-
-------
INJECTION
PRESSURE GAUGE
ANNULUS
PRESSURE GAUGE
SURFACE CASING/
INJECTION CASING
ANNULUS
PRESSURE GAUGE
POTABLE
WATER
NON-POTABLE
WATER
UNDIFFERENT1ATED
ROCKS
CONFINING BED
INJECTION ZONE
INJECTED
LIQUID
>WELLHEAD
1^ ANNULAR
ACCESS
SURFACE CASING
CEMENT
INJECTION TUBING
LONG STRING CASING
ANNULAR FLUID
CEMENT
PACKER
PERFORATIONS
FI0URE 1. TYPICAL INJECTION WELLBORE CONFIGURATION FOR
RULE AUTHORIZED CLASS II WELLS
-552-
«L-P«D-«7-«l4.23.a
-------
Rl 4W
13
6
A
5
20
f§ e « »|§ *
7 16 15 14 13
X
RI3W
18
19
20 21
24
22 23
| 35 34 33 32
36 37
30
38
X
5
24
25
29
28
*39 40
<§> «
42 41
49 48
26
27
46 47
43
44
Jj
LEGEND:
• PRODUCING OIL WELL
^ ABANDONED OIL
-6- DRY HOLE
® INJECTION WELL
j§f ABANDONED INJECTION WELL
FIGURE 2. SPRIK«FIII.B NORTH UNiT FIELD
Mobil Oil Corporation
DENVER AFFILIATE
DATA MAP
SPRINGFIELD NORTH UNIT
POSEY CO., INDIANA
DRAWN C . ARCHER
CHECKED CHECKED
SCALE NONE
DATE 03/02/37
DWG NO.
IN-PRG-DM-
-553-
-------
APPtOX.
THICKNESS
aoo-
4V-TO'
SURFACE
•AND, (HALES *
INTEHSEDDCD
(AMD * *HAUS
100- 10V CLOH « NCOIIIA (MAUI
•-12'
e-tas* or SUMACI
CASINO (CIBINT
CUICULATU TO
SUWAd V SUWAd
CASINO MISMT)
TVSINO- CASINO
ANNUtVS FILLED WITH
COOMSION INNISITID
a- 1/i CIMNT. LINED
MMCTIOII TVMNO
CEMENT TOPS SOS-taTl'
ASOVE THE INJECTION
INTIBVAL
•mi niie w * rrrnu.
m >rtan« •IIIJIM acufci,
PACHEB SET AT AN
AVEBAOE DEPTH OF I MO1
PDODUCTION CASINO SET
AT AN AVEBAOE DEPTH
OF isso*
FIOURE X. TYPICAL INJECTION WCtLIORI CONFIQURATION IN
THE SPRINQFIELD NORTH UNIT
•C-U-A-Gft-Ul i 3( .2
-554-
-------
TABLE 1
SUMMARY OF THREE COMPLIANCE CASES
COMPLIANCE
RELATED INVESTMENT PAYOUT
WORK REQUIRED (YRS.)
CASE I 2 CA LOGS RAN ON $ 7,000 0.4
REPRESENTATIVE
BASIS
CASE II CA LOGS RAN ON $ 45,500 DOES NOT
ALL INJECTORS PAYOUT
IN UNIT
CASE III SQUEEZE USDW'S $286,000 DOES NOT
IN ALL INJECTORS PAYOUT
IN UNIT
-555-
-------
A METHOD TO CONVERT MULTIPLE-SHOT SECTION OPENHOLE
COMPLETIONS INTO CASED-HOLE COMPLETIONS WITH ZONAL ISOLATION
Authors
C.D.K. Darr and E.K. Brown
Conoco, Inc. & J.R. Murphey, Halliburton Services
Presented
at
THE UNDERGROUND INJECTION PRACTICES COUNCIL\EPA
INTERNATIONAL SYMPOSIUM ON SUBSURFACE INJECTION
OF OILFIELD BRINES
THE ROYAL SONESTA HOTEL, NEW ORLEANS, LOUISIANA
MAY 4-6, 1987
-556-
-------
Introduction
The "Puddle-Pack" completion process was developed to convert old
wellbores, not originally designed for fluid injection, into usable
wellbores for fluid injection. This process was applied to Conoco's MCA
Unit to prevent fluid loss and provide zonal isolation in injection
wells. Furthermore, this process has been used on producing wells to
increase productivity by selective stimulation. The MCA Unit, located in
southeastern New Mexico, is currently under waterflood and has tertiary
oil recovery potential (see Figure 3).
Production within the unit is from the Grayburg Sixth sandstone and
the San Andres Upper Seventh, Upper Ninth and Lower Ninth Massive dolomite
formations. Formation depths range from 3650 to 4050 feet (see sample log
Figure 2). Most of the 366 active wells are over 30 years old and were
open hole completed over a 300 foot interval. Most wells have at least
two and often three shot sections which generally exceed 20" in diameter
(Figure 1). These factors have combined to make remedial attempts very
difficult.
The feasibility of implementing a carbon dioxide miscible flood
within the unit is presently being evaluated. A major operational,
environmental, and economic concern was whether old well problems could be
repaired to maximize C02 efficiency during C02 and post-flush
injection. Wells drilled four decades ago without any thought of use for
injection could not be expected to prevent injection fluid loss and
provide zonal isolation. These conditions are not tolerable during CC>2
injection because only a small loss of CC>2 would significantly reduce
the profitability of the project; thus zonal isolation would be required.
A liner with a simple cement job would not satisfy this requirement since
it would be impossible to perforate through the thick cement sheaths in
-557-
-------
Page No . 2
the shot hole sections. Therefore, a different method has to be developed
to convert the shot hole wells to cased hole completions with zonal
isolation. Otherwise replacement wells for all the existing open hole
injection wells would have to be drilled and the old wells plugged and
abandoned. An estimated 14 million dollars could be saved if the current
injectors would not have to be replaced with new wells.
Discussion
The problem of injection fluid loss and zonal communication in old
wells (40 years old or greater) has been recognized for many years.
Injection fluid control has been attempted by running and cementing liners
across the shot open hole section. These attempts failed because
conductivity with the formation could not be re-established after
perforating the liner and acidizing in the shot open hole section. This
is due to the fact that the cement, after completely filling the shot hole
could not be totally penetrated. Other attempts at casing the shot open
holes failed because zonal isolation was not achieved. The "Puddle-Pack"
was designed to solve these problems.
Criteria for Success of the
"Puddle-Pack" Completion
The four requirements listed below had to be met for the
"Puddle-Pack" process to be considered successful.
-558-
-------
Page No. 2
the shot hole sections. Therefore, a different method has to be developed
to convert the shot hole wells to cased hole completions with zonal
isolation. Otherwise replacement wells for all the existing open hole
injection wells would have to be drilled and the old wells plugged and
abandoned. An estimated 14 million dollars could be saved if the current
injectors would not have to be replaced with new wells.
Discussion
The problem of injection fluid loss and zonal communication in old
wells (40 years old or greater) has been recognized for many years.
Injection fluid control has been attempted by running and cementing liners
across the shot open hole section. These attempts failed because
conductivity with the formation could not be re-established after
perforating the liner and acidizing in the shot open hole section. This
is due to the fact that the cement, after completely filling the shot hole
could not be totally penetrated. Other attempts at casing the shot open
holes failed because zonal isolation was not achieved. The "Puddle-Pack"
was designed to solve these problems.
Criteria for Success of the
"Puddle-Pack" Completion
The four requirements listed below had to be met for the
"Puddle-Pack" process to be considered successful.
-559-
-------
Page No. 3
1) Zonal isolation. Zonal isolation aids in the stimulation of
individual zones and the profile control of injection fluids.
2) No fluid loss to non-pay intervals. Injection fluid loss to non-pay
intervals can cause collapsed casing and in extreme cases surface
waterflows.
3) No loss in injectivity. Based upon a bbl/psi/NEP criterion,
injectivity after the "Puddle-Pack" process should be greater than or
equal to (within 15%) the injectivity before the "Puddle-Pack".
4) Capability of running an injection profile log. To properly manage a
waterflood or tertiary recovery project the injection profile must be
regularly monitored. Therefore, it is important that an interpretable
injection profile log can be run.
Requirements of the Resin Coated Fill Material
Resin coated fill material used in the "Puddle-Pack" process must
have the properties of (1) permeability, (2) strength, (3) chemical
inertness to formation fluid and injected fluid, and (4) feasible cost.
Additional desirable features are (1) inertness to projected treating
chemicals and (2) ease of handling.
PERMEABILITY. A reasonably wide range of permeability is acceptable
provided (1) communication to the formation is not lost or restricted and
(2) fluid loss during cementing does not result in excessive cement
dehydration. Vugs and mud channels in the fill material are unacceptable,
therefore, systems requiring long reaction or settling times and materials
of widely varied densities were not considered.
-560-
-------
Page No. 4
STRENGTH. An upper strength limit of 5000-6000 psi is suggested for
maximum penetration rates based on experience with drilling cements (see
Table 1). Although the compressive strength of the native formation is
approximately 9700 psi, fill material need only be strong enough to hold
its form while being drilled out.
CHEMICAL INERTNESS. Resin coated fill material must be chemically
inert to formation crude oil and formation brine as well as the
combination of injected C02 and brine. This combination forms carbonic
acid under downhole temperature and pressure. Inertness to common
workover fluids is also desirable. Fluids used on the MCA Unit project
include hydrochloric acid, mutual solvent flushes, and aromatic solvent
flushes.
EASE OF APPLICATION. Ideally, synthetic formation slurries can be
placed with conventional bulk mixing equipment and should require no
unusual steps in either preparation or cleanout. Rapid cure time and easy
drillout are desirable qualities.
Selection and Optimization of Fill Material
Two resin systems were evaluated to determine the optimum fill
material for the project. The first system has been widely used in resin
cementing of disposal wells where highly corrosive fluids could attack a
conventional cement-^. The second system has been widely used for
consolidated pack sand control jobs on the Gulf Coast and West Coast.
Both systems proved to have good chemical resistance, but the Gulf
Coast system was chosen for its better retained permeability.
-561-
-------
Page No. 5
With the system selected, optimization of permeability was approached
in a manner similar to that used in gravel packing; i.e., intermixing of
fine sand with the pack sand to achieve the desired porosity and
permeability1. Table 2 demonstrates this effect when 70-170 mesh sand
is mixed with 10-20 rounded pack sand and 20-40 mesh angular sand.
Eventually, the method selected was to mix small matrix sand (70-170
mesh) and graded silica flour. This mixture yielded the desired
permeability, a more uniform graduation, and minimal separation by
sedimentation.
Since any wellbore is an excellent sedimentation column, a mixture
with a wide range in particle size would have the tendency to settle out.
Introduction of a gelled carrier fluid moderately reduced this tendency
but did not eliminate it. Experimentation with laboratory samples showed
that the addition of the selected resin greatly reduced the tendency to
form sediment. Microscopic examination showed that the fine grained
silica flour clustered about on the larger grains of resin coated sand.
Further experimentation proved that the order of addition of ingredients
influenced the character of the mix.
The sequence used is as follows:
1. Add the large sand to the gelled carrier fluid.
2. Then add the resin, which coats the large sand grains and leaves
little resin in the carrier fluid phase.
3. Then add the silica flour. The silica flour particles are not
initially coated, but rather attach themselves to the larger
grains, and are then coated with resin.
-562-
-------
Page No. 6
Contrary to the initial assumption that the fine silica flour would
substantially increase the requirement for resin due to increased surface
area of the flour over the 10-20 mesh sand used in control experiments,
the increase was only on the order of 20%. Final strength and
permeability of the resin coated gravel pack depend upon optimization of
the resin volume.
