EPA/530/UST-89/012
      DETECTING LEAKS
Successful Methods Step-by-Step
              November 1989
       Office of Underground Storage Tanks
       U.S. Environmental Protection Agency
           Washington, D.C. 20460

-------

-------
PREFACE
            This handbook provides basic information on release detection that
            you—as state and local regulators—will find useful in developing and
            implementing the release detection portion of your program for regulat-
            ing underground storage tank systems (USTs).

            This information is meant to foster your understanding and use of the
            release detection methods appropriate for your individual UST
            programs.  The handbook contains information on the methods of UST
            release detection that were the most widely used at the time of publica-
            tion; inventory control, manual tank gauging, tank tightness testing,
            automatic tank gauging, vapor monitoring, ground-water monitoring,
            secondary containment with interstitial monitoring, and piping release
            detection methods.
                                          111

-------

-------
ACKNOWLEDGMENTS
       This document was written by Cecily Beall, Linda McConnell, Albert
       Nugent, and Julie Parsons for the U.S. Environmental Protection Agency's
       Office of Underground Storage Tanks (EPA/OUST) under contract
       No. 68-01-7383. Production assistance was provided by Maridene Amdt,
       Keith Cox, Donna Kirk, and Doris Nagel. The Work Assignment Manager
       for EPA/OUST was Thomas Young, and the EPA/OUST Project Officer
       was Vinay Kumar. Technical assistance and review were provided by the
       following people:

           Tom Bergamini - Wisconsin Department of Natural Resources
           Fermin de la Camara - Dade County, Florida, Environmental Resources
             Management
           Jon Gross - Nebraska State Fire Marshall
           Sav Mancieri - Rhode Island Department of Environmental
             Management
           Enemute Oduaran - Delaware Department of Natural Resources and
             Environmental Control
           Michael Randolph - San Jose, California, Fire Department
           Mike Scoggins - U.S. EPA Region VI
           Helga Butler - U.S. EPA Office of Underground Storage Tanks
           Steve McNeely - U.S. EPA Office of Underground Storage Tanks
           Peg Rogers - U.S. EPA Office of Underground Storage Tanks
           Phil Durgin - U.S. EPA Environmental Monitoring Systems Laboratory,
             Las Vegas, Nevada
           Tony Tafuri - U.S. EPA Risk Reduction Engineering Laboratory,
             Edison, New Jersey
           The Leak Detection Technology Association
           American Petroleum Institute

-------
                          Disclaimer

This document has been reviewed in accordance with U.S. Environ-
mental Protection Agency policy and approved for publication.
Mention of trade names or commercial products does not constitute
endorsement or recommendation for use. Other alternatives may exist
or may be developed.
                               VI

-------
TABLE OF CONTENTS
  I. INTRODUCTION
        Purpose of the handbook
        Contents of the handbook
        Use of the handbook
 1
 2
 3
 II. INVENTORY
        Summary
        Brief description
        Potential problems and solutions
        Ensuring effective manual tank gauging
        References
13
13
14
30
31
 III.  MANUAL TANK GAUGING
        Summary
        Potential problems and solutions
        Ensuring effective manual tank gauging
        References
33
34
43
45
 IV.  TANK TIGHTNESS TESTING
        Summary
        Brief description
        Potential problems and solutions
        Ensuring effective testing
        References
47
48
52
73
76
                                vu

-------
   V.  AUTOMATIC TANK GAUGING
         Summary
         Brief description
         Potential problems and solutions
         Ensuring effective automatic tank gauging
         References
 77
 78
 83
 92
 94
  VI.  VAPOR MONITORING
         Summary
         Brief description
         Potential problems and solutions
         Approaches to ensuring effective vapor monitoring
         References
 95
 95
 96
123
124
 VII. GROUND-WATER MONITORING
         Summary
         Brief description
         Potential problems and solutions
         References
127
128
130
159
Vin. SECONDARY CONTAINMENT WITH
       INTERSTITIAL MONITORING
         Summary
         Brief description
         Potential problems and solutions
         Ensuring effective secondary containment monitoring
         References
161
162
169
180
181
                                 viu.

-------
  IX.  PIPING RELEASE DETECTION METHODS

        Summary
        Brief description
        Potential problems and solutions
        Ensuring effective release detection for piping
        References
183
184
193
204
206
SUBJECT INDEX

APPENDIX A—LIST OF FIGURES
APPENDIX B—LIST OF TABLES
207


212


215
                              rx

-------

-------
 Chapter I
Introduction

-------

-------
INTRODUCTION
PURPOSE OF THE HANDBOOK
            This handbook provides basic information on release detection that
            you— as state and local regulators—will find useful in developing and
            implementing the release detection portion of your program that
            regulates underground storage tank systems (USTs).

            Even though tens of thousands of UST systems have had leaks or are
            currently leaking, most existing UST systems are not currently being
            monitored for releases. As a result, Federal regulations now require all
            UST systems containing petroleum or certain hazardous chemicals to
            have effective release detection.

            Although some states and local governments have had active UST
            programs for several years, most regulators are just beginning to  acquire
            the knowledge necessary to develop and implement UST management
            programs. This handbook supplies information useful in acquiring that
            knowledge.

            Because the Federal release detection requirements will be implemented
            through your state and local regulatory agencies, you need information
            during the development of your UST program to answer the following
            questions:

            •    What release detection methods are allowed?

            •    How do these methods basically work?

            •    What potential problems does each method have?

            •    What solutions to these problems are available?

            •    How can you make sure UST owners and operators are aware of
                these potential problems and respond to them?

            •    How can you make sure that the providers of release detection
                follow proper protocols and practices?

-------
            You will find enough basic information in this handbook to answer
            these kinds of questions as you develop and implement effective release
            detection programs.

            The handbook is not intended, however, to be a technical "how to"
            manual for the release detection methods. It should not be used to
            become familiar with the intricacies of different brands of test
            equipment.

CONTENTS OF THE HANDBOOK
            The handbook provides information about the following methods of
            release detection for tanks allowed in the final rule
            (53 FR 37196-37212):

            •    Inventory Control

            •    Manual Tank Gauging

                Tank Tightness Testing

            •    Automatic Tank Gauging

            •    Vapor Monitoring

            «    Ground-water Monitoring

            •    Secondary Containment with Interstitial Monitoring

            The handbook also contains information on the release detection
            methods allowed in the final rule for underground piping.

            The specific release detection methods allowed in the Federal rule have
            been demonstrated to successfully detect petroleum releases from UST
            systems and are currently allowed under several established state and
            local programs.

-------
            The chapters that follow describe the release detection methods allowed
            under the Federal regulation. Each chapter focuses on a series of topics
            presented in the same general order:

            •   Summary of the chapter

            •   Brief description of the release detection method

            •   Potential problems associated with the method

            •   Solutions to the problems

            •   Ways to oversee or enforce that proper release detection takes
                place

            •   Technical references

            The information in this handbook was initially gathered by EPA to
            support development of the final federal regulations. It was gathered, in
            part, from numerous visits to state and local programs and from
            experienced vendors of the various methods. Additional important
            technical information came from an extensive release detection research
            and development program conducted by EPA laboratories in Edison,
            New Jersey, and Las Vegas, Nevada, from 1985 through 1988.
USE OF THE HANDBOOK
            On the next page, you will find a generalized flow chart (Figure 1) of
            the process used by state implementing agencies to develop UST release
            detection programs.  This chart is based on information from several
            states that have already developed UST programs. Agencies typically
            begin by writing regulations on leak detection. The crucial decisions in
            this area are what methods to allow and what design and performance
            standards are necessary to ensure that the methods are effective.  Once
            the regulations are in place, the states face the issue of how to
            implement their requirements. States have developed a variety of
            oversight mechanisms to ensure that testers or installers follow the
            design and performance standards.

            The following sections provide expanded discussions of how you can
            use the handbook in developing a leak detection program as  outlined in
            Figure 1.

-------
  Regu ation
 Development
Implementation
  Activities
                       Write and Finalize Regulations
                                        Which methods to allow
                                        What restrictions on the methods
        What
   additional work
      is needed
to implement effective
  release detection
1
f

1
Develop
Procedures
for SITE
INSPECTIONS


1


REVIEW DATA
from Tests or
Monitoring Efforts


i

r
Develop
GUIDANCE
Materials


1
r
Develop
APPROVAL/
CERTIFICATION
Procedures

         Figure 1. Development of a state or local leak detection program

-------
 Which Release Detection Methods Should You Allow?
In writing the federal regulation, the U.S. Environmental Protection
Agency reviewed existing release detection methods and state UST
programs and selected the methods that were demonstrated to be
effective. States have the option of being more restrictive when
deciding which methods to allow within their jurisdiction. EPA
believes that, with appropriate restrictions and oversight, most, if not all,
release detection methods discussed in this handbook can be successful
for any given set of site conditions. Therefore, there should seldom be a
reason to exclude a type of release detection from a state program.
Many of the problems that are rumored about a method, such as
background contamination interfering with vapor monitoring, can often
be avoided or overcome. The chapters on the specific release detection
methods provide detailed discussions of solutions to potential problems
for both contractors and state officials.

None of the methods (including secondary containment with interstitial
monitoring) is fail-safe and assures detection of all leaks. EPA research
and state and local experiences in the field have shown, however, that
each of these methods has proven to be effective in detecting UST
releases when  used properly and within the inherent design limitations
of the method.  EPA has developed this handbook in the belief that a
better understanding of potential key limitations of each method will
convince you that each one can have an important role in your UST
program for detecting releases.

Although the handbook discusses the potential limitations or problems
with each method, it does not provide information on how frequently
these problems actually occur today in the field.  In fact, many of the
reservations identified in the handbook concerning a method's proper
use are already well known and carefully controlled for by experienced
providers of release detection.  Thus, many of the problems  identified in
the handbook may not need special state or local consideration or
oversight to ensure that a particular method is being effectively applied.

For example, numerous tightness testers have recently adjusted their
protocols and training materials to incorporate the lessons learned from
EPA's tank testing research completed at the Edison, New Jersey,
laboratory in 1988. Accordingly, many of the concerns raised in the
tank tightness testing and automatic tank gauging chapters of the
handbook are now well known in the release detection service industry

-------
and have been used recently to increase the quality of the testing
services and equipment being provided nationwide.

In spite of the basic technical soundness of the different methods,
however, new entrants or otherwise inexperienced members of the
rapidly developing release detection service industry may need some
state and local oversight and guidance to assure their methods are
properly applied. Likewise, typical UST owners and operators may
need some help in selecting and using the right release detection method
for their sites. This handbook will help you develop regulatory
programs that assure release detection is used effectively at all UST
systems.

Table 1 on pages 8 and 9 presents a summary of the important factors
that affect the success of each release detection method. Included in the
table are summaries of the important design or operational elements that
each method should include to account for specific site conditions.
These elements are explained more completely in the chapters on each
release detection method. As can be seen in the table, there are very
few site conditions that completely eliminate from consideration any of
the approaches to release detection.  Most methods will work at a site
given the selection of appropriate equipment and proper installation and
operation. Different site conditions favor different methods.  For
example, ground-water monitoring is more effective in areas with
shallow ground water and with a product that floats.  Methods that are
external to the tank itself, such as vapor monitoring or ground-water
monitoring, are more effective for large tanks. Methods that are internal
to the tank or its containment (interstitial monitoring, tank tightness
testing, automatic tank gauging, manual tank gauging, and inventory)
are more effective for highly contaminated sites. Tank owners and
operators will be best able to meet the regulations considering their
specific site conditions if all choices are left open to them. The
availability of a wide selection of release detection methods allows the
owner  or operator to get an effective leak detection system for the
lowest cost.

The information listed in the table is applicable to both tank and piping
release detection. However, while manual tank gauging and inventory
control will effectively detect leaks from tanks, the sensitivity of these
methods is very low when applied to piping.  Even the procedures
outlined in the table will not improve the performance of these methods
sufficiently to make them acceptable stand-alone release detection
methods for piping. The sensitivity of automatic tank gauging to piping
releases is still unproven.

-------
 What Restrictions Should You Place on the Methods?
As illustrated in Table 1, most methods will work in most situations as
long as proper procedures are followed. In developing their UST
programs, implementing agencies may elect to place restrictions on
allowable release detection methods to ensure maximum effectiveness.
Each chapter contains a table titled "Indicators and Solutions for
Problems" that summarizes the information presented in the chapter,
and the column titled "Solutions" summarizes the actions that are
necessary to ensure effective release detection with that method. This is
the information that should be evaluated when an implementing agency
is considering regulatory limitations on the use of a method. For
example, the chapters on ground-water and vapor monitoring include
monitoring well network design requirements from several existing state
UST programs that have been  confirmed as effective through a
combination of field experience and EPA research.
 What Agency Oversight Mechanism Should You Use?
Once leak detection regulations have been adopted, there are four basic
approaches that an implementing agency can use to oversee the work of
release detection testers and installers to ensure that appropriate
methods have been chosen and that the correct procedures are followed
by testers and installers (see Figure 1).  The "Indicators and Solutions
for Problems" table in each chapter contains a column titled "Agency
Oversight Options" that lists possible applications of these approaches
to the specific release detection method; the oversight options are
discussed in more detail in the text of the chapters. The following
discussions present general descriptions of the approaches and their
advantages and disadvantages.

Site inspections

     Some local agencies have found that having their staff present
     during release detection tests effectively ensures that tests are
     conducted properly. Before such an approach can be implemented,
     agency personnel have to be trained in proper procedures and
     develop either a checklist of important features to be examined at
     each site or an inspection procedures manual.

-------
Table 1.  Effect of Site Conditions on Success of Release Detection Methods for Tanks and Piping
                 Manual Tank
                 Gauging and
Site Conditions  Inventory
                      Automatic Tank
                      Gauging
                      Tank Tightness
                      Testing
                      Secondary Contain-
                      ment with Interstitial  Ground-Water
                      Monitoring           Monitoring
                     Vapor Monitoring
Ground water
Product type
Tank size
Shallow ground water
may interfere under
some conditions. Use
water-finding paste to
detect water in tank.
N/A
Less effective for
large tanks. Manual
tank gauging limited to
less than 550 gallons
when used alone or
less than 2,000 gal-
lons when combined
with tightness testing.
Shallow ground water
may interfere under
some conditions. Use
sensor to detect water
level in tank.
Effective for products
with viscosity and
thermal properties
similar to gasoline
and diesel fuel.
Shallow ground water
may mask a leak.
Can be used in high
ground water if
ground-water level is
measured and  product
level in tank is  raised
to overcome ground-
water pressure. Do
not test when ground
water is fluctuating.
Effective for products
with viscosity and
thermal properties
similar to gasoline and
diesel fuel.
Less effective for
large tanks. Generally
applicable to tanks
< 12,000 gallons.
Applicability to larger
tanks depends on
method and must be
demonstrated.
Less effective for
large tanks. Generally
applicable to tanks
< 12,000 gallons.
Applicability to larger
tanks depends on
method and must be
demonstrated.
                                                                                 Shallow ground water
                                                                                 may interfere with
                                                                                 sensors. Usefully
                                                                                 double-walled USTs,
                                                                                 excavation liners, or
                                                                                 jacketed tanks or
                                                                                 methods that are not
                                                                                 affected by water.
                                                                Depends on sensor
                                                                construction.
                                                                                 N/A
Deep ground-water    Shallow ground water
delays detection. Do  may interfere with
not use for sites where sensor. Do not use in
(a) ground-water depth saturated sites.
is > 20 ft or < 3 ft or (b)
ground-water fluctu-
ation exceeds well
screen interval for
more than 30 consec-
utive days. Place some
wells downgradient to
UST, if grade can be
determined.
Product must float on
ground water to be
detected (most petro-
leum products do).
N/A
Lower volatility products
delay and may prevent
detection. Effective for
gasoline. Response to
other products should be
verified. For less volatile
products, add tracer
compound, use aspir-
ated sensors and more
and larger diameter
monitoring wells, or set
lower alarm levels.

N/A

-------
 Backfill
 permeability
 N/A
 N/A
 N/A
Existing site
contamination
N/A
N/A
N/A
 With excavation liners,
 backfill must be
 permeable to allow
 release to be detected
 Use sand or pea
 gravel as backfill.

 Use monitoring
 method that can
 distinguish leak from
 existing contamina-
 tion or use full double-
 walled tank or
 completely sealed
 excavation liner or
 tank jacket.
                                                                                        Use sand or pea
                                                                                        gravel.
                                                                                        Floating product from
                                                                                        prior releases may
                                                                                        cause false alarm. Do
                                                                                        site assessment in
                                                                                        areas of suspected
                                                                                        background contam-
                                                                                        ination,  and use other
                                                                                        methods if contami-
                                                                                        nation would  cause
                                                                                        false alarm.
                                                                                        Use sand or pea
                                                                                        gravel.
                                                                                        Do not use at site with
                                                                                        high (background
                                                                                        concentration < 15,000
                                                                                        gallons ppm for gas-
                                                                                        oline) unless a manu-
                                                                                        facturer can show how
                                                                                        their device works at
                                                                                        high background levels.
                                                                                        Tracer compound may
                                                                                        also be used.
Temperature
Subsurface
conduits
Temperature differ-
ence between
newly delivered
product and product
in tank limits
accuracy. May use
simple temperature
measurement to
partially compensate.
N/A
Measure product
temperature for at
least three levels in
tank. Use longer
waiting times before
testing as tempera-
ture difference
between newly
delivered product
and product in tank
increases.
N/A
Requires frequent
measurement of
product temperature
for at least three
levels in tank; or that
the product mixed.
The greater the
temperature differ-
ence between added
product and tank,
the longer the wait
before testing.  A few
methods are indepen-
dent of temperature.

N/A
Freezing of interstitial   N/A
fluid in hydrostatic
systems will prevent
detection while high
temperatures may
cause false alarms
because of loss of
fluid through
evaporation.  Add
antifreeze in cold
weather and add
more fluid in hot
weather.
                                                                 Subsurface conduits    Subsurface conduits
                                                                                  should not be in
                                                                                  backfill. May cause
                                                                                  leak to migrate so
                                                                                  that it will not be
                                                                                  detected.
                                                                                       should not be in
                                                                                       backfill. May cause
                                                                                       leak to migrate so
                                                                                       that it will not be
                                                                                       detected.
Low temperatures
reduce sensitivity.
Install sensors below
frost line.  Use more
and larger wells or
aspirated systems to
compensate for reduced
volatility of product.
                                                                                       Subsurface conduits
                                                                                       should not be in
                                                                                       backfill. May cause
                                                                                       leak to migrate so
                                                                                       that it will not be
                                                                                       detected.
*N/A = No significant impact on test method.

