«*EPA
United States
Environmental Protection
Agency
Industrial Environmental Research
Laboratory
Cincinnati OH 45268
EPA-600/7-80-024
February 1980
Research and Development
Surface
Containment for
Geothermal Brines
Interagency
Energy/Environment
R&D Program
Report
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND DEVELOPMENT series. Reports in this series result from the
effort funded under the 17-agency Federal Energy/Environment Research and
Development Program. These studies relate to EPA's mission to protect the public
health and welfare from adverse effects of pollutants associated with energy sys-
tems. The goal of the Program is to assure the rapid development of domestic
energy supplies in an environmentally-compatible manner by providing the nec-
essary environmental data and control technology. Investigations include analy-
ses of the transport of energy-related pollutants and their health and ecological
effects; assessments of, and development of, control technologies for energy
systems; and integrated assessments of a wide range of energy-related environ-
mental issues.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/7-80-024
February 1980
SURFACE CONTAINMENT FOR
GEOTHERMAL BRINES
by
R. Sung, W. Murphy, J. Reitzel
L. Leventhal, W. Goodwin, L. Friedman
TRW, Inc.
Redondo Beach, California 90178
Contract No. 68-03-2560
Project Officer
Robert P. Hartley
Power Technology and Conservation Branch
Industrial Environmental Research Laboratory
Cincinnati, Ohio 45268
INDUSTRIAL ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
Cincinnati, Ohio 45268
-------
DISCLAIMER
This report has been reviewed by the Industrial Environmental Research
Laboratory, U.S. Environmental Protection Agency, and approved for publication.
Approval does not signify that the contents necessarily reflect the views and
policies of the U.S. Environmental Protection Agency, nor does mention of trade
names or commercial products constitute endorsement or recommendation for use.
ii
-------
FOREWORD
When energy and material resources are extracted, processed, converted,
and used, the related pollutional impacts on our environment and even on our
health often require that new and increasingly more efficient pollution
control methods be used. The Industrial Environmental Research Laboratory-
Cincinnati (lERL-Ci) assists in developing and demonstrating new and improved
methodologies that will meet these needs both efficiently and economically.
Heat and chemical constituents in geothermal brines could be very damag-
ing in an uncontrolled release to the surface environment. This report pro-
vides a preliminary evaluation of the probability of unplanned brine releases
from geothermal power plant operations and describes measures that may be
used to contain such releases.
Further information on the subjects of this report can be obtained from
the Power Technology and Conservation Branch, Industrial Environmental
Research Laboratory, Cincinnati, Ohio 45268.
David G. Stephan
Di rector
Industrial Environmental Research Laboratory
Cincinnati
iii
-------
ABSTRACT
Planning is currently underway for approximately ten power plants in the
United States, utilizing liquid-dominated geothermal resources. The extrac-
tion of heat from these geothermal brines will inevitably produce a large
quantity of spent fluid that requires safe and economical disposal. Because
of the large volume and potentially toxic nature of the spent fluid, environ-
mental degradation may occur from rupture of the transport mechanism. The
objective of this study has been to determine measures to minimize environ-
mental damage from unplanned or accidental surface release of geothermal brine.
These measures primarily involve methods of containing releases in time and
area.
Six types of geothermal energy conversion systems were considered initi-
ally and three were selected for detailed analysis. These are:
• a 360°F (182°C) double flash system;
t a 360°F (182°C) brine binary system;
• a 550°F (288°C) multiflash binary system.
Flow rate and component sizes in these systems were selected for consistency
with the requirements of a 50 MWe power plant. Maximum brine flow rates range
from 4000 1pm (1000 gpm) to 53,000 1pm (14,000 gpm) in the well-to-plant pip-
ing. Accordingly, the rate of flow in an inadvertent release could range from
a trickle to approximately 53,000 1pm (14,000 gpm), depending upon the com-
ponent failure mode that causes the release.
The probability of an inadvertent brine release ranges from about one in
500 that a large spill will occur during a 40 year plant life to a virtual
certainty that trickle spills will occur during the 40 years.
There are three means of containing an inadvertent release and minimizing
the resultant environmental damage: minimizing the potential for component
failure, limiting release duration, and limiting the affected area.
Minimizing the potential for component failure is a matter of:
• using materials of good engineering design, incorporating component
redundancy, and minimizing the number of components;
• proper maintenance and component replacement; and
• security to prevent wanton damage.
IV
-------
Release time can be minimized with an adequate brine shut-off system.
Automatic flow or pressure-actuated shut-off systems could limit brine release
to a few hundred or thousand gallons, but such systems are complex and expen-
sive. Given the relatively low probability of a major release, a shut-off
system requiring manual closure of appropriate valves appears more attractive.
If an adequate alarm system is present and free movement of personnel is as-
sured through the plant in event of a release, alert and well-trained plant
operators could manually close wellhead valves in an estimated maximum time
of two hours. Accordingly, a maximum release could spill approximately 6,300
cubic meters (5 acre-feet) of brine.
Ponding can limit the areal extent of a brine release. Environmental
damage would be minimized by locating the pond within the plant perimeter.
Since significant spills to this pond will be rare and only temporary, unused
portions of the plant site and non-critical areas such as parking lots could
serve as locales. Since a 50 MWe geothermal plant will occupy approximately
40,000 square .meters (10 acres), 6,300 cubic meters (5 acre-feet) of brine
could be accommodated in most circumstances. Accordingly, areal containment
will involve constructing a dike with its location and dimensions dependent
upon available plant area and the grading profiles of the plant site.
-------
CONTENTS
Foreword iii
Abstract iv
Figures viii
Tables ix
1. Introduction 1
2. Conclusions 2
3. Recommendations 4
4. Energy Conversion System and Component Characterization 5
Energy Conversion System Descriptions 5
Potential Brine Loss Points 15
5. Fluid Release Analysis 19
Fluid Release Causes 19
Brine Release Rates 26
Failure Mode Analysis 27
Failure Rate Analysis 31
6. Containment Methods and Fluid Release Causes 34
General Containment Evaluation 34
Specific Containment Evaluation 37
References 43
vii
-------
FIGURES
Number
1 Double flash energy conversion system diagram - system 1 8
2 Well layout - system 1 (flash) 9
3 Brine binary energy conversion system diagram - system 2 11
4 Well layout - system 2 (binary) 12
5 Multiple flash binary energy conversion system diagram - system 3. 13
6 Well layout - system 3 (multiflash/binary) 14
7 Potential brine loss points - system 1 16
8 Potential brine loss points - system 2 16
9 Potential brine loss points - system 3 17
vm
-------
TABLES
Number Page
1 Numbers of Brine Transporting Components 18
2 Corrosion Rates (MM/YR) of Various Alloys Versus Brine Composition 23
3 Brine Flow Rates 26
4 Failure Rate Data 28
5 Geothermal Brine Release, Failure Modes and Effects 30
6 Geothermal Components and Failure Causes 31
7 Brine Release Probabilities 33
8 Surface Brine Release Containment Methods 35
IX
-------
SECTION 1
INTRODUCTION
The major commercial development in geothermal energy in the United
States is The Geysers dry steam field, operated by Pacific Gas and Electric
Company for electric power production. Commercial development of the much
more prevalent liquid-dominated geothermal resources can be expected within
the next five to ten years. Approximately ten power plants using liquid-
dominated resources are now under construction or being planned.
The extraction of heat from geothermal fluids will inevitably produce a
large quantity (several million gallons a day at each site) of spent fluid.
The most likely method of liquid disposal will be subsurface injection. Be-
cause of the large flow and potentially toxic nature of the spent brine,
environmental degradation may result from rupture of the brine distribution
system.
The objective of this study is to determine the measures necessary to
minimize environmental damage from unplanned or accidental surface releases of
geothermal brine. These measures primarily involve methods of preventing re-
leases and containing those that do occur. Particular emphasis has been
placed on power generation systems expected to be operational within the next
five years.
-------
SECTION 2
CONCLUSIONS
The conclusions that can be drawn from this study follow:
• In their chemical make-up, geothermal liquids in the United States
range widely from nearly pure water to high salinity brines. Dilution
of escaped geothermal liquid may occur by mixing with natural waters
to limit the extent of damage, (although the damaged area may sometimes
be increased), depending on the fluid characteristics and the local
hydrological conditions.
• The conversion systems that are most likely to be used in the next five
to ten years for generating electricity from geothermal liquids are
flash and binary systems or a combination of these. In these systems,
the geothermal liquid will be brought up from wells, passed through
flash chambers or shell-and-tube heat exchangers, and returned to the
subsurface by injection wells. The plant components that handle brine
above ground will be comparatively few, simple, and of standard design.
Independent units producing about 50 to 100 MWe from brine flows of
114,000 1/min (30,000 gpm) or less, depending on the fluid temperature,
will be common. Escape rates in the event of a serious accident will
be limited to about 53,000 1/min (14,000 gpm), if the brine flow is
split into two parallel flows for engineering or safety reasons.
• Using individual component failure rates taken from nuclear safety
studies, it is estimated that the overall probability of one large
brine spill during 40 years operation of a 50 MW plant is about 1 in
500, or 0.2 percent. (The probability of two or more spills is very
much less). A major portion of this risk is from failures in flash
chambers, steam separators, heat exchangers or flow lines; wellhead
failures are not included. The estimated probability that one wellhead
will release brine during 40 years operation ranges downward from 0.8
percent, depending on the number of wells needed to produce 50 MWe;
the probability decreases with decreasing number of wells.
a The successive lines of defense against brine spills are:
- Good design and materials selection in building the plant, together
with systematic monitoring and maintenance.
- Systems to limit the duration of spills with automatic pressure-drop
alarms and shut-off or bypass valves utilized where possible.
-------
Dikes to contain spills, designed to keep the largest expected spill
within a small area inside the plant perimeter. Moderate spills will
obviously be contained, if containment facilities are designed for
the largest potential spill. Impervious ponds with clay, bentonite,
or other lining material should be used for containment of spills
within a plant facility.
-------
SECTION 3
RECOMMENDATIONS
Recommendations arising from this study are as follows:
t Geothermal plant designers and operators should be caused to recognize
the possibility of environmental damage resulting from a brine spill,
and means of minimizing this damage should be routinely included in
plant design and operations. Guidelines for materials, equipment,
maintenance and special containment measures should be established
to secure a uniformly low risk of spill-induced environmental damage.