A minimal resin level is desirable to force the silica flour to
cluster on the sand grains and since resin is the single most expensive
ingredient of the fill, keeping the minimum level is even more desirable.
Gelling agent level was the final variable in the fill formula. The
proper gelling agent in a fill of this type must be sufficient to suspend
the material, but provide no additional restriction to injection.
Functions of the gelling agent are (1) to suspend the sand and silica
flour in bulk equipment before pumping, and (2) suspend the slurry
downhole while the slurry is being pressure packed.
The final product showed uniform permeability from 1.8 to 8 darcies
with most falling around 6 darcies (see Table 3).
Bulk Mixing Tests
Large scale mixing tests were performed to determine pumping
characteristics of the 14 Ib/gal optimized slurry as developed in the
laboratory. The procedure was as follows:
1. A clean 12 bbl conventional stirring blender was loaded with 3
bbls water and gel. Ambient temperature was 35° F.
-563-
-------
Page No. 7
2. An addition of 1 cu ft 70-170 mesh sand was made.
3. The resin (less than 1 gal/cu ft sand) was then added.
4. Then 1 sk of silica flour (1 cu ft) was added.
5. The slurry was then circulated for 1 hour with a standard
centrifugal pump and triplex pump. The slurry properties
resembled a cement slurry.
6. The sample was withdrawn and breaker was added to the sample.
Break time was slow due to cold temperature.
7. To stimulate bottom hole temperatures, samples similar to No. 6
were cured in an 80° F water bath which resulted in a 1200 psi
compressive strength within 24 hours.
The Remedial Procedure
MCA Unit No. 61 was identified as a problem well for injection fluid
loss control in the MCA Unit. Past injection profile logs indicated
abnormal injection fluid distribution and possible fluid loss to
stratigraphically higher horizons. After reviewing several wellbores in
the MCA Unit, MCA Unit No. 61 was identified as a worst case condition due
to the three large shot holes, hole sloughing, and the large volume of
resin coated gravel required (see Figure 1). Thus, it was felt that if
MCA Unit No. 61 could be successfully repaired using the "Puddle-Pack"
process, it would be applicable to the other shot open hole completions in
the MCA Unit.
The following steps were taken to convert MCA Unit No. 61 to a cased
hole completion with zonal isolation:
-564-
-------
Page No. 8
1) To prepare the wellbore for squeeze cementing the production casing
shoe, all injection equipment was pulled from the well and the shot
open hole was plugged back to within 20' (35701) of the production
casing shoe with crushed oyster shells. The volume of crushed oyster
shells required to fill the open hole was recorded to verify the
volume of resin coated gravel required. Crushed oyster shells were
used because they were tested to be 96% acid soluable. Therefore,
when the shells were drilled out, the oyster shells remaining in the
open hole section could be removed with acid. A 100 Ib quick setting
cement plug was then placed on top of the crushed oyster shells @
3570' .
2) The production casing shoe at 3550" was cement squeezed with 20 sacks
of Class "C" cement with 2% CaCl2 and 30 sacks Class "H" thixotropic
cement.
3) After WOC time, the cement and oyster shells were drilled out to a TD
of 4024'. The shot open hole sections were then jet washed. Jet
washing uses a sub above the bit that has an orifice which directs
hydraulic impact force towards the open hole walls. The hydraulic
impact force removes scale and loose formation rock from the shot open
hole sections.
4) The wellbore was then prepared for the resin coated gravel placement
by:
a) Rattling and pickling the tubing by spotting 32 bbls of a mutual
solvent and scale converting chemical solution in the open hole
section for 13 hrs. Every two hours after the chemical solution
was in place, the downhole assembly was worked up and down 60'.
-565-
-------
Page No. 9
This would agitate the chemical solution in the open hole section.
The scale converter chemical was used because samples from
drilling scale bridges in the well had indicated the bridges to be
calcium sulphate (CaSO^ scale, which is only moderately
soluable in acid without conversion.
b) The chemical solution was reversed out and 37 bbls of 15% HC1
treated with scale inhibitor was then spotted in the open hole
section. The workstring and bottom hole assembly was stroked 60'
after 30 minutes of shut-in time to agitate the acid solution in
the open hole. The acid was then allowed to soak an additional 30
minutes. After the 1 hour soak period, the acid was reversed out
of the hole with 233 bbls of 8.4 Ib/gal KC1 water filtered to two
microns.
The pH of the circulating fluid was adjusted with clay stabilizers
to 6.8. It was important that the pH of the fluid in the wellbore
be in a range of 6-8. If the pH was too low, the resin would
prematurely harden, and if the pH was too high, the resin would
not harden.
The mixing procedure was: 1) 50 bbl of gelled brine (2% KC1, 40
Ibs hydroxyethylalcohol/1000 gals) was prepared; 2) 33.125 bbls of
this gelled brine was actually used for mixing the slurry; 3)
23,875 Ibs of sand, 209 gallons of resin, and 1375 Ibs of silica
flour were added to the gel in succession. The mixing process
took approximately 2 hours.
c) After circulating the 8.4 Ib/gal KC1 water, total depth was tagged
at 4024' and the workstring was picked up 10' off bottom.
-566-
-------
Page No. 10
5) The resin coated gravel mix was placed in the open hole by:
a) Pumping 52 bbls of resin coated gravel slurry and displacing the
slurry with 18 bbls of 8.4 Ib/gal KC1 water. The calculated open
hole volume was 37 bbls. 30% additional volume was calculated for
slurry shrinkage and 11% was calculated for excess open hole
volume. Laboratory work had indicated that when the gel broke in
the resin slurry, a 30% volume reduction occurred. The 11^ excess
volume was used due to jet washing the open hole after drilling
out the oyster shells. This jet washing increased the shot open
hole size by removing scale and loose formation.
By displacing the resin coated gravel with only 18 bbls of 8.4
Ib/gal brine, the tubing was left with a calculated overbalance of
344 psi. This overbalance allowed the tubing to be pulled dry
immediately after shutting down the displacing pumps. Initial
pumping was at 2 bbl/min at 600 psi.
b) Immediately after displacing the resin coated gravel, the tubing
was pulled 900", 150 psi was applied at the surface, and the well
was shut-in.
6) After waiting on resin for 30 hours, the resin coated gravel was
drilled out to a total depth of 4024' with 6-1/4" milled tooth bit.
The weight on bit was 4000 Ibs and the penetration rate was 150'/hour.
At this point, the shot open holes were filled with the permeable
resin coated gravel and a 0.375" sheath of permeable resin coated
gravel existed between the shot sections in the wellbore.
The intervals from 3560'-3575' and 3670'-3850' were underreamed to 7"
-567-
-------
Page No. 11
to remove the permeable resin fill sheath from above and between the
shot sections.
7) A 4-1/2", 10.5 Ib/ft, K-55, ST&C liner was run and cemented in place
with 120 sxs of a 50/50 Pozmix Class "C" cement mixture treated with
fluid loss additive. The density of the slurry was 15.5 Ib/gal.
Centralizers were used and during cement displacement, no pipe
movement was used. The minimum displacement rate was 6 BPM and the
maximum displacement rate was 7 BPM.
8) After WOC time, the cement and cement plugs were drilled out and the
liner top was tested to 1200 psi.
9) The well was then logged with a CBL, CCL-GR and perforated from
4005'-3890' and 3636'-3595' with a 3-1/2" hollow steel carrier gun
loaded with 1 JSPF (see Figure 4). A total of 158 shots was fired.
Testing
The perforations where the shot open holes existed were straddled
individually with bridge plugs and packers. Injection rates and pressures
were recorded and the intervals were tested for communication. The
testing revealed that no communication existed between shot sections and
injectivity had increased from 0.00403 bbls/psi/NEP to 0.00857
bbls/psi/NEP.
After seven days of continuous injection, an injection profile and
temperature log were run. These logs revealed that 15%-28% of the
injected fluid was entering the Grayburg 6th, 10% was entering the
-568-
-------
Page No. 12
San Andres Upper 9th and 72% to 75% was entering the San Andres Lower 9th
Massive. The injection profile had changed dramatically in comparison to
the profile run before the "Puddle-Pack" (see Figures 5 and 6). The
profile of MCA Unit No. 61 run after the "Puddle-Pack" closely resembled
the injection profile of MCA Unit No. 257, a cased hole injector 3
locations to the east (see Figures 6 and 7). An injection profile survey
run 7 months after the "Puddle-Pack", revealed that the injection profile
had not changed significantly as compared to the original injection
profile log run 7 days after the "Puddle-Pack" (see Figure 8).
After two weeks of injection, injectivity declined to .0052
bbls/psi/NEP, still higher than the injectivity before the "Puddle-Pack".
After five weeks of continuous injection, the injectivity stabilized at
0.0052 bbls/psi/NEP. The injectivity has been monitored once a month
since that time. Table 4 lists the injectivity for MCA Unit No. 61.
The cost to "Puddle-Pack" MCA Unit No. 61 was approximately $141,000.
This represented a savings to Conoco of approximately $275,000 vs. new
well drilling costs to achieve the same goals.
Also, it is believed that potential exists to "Puddle-Pack" shot
open hole producing wells. The economic advantages would be:
1. Reduction of clean out frequency and costs caused by open hole
sloughing.
2. Improved stimulation of producing wells by allowing mechanical
isolation of pay horizons during stimulation.
3. Reduction of chemical volumes required to stimulate the wells compared
to the volumes required to treat the shot open holes.
-569-
-------
Page No. 13
4. Allowing more efficient beam lifting techniques by placing the seating
nipple below the producing horizon instead of above the producing
horizon as in shot open hole wells.
Conclusions
1. A method has been developed and tested to convert multiple shot
section open hole completions into cased hole completions.
2. The "Puddle-Pack" method has provided zonal isolation between shot
sections in MCA Unit No. 61.
3. The "Puddle-Pack" method has controlled injection fluid loss to non-
pay horizons in MCA Unit No. 61.
4. The "Puddle-Pack" process is mechanically and economically feasible.
5. Interpretable injection profile logs can be obtained after an
injection well has been successfully repaired using the "Puddle-Pack"
process.
-570-
-------
Page No. 14
Acknowledgements
The authors wish to thank Conoco Inc. and Halliburton Services for
allowing the publication of this paper. They would also like to thank the
management and Engineering staff of the Conoco, Hobbs, New Mexico,
Division office, for without their support and assistance the work would
not have been done.
SI METRIC CONVERSION FACTORS
bbl x 1.589 873 E-01 = ra3
ft x 3.048 E-01 = m
gal x 3.785 412 E-03 = m3
Ibm x 4.535 924 E-01 = kg
Ibm/gal x 1.198 264 E+02 = kg/ra3
psi x 6.894 757 E-03 = MPa
REFERENCES
1. Saucier, R. J.: "Gravel Pack Design Considerations", SPE 4030,
presented at the 47th Annual Meeting of the Society of Petroleum
Engineers, San Antonio, Tex., Oct. 1972.
2. VanPoollen, H. K.; Tinsley, J. M.; and Saunders, C. D.: "Hydraulic
Fracturing: Fracture Flow Capacity vs. Well Productivity:, paper
number 890-G presented at the 32nd Annual Meeting of the Society of
Petroleum Engineers, Dallas, 1957.
3. Cole, R. C.: "Epoxy Sealant for Combatting Well Corrosion", SPE 7874
presented at the SPE International Symposium on Oilfield and
Geothermal chemistry, Houston, TX., Jan. 1979.