-------
                 There are several advantages to having a knowledgeable regulator
                 observe the installation and/or operation of the test equipment.
                 Observation may cause the testers to be more conscientious about
                 following procedures.  If deviations from proper procedure occur,
                 the inspector can catch them immediately and require a correction
                 or retest.  An observer may be able to provide suggestions when
                 problems arise.  A third-party observer may provide additional
                 credibility to the test results, in essence "certifying" that the test
                 results are valid. Finally, if a release is detected, the regulator is
                 on-site and can begin investigation immediately. Site inspections,
                 however, are expensive and labor intensive and may create
                 considerable scheduling difficulties when a limited number of
                 people are available to conduct the site visits. The following
                 discussion on data reviews indicates a possible approach that
                 requires less resources.

              Data review

                 It is possible to requke that release detection installers, testers, and
                 tank owners keep detailed records of site conditions, events, time,
                 results, etc., during a test and submit these records to the
                 implementing agency. Regulators can then review all of the
                 reports, which would be somewhat time-consuming and expensive,
                 or a certain percentage of the reports, possibly on  a random basis.
                 All aspects  of the tests may be reviewed or, to save time, only
                 those aspects that have been determined to be most crucial for that
                 method in that locale.  The most important things to check for in
                 the installation, operation, and interpretation of each method are
                 listed in the method-specific chapters.  For example, jurisdictions
                 with a lot of clay deposits in the area may want to check the boring
                 logs for ground-water or vapor monitoring wells to verify that
                 product could get to the well if a leak occurred.  For test methods
                 that requke calculation, a manual or computerized check of the
                 calculations may be performed.

                 A program to review test data costs less in time  and money than a
                 site inspection program and has some of the same benefits,
                 although to a lesser degree. The obviously bad tests will be
                 identified with less effort than an inspection program would take.
                 Discovery of questionable results in a data review may be used to
                 target scarce inspection resources. There are two main drawbacks
                 to such an approach. Fkst, test reports may accumulate unread
                 while regulators work on problems  of more urgency. Second,
10

-------
  correcting a problem "after the fact" is more difficult than
  correcting it on-site as soon as it happens, both in terms of the
  regulator finding the time to do it and in enforcing additional action
  by the owner/operator or tester after the tester has left the site.  A
  program to mail back deficient reports, inspect sites with deficient
  reports, enforce retesting, etc., may be as expensive as doing the
  site inspections initially.

Guidance or training

  Release detection equipment installers and testers may conduct
  tests improperly out of ignorance of the proper procedures.
  Several implementing agencies have issued guidance to owners
  and testers that identifies the worst procedural violations and helps
  to ensure that release detection is effective. Guidance for testers
  and installers may be in the form of booklets, videos, or training
  seminars. Another approach to guidance would be to provide a
  checklist of important procedures for each release detection
  method to owner/operators for use during a test, so that they can
  knowledgeably observe installation and testing at their sites.

  This approach attempts to correct the problems before they occur
  and can reach a fairly wide audience at relatively small expense.
  By educating one tester or testing company, procedures at many
  future tests have been improved.  However, review of the guidance
  material and adherence to its recommended procedures is
  voluntary.

Approval/certification of release detection methods and personnel

  Licensing or certification of release detection equipment and/or
  operators is one potential means of ensuring valid test results.  The
  regulatory agency would have to set up some mechanism for
  reviewing and approving equipment and/or personnel. There are
  several ways in which equipment might be approved. A board of
  knowledgeable regulators can review written evaluations of a
  method's performance or hear presentations made by
  manufacturers, during which any questions can be answered. A
  regulatory agency may accept the findings of an independent
  third-party evaluation.  EPA is developing standard procedures for
  testing leak detection equipment that can be used by independent
  laboratories to evaluate the performance of commercial methods.
                                                                     11

-------
                 Approving leak detection equipment has two main advantages.
                 First, it allows the agency to make clear and helpful
                 recommendations to owners about what methods are acceptable.
                 Second, it allows the agency to evaluate the claims made by
                 manufacturers and ensure that owners and operators use effective
                 equipment. The EPA test procedures should allow regulators to
                 compare the performance of various methods that have each been
                 tested the same way. Licensing of leak detection testers or
                 installers is more difficult than approving equipment. There are
                 many more contractors than manufacturers and the skills required
                 are diverse.  Possible methods for screening personnel for licensing
                 include written examination, field observation, field test,
                 apprenticeship, and training requirements (manufacturer
                 certification). Any effective licensing/certification program also
                 requires a follow-up program to identify equipment or personnel
                 that is operating without a license. Also, some system of sanctions
                 against violators is needed.  Such follow-up programs also may be
                 time-consuming and expensive.  Studies of occupational licensing
                 programs in many areas have shown little improvement in service
                 quality after the program has been initiated, especially when an
                 examination is the only requirement. Controlling testers and
                 installers is important, however, because of the crucial role
                 procedure plays in determining leak detection effectiveness. Each
                 of the following chapters describes the available leak detection
                 methods and important aspects of procedure in greater detail.
12

-------
   Chapter II
Inventory Control

-------

-------
  INVENTORY  CONTROL
II
 SUMMARY
              To date, the practice of underground storage tank (UST) inventory
              control primarily has been limited to use as a business practice to keep
              track of product, and indirectly as a release detection method, for
              motor fuels stored at retail distribution facilities. Inventory control,
              also called inventory reconciliation is claimed to be the simplest and
              most economical leak detection method. The technique is effective for
              finding larger leaks (over 1 gal/h) if "recommended practices" are
              followed. Because inventory control has low sensitivity, EPA requires
              that it be combined with tank tightness testing. Recommended
              practices for inventory control can be found in the American
              Petroleum Institute's Publication API  1621, Recommended Practice
             for Bulk Liquid Stock Control at Retail Outlets. Parts of the following
              discussion are largely based on this publication.

              The discussion presented in this chapter covers many of the possible
              problems that may occur during inventory control. This does not mean
              that all, or even most, of these problems will occur.  Nor does it mean
              that all of the problems are of equal importance, in terms of frequency
              of occurrence or severity of impact on the effectiveness of inventory
              control. Some problems occur infrequently, while others have limited
              impact. This chapter presents a range of potential problems for
              educational purposes, not to imply that they will always occur.
BRIEF DESCRIPTION
             Inventory control is basically an ongoing accounting system, similar to
             a check book.  Its objective is to reconcile the inputs and outputs of a
             stored substance in a given UST with the volume remaining in the
             UST. Careful records of all product delivered, product dispensed, and
             daily tank inventories are recorded on a ledger-like form and
             reconciled on a monthly basis. The system "imbalance" at the end of a
             month, the difference between book inventory and measured
             inventory, is compared to a threshold value to help determine whether
             the imbalance signifies a leak.
                                                                                 13

-------
             Daily UST inventories are determined by using a gauge stick (similar
             to a yard stick). The stick is inserted vertically through the fill tube
             until the end of the stick touches the tank bottom.  When the stick is
             removed, the level of the product corresponds with a number on the
             stick (similar to the way a car's oil level is indicated on its dipstick)
             which, by using a calibration chart, can be translated to a volume of
             product in the tank. A calibration chart often is furnished by the UST
             supplier. The chart shows the number of gallons represented by each
             inch on the gauge stick. Each chart is calculated for a specific brand of
             tank with particular dimensions and capacity, and the chart used must
             correspond to the tank being gauged.

             Once the volume of product in the UST is determined, it is recorded on
             a ledger form as the UST's daily inventory.  The amounts  of product
             delivered to and withdrawn from the UST each day are also recorded.
             At least once each month, these data are compared to determine if the
             volume measured in the tank corresponds with sales and delivery
             records.

             The process  of obtaining inventory information and its reconciliation,
             can be divided into five steps:  (1) tank gauging—the process of
             measuring the stored substance or the water in an UST;
             (2) calibration—the correlation of a gauge reading with the proper
             calibration chart to determine the volume of the product in the UST;
             (3) tank stock control—the determination of the amount of product that
             was added to and withdrawn from the UST; (4) recording and
             reconciliation—the use of an accounting form to record and reconcile
             the information gathered;  and (5) interpretation—the determination of
             whether the result of a month of inventory records signifies an UST
             release. The relationships among these five stages are shown in
             Figure 2.
POTENTIAL PROBLEMS AND SOLUTIONS
             A number of factors can affect the accuracy of inventory control as a
             release detection method.  Often an apparent loss of inventory may
             occur even though the UST is sound; inventory records also may show
             an overall increase in product.  The following sections discuss
             problems and solutions related to each of the five steps involved in
             implementing inventory control. The order of discussion is not
             intended to prioritize the importance of the problems, rather it is
             intended to follow the order in which they would occur according to
14

-------
Testing
Analysis
                           Tank Gauging
                                     Product gauge
                                     Water gauge
                            Calibration
         Volume of product determined
          from calibration chart
                         Tank Stock Control
                                     Withdrawals
                                     Receipts
                      Recording & Reconciliation
                                 \
Interpretation
Leak

No Leak
    Figure 2.  General procedure for inventory control
                                                          15

-------
             the flow chart in Figure 2. The discussion of problems and solutions is
             summarized in Table 2 on pages 18 and 19, in the order of the
             discussion, and the most serious concerns have been marked by an
             asterisk. Some agency oversight options are offered for problems,
             when applicable, but not all of them need be undertaken.
              Tank Gauging
   "testing; f[  Assure the tank is gauged properly
   Analysis
If a tank is not gauged properly, the gauge reading will not
correspond well to the actual volume of product in the UST.
Accuracy will decrease if a gauge reading is inadvertently taken
with the stick slanted or resting upon a protrusion (e.g., an internal
support ridge) in the tank, or if the stick is "bounced" off the
bottom of the tank. Accuracy is increased by careful, vertical
insertion of the stick, by taking more than one gauge stick reading
and wiping off the gauge stick between readings, then carefully
noting and averaging the results.

According to API, to properly gauge a tank, the stick is placed into
the tank through one of the openings in the tank until its tip touches
the tank bottom. Some tanks have a separate opening, called a
gauge hole, used for gauging the tank. The tank fill pipe opening
can be used to insert the stick if there is no gauge hole. The stick
should be inserted at the same point in the gauge hole each time a
gauge is taken and should be held in a vertical position. After the
stick is quickly withdrawn, the product "cut" (wet mark left by the
product on the gauge stick) is read on the graduated scale to the
nearest 1/8 inch. Once the stick is cleaned by wiping the "cut"
with a cloth, the procedure is repeated. Both readings should be
recorded. The average of the measurements should be used to
calculate the product volume in the tank.

Some UST owners prefer that two, consecutive gauge readings be
taken at both open and close each day or before each change in
working shifts.  However, an accurate inventory reconciliation can
be obtained with only consecutive closing or opening gauges on  a
daily basis. That is, one or the other should be chosen and the tank
gauged consistently at that time. Additionally, the tank meter
should be read at the same time of day that the  tank gauges are
taken. Implementing agencies could provide some type of training
to owners and operators teaching a preferred method.
16

-------
Analysis
  I
Analysis
Analysis
Avoid damaging the tank from careless gauging

    Repeated gauging of a tank may wear a hole in the bottom of the
    UST.  When the gauge stick strikes the bottom of the tank, rust can
    be chipped off of the surface and expose new metal, thus allowing
    for quicker corrosion and the development of a hole in the tank
    base.  This can be protected against by careful gauging and by
    outfitting the tank with a striker plate. A striker plate is simply a
    layer of metal added to the tank to increase its strength in the area
    that comes into contact with the gauge stick. Unless the striker
    plate is of negligible thickness, whenever a striker plate is added to
    an UST, the end of the gauge stick must be modified by cutting off
    a length equal to the thickness  of the striker plate so that true
    volume conversions can be obtained.

Ensure accuracy of the product reading

    If the gauge stick is used to gauge gasoline or other volatile
    products, the side adjacent to the graduated side should be grooved
    every  1/8 inch to keep the product from moving up the stick past
    the measured level (creepage). Additionally, product-finding
    pastes, applied over the stick in a light, even film, can significantly
    improve the accuracy of gauging. Pastes improve adherence of
    product to the gauge stick and prevent creepage.  Product-finding
    pastes change color in the presence of product.

Water in the tank must be identified

    The presence of water in  an UST results in an inaccurately high
    gauge reading.  Water intrusion may indicate a leak, especially in
    high ground water situations. In addition, water may enhance
    corrosion.
              To identify the presence and measure the amount of water, a water
              gauge must be taken at least once a month.  A water-finding paste,
              which is unaffected by the stored product but changes color in
              water, is used to check for water at the bottom of USTs.
              Information on satisfactory pastes may be obtained from an
              equipment supplier.

              A water gauge is taken in the same manner as a product gauge;
              however, the length of time the stick remains in the tank should be
              monitored carefully. The immersion time for a water "cut" is
                                                                                17

-------
00
      Table 2. Indicators and Solutions for Problems Encountered During Inventory Control
      Problem
Indicators
 Tester Solutions
Agency Oversight Options
      * Assure tank is gauged
      properly.
Slanted stick. Dipstick not
wiped.  Large discrepancy
between consecutive gauges.
 Take consecutive gauge
 readings. Wipe dipstick
 between gauges.
 Provide training.
      Avoid damaging tank from
      careless gauging.
Loss of product.
 Add a striker plate to tank base.
 Gauge tank carefully.
      Assure accuracy of product
      reading.
No clear product line on gauge
stick.
 Use notched gauge stick. Use
 product-finding paste. Withdraw
 pole quickly.
      Water in the tank must be
      identified.
      Need to keep track of the
      temperature when taking
      gauges.
Water in gas.  Unexplained gain
in product amount.
 Large temperature changes.
 Take a water gauge using water-
 finding paste. Take water gauges
 after delivery.
 Record daily temperature next
 to gauge readings. Do not take
 meter readings shorty after a
 delivery.
 Require monthly water gauge.
 Check UST for water during
 site visits. Track customer
 complaints of water in gas.
      Account for product
      evaporated during delivery.
 Unexplained product losses.
 Use vapor control.  Use pressure
 relief valves.
 Check weights and measures
 seal at retail stations. Check
 calibration before further leak
 investigation.
      Assure the calibration
      chart corresponds to UST.
 Specifications on chart do not
 match tanks.
Ask tank manufacturer to provide
a chart that corresponds with their
tank.

-------
 Need to use the calibration
 chart correctly.
Assure the accuracy for the
pump meter.
 Need to unquantify
 withdrawals or additions to
 the LIST.
Imbalance in inventory.



Unexplained losses or gains.



Unexplained losses.
Use chart according to API
recommendations.
Calibrate pump meter. Read
meters when gauges are taken.
Quantify all losses as closely
as possible.
Review inventory forms.
Require pump meter
calibration.
Assure the data are recorded       Imbalance in inventory.
completely and correctly.
*Assure proper reconciliation     Imbalance in inventory.
of data.
Need to interpret data correctly.     Imbalance in inventory.
                                 Ask supply company for
                                 recommended recording
                                 practiice.
                                 Follow proper reconciliation
                                 process. Double-check
                                 calculations.
                                 Follow recommended
                                 interpretation process.
                                       Review inventory forms.
                                      Review inventory forms.
                                      Review inventory forms.
'Indicates the most significant problem.

-------
   Analysis

   Analysis
   approximately 10 seconds for light products such as gasoline and
   kerosene and 20 to 30 seconds for heavier products.

   The quantity of water in the tank is calculated using the same
   procedure described for product calculations using a tank
   calibration chart. If the test shows more than 1/2 inch of water,
   arrangements should be made for its immediate removal, the
   product supplier should be notified that their product may contain
   significant amounts of water, and further tests should be conducted
   to ensure that the tank is not leaking.

Need to keep track of the temperature when taking gauges

   Temperature increases or decreases can cause an expansion or
   contraction, respectively, of the product within an UST.
   Expansions and contractions cause level changes that may mask or
   imitate a level change due to a leak.  The difference between the
   temperature of the stored product and that of the delivered product
   will have the largest effect on the volume of product in the UST.
   However, the outside daily ambient temperature will also affect the
   system. On a daily  basis it is important to be aware of the apparent
   losses and gains of product that temperature changes may cause.
   Ambient temperature should be noted when recording gauge
   readings, for reference when interpreting results. Effects may also
   be minimized by gauging the tank each day at the same time and
   by not taking gauges immediately after product delivery.

Account for product evaporated during delivery

   The more volatile (tendency of a liquid to change to vapor) the
   stored product is, and the higher the temperature is, the greater
   effect evaporation will have on the system. Evaporation  losses
   primarily occur during UST filling.  For gasoline, these losses
   average about 0.0012 gallon lost per gallon of throughput. This
   effect can be minimized by using vapor control, such as  vapor
   recovery during filling, or by placing pressure relief valves on the
   tanks to reduce the pressure within the UST during filling, which
   reduces the amount of product that volatilizes.
20

-------
            Calibration
         J  Ensure the calibration chart corresponds with the UST
 Analysis
 Underground tanks are fabricated as production items. The tank
 manufacturer supplies charts intended to be used for all tanks of
 the same nominal dimensions and capacities.  Charts cannot
 account for variations in the position of a tank in the ground. For
 example, a tank may be tilted in the ground due to poor
 installation, settling, frost heave, etc. A tilted tank will vary from
 the chart in proportion to the degree it is tilted, and calibration
 charts cannot account for such variations. If a tank is tilted and the
 tank is being gauged from the middle, there is no effect. The
 amount of product will be overestimated if the tank is gauged at the
 low end and underestimated if the tank is gauged at the high end.
 The amount of overestimated or underestimated product will,
 however, be consistent as long as gauging is always done at the
 same end.  If the tank is tilted, it should be remembered as a source
 of error. It is possible to create a tank chart specific to a tank by
 adding small known volumes of product to an empty tank and
 measuring the product level and repeating these steps until the tank
 is full. This approach is very time consuming.