Given the wide variation in fluid chemistry at different sites and
the generally small probability of large spills resulting from normal
operations, numerical standards may not be appropriate.
• Given the lack of geothermal plant operating experience in the United
States, estimating failure rates for geothermal components has
required reliance upon nuclear data. Although these data are con-
sidered adequate for this preliminary estimate, it is recommended
that confirmation of these estimates be made as early as possible
from actual geothermal operating experience.
• With the lack of operating experience in geothermal hot water plants
in the United States, data for establishing proper plant maintenance
procedures, schedules and costs are not readily available. In
particular, the workover requirements for long-term operation of
geothermal wells are not defined. Maintenance data are available in
other countries with data on the Cerro Prieto development in Mexico
being perhaps the most available. Collecting, analyzing and applying
these data to United States conditions are suggested.
• Most working fluids favored for use in geothermal binary conversion
systems are flammable when mixed with air. Accordingly, an inad-
vertent spill of these materials could cause significant environ-
mental damage as well as presenting a severe safety hazard.
Additional investigation of the consequences of releasing working
fluid into the environment is suggested.
-------
SECTION 4
ENERGY CONVERSION SYSTEM AND COMPONENT CHARACTERIZATION
ENERGY CONVERSION SYSTEM DESCRIPTIONS
Many energy conversion system configurations can be applied to geothermal
energy and the optimum configuration for any single generating plant depends
upon the characteristics of the geothermal fluid, the cost, and the degree of
experimentation that the plant owner is willing to include in the facility.
The conversion systems that are now finding application fall into two general
types, flash and binary, or combinations of these types. The detailed work-
ings of these systems have been described in detail in the geothermal litera-
ture.
A flash system involves bringing geothermal brine, which is residing in
the subsurface reservoir at a pressure above saturation, to the surface where-
upon the pressure is lowered and steam is separated from the brine. The steam
is then passed through a turbine coupled to an electric generator. Multiple
flash stages may be used wherein steam is separated more than once from the
brine residual and expanded through lower pressure turbine stages.
A binary system involves transferring thermal energy from the geothermal
brine to a working fluid, commonly isobutane, in a series of heat exchangers.
High pressure working fluid vapor is then expanded through a turbine/generator.
Binary systems may be designed to maximize conversion efficiency by including
economizers and regenerators. Both flash and binary systems condense the
vapor (steam or working fluid) at the turbine exhaust and both systems
usually inject the spent brine back into the subsurface.
There are no full-scale operating geothermal plants yet in the United
States that are based on liquid geothermal resources. The most specific and
realistic designs currently available for United States conditions are con-
ceptual, including some detail at the component level, for plants at particular
sites, using particular geothermal fluids.
The optimum energy conversion process and equipment for a given geothermal
field will be closely related to the characteristics of the geothermal fluid
The most important parameters are temperature, pressure and concentration of
dissolved solids. Three conversion systems have been selected for detailed
analysis in that they have received enough attention to make them likely can-
didates for power plants within the next 5 to 10 years. The selected systems
are:
-------
System 1. Double flash, 180°C (36QOF) reservoir temperature.
System 2. Brine binary, 180°C (360°F) reservoir temperature.
System 3. Multiflash binary, 290°C (550°F) reservoir temperature.
Any type of conversion system will include wellhead assemblies on the
production and injection wells. Each assembly is comprised of a concrete
structure, usually below ground level, containing a complement of valves and
pipes. The assembly allows well shutdown, redirection of the fluid flow,
restricting the flow, and the insertion of probes into the well. The assembly
is subject to stresses caused by thermal expansion and contraction as the geo-
thermal wells are flowed and shut down. Excessive thermal cycling can cause
ruptures in the valves and piping assembly.
The well layouts, piping layouts, and component arrangements for all of
these systems have many similarities with the main differences being in the
number of wells required for a given power output. In each of the systems,
there is an appropriately valved bypass line from the production wells to the
injection wells. The valves are opened in case of a power plant failure or
the flow rate being too high for the power demanded of the system. As typical,
a 50 MWe plant has been selected for analysis for a number of reasons:
• it represents the installation of minimum size that would be of interest
to a large utility;
• design work on such units has been widely reported in the literature;
• the components and equipment are, in general, commercially available;
and
• estimates can be made on failures and failure rates and prevention
measures.
As will be discussed further below, this analysis assumes a production
well flow of 3785 1/min (1000 gpm) from each well and a production to injec-
tion well ratio of two to one. Accordingly, each injection well flow is 7570
1/min (2000 gpm). Reference 1 contains the rationale for these flow rates and
for the number of wells per system.
All production and injection wellheads are assumed to be located in a
drilling island. The wells are bottomed on an approximate 25-acre spacing
with producers and injectors located in a pattern approximating the five-spot
pattern commonly used in petroleum development. Drilling from a drilling
island is costlier than in a dispersed field because wells are slant-drilled.
Slant drilling, however, offers definite advantages, such as minimizing the
overland piping and allowing containment of a brine spill in a small area.
The economic trade-offs of the two alternatives are not obvious and were not
investigated in this study.
The geothermal field, as assumed here, will be developed around a central-
ly located power plant and energy center. Production and injection wellheads
-------
will be manifolded into aboveground brine supply and injection piping systems.
Other components such as double block valves, bleed valves, bypasses, relief
valves, expansion loops, pipe anchors and pipe supports, containment trenches,
curbs, pits and sumps are included as good design practice to minimize thermal
stresses, pressure surges, mechanical or material failures, or to allow access
to the system during maintenance operations.
System 1. Double Flash
A double (two-stage) flash power generation system is shown schematically
in Figure 1. The diagram shows flash and turbine redundancy and a direct con-
tact condenser. A similar system could be constructed without redundancy and/
or with a surface condenser.
System 1 includes distribution lines, separators, demisters, turbines,
condensers, cooling towers and generators. The possibility of failure and con-
sequent brine spill exists in all of the components in the brine loop and in
the connecting pipes, joints, and valves. The condensers, while not in the
brine loop, merit special attention because non-condensible gases, that are
separated with the steam from the geothermal brine, must be removed to prevent
back pressure build-up and a deterioration in turbine performance. The non-
condensible gas is mostly C02» a relatively innocuous compound. However,
other gases, making up to 10 percent of the total, include toxic compounds
such as H2S, NHo, and SO?. Currently, the most common recognized problem is
H2S, a major pollutant that even in small amounts can cause serious impacts,
including deterioration of habitability (from noxious odors) in the area and
corrosion of metal surfaces.
The steam is usually separated in vessels in which the well fluid is in-
jected tangentially to the wall. The resulting centrifugal action improves
the separation efficiency with the steam moving to the center. Since the well
fluid usually carries solid particles, the points of impact on the wall must
be protected from abrasion that could cause perforation.
System 1 is assumed to require 28 production wells and 14 injection wells
with a layout as shown in Figure 2. The brine-carrying components of System 1
are as follows:
t Two first stage flash tanks with valves, pipes, and fittings necessary
for bringing brine from the production wells, conducting brine to the
second stage flash tanks, and conducting steam to the demister and
turbine.
0 Two second stage flash tanks with the valves, pipes, and fittings neces-
sary for bringing brine from the first stage flash tanks, conducting
brine to injection lines, and conducting steam to the demister and low-
pressure turbine.
• Three branch lines with pumps, valves, and fittings required to carry
brine from the second stage flash tanks to the main line to the injection
wells. In this system and in Systems 2 and 3, one of the fluid convey-
ance pumps can handle one-half of the total flow and one is on standby.
-------
00
Figure 1. Double flash energy conversion system diagram - system 1.
-------
24° DEVIATION
BOUNDARY
SECTION LINES
(TYPICAL)
-V
/[
1
,
V
' _/
>
/
X
5280' j
A
\
~\
^
\
-T
y
^v
^'
\
_^-—-"
^•^ ^
~ 1
^
^~
^
B /
*\ '''' x
P:
/ \ ^
^^
/
^
•^
_— ^
\
)
/
•
^
>— 48° DEVIATION BOUNDARY
\£
'^
X
^\
,\
\
jr "
/
/
/ ^» WELLh
V W " (ENER
„ PRODU
* "BOTTO
. , INJECT
r "BOTTO
EAD LOCATION
3Y SUPPLY CENTER)
SING WELL
M HOLE LOCATION
ION WELL
M HOLE LOCATION
TO OTHER VI ELLS
8" VALVE
3" VALVE
1/2" NEEDLE VALVE
-TO OTHER WELLS
- 8" VALVE
- 3" VALVE
-1/2" NEEDLE VALVE
PLAN VIEW
PLAN VIEW
1/2" NEEDLE VALVE
. 3" VALVE
8' TEE
- 8" VALVE
- SPOOL, HANGER
. 12" VALVE
-1/2" NEEDLE VALVF
- 3" VALVE
- 8" TEE
• SPOOL ADAPTER
- 8" VALVE
(28 PRODUCERS - 14 INJECTORS)
DIRECTIONAL DRILLED BOTTOM HOLE LOCATIONS
FROM
POWER
PLANT
If-
EXPANSION
LOOP
(TYPICAL)
"A" "A"
20'-0'_
(TYP.)
ELEVATION B-B
PRODUCTION WELL
ELEVATION A-A
INJECTION WELL
"B" "B"
TO
POWER-
PLANT
ii mi i
SURFACE WELLHEAD LOCATIONS
Figure 2. Well layout - system 1 (flash).
-------
The cooling water loop, including the cooling tower and condensers with
associated components, does not carry brine. Among the noncondensable gases
removed from the condenser, H2S is a serious pollutant. A two-stage ejector,
and auxiliary vacuum pump and an l^S scrubber with steam, vapor and water
lines, valves and fittings are the usual requirements for handling these gases.
System 2. Brine Binary
A binary system (Figure 3), System 2, involves transferring thermal energy
from the geothermal brine to a working fluid, commonly isobutane, in a series
of heat exchangers. High pressure working fluid vapor is then expanded through
a turbine/generator and recirculated. Both the brine and the working fluid
flow in closed systems. The brine is maintained under pressure from production
through injection so that flashing is prevented.
Binary systems are subject to the possibility of brine loop component
failure and consequent brine spill in the same manner as flash systems. Here,
however, the heat exchangers warrant special attention because, when the geo-
thermal brine contains large amounts of dissolved solids, deposition (scaling)
will occur in the heat exchanger tubes. To minimize this, acid (hydrochloric
is most commonly used) may be added to the brine, but it may be corrosive to
the materials used in the heat exchangers and pipes. Material must be resis-
tant to corrosion from both brine chemicals and additives.