-571-
-------
Page No. 15
Table 1
Compressive Strength of Resin Consolidated Formation Exposed to a
Saturated CC^-Water at 120° F
Compressive
Test Strength, psi
First Day 1675
8 days 1675
16 days 1835
34 days 1910
Table 2
Permeability of Mixtures of 70-170 Mesh Sand in 10-20 and
20-40 Mesh Sands
Permeability in Darcies
10-20 Mesh 20-40 Mesh
Rounded Sand Angular Sand
0% Fine Sand Added 310 121
10% Fine Sand Added 90 70
20% Fine Sand Added 60 25
*From Ref (a) VanPoollen, Tinsley and Saunders
Table 3
Compressive Strengths and Permeabilities of Various Tests
Compressive Permeability
Strength, psi Darcies
Laboratory, 5% Silica Flour; 44755
80 lb/1000 gal gel carrier
Laboratory, 5% Silica Flour;
40 lb/1000 gal carrier 4760 2.1
Resin Cement System 8000 .01
Mixing Test 1890 12*
MCA No. 61 1960 N.R.
est. 6*
*These were permeabilities of samples settled from the slurry, not
pressure compacted. The samples were taken from the blenders after the
test.
-572-
-------
Page No. 16
Table 4
Injectivity of MCA Unit No. 61
Date
7/84
8/84
9/84
10/84
11/84
12/84
1/85
4/85
5/85
6/85
7/85
8/85
9/85
10/85
Injectivity
bbl/psi/NEP
Comments
0.0031
0.0029
0.00366
0.00403
0.00401
0.00398
0.00403
0.00857
0.0052
0.0083
0.0051
0.0052
0.0052
0.0051
MCA Unit No. 61
"Puddle-Packed"
-573-
-------
Table 1
Compre3slve Strength of Resin Consolidated Formation Exposed to a
Saturated CO -Water at 120°F
Compressive
Test • Strength, psi
First Day . 1675
8 days 1675
16 days 1835
34 days 1910
Table 2
Permeability of Mixtures of 70/170 Mesh Sand in 10/20 and
20/40 Mesh Sands
Permeability in Darcies
10/20 Mesh20/40 Mesh ^
Rounded Sand Angular Sand
0% Fine Sand Added 310 121
10% Fine Sand Added] 90 70
20% Fine Sand Added 60 25
From Ref (4) VanPoollen, Tlnsley and Saunders
Table 3
Compressive Strengths and Permeabilities of Various Tests
Compressive Permeability
Strength, psi Darcies
Laboratory, Test 1 4475 6
Laboratory, Test 2 4760 2.1
2
Resin Cement System 8000 .01
Bulk Mixing Test 1890 12
MCA Ib 61 I960 N.R. ^
est. 6
These were permeabilities of samples settled from the slurry, not
pressure compacted. The samples were taken from the blenders after the
test.
Table 4
Date
7/84
8/84
9/84
10/84
11/84
12/84
1/85
4/85
5/85
6/85
7/85
8/85
9/85
10/85
Injectivity of MCA Unit No. 61
Injectivity
bbl/psi/NEP Comments
0.0031
0.0029
0.00366
0.00403
0.00401
0.00398
0.00403
0.00857 MCA Unit No. 61
"Puddle-Packed"
0.0052
0.0083
0.0051
0.0052
0.0052
0.0051
-574-
-------
T.D. 4024' P.B.T.D. 4020'
10 3/4' AT 721
GAMMA RAY NEUTRON
T AT 3558'
SHOT W/80 QTS. NITRO
AVG. HOLE DIA. 19'
SHOT W/70 QTS. NITRO.
AVG. HOLE DIA. 20"
SHOT W/170 QTS. NITRO
AVG. HOLE DIA. 17"
GRAYBURG-6th ZONE-
TOP SAN ANDHES-7th ZONE -
Sth ZONE
9th ZONE-
9-M ZONE —
Fig. 1—MCA Unit 61, a typical Injection well.
Fig. 2—Typical tog aectlon.
XXXXXXXXXXXXXXX/ XXXXXXXXXX
MCA UNIT
SCALE
1 MILE
r_.
*
PILOT AREA
XXXXXXXXXXXXXXXX,
PROPOSED 1st STAGE EXPANSION
/
\xxx\xxx\\xxxx
Fig. 3— MCA unit.
-575-
-------
T.D. 4024' P.B.T.D. 4020'
10 3/4- AT 72'
AT 3558'
SHOT W/80 QTS. NITRO
AVG. HOLE DIA. 19'
SHOT W/70 QTS. NITRO.
AVG. HOLE DIA. 20'
SHOT W/170 QTS. NITRO
AVG. HOLE DIA. 17'
4 1/2' LINER
RESIN COATED GRAVEL
Fig. 4—MCA Unit 61 after puddle pick.
CASING
SHOE
GAMMA
' RAY
58%
I
1
^VELOCITY
I PROFILE
I SURVEY
I
38%
TEMPERATUR
SURVE
Fig. 5— Injection profile survey before puddle pick.
TRACER VELOCITY
15%
16%
VELOCITY-
TRACER
Fig. 6—Injection profile survey after puddle pack.
-576-
-------
TRACER
VELOCITY
TEMPERATURE
VELOCITY
% OF LOSS
TRACER
0 20 40 60 80
1 ' J '
% OF LOSS
VELOCITY
0 20 40
11%
10%
7%
3%
INJECTING
TEMPERATURE
|-30 MIN. SHUT IN
1 HOUR SHUT IN
Fig. 7—Injection profile cased hole MCA unit.
Fig. 0— Injection profile lurvey 7 monthi after puddle
pack.
-577-
-------
HOW TO LOCATE ABANDONED WELLS
by
J. Jeffrey van Ee and Eric N. Koglin
Environmental Monitoring Systems Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Las Vegas, Nevada 89114
ABSTRACT
Record searches are typically used to locate abandoned oil and gas wells
within the area of review for injection wells; however, the accuracy and success
in locating all of the abandoned wells often is questionable. In some cases,
the records may be incomplete, or inaccurate; in other cases, a thorough search
of the records may be quite time consuming, particularly when large areas and
multiple record bases must be searched. Other methods for locating abandoned
wells are frequently sought when the risk in missing an abandoned well from a
record search appears to be significant.
The U.S. Environmental Protection Agency (EPA) has conducted several
studies to determine if other means exist to locate abandoned wells. The R. S.
Kerr Laboratory conducted a literature search of alternate methods for locating
abandoned wells. Field, geophysical, and aircraft-based remote sensing surveys
were some of the methods that were highlighted in the final report. The
Environmental Monitoring Systems Laboratory in Las Vegas evaluated two of the
most promising methods in a survey of central Oklahoma for abandoned oil and
gas wells. The evaluation of geophysical methods began with the development of
a mathematical model for the magnetic anomaly produced by steel casing. The
United States Geological Survey determined from the mathematical modeling that
airborne magnetometry offered the greatest potential of success in surveying
large areas for abandoned wells. The EPA's Environmental Photographic Inter-
pretation Center evaluated historical aerial photographs as the second means
for locating abandoned wells. Photographs dating back to the 1930's were
examined. The data from the aerial magnetometer survey were compared against
the historical photographs, and the results from these two methods were then
compared against a search of the records. The record search was conducted by .
the University of Oklahoma's Environmental and Ground Water Institute.
All three methods were successful in locating abandoned wells. Each has
its own advantages and disadvantages. Used alone, each method was useful in
-578-
-------
locating abandoned wells. Used together, the methods were able to locate a
higher percentage of wells than any one of the methods used alone.
INTRODUCTION
It has been estimated that over two million abandoned wells exist in the
United States (Aller, 1984). Numerous problems are created by these wells, and
documented cases of pollution from abandoned wells are widespread. Improperly
plugged and abandoned wells may allow fluids to migrate between aquifers
especially when those wells are located within the zone of influence of under-
ground injection wells. When the piezometric surface is greater than the land
surface, brine may contaminate the land and surface waters. Abandoned oil and
gas wells may also allow gases to migrate toward the surface and into structures
where explosive levels may lead to fire and explosion. Abandoned agricultural
wells in Silicon Valley are a problem in conveying contaminated water from
shallow aquifers to deeper, drinking water aquifers. Knowing the location of
abandoned wells is an important first step in characterizing the potential for
pollution of underground sources of drinking water. Once the well has been
located, an assessment is usually made of the condition of the well to determine
whether it was properly plugged and abandoned.
Federal regulations developed in response to the Safe Drinking Water Act
and the Resource Conservation and Recovery Act require a search for abandoned
wells within an "area of review" of underground injection wells (see 40 CFR
Part 146). Typically, these searches are of records. Other data bases and
methods may be used when the risk in not locating all the wells within the area
of review is high, and when the location of a well in the field is complicated
by either a lack of surface features, or poor, incomplete, or nonexistent
records. In those instances where large areas must be surveyed (such as
counties where reservoirs or injection wells may be located) a search of the
records by itself may not be sufficient. Other methods must be examined.
The Environmental Protection Agency's R. S. Kerr Laboratory reviewed the
literature to determine what methods have been, or may be used to search for
abandoned wells (Aller, 1984). A variety of methods were identified with some
being routine and straight forward, such as a search of records or consulting
long-time property owners, and others, such as thermal mapping, being less
feasible. The EPA's Environmental Monitoring Systems Laboratory in Las Vegas
(EMSL-LV) chose to evaluate three of the most promising techniques: record
searches, historical aerial photographic analysis, and magnetometry. Much of
the research centered on four test areas outside of Oklahoma City where the
three methods were compared. The purpose of this paper is to summarize the
results of the research and to outline a strategy for locating abandoned wells.
A bibliography of EPA-funded research publications is provided for further
references.
RECORD SEARCHES
Written records for oil and gas, mineral exploration, water, and injection
wells reside in numerous locations throughout the country. Searches of those
records are a starting point in the search for abandoned wells. Where records
are easily accessible, such as computer data bases or maps, little effort is
-579-
-------
required. Where records are scattered or incomplete, more effort is required
and the pay back is reduced. When records do not exist, such as for wells that
were drilled many years ago, a search of the records will not locate all of the
wells within the search area. In many cases, particularly with the older
wells, descriptions are poor of where the wells are located, and how they were
drilled and plugged. A well may have been described as being two hundred feet
from the big oak tree; however, the tree may no longer be present. Even with
more modern day records, the accuracy in which the location of a well may be
pinpointed can be poor. A well may simply be located in 1/64 of a section, and
the area where the well may be located can be on the order of hundreds of feet.
When no surface features are left to identify the location of the well bore,
locating an abandoned well can be quite difficult from a search of the records.
One reason why it is important to first search the records is that infor-
mation presumably exists on how the well was drilled and whether the well was
properly plugged and abandoned by modern day standards. A search of historical
photographs and the use of magnetometers cannot provide this information
(Fairchild and others, 1983).
HISTORICAL AERIAL PHOTOGRAPHIC ANALYSIS
EMSL-LV began a research program in 1982 to devise a method of locating
abandoned wells cost effectively and quickly. This research program, conducted
by the Environmental Photographic Interpretation Center (EPIC) of EMSL-LV
located at Vint Hill Farms Station, Warrenton, Virgina, tested a method of
locating abandoned wells using historical aerial photography to locate old
wells during or close to their period of production, when well site features
are most recognizable. Photographic analyses are particularly useful in areas
where commercial or residential development in agricultural or oil-producing
areas have virtually obliterated the old wells.
Abandoned wells are located from aerial photographs through the development
of "signatures." A signature is a combination of characteristics or features
by which an object or activity can be identified on an aerial photograph.