 The calibration chart's specifications should correspond to the
 UST's brand, size in gallons, and dimensions. Manufacturers
 should provide a calibration chart for their particular tank.  A
 sample calibration chart is shown in Figure 3.
 illMi&iil  Need to use the calibration chart correctly
Analysis
If the calibration chart is read improperly or the calculations are
done incorrectly, the volume of product in the tank may be
estimated inaccurately resulting in false alarms or missed
detections.
                                                                                21

-------
                                                     zz
essus
               *
                      M
                                       I:
                **. ^*
                             SB
                                          
                                             1
^8
S
        0
                             S



               s
                             >a


                   H

-------
33
34
35
36
37
38.
39
40
41
42
43
44
45
; 46
4?
48
43
SO
51
52
*».,, .
54
55
56
&
58
59
60
61
62
63
64
'. 36$
402,
415
428
440-
452
464
478
486
497
507
516
524
532
539
544
548
549














'm
.'., 73%
755
,776
, .80*
825
844
, , 865
i . ,. 885
90*
922
338
954
, 968
• 380
,990
997
999

"•
"











52*
341
56*
580
600 .
620
63S
659
678 r
69?
?1£
734
753
774
7S9:
806
823
849
856
872
887
302
&&
930
943
955 -
966
976
985
393
999
1002
* 784
at*
841
en '
901
930
953
988
•tet?
, -KJ46
1«74
tf02
1130
«S7
$183
tate
t23S
1260
1285
1308
1331
t353
1375
1895
1414
1432
1449
1464
1478
1490
1439
1504
1042 - - :
1982
1122
1161
12Ot
1240
1279
13*8
1 1356
, *39S
1432
1469
1506
1542
1578
1613
f647
1680
1713
1745
1775
18Q5
1833
i860
1886
1910
*932
1953
. 1S7t
1£8B
1998
2005
1303-
1353,
1482
1452
1501
I55G
1593
1648
f'O9&
1743
179O
1837
1883
1928
1973
2016
2059
2*01
2141
2181
22te
225$
2292
2325
2357
2388
2410
2441
2464
2483
2498
2506
1563 V
1623
1683:
1742
1802
1S6S
ISfSf
1977
2035
, 2092 , '•
2149
2204
2263
2314
. 2367
2420
'2471
2521
, 2570
26*7,
2663
2707
275ft
2794
2829
2865
2899:
2929
2957
2983
2998
3008
,2d8$
-2*6*
2244
23©3
2402
2481
1 2553
f 2637
[ 2713
, 2890
2865
2939
t 3013
1 3085
3456
3226
3295
3364 -
3426;
3490
355?
3610
3667
3721
3772
3823
3865
3906
., 3942
3973
• ^ 3937
4010
Figure 3. Sample Calibration Chart  Source: Buffalo Tank

-------
             The following procedure is excerpted from API publication 1621:
               , " l^e obgrtsbdBjd, be «#d directly Sot & gauges which we ta the exact inch or
               /Jeitifib^ove^icfa^oww^etlocbipafc,   \   ,   /'          ^   T

               ,- ' For gauges of 1/& inch (or wore) over or oflder iftte ex^ot Jacli the toll Wiag
                                   '                  "    *        "
                  .  ,           --"  \-              ^-    ••'
                  a. ^jU , for a 1,000-gaiioft tanit (diameter ^4 inches; lengtb
                                                            '
                             s Chart reading at 47 inches » 7J
                        *'  ,  CSiait reading at 46 iaches « 771 gaKlons
                                  ,     ,%,^,  ^      ,
                            ,-  IS gallons times 3/4* 13-5
                        '             %             ''"'•••••£
                   d. Tbis gattoaage Calculated jn step cVis added to the ga^oaage shown'ba
               tt& chart jfor the Jower whole inch reading, i^
"  ' "Total
                                                  7845 gallons
              Therefore, a tank gauge of 46 3/4 inches, for the given UST, represents
              784.5 gallons of product. Some companies offer computer programs
              that perform these calculations automatically.

              If a water gauge has been taken, the quantity of water contained in the
              tank also is determined by using the above procedure. The total
              amount of water should be subtracted from the total amount of liquid
              in the tank (as determined in step d) to determine the net gallons of
              product contained.
24

-------
             Tank Stock Control
  testing 'I  Assure the accuracy of the pump meter
    I
|  Analysis
  Analysis
    It is impossible to calibrate dispensing meters to be 100 percent
    accurate. Pump meter inaccuracy will cause a consistent, apparent
    loss or gain in inventory depending upon whether the meter is slow
    or fast.  Once a meter has been calibrated, the system error can be
    limited during reconciliation.  The pump meter must be maintained
    properly and calibrated frequently to keep it as accurate as possible
    and to determine the degree and direction of error the meter may be
    causing. All dispensing meters at retail outlets must be calibrated
    to weights and measure standards of the locality. A procedure for
    testing the accuracy of a dispensing meter is included in API
    publication 1621.

Need to quantify all withdrawals or additions to the UST

    Unaccounted for additions and withdrawals to the UST will cause
    an imbalance in inventory reconciliation. Withdrawal sources
    include all product withdrawn for personal use, any spills during
    delivery of product or at other times, and any thefts. For the sake of
    accuracy, withdrawals not shown by a meter should be quantified
    as carefully as is possible and included in the inventory records and
    reconciliation process.

    All deliveries to the tank must be carefully recorded. A receipt
    showing the delivery amount should be kept for the inventory
    records. To check that the received product amount corresponds
    with the receipt amount, the delivery receipt should be reconciled
    with tank gauges taken immediately before and after delivery,
    noting any withdrawals that occurred during the delivery.
             Recording and Reconciliation
   1
 Analysis
Ensure the data are recorded completely and correctly

   Recording inventory data is the first step in the reconciliation
   process.  The format for data entry varies greatly depending upon
   owner preference, number of USTs at a facility, and the
   inter-relation of these USTs. Because of the many different
   accounting systems that can be used to record and analyze the
                                                                                 25

-------
                 inventory data, numerous different accounting forms are available.
                 Many oil supply companies advise operators as to the proper
                 accounting procedures to be used, including the use of suitable
                 accounting forms and where they may be obtained. All inventory
                 records should have a place to record daily receipts, daily
                 withdrawals, and the volume of product associated with the closing
                 inventory stick reading.  A sample accounting form is shown in
                 Figure 4.

             Ensure proper reconciliation of inventory data

                 During inventory reconciliation it is easy to make mistakes that
                 may improperly indicate that a tank is or is not sound.  The basic
                 formula for daily reconciliation of an UST's inventory is as
                 follows:

                    Opening Inventory + Deliveries
                    - Sales - Unmetered  Use - Closing Gauged Inventory
                    = Daily Overage or Shortage

                 Overage means that the gauged inventory is more than that which
                 is accounted for by deliveries, sales, and other use. An overage is
                 indicated by a final positive number. A shortage is indicated by a
                 negative result, and suggests that the inventory remaining in the
                 tank is less than that accounted for by deliveries, sales, and other
                 use.  This overage or shortage is recorded on a monthly
                 reconciliation worksheet (see Figure 5 for an example).  Inventory
                 monitoring records from an UST that does not have a leak should
                 have daily overages and shortages, that fluctuate randomly around
                 zero.  Large overages or shortages for one day, when no history of
                 overages or shortages exists, should not be a cause for alarm unless
                 similar results are obtained in future testing.

                 Sophisticated statistical  analyses can be performed to reconcile the
                 daily inventory records.  These methods increase the sensitivity of
                 inventory control and may identify tank conditions interfering with
                 accuracy (e.g., tilted tanks).  Owners/operators can purchase
                 statistical reconciliation  services from outside companies; such
                 analyses typically cannot be performed by in-house personnel.
26

-------
Sample Inventory Control Program DAILY Reconciliation Form
Location Date:
•:, , Regular Regular Premium
Spofcnljnvantprsr,,,,,. Leaded Unleaded Unleaded Diesel
1,: Closing " 1 4-
2, 2 *
3. . 3 4-
4, ' 4 4-
*> 	 , - 5 *
& 	 B 	 , 4-
7, "~ 7 " +
8. 8 +
9. 9 4-
10. ' fO 4-
1.1, Total Meters ±=
12. Meters Out -f
13, Meters In -
14. Dispenser Cal Test
^allonsi
I5ti Jotal Cfosfna Meters
16. Opening Meters -
17, Today's Sales -
••

































•.











'






-

''





*






Regular Regular Premium
Leaded Unleaded Unleaded Diesel
Physfcaltnventory mk Gal, In. Gal. In. GaL In. Gal.
i.$.,.T.anki Product ,
i&tftiuri Water ,
20. Tank 1 Met
21. Tank a Product
22. Tank 2 Water
23. Tank 2 Net
24* Physical Inventory




























Regular Regular Premium
Tank Reconciliation (gallons) Leaded Unleaded Unleaded Diesel
25,, Opening Physical Inventory
26, Today's Sales
27. Product Receipts 4-
28, Inventory Balance «
29> Pnvsical Inventory -
30. Tank Over (if 4-)
31, Tank Short (if -)



























-
Figure 4. Sample of daily reconciliation form
                                                                    27

-------
Sample inventory Control Program MONTHLY Reconciliation Form
-ocatlon , - "''"''', Dalev """ ,: --"-'.

-

s - „ . ~ "^ Dally Overage/Shortage
Regular .Regular
Jn'e Day Headed. . Unleaded

^


















- • • 5










1
2
3
4
1
2
3
4
5
6
7
8
9 "'
10
11
12
13
14
1S ^
16
17
18
id
20
21
22
23
24
25
26
27
28
29
30
31
Cum. Over. Total
% thru.
Cum. Shrt.Total
% Thru.
••
« -.;-,/

-
••

- - ,
< -, f

-
- ,^.
,,v,,.



" ' '
- '
•* f
f

-
,v ,

. " '

:
>









' "' S
" '
' ,'/'" '
'" '

"
... svJ^
" '" , ' '
"
-

,,"•-,.,
- v ^'4- -;
f S /
,,,, + +
f fffff
•i '"•
% •> •> •>* I, ^
1J < s s'
•<• f
„«-,'"„,'




"

,


s, i



Gallons i
Premium
Unleaded Diesel

•.


_.
s s
,
--"
f *

•>


'
,
"•/,
V.


_,_,
-

'























,

-
<










'









Attention: the cumulative sum of monthly overages or shortages should I not exceed 1,0%
Df the monthly throughput plus 130 gallons! 	 " 	 »-•—.«.' ^^, <^.~ 	 	
   Figure 5. Sample of monthly reconciliation form. Source: API
28

-------
           Interpretation
Analysis
Need to interpret the reconciled data properly

   Improper interpretation may result in an inaccurate indication of
   the UST's status.  The federal regulation requires that an UST be
   reported to the local implementing agency as leaking when
   monthly reconciliation for two consecutive months indicates there
   is a cumulative monthly overage or shortage equal to 1.0 percent of
   total flow-through for the month plus 130 gallons. Using the form
   shown in Figure 5, the following explains the steps for determining
   this result:

   Step 1.  Enter the daily overage or shortage for each product, as
   determined during manual reconciliation, for each operating day of
   theanonth.

   Step 2.  When the daily overages and shortages from 31 operating
   days (or one continuous month) have been entered, add all
   overages and shortages and enter the value on line I or 3, as
   appropriate.

   Step 3. Calculate the total flow-through for the month for each
   tank. (The total flow-through volume can be either the sum of the
   monthly pump readings or the total amount of product delivered in
   a month. Whichever method is chosen should be used
   consistently.)

   Step 4. Calculate one percent of the total flow through and add
   130 gallons to it to determine a comparison number, using the
   following formula:

      (0.01 x flow - through) + 130 gallons = comparison number

   Step 5.  If the cumulative overage or shortage (determined in
   step 2) exceeds the comparison number (calculated in step 4), a
   leak in the UST system may  be present. (It is likely that a
   mathematical error, pump miscalibration, unaccounted for delivery,
   or unaccounted for water is responsible for the discrepancy.)
   These results should be rechecked carefully, remembering that
   water may still be an indication of a leak. If the subsequent
   month's results again exceed the comparison number, the results
                                                                              29

-------
                must be reported to the local implementing agency as a possible
                leak.

                This process should be performed individually for each tank.

ENSURING EFFECTIVE MANUAL TANK GAUGING
             Chapter 1 provides a general description of the types of oversight that
             can be used. The following sections discuss how these approaches
             may be applied specifically to inventory control.
             Site Inspections
             Site inspections to review the inventory records is a possible oversight
             mechanism. It would also be possible to observe the staff performing
             the actual measurements, to see if the tank gauging is performed
             correctly.
             Data Review
             Although all inventory control recording forms could be reviewed for
             proper recording and interpretation of gauging data, this would entail
             reviewing numerous forms. However, a smaller number of forms
             would need review if submission of forms was only required when a
             leak is suspected. Forms should be reviewed for accuracy in
             recording, in conversion of gauge measurements to volumes and to
             ascertain correct reconciliation.
             Guidance and Training
             Training could be provided to teach owner/operators proper tank
             gauging techniques and how to conduct proper inventory control. An
             alternative to training classes is guidance materials (e.g., a manual or a
             video tape). The API 1621 is an excellent resource for tank owners
             using inventory control.
              Approval and Certification
             Most errors in inventory control take place because the tank is gauged
             improperly. An implementing agency could provide training and
30

-------
             certification for all persons using inventory control. Training should
             cover the proper tank gauging technique, recording, reconciliation and
             interpretation of data.
REFERENCES
              1. American Petroleum Institute. 1987. Recommended Practice 1621,
                Bulk Liquid Stock Control at Retail Outlets.

              2. American Petroleum Institute. June 5,1987.  Review and Analysis
                of Existing and Proposed Underground Storage Tank Inventory
                Control Procedures, Vol. 1. Report by Radian Corporation, for the
                American Petroleum Institute.

              3. Mobil. 1984. Motor Fuel Inventory Verification Procedures.

              4. Radian Corporation. August 1984. Analysis of Factors Affecting
                Service Station Inventory Control.

              5. Schwendeman, Todd G. and H. Kendall Wilcox. 1987.
                Underground Storage Systems. Lewis Publishers, Inc., Chelsea,
                Michigan.

              6. U.S. EPA. January 1986. Underground Tank Leak Detection
                Methods: A State-of-the-Art Review. Report by Shahzad Niaki and
                John A. Broscious for Hazardous Waste Engineering Research
                Laboratory, Office of Research and Development, U.S. EPA.

              7. U.S. EPA. September 1988. Analysis of Manual Inventory
                Reconciliation. Report by Richard F. Eilbert of Entropy Limited
                for Midwest Research Institute for the Office of Underground
                Storage Tanks, U.S. EPA.
                                                                                31

-------

-------
     Chapter III
Manual Tank Gauging

-------

-------
MANUAL TANK GAUGING
III
SUMMARY
            Manual tank gauging, also called static tank testing, is an effective,
            easy, and inexpensive release detection method for small volume USTs.
            A study by EPA shows that manual tank gauging can detect leaks as
            small as 0.2 gal/h for tanks less than or equal to 550 gallons in capacity.
            The same study shows that for tanks of 551 to 2,000 gallons, manual
            tank gauging has about the same sensitivity as inventory control. These
            attributes make it a very appealing release detection method for smaller
            UST operators.

            The discussion presented in this chapter covers many of the possible
            problems that may occur with manual tank gauging. This does not
            mean that all, or even most, of these problems will occur. Nor does it
            mean that all of the problems are of equal importance, in terms of
            frequency of occurrence or severity of impact to the effectiveness of
            manual tank gauging. Some problems occur infrequently, whereas
            others have limited impact.  This chapter presents a range of potential
            problems for educational purposes, not to imply that they will always
            occur.

            Manual tank gauging is a weekly, short-term static test in which the
            liquid level is measured in a quiescent tank at the beginning and end of
            a 36-hour time period. Any change in liquid level is used to calculate
            the change in volume, which is compared against established guidelines
            to determine whether any disagreement in the measurements is
            significant enough to indicate a leak in the UST system.  Manual tank
            gauging is sometimes confused with inventory control. Although both
            methods involve "sticking the tank," manual tank  gauging is a
            short-term static test, while, in contrast, inventory control is an ongoing
            record of all the activities at an operating UST for an entire month (for
            more information on inventory control, see Chapter 2 of this manual).
            Because the problems with inventory control and manual tank gauging
            are similar, Chapter 2 can used as a cross reference for this discussion.
                                                                                33

-------
            The process of manual tank gauging involves the following steps:
            (1) tank gauging—the process of measuring the product level in an
            UST; (2) calibration—the correlation of a gauge reading with the
            proper calibration chart to determine the volume of the product in the
            UST; (3) recording—accurately recording gauging results; and
            (4) interpretation—the determination of whether the result of tank
            gauging signifies an UST release. The relationships among these four
            steps are shown in Figure 6.

POTENTIAL PROBLEMS AND SOLUTIONS	
            A number of factors can affect the accuracy of manual tank gauging as a
            release detection method.  The following sections discuss problems and
            solutions related to each of the four steps involved in implementing
            manual tank gauging. The order of discussion is not intended to
            prioritize the importance of the problems, rather it is intended to follow
            the order in which they would occur according to the flow chart in
            Figure 6. The discussion of problems and solutions is summarized in
            Table 3, in the order of the discussion in the text, and the most serious
            concerns are marked with an asterisk. Some agency oversight options
            are offered for problems, when applicable, but not all of them need be
            undertaken.
             Tank Gauging
    I
  Analysis
Ensure the tank is gauged properly

       If a tank is not gauged properly, the gauge readings will not
       accurately reflect the amount of product in an UST. An
       inaccurate measurement will occur if the gauge is read
       incorrectly, or if the gauge is improperly taken with the stick
       slanted or resting upon an extension in the tank. To take a gauge
       properly, the stick is placed carefully into the tank through one
       of the tank openings until its tip touches the tank bottom.  Some
       tanks have a separate opening, called a gauge hole, that should
       be used for this procedure; otherwise the fill pipe can be used.
       The stick should be inserted at the same point in the gauge hole
       each time a gauge is taken and should be held in a vertical
       position. The stick should not rest on a projection on the tank
       bottom (e.g., a reinforcement rib in the base of a fiberglass tank).
       After the gauge stick has been wiped off, the gauging procedure
       should be repeated and both readings should be recorded. The
34

-------
 Testing
Ana ysis
                      Tank Gauging
                                Measure product level
                       Calibration
          Determine volume of product
          from calibration chart
                       Recording
                                Gauging results
                                Temperature
Interpretation
Leak

No Leak
Figure 6. General procedure for manual tank gauging
                                                        35

-------
OJ
o\
Table 3. Indicators and Solutions for Problems Encountered During Manual Tank Gauging

Problem	Indicators	Tester Solutions	Agency Oversight Options
      'Assure tank Is gauged
      properly.
      Avoid damaging tank from
      careless gauging.