The favored working fluids for binary systems are flammable when mixed
with air. These fluids, if inadvertently released, would pose a major hazard
both to the environment and to the safety of personnel and equipment.
System 2 is assumed to require 25 production and 13 injection wells
(Figure 4). The wells house downhole pumps to pressurize the brine and pre-
vent flashing in either the supply pipes or the heat exchangers. Only sen-
sible heat is removed from the brine which stays in the liquid phase throughout
the process as illustrated in Figure 3. After exiting the preheater/economizer
units, the brine is discharged to injection wells. Again, a drilling island
is assumed.
t Two boiler-superheaters with the associated valves, pipes, and fittings.
As assumed here, the boiler-superheaters and the preheater-economizers
are tube-and-shell heat exchangers with brine flowing through the tubes
and working fluids flowing through the shell.
• Two preheater-economizers with valves, pipes, and fittings to carry the
brine from the boiler-superheater and to the injection lines.
• Three branch lines, two from economizers and one from brine main, with
pumps, valves, and fittings to transport the brine to the main injec-
tion line and wells.
Nearly all pollutants will stay in the brine as it passes through the
system and will be injected. Noncondensable gases cannot be released.
10
-------
LOW
-, PRESSURE
TURBINE
HIGH PRESSURE TURBINE
SURFACE CONDENSER
BRINE
g
UJ
ct
CO
-•i-^-Cc
GENERATOR
COOLING WATER
-•-FROM COOLING TOWER
COOLING TOWER
BOILER/
SUPERHEATEF
PREHEATER,
ECONOMIZER
BOILER/
SUPERHEATED
PREHEATER/
ECONOMIZER
GEOTHERMAL BRINE
WORKING FLUID
CIUDOWNHOLE PUMPS
-/WORKING FLUIDN
I RECEIVER /
O-j REINJECTION PUMPS
I
TO REINJECTION
I
DRILLING ISLAND
Figure 3. Brine binary energy conversion system diagram - system 2,
-------
5280
«° DEVIATION BOUNDARY
FROM
POWER •
PLANT
TO
POWER
PLANT
TO
POWER
PLANT
20'-0'__
r
111111r
EXPANSION
LOOP (TYPI
5" 5:
St
SECTION LINES
(TYPICAL)
• WELLHEAD LOCATION
~ (ENEfjGY SUPPLY CENTER)
.— PRODUCING WELL
BOTTOM HOLE LOCATION
*— INJECTION WELL
BOTTOM HOLE LOCATION
I I II 11
SURFACE WELLHEAD LOCATIONS
DIRECTIONAL DRILLED BOTTOM HOLE LOCATIONS
(25 PRODUCERS - 13 INJECTORS
Figure 4. Well layout - system 2 (binary).
System 3. Multiflash Binary
The multiflash binary system (Figure 5), of System 3, is similar to the
brine binary system except that the geothermal fluid is flashed prior to heat
exchange with the secondary fluid. The intent is to prevent scaling in the
heat exchangers that would occur if the brine were passed directly through
the exchangers.
System 3 is designed for use in the Salton Sea geothermal area of Cali-
fornia, a resource in which temperature and salinity are very high. While
the characteristics of this resource may be unique in this country and the
characteristics of the applicable conversion system may also be unique, the
great power generating potential of the area attracts special attention and
warrants inclusion of the system in this study.
In System 3, as shown in Figure 5, the brine vaporizes and condenses in a
series of chambers operating at successively lower pressures. The produced
steam condenses over tubes located in the upper part of the chamber and the
working fluid flows through the tubes absorbing the heat of condensation.
Accordingly, the brine remains at the constant temperature corresponding to
the pressure in each chamber. The condensate falls and mixes with the brine.
The net effect is that the brine salinity remains nearly constant throughout
the process and the brine does not come in contact with the heat transfer
surfaces.
System 3 is assumed to require 8 production wells and 4 injection wells
as shown in Figure 6.
12
-------
•VBH-L*
Figure 5. Multiple flash binary energy conversion system diagram - system 3.
-------
24° DEVIATION
BOUNDARY
SECTION LINES
(TYPICAL) "
5280'
• 48° DEVIATION BOUNDARY
FROM
POWER
PLANT
TO
POWER-
PLANT
- EXPANSION
LOOP (TYP.)
A" * >—I
20MT
(TYP.)-H-
•H-H
• WELLHEAD LOCATION
~ (ENERGY SUPPLY CENTER)
• _ PRODUCING WELL
BOTTOM HOLE LOCATION
*— INJECTION WELL
BOHOM HOLE LOCATION
SURFACE WELLHEAD LOCATIONS
DIRECTIONAL DRILLED BOTTOM HOLE LOCATIONS
(8 PRODUCERS - 4 INJECTORS)
Figure 6. Well layout - system 3 (multiflash/binary).
Specific brine-carrying components of System 3 are:
• One steam separator-heat exchanger with the valves, pipes, and fittings
necessary to transport brine from the production wells, conduct brine
to the multistage flash units, and conduct high pressure steam to the
jet ejector. A high pressure steam line is included.
• Four multistage flash tanks operating in series, with valves, pipes,
and fittings necessary to conduct brine from the separator/heat ex-
changer, to interconnect the tanks for the transfer of brine and steam,
and to remove brine and low pressure steam from the last stage tank.
The brine is moved to the injection lines, with a small fraction first
going to an H2S scrubber. Steam flow rates are small when compared to
the total brine flow since most of the steam is condensed and mixed
with the brine in each of the tanks.
• Three branch lines with the valves and pumps required to carry brine to
the injection wells.
14
-------
POTENTIAL BRINE LOSS POINTS
Virtually every component in the system, with the exceptions noted below,
is a potential brine loss point in that there is some possibility of failure
(albeit very small) associated with each of these components. The most sig-
nificant potential loss points are pipe joints, either sleeve fittings or
flanges, and points where the brine flow impinges directly on a metal surface,
such as pipe elbows and tees. Since the input and output of every component
requires a pipe joint and since lengthy straight runs of pipe are rare in a
conversion system, the potential for loss exists at many points in the system.
Components that are not potential loss points are:
• Downhole pumps - failure will cause a reduction in produced brine in
System 2 but, in all reasonable failure modes, effects will be con-
tained in the wellbore with no brine escape.
• Heat exchanger tubes - System 2 and 3 consider the brine to pass
through the tubes of conventional shell-and-tube heat exchangers. The
working fluid passes through the shell. Accordingly, in the event of
tube failure, the brine will be contained by the shell and will not
escape into the environment. Of course, mixing the brine and working
fluid would necessitate shutting down that portion of the system, but
brine spillage would not ordinarily occur. However, for conservatism,
heat exchangers are included in the failure rates discussed in the
next section.
t Brine treatment - pretreatment of spent brine is sometimes required
prior to injection. This would constitute a potential brine loss
point and subsequent containment may be required. However, the need
for brine treatment is dependent largely on the salinity of the geo-
thermal fluid under investigation. Geothermal fluids with salinity
less than 5000 mg/1 will most likely not require any pretreatment
prior to subsurface injection.
Potential loss points and associated interconnections for the three sys-
tems of interest are indicated in the networks of Figures 7, 8, and 9. Table 1
shows the length of pipe and the numbers of valves, elbows, and flanges re-
quired by the three systems analyzed. System 1 contains the greatest number of
brine-carrying components and System 3 contains the least.
The cooling water loop, including the cqoling tower and condensers with
associated components, does not carry brine. Among the noncondensable gases
removed from the condenser, H£$ is a serious pollutant. A two-stage ejector,
and auxiliary vacuum pump and an H2S scrubber with steam, vapor and water
lines, valves and fittings are the usual requirements for handling these gases.
15
-------
STEAM/NONCONDENSABLE
VENT
Figure 7. Potential brine loss points - system 1.
Figure 8. Potential brine loss points - System 2,
16
-------
VENT
CQ
C
tt>
vo
-o
o
ft)
Q>
o-
-S
ro
o
in
-a
o
C+
(/)
I
OJ
-------
TABLE 1. NUMBERS OF BRINE TRANSPORTING COMPONENTS
No
No
Item
. of production wells
. of injection wells
System 1
28
14
System 2
25
13
System 3
8
4
Feet of piping
No
No
No
No
No
8"
16"
30"
34"
. of valves
1/2"
3"
8"
12"
. of elbows
8"
16"
30"
34"
. of flanges
3"
8"
12"
20"
. of tees
8"
. of stub-ins
3" to 12"
8" to 16" - 34"
1680
1050
168
168
112
42
126
12
374
84
84
42
42
168
42
1520
980
152
152
101
38
114
342
76
76
38
38
152
38
480
140
220
48
48
32
12
36
4
4
108
64
24
12
12
48
12
18
-------
SECTION 5
FLUID RELEASE ANALYSIS
FLUID RELEASE CAUSES
The various potential causes of fluid release are described in the follow-
ing, together with the means of minimizing the chances of release:
Blowouts
Geothermal production wells can blow out, i.e., flow uncontrolled, when
proper drilling, casing, or completion practices are not followed. There are
several reasons for blowouts as described in Reference 2.
• Unconsolidated earth blowouts may occur when the well penetrates soft,
unconsolidated rock (identified as "Punky" in the reference), reservoir
pressures exceed hydrostatic, and surface casing is not long enough.
The pressure is sufficient to force the brine through the rock to dis-
charge on the surface at some unpredictable point away from the wellhead,
• Landslide blowouts may occur when the well is located in unstable soil.
The soil slips, the casing shears, and the brine reaches the surface,
either along the slippage zone or some other route. Here again, the
brine may discharge on the surface at some unpredictable point away
from the wellhead.
• Wellhead blowouts may occur when the wellhead assembly is incorrectly
installed. Vibration and temperature cycling may fracture the wellhead
assembly or crack surface casing cement, allowing brine access to the
surface through the wellhead.
• Improper downhole casing and/or cementing blowouts may occur when
defective casing is used and/or the casing cement is permeable. Brine
may infiltrate the cement either through defects in the casing or from
below the casing shoe, reaching the surface through the permeable
cement or the surrounding rock. The brine may discharge on the surface
either at the wellhead or at some unpredictable point away from the
wellhead.