Depending upon the land use and history of the area being analyzed (agricul-
tural, oil or gas production), these signatures may include pump houses, storage
tanks, derricks, impoundments, or depressions in the earth left by storage
tanks (Figure 1). Sites in which these signatures are very clear in successive
years of imagery are classified as "active/abandoned" wells. Signatures whose
origin is less certain are classified as "probable abandoned" or "possible
abandoned" wells, depending upon the degree of certainty.
As an example, the first application of this method was at sites located
around the Oklahoma City, Oklahoma, area (U.S. EPA, 1983). Signatures for the
well sites were developed through researching early petroleum publications,
personal communications with individuals familiar with old drilling techniques,
and preliminary field work. Signatures for producing wells were found to
include various combinations of the following features: maintained roads,
brine pits, derricks, power houses, ground stains, ground scars, walking beams
and scars from pipelines. The actual well locations were determined by knowing
the general spatial relationship between the wells and these recognizable
features. In addition, associated oil extraction activities such as storage
-580-
-------
oo
Figure 1. Active oil wells in 1951 located in the Arcadia, Oklahoma, study area.
-------
tanks, water/oil separation ponds and well spacing patterns aided in well site
identification. The analysts found that signatures varied from one Oklahoma
area to another, depending on the time the oil field was developed and the
technology used (Stout and Sitton, 1984).
Signatures for agricultural and water supply wells have features different
from those for oil and gas wells. The principal features include pump houses,
power poles, water tanks, shade trees, access roads and irrigation water flow
patterns. Usually, the features associated with agricultural or water supply
wells are not as prominent on aerial photographs because of their size and the
minimum ground resolving capability of the historical photography (U.S. EPA,
1987).
Once a signature is identified in a photograph, its location can be manu-
ally transferred from the historical photograph to a recent photograph or with
the aid of a computerized interactive graphics system. With the aid of this
current photograph and overlay, a field crew can inspect a site and identify
likely well locations within a relatively short period. In cases where the
well is not visible from the surface, the crew can use a magnetometer to locate
the metal well casing.
This historical photographic analysis method of locating abandoned wells
has advantages over the traditional method of record searches. First, it is
relatively quick. In about one month, a photointerpreter can analyze an area
of several square miles and provide maps and overlays to field crews. Gen-
erally, if the wells are in a nonurbanized area with moderate vegetation, the
field crews can locate the wells easily.
A second advantage of the aerial photographic analysis method is that it
provides confirmation of the location and number of wells in an area with one
common method used to identify past well drilling activity. Though it may not
be possible to locate all wells identified on the photos, it is still advan-
tageous for the purpose of risk assessment to know where and how many wells
exist.
Finally, there appears to be significant cost savings in using photographic
analysis. In a Santa Clara County, California, study area where the landscape
has changed drastically due to rapid growth and urbanization, costs to local
agencies for personnel to identify a single abandoned well have been as much as
$4,000. In the same area, the photo analysis cost as little as $40 per well.
As this technique can narrow the field of search to within a 50-foot diameter,
time spent in the field is greatly reduced, resulting in further cost savings.
This difference is especially dramatic in urban areas where literally thousands
of abandoned wells may exist.
No method for locating abandoned wells is perfect. A disadvantage of the
photographic method is that some abandoned wells will not be found, particularly
when photo coverage of the area has been poor, and rapid changes have occurred
on the surface. When combined with other methods, such as record searches, the
use of historical aerial photographs can increase the level of success in
finding all the abandoned wells that may exist in an area.
-582-
-------
SOURCES FOR HISTORICAL PHOTOGRAPHS
Historical aerial photography is available through a number of different
sources, some of which are listed in Table 1.
TABLE 1. SOURCES FOR HISTORICAL AERIAL IMAGERY
° Limited air photo library maintained by EPIC and the Remote and Air
Monitoring Branch at EMSL-LV
° National Cartographic Information Center, Reston, Virginia, (will
research availability of aerial photo coverage). Telephone number:
(703) 860-6045
° EROS Data Center, U.S. Geological Survey, Sioux Falls, South Dakota,
(maintains photography for the following agencies: USGS, USAF, USA,
USN, BLM, COE, and NASA). Telephone number: (605) 594-6511, ext. 151
° National Archives and Records Service, Alexandria, Virginia, (photog-
raphy from the 1930's and early 1940's). Telephone number: (703)
756-6700
° Agriculture Stabilization Conservation Service, U.S. Department of
Agriculture, Salt Lake City, Utah. Telephone number: (801) 524-5856
° National Oceanic and Atmospheric Administration, National Ocean
Survey, Rockville, Maryland. Telephone number: (301) 443-8661
° State Department of Transportation Offices
° County Tax Assessors
0 Private companies which specialize in taking aerial photographs
MAGNETOMETER SURVEYS
Abandoned wells with few visually evident surface features can be located
with magnetometers. Ferrous metal scrap and trash located on the surface near
the well bore and steel casing in the hole can be used to locate the well in
areas where cultural features, such as metal tanks, fences, and houses are few
and far between. The earth's magnetic field averages approximately 53,000
gammas in the U.S. and a proton precession magnetometer is able to measure
changes in the field intensity of a few gammas. By mapping the magnetic field
in an area, it is possible to locate the well bore of a steel-cased well to
within a few feet.
-583-
-------
The United States Geologic Survey (USGS) performed several studies for the
EMSL-LV to determine whether airborne magnetometry could be used to locate
abandoned wells. It was first necessary to develop a mathematical model of the
magnetic anomaly to determine the optimal altitude and spacing of the aircraft
flight paths. Ground-based magnetometer measurements were made in the vicinity
of steel-cased oil wells to verify the model. Good agreement was obtained
between the calculated and observed magnetic anomaly on the ground (Figures 2a
and 2b). This provided some assurance that the extrapolation of the model to
the airborne case would be valid (Frischknecht and Raab, 1984).
The model indicated that abandoned wells with a minimum of several hundred
feet of casing could be located from an aircraft at 200-foot altitudes (Figure
3). The spacing between the flight paths would have to be approximately 300-400
feet to adequately map the magnetic anomaly with a proton precession magnetom-
eter.
The USGS possessed a small private plane that was instrumented and used
for magnetometer surveys. The magnetic field has been extensively mapped by
the USGS across most of the U.S. Small features such as wells were not observed
because they were not of interest, and they would contribute "noise" to the
magnetic field of the underlying geologic material of interest. Ferrous metal
materials in the plane had been removed, and the much of the remaining magnetic
field had been compensated by the use of coils and an electrical current to
produce an opposing magnetic field. A radar altimeter was used to record the
altitude above the ground and to ensure that the plane kept a constant above
the ground. A ground-based radio navigation system was deployed on the perime-
ter of each test area in Oklahoma to allow the pilot to maintain precise flight
paths and to allow the magnetic data to be referenced to an accurate location.
The orientation of the aircraft was also recorded to permit compensation of the
magnetic data after the flight. Figure 4 graphically depicts this airborne
profile data from one of the Oklahoma study areas. While the use of airborne
magnetometry by the USGS would seem to be an involved, complicated process, the
general process and most of the equipment could be readily acquired and used by
commercial airborne magnetometer firms.
Measurements of the magnetic field with a ground-based magnetometer can be
complicated by nearby, small pieces of metal. Figures 2a and 2b illustrate the
magnetic profile generated by a ground-based magnetometer. The response drops
off rapidly within a short distance from the well location, therefore, metal
debris in the subsurface can mask a location or confuse an interpretation. An
airborne magnetometer is not as sensitive to ground clutter in the mapping of
the magnetic anomaly from larger objects. The magnetic anomaly from a well
will be reduced in intensity with altitude (Figure 5); however, the anomaly
will broaden in size and fewer survey lines will be required to detect the well
casing. Ground-based magnetometers are able to pinpoint the location of a
buried well casing to within a few feet.
The airborne magnetic surveys in Oklahoma found a significant number of
abandoned wells (Figure 6). Some of the anomalies could easily be associated
with a visible feature such as the well head or concrete pad; other anomalies
could not. Ground-based magnetometer measurements were made where an abandoned
well could not be observed from the surface. In many instances, the anomaly
-584-
-------
N-S MAGNETIC TRAVERSE
54000.
54720.
54060.
I 64400.
an
" 54240.
a
« 04080.
— 53920.
N
53760.
53600.
53440.
53280.
63200.
T
I ' I
o o o o o
o o o o
MM— —
1 I I
DISTANCE (Fe»t)
Figure 2a. Observed and calculated north-south magnetic profiles
over well No. 17, Horseshoe Lake test area, Oklahoma.
54*0™
54720.
54560.
54400.
54240.
« 54080.
U.
CJ
« 53920.
2
" 53760.
^ 53600.
*~ 53440.
53280.
63200.
E-W MAGNETIC TRAVERSE
8 ° S S S 8
y - S » •»
DISTANCE (F*«t)
8
Figure 2b. Observed and calculated east-west magnetic profiles
over well No. 17, Horseshoe Lake test area, Oklahoma.
-585-
-------
0)
(U
o
c
CO
**
CO
-400
-300
-200
-100
100
200
300
400
N "
-400 -300 -200 -100 0 100 200 300 400
Distance (feet)
Figure 3. Calculated center map of total intensity at a height of
200 feet above a well (the lines show 200 foot spacings on north-south
or east-west flight (has to measure a two gamma anomaly in the worst case),
-586-
-------
MANEUV. NOISE
COR.-GAMMAS
DIFFERENTIAL
ROLL-DEG.
DIFFERENTIAL
PITCH-DEG.
DIFFERENTIAL
HEADING-DEG.
BARO. ALTMTR
METERS
RADAR ALTMTR
METERS
0.76
-0.76
376
300
26
COR. MAG
FIELD-GAMMAS
4620
1
4480
4440
-3
'100
26
26
4620
4480
4440
1
H MILE
Figure 4. Airborne profile data from Arcadia area (the numbers at the top
and bottom are identification numbers associated with each reading, and the
numbered anomalies correspond with those on Figure 6).
-587-
-------
N
o>
SB'
•o
0)
(0
*»
o
200-n
100-
0 J
Height of Plane
250 feet
200
150
100
5000
10000 feet
I
Figure 5. Aeromagnetic profiles for different aircraft heights
over Well No. 4, Piney Creek test area, Colorado.
-588-
-------
Figure 6. Total intensity contour map for part of
the Arcadia, Oklahoma, test area.
-589-
-------
measured on the ground could be associated with an abandoned well. Sometimes,
the well may have been a water well. In other instances, the anomaly which
appeared to be from a well was from a pipeline that traversed a hill. The bend
in the pipeline produced the magnetic anomaly that appeared to be from an
abandoned well (Frischknecht and others, 1984).
Ground-based magnetic surveys for abandoned agricultural wells have been
conducted by the USGS in Silicon Valley. A mathematical model was developed
and verified with field data to determine the minimum size of casing that could
be observed in an urban area. The mathematical model indicated that most
agricultural wells could be located in theory, but the difficulties in making
measurements in an urban area remained to be investigated.
Where historical photographs were able to locate a probable abandoned well
in a vacant lot or in a backyard, ground-based magnetometer measurements were
usually successful in locating abandoned agricultural wells in the urbanized
areas of Silicon Valley. When a well was thought to exist in a parking lot,
ground-based magnetometer measurements were complicated by the presence of
buried utilities, reinforcing steel, and nearby automobiles and buildings.
When abandoned wells were thought to exist under buildings, no magnetometer
measurements were made, nor is it believed that they could have been made with
the interfering utilities and nearby metal objects. Without the use of his-
torical photographs to identify search areas for magnetometer measurements,
magnetometer measurements for abandoned wells in urbanized areas are likely to
be less effective and more costly than searches for abandoned wells in less
developed areas (Jachens and others, 1986). Further details may be obtained
from the USGS and the publications listed in the bibliography.