      Assure accuracy of product
      reading.
      Assure sufficient testing
      period.
      Identify water in the UST.
      Need to keep track of
      temperature fluctuating
      during the test period.

      Assure the calibration chart
      corresponds to UST.
      Need to use the calibration
      chart correctly.

      Record the data completely
      and correctly.
      Need to interpret the data
      property.
                                Slanted stick. Gauge stick not
                                wiped. Large discrepancy
                                between consecutive gauges.

                                Loss of product.
                                No clear product line on gauge
                                stick.
                                Time elapsed < 36 hours.
                                Gain in product amount.
                                Large temperature changes.
                                Specifications on chart do not
                                match tanks.
                                Imbalance in results.
                                Imbalance in results.
                                Imbalance in results.
Take consecutive gauge
readings. Wipe dipstick
between gauges.

Add a striker plate to tank
base. Gauge tank carefully.

Use notched gauge stick. Use
product-finding paste.
Withdraw pole quickly.

Wait 36 hours or more between
beginning and ending gauge.
Take a water gauge using water-
finding paste.

Take gauges at same time: of
day. Record daily temperature
when  gauging.

Ask tank manufacturers to
provide a chart that
corresponds with their tank.

Use chart according to API
recommendations.

Ask supply company for
recommended recording
practice.

Follow recommended
interpretation process.
Provide training.
Set a mandatory waiting period
Check beginning and ending
times of test.

Check during site visits for water
 in tank.

Require recording of
temperature on gauging
records.
Review gauging forms.


Review gauging forms.



Review gauging forms.
      "Indicates the most significant problem.

-------
Analysis
       average of the measurements should be used in conjunction with
       a calibration chart to calculate the product volume in the tank.
       After at least 36 hours, two more measurements should be taken.
                       ';•"••    -              •   \   * •
       Because two gauges must be taken consecutively at both the
       beginning and end of the time period, an error in stick placement
       will be apparent if the gauges differ by more than a few gallons.
       In this case, a third gauge should be taken to determine which of
       the first gauges is correct.

Avoid damaging the tank from careless gauging

       Repeated gauging of a tank may wear a hole through the bottom
       of the UST. When the gauge stick strikes the bottom of the tank,
       rust can be chipped off of the surface and expose new metal,
       thus allowing for quicker corrosion which may result in a hole.
       This can be protected against by careful gauging and by,
       outfitting the tank with a striker plate. A striker plate is simply a
       layer of metal, added to the tank to increase its strength in the
       area that comes into contact with the gauge stick.  Unless the
       thickness of the striker plate is negligible, when a striker plate is
       added the end of the gauge stick must be modified by cutting off
       the exact length as the thickness of the striker plate so that true
       volume conversions can  be obtained.          :  . *        ,
           Ensure accuracy of the product reading
^

Analysis
                 If the stick is used for gauging gasoline or other volatile
                 products, the edge of the stick adjacent to the graduated side
                 should be grooved every 1/8 inch in order to keep the product
                 from moving up the stick past the measured level (referred to as
                 creepage). If desired, product-finding pastes can be used to
                 improve the accuracy of gauging. These pastes improve
                 adherence of the product to the gauge stick and prevent creepage
                 that would distort the reading. Product-finding pastes change
                 color in the presence of product, making it easy to identify the
                 line left on the stick by the product. Information on satisfactory
                 pastes may be obtained from an equipment supplier.

                 If product-finding paste is used, it should be applied in a light,
                 even coat to the stick before insertion into the UST. After the
                 stick is quickly withdrawn (to avoid creepage of the product),
                 the product "cut" (the mark left by the product on the stick) is
                 read on the graduated scale to the nearest 1/8 inch.
                                                                                  37

-------
  Analysis |
f

|  Analysis |
    I
  Analysis \
Assure sufficient testing period

       Manual tank gauging should take place weekly over at least a
       36-hour period during which no liquid is added or subtracted to
       the tank. Because a shorter time period for testing reduces
       manual tank gauging's ability to detect UST leaks, the 36-hour
       period is required by the federal regulations. As the testing time
       period is extended, manual tank gauging is able to identify
       smaller leaks. Over a 36-hour testing period, manual tank
       gauging should be able to accurately identify a leak of 0.2 gal/h.

Determine presence of water in the UST

       A water gauge may be taken to determine whether water is
       present in the UST (see Chapter 2). The presence of water in an
       UST may indicate a leak. In addition, good management
       practice dictates the detection and removal of any water in an
       UST because  water may enhance corrosion.  A water-finding
       paste, which is unaffected by the stored product, but which will
       change color in water, is used to check for the presence of water
       at the bottom  of USTs.

Keep track of temperature fluctuations during the test period

       Changes in temperature can affect the volume of the stored
       product and the UST; temperature increases and decreases can
       cause expansion or contraction, respectively, of the product
       within an UST. An increase in volume due to a temperature
       increase may  mask a leak. Similarly, a decrease in volume due
       to  a temperature decrease may imitate a leak. Temperature will
       have the least effect on dense liquids (e.g., used oil), which
       expand and contract a very small amount per degree of
       temperature change.

       Because tank gauging is done over at least a 36-hour time
       period, during which product is not delivered or removed,
       temperature effects should be due only to ambient temperature
       changes.  To minimize the impact of temperature, if the change
       in temperature is great, the testing period could be lengthened to
       48 hours so that the beginning and ending measurements of the
       gauge can be  taken at the same time of day.  Temperature effects
       should be kept in mind as a potential source of error.
38

-------
             Calibration
  I
Analysis
Analysis
Ensure the calibration chart corresponds to the UST being tested


       A calibration chart shows the number of gallons in the tank as
       represented by inch marks on the gauge stick. Each chart is
       calculated for a specific brand of tank of particular dimensions
       and capacity, and the chart used must correlate to the tank being
       gauged. If a chart is used that does not correspond properly, the
       tank volumes determined will not be accurate. Manufacturers
       will usually provide a calibration chart for their particular tank.
       Figure 7 on pages 40 and 41 shows a sample calibration chart.
       See Chapter 2 for additional discussion of calibration charts.

Use the calibration chart correctly


       If the calibration chart is read improperly, the volume of product
       in the tank may be estimated inaccurately.


       API publication 1621 recommends the procedure shown below:
                    I -.  the chart should be read directly for all gauges which aw to the exact inch or
                    1/1$ inch above or below an exact inch mark,

                    %.   JFof gauges of 1/8 inch (at more) over o=f under the exact inch the following
                    procedure should be used;

                        «.  the chart should be read fo* toe exact inch on the scale above and below
                    tfoe actual gauge stickjeading. Pot example^ if $» gauge sttck *e*te 46 3/4 inches,
                    the chart should be read at both 46 and 47 tacfaes.

                        b,  The smaller gallonage shown oft the scale at these two readings should be
                    subtracted from the larger ie., fo* a 1,000-gallott tank (diameter 64 laches; length,
                    72 inches);
                                    Chart reading at 47 inches s* 7$9 gallons
                                    Chart reading at 46 inches «* 771 gallons
                                    Subtracting             s* 18 gallons

                        c.  tliis gallouage is then multiplied by the fraction of an inch shown on the
                    original gaugevi.e.;
                                     18 gallons times 3/4 * 13.5 gallons.

                        d,   This gallonage (calculated in step c) is added to the gallonage shown on
                    the chart for the lower whole inch reading, ie<:

                                      Gallons at 46 inches * 771  gallons
                                      Gallons at Ifl inch **.1§«5 gallons
                                     Total          '  *> 784J gallons
                                                                                        39

-------
                13:
                                                         r*
                                                                                                 ft
            I
O
1
                                                                                 I
                 I
                                 
-------
33
34
3S ,
3S
37
35
39
4ft
41
, &
43
44
4S
46
, 4T
4ft
49
50
51
52
53 i
54
SS
56
S7
58
sn
€0
s fit
.62
63
64
* 389
402
415
428.
44ft
45?
464
47«
486
ACff
SJ7
516
524
so*
533
S44
548
549













:
7Q7 . '
733
755
77S
an*
823
844
res
885
fl(14
9?S»
938;
€64

916
936
943
ass
fl«ft
S7«
flfiR
993
fl9ft
toee
7W
Hit
841
871
901
S3?
959
98S
101T
^fl4ft
1P74
11S2
1130-
1157:
visa
191ft
1535
126ti
fsas.
^sflft
1331
43S3
1375
1395
*4t4
143>
1449
1464
147S
149^
1499
J504
4042-
fS8£ .
11S&
1*S»
126*..
1240
1279
131$
1356 .....
13Q«i
1435
1469
1506 ,
154?
157S
•Jft*S
1647
168Q:
1713
1745
177S
1885
1833
1860
1886
iffto
ia<^
««a
197*
1986
199R
2005
1303
1353
1402
1452
1SQ1
1550
1599
1648
1696
1743
179O
1837
1883
1938
1973
9f«R
3059
2131
3141
21fft
2219
2256
2292
2325
2357
2388
941R
2441
2454
2483
5498
2505
1563
1623
1683.
1742
180?
1860
1919
1977
2035 .
9
214S
2204
226S
2314
2367
5>45>O
2471
2521
257O
, 26*7
2669.
2707
27SQ
2794
2829
2865
9ff9O
292S
295T
2980
299^
3008
2085
31$4
2244
2323
24t»
2481
?55ft
P&S7
271$
9ftOft
28S5
2939
3013
3O^
3156
f&Ofi
329S
3361
342,5
349O
3551
3610
^667
3721
3772
382p
SftRfi
390$
3S4«
«973
3897
40*6;
Figure?. Sample of calibration chart. Source: API

-------
                   Therefore, a tank gauge of 46 3/4 inches, for the given UST,
                   represents 784.5 gallons of product. Computer programs are
                   available to perform or check these calculations.
             Recording
  Analysis
Record the data completely and correctly

      Tank gauging data should be carefully recorded. The format for
      data entry will vary greatly depending upon preference, number
      of USTs at a facility, and the interrelation of these USTs. All
      tank gauging recordkeeping systems should have a place to
      record all tank gauge readings, thek average reading and
      associated volume, the time the gauge readings were taken, and
      a final comparison of the beginning and ending volumes. It may
      also be helpful to record the ambient temperature at the time of
      the readings. This information might be useful when
      investigating a possible release because temperature changes can
      be a source of error.
             Interpretation
  .Testing :J   Need to interpret the data properly
  Analysis
      If tank gauging data are not interpreted properly, the tank may
      falsely be considered sound or leaking. All tanks using manual
      tank gauging must test weekly and check the average of the
      differences from the four previous weeks against the chart in
      Table 4.
                                            Table 4
                                    Monthly and Weekly
                             Manual Tank Gauging Standards
                                                           Monthly Standard
                     Nominal Tank      Weekly Standard    (four-test average)
                   Capacity/gallons      (one test) gallons         gallons
<550
551 - 1,000
1,001 - 2,000
10
13
26
5
7
13
42

-------
                   After the tank is gauged for a given week, and the beginning and
                   ending volumes for the time period have been determined, these
                   two volumes must be compared to one another.  The beginning
                   volume should be subtracted from the ending volume, and the
                   difference (including if it is a positive or negative number)
                   should be recorded. If the difference is positive, the UST has an
                   apparent gain. If the difference is negative, the UST has an
                   apparent loss. The difference should then be compared to the
                   information provided in Table 4 which correlates to weekly
                   results. If either a positive  or negative difference is greater than
                   the numbers provided in Table 4, the tank has failed for that
                   week.

                   Manual tank gauging results must also be checked against a
                   monthly standard. To calculate an average monthly value, the
                   four previous weekly differences are added (the positive or
                   negative sign of the difference should be retained for this
                   addition) and the sum is divided by four. The result of this
                   calculation should then be compared to the monthly standard
                   provided in Table 4.

                   If the difference in the weekly beginning and ending volume
                   measurements, or the monthly average difference, is equal to or
                   greater than those shown for the appropriate size of tank in
                   Table 4, the UST may be leaking or may have holes that are
                   allowing water to enter.

ENSURING EFFECTIVE MANUAL TANK GAUGING


            Chapter 1 provides a general description of the types of oversight that
            can be used. The following sections discuss how these approaches may
            be applied specifically to manual tank gauging.
            Site Inspections
           Implementing agencies could perform site inspections to review the
           manual tank gauging records and to observe the staff performing the
           tank gauging. The records could be reviewed for correct length of test
           and proper recording and analysis of data.
                                                                               43

-------
             Data Review
            Implementing agencies could require that all manual tank gauging forms
            be submitted by the owner/operator to the agency to be reviewed for
            accuracy in recording and conversion of gauge measurements to
            volumes and to ascertain correct interpretation of results. Such an
            approach, however, would be very labor intensive. Computer programs
            could be developed to recalculate the submitted data.  The records could
            be reviewed for only the one or two most important factors, such as the
            time between initial and ending measurements, to be sure the test was at
            least 36 hours long. Another approach would be to require submission
            of only some of the forms, such as twice each year.
             Guidance and Training
            Most errors in manual tank gauging take place because the tank is
            gauged improperly.  The implementing agencies could hold seminars or
            training classes in proper gauging procedures for persons electing to use
            this release detection method. An alternative to providing training
            classes  would be to provide guidance materials (e.g., a manual or a
            video tape) explaining the proper process.
             Approval and Certification
            An implementing agency could provide testing and certification for all
            persons using manual tank gauging. Such an approach would be time
            consuming.
44

-------
REFERENCES
            1.   American Petroleum Institute. 1987. Recommended
                Practice 1621, Bulk Liquid Stock Control At Retail Outlets.

            2.   American Petroleum Institute. February 2,1987. Analysis of
                Static Tank Testing as a Leak Detection Technique for Used
                Oil Tanks at Retail Outlets.

            3.   Mobil. 1984. Motor Fuel Inventory Verification Procedures.

            4.   U.S. EPA. January 1986. Underground Tank Leak Detection
                Methods: A State-of-the-Art Review. Report by Shahzad Niaki
                and John A. Broscious for Hazardous Waste Engineering
                Research Laboratory, Office of Research and Development,
                U.S. EPA.

            5.   U.S. EPA. April 1,1988. Review of Effectiveness of Static Tank
                Testing. Report by Midwest Research Institute for Office of
                Underground Storage Tanks, U.S. EPA.
                                                                               45

-------

-------
      Chapter IV
Tank Tightness Testing

-------

-------
TANK TIGHTNESS TESTING
IV
SUMMARY
             EPA studies show that tank tightness testing can be done reliably and
             affordably.  For many existing tanks it is the best available release detection
             option because permanent installation of equipment is not necessary, capital
             requirements are limited, and many commercial methods are available.

             Release detection methods are a combination of equipment and procedure.
             EPA studies show that, for tank tightness testing, the equipment is generally
             reliable and meets the manufacturers' specifications.  However, tank tightness
             testing procedures can be problematic. Getting effective tank tightness tests is
             a matter of focusing on the manufacturer's protocol and the testers adherence
             to the protocol. Consequently, this chapter explains key procedures, why they
             are important, and how some existing state UST programs have been ensuring
             that proper testing procedures are followed in their jurisdiction.

             The discussion presented in this chapter covers a wide range of possible
             problems that may occur with tank tightness testing. This does not mean that
             all, or even most, of these problems will occur at the same time.  Nor does it
             mean that all of the problems are of equal importance, in tenns of frequency of
             occurrence or severity of impact to the effectiveness of tank tightness testing.
             Some problems, such as poor access, seldom occur, while other problems,
             such as interference from evaporation/condensation have limited impact.
             Experienced testers are well aware of these problems and how to deal with
             them.  For example, an experienced  tester can recognize the presence of vapor
             pockets. Release detection, however, is a growing industry, and new
             companies are being formed with less experience. This chapter presents the
             full range of potential problems for educational purposes, not to imply  that
             they will always occur.

             Many of the descriptions and principles provided below for tank tightness
             testing are also applicable to automatic tank gauging and to piping tightness
             tests.  Those release detection methods are discussed separately in Chapters 5
             and 9, respectively.
                                                                                  47

-------
BRIEF DESCRIPTION
             There are many commercial tank tightness test methods available. While there
             is great diversity in the equipment and analysis schemes used, all tightness test
             methods are based on the same general approach. This section focuses on the
             general principles behind tightness testing and on procedures, not on specifics
             or equipment. Details on many methods and their performance are available
             in other publications (Ref. 1 and 6), and manufacturers and vendors of
             equipment will provide literature on their equipment.

             There are two main types of tank tightness tests: non-volumetric and
             volumetric.  There are only a few non-volumetric methods available, and they
             are not covered in this document. Volumetric tightness test methods measure
             the change in product volume or level over time to determine if there is a leak.
             When a leak occurs in an UST, the loss of product causes a decrease in the
             volume of product in the tank and, thus, a decrease in the level of product.
             Other factors, such as changing product temperature, can also cause product
             volume or level changes. Tightness test methods differ in how they measure
             the volume or level change and how they account for the interferences.

             The general procedure for conducting a volumetric tank test is quite similar
             from one test method to another (see Figure 8). The three procedural aspects
             common to all volumetric test methods are preparation, testing and analysis.

             Fkst, the physical layout and condition of the tank and piping must be
             evaluated to determine if tightness testing can be performed on that UST
             system.  There are situations in which the test equipment cannot physically be
             placed into the tank.  Other configurations may cause problems in obtaining
             meaningful results. Details of physical concerns are discussed in the problems
             section below.