• Seismicity may cause blowouts where the well crosses an active fault.
The brine discharge rate in a blowout is the flow rate generated by down-
hole pressure and flashing in the wellbore. Geothermal wells in the East Mesa
area of California typically flow at a few hundred 1pm at brine temperatures
19
-------
above 300°F, Reference 3. Flow rate can be expected to increase with increas-
ing temperature, but natural flow rates of more than 3785 1pm (1000 gpm) are
likely to be relatively rare.
Blowouts are of particular concern because of potential difficulties in
controlling them. Oilfield blowout preventers (BOPs) are routinely used in
geothermal well drilling and required by regulation. While BOPs are useful in
controlling the large, abrupt pressure surges that are typical of oilfield
blowout occurrences, their application to geothermal drilling, where pressure
surges need not be as great or abrupt, is less well defined. Many of the
blowouts encountered to date in geothermal operations could not have been
controlled by BOPs.
In a geothermal blowout, if the brine is discharging through the wellhead
and is uncontrolled by the BOP, it may be possible to fasten a cap or closing
valve to the assembly stub. However, if the brine is discharging away from the
wellhead, flow may be stopped by one of two methods, i.e., cooling or flow-
inhibiting materials can be injected directly into the wellbore, or coolant
can be injected into the reservoir very near the uncontrolled wellbore through
a relief well. The first possibility may be impractical in that the brine
discharge may be located so that there is no access to the wellhead and the
second possibility, requiring a special well, is expensive. Either of these
control methods will be time-consuming in that special supplies, equipment and
personnel must be obtained and transported to the drill site. Accordingly,
while the rate of brine discharge in a blowout may not be great, the flow can
continue for a significant length of time and the volume of discharged brine
can be large. As an example, Well Thermal 4 in The Geysers dry steam field
blew out while drilling in 1957 because of inadequate casing and a minor land-
slide. To date, this well is still only partially controlled. It has been
estimated that from 1957 to 1972 the well emitted more than 9 million tons of
steam, about 4,000 tons of H2S, 5,000 tons of NHo, and 6,000 tons of CH4,
Reference 4. The Geysers is the only operational geothermal field in the
United States. Over 100 wells have been drilled there with four blowouts.
Blowouts have also occurred at Beowawe, Nevada, where three capped wells,
drilled between 1959 and 1965, were dynamited by vandals. Strong ejections
of steam and water resulted and these wells are still uncontrolled but adequate
control measures have not been applied. The discharged water is of low'
salinity and is currently used for pasture irrigation.
Outside the United States, blowouts have occurred at Cesano in Italy,
Cerro Prieto in Mexico and Wairakei in New Zeland. In the Cesano field, the
first exploratory well blew out in January, 1975. The well flowed uncon-
trolled and then geyser-type eruptions occurred. The well eventually spewed
both mud and water and, after 10 hours, closed itself by scale blocking
valves and drilling equipment.
At Cerro Prieto, over 40 wells have been drilled, with two blowouts, in
1961 and 1972. The 1961 blowout was controlled by directional drilling and
injection of cement, while the 1972 blowout was eventually controlled after
blowing wild for four months, following a violent eruption.
20
-------
At Wairakei, of the 100 geothermal wells drilled, three have been subject
to blowouts. One of these resulted in a large crater which emitted steam for
several years.
Corrosion
Components in a geothermal power plant can fail because of corrosive
effects that are functions of the material used, physical characteristics of
the brine, chemical composition of the brine, the presence of entrained solids
or gases in the brine, and the time involved. It is important to note that
the presence of oxygen in the brine will increase corrosion rates in nearly
all metals. Data is currently being compiled on rates of corrosion versus
metal type; Reference 5 contains such typical data.
General corrosion occurs by removal of material over the entire exposed
surface. The result is a diminishing of wall thickness with a consequent
structural weakening of the component. Products derived from corrosion can
also "jam" movable parts such as valves.
Physical factors which affect corrosion and corrosion control are tem-
perature, velocity of fluid flow, changes in flow direction and velocity,
and contact with a second metal. As a general rule, higher temperatures
generate higher corrosion rates, higher velocities usually increase corrosion
rates, and bimetallic contact intensifies corrosion of one of the metals.
Extreme temperatures are the greatest accelerators of corrosion in iron
and steel plumbing systems in that the higher the temperature, the more rapid
is the rate of chemical reaction. In general, for each 25°C (45°F) rise in
water temperature, the rate of corrosion doubles. An increase in flow velocity
induces higher frictional resistance on the pipe; hence the potential for
stress corrosion is generally increased. This is particularly true for ag-
gressive waters in which high velocities are conducive to rapid pipe deteri-
oration.
Bimetallic contact is the usual cause of galvanic corrosion. When two
metals are in contact, as in the case of coupling copper to iron pipes, the
difference in electrochemical potential results in current flow. The iron
becomes the anode and corrodes to protect the copper (the cathode). The
cathodic metal is said to be protected at the expense of the anode. In
general, the rate of galvanic corrosion is increased by greater differences in
potential between the two metals; large areas of cathode relative to the
anodes; by proximity of the two metals; and increased mineralization or con-
ductivity of the water.
Failures also can occur in materials subjected to stress over a long
period of time while exposed to brine and the failure stress can be much
smaller than would be expected from straightforward considerations. Some
materials are prone to pitting, and deep penetration can occur in a short
time. Intergranular corrosion is a localized type occurring at grain bound-
aries within a metal. There are, however, several types of corrosion that
are particularly important in geothermal facilities and these are discussed
in the following.
21
-------
Stress corrosion cracking can occur when a metal under tensile stress is
exposed to a specific corrosive environment; for example, carbon steels and
aluminum alloys, commonly used in geothermal power plants, can be affected by
sea water and marine and industrial atmospheres as well as other solutions.
Copper alloys can be affected by hydrogen sulfide. Stress corrosion cracking
can be prevented by cathodic polarization.
Corrosion fatigue can occur when a metal in a corrosive medium is sub-
jected to repeated stresses. There are many corrosive fluids that can cause
this failure and they are not necessarily specific for a given metal. Hydrogen
embrittlement occurs when metal is exposed to hydrogen gas although one type
is evidently the result of hydrogen being in solid solution in the steel lat-
tice. It is marked by delayed failure under low temperature, long term load
conditions; failure occurs at loads less than the expected. Hydrogen embrittle-
ment is more prevalent in high strength materials.
Erosion-corrosion is very important in areas of high brine velocity,
especially when a directional change in the flow occurs. Here, protective
films are continuously removed by impingement of the fluid and the corrosion
rate is accelerated. Entrained solids in the stream accentuate the problem.
A similar mechanism is fretting-corrosion, where small mechanical displace-
ments (due for instance, to a vibrational load) cause the protective films to
be continuously removed and fresh metal exposed.
The corrosive effects of geothermal brines have been and continue to be
the subject of much attention within the geothermal community. Since the
physical and chemical characteristics of brine can vary significantly between
and within reservoirs, extensive analyses and tests of brine-induced corrosion
have been and are being conducted. Past work has been summarized by Banning
and Oden (Reference 6). Other studies pertaining to specific geothermal
reservoirs have been documented by Miller (Reference 7) and the Bureau of
Reclamation (Reference 8). Much of this investigation has been oriented toward
materials selection, i.e., identifying the construction material that is least
susceptibe to corrosion by determining the amount of material lost per time
(commonly in millimeters per year) through corrosion under the test conditions.
Table 2, as derived by Shannon (Reference 9), indicates typical test results.
The materials investigated are usually metals, although concretes and elasto-
mers have also been examined. Kukacka (Reference 10) and Hirasuna (Reference
11) conducted studies in these areas. Still other studies, such as that by
the Lawrence Berkeley Laboratory (Reference 12) have examined methods of treat-
ing the brine to control corrosion of a given material.
Based on the above and many similar studies, there appears to be a com-
plex correlation between corrosion and component failure rate but quantitive
correlations have yet to be established.
Abrasion
Flowing brine can abrade the inside walls of brine carriers; suspended
particulate matter increases the abrasion rate. The effect is to reduce the
carrier wall thickness with a consequent lessening in a structural strength.
Soft, elastomer valve seats are particularly vulnerable to abrasion with the
result that the valve will not close properly.
22
-------
TABLE 2. CORROSION RATES (MM/YR) OF VARIOUS ALLOYS VERSUS BRINE COMPOSITION*
Alloy
A570
A53B, Heat 1
A53B, Heat 2
C75
1010
4130
2 1/4
410, Heat 1
410, Heat 2
E-Brite 26-1, Heat 1
#-Brite 26-1, Heat 2
Has ta Hoy C-276
Inconel 625
Inconel 600
Incoloy 825
29Cr-4-2
6X
1% NaCl
0.30
0.30
-
0.20
0.40
-
0/08
0.08
-
0.05
-
0.02
0.01
0.02
0.01
0.08
0.01
5% NaCl
0.90
0.60
0.40
0.40
1.00
-
0.60
0.10
0.20
0.01
0.02
0.03
0.01
0.02
0.01
0.01
0.01
10% NaCl
1.10
0.90
0.60
0.20
1.00
0.20
0.30
0.80
0.50
0.02
0.02
0.04
0.01
0.05
0.01
0.02
0.01
20% NaCl
2.80
2.80
3.20
2.00
3.50
2.30
1.80
3.50
3.80
0.08
0.02
0.04
0.02
0.04
0.05
0.01
0.03
*Test Conditions: T = 250°C; P = 68.9 Bar (1000 psi); Oxygen = <0.01 ppm;
pH = 4.6 to 4.8
Since the most common source of particulate matter is the production
reservoir (rock particles are commonly entrained in the produced brine), com-
ponents at and near the production wellheads are most affected by abrasion.
Scaling
Scale can form on the inside surfaces of pipes and other brine carriers.
It is formed, in part, by an oxidation reaction wherein a metal component in
the system is dissolved and redeposited as an oxide. Also, in addition, re-
actions will occur to form insoluble compounds which precipitate on exposed
surfaces under changing temperature, pressure, fluid composition and velocity
conditions. The most common of the latter types are sulfates and carbonates
of calcium, magnesium and sodium and also metallic sulfides.