COST COMPARISON
Costs for locating abandoned wells vary with the area and the elapsed time
since the well was drilled. It has been estimated that a search of records to
"locate" an abandoned well costs approximately $50 (Arthur D. Little, Inc.,
1979 in van Ee, 1984). To actually locate the well may require the use of other
methods and data sources. The cost of using historical photographs has been
estimated at approximately $600 per square mile (Stout and Sitton, 1984). As
noted previously, "probable" or "possible" abandoned wells will require field
verification. The cost for conducting magnetometer surveys is more difficult
to estimate. The size of the search area is an important factor because the
deployment costs for an airborne magnetometer survey can be significant no
matter what size an area is to be surveyed; thus, the cost on a "square mile"
basis will be lowered as the number and size of areas increases. A cost-figure
obtained for airborne magnetometer surveys from the Oklahoma studies was between
$1,000 to $2,000 per square mile. For ground-based magnetometer surveys,
estimates of the cost are relatively fixed. On a lineal basis, the costs range
from $50 to $121 per line-mile, or approximately $3,100 to $12,100 per square
mile. While the equipment required to perform an airborne survey is more
sophisticated and expensive than required for ground-based measurements, the
increased time required to perform a ground-based survey of a large area leads
to higher costs (Frischknecht and Raab, 1984).
-590-
-------
IMPLEMENTATION STRATEGY
Airborne magnetometer studies proved to be cost effective in areas where
low-level flights could be made and where little development of the land occur-
red. Federal Aviation Administration (FAA) regulations limit the type of
aircraft and flight patterns that can be flown at low altitudes. Tall, man-made
objects such as radio towers, electrical transmission towers, water towers,
buildings, and silos need to be located before an assessment can be made on the
practicality of airborne measurements. Low-level overflights of farms and
dwellings are permissible with certain restrictions; however, as the number of
proposed flights over these features increase, the difficulties in complying
with the FAA regulations and the likelihood of complaints also increase.
Aircraft magnetometer measurements can only be considered after other factors
have also been considered.
Typically, the first approach is to consult the records. The next approach
would be to use historical photographs, and the third approach would be to
consider magnetometry with airborne measurements being a consideration for
surveying large areas. Ground-based magnetometer measurements should always be
considered in locating those wells that have little, if any, visible surface
features.
Study Areas - Lessons Learned
Table 2 lists all the abandoned wells projects which have been conducted
by EMSL-LV. These eight projects provided great insight into the application
and limitations of the above-mentioned methods. The following sections discuss
some to the lessons learned from selected projects.
OKLAHOMA AND CLEVELAND COUNTIES, OKLAHOMA
The objective for the studies conducted in Oklahoma and Cleveland Counties,
Oklahoma, was to test, evaluate, and compare the three previously discussed
methods for locating abandoned wells. Four areas were selected within these
counties because of the presence of underground injection wells in each area.
These study areas represented ideal locales in which each method worked
very well. The Oklahoma Corporation Commission records were adequate; the
aerial photographic method was very successful because signatures were well
defined and not obscured by urban growth or revegetation; and, the ground and
airborne magnetic surveys did not suffer from interference effects due to
cultural features (U.S. EPA, 1983).
As an example, in the Arcadia, Oklahoma, study area 36 wells were identi-
fied from photos, 41 were wells identified from the record search and 37 were
wells identified with magnetic methods. Frischknecht and Raab (1984) concluded
that 95 to 98 percent of the magnetic anomalies identified in the four study
areas were associated with abandoned wells. Stout and Sitton (1984) concluded
that 91 percent of the abandoned wells in the four study areas were identified
with the aerial photographic method, using the results of the record search as
a measuring stick. They believe that some additional wells may not have been
-591-
-------
TABLE 2. SUMMARY OF EMSL-LV ABANDONED WELLS PROJECTS
State EPA Region Year Completed
Oklahoma 6
Cleveland and
Oklahoma Counties 1983
Kay County 1985
Washington County
Pennsylvania 3
Elk, McKean, and
Warren Counties 1985
Allegheny Reservoir 1987
New York 2
Chautauqua County 1985
(Levant)
California
Farmers Market 9 1985
Santa Clara County 9 1985 (Phase I)
1987 (Phase II)
identified, but these represent a very small minority and would not signifi-
cantly change the accuracy rate of the photo analysis.
PENNSYLVANIA STUDY AREAS
Only the photographic analysis method was applied in the study areas in
Pennsylvania (Elk, McKean, and Warren Counties and the Allegheny Reservoir).
Because of the rapid revegetation, and much of the oil exploration predated the
earliest aerial imagery, the photo analysis method was not as successful com-
pared to the Oklahoma experience. The signature developed for these areas had
some similarities to the Oklahoma areas; however, unique oil field attributes
were identified in Pennsylvania (U.S. EPA, 1985 and U.S. EPA, in progress).
-592-
-------
SANTA CLARA COUNTY, CALIFORNIA
In this study area, aerial photographic analysis and hand-held magnetometer
surveys were used to locate abandoned agricultural and domestic water supply
wells. The signature developed for this study area included site features
which are different from those used in oil and gas production areas. Using
these signatures, 805 wells were identified in the 26-square mile study area.
This number may not account for all abandoned wells since some may not have
exhibited any surface features visible on historical photography, or the wells
may have been obscured by vegetation.
Field work failed to reveal the degree of accuracy of the historical photo
analysis method because of the small sample of wells visited and the difficul-
ties encountered in verifying their locations in an urban environment. Many of
the photo-identified wells are now located under buildings, parking lots, and
highways. Geophysical methods proved less successful in Santa Clara County
because of the abundance of metal objects and structures present in the study
area. It is also conceivable that some well casings may have been removed
during the construction of highways and buildings (U.S. EPA, 1987).
SUMMARY AND CONCLUSIONS
Three methods were used to locate abandoned oil, gas, agricultural, and
water supply wells in various areas around the U.S. As each method was applied
in the Oklahoma study areas, the level of confidence that all abandoned wells
had been located increased, but each method also raised the total cost of the
investigation. The records search provided information on well construction
which the other techniques cannot supply; therefore, it is likely that records
search will always be required to assess the pollution potential from abandoned
wells. Unfortunately, the information contained in the records on both well
location and construction may not be complete or accurate. Additional location
techniques are desirable to supplement the data.
Historical aerial photographs are particularly valuable for those periods
when records are not complete or accurate. This particularly true for the
period from the 1930's through 1950's during which improved, wide-spread photo-
graphic coverage became available and accurate records were not often required.
In areas where rapid land use changes have occurred, it can be difficult to
locate abandoned wells when the length of drilling time at a site was short in
relation to the period of time between photos. Even wells drilled in the
recent past, when frequent photographs are likely to exist, can escape detection
because the length of time that modern-day rigs spend on a site can be less
than in the past when the drilling derrick had to be constructed at the site.
Fortunately, the increased emphasis on developing good records has made the
problem of locating recently abandoned wells much easier.
The aeromagnetic method, like the photographic method, can be readily used
to locate abandoned wells for many areas where there has been no surface evi-
dence of the well. Large areas can be surveyed rapidly from the air without
need for access to the property. While the method allows a well casing to be
located to within 3 to 6 feet with the aid of a ground-magnetometer, the method
is costly. An aeromagnetic survey requires more sophisticated equipment and
-593-
-------
technical expertise than the other two methods. However, more wells were
detected by the aeromagnetic surveys than by the initial photointerpretation.
For any survey method, or methods, selected will depend on the available
resources and the potential threat posed by unknown locations of abandoned
wells in an area.
ACKNOWLEDGEMENTS
The authors would like to recognize all the individuals who have con-
tributed to the EMSL-LV abandoned wells studies. Chief among them are Kristen
Stout of the Bionetics Corporation and Frank Frischknecht of the United States
Geological Survey. Their efforts and resulting publications have provided
important contributions to the 4-year abandoned wells research program.
NOTICE
Although the research described in this article has been funded wholly or
in part by the United States Environmental Protection Agency, it does not
necessarily reflect the views of the Agency and no official endorsement should
be inferred.
ERRATUM NOTICE
The authors found a few omissions and oversights in the reference and
bibliography sections after the final copy was submitted to the UIPC.
Citations highlighted with asterisks {*) are the replacements for the pre-
ceeding citation. The body of the paper does not cite the replacements.
The citation highlighted with the pound symbol {#) was inadvertantly
omitted from the bibliography.
REFERENCES
Aller, L. 1984. Methods for Determining the Location of Abandoned Wells.
EPA-600/2-83-123. Available through NTIS, Publication No. PB84-141530 and
through NWWA.
Fairchild, D. M., C. M. Hull, and L. W. Canter. 1983. Selection of Flight
Paths for Magnetometer Survey of Wells. Environmental and Ground Water
Institute. The University of Oklahoma, Norman, Oklahoma. EPA Unpublished
Report.
Frischknecht, F. C. and P. V. Raab. 1984. Location of Abandoned Wells with
Geophysical Methods. EPA-600/4-84-085. Available through NTIS, Publica-
tion No. PB85-122638. Frischknecht, F. C., L. Muth, R. Grette, T. Buckley,
and B. Kornegay. 1984. Geophysical Methods for Locating Abandoned Wells.
EPA-600/4-84-065. Available through NTIS, Publication No. PB84-212711.
-594-
-------
Frischknecht, F- C., P- V. Raab, R. Grette, and J. Meredith. 1984. Aeromag-
netic Surveys for Locating Abandoned Wells. USGS unpublished report.
Jachens, R. C., M. W. Webring, and F. C. Frischknecht. 1986. Abandoned-Well
Study in the Santa Clara Valley, California. USGS Open-file Report 86-350.
Stout, K. K. and M. D. Sitton. 1984. Locating Abandoned Oil and Gas Wells
with Historical Aerial Photos. Proceedings of the First National Confer-
ence on Abandoned Wells: Problems and Solutions, held May 30 to 31, 1984.
Environmental and Ground Water Institute, University of Oklahoma, Norman,
Oklahoma.
U.S. EPA. 1983. Abandoned Wells Study: Oklahoma and Cleveland Counties,
Oklahoma. TS-PIC-83051.
*Stout, K. K. and M. D. Sitton. 1983. Abandoned Wells Study: Oklahoma and
Cleveland Counties, Oklahoma. The Bionetics Corporation for the U.S. EPA.
Report Number TS-PIC-83051.
U.S. EPA. 1985. Abandoned Wells Study: Elk, McKean, and Warren Counties,
Pennsylvania. Technical Support to Region III. Two Volumes. TS-PIC-
85008.
*Sitton, M. D. 1985. Abandoned Wells Study: Elk, McKean, and Warren Counties,
Pennsylvania. The Bionetics Corporation for the U.S. EPA. Technical
Support to Region III. Two Volumes. Report Number TS-PIC-85008.
U.S. EPA. In progress. Abandoned Wells Study: Allegheny Reservoir,
Pennsylvania. Technical Support to Region III. Two Volumes.
*Stouts K. K. and L. M. Fauss. In progress. Abandoned Wells Study: Allegheny
Reservoir, Pennsylvania. The Bionetics Corporation for the U.S. EPA.
Technical Support to Region III. Two Volumes.
U.S. EPA. 1987. Abandoned Agricultural Wells: Santa Clara County, California.
Technical Support to Region IX. Two Volumes TS-PIC-86046.
*Stout, K. K. and L. M. Fauss. 1987. Abandoned Agricultural Wells: Santa
Clara County, California. The Bionetics Corporation for the U.S. EPA.
Technical Support to Region IX. Two Volumes. Report Number TS-PIC-86046.
van Ee, J. J., L. Aller, K. K. Stout, F- Frischknecht, and D. Fairchild. 1984.