             To prepare for a test, the tank must first be filled to a gross approximation of
             the level requked for testing. A waiting period must be observed to ensure
             that thermal effects and structural deformation resulting from filling the tank
             have stabilized.  The instrumentation to measure product level and
             temperature can be installed during or after the waiting period. Next, fine
             adjustments may be made to the fluid level by adding or removing small
             amounts of product.  If fine adjustments of product level are made, a second
             waiting period should follow.
48

-------
Preparation
 Testing
 Ana
ysis
                              Fill Tank
                                 i
                                 Wait for tank to stabilize
                        Install Test Equipment
                                 Determine height-to-volume
                                 conversion factor
                                 Determine coefficient of
                                 thermal expansion
                                 Measure ground-water level
                              Top Tank
                                      Walt
                           Conduct Test
                                 1
           Leak    No Leak
                                • Temperature measurement
                                • Level or volume measurement
Make calculations
Plot graph
Apply detection criterion
  Figure 8.  General procedure for conducting a volumetric tank test
  Source: U.S. EPA 1989
                                                                49

-------
             Volumetric tank tests can be divided into two categories: overfilled and
             partially filled. Figure 9 presents the two types of tests. In an overfilled-tank
             test, the tank is filled until the level of the product reaches the fill tube or a
             standpipe located above grade. Level changes occur in a small surface area,
             so that small changes in volume cause relatively large changes in level.  For
             example, in a 4-inch fill pipe, a volume change of 0.05 gallons will cause a
             level change of about 1 inch. In a partially filled tank, the test is conducted
             with the product level somewhere below the top of the tank.  Level changes in
             these tests occur in large surface areas, where small changes in volume cause
             very small changes in level. For example, in a half-filled 10,000-gallon tank
             8 feet in diameter, a volume change of 0.05 gallons will cause a level change
             of about 0.00006 inches.  Level sensing devices must be considerably more
             sensitive for partially-filled-tank tests than for overfilled-tank tests in order to
             achieve the same accuracy.

             During the waiting periods(s), the test operator must determine values for the
             height (level)-to-volume conversion factor and the coefficient of thermal
             expansion (see section below on problems  during preparation for discussion of
             these terms). Finally, the tester should determine the height of the water table.
             The preparations are now complete, and testing can begin.

             During the test, sensors take measurements of both the temperature and the
             level of the fluid in the tank. The respective measurements are taken
             repeatedly at specified intervals and are recorded for analysis at a later time.
             Data collection can be manual or automated. The end of a test is based on a
             criterion pre-determined by the manufacturer.  Usually this criterion is
             expressed in terms of time; for example, the test ends 60 minutes after the start
             of the data collection.

             By the time the test is complete, a considerable amount of data may have been
             gathered. Generally, the more data that are gathered, the better the test.
             Procedures for averaging the data, compensating for temperature, and
             computing a volume leak rate are usually well defined by the manufacturer of
             the test method. The end result of the analysis is a calculated volumetric
             "flow rate" that indicates how fast fluid is escaping from the tank.

             The detection criterion (usually a single threshold value) is applied after the
             analysis has been completed and is used to determine whether the level
             changes are due to a leak or to normally occurring volume fluctuations.  If the
             temperature-compensated volume change exceeds the detection criterion, a
             leak is suspected;  if not, it is assumed that the tank is not leaking. The most
             common criterion is 0.05 gal/h.
50

-------
          Partially Filled Tank
            Overfilled Tank
                                                                                        A
  (A) Adding a quart of liquid to this tank
      would produce a barely noticeable rise In the
      level of fluid. Level changes are distributed
      over a large surface area, so that even large volume
      changes produce only very small level changes.
Adding a quart of liquid to this tank would
cause the fluid to rise many Inches. Here the
surface area is very small. Thus, even a small
volume change can mean a drastic level change.
Figure 9.  Comparison of partially filled and overfilled tanks.
Source:  U.S. EPA (1989)
                                                                                          51

-------
POTENTIAL PROBLEMS AND SOLUTIONS FOR TIGHTNESS TESTS
             From the practice of tightness testing and studies performed by EPA,
             important procedural aspects have been identified. This section discusses the
             most important aspects of tightness testing and the errors and problems
             encountered.  Table 5 presents a summary of the problems encountered during
             tank tightness testing, the indicators and solutions for these problems, and how
             implementing agencies might provide oversight to prevent or correct the
             problems. A number of agency solutions are offered for each problem, but not
             all of them need be undertaken. In the table, the problems that are most
             serious, either because of frequency or severity of impact, have been marked
             by an asterisk. The problems are discussed below in the order of the testing
             procedure presented in Figure 8 (page 49).  There is no ranking implied by  the
             order of discussion.
              Site Considerations
    Testing
   Analysis
Many "real world" factors concerning the installation and condition of the
UST system can influence how well a tightness test performs and how much
effort is involved in obtaining meaningful test results. Most physical factors
can be accounted for with proper equipment and experienced testers.
Tightness testing will be most effective and efficient if as much as possible is
known about the setup and condition of the UST system prior to selecting the
method or beginning the test. The correct equipment can be assembled
beforehand, and any necessary modifications to the equipment, protocol, or
site can be made.

Joints, bungs, and manways must be tight for overfill methods

      Common problems with UST systems are (1) fittings and joints that
      were not tightened properly during installation; (2) gaskets at joints that
      were not installed, were installed incorrectly, or have deteriorated over
      time; and (3) joints that have become loose over time, such as piping
      connections dislodged during frost heaves. Often gaskets are not
      installed on manways. Also, temporary covers on openings used during
      delivery are frequently not replaced with permanent bungs during
      installation. During the overfill type of tightness test, the top of the tank
      and the piping are filled with liquid, and product will leak from any
      loose fittings or joints. Identification of these problems is desirable;
      however, leaks due to loose fittings obscure identification of possible
      leaks due to corrosion holes. Therefore, leaks due to loose fittings must
      be stopped before testing of the tank for corrosion holes can begin.
52

-------
     Table 5.  Indicators and Solutions for Problems Encountered  During Tank Tightness Testing
Ui
      Problem
Indicators
Tester Solutions
Agency Oversight Options
     Joints, bungs, and manways
     must be tight for overfill
     methods.
     Assure that manifolded tanks
     are tested appropriately.
     Abandoned piping may require
     special treatment (for overfill
     methods).
      Poor access may require use
      of another method.
     Assure that drop tubes do not
     interfere with test method.
     *Assure waiting time between
     delivery and testing is
     adequate.
Top of tank or fittings are wet with
product.
Tank drawings. Owner/operator
knowledge.
Tank drawings. Owner/operator
knowledge. Unexplained piping
from top of tank. Very large
losses when tank is overfilled.

Drawing of site showing tank near
or under large structure.
Object such as inventory stick
placed through fill pipe hits
sides of tube.
Erratic and large volume changes;
short-term volume decrease that
levels out.
 Check fittings and gaskets before
 testing. If results of test indicate
 possible leak, uncover tank,
 check fittings, and retest.

 Disconnect tanks and test
 separately.
 Find out purpose of all piping.
 Install shutoff valves on piping
 with unknown purpose. Dig up
 runs of piping.

 Inspect  site before testing to make
 necessary equipment modifica-
 tions. Use other opening in tank.
 If unable to make necessary
 physical modification, decline to
 test.

 Replace with temporary drop tube.
 If permanent drop tube cannot be
 removed easily or cost-effectively,
 do not test.

 Wait at least 6 hours after large
 product additions to tank and at
 least 3 hours after topping off.
Observe test. Inspect site after
test for new gaskets, etc., if
claimed by tester.
Review data sheets. Check
site after test for evidence of
disconnect.

Review data sheets for test
results. Check site after
test for evidence of digging
or shutoff valves.

Review site plans. Observe test.
Site inspection.
Review data sheets for volume
changes and testing times.  Plot
test data to observe trends.
                                                                                                                     Continued

-------
(J\
      Problem
Indicators
Tester Solutions
Agency Oversight Options
      'Need to Identify and remove
      vapor pockets.
Erratic or sudden large volume
changes; bubbles in fill pipe or
stand pipe.
 Add known volume to tank, and
 calculate expected and observed
 level changes.  If different,
 bleed off vapor and start test over.
Review data sheets for volume
changes.
      *Accounting for presence of
      ground water.
Water in the tank.
 Check existing observation well
 or drill hole to level of bottom
 of tank.  If present, raise liquid
 level in tank to counteract
 pressure of ground water, or do
 not test.
Review data sheets for ground-
water data. Inspect site to be sure
an observation well was available
or drilled.
      Assure accuracy of height-to-
      volume conversion factor for
      each tank.
 None.
 Add or withdraw known volume
 from tank, and observe level
 changes.
Review data sheets for calculation
of conversion factor.
      Use correct coefficient of
      expansion of product.
 Used average coefficient for product
 type from published source.
 Determine for each tank using
 API hydrometer and tables.
 Review data sheets for calculation
 of coefficient.  Compare coefficient
 to published average values.
      Assure collection of sufficient
      test data.
 Fewer than 25 data points collected.
 Sample at least every 5 minutes
 for at least 2 hours.
 Review data sheets for duration
 of test and sampling frequency.
      *Provide adequate tempera-
      ture compensation during
 None.
 Use at least three temperature
  equally spaced vertically, or
 mix the product and use one
 sensor.  Or use method that
 measures a factor independent
 of temperature.
 Observe test equipment as it is
 installed in the tank.  Review
 product literature.

-------
 *Need to maintain constant
 product level during test
 (overfill methods).
 Need to compensate for
 condensation and evaporation
 in very hot conditions.

 *The number of tests per tank
 should be fixed.
 Decisions left to tester should
 be minimized.
Need to follow protocol and
use correct threshold value.
 Constantly changing product level
 in fill pipe or stand pipe.
 Vapors above stand pipe or fill pipe
 or condensation on sides of pipe.
 High percentage of tanks declared
 "tight."
None.
High percentage of tanks declared
light" or leaking.
Assure calculations are
performed correctly.
A few data points are extremely
different from the rest of a test.
 At frequent intervals, add or
 remove product to keep level
 constant.
 Minor problem. No need to
 compensate. Shade from direct
 sunlight.

 Protocol must include fixed
 number of tests.
Protocol should be as explicit as
possible and cover as many
potential situations as possible.
Analysis scheme must be well
defined. Criteria for declaring
"tight" or leaking" must be clear.
Threshold value for declaring
leak should be smaller than
minimum detectable leak rate by
a factor of at least 2.

A second person double-checks
manual calculations. Computer
program is reviewed and tested
using known and verified data.
 Approve only methods whose
 protocol includes maintaining
 constant level.  Observe tests.
 Provide guidance and education
 on the importance of constant
 product level.  Review data
 sheets for evidence of additions
 and withdrawals.

 None.
 Review data sheets for results
 of all tests.  Compare tests
 conducted to protocol.  Keep
 track of pass/fail ratios for each
 testing company.

 Compare activities on data sheets
 with published protocol.  Observe
 during test.  Approve for use only
 those methods with adequately
 defined protocols.

 Review data sheets and
 calculations to see if steps agree
 with protocol.  Keep track of
 pass/fail ratio of each testing
 company.
Manual or computer recalculation,
from raw data to tightness
declaration.
* Indicates the most significant problems.

-------
   Testing
   Analysis
   Testing
   Analysis
      Testing the tank separately from the piping will help to distinguish some
      of these causes of release. Any fittings on top of the tank that are
      visible without removing backfill should be checked before beginning
      the test for tightness and adequate seals/gaskets (e.g., on the manway)
      should be installed; any visibly deteriorated gaskets should be replaced.
      Checking all visible fittings on top of the tank during the test to see if
      any product is present will help to discriminate sources of release. If the
      visible fittings are not wet with released product but the test indicates a
      leak, it may be necessary to expose the rest of the fittings on top of the
      tank by removing the backfill. If an UST fails a tightness test and the
      fittings are subsequently found to be loose and are tightened, the UST
      may be tested again; if the UST is found to be tight, no report of a
      suspected release is  necessary.

Manifolded tanks must be tested appropriately

      Sometimes tanks at  a site that hold the same product are connected
      together by piping.  The siphon effect allows product to be drawn from
      all of the tanks that are manifolded together. For overfill types of
      tightness testing, it is possible to test manifolded tanks at the same time.
      However,  the uncertainty associated with the necessary temperature
      measurements is high because it is unclear how to apply the temperature
      measurements from two or more different tanks to one common volume
      measurement.  For test methods using partially filled tanks, there may
      be some slight "wave action" in the tanks when they are connected due
      to the siphoning effect of the connecting pipe. This variation in the level
      of the liquid in the tank interferes with accurate level determinations.
      For these reasons, manifolded tanks generally should be disconnected
      from each other and tested separately.  If the piping between the tanks
      can be disconnected, either type of tightness testing can be used. This
      approach can be very convenient for the overfill methods because the
      extra product needed to overfill each tank is available from the tank
      manifolded to it.

Abandoned piping may require special treatment

      At older sites or sites where the use of the UST system has changed
      frequently, there may be piping connected to the tank that is no longer
      used. Rather than digging up the entire run of old piping, only the end
      connected to the old pump or delivery source may have been removed.
      This abandoned piping may be left open-ended. When the overfill type
      of tightness test is used, this extra piping can cause several problems.
      First, vapors can become trapped in the piping, and locating vapor
      pockets in piping and removing them is very difficult.   Second, if some
56

-------
 Testing
Analysis
 Testing
Analysis
      of the piping is open to the ground, large volumes of product will leak
      to the environment when the tester tries to fill the system. These
      problems can be overcome by using a nonoverfill tightness test method
      or by isolating the tank from the piping. Isolating the tank from old
      piping may involve digging up part of the piping to install some type of
      device that closes off the piping. Because an open-ended pipe is always
      a potential source of large product losses, abandoned piping should be
      removed or closed any time it is discovered, regardless of the type of
      release detection being used.

Poor access may require use of another method

      Some UST systems are located in sites where it is difficult to set up the
      test equipment, such as near buildings or under other structures. The
      space available to set up the test equipment and maneuver during the
      test may be too limited for some tightness test methods. Another access
      problem is the location of the fill pipe relative to the tank itself. Remote
      fill pipes are sometimes used, where the actual opening is a long way
      from the tank and connected to the tank by a run of horizontal piping.
      Most tightness test methods rely on using the fill pipe to insert all of the
      necessary equipment and sensors into the tank and, therefore, such
      methods may be infeasible for a tank with a remote fill pipe. It may be
      possible for such a test method to use other openings on the top of the
      tank, such as the manway. Otherwise, another release detection method
      should be used.

Drop tubes must not interfere with test method

      To avoid agitation, wave action, and splashing during filling, many
      tanks are equipped with a drop tube. This is a tube the diameter of the
      fill pipe that extends from the opening to near the bottom of the tank.
      The presence of a drop tube can interfere with proper temperature
      measurement because the product in the tube is isolated from that in the
      rest of the tank. The tube also can be an obstruction to placing
      equipment and sensors in the tank.  Drop tubes interfere with product
      circulation for those tightness test methods that try to achieve even
      temperature distribution using a circulating pump. Some drop tubes are
     removable and, obviously, should be removed before a test is begun. A
     permanent drop tube can be replaced with a temporary tube. Such a
     replacement is difficult and expensive and may only be worth the effort
     if it is expected that tightness testing will be performed routinely on the
     tank for a long time.  Otherwise, the tightness testing company should
     be questioned about the ability of the method to perform well with a
     permanent drop tube.
                                                                                  57

-------
              Testing Considerations
| Preparation]   Adequate waiting time between product delivery and testing
   X
  Testing
  Analysis
The most common problem with tightness testing is not waiting long
enough between adding product to the tank and beginning collection of
temperature and level data. The fluctuating temperature and structural
changes of the tank following the addition of product cause volume and
level changes that are unrelated to changes caused by a leak. Not
waiting for these changes to stop results in erroneous test results.

There are two times when product may be added to the tank: (1) gross
product delivery, where thousands of gallons may be added to bring the
product to approximate testing level; and (2) "topping off', where
several gallons may be added to achieve the final level for testing.  For
the reasons discussed below, each addition of product and increase in
product level causes changes within the tank that interfere with accurate
test results.

When product is added to a tank, its temperature is at or near ambient
air temperature. The temperature of the product already in the tank is at
ground temperature. The difference between ambient air temperature
and ground temperature varies with the season and location, but
differences of 10 to 20 degrees Fahrenheit are not uncommon. For
some time after delivery, the temperature of the product will fluctuate
rapidly and widely as the product mixes and eventually achieves an
equilibrium near ground temperature. As the temperature of the liquid
in the tank increases or decreases, the volume of the liquid will increase
or decrease, respectively (see Figure 10). For example, 1000 gallons of
gasoline will shrink by 0.7 gallons when the temperature drops by one
degree Fahrenheit. Increasing volume due to temperature increase may
mask a leak while decreasing volume due to temperature decrease may
falsely indicate a leak.

All tightness test methods must account for temperature changes.
However, for a period of time following delivery, the thermal chaos in
the stored product is too extreme to be adequately measured and
accounted for.  In addition, the temperature changes immediately
following delivery are not the same at the ends of the tank as at the fill
pipe. Consequently, no matter how many temperature sensors are added
at the fill hole, the measured temperature at that point does not reflect
the average temperature in the tank.
58

-------
               CAP
                             CONCRETE
                                                                   CAP

f LEVEL



^^- FILL TUB
                            i—CONCRETE
                                                   PRODUCT LEVEL
                                                                                   •FILL TUBE
            NEW, COOLER PRODUCT
 (A)  Product has Just btien added to an underground
     tank that was already partly filled. The new
     product Is cooler than the resident product, and
     temperatures fluctuate greatly.
'  '
    As the old product cools and the new warms,
    equilibrium Is reached. But tha temperature as a
    whole Is cooler, causing the product to contract
    and the level to go down.  (The Inverse Is true
    when warmer product Is added, causing the product
    io expand and the level to rise.)
Figure  10. How temperature changes can be mistaken for a leak.
Source: U.S. EPA (1989)
                                                                                             59

-------
                In addition to fluctuating temperature, the addition of product causes
                structural deformation of the tank. Whether it is constructed of steel or
                fiberglass and whether it is embedded in a dense backfill or in a loose
                one that has more "give,"  the tank itself expands and contracts in
                response to both temperature and level changes. When the tank
                expands, the level of the fluid inside it goes down; conversely, when it
                contracts, the level goes up (See Figure 11). The amount of volume
                change due to tank deformation varies with the material of construction
                and the type of backfill, but effects can easily be in the range of 1  to 10
                gallons. Distinguishing between real volume changes and the apparent
                changes brought on by structural deformation is generally not possible,
                regardless of how accurate the equipment is.