23
-------
Saturation index (S.I.) is normally used to measure the corrosiveness or
scaling potential of a water. This index is based on the concentration of the
carbonate and calcium ions exceeding the solubility product as a criteria for
deposition. Since the index is dependent upon the law of mass action between
the two ions, the effect of dissolved oxygen on corrosion rates is not con-
sidered. In general, when S.I. is positive, scaling or deposition of calcium
carbonate will occur. An S.I. >0.5 indicates excessive scale formation. When
S.I. is negative, the water is corrosive and scale will dissolve in solution.
'-Ideally, to prevent corrosion, the water should have a S.I. of zero at
all times. In reality this is not possible because of the dynamic system; the
general rule of thumb is, therefore, to maintain the S.I. slightly positive
(0.1 to 0.3) so that a protective film of calcium carbonate is always there to
minimize corrosion effects.
While the scale itself does not present a significant problem in brine
release, the scale will constrict flow channels either reducing the flow in a
fixed pressure delivery system or, under certain special circumstances, cause
an increase in pressure in the system. This pressure increase combined with
structured! weakness caused by corrosion or erosion could cause failure. When
the scale breaks loose, it can clog valve seats causing valves to freeze open
allowing fluid to pass with possibilities of fluid release. A case in point
would be the failure of a relief valve to reseat due to scale on the seat and
continuing to discharge fluid to the atmosphere.
A variation of this can take place in the injection process. A very high
brine injection rate is desirable to minimize the number and cost of injection
wells. Precipitation can occur in the reservoir in the vicinity of the well-
bore, necessitating a rise in injection pressure which, again in combination
with other factors, could cause a filaure in the system.
Similar to corrosion, the subject of scaling is receiving much attention
by the geothermal community and many tests of scaling in geothermal systems
have been and are being conducted. These studies involve both examining the
mechanisms of scale accumulation, of which Makrides (Reference 13) is an
example, and methods of removing the scale after it is formed, exemplified by
Daedalean Associates (Reference 14).
The three foregoing processes, corrosion, erosion and scaling, are po-
tential failure mechanisms that have a high probability of occurrence in
geothermal systems. Minimizing the effects of these requires good engineer-
ing practices, periodic inspections during operations, and an adequate main-
tenance program. Good engineering practices would include proper choice of
materials and components (e.g., some valves are more resistant to abrasion
than others), inclusion of inspection ports in vessels, and, to minimize
abrasion, use of long radius elbows, heavy walled elbows and expendable com-
ponents such as replaceable impingement baffles used in vessels at Cerro
Prieto and Westmorland.
24
-------
Improper Design/Poor Workmanship
Proper design of a system, good workmanship in the installation, and the
use of good materials are normally expected in any piping and equipment system
such as refineries, power plants and chemical plants. Design, installation,
and materials necessary to assure a reliable geothermal system are all within
the state-of-the-art. Design inadequacies that may allow failures and rup-
tures and consequent brine release include pressure build-ups between shut-off
points with no relief valves, improperly selected pumps (shut-off heads ex-
ceeding allowable stresses), and inadequate flexibility in piping to allow for
temperature growth. Improper selection of class or rating as well as materials
of seals, gaskets, bonnets, seats of valves and equipment is possible if the
designer is not thoroughly familiar with the system and all its extremes.
Poor workmanship with improper or non-existent inspection and testing can re-
sult in loose leaking joints. A check valve or a relief valve installed back-
wards can have disastrous effects with leaking brine and/or complete rupture
of pipe or equipment. Pumps have been installed backwards as well as binary
vessels, such as exchangers, having connections reversed.
Designs should be reviewed by competent, knowledgeable individuals,
similar to building and safety and environmental engineers. Construction
should be inspected to assure compliance with the plans and specifications of
the design. Each system and subsystem within a geothermal loop should be
tested under controlled conditions prior to operation.
Natural Disasters
The geographical correlation between geothermal resources and zones of
tectonic activity is well known. Accordingly, geothermal power plants are
inherently associated with areas of higher seismicity. In fact, locating
areas with an unusually high level of micro-earthquake activity is an important
geothermal exploration method, and determining the geometric shape of these
areas is becoming an important geothermal reservoir development factor. Con-
sequently, the possibility of earthquake damage appears greater during the
life of a geothermal installation than the more conventional power plants.
Damage can be minimized through proper design.
Flood, landslide, high wind and/or fire damage during the life of the
power plant is a possibility. The probability of flood and slide damage can
be minimized through proper siting, i.e., avoidance of flood plains and un-
stable slopes. The possibility of subsidence damage may be minimized by
injection of spent brine back to the geothermal reservoir.
Vandalism/Terrorism
Terroristic attacks, wanton destruction and inadvertent damage caused by
demonstrators for otherwise peaceful causes are becoming more of a problem,
particularly for projects that are controversial. The mitigative action here
is an adequate security system including controlled access to the facility, in
addition to a realistic public information program.
25
-------
Accidents/Personnel Errors
Accidents, such as a vehicle being driven inadvertently into a brine
carrier, and personnel errors, such as the wrong valve being opened, are ever-
present possibilities. A well trained, motivated work force is the best re-
sponse to this. Protective structures and devices should be installed in
higher risk areas.
BRINE RELEASE RATES
The rate of fluid release will be dependent upon failure mode. A leaking
gasket in a flange will allow a trickle release. Failure of valve packing or
pump seals may allow a few percent of the brine flow to escape. Brine release
from perforations in the wall of steam separators or other vessels will range
from a trickle to a few percent of the flow. Release from a burst pipe will
depend upon the location and size of the break. A longitudinal split is the
normal mode of failure. A small split on the upper surface of a level pipe
may release a few tenths of a percent of the pipe flow while a large split on
the undersurface of a sloping pipe may release the entire flow. Given this
variability in potential brine release rates, a containment method should be
designed for the worst case.
Table 3 indicates the flow rates in specific parts, comprising groups of
potential loss points, depicted in Figure 7, 8, and 9, that have a common
flow rate. The indicated flow rates are the maximum brine spillage rates that
can occur within each group. The spillage rates associated with the production
and injection manifolds are a function of where in the manifold the pipe bursts.
The maximum potential release rate of 53,000 1/min (14,000 gpm) is found in
System 1.
TABLE 3. BRINE FLOW RATES
o
Potential Loss Point Groups (m /min)
System 1
System 2
System 3
Production
wellhead
4
4
4
Common pipe
(including
manifold)
4-53
4-49
4-30
Steam sepa-
rator/heat
exchanger
—
49
30
Flash Common
units pipe
53* 53*
49
30** 15**
Injection
manifold
8-53*
8-49
8-15**
Injection
wellhead
8*
8
8*
* Less fluid flashing to steam
** Less noncondensables
26
-------
FAILURE MODE ANALYSIS
A failure mode and effects analysis was performed on each identified com-
ponent in the three systems of interest; results are tabulated in Table 4. Of
primary interest were those components and failure modes that could contribute
to brine spillage. Although the loss of working fluid in a binary system and
the venting of non-condensable gases are hazardous, they are outside the scope
of this study.
Failure Rate Data
There are no compilations yet of failure rate data that are specifically
applicable to geothermal power conversion and there is no extended history of
liquid-dominated geothermal operations in the United States upon which to base
such compilations.
There is a growing body of data on the corrosion and scaling of materials
when exposed to geothermal brines. While these data are no doubt influencing
the design and selection of components for geothermal power plants with a con-
sequent increase in overall reliability, extrapolating these data into quanti-
tative failure rates at the present time would be pure speculation.
The source for the component failure rates used here is Appendix III of
a commercial reactor safety study, the "Rasmussen Report" (Reference 15). This
report contains failure rates that reflect both nuclear and non-nuclear indus-
trial experience. The data were derived from Department of Defense data, NASA
data and general industrial and commercial operating experience. Approximately
50 sources of failure data are cited. The operations that provide the basis
for the failure rates involve high pressure, high temperature steam, and the
components involved are of a physical size comparable to those used in geo-
thermal power plants. The Rasmussen Report treats failure rates as random
variables as opposed to single, best values and, consequently, provides
failure rate in a variety of forms. These are summarized in Table 4 for
several components that are applicable to geothermal power plants. The speci-
fic type of failure is also provided. It is important to note that failure
rates are provided in terms of both failures per demand and failures per hour.
The only data considered herein are failures per hour. Maximum and minimum
failure rates obtained from raw industrial data, the median values of those
data and entirely nuclear failure rates are listed. The failure rates used
here are the median values and represent a combination of nuclear and non-
nuclear data. While the study does not list entirely non-nuclear failure
rates (and these are the failure rates that might be considered best appli-
cable to geothermal usage), the Rasmussen Report does state (p. 111-77) that,
in the case of pipe failure (the single most controlling factor in determining
the probability of brine release in a geothermal power plant), there is general
agreement between nuclear and industrial data. Further, a comparison of the
median values and the nuclear experience values indicates that, for a given
component, the two are equivalent or the median failure rates are higher.
Since these failure rates are implicitly valid for components and materials
that meet good engineering practices for the conditions of temperature, pres-
sure and chemical environment in which they operate, use of the median rates
in assessing geothermal brine release probabilities appears valid.
27
-------
TABLE 4. FAILURE RATE DATA
Components
Pumps
Fall to start/d*
Fall to run/hr**
Valves, Motor Operated
Fall to opera te/d
Plugged/d
Leak/rupture/hr
Valves, Solenoid Operated
Fail to opera te/d
Valves, Check
Fail to open/d
Leak/rupture/hr
Valves, Relief
Fail to open/d
Valves, Manual
Plugged/d
Pipe
Plug/rupture/hr
S3" diameter
>3" diameter
Gaskets
Leak/hr
Failure Rates - Rasmussen Report
(Reference 15)
Extreme Range
5 x 10"5 - 5 x 10"3
1 x 10"7 - 1 x 10"4
2 x 10"4 - 7 x 10"2
6 x 10"5 - 3 x 10"4
...
2 x 10"5 - 6 x 10"3
2 x 10"5 - 3 x 10"4
___
1.4 x 10"5 - 3.6 x 10"5
—
2 x 10"9 - 5 x 10"6
1 x 10"10 - 5 x 10"6
___
Median
1 x 10~3
3 x 10"5
1 x 10"3
1 x 10"4
1 x 10'8
1 x 10"3
1 x 10"4
1 x ID"8
1 x 10"5
1 x 10'4
1 x 10~9
1 x ID'10
3 x 10'6
Nuclear
Experience
1 x 10"3
1 x 10"6
1 x 10"3
3 x 10"5
...