Summary and Comparisons of Three Technologies for Locating Abandoned Wells
in Central Oklahoma. Proceedings from the Seventh National Ground Water
Symposium, September 26 to 28, 1984, Las Vegas, Nevada. Available through
the NWWA.
BIBLIOGRAPHY
Aller, L. 1984. Abandoned Wells: How to Find Them. Proceedings from the
Seventh National Ground Water Symposium, September 26 to 28, 1984,
Las Vegas, Nevada. Available through the NWWA.
-595-
-------
Environmental and Ground Water Institute. 1984. Proceedings of the First
National Conference on Abandoned Wells: Problems and Solutions, held May
30 to 31, 1984. Environmental and Ground Water Institute, University of
Oklahoma, Norman, Oklahoma.
Frischknecht, F. C. et al. 1983. Geophysical Methods for Locating Abandoned
Wells. USGS Open-File Report 83-702.
#Frischknecht, F- C., L. Muth, R. Grette, T. Buckley, and B. Kornegay. 1984.
Geophysical Methods for Locating Abandoned Wells. EPA-600/4-84-065.
Available through NTIS, Publication No. PB84-212711.
Frischknecht, F. C., D. P- O'Brien, R. Grette, and P. V. Raab. 1985a. Location
of Abandoned Wells by Magnetic Surveys: Acquisition and Interpretation of
Aeromagnetic Data for Five Test Areas. USGS Open-File Report 85-614A.
Frischknecht, F. C., D. P. O'Brien, R. Grette, and P. V. Raab. 1985b. Location
of Abandoned Wells by Magnetic Surveys: Location Maps and Aeromagnetic
Contour Maps. USGS Open-File Report 85-614B.
U.S. EPA. 1985a. Abandoned Wells Study: Kay County, Oklahoma. Technical
Support to Region VI. Two Volumes. TS-PIC-85008D.
*Stout, K. K. and L. M. Fauss. 1985a. Abandoned Wells Study: Kay County,
Oklahoma. The Bionetics Corporation for the U.S. EPA. Technical Support
to Region VI. Two Volumes. Report Number TS-PIC-85008D.
U.S. EPA. 1985b. Abandoned Wells Study: Washington County, Oklahoma. Tech-
hnical Support to Region VI. Two Volumes. TS-PIC-85008F-
*Stout, K. K. and L. M. Fauss. 1985b. Abandoned Wells Study: Washington
County, Oklahoma. The Bionetics Corporation for the U.S. EPA. Technical
Support to Region VI. Two Volumes. Report Number TS-PIC-85008F.
U.S. EPA. 1985c. Abandoned Wells Study: Chautauqua County, Levant, New York.
Technical Support to Region II. Two Volumes. TS-PIC-85008D.
*Stout, K. K., L. M. Fauss, and M. D. Sitton. 1985c. Abandoned Wells Study:
Chautauqua County, Levant, New York. The Bionetics Corporation for the
U.S. EPA. Technical Support to Region II. Two Volumes. Report Number
TS-PIC-85008D.
U.S. EPA. 1985d. Abandoned Wells Study: Farmers Market Area-Los Angles,
California. Letter Report to Region IX.
*Stout, K. K. 1985d. Abandoned Wells Study: Farmers Market Area-Los Angles,
California. The Bionetics Corporation for the U.S. EPA. Letter Report to
Region IX.
-596-
-------
BIOGRAPHICAL SKETCHES
J. Jeffrey van Ee is an electronics engineer with the Aquatic and Subsur-
face Monitoring Branch at the EPA Environmental Monitoring Systems Laboratory
in Las Vegas, Nevada. Mr. van Ee is an EPA Project Officer who has been
involved in several major EPA Programs during his 15 years with the Agency. He
was involved with the measurement of air pollution in the 1970's, and he became
involved in the development of quality assurance procedures for the calibration
of air pollution instruments. His work with the National Eutrophication Survey
involved the assessment of the water quality of lakes and reservoirs. His
recent duties include the assessment of monitoring systems for the detection of
leaks from underground storage tanks, the development of monitoring strategies
for hazardous waste site assessments, and the development of quality assurance
guidelines for ground-water studies.
Eric N. Koglin is a hydrogeologist with the Aquatic and Subsurface Monitor-
ing Branch at the EPA Environmental Monitoring Systems Laboratory in Las Vegas,
Nevada. He holds a B.S. in geology from Indiana State University and an M.S.
in hydrology from the University of Arizona* Prior to joining EMSL-LV,
Mr. Koglin was an environmental scientist working for U.S. EPA Region 9 in the
Superfund Programs Branch. From 1979 to 1982 he worked for the South Dakota
Geological Survey as a mud rotary drill rig operator and geologist. Since
joining EMSL-LV, he has been involved with a variety of research projects
including the placement of ground-water monitoring wells, ground-water flow and
contaminant transport in fractured rocks, and the application of geographic
information systems to ground-water resource management and contamination
issues.
-597-
-------
"ADA" PRESSURE TEST
Richard C. Peckham
Environmental Protection Agency - Region VI
Dallas, Texas
and
Everett M. Wilson
Environmental Protection Agency - Region VI
Pawhuska, Oklahoma
ABSTRACT
Since December 30, 1984, the Environmental Protection Agency's
Region VI has been implementing the Underground Injection Control (UIC)
program in Osage County, Oklahoma. There are approximately 3500 injec-
tion wells in the county which must demonstrate mechanical integrity
before January 1, 1990 or be plugged. There are a number of wells with
open perforations above the packer which cannot be tested by the standard
annulus pressure test.
A special pressure test was developed to test these wells and the EPA
Robert S. Kerr Laboratory's (RSKERL) "leak test" well was used to test the
principle of the method before using it in the field. The method was
designed on the same principle used to measure water levels by an air line
in water wells. After measuring the fluid level in a well to determine
the height of the water column above the perforations, the pressure
required to depress this column of water to the top of the perforations
is calculated. Nitrogen is added to the annulus until the pressure no
longer increases. If the pressure reached is approximately the same as
that calculated and it remains constant for 30 minutes, after closing the
valve to the nitrogen source, there are no leaks in the casing above the
perforations.
-598-
-------
The same test may be used inside the tubing to demonstrate the integrity
of the tubing and packer.
The method was used to test 13 wells in Osage County in 1986. All
were witnessed by EPA inspectors and the results were conclusive. Five
of the wells passed and eight failed.
Three case histories covering the basic spectrum of conditions that
will be encountered on wells with this type of test are presented as
operational examples of the Ada Pressure Test.
From these examples, it can be demonstrated that the Ada Pressure
Test is a simple inexpensive, reliable and viable test for establishing
the mechanical integrity of a well. In addition, the conditions for
application of the test assures the protection of the USDW.
-599-
-------
INTRODUCTION
In EPA's Region VI, primacy for the UIC program has been
delegated to all five of the states in the Region (Arkansas, Louisiana,
New Mexico, Oklahoma, and Texas).
The Osage Nation consists of the entire county of Osage in Oklahoma
(Figure 1) and as required by the Safe Drinking Water Act of 1974 (PL93-523),
Region VI has direct implementation of the UIC Program on Indian Lands.
Accordingly, the Osage UIC regulations (40 CFR Part 147, Subpart GGG)
were established and became effective December 30, 1984. These regulations
require that all injection wells demonstrate mechanical integrity by
December 30, 1989 and at least once every five years thereafter.
Osage County Oklahoma has approximately 3500 injection wells ranging in
depth from 500 to 3000 feet. In order for these wells to have mechanical
integrity it must be demonstrated that:
(1) There is no significant fluid movement into an underground source
of drinking water (USDW) through vertical channels adjacent to the
well bore, and
(2) there is no significant leak in the casing, tubing or packer.
The demonstration of (1) above can be through any of the following:
(a) Cementing records (need not be reviewed every five years);
(b) Tracer survey (in appropriate hydrogeologic settings; must
be used in conjection with at least one of the other
alternatives);
(c) Temperature log:
(d) Noise log; or
(e) Other tests deemed acceptable by the Regional Administrator.
-600-
-------
This demonstration is usually accomplished through a file review of
Bureau of Indian Affairs (BIA) records. The BIA has regulated the oil
and gas production in Osage County almost since the first discovery of
oil in the County and maintain a comprehensive file on all well compl-
etions dating back to the early 1900's.
The demonstration of (2) above can be through any of the following:
(a) Performance of a pressure test of the casing/tubing
annulus to at least 200 psi, or the pressure specified by
the Regional Administrator, to be repeated thereafter, at
five year intervals, for the life of the well (pressure
tests conducted during well operation shall maintain an
injection/ annulus pressure differential of at least 100
psi through the tubing length); or
(b) Maintaining a positive gauge pressure on the casing/tubing
annulus (filled with liquid) and monitoring the pressure
monthly and reporting of the pressure information annually; or
(c) Radioactive tracer survey; or
(d) for enhanced recovery wells, records of monitoring showing the
absence of significant changes in the relationship between
injection pressure and injection flow rate at the well
head, following an initial pressure test as described by
(a) above; or
(e) Testing or monitoring programs approved by the Regional
Administrator on a case-by-case basis.
Over 90 percent of the injection wells in Osage County demonstrate
the presence or absence of a significant leak in the casing, tubing, or
packer through the standard pressure test (2a. above).
-601-
-------
However, early into the mechanical integrity test (MIT) program, it
was discovered there were some wells which had open perforations above
the packer. The operators and the BIA were reluctant to squeeze off
these perforations, both because of the economics of the remedial work
and the possibility that these zones might once again become commercially
productive or could be used for an injection well in an enhanced recovery
project. Thus, the problem of being able to demonstrate the mechanical
integrity of such wells.
Region VI was not the only region or State to face this problem.
Kansas has similar types of completions in S.E. Kansas, which is adjacent
to Osage County. Their program, having been in operation several
years ahead of the Osage UIC program, had already discovered they had no
practical means of testing these wells. A memorandum, written in February
1984 by Harold Owens of EPA Region VII, suggested the possibility of
pressuring the annulus with air (or gas) and forcing the fluid level
down to the perforations.
In search of a practical and reliable method of testing these wells,
Owens' suggestion was evaluated and it was determined that the principal and
procedure was very similar to the air line method used to measure fluid levels
in some municipal wells with deep water levels. The following method is quoted
from the Missouri Water Well Handbook (Reference 1) and Figure 2 illustrates
the application of this method.
-602-
-------
Air Line: One of the best methods is the air line which
can be installed easily and permanently. The air line is
usually 1/8 or 1/4-inch copper tubing or galvanized pipe,
long enough to extend below the lowest water level to be
measured. The air line may be fastened to the pump bowls
or cylinder and installed with the pump. The pipe must
be airtight and care should be taken in making up all
joints. The vertical length of the air line (A) from the
pressure gauge to the bottom of the line should be measured
carefully at the time of installation.
A pressure gauge is attached to the air line at the surface
with an ordinary tire valve to permit attaching a tire pump
or air compressor hose.
To measure the depth to water at any time, pump air into the
air line until the maximum reading on the gauge is obtained.
This reading is equal to the pressure exerted by the column
of water (B) standing outside of the air line. It is custo-
mary to use an altitude pressure gauge reading directly in
feet of water. If the gauge reads in pounds per square inch,
multiply by 2.31 to convert to feet (or use the conversion
table in Chapter I).
The gauge reading in feet (which equals the height B) is then
subtracted from the total vertical length of air line (A) to
obtain the depth to water (C) in feet below the center of the
gauge.
The procedure was presented to a number of engineers, geologists, and
hydrologists for their opinions. The opinions were equally divided as to
whether it would or would not work in the situation for which it was being
proposed.