                The solution to the problems of temperature fluctuation and tank
                deformation following addition of product is to wait until these changes
                have stopped, i.e., until the product temperature has stabilized and the
                tank has completely expanded.  The exact waiting time that is necessary
                will vary with the amount of product delivered and the temperature
                difference between added and original product. As a guideline, a
                minimum waiting time of about 6 hours should elapse after delivery of
                product in the range of hundreds or thousands of gallons and about 3
                hours after topping off of the tank. These minimum times should be
                sufficient for both thermal fluctuations and tank deformation to
                stabilize. A few tightness test methods avoid the problem of
                temperature effects by measuring some aspect that is independent of
                temperature, such as the mass of the product.

                To determine that sufficient time has elapsed for the tank to stabilize,
                the tester should watch the temperature and level changes.  Preferably,
                temperature and volume measurements versus time should be plotted on
                a graph as the measurements occur. If the readings show large and
                erratic changes, conditions in the tank are probably still fluctuating.  If
                the temperature-compensated level changes decrease and eventually
                level off, it is an indication that the tank ends were continuing to relax
                early in the test but finally stabilized.  Some tightness test methods
                include statistical analyses of the data as they are collected and, from the
                randomness of the data, can determine that tank conditions  are not
                stable enough to begin the test. Regardless of which method is used to
                determine tank conditions, any initial data indicating instability should
                not be used in the final evaluation, and the length of the test should be
                extended so that the minimum acceptable test duration occurs after the
                tank has stabilized. That is, if the test protocol says that
60

-------
               CAP	W    pCONC
RETE
    PRODUCT.
    LEVEL
                                     FILL
                                     TUBE
                  PRODUCT
   (A)  An empty underground tank has Just been filled
       with product.
                       CAP
CONCRETE-
                                                PRODUCT
                                                LEVEL
                                            FILL
                                            TUBE
                           PRODUCT
                  In response to the pressure and/or temperature
                  of the product, the ends of the tank begin to
                  deflect ("structural deformation"), and the
                  level of the product goes down.
Figure 11.  How structural deformation of the tank
can be mistaken for a leak.  Source: U.S. EPA (1989)
                                                                                  61

-------
               the test should last for 2 hours, the 2 hours of test data used in the
               analysis should be obtained after the tank has stabilized.
I TeaHng  |
~~
|  Analysis  |
Vapor pockets in overfilled-tank tests identified and removed

      In overfilled-tank tests, vapor pockets may form. Although most test
      operators claim to be able to identify vapor pockets easily, many tests
      continue to be invalid or yield incorrect results because the problem was
      not recognized or corrected.

      Vapor pockets are almost always present after a tank has been filled to
      or above the top of the tank because vapor becomes trapped in the
      manways, deadend piping, etc. (see Figure 12). Temperature
      fluctuations change the volume of the vapor pocket, and the expansion
      and contraction of the vapor pockets changes the liquid level in the tank.
      For example, a temperature change of about 0.25 degrees Fahrenheit
      may change the volume of a 100-gallon vapor pocket by about 0.05
      gallons. To a lesser extent, changes in barometric pressure also cause
      vapor pockets to expand and contract, causing changes in liquid level.
      An increase in liquid level due to expansion of a vapor pocket may
      mask a decrease  in liquid level due to a real leak, while a decrease in
      level due to contraction of a vapor pocket would falsely indicate a leak
      or exaggerate the rate of an actual leak. Vapor pockets in quantities as
      small as 10 gallons can influence a test result.

      The first step in solving this problem is to identify the presence of a
      vapor pocket. While it is virtually impossible to determine the exact
      size of the vapor pockets, there are several methods that can be used to
      check for their presence. The most easily identifiable indication of a
      vapor pocket is the presence of bubbling in the fill pipe or stand pipe
      during the test. Vapor pockets may also be indicated by a sudden large
      drop in product level, indicating a vapor pocket that just "released". If
      the temperature-compensated volume changes fluctuate over time with
      no obvious trend, then there may be a vapor pocket that is expanding
      and contracting,  thus confounding the results. This indicator, however,
      is not conclusive unless sufficient waiting time has elapsed since
      addition of product for temperature and structural deformation changes
      to subside. Another method of identifying vapor pockets is to add a
      known volume (of product or a solid object) to the tank and compare the
      actual increase in product level to the increase that would be expected
      from the geometry of the tank.  If the actual level change is less than the
      expected change, a vapor pocket may be present that compressed from
      the pressure of the added product.
62

-------
                     Dispenser
        Trapped
         Vapor
                                                                                         Vent Pipe
o\
Figure 12. Location of vapor pockets in an overfilled tilted tank.
Source: Schwendeman and Wilcox (1987)

-------
                   If a vapor pocket is shown to be present, it must be removed. One
                   method of removal is to uncover the tank and install a bleed valve
                   on the high end of the tank.  As the vapor is bled off, product will
                   fill the void. A relatively new method of removing vapor pockets
                   involves inserting a hose and bladder into the tank, inflating the
                   bladder so that it rises to the high end of the tank (where the vapor
                   pocket is), and suctioning out the vapors. After a vapor pocket is
                   removed, the test should be started over. Checking for signs of a
                   vapor pocket should continue because not all of the vapor may have
                   been removed or another pocket may have formed.
             Accounting for the presence of ground water
  T«atlng

  Analyala  |
The presence of ground water around the tank may completely mask an
actual leak or at least slow the rate at which product is leaking.  Failure
to check for the presence of ground water and to take action when it is
present make the results of a tightness test questionable.  In the National
Survey conducted by EPA at about 500 randomly selected sites around
the U.S., ground water was above the bottom of the tank at about 25
percent of the sites.

The water table of the soil in which a tank is buried can vary in height
depending on factors such as geographic location, season, and amount
of precipitation. As the illustrations in Figure 13 show, the height of the
water table in relation to the tank can have a direct effect on the leak
rate measured during a test. If the water table is above the location of a
hole or fissure in an underground tank, the ground water exerts a
pressure on that hole which counteracts the pressure exerted on the
same hole by the fluid in the tank.  The best test results are obtained
when the water table is below the level of the tank. Flow of the leak
through the hole is then unrestricted, and measurement of the flow rate
will not be influenced by ground water.

Because it is virtually impossible to determine the location of a hole in
an underground tank, efforts must  be concentrated instead on
monitoring the ground-water level. The tester should determine the
ground-water level or at least determine if it is below the bottom of the
tank. Hydrogeological information may be available from agencies
such as the U.S. Geological Survey or from boring logs from nearby
sites, but such data do not necessarily indicate conditions at the tank
being tested.  There can be very localized hydrogeological formations
that result in a "pocket" of water in a small area of a region otherwise
characterized by deep ground water.  The only sure way to determine if
64

-------
        NO FLOW. Pressure exerted by
        the product Is exactly balanced
        by pressureof ground water
        at hole. Product Is less
        dense than water, so there
        Is no flow in either direction
        even though product level
        Is higher than water table. The
        dotted line shows the product height
        required to produce an equal balance
        of pressure given the height of the
        water table.
FLOW INTO TANK. Pressure exerted
by ground water Is greater than
that of product, so water flows Into
tank. Product level would have to be at
the "equal pressure" line In order to
achieve an exact balance with the
ground water, given the height
of the water table.
       FLOW OUT OF TANK. Water table
       Is below the tank, so there is no
       counter-pressure against the product
       at the hole. Therefore, product flows out.
  FLOW OUT OF TANK. Here, the pressure exerted
  by the groundwater is less than that
  of the product; therefore, the product
  flows out.
Figure 13. Effect of ground water on the rate of flow through
a hole in an underground tank.  Source:  U.S. EPA (1989)
                                                                                             65

-------
the tank is surrounded by water is to check observation wells near
the excavation zone. If there are no existing wells, one can be made
fairly easily with a hand auger. A well drilled just for checking
water level during a tightness test does not have to be a formally
installed observation well. The important point is to determine
whether the tank is surrounded by water, so the well does not need
to be completed to the water table, only to the bottom of the tank.

A tightness test method should include a formal procedure for
dealing with high ground-water levels.  For example, a test can be
postponed until the water table drops below the level of the tank.
This is often impractical, however.  For overfilled-tank tests, another
approach is to raise the level of the product above grade until the
pressure on the bottom of the tank reaches  a given level, such as 4 or
5 psi (the pressure must be within the safety margins of the design
of the tank). Under this approach, the pressure of the liquid within
the tank will be greater than that of the ground water surrounding
the tank.

Some tightness test methods conduct two consecutive tests, each at a
different fluid level within the tank, and compare the leak rates. If
there is a leak in the tank, the leak rates will be different because the
head pressures are different. Theoretically, no measurement of
ground water is necessary because the test is independent of ground
water, that is, the difference in leak rates will show up whether
ground water is present or not. However, the difference in head
pressure between two different test levels causes only a very small
change in leak rate, and the differences in leak rate from changes in
head pressure probably will be obscured by variations introduced by
other factors such as temperature.

Effective tightness test methods should include a procedure for
determining the presence of ground water and compensating for it.
To ensure that testers follow these procedures, the state of Rhode
Island checks all test reports for the information that the ground-
water level was checked and how it was compensated for. Some test
sites are visited after the test to check that a well is, in fact, present.

-------
 Testing
 Analysis
Accuracy of height-to-volume conversion factor for each tank

      Most tightness test methods measure changes in product level over time.
      To calculate a leak rate in gallons per hour, these level measurements
      must be converted to changes in volume. The value used to make this
      conversion is called the height-to-volume conversion factor and will be
      different for each tank. If the wrong conversion factor is used, the
      volume change that is calculated will be wrong, resulting in an
      erroneous leak rate or even a false decision regarding the integrity of the
      tank.

      The height-to-volume conversion factor should be determined
      specifically for each site because the geometry of each UST system is
      slightly different. The conversion factor can be derived mathematically
      based on the geometry of the tank. However,  there can be differences
      between the manufacturer's specifications for the general type of tank
      and the actual tank that was installed, and there may be unknown factors
      that change the internal volume of the tank, such as tank end deflection
      or old piping that is still attached. For these reasons, theoretical
      calculations of the conversion factor are considerably less accurate than
      direct measurement. The conversion factor can be determined by
      adding and withdrawing known volumes to the tank and measuring the
      actual change in height of the product. The known volume added can
      be either product or some solid object such as a metal bar. For example,
      if 5 gallons is added to the tank and the height of the liquid in the fill
      pipe increased by 15 inches, a height of 3 inches would equal a volume
      of 1 gallon, making the height-to-volume conversion factor equal to 3
      inches per gallon.
          Correct coefficient of thermal expansion of product
Testing
Analysis
      The coefficient of expansion of a liquid relates changes in its volume to
      changes in its temperature. The coefficient of expansion is different for
      each product. The units of the coefficient are change in gallons per
      degree Fahrenheit.  If the wrong coefficient of expansion is used, the
      calculated leak rate will be incorrect.

      Some tightness test methods use average coefficients of expansion for
      general classes of product, such as gasoline or kerosene. However, the
      coefficient of expansion varies within each type of product, and the
      uncertainty inherent in assuming an average coefficient based on
      product type is at least 10 percent. To more accurately determine the
                                                                                    67

-------
                     coefficient, the tester should take a product sample, measure its
                     density using an API hydrometer, convert the density to standard
                     temperature and pressure, and then read the coefficient from a set of
                     API tables. The uncertainty inherent in a coefficient of expansion
                     determined by measurement of the product density is about
                     5 percent.
             Testing
| Preparation!

  Analysis
Collection of sufficient test data

         Another common problem with tightness testing is not collecting
         enough data to make an accurate and statistically significant
         determination of the status of the tank. Some tightness test
         protocols requke sampling only every 15,30, or 60 minutes.  As
         discussed above, some important interferences to accurate tightness
         testing such as tank end deflection and thermal fluctuation should be
         monitored using the level and temperature data. If these data are
         taken only infrequently, important trends in the data may be missed
         and, thus, problems may not be identified. If the test does not last
         long enough, very small leaks may not be identified.

         More data may be collected by either sampling more frequently or
         conducting longer tests. As a rule, obtaining more data increases the
         probability of correctly identifying the presence of a leak. Ideally,
         level data should be sampled at intervals of approximately one
         second; however, such a sampling frequency is not practical with
         most equipment. For manual tightness test methods, a sampling
         frequency on the order of 5 minutes is generally adequate.  Studies
         have shown that tests at least 2 hours in duration (after appropriate
         waiting periods have been observed) provide more accurate results.

Adequate temperature compensation during  test

         Measurement of the average temperature of the product throughout
         the tank is important because the total volume of the product will
         change in response to changes  in temperature. If the temperature
         changes in only one or two points  in the tank are measured,
         incorrect total volume changes for the tank will be calculated.

         The stored product will expand and contract in response to
         temperature changes.  Once the appropriate waiting time has passed
         following product delivery, the product in the tank will usually be
   68

-------
                      stratified into layers of different temperatures. Very small
                      temperature changes continue to occur in each of these layers, even
                      after the tank is essentially stabilized. The extent and rate of
                      temperature changes will vary layer by layer, and thus, the change in
                      product volume will be different for each layer.  The temperature in
                      one layer is not indicative of changes in the other layers.  Therefore,
                      it is necessary to measure the temperature in as many layers as
                      possible to obtain an accurate total volume measurement.

                      Because of the problem of temperature gradients within the tank, the
                      tightness test method must have temperature sensors that provide
                      adequate spatial coverage of the tanks, so that the data they record
                      are representative of product conditions throughout the tank. A
                      vertical array of at least five thermistors provides the best spatial
                      coverage.  An array of three sensors, arranged vertically, at the top,
                      middle, and bottom of the tank is considered adequate. Although
                      one sensor typically is not sufficient, methods that circulate the
                      product in the tank can obtain satisfactory results with a single
                      probe. By mixing the product, the problem of uneven temperature
                      distribution in the tank is eliminated.  Methods using temperature
                      probes that average the temperature throughout the tank into a single
                      value also eliminate this problem.
Analysis
[Preparation]  Constant product level maintained during overfill tests
     *                                            - -
                      Adding or removing product changes the hydrostatic pressure within
                      me tank' causin& me sides of me tank to expand or contract,
                      changing the apparent volume of product in the tank. Even relatively
                      small level changes that occur during data collection can cause some
                      tank deformation, leading to erroneous test  results.

                      Tank tightness test methods can be categorized into those that
                      maintain a steady product level and those in which the product level
                      is allowed to fluctuate. In constant-level tests, small  amounts of
                      product are added or removed periodically to maintain the product at
                      a specified level. The product can be added or removed either
                      manually or automatically.  In variable-level tests, no such
                      adjustments are made. When the fluid level is kept constant during
                      the test, the tank will neither expand nor contract during the test in
                      response to the level changes, thus removing an interfering factor,
                      and measured volume changes will accurately represent actual
                     volume changes. The amount of product added during
                     constant-level tests is too small to introduce any error in the test
                     results due to temperature-related effects.
                                                                                        69

-------
                      Variable product level is a problem primarily with overfill tightness
                      test methods.  For underfill test methods, the changes in product
                      level are too small to cause enough change in head pressure and,
                      thus, deformation of the tank ends.

                      If the product level fluctuates during a test, it is impossible to
                      compensate for the effects of tank deformation on volume. It is
                      advisable, therefore, to eliminate from consideration for use any test
                      conducted under variable hydrostatic pressure (i.e., variable product
                      level).
[Preparation!
    ..*....
\ •: Tasting ¥[
  Analysis
Compensating for evaporation and condensation

         Unless a tank and its fill tube are completely filled and no air or
         vapor pockets are present, it is likely that, as temperature changes,
         fluid will evaporate from the free product surface into the air space
         and/or will condense along the inner surface of the tank walls and
         drip back down.  This activity produces volume fluctuations as
         liquid product evaporates and condenses. Although a few test
         methods attempt to account for the effect of evaporation and
         condensation, these effects are not believed to be significant enough
         to warrant special control or measurement during a tightness test.  In
         an overfilled-tank test conducted under extremely hot conditions or
         with the standpipe in direct sunlight, it is possible that product may
         evaporate in the standpipe; however, such conditions are very rare.
         Shade could be provided for the standpipe, or testing delayed to a
         cooler day.  A record of ambient air temperature during the test may
         help to pinpoint possible reasons for unusual test results.
I Preparation]  Fixed number of tests in the testing protocol
  Testing
  Analysis
   70
         When the results of a tightness test indicate that a tank is leaking but
         the leak rate is only slightly above the threshold value for declaring
         a leak, some testers repeat tests on the tank until the results of one
         test indicate that the tank is tight. This approach is a misuse of the
         "more data are better" axiom and reduces the probability of
         detecting a leak. A multiple-testing strategy is a valid approach to
         tightness testing but it must be performed correctly and all of the
         data must be considered, not just the data from the one test that
         gives the desired answer. For example, the testing protocol could
         require that five tests be performed and the tank declared tight if the

-------
                      results of three of the tests are below the threshold. Under this
                      approach, the minimum detectable leak rate is less than that for a
                      single tightness test. A multiple-testing strategy must be a
                      well-defined part of the testing protocol, and no deviation should be
                      allowed from the total number of tests conducted.

             Minimizing decisions left up to the tester

                      In some test methods, the procedures to be followed by the tester are
                      ill-defined or are left to the tester to decide, such as the point at
                      which to end the test. Some circumstances may not be included in
                      the protocol at all, such as how to recognize vapor pockets. Much is
                      left to the discretion and experience of the tester.  However, not all
                      testers have sufficient experience to make informed decisions. In
                      addition, even relying on experienced testers may result in different
                      decisions being made by different people in the same situation.
                      These factors decrease the likelihood of a test method achieving
                      accurate and consistent results.