1 x 10"3
1 x 10"4
___
1 x 10"5
3 x 10"5
1 x 10"9
1 x ID'10
— _
Failure Rates - RADC***
(Reference 16)
Components
Pumps
Boiler feed
Centrifugal
Valves
General
Butterfly
Solenoid
Diaphragm
Check
Relief
Gaskets
General
Packing
Failure
Rate/Hr
4.2 x 10"7
5.8 x ID"6
1.5 x 10"5
1.3 x 10"6
1.6 x 10"6
2.6 x 10"6
3.2 x 10"6
1.6 x 10"6
1.3 x 10"6
3.5 x 10"6
LEGEND
* /d = per demand
** /hr = per hour
*** RADC = Rome Air
Development Center
ro
00
-------
Additional failure rate data have been compiled in Table 4 from the
Reliability Analysis Center, Rome Air Development Center, Reference 16. This
document is a compilation of failure rates for nonelectronic parts for both
military and commercial applications. Parts that are applicable for geothermal
applications include pumps, valves and gaskets and pertinent data on the
failure rates of these are also tabulated in Table 4. These rates, however,
consider components generically and include failures from all causes; accord-
ingly, comparison of data from this document and the Rasmussen Report is dif-
ficult. The only direct comparison is that of gaskets, and the failure rate
is comparable. A comparison of other components that are not included in this
geothermal examination, i.e., clutches, motors, relays, switches and power
supplies, suggest that failure rates from entirely nuclear parts are higher
than commercial and military counterparts.
The specific failure rates used in this analysis are tabulated in Table
5. The pipe rupture failure rate, as extracted from the Rasmussen Report, is
applicable to a section of pipe where a section is defined as "an average
length between discontinuities such as valves, pumps, etc. Each section can
include several welds, elbows and flanges." The pipe rupture rate was also
used as a basis for estimating the failure rates for other components. The
rate for the boiler/superheater, preheater/economizer, separator/heat ex-
changer, multistage flash units and flash tanks were derived by applying a
complexity factor of ten to the large diameter pipe failure rate. This
factor was deemed appropriate due to the number of welds, joints, valves and
flanges involved. Similarly, a complexity factor of five was applied to the
injection pump section and a factor of two was applied to the hydrogen sulfide
scrubber. Wellhead assembly failure rates were obtained by combining failure
rates for an appropriate number of valves and pipe sections.
In relating the causes of fluid releases to the components in a geother-
mal brine system as shown in the matrix of Table 6, it becomes apparent that,
except in the special cases of wells and gaskets, all components are subject
to most of the generic causes of failure and/or rupture as shown. A matrix
such as this can be used by the designer in the selection and design of the
individual components in his attempts to minimize the number of spills of
brine. Gaskets, with the lowest reliability, must be selected of the proper
type and material, and close control of installation must be exercised. Pro-
per maintenance is also indicated to prevent leaking. A leak is distinguished
from a rupture, as used here, by the degree of fracture. A rupture is a major
violation of the structural integrity of a component leading to the release
of a significant quantity of fluid per unit time. A leak, on the other hand,
is a minor violation of structural integrity and the quantity of fluid re-
leased per unit time is relatively small. From Table 5, although the failure
rate of gaskets is high, the failure mode is leakage and not rupture; contain-
ment of brine from this cause and component is not warranted. Conversely,
boiler/superheaters, and flash units, which have a relatively high failure
rate, a number of uncontrolled failure causes, and a rupture failure mode,
will require either positive brine physical containment methods or failure
prevention methods, whichever is the more economical.
It should be recognized that the failure rates included here are measured
rates that apply to specific component parts operating in specific environments.
29
-------
TABLE 5. GEOTHERMAL BRINE RELEASE, FAILURE MODES AND EFFECTS
Item
Wellhead assembly
Valves
Piping
Boiler/superheater
Preheater/economi zer
Injection pump/ valves/
piping
Injection wells/valves
Separator/heat exchanger
Multistage flash units
Hydrogen sulfide (hLS)
scrubber
Flash tanks
Gaskets
Failure mode
mechanism
Valve leakage
Pipe rupture
Failure to operate
Leakage
Leakage
Rupture
Tube failure
Nozzle leakage
Rupture
As above
Blade defect
Bearing wear
Seal wear/ leakage
Rupture
Leakage
Valve/piping rupture
Nozzle leakage
Rupture
Nozzle leakage
Rupture
Leakage
Rupture
Nozzle leakage
Rupture
Leakage
Failure effect
Blow out
Brine spill
Control loss
Control loss
Brine spill
Brine spill
Leakage
As above
Brine spill
Brine spill
Reduced output
Brine spill
Reduced output
Brine spill
Brine spill
Reduced output
Brine spill
Brine spill
Failure rate
(failures/hour)
8 x 10-10
1 x 10"8
1 x ID'10
1 x 10"9
As above
5 x ID'10
8 x 10~10
1 x 10"9
1 x 10"9
2 x ID'10
1 x 10"9
3 x 10"6
Accordingly, some of the causes of brine release discussed in the previous
section may not be adequately represented. In particular, accidents and per-
sonnel errors will probably cause a significant percentage of the total fluid
releases, but the above quantitative rates may not reflect this. Consequently,
the cited failure rates and the conclusions drawn from them are probably
conservatively low.
30
-------
TABLE 6. GEOTHERMAL COMPONENTS AND
GEOTHERMAL COMPONENT FLUID
FAILURE CAUSES
RELEASE CAUSES
ITEM FAILURE RATE
(FAILURE/HR)
Gaskets 3 X 10'6
Valves 1 X 10'8
Boiler/Superheater 1 X 10"
Preheater/Economizer 1 X 10"9
-Separator/Heat Exchanger 1 X 10"9
Flash Units 1 X 10"9
Flash Tank 1 X 10"9
Wellhead Assembly 8 X 10"8
Injection Wells 8 X 10"10
Injection Pump 5 X 10"10
H2S Scrubber 2 X 10"10
Piping 1 X 10"10
i-
ra
o
3
o
_1
CO
X
X
3Z
0
*-H
(/»
O
ex.
a
o
c_>
X
X
X
X
X
X
X
•z.
0
1— <
in
O
en
UJ
~^
sz
0
to
00
X
X
D£
UJ
t-
tf>
<
-------
• Minor release - a single-point gasket or seal leak.
A critical release will result in the spillage of large amounts of brine,
ranging from 4000 to 56,000 liters per minute. A major release will allow
spillage rates ranging from approximately 40 to 4000 liters per minute, depend-
ing upon the size of the component that ruptures. A minor release will usually
involve less than 40 liters per minute.
Each system under consideration was diagrammed (Figures 7, 8, and 9) to
highlight the potential fluid release sections within the system. The most
likely areas of concern, i.e., a critical release, would be at the brine entry
side of the plant. As the flow progressess through the plant, more faults (or
failures) are required for critical brine release. Critical fluid releases and
major fluid release probabilities were computed for both an assumed forty-year
design life and at a five-year point. The numbers reflect the probability that
at least one fluid release would occur during the designated time periods. For
example, the probability that at least one critical fluid release will occur
in the double flash system during the forty-year design life is 3.7 x 10-5. it
can be seen, from the table, that approximately an order of magnitude decrease
in the probability of a brine release can be realized by substituting a compre-
hensive preventive maintenance program, i.e., essentially renewing the system
at five-year intervals.
The overall probability of a major fluid releast somewhere in the system,
as shown in Table 6, is about the same for all three systems, approximately
2 x 10-3, or one chance in 500 during 40 years operation of a plant. This
figure excludes wellhead ruptures, which are discussed below.
So that the possibility of inducing additional failures by rapid flow re-
duction is reduced, consideration should be given to having the capability to
bypass all brine from the production wells to the reinjection wells. Such a
bypass line is shown in the system configurations of Figures 1, 3, and 5. A
review of Table 6 clearly shows that the major contributor to the overall plant
critical brine release probability is the well-to-plant sections. Adding a
valve at the well-to-plant section to facilitate isolation of that section
would change the well-to-plant critical release probability from 3.5 x 10~5 to
3.5 x 10"8 and, consequently, the overall system critical fluid release prob-
ability would decrease significantly.
Production well brine release probabilities vary with the number of pro-
duction wells required in each system. The limited historical data available
suggests that approximately two percent of geothermal wells will blow out
during drilling. . Each wellhead contains eight valves and associated piping/
connections that are potential rupture release points. Based upon data
presented previously, the probabilities of a wellhead rupture during the forty-
year design life are: System 1 - 7.8 x 10-3; System 2 - 6.9 x 10~3; and
System 3 - 2.2 x 1Q-3.
Minor leaks are to be expected. Based upon available gasket failure data,
each gasket in the system would be expected to fail at least once during the
assumed forty-year design life.
32
-------
TABLE 7. BRINE RELEASE PROBABILITIES
System/section
System 1 - total
Well-to-plant
First stage flash
Second stage flash
H9S scrubber
-------
SECTION 6
CONTAINMENT METHODS AND FLUID RELEASE CAUSES
GENERAL CONTAINMENT EVALUATION
The previous section describes the causes and probabilities of a geo-
thermal brine release and points out that the volume and rate of brine released
can be extremely variable and unpredictable and that the location of release
and the direction of brine emission from the release point is also unpredict-
able. Accordingly., the containment measures described in this section are in
keeping with good engineering practice and consider the worst case release con-
dition.
There are no firm guidelines to indicate when a containment system should
be incorporated in a geothermal facility and so the criteria for including a
containment system is to minimize environmental damage from a brine spill. It
should be recognized, however, that in some cases, environmental damage from
spills could be minimal where geothermal brines are relatively low in salt con-
tent and the area is extremely arid. Accordingly, the spilled water flowing
over the terrain might be beneficial to plant and animal life. A containment
system appears to be most desirable in those facilities utilizing a brine with
a high salt content and where flow from a large spill would cause erosional
damage. Further, thermal pollution from a large spill can be mitigated by a
containment system with delayed release of the brine into the environment.
Decisions as to the use of a containment system and the type of system to be
used will be judgmental, based on trade-offs between cost and the positive and
negative environmental impacts of a spill.
Measures or methods for containing surface releases of geothermal brine
fall into three general categories, (1) minimizing release possibility; (2)
minimizing release duration; and (3) minimizing release area. Table 8 has been
developed to indicate the applicability of these three containment methods to
the causes of fluid release identified and discussed in the previous section.