Early in 1985, EPA's RSKERL in Ada, Oklahoma had constructed a "leak
test" well for the purpose of providing a facility to develop methods for
testing the integrity of the tubing, casing and packer of injection wells, as
required in EPA's UIC regulations.
-603-
-------
The "leak test" well was designed to represent and to operate like a
typical injection well, with a few exceptions that were added to permit
the simulation of numerous different test conditions. In addition to the
standard surface casing, longstring casing, tubing and packer, the well
is equipped with a second packer and a sliding sleeve on the injection
tubing and a 2 3/8" tubing attached to the outside of the long string
(Figure 3). This rather unorthodoxed configuration permits the control
and monitoring of the desired conditions from the surface.
A more detailed description of this well may be found in a paper by
Thornhill and Benefield (Reference 2) presented at the International
Symposium on Subsurface Injection of Liquid Wastes in New Orleans, March
1986.
Development of the Ada Pressure Test
A test was designed to demonstrate the principle in the RSKERL test well
in December 1985. With the well perforated from 1120 to 1130 feet and
the hole in the longstring casing, leading to the outside tubing, at
1070', the packers were set straddling the hole with the upper packer at
1057' and the lower packer at 1084'.
The test was performed on the tubing in two parts: the first (test
A) with the sliding sleeve open, to represent a leak in the tubing at a
depth of 1070 feet; and the second (test B) with the sliding sleeve
closed, to represent a no leak situation. The fluid level was measured at
360 feet below the land surface with an acoustic fluid level instrument.
This gave us 710 feet of hyrostatic head above the open hole at 1070
feet and 760 feet of head above the top of the perforations at 1120
-604-
-------
feet. Using 2.31 feet* of water per psi, it was calculated that it
would require 307 psi to depress the water level to a depth of 1070 feet
and 329 psi (the formation pressure of the perforated zone) to depress
it to a depth of 1120 feet:
710 760
"273"! = 307 psi "Ol = 329 psi
*Note: 1 psi = 2.31 feet of fresh water was used because the casing had been
filled with fresh water before perforating and the same water was still in the
well.
The following table represents what is theoretically taking place in
the well during the tests as air is added to the tubing.
(1)
Tubing
Gauge
Reading
(psi)
0
100
200
300
307
329
(2)
Depth to
Fluid
Level
(feet)
360
591
822
1053
1070 (Hoi
(3)
Hydrostatic
Head Above
the Perforations
(feet)
760
529
298
67
e) 50
1120(Perforations)0
(4)
psi
P
Fluid
Level
0
100
200
300
307
329
(5)
psi
§
hole
(1070')
307
307
307
307
307
329
(6)
psi
P
perf.
(1120')
329
329
329
329
329
329
With a static fluid level of 360 feet (column 2) below the land surface, the
hydrostatic head above the perforations would be 760 feet (column 3). The
tubing gauge pressure (column 1) and the psi at the fluid level (column 4)
would both be zero. This hydrostatic head would exert 307 psi (column 5) at
the hole (1070 feet of depth) and 329 psi (column 6) at the perforations (1120
feet of depth). As air is added from cylinders of compressed air, the gauge
pressure (column 1) increases and depresses the fluid level (column 2)
-605-
-------
2.31 feet for every psi added, thus reducing the hydrostatic head (col. 3) by a
corresponding amount. The amount of pressure at the gauge (column 1) and
pressure at the fluid level (column 4) remain equal to each other throughout
the procedure. The pressure at the 1070-foot hole (during test A), 307 psi
(column 5), and at the 1120-foot perforations, 329 psi (column 6), remain
constant throughout even though air pressure is being added. The added air
pressure simply replaces the lost hydrostatic pressure caused by depressing
the fluid level.
In test (A), with the sliding sleeve open, when 307 psig of air
has been reached, the fluid level should be at a depth of 1070 feet and you
would not be able to add any more pressure because any addition of air will be
lost through the hole. If the source of air (cylinders) is shut off, and
there are no leaks in the system above the 1070-foot hole, the pressure should
remain 307 psi.
In test (B), with the sliding sleeve closed, you should be able to reach
329 psig before you could not increase pressure by adding more air. At this
point the fluid level should be at the top of the perforations and any additional
air added would be lost into the formation. Again with the air source closed,
the pressure gauge will continue to read 329 psig as long as there are no
leaks in the system.
During test (A), using cylinders of compressed air, air was added to the
tubing until the pressure would no longer increase. This occurred at 300 psig,
a little less than calculated, but considering the accuracy of the acoustic
fluid level instrument we were close to getting the results we were looking for
and once the air source valve was closed, the pressure gauge remained at 300 psig.
Test (B) was a different story. After closing the sliding sleeve, compressed
air was again added to the tubing. An excessive pressure (380 psig) was achieved
-606-
-------
without reaching a maximum. The cylinder valve was closed and the pressure
dropped to a point less than 329 psig. The procedure was repeated several
times and each time the excessive pressure was added, the ensuing pressure
drop became less, but it never did stablize at 329 psig before we ran out of air
cylinders and aborted the test. This indicated that the permeability of the
injection zone was probably extremely low and that even though the added pressure
was more than enough to depress the fluid level to the 1120-foot level, the
formation would not accept the water fast enough and the fluid level was not
as deep as the pressure indicated it should be. Since the well had been filled
with fresh water at the time the well was perforated, creating a pressure
inside the well higher than that of the formation, it was hoped that debri
clogging the perforations rather than a formation with extremely low permeability,
was responsible for the situation.
Even though the tests did not go perfectly as planned, the results showed
that the principle was sound and that a practical, economical, and reliable
test could be developed.
Several months later, the well was acidized and injectivity tests showed a
permeability of 125 md. Test (B) was then successfully run without the problems
encountered on the original test. Nitrogen was substituted for the compressed
air because when used on a formation which contains hydrocarbons, the compressed
air will cause a combustible mixture. Also, it took less cylinders of nitrogen
to achieve the desired pressure and the cost was comparable to that of compressed
ai r.
Development of Procedures
Based on the results of the tests conducted on the RSKERL "leak test" well
and the operational considerations learned through trial and error while performing
these tests, we developed procedures for an annulus pressure test on wells
with open perforations above the packer (the "Ada" Pressure Test). Those
procedures are as follows:
-607-
-------
Test Requirements
1. Must have at least 100 feet of cement immediately above the uppermost
perforations.
2. Must have at least 200 feet of water above the uppermost perforations in
the annul us (must have an accurate static fluid level measurement and know the
depth to uppermost perforations).
3. Must know the specific gravity or total dissolved solids (IDS) of the water
in the annulus.
4. There can be pressure on the tubing, but injection must be shut-in and the
pressure stabilized. The well should be shut-in long enough before the
test for temperatures to stabilize.
5. Must have at least a 500-foot interval between base of USDW and the uppermost
perforations, or a total of at least 100 feet of good shale (not silty or
sandy shale), as determined from an electric log.
6. Annulus water level may not be above the base of USDW unless the casing is
cemented from the land surface through the base of the USDW.
7. With the tubing and packer set at their normal injection depth, (a) tracer
survey must be run through tubing, while injecting, to demonstrate no leaks
in the tubing or packer below the uppermost perforations, or (b) this same
type pressure test can be run in the tubing if: distance between injection
perforations and bottom of tubing is at least 50 feet; water level in tubing
is at least 200 feet above perforations; fluid level is measured; and the
specific gravity or TDS of fluid and depth to perforations are known.
To make sure the test is reliable in demonstrating the protection of the
USDW's, we feel the above requirements are necessary. The rational for each are:
1. The cement above the uppermost perforations is required to prevent
injected fluid from moving out the perforations and up the well bore in the
event of a tubing or packer failure.
-608-
-------
2. The 200 feet of water above the perforations is needed to require
enough added pressure to adequately test the casing. An accurate static
fluid level and known depth to perforations is needed to determine the
hydrostatic head.
3. The specific gravity or dissolved solids of the water is needed to
accurately calculate the pressure needed to depress the fluid level.
4. Shut-in during the test is needed to stabilize the effects of temperature.
5. The depth requirement below the base of USDW is to assure an adequate
confining layer above the upper perforations in the event there is a tubing
or packer leak and injected fluid is injected into the shallower zone with
the upper perforations.
6. If the water level in the annulus is above the base of the USDW, the
casing or surface casing must be cemented from the land surface through the
base of fresh water to protect the USDW in the event of a corrosion hole
developing in the casing.
7. The tubing and packer must be tested independently from the annulus
test because the annulus test does not tell you if you have any leaks in the
tubing, packer, or casing below the uppermost perforations. The reason for
the 50 foot distance between the end of the tubing and injection perforations
is so the pressure differential between the two points is sufficient that
you will be able to recognize (interpret) a packer leak.
Test Procedures
1. Calculate the pressure required to depress the fluid level to top of per-
forations: Sp. Gr. X .433 = Gradient (psi/ft of head) X water column = psig
2. Pressure the annulus (the tubing, if testing the tubing and packer) using
compressed nitrogen cylinders. Be sure the hoses and gauges are rated to
handle the high pressures of the cylinder. The number of cylinders required
-609-
-------
will depend on the volume of the space above the perforations.
3. When pressure at the wellhead will not increase any more (be sure there is
still gas flowing from the cylinder into the well), shut off the valve to the
..cylinder.
4. Record the time and pressure. Monitor the pressure for 30 minutes. Record
pressures after 5, 10, 20 and 30 minutes.
Test Interpretation
1. If you cannot pressure up - indicates a hole in the casing or tubing above
the fluid level.
2. If you cannot obtain the calculated pressure (step 1 of the procedures) -
indicates hole in tubing or casing between the static fluid level and
perforations and a lower pressure in the tubing or formation.
3. If the pressure calculated to force the fluid level down to the open
perforations is obtained and can be held for 30 minutes (no increase or
decrease) - indicates integrity of the tubing and casing above the
perforations.
4. If you reach desired pressure or greater, and the pressure decreases below
the calculated value during the 30-minute hold period - indicates a small
leak above the perforations and a lower pressure in the tubing or formation.
Note: It may take more pressure than calculated before you reach the point
where you cannot increase the pressure any more and, when shut-in, the
pressure will decrease to the calculated value, but this should be within 5
minutes of shut-in and it should stabilize at or near the calculated value.
-610-
-------
In addition to testing wells with perforations above the packer, the
"Ada" pressure test can be used:
1. To test the casing in wells without packers.
2. To test the tubing in wells in which the tubing has been cemented in
the casing.
3. To test the tubing and packer as described under "Test Requirements"
7.(b) above.
It has been suggested by some that we should take into account the weight
of the gas and temperature changes in the gas. To do so, would require a lot
of assumptions and calculations which would complicate the interpretation and
thus reduce its usability in the field. It may be you cannot use the test on
deep wells, but for the shallow wells of Osage County and S.E. Kansas, it
works, it's simple, it's easy to interpret, it's relatively inexpensive, and
it is reliable. If there are errors in this simplicity, we feel that it is on
the conservative side. That is, if the well passes this test, we feel that it
has demonstrated that the casing has no holes above the uppermost perforations
and no leaks in the tubing or packer.
Case Histories or Field Application
Using the above procedures, we began using the "Ada" Pressure Test in the
Osage UIC program in January 1986. During 1986, we tested 13 wells using
this method; 8 failed and 5 passed. The following 3 case histories are examples
of the Ada Test as applied in the field.