                      The most reliable test methods are those least subject to operator
                      influence. Any test method which requires or allows the operator to
                      make subjective decisions during the test should be avoided. For
                      example, the test operator should not be allowed to decide when to
                      make a product-level adjustment or to decide how much product
                      should be added or removed.  Adjustments of any kind should be
                      accomplished using a set, repeatable procedure, not an arbitrary one.
                      As discussed above, the number of tests and their length should be
                      specified by the test protocol and not left to the discretion of the
                      tester.
             Analysis
| Preparation] Protocol followed and correct threshold value used
     *
I—es*ng  I           Many test methods lack a defined data-analysis protocol and a clear
                      criterion for deciding if a tank is leaking. This deficiency allows
          *           testers to make subjective decisions and leads to unclear or even
                      false determinations of the status of the tank.

                      A reliable test method will have a well-defined procedure for
                      analyzing the data, either manually or by computer.  The necessary
                      steps in a data analysis are presented here. The first step is to
                      calculate the volumetric flow rate, which can be accomplished in
                                                                                        71

-------
                   different ways. The goal of any calculation is to compute the
                   average rate of change, or flow rate, that indicates how fast liquid is
                   escaping from the tank. First, the level changes measured during the
                   test are converted to volume changes using the height-to-volume
                   conversion factor. Using the coefficient of thermal expansion,
                   compensation is made for any temperature changes that were
                   recorded.  The temperature-induced volume changes are then
                   subtracted from the measured volume change, yielding the
                   temperature-compensated volume changes. These are used to
                   determine the volumetric flow rate. If performed manually, these
                   calculations must be explicitly defined for the tester.

                   The manufacturers of some test methods use all the data available
                   when they perform the analysis. Others employ averaging
                   techniques. One example of an averaging scheme is to subtract
                   end-of-test data from start-of-test data and divide by the time that
                   has elapsed between the two. Another example is to add all the
                   cumulative volume changes and then divide this sum by the duration
                   of the test. Finally, some manufacturers fit a line to the data; that is,
                   the data points are expressed as dots on a graph, and a line is drawn
                   as closely as possible through the points. Whatever analysis scheme
                   is used, it must be well defined, and the tester must adhere to that
                   protocol.

                   To determine if the tank is leaking or not, the volumetric flow rate
                   must be compared to a threshold value, which has been
                   predetermined as part of the test design. In order for a test method
                   to perform well against small leaks, the threshold value must be
                   smaller by a factor of 2 or more than the smallest leak to be
                   detected.  Let us  assume, for example, that a tank is leaking at the
                   rate of 2 gal/h. If the test method in question can discern leaks as
                   small as 1 gal/h,  it will almost certainly detect the 2-gal/h leak.
                   However, if the tank is leaking at a rate of 0.5 gal/h, less than what
                   is discernible by this test method, the leak will go undetected.

                   The most commonly used threshold is 0.05 gal/h.  This threshold is
                   often confused with the leak rate to be detected. If the threshold is
                   equal to the leak rate to be detected, the probability of detecting a
                   leak of that size is only 50 percent. The final federal regulation
                   requires a tightness test method to have a minimum detectable leak
                   rate of 0.1 gal/h.  For a test method to meet this requirement, its
                   threshold must be less than 0.1 gal h.  For most test methods, this
                   threshold will be around 0.05 gal/ h.
72

-------
[Preparation]  Calculations performed correctly
  Testing
                     For test methods in which the volumetric flow rate is determined
                     manually, one problem can be simple calculation errors, leading to
                     an incorrect conclusion.  A tester should always have someone else
                     double check the calculations. In some local programs such as
                     Nassau County, New York, copies of all tightness test records must
                     be submitted to the implementing agency. Personnel at the agency
                     double check the calculations, either by hand or using computer
                     programs developed for the tightness test methods allowed in the
                     county.
  ENSURING EFFECTIVE TESTING
              To ensure that testers follow the procedures necessary to prevent the
              problems described above from occurring, a number of state and local
              implementing agencies have developed programs to oversee the practice in
              their jurisdictions.  Chapter 1 provided a general description of the four
              types of oversight that can be used.  The following section summarizes
              briefly some of the specific actions taken by implementing agencies to
              ensure effective tank tightness testing.
               Site Inspections
              In Rhode Island, the ground water is frequently very high and, therefore,
              checking and compensating for its presence is very important. State agency
              personnel routinely follow up a number of tests by visiting the site and
              making sure that there is some way in which the ground-water level could
              have been checked. Either an observation well was already on-site for some
              other purpose or a hole was drilled by hand specifically for the tightness
              test.

              In Nassau County, New York, testers must call county officials when level
              and temperature data collection is about to begin so that the official may
              visit the site and observe the test.  Which tests to observe is left to the
              discretion of the regulator. In Austin, Texas, the owner/operator must
              notify the implementing agency before a test is performed and agency
              personnel observe all tests.
                                                                                     73

-------
             Data Review
             Several states and counties, such as Rhode Island, Nassau County, New
             York, and Madera County, California, require tightness testers to submit
             test reports to the implementing agencies. Regulators then review these
             reports for adherence to key procedural elements such as: ground-water
             level, actual calculation of coefficient of expansion, length of test, number
             of data points, and appropriate testing levels. Recalculation of the test data
             is also performed on all or part of the test reports. Personal computers have
             been used to cut down on the time necessary for a review of test data.
             Nassau County, New York, has written a computer spreadsheet to be used
             with the data from two common test methods that double checks the tester's
             calculations using the tester's raw data. Rhode Island enters  all test results
             into a computer program that performs statistical analysis on the pass/fail
             ratios of the test companies. Whenever a company is passing or failing a
             disproportionate number of tanks, the agency investigates.
             Guidance and Training
            Because proper procedure is so important to effective tank tightness testing
            and because it is the major source of error as currently practiced, training
            and guidance can be an important tool.  Guidance can be aimed at
            implementing agency personnel, so that they can provide effective
            inspection and review, at owner/operators, so they can select and oversee
            effective testers, and at testing personnel, to ensure that they perform the
            tests correctly.
             Approval and Certification
            Some implementing agencies have tried to prevent tightness testing
            mistakes by only allowing methods and personnel that they feel are
            acceptable.  Several different approaches are being used.  Maryland requires
            that manufacturers submit performance evaluation results to the state for
            review, physically demonstrate the method to state personnel, provide
            information on their personnel training/certification program, and then, if
            accepted, train state agency personnel.  In Nassau County, New York,
            regulators review the test procedure manual for specific directions
            addressing the key elements identified in NFPA 329. In Los Angeles
            County, California, all tightness testers must be tested and approved by an
            independent party.
74

-------
            In other states, one step in the method approval process requires that the
            manufacturer train agency personnel in the operation of the test method.
            Massachusetts requkes that the manufacturer provide a video to be used for
            training. Rhode Island and Maryland requke that testers train the
            implementing agency personnel.
REFERENCES
             1.   U.S. Envkonmental Protection Agency. November 1988.
                 Evaluation of Volumetric Leak Detection Methods for Underground
                 Fuel Storage Tanks, Vol.1. EPA/600/2-88/068a. Prepared for U.S.
                 EPA by Vista Research, Inc.

             2.   Schwendeman, T.G. and H.K. Wilcox. 1987.  Underground
                 Storage Systems - Leak Detection and Monitoring.  Lewis
                 Publishers, Chelsea, Michigan.

             3.   National Fke Protection Association. 1987. NFPA329-
                 Underground Leakage of Flammable and Combustible Liquids.
                 Quincy, Massachusetts

             4.   U.S. Envkonmental Protection Agency. 1986.  Development of a
                 Tank Test Method for a National Survey of Underground Storage
                 Tanks. EPA-560/5-86-014. Prepared for U.S.  EPA by Midwest
                 Research Institute and Vista Research, Inc.

             5.   U.S. Envkonmental Protection Agency. May 1986. Underground
                 Motor Fuel Storage Tanks: A National Survey, Vol. 1. Technical
                 Report. EPA-560/5-86-013.

             6.   U.S. Envkonmental Protection Agency.  1986. Under ground Tank
                 Leak Detection Methods: A State-of-the Art Review.
                 EPA/600/2-86/001. Prepared for U.S. EPA by IT Corporation.

             7.   U.S. Envkonmental Protection Agency. July 1988. Common Human
                 Errors in Release Detection Usage. Prepared for U.S. EPA by Camp
                 Dresser & McKee, Inc.
                                                                                 75

-------
             8.   40 CFR Part 280. Technical Standards and Corrective Action
                 Requirements for Owners and Operators of Underground Storage
                 Tanks (UST).  53 FR 37194-37212.

             9.   Maresca, J. W. and M. L. Seibel.  September 23,1988. Volumetric
                 Leak Detection Systems. Vista Research, Inc.

            10.   U.S. Environmental Protection Agency. April 1989.  Volumetric
                 TankTesting: An Overview. EPA/625/9-89/009.
76

-------
      Chapter V
Automatic Tank Gauging

-------

-------
AUTOMATIC TANK GAUGING SYSTEMS
SUMMARY
           Automatic tank gauging (ATG) systems are permanently installed in
           underground storage tanks and provide both leak testing and inventory
           information. When the two modes are used together, ATG is an
           effective release detection method. Two studies conducted on ATG
           systems indicated that the leak test mode is capable of detecting leaks as
           small as 0.2 gal/h with a probability of detection of 95 percent and a
           probability of false alarm of 5 percent. The inventory mode of ATG
           systems is more accurate than the comparable manual method.

           ATG systems are selected by owner/operators because such systems
           require minimal operator involvement, cause few service interruptions,
           and can provide frequent automated release detection. In addition, the
           inventory mode provides continuous product information helpful in
           business management. For UST operations that close at least once a
           month, which includes many sites, the leak test mode does not interrupt
           operation. The inventory mode provides nearly continuous monitoring
           for the large losses typically discovered by inventory methods and,
           depending on how the owner/operator elects to  operate the ATG system,
           the tightness test mode can be used every time operations cease (e.g.,
           nightly, at many service stations).

           The discussion presented in this chapter covers  a range of problems that
           may occur with ATG systems.  This does not mean that all, or even
           most, of these problems will occur at the same time or at the same site.
           Nor does it mean that all of the problems are of equal importance, in
           terms of frequency of occurrence or severity of impact on the
           effectiveness of ATG systems.  Some  problems, such as blended fuel
           dispensers, seldom occur. Experienced ATG system vendors and
           installers are well aware of these problems and how to deal with them.
           For example, a reputable installer knows the proper wiring installation
           materials and methods and will follow them.  Release detection is a
           growing industry, however, and new companies are being formed that
           bring less experience to the field. Presented in this chapter is a range of
           potential problems for educational purposes, not to imply that  the
           problems will always occur.
                                                                             77

-------
BRIEF DESCRIPTION
             General
            Figure 14 presents a general schematic of an ATG system. All ATG
            systems are based on the same general approach. This section focuses
            on the general principles behind ATGs and not on specific equipment.
            For details on the specific equipment, manufacturers and vendors will
            provide literature.

            ATG systems measure the change in product level within the tank over
            time to determine if there is a leak.  When a leak occurs in an UST, the
            loss of product causes a decrease in the level of product. Other factors
            also can cause product level changes. The most important of the
            interferences are product temperature and tank deformation caused by
            addition or withdrawal of product. ATG systems differ primarily in
            how product level and temperature are measured.

            ATG systems have two modes of operation: inventory control and leak
            testing.  When the system is on, it is in one of these modes and can be
            switched to the other. The same equipment is used for both operations.
            Installation of the equipment and the operation of the inventory and test
            modes are described in the following sections. Figure 15 is a flow chart
            of the general operation of an ATG system.
             Installation
            Installation of an ATG system involves equipping each tank with a
            probe to measure product level and temperature.  The probe is inserted
            into the tank through a separate fitting (not the fill pipe) on top of the
            tank.  For most ATG systems, the fitting must be 3 or 4 inches in
            diameter. Some older USTs do not have extra openings or the existing
            bungs are too small; in this case, an opening of appropriate size for the
            probe must be made by the installer.

            A remote monitor and microprocessor are installed in a nearby building
            to record the information read by the probe. The monitor usually has a
            keyboard for programming and a display for presenting the required
            data.  Underground wking is installed between each probe and the
            remote monitor.  In some ATG systems, wking is also installed between
            the dispensers and the monitor or the pump control console/point of
            sales device and the monitor. National electrical codes requke that the
78

-------
               Remote ATG Monitor
   Fill Pipe
Pump or Pump Control
                Console
       Temperature, and Water
              Sensor)
Figure 14. Schematic of automatic tank gauging system.

-------
                             Install Equipment
   installation
                                    J
      Probes in tanks
      Cables in conduits
      Monitor nearby
    Testing
       or
    Inventory
                             Program Monitor
                                         Coefficient of thermal
                                          expansion/product type
                                         Tank depth and volume chart
                                         Threshold criteria
                                         Test times
                                         Alarm levels
                         Shut Down Tank Operations
  I
  /ai

  I
                                       • Temperature to reach
                                          equilibrium
                                       • Tank deformation to subside
                  Take Temperature and Product Level Measurements
                                                    \
                  Reconcile Inventory Data
    Ana ysis
Detection
 Criteria
             Analyze Test Data
                                     Leak   No Leak
               Figure 15. General procedure for ATG Systems
80

-------
 wiring be installed in some kind of conduit to protect it from dirt and
 moisture and isolate it from the surroundings. For most ATG systems,
 four to eight.probes can be connected to one monitor. The height-
 to-volume conversion factor and the coefficient of thermal expansion
 (see Chapter 4 for definitions of these terms) must then be entered into
 the monitor by the installer.

 Installation time varies with conditions at the site and the number of
 tanksrbeing fitted with probes. For a four-tank site with no unusual
 problems, it typically takes about a day to install the conduits and
 wiring and a day to install the probes.
 Leak Detect Mode
Tanks must be taken out of service when the ATG system is in leak
detect mode. Most tanks, therefore, are tested at night when operations
at the site typically shut down. Most ATG systems can be programmed
at the monitor to automatically switch to the test mode at a preset time.
Alternatively, on-site staff can manually switch the ATG system to test
mode at the appropriate time. The test mode can be run as frequently as
the owner/operator determines.

The test is conducted at whatever product level is in the tank at the time
of the test. The large surface area of the product in an underfilled tank
(product level below the top  of the tank) means that any product loss
due to a leak would result in  a very small change in level. Because
testing is performed on underfilled tanks, the product level is essentially
constant during the test, which is necessary for an effective tank test
(see Chapter 4 for a detailed  discussion of constant-level vs.
variable-level tests).

When the ATG switches to leak detect mode, temperature and level
readings are taken automatically. The test can be programmed to last a
predetermined length of time. Also, some ATG systems can be
programmed for a desired minimum detectable leak rate; the
microprocessor then determines the necessary sampling frequency for
product level and temperature and for test duration. If a detectable leak
rate below 0.2 gal/h is selected, the probability of detecting such a leak
will probably decrease somewhat.  Level and temperature readings are
usually taken every 1 to 2 seconds and  averaged every 30 to 60 seconds.
The length of the test varies with the system and the level of sensitivity
desired. Tests generally range from 1 to 6 hours in length, with most test
lengths falling in the lower end of the range. As the number of probes
                                                                      81

-------
            connected to the monitor increases, the frequency of readings and the
            test length decrease because the probes are "read" in series and it takes
            longer to complete a circuit of eight probes than to read two probes.

            The level and temperature readings are fed to the microprocessor, which
            converts these readings to temperature-compensated volume
            measurements, analyzes the data according to some statistical algorithm,
            and determines a rate that indicates how fast product level is changing
            in the tank.  This rate must then be compared to a threshold value to
            determine if the level changes are due to a leak or to normally occurring
            volume fluctuations. If the temperature-compensated volume change
            exceeds the threshold, a leak is suspected; if not, it is assumed that the
            tank is not leaking.
            Inventory Mode
            The same level and temperature readings taken during the test mode are
            taken in the inventory mode, which is in operation any time a test is not
            being run. In addition to taking product level and temperature readings
            in the tank and converting them to volume measurements, some ATG
            systems measure and record the amount of product dispensed. For other
            ATG systems, the dispenser information must be recorded manually by
            on-site staff. Product deliveries also are recorded by the ATG system.
            Increases in volume in the tank that are above a minimum rate and
            volume are interpreted as a delivery.

            If the dispensing information is collected by the ATG system, the
            microprocessor automatically reconciles the inventory data at the
            interval programmed into the monitor; one hour is often selected as the
            interval. If staff record the dispensing information manually, the
            volume and delivery data from the ATG must be combined with the
            dispensing data and reconciled manually (see Chapter 2 on Inventory
            Control).

            As part of the inventory mode, the probe in most ATG systems also
            measures the level of water in the bottom of the tank. This information
            is converted to a volume and used in the inventory reconciliation. Other
            features included in many ATG systems are alarms for high product
            level, low product level, high water level, and theft (indicated by sudden
            large loss of product). The levels at which these alarms are triggered
            are programmed into the monitor.
82

-------
POTENTIAL PROBLEMS AND SOLUTIONS


            This section presents a discussion of problems that have been
            encountered with ATG systems. Many of the problems and solutions
            are similar to those for tank tightness testing, which are discussed in
            Chapter 4.  To avoid repetition, this chapter includes only a brief
            summary plus a reference to the relevant section of Chapter 4 for those
            problems that are common to both methods. Table 6 presents a
            summary of the indicators and solutions to problems with ATG systems
            as well as possible approaches that implementing agencies can use to
            prevent or overcome the problems.  A number of agency solutions are
            offered for each problem, but not all of them need be undertaken.
            Table 6 and the discussion below are presented in the order of the
            flow chart (Figure 15 on page 80).  There is no ranking implied by the
            order.  In Table 6, the most serious concerns are indicated by an
            asterisk.
            Installation
           Ensure that equipment is installed properly

                  Most problems with the installation of ATG systems are
                  associated with the installation of conduits for the wiring.
                  National electrical codes require that the cables be isolated from
                  other electrical wires.  Most UST sites already have conduits
                  containing wiring for other equipment, such as the dispensers,
                  and this wiring is often contained within one main conduit.
                  Installers of ATG systems sometimes use existing conduits for
                  the ATG wiring because it is easier and less expensive than
                  installing new conduits. If existing conduits containing wiring
                  are used, the ATG wiring must be isolated in some manner, such
                  as with flexible plastic casing. Any time wiring is found that is
                  not isolated, power to the wiring should be immediately shut off
                  and the problem fixed before the system is allowed to operate
                  again.