It is quite obvious that the number of failures and resulting brine releases
can be reduced by good design practices and by use of proper operating and
maintenance techniques but it must be realized that failures and releases can-
not be totally eliminated. Accepting that spills will occur, it then becomes
a question of reducing the quantity of brine released and minimizing the area
affected by the spill. This will, in turn, minimize the environmental impact
resulting from the spill. Correct operating procedures, inclusion of alarms
in the design and proper personnel training programs are necessities if the
time of brine flow through a rupture (and, therefore, the quantity of brine
released) is to be minimal. Establishing procedures and personnel training
are inexpensive and should be part of routine operations at any power plant.
34
-------
TABLE 8. SURFACE BRINE RELEASE CONTAINMENT METHODS
CO
CTT
FLUID RELEASE
CAUSES
BLOW OUTS
CORROSION
ABRASION
SCALING
NATURAL DISASTERS
VANDALISM/TERRORISM
ACCIDENTS
PERSONNEL ERROR
TO MINIMIZE
NUMBER OF FAILURES
UJ
in
g
3
UJ
?
ID
t—
O.
O
X
X
X
t-
•Q
f-
Z
UJ
z z
§!/>
UJ
<_) 0
X
X
X
10 =T
UJ H-
11
UJ 0
X
X
UJ
3
UJ
_J
Q*
UJ
c*
PRESSU
X
X
o
u>
CO
z £
3C UJ
HINIHU
COHPON
X
X
X
X
>•
t—
_J
CO
<
Z
MAINTA
X
X
X
LU
o
5
UJ
h-
z
?
1— <
o <^*
UJ O
u-0
u. «
UJ Q.
X
X
X
X
*/>
UJ
0 £
UJ C?
(J O
o: a.
a.
ss
UJ t—
a.
<=>•»
X
X
X
X
ECURIT
i/>
-j
PHYSIC
SYSTEM
X
X
TO REDUCE TIME
DURATION OF RELEASE
O
t—
=3
f)
h- .J
Sri
o at
i—
X
X
X
X
X
X
X
u.
u.
o
i
H-
3C
VI
K
X
X
X
X
X
X
X
QC
Ul
-J
UJ
X
(/>
O tt
UJ t3
O O
o ^
Oi t
a.
>• 5=
o »-«
Q£ S
S-
X
X
X
X
X
X
X
TO REDUCE AFFECTED
AREA OF RELEASE
>•
0
z
rs
0
UJ
a:
^-
X
X
E:
EOUNCA
h-
3Z
X
X
X
X
X
o <:
o a:
i . i ^
0 O
o a:
Of £L
a.
>• ^
n: S
IAJ H-
LJ fctJ
X
X
X
X
X
X
X
-------
The methods, or combinations of methods, that are optimal for reducing the
duration of release and the area affected is a function of the cause of re-
lease and this is discussed in the following by reference to Table 8.
Blowouts
Blowouts are best prevented by employing good operating procedures and
proper equipment during drilling. An effective maintenance program involving
periodic inspection and reworking of the wells will detect and repair those
factors, e.g., corroded casing or failed cement that may generate blowouts
during the life of the well. Proper rework of the cement can also increase
the probability of post-blowout brine flow being contained in the wellbore so
that capping is expedited and the duration of release is reduced. Since geo-
thermal wells are inherently located in seismically-active areas, seismically-
induced failure of casing and wellhead equipment may become a problem and in-
cluding seismic considerations in the well design may be desirable. An example
of this might be more prevalent use of uncemented hung liner that would be more
amenable to lateral longitudinal motion than would be cemented casing. Area!
effects of a blowout-caused spill could be reduced by diked areas, but this is
probably not generally practicable.
Corrosion and Abrasion
Containment methods useful for corrosion and abrasion-caused releases are
essentially identical. Minimizing the probability of failure through these
causes is a matter of employing good design practices including the use of op-
timal materials, minimizing the number of components used and directly apply-
ing the necessary design techniques such as specifying increased wall thick-
nesses where abrasion in piping could be significant. Proper maintenance
including periodic inspections is a requirement for minimizing failure pro-
bability during the plant life. Reducing the time duration of release due to
these causes is a matter of a shut-off system (either automatic or manual)
being available at the wellhead, having emergency procedures defined and having
a well-trained work force available to implement these procedures. It is in-
teresting to note that these factors are also applicable to reducing the time
duration of release from any of the causes. Areal effects are reduced by in-
cluding several points of shut-off in the system, again having procedures and
trained personnel available and using diked areas of a size appropriate to the
expected worst-case release.
Scaling
Seal ing-caused release containment is very similar to that of corrosion
and abrasion except that including scaling effects in design considerations
is less practical and the use of pressure release components at appropriate
places in the system becomes mandatory. Areal containment methods are also
very similar to those of corrosion and abrasion.
Natural Disasters
The releases induced by natural disasters such as quakes promise to be
36
-------
worst-case in nature so that larger diked areas may be necessary. Design tech-
niques for limiting seismic effects on piping systems are well known and could
be employed to minimize spills from this cause.
Vandalism and Accidents
Methods of minimizing failures caused by vandalism and accidents are simi-
lar in that the necessary operating procedures plus a well-trained work force
to implement these procedures are required. Further, a physical security sys-
tem that would monitor the presence of intruders, both purposeful and accident-
al, within the plant boundaries is necessary. Areal limitation of resulting
spills is a function of having multiple shut-off locations available, the neces-
sary procedures and trained people being available, and appropriate diking.
Personnel Errors
Releases through personnel errors can be reduced primarily by employing
a trained, motivated staff. Using the minimum of components and appropriate
pressure release mechanisms (to be used in event that a valve is mistakenly
actuated and a pressure buildup is generated) are also significant. Areal
containment is primarily a function of appropriate diking.
Recommenda tio ns
Examination of Table 8 indicates that certain containment methods are
essentially common to all release causes once fluid has been released. These
methods are: the availability of a shut-off system, either automatic or manual,
at the wellhead, an appropriate diking system, the existence of emergency oper-
ating procedures and a work force that is sufficiently trained and motivated
to exercise these procedures on demand.
SPECIFIC CONTAINMENT EVALUATION
As stated above, there are three general categories of containment methods,
i.e., minimizing the possibility of brine release, containing a release in time,
and containing a release in area. These are evaluated in the following discus-
sion. Finally, a step-by-step procedure for determining containment designs
and specifications is presented.
Minimizing Brine Release Possibilities
The most significant factor in minimizing the possibility of brine release
is the use of good engineering practices in the design and operation of the geo-
thermal energy conversion system. The possibility of brine release should be
considered from the onset of the program. Some specific practices to lessen
effect on brine release possibility are discussed below. (It should be recog-
nized that other system considerations may be sufficient to negate the actual
use of some of these practices).
37
-------
• The facility should be designed so that components are visible and
accessible. Piping should be above ground, pipe insulation should
cover flanges, and pressure gauges should be located for easy viewing.
t Redundance should be incorporated into the design. Systems 1 and 2,
described previously, have fully redundant brine loops. (System 3 is
more experimental and is patterned closely to a specific installation).
Incorporating redundancy in a system is now a good, standard design
practice. Subsequent examination of the failure rate data for these
systems shows no obvious need for more redundancy than has been used.
t The numbers of components should be minimal while keeping with the re-
dundancy principle. This is an obvious desirability, also reflecting
into system costs.
• Optimum materials should be used. Existing facilities rely heavily
on carbon steel but a body of data on the corrosive effects of geo-
thermal brine on various materials is being generated and other, pro-
bably more expensive materials may find application. Titanium heat
exchanger tubes are now commonly accepted.
• The facility design should recognize the relatively higher seismicity
of geothermal areas, and pipe supports and other installations should
be so designed and installed. Unstable soil areas should be identified
and avoided.
• Blowout preventers should be used when drilling and casing and cement
should be selected to withstand high temperatures.
Another significant factor in minimizing brine release probabilities is
an adequate maintenance program during facility operation. The failure analy-
sis in the previous section points this out. The program should have an ade-
quate budget and the importance should be recognized. However, the data neces-
sary for devising a complete maintenance program does not now exist. While
data pertaining to brine carrying and power conversion components are avail-
able, as discussed in the previous section, wells present a more formidable
problem. There is not yet sufficient experience in this country to determine
the workover procedures and schedules necessary to keep production and injec-
tion wells operating in an optimum manner. Experience along these lines does
exist in other countries. A large body of well maintenance data exists for
the Cerro Prieto field in Mexico. A program of collecting and analyzing main-
tenance data from these sources is suggested.
Finally, security of the area should be considered. The system designs
considered here are inherently secure in that the facility is limited in size
and it will be well lighted and manned around the clock. Other systems that
utilize vertical drilling will occupy a larger area and a longer perimeter
fence and, perhaps, patrols will be required.
38
-------
Brine Release Containment in Time
The time during which brine is released should be minimized, thereby mini-
mizing the volume spilled. Accordingly, flow through the loss point should be
shut off quickly. There are several means of doing this, as follows:
• A fully-automatic shut-down system may be used. A brine release will
generate a pressure loss that can be sensed and used to trigger valve
actuators at appropriate places in the system. One configuration of
this system would have automatically-actuated valves on the production
wellheads that would close upon signal. Another configuration would
use a bypass producer-to-injection line, as shown in the systems pre-
viously described and automatically-actuated valves would open this
line while closing the lines to the conversion system. This would
present some advantages in allowing the wells to continue to flow.
The actuating time of motor-driven valves is measured in seconds and
the maximum brine spillage, associated with the maximum flow rate of
53,000 1/min, (14,000 gpm) would be less than 8,000 liters (2,000 gal-
lons). Fully-automatic systems are complex and expensive. An eight-
inch motor-driven valve will cost approximately $6,000 installed and
the cost of so equipping 28 wellheads would be approximately $170,000.
A 36-inch motor driven valve, such as would be used in the bypass con-
figuration, would cost approximately $30,000 installed and two of these
would be required. A data processing unit as well as switching units
will also be required. In summary, designing, procuring and install-
ing such a fully-automatic system would cost approximately $250,000,
a significant cost figure. This cost refers to the systems of inter-
est here where electrical power can be easily supplied to the wellheads.
In systems utilizing vertical wells, the cost of supplying power to the
dispersed wellheads would make the automatic feature more expensive.