-611-
-------
Case history #1 (illustrated in Figure 4) represents a well that has
been re-entered, has no pressure on the tubing and has minimum casing around
the tubing through which the salt water is injected. Since there is 1560 feet
of open hole surrounding the injection string, it would be impossible to apply
pressure on the annul us between the 2 3/8" tubing and 8" casing and be sure
that the entire length of tubing was being adequately tested for integrity.
The Ada Pressure test allows the tubing to the tested internally throughout
it's length and the depth of a leak (if any) to be determined by simple math-
ematical calculations.
The pressure required to push the fluid level from 180 feet to 1836 feet
was calculated to be:
(18361 - 180') x 1.13 S.G. x .433 psi/ft = 810 psig.
The operator reached pressure of 432 psig before running out of Nitrogen.
This amount proved to the sufficient as the pressure began dropping immediately
upon the well being shut-in. Figure 5 illustrates the corresponding relationship
between the pressure, shut-in time and fluid depth in the tubing during the test.
Eighty minutes into the test the pressure reached 370 psig and held steady for
the next 30 minutes indicating through calculations that the fluid level and
corresponding leak was at:
370 psig 7 1.13 S.G. f .433 psi/ft + 180' = 936 feet
It is significant that the operator found the leak at 938 feet
during preparation for remedial work to bring the well into compliance, there-
fore demonstrating the reliability of the test in determining the depth of the
leak. It should further be noted that this test can only determine the existence
and location of the uppermost leak should there be more than one present in the
we! 1.
-612-
-------
Case history #2 (represented in Figure 6) is a well with pressure on the
tubing and known perforations in the casing. The mechanical integrity of the
casing was previously demonstrated by setting the packer above the uppermost
perforations and performing the standard pressure test. The following calcula-
tions were made to determine the pressure requirements to force the fluid
level from surface to the injection zone and demonstrate mechanical integrity
of the tubing and packer:
1024.0 feet x 1.10 S.G. x .433 + 100 psig (tubing pressure) = 588 psig
The tubing was pressured to a maximum of 625 psig with Nitrogen. Upon
shut-in, the tubing pressure decreased immediately to 600 psig and held for
30 minutes. At this point, a fluid level was acquired by an acoustic fluid
level instrument which confirmed that the liquid had been depressed to the top
of the injection interval at 1024 feet. After the test was run it was determined
that the actual specific gravity of the injection fluid was 1.13 instead of the
1.10 value used in the initial calculations. Using a specific gravity of 1.13*
the calculated pressure for the test to push the fluid level down to 1024 feet
is 601 psig. This is a difference of 1 psig as opposed to 12 psig under the
original calculations.
It is desirable that it be standard procedure to shoot a fluid level when
applying this test in the field since there are inherent variables both in the
well construction and fluid properities in the wellbore that can affect the
relationship between the calculated pressure and the actual final test pressure.
However, when specific gravity of 1.10 is used for salt water in the initial
calculation, it can be assumed that any pressure that exceeds the calculated
value and holds steady for 30 minutes is sufficient to demonstrate mechanical
integrity.
-613-
-------
Case history #3 represents one of the first wells to have this test
applied to it and is an excellent example as to the reliability and simplicity
of the test.
Figure 7 details the construction of the well and defines the conditions
in the well at the time the Ada Pressure test was run. Figure 8 illustrates
the pressure-time-fluid level relationship as it evolved during the course of
the test. From this data the following test analysis can be made:
Tubing Test
1. The pressure required to push the fluid level from 150 feet
to 1614 feet was calculated to be:
(1614' - 150') x 1.13 S.6. x .433 psi/ft = 716 psig
2. The maximum pressure achieved was 705 psig. Upon shut-in,
the tubing pressure steadily decreased, as shown on Figure 8,
until the pressure stabilized at 670 psig. The test was repeated
with the same results.
3. The leak in the tubing was calculated to be at:
670 psig i 1.13 S.G.j .433 psi/ft + 150 feet = 1519 feet
Although we know at this stage that the tubing has a leak,
we do not know if the packer is also leaking since the fluid level
is above the packer depth.
Casing Test
1. This is a dual completion well with 35 psig of gas pressure on the
casing annulus. The pressure required to push the fluid level from
562 feet to 748 feet was calculated to be:
(7481- 562') x 1.08 S.G. x .433 psi/ft + 35 psig = 122 psig
-614-
-------
2. The maximum pressure achieved for the test was 180 psig. Five minutes
after shut-in the pressure had dropped to 120 psig and continued to
decrease for the next 30 minutes to 104 psig, at which time the test
was terminated. This indicated a leak in the casing at or above:
(104 psig - 35 psig) 7 1.08 S.G. f .433 psi/ft + 562 feet = 710 feet
3. The operator contended that the annul us contained oil instead of
salt water. If a 39 API gravity of 0.830 S.G. was assumed to be
present in the casing/tubing annulus, the calculated test pressure
would have been:
(748'- 562') x .830 S.G. x .433 psi/ft + 35 psig = 102 psig
At the point where the test was terminated, the inspector
could have attempted to add more pressure to the annulus.
If the operator's contention was correct, he would not have been
able to increase the pressure and when shut-in, the pressure should
remain stabilized at about 102 psig for 30 minutes if there were no
leaks in the casing above the upper perforations. In this case,
however, a fluid level was obtained with an acoustic measuring instru-
ment which showed the fluid level to be above the perforations. When
the specific gravity of the fluid (or fluids) is not known, it is
essential that an acoustic fluid level instrument be used to confirm
the validity of the calculations to the actual test results.
CONCLUSIONS
It is a viable test for demonstrating whether a well has mechanical
integrity. This relatively simple test will not only assure the protection of
the underground sources of drinking water, but will provide the oil industry
with a relatively inexpensive means of demonstrating integrity of their wells.
-615-
-------
A demonstration which, because of their construction (i.e. perforations above the
packer), could not be made through the conventional standard pressure test.
In many respects the "Ada" Pressure Test gives you more than the standard
annulus pressure test. If the well fails the test, it not only tells you that
you have a leak, but it will tell you at what depth the leak is occurring and
whether it is relatively small or large.
The test will not tell you if you have casing leaks below the upper
perforations, but if the well later (after testing) develops a leak in the
tubing or packer, the leaking fluids will go out those casing leaks or the
upper perforations. By requiring the well bore to be cemented above the upper
perforations, this fluid is prevented from moving up the well bore.
The test may not include everything or be as sensitive as some regulators
would like, but:
1. It does meet the requirements of the regulations;
2. It is simple to run and interpret;
3. It is reliable;
4. It is relatively inexpensive; and
5. It does provide a practical test for demonstrating mechanical integrity
for a well which could not otherwise be tested.
REFERENCES
1. Missouri Water Well Drilling Association, "Water Well Handbook", 1959, 199
pages (Edited by Anderson, Keith E.).
2. Thornhill, Jerry T. and Benefield, Bobby G., "Mechanical Integrity Research"
Proceedings of the International Symposium Subsurface Injection of Liquid
Wastes, pp. 241-278, March 3-5, 1986.
-616-
-------
ILLUSTRATIONS
Figures:
1. Map of Region VI Showing Location of Osage County
2. Air Line Method of Measuring Water Levels
3. Sketch of the RSKERL "Leak Test" Well
4. Case History #1: Well bore Schematic
5. Case History #1: Pressure - Time - Fluid Depth Plot
6. Case History #2: Wellbore Schematic
7. Case History #3: Wellbore Schematic
8. Case History #3: Pressure - Time - Fluid Level Relationships
-617-
-------
MAP OF EPA REGION VI
Osage County, Oklahoma
FIGURE 1
-------
SCHEMATIC OF AIRLINE METHOD FOR MEASURING FLUID LEVELS
Pressure Gauge
Compressor \ ^
— S~
i
I
\
't
C
i
•MM
/
^
£
t \
•*
r
•09
t
i
r
*• C
•
Air Line ^
Tubing -
Electric Pump -
J*
~s*
j>
N^X"
JT
*-»-
jr-**-
N
J
-1
ooo
I —
-
t
C
«
a
X
-V»X
Tramr
"* £— Fluid Level
_ Electric Line
^ «t-»Casing
Perforations
FIGURE 2
-619-
-------
RSKERL "Leak Test" Well
680*
.•,,.,.,,..,,.....,..,.... 905'
Bifiji liiiii:
vx:w:m:? Saffiffiffi g35,
1057' Deptn of
upper packer
Cement
1070'
1084' Deptft of
lower packer
1. Baker nodel *C-1" landen Tmsion Pecker
2. 2 J/a" luolng
3. Wur nodel V Sliding Slew
4. Mur Hod«l *R' Profile Hippie
5. Mur Hodel 'Ad-l* Tension Packer
6. 2 VT tubing
7. Befctr flodti f Profile Hippie
I. leker Hodel f Profile Hippie
9. » 1/2" Long suing
INJECTION ZONE
FIGURE 3
-620-
-------
CASE HISTORY II
Wellbore Schematic
1
V.-.i
p
<£
111
pit
^•a
i
8
»*i
m
fe:
85
^
fc
<«-
FIGURE 4
-621-
i
&
^^i
^•^ — ^ 140' TD of 8" casing
—.180' Static Fluid Level in Tubing
* ,£—,140' - 1600' Open Hole
— 2 3/8" Tubing
•*"~X— 1631' Top of Cement
"Z 1800' Packer
**TL_»1836' to 1600' 5h" Casing
"2L-. 1872' TD
-------
(psig)
PRESSURE
434§
432
430
428
426
424
422
420
418
416
414
412
410
408
406
404
402
400
398
396
394
392
390
388
386
384
382
380
378
376
374
372
370
368
366
364
362
360
0 CASE HISTORY #1
o
o o
Pressure-Time-Fluid Depth Plot
o
o
0
0
0
0
0
0
0
0
o
0
o o o o o o <
1063
1059
1038
1018
998 _
*j
0)
0)
«w
£
E-«
O,
Ed
977 °
Q
M
D
J
fc,
957
936
916
5 15 25 35 45 55 65 75 85 95 105
10 20 30 40 50 60 70 80 90 100 110
TIME (minutes)
Figure
-------
PASE HISTORY 12
Wellbore Schematic
#
r-
t~
i •
(:
n i » 100 psig on Tubing
Fu
, Fluid Level at Surface
23/8" Tubing
1
o
480' Top of Perforations
610' Top of Perforations
%%-*-
958' Packer
1024' Top of Perforations
7" Casing
15391 TD
FIGURE 6
-623-
-------
CASE HISTORY #3
Wellbore Schematic
C'"
c*.
t-i
§
-•» 35 psig Casing Pressure
\::i
<£
150' Static Fluid Level
in Tubing
-5
562' Static Fluid Level in
Casing
748' Top of Perforations
2 3/8" Tubing
Casing
Sliding Sleeve
1562' Packer
1580' Bottom of Tubing
1614' Top of Perforations
1648' TD
FIGURE 7
-624-
-------
Tubing Test
Case History #3
PRESSURE-TIME-FLUID LEVEL RELATIONSHIP
Calculated
Calculated
Time
(minutes )
0
5
10
20
25
30
Test Pressure
(PSIG)
705
695
686
674
670
670
Fluid Level
Above Perforations
( Feet )
23
44
62
86
95
95
Fluid Level
Below Land Surface
(Feet)
1591
1570
1552
1528
1519
1519
Casing Test
Calculated
Time
(minutes )
0
5
10
15
17
23
25
35
Test Pressure
(PSIG)
180
120
117.
114
112
108
108
104
Fluid Level
Above Perforations
(Feet)
(maximum pressure
4
11
17
21
30
30
38
V«-*-l- T_l_» -1_CA |_^U
Fluid Level
Below Land Surface
( Feet )
achieved for test)
744
737
731
727
718
718
710
Figure 8
-625-
------- |