• A semi-automatic system may be used where a pressure loss, signalling
a brine release, is presented as an alarm to the plant operator who
manually triggers the actuators. Here, the volume of brine released
is a function of the reaction time of the operator, which will be, in
turn, a function of his training and alertness. A maximum reaction
time of two minutes might be expected, resulting in a maximum spill of
approximately 106,000 liters (28,000 gallons). This system has some
advantages in eliminating the data processing system, about 15 percent
of the cost, and in reducing the possibility of false alarms.
t A semi-manual system may also be used in which a pressure loss alarm
is presented to the plant operator who informs and instructs other
operational personnel. These personnel then operate the required
valves manually. The brine release will be determined by the time
required to locate and inform the proper personnel, for these person-
nel to access the appropriate valves and to operate them. A maximum
total reaction time of two hours might be expected, resulting in a
maximum spill of approximately 6.4 million liters (1.7 million gallons).
39
-------
This system requires personnel to be able to approach and operate the
correct valve hand wheels without being significantly deterred by the
escaping hot fluid and steam. This implies that there must be several
routes of access to the valves, that at least some of the valves must
be elevated above potential surface fluid flows and that valve stems
must be extended so that hand wheels lie behind protective bulkheads.
This system could be implemented in the three conversion systems of
interest for approximately $8,000, assuming two low-lying catwalks,
pressure detectors with alarms, and a wooden bulkhead near the produc-
tion wellheads with valve stems extended to pass through it.
• Still another system would be completely manual with the pressure
detectors eliminated and operating personnel being required to visual-
ly detect a brine release and to operate the valves necessary to con-
trol it.
Of these time-containing systems, the semi-manual system, backed up by
visual release detection by operating personnel, appears the most likely to
be implemented.
Bri ne Conta i nment Pi kes
Areal containment of brine release involves locating, designing and con-
structing a pond of a size to hold the maximum potential volume of released
brine. If used only for spill containment, factors to be considered will be
the potential spill volumes and the fluid retention time required. Should the
pond be constructed for other purposes in addition to spill containment, water
treatment for example, factors pertaining to these other purposes would have
to be considered also.
The maximum expected release of 6.8 million liters (1.7 million gallons)
of brine is equivalent to approximately 5.2 acre-feet. Since a 50 MWe geo-
thermal power plant of the systems of interest here occupies approximately ten
acres, a maximum expected release would cover the entire plant site to a water
depth of approximately 15 cm (six inches).
It is desirable to cause the spilled brine to flow away from the wellheads
and conversion system so that water damage may be minimized and repairs may be
speeded. Accordingly, the plant site should be graded to drain to an appro-
priately placed pond. The desired surface area of the pond, and the resulting
length and height of the berrn, will be a function of the plant layout. It
seems that, given the rarity of significant spills, areas of the site not oc-
cupied by facilities or planned operations would be available for containment
pond usage. Such areas might normally serve as parking lots. If one-third
of the ten acre-site is available for the pond and if the pond bottom is formed
by a single-slope grade to the perimeter fence an average dike height of ap-
proximately one meter (three feet) will be required to contain the maximum
potential spill. The dike length, along the perimeter fence, would be a func-
tion of the shape of the available land area. This type of containment has
the advantages of reduced cost and adding very little environmental damage to
40
-------
that already incurred by constructing the plant.
There may be situations in which grading the site to route the released
brine into a nearby holding pond is necessary. Reference 17 presents invest-
ment costs for such holding ponds; a pond with one-half acre of surface area
and ten-foot dikes would cause environmental damage during construction and
would cost approximately $25,000. The perimeter dike discussed above would
cost approximately one-tenth of this figure.
Under ordinary circumstances, the containment will act as a settling pond
with the brine eventually disappearing through evaporation and percolating in-
to the subsurface. Under conditions of very saline brine and rigid environ-
mental protection requirements, where contamination of ground water aquifers
is feared, percolation into the subsurface may not be allowable and the con-
tainment pond may require lining. This would add substantially to the costs.
Step-by-Step Procedures
• From the system design, locate the component having the greatest
brine flow and note the flow.
• Compute costs of automatic and semi-automatic brine shutoff systems.
- Valves and actuators
- Supplying power to actuators
- Pressure sensors
- Signal conditioning and processing unit
- Switching units
(Costs should include engineering, installation and maintenance
over the life of the plant).
• Collect information on local environmental effects, regulatory require-
ments, community attitudes and political climate regarding a brine
spill.
• Decision: Is automatic/semi-automatic shutoff system feasible and
desirable?
If yes, no containment area is necessary.
If no, implement manual system and proceed to the next step.
• From plant layout, around-the-clock staffing plans and internal plant
communications and transportation plans, determine maximum reaction
time for personnel to access and operate wellhead valves upon alarm.
Influence communications and transportation plans to reduce reaction
time.
41
-------
t Multiply steps to determine maximum volume of an expected brine re-
lease. Convert to acre-feet.
• From the plant layout, determine areas within the plant perimeter
where temporary submergence by a containment pond is tolerable. In-
clude non-critical facility areas such as parking lots.
• From the areas defined in the preceding step, determine the average
berm height and the berm length, as located on the site perimeter, to
contain the maximum expected brine release volume.
• From the areas defined previously and site grading plans and profiles,
determine actual berm heights, approximating the desired average,
around the site perimeter. Compute the volume of fluid that will be
contained by this berm. Influence site grading plans to maximize this
volume.
• Decision: Will the pond contain the maximum expected brine release
volume?
If yes, go to the next-to-last step.
If no, proceed to next step.
• Locate topographically feasible areas outside plant perimeter into
which the pond can be expanded with minimum environmental damage,
compute berm height and length necessary to contain the maximum ex-
pected brine release volume.
• From the brine characteristics and the local environmental require-
ments determined previously, determine if the brine settling into the
subsurface is satisfactory. If not, add a bottom liner to the pond
area.
0 Influence construction schedules so that the berm is constructed
simultaneously with plant site grading.
42
-------
REFERENCES
1. TRW, Incorporated. Experimental Geothermal Research Facilities Study.
TRW Final Report 26405-6001-RU-00. Redondo Beach, California, 1974.
2. Campbell, G.E. Geothermal Well Drilling and Completion Practices in
California Including Casing and Abandonment Programs and Examples of
Blowouts. Publication of the Geothermal Unit, California Division of
Oil and Gas.
3. TRW, Incorporated. Study of the Geothermal Reservoir Underlying the East
Mesa Area, Imperial Valley, California. TRW Final Report 28859-6001-RU-00.
Redondo Beach, California, 1976.
4. Final Environmental Statement for the Geothermal Leasing Program, Volume
1. United States Department of the Interior, 1973.
5. Miller, R.L. Results of Short-Term Corrosion Evaluation Tests of Raft
River. Idaho National Engineering Laboratory. Report TREE-1176, 1977.
6. Banning, L.H. and Oden, L.L. Corrosion Resistance of Metals in Hot
Brines - A Literature Review. U.S. Bureau of Mines Report IC-8601, 1973.
i
7. Miller, R.L. Results of Short-Term Corrosion Evaluation Tests at Raft
River. Idaho National Engineering Laboratory Report TREE-1176, 1977.
8. Status Report, Geothermal Resource Investigations, East Mesa Test Site,
Imperial Valley, California. U.S. Bureau of Reclamation, 1977.
9. Shannon, D.W. Corrosion of Iron-Base Alloys Versus Alternate Materials
In Geothermal Brines. Battelle Pacific Northwest Laboratories Interim
Report PNL-2456, 1977.
10. Kukacka, L.E. et.al. Alternate Materials of Construction for Geothermal
Applications. Brookhaven National Laboratory Progress Report BNL 50834,
1978.
11. Hirasuna, A.R. et.al. Geothermal Elastomeric Materials. L'Garde Inc.
Progress Report SAN/1308-1, 1977.
12. Lawrence Berkeley Laboratory. A Study of Brine Treatment. Electric Power
Research Institute Report ER-476, 1977.
43
-------
13. Makrides, A.C. et.al. Study of Silica Scaling from Geothermal Brines.
EIC Corporation Report COO-2607-5, 1978.
14. Daedalean Associates Inc. The Results of the Initital Feasibility Program
on Cavitation Descaling Techniques for Pipes and Tubes Used in Geothermal
Energy Plants. Daedalean Associates Report HCP/T2289-01.
15. Reactor Safety Study. United States Department of Commerce Report
WASH-1400 (NUREG 75/014), 1975.
16. Fulton, D.W. Non-electronic Parts Reliability Data. Rome Air Defense
Center Report NPRD-1, 1978.
17. Sung, R. et.al. Preliminary Cost Estimates of Pollution Control Techno-
logies for Geothermal Developments. EPA Contract No. 68-03-2560, T-5004,
U.S. Environmental Protection Agency, Cincinnati, Ohio, 1979. 131 pp.
44
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/7-80-024
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
SURFACE CONTAINMENT FOR GEOTHERMAL BRINES
5. REPORT DATE
February 1980 issuing date
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
R. Sung, W. Murphy, J. Reitzel, L. Leventhal,
W. Goodwin, and L. Friedman
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
1O. PROGRAM ELEMENT NO.
TRW, Inc.
One Space Park
Redondo Beach, CA 90178
1NE827
11. CONTRACT/GRANT NO.
68-03-2560
Work Directive T5003
12. SPONSORING AGENCY NAME AND ADDRESS
Industrial Environmental Research Lab. - Cinn, OH
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, OH 45268
13. TYPE OF REPORT AND PERIOD COVERED
Final Report
14. SPONSORING AGENCY CODE
EPA/600/12
15. SUPPLEMENTARY NOTES
16. ABSTRACT
This report examines the probability of significant releases of geothermal brine to
the surface environment through unplanned or accidental events. It then evaluates
the containment measures that may be used to prevent environmental damage. The
results indicate that major spills are likely to be very rare and that simple
warning systems and diked containment areas should provide adequate protection.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS C. COSATI field/Group
Geothermal prospecting
Electric power plants
Water pollution
Geothermal energy
Pollution control
Surface spills
8H
10A
13B
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (ThisReport)
UNCLASSIFIED
21. NO. OF PAGES
55
20. SECURITY CLASS (Thispage)
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (Rev. 4-77)
PREVIOUS EDITION IS OBSOLETE
:45
4 U S eOVEIWMEIIT PRINTING WFICE: I960 -657-146/S578
------- |