                 Installers sometimes neglect to continue the conduit/isolating
                 material around the wiring where the wiring enters the building
                 and connects to the monitor. In addition, installers sometimes
                 use sealing compound for the last several feet of wiring instead
                                                                               83

-------
00
  Table 6.  Indicators and Solutions for Problems Encountered With Automatic Tank Gauging Systems
   Problem
Indicators
Tester Solutions
Agency Oversight Options
  Assure correct installation
  of wiring.

  Use correct coefficient of
  thermal expansion.
   *Need for adequate waiting
   time before testing.
No readings or erratic readings.
False alarm.
Erratic and large volume
changes. Short-term volume
changes that level out.
   *Need for adequate temperature   None.
   compensation during the test.
   'Preventing Interference due to   None.
   evaporation and condensation.
Isolate wire in separate conduits,
or encase in isolating material.

Whenever type of product
changes, reprogram monitor.
Wait at least 6 hours after fuel
delivery before testing. Wait
as long as possible after UST
operations stop to begin test.
At 24-hour sites, test following
period of lowest use.

Use at least three temperature
sensors equally spaced vertically,
or use temperature-averaging
probe.

Do not open any fittings on top
of the tank for at least 6 hours
before the test and during the test.
Observe installation. Review
installation plans.

Compare product type to
program value during site
inspecion or review of reports.

Review printouts for product
delivery and test times.  Review
and approve testing schedule.
                                                                   Observe installation. Review
                                                                   product literature. Approve
                                                                   only ATG systems with
                                                                   appropriate designs.

                                                                   Observe test Train/educate
                                                                   staff at UST site.
   *The number of tests to be
   conducted at a tank must be
   fixed.
 High percentage of tanks declared
 "tight."
Protocol must include fixed
number of tests. Use
progammable monitor that can
be locked.
 Review printouts of tests;
 compare to protocol. Approve
 only programmable systems that
 cannot be overridden. Train/
 educate staff at UST site.

-------
   Correct threshold must be
   used to determine leak status.
   (A threshold is a predetermined
   value; measurements made
   during a test are compared to
   this value.)
  Accounting for the presence
  of ground water.
High percentage of tanks declared
"tight."
  Assuring testing at a range
  of product levels.
  Use correct height-to-volume
  conversion factor.
  Assuring appropriate treatment
  at UST sites with blended fuel
  systems.

  Assure calculations are
  correct.
Water in the tank.
Monthly tests are always
conducted at low product level.
None.
None.
Improbable results.
Threshold value should be
smaller than the minimum
detectable leak rate by a factor
of at least 2. Program monitor
compares leak rate to threshold
and triggers alarm if leak is
suspected.

Water sensor in the tank as
part of the ATG system.
Program tests to take place
at varying product levels that
cover the range typically stored.
Add or withdraw known volume
from tank, and observe level
changes.

Treat multiple tanks as one
unit for purposes of
reconciliation.

A second person double-checks
manual calculations.  Computer
program is reviewed and tested
using known and verified data.
Keep track of pass/fail ratio
of each type of ATG system.
Approve systems with adequate
thresholds or with alarm systems.
Approve only ATG systems
with water sensors.  Observe
installation.  Review release
detection plans and  manufac-
turer's product descriptions.

Review printouts to compare
range of product levels during
the month to levels during tests.
Review testing schedule with
knowledge of delivery schedule
and UST operations.

Review data sheets  for
calculation of conversion
factor.

Inspect site before system
installed. Review site plans
before installation.

Manual or computer
recalculation. Approve only
systems with automated
inventory.
  indicates the most significant problems.
00

-------
                    of for the last few inches. Both of these practices violate
                    electrical wiring codes and are unsafe.

                    Sometime the conduits that are installed are too small for the
                    amount of wiring that must run through them.  It takes extra time
                    and effort to install the wiring and creates the risk of damaging
                    the cables.  Generally, conduits should have an internal diameter
                    of at least 3/4 inch.
1
r
Testing or
Inventory
  Analysis
 installation |  Use correct coefficient of thermal expansion of the product
The coefficient of thermal expansion of a liquid relates changes
in its volume to changes in its temperature. The units of the
coefficient are change in gallons per degree Fahrenheit. The
coefficient is used to convert changes in temperature readings
taken by the probe to changes in volume as part of the
determination of leak rate. The coefficient of expansion is
different for each product. When an ATG system is installed,
the "average" coefficient for the product type in the tank is
programmed into the monitor.  If the type of product stored in
the tank is later changed (e.g., from gasoline to diesel) and the
coefficient is not changed, the calculated leak rate will be
incorrect. When the ATG system is installed, it is important that
the installer inform the on-site personnel of the need to
reprogram the monitor and inform them how to accomplish this
should the product be changed.
             Leak Test Mode
             Allow adequate waiting time before beginning test

                    When product is added to a tank, the product in the delivery
                    truck is at or near the air temperature while the product in the
                    tank is at or near ground temperature.  For some time after
                    delivery, the temperature within the tank fluctuates rapidly and
                    widely as the product mixes and eventually achieves
                    equilibrium. The addition or withdrawal of product causes the
                    ends of the tank to move outward or inward, respectively, in
                    response to the changing head pressure of the liquid on the tank
                    walls. The fluctuating temperature and structural changes of the
                    tank following the addition of product cause volume and level
                    changes that can be misinterpreted as changes caused by leaks.
                    See Chapter 4 for additional discussion of these phenomena.
86

-------
      Unlike the tank tightness testing methods discussed in
      Chapter 4, product is not deliberately added to a tank before a
      test is conducted with an ATG system.  Instead, the test is
      performed at the product level in the tank when the monitor is
      switched to the test mode. The level in the tank depends on
      routine product deliveries and withdrawals. No standard
      minimum waiting time between stopping tank operations and
      beginning to test will be applicable to all situations because the
      waiting will vary with the amount of use (i.e., number and
      amount of withdrawals) at that tank before operations ceased.
      The ATG system should be programmed to begin testing as long
      as possible after UST operations have stopped. If the UST site
      operates 24 hours per day, the test should be scheduled
      following the period of least use, such as late at night (at service
      stations).  The tank to be tested must be shut down during the
      test.  Any time product is delivered to an UST, at least 6 hours
      should elapse between delivery and testing.

      Another temperature-related problem can occur when product is
      added that is significantly higher in temperature than the product
      in the tank. In this situation, ATG systems that determine
      product level using capacitance probes may respond with a false
      alarm during a leak test.  Differences in temperature cause
      differences in densities, which in turn cause differences in the
      capacitance of the products. The waiting period discussed above
      should reduce the chance of false alarms caused by this problem.
      Most ATG systems using a capacitance probe also have a
      program built into the analysis that can usually detect this
      problem and declare the test invalid.

Adequate temperature compensation during the test

      After the temperature in the tank stabilizes, the product is
      usually stratified into layers of different temperature. Each layer
      will continue to undergo small changes in temperature, the
      extent and rate of the change being different for each layer. If
      the temperature changes of only one or two layers is recorded
      during a test, the temperature-compensated volume changes
      calculated for the entire tank will be incorrect, resulting in an
      erroneous leak rate determination. For additional discussion of
      this issue, see Chapter 4.
                                                                       87

-------
                   For the reasons given, the temperature sensors in the ATG
                   system probe must provide adequate spatial coverage of the
                   tank, i.e., the data recorded must be representative of product
                   conditions throughout the tank. At least five temperature
                   sensors arranged vertically provide the best coverage, although
                   three vertically arrayed sensors usually are adequate.  One
                   temperature sensor generally does not provide sufficient
                   coverage of the tank, with one exception. At least one ATG
                   system uses a temperature-averaging sensor that provides a
                   single temperature value over the entire depth of the liquid.

            Preventing interference due to evaporation and condensation

                   During a test with an ATG system, the tank is only partially
                   filled with product.  If there are at least 15 cm between the level
                   of the liquid and the top of the tank, which is the case with most
                   ATG tests, evaporation and condensation are potential problems.
                   As temperatures change, fluid will evaporate from the product
                   surface into the air space at the top of the tank and/or will
                   condense along the inner surface of the tank walls and drip back
                   down.  Evaporation and condensation of the liquid product cause
                   fluctuations in the product level that could mask or mimic a leak.
                   Eventually, the evaporation and condensation between the liquid
                   product surface and the air space above it will come into
                   equilibrium and no further change in the liquid level will occur.
                   If this equilibrium is upset, such as by opening the top of the
                   tank or adding product to the tank, the liquid level will fluctuate
                   until the tank regains equilibrium.

                   Before conducting an ATG test, the headspace of the tank must
                   be allowed to come  to equilibrium. To achieve this, none of the
                   fittings on the top of the tank should be opened for at least 6
                   hours before the test begins. The equilibrium must be
                   maintained during the test, again by not opening any of the
                   fittings on the tank.

            Fixed number of tests to be conducted

                   When the results of a test indicate a leak rate that is only slightly
                   above the threshold value for declaring a leak, some owners may
                   rerun the leak test mode on the tank until the results of one test
                   indicate that the tank is tight. Because conducting a test with an
                   ATG system is relatively simple and can be done during
                   nonbusiness hours when the tank is not in use, this approach to
88

-------
       testing may be tempting. This approach, however, is a misuse of
       the system and reduces the probability of detecting a leak.

       The same ease of testing that leads to this type of misuse also
       makes a valid multiple-testing strategy a likely option for ATG.
       In a multiple-testing strategy, a fixed number of tests are run,
       and the results of all the tests are used to determine the leak
       status of the tank. This strategy increases the sensitivity of the
       test and the likelihood of detecting a leak.  An UST site that
       closes operations for the  night or weekend could easily run
       several tests with the ATG system, and the microprocessor can
       be programmed to perform the necessary statistical analysis.
       Whether the ATG system uses a multiple-testing strategy or a
       single test, the key to successful leak detection is to define
       explicitly the number of tests and to carry them out.  Many ATG
       systems can be locked after they are initially programmed so that
       the protocol cannot be changed except by the person with the
       key. The manufacturer of the ATG system must determine the
       number of tests to be run to meet the regulatory performance
       standards and provide  this information.

Correct threshold value used to determine leak status

       After reading and analyzing the product  temperature and level
       data, the calculated volumetric leak rate  in gal/h must be
       compared to a threshold value  to determine if the tank is leaking.
       Some ATG systems just display the calculated leak rate and the
       comparison and determination of leak status is made by on-site
       personnel. Other ATG systems compare the calculated leak rate
       to a programmed threshold and display PASS or FAIL along
       with the leak rate. If the  wrong threshold value is used, an
       incorrect determination will  be made. For a test method to
       perform well in detecting small leaks, the threshold value must
       be smaller by a factor of two or more than the smallest leak to be
       detected. The federal regulation requires an ATG system to
       have a minimum detectable leak rate of 0.2 gal/h. For an ATG
       test to meet this requirement, its threshold must be less than 0.2
       gal/h. For most systems, the threshold will be around 0.1 gal/h.
       The exact value for each  system should be determined by the
       manufacturer and supplied to the owner/operator.
                                                                       89

-------
             Accounting for the presence of ground water

                    The presence of ground water around any portion of the tank
                    may lead to erroneous test results if there is a hole in the portion
                    of the tank under water. The pressure of the ground water
                    inward counteracts the outward head pressure of the leaking
                    product, either slowing or completely stopping the leak. The
                    ground-water level at a site generally fluctuates over time with
                    the seasons and the amount of precipitation.  For further
                    discussion of the influence of ground water on leaking tanks, see
                    Chapter 4.

                    ATG systems conduct tests at the level of product in the tank at
                    the time of the  test; therefore, the product level will be different
                    for each test. The ground-water level is  also likely to be
                    different during each test. It is unlikely that for each test the
                    levels of product and ground water will be such that the
                    pressures are equal, preventing product from leaking out of the
                    tank and water from entering the tank, particularly if tests are
                    conducted daily or weekly. Therefore, the test mode will detect
                    either a decrease in product level due to product leakage or an
                    increase in water level inside the tank due to ground-water
                    incursion. If at all possible, it is still preferable, however, to
                    conduct the leak test at a time when the ground water is below
                    the bottom of the tank.

                    It is also important to have a water level sensor as part of the
                    ATG system and to program the monitor to trigger an alarm at
                    some preset water level. Accumulation of water in the tank
                    indicates a possible hole in the tank that must be investigated.

             Ensuring testing at a range of product levels

                    ATG is conducted with the product level below the top of the
                    tank.  Some portion of the tank walls will not be below the
                    product surface during a test and, therefore, will not be checked
                    for holes. Either deliberately or by chance, the monthly tests
                    may always occur at a very low product level, allowing a
                    significant portion of the tank surface to  go unmonitored.

                    Given the variability in tank use, it is unlikely that, by chance,
                    the monthly ATG test would always be conducted at a low
90

-------
       product level. Over the months, the tests probably will be
       conducted over the range of product levels normally held in the
       tank.

       To counteract deliberate testing at low product levels,
       implementing agencies could require that data on the product
       levels in the tank throughout the month be submitted along with
       the product level during the test. A review of this information
       would show whether the tests are being conducted over the full
       range of product levels.

Correct height-to-volume conversion factor used

       ATG systems measure changes in product level over time.  To
       calculate a leak rate in gallons per hour, these level
       measurements must be converted to changes in volume. During
       installation of an ATG system, the tank manufacturer's tank
       calibration chart giving depth measurements and corresponding
       volumes is programmed into the monitor. Slight variations in the
       manufacturing process means that the chart is not precisely
       accurate for all tanks (see Chapter 2 for further discussion of
       tank charts  and how to use them). Usually, these minor
       variations do not interfere with the accuracy of the ATG results.
       Occasionally, however, the tank can differ enough from the
       calibration chart to cause significant error. If repeated false
       alarms occur, one possible factor to check is the height-to-
       volume conversion. To do this, product should be added to the
       tank in known increments, usually of 100 to 500 gallons, and the
       level measured using the probe.  These measured
       height-to-volume values should then be compared to those
       programmed into the monitor using the manufacturer's
       calibration chart. If they are substantially different, the actual
       measured values should be keyed into the monitor, replacing the
       calibration chart values. "Strapping" the tank is not performed
       at each ATG system installation because it is very time
       consuming.
Ensure appropriate treatment at sites with blended fuels

       At some service stations, the customer can select different
       blends of fuel at the same dispenser.  One dispenser is connected
                                                                      91

-------
                   to more than one tank. The dispenser has several "blenders" that
                   mix different fuels in preset ratios, such as 60/40.  To perform
                   accurate inventory control, the ATG system must have accurate
                   values of the amount of product withdrawn from each tank.
                   Some blenders do not have flow meters in the lines leading to
                   the blender to provide accurate product volumes supplied by
                   each tank.  The blenders may have 1 to 2 percent error, so that
                   applying the stated ratios to the total volume passed through the
                   blender will not determine the actual volumes withdrawn from
                   each tank.

                   If blended fuel dispensers are encountered that are not
                   sufficiently accurate, one solution is to treat the product in all of
                   the tanks connected to one dispenser as a single volume. Each
                   tank must have a probe to measure product level and deliveries
                   as usual, converting the level readings to volumes.  These
                   readings can then be combined and treated as one tank for
                   purposes of inventory control. The total amount withdrawn
                   from the dispenser is used to perform the reconciliation; no
                   determination of the amount attributable to each tank is
                   necessary because the tanks have been "combined." Combining
                   the tank readings is applicable only to the inventory mode; the
                   tanks must be tested separately in the leak test mode.

            Ensure correct calculations

                   Some ATG systems are not connected to the dispensers. In this
                   situation, the dispenser readings must be recorded manually.
                   The delivery and tank volume data are taken from the ATG
                   monitor, combined with the manual dispenser data, and the
                   inventory reconciliation is performed by hand or the data may be
                   entered into a computer spreadsheet.  Chapter 2 provides
                   information on how inventory data should be collected and
                   analyzed. Math errors may result from the hand calculations or
                   entry errors can occur when the data are input into the computer.
                   Any reconciliation other than that performed entirely by the
                   ATG system should be double-checked by another staff
                   member.
ENSURING EFFECTIVE AUTOMATIC TANK GAUGING
            The four general approaches that implementing agencies can use to
            ensure effective release detection are discussed in Chapter 1. The
92

-------
following sections discuss how these approaches may be applied
specifically to ATG systems.
 Site Inspections
The site may be visited during installation of the ATG equipment. The
most important features to inspect would be that (1) proper conduits are
constructed for the cables; and (2) the proper values for the coefficient
and the threshold were programmed into the microprocessor. After
installation, random site visits might be made to check test records and
inventory results.  The microprocessor and monitor also could be
inspected to ensure that no changes have been made in the
programming.
 Data Review
The implementing agency could require that information on how the
ATG system is installed and programmed be submitted and approved
prior to actual startup of the system. In addition, copies of the test
and/or inventory results could be mailed to the agency for review. Most
ATG systems can be connected to a printer so that little extra effort is
needed on the part of the owner/operator to obtain copies. Agency
personnel then could compare the leak rate values to the threshold value
for the brand of ATG in use.

Similar to an approach used for tank tightness testing results
(Chapter 4), the implementing agency could keep a tally of the number
of passes and fails for each type of ATG system and investigate those
systems with abnormally high proportions of passes.
 Guidance and Training
Guidance materials aimed at the owner/operator should emphasize the
proper timing of a test and the need to keep the tank closed both before
and during a test.  Guidance material aimed at manufacturers should
emphasize the design needs, such as the number of temperature sensors
and the water level sensor.
 Approval and Certification
An implementing agency can elect to review ATG systems and approve
for use in its jurisdiction only those systems that pass a review and
                                                                     93

-------
            approval process.  Los Angeles, California, requkes data from an
            independent third party demonstrating that the system meets the
            performance standards.  Massachusetts lists all of the elements that must
            be contained in the written request for approval, such as the principles
            of operation and whether the method tests the whole tank surface or just
            the surface below the product surface.

            Another approach an implementing agency can use is to require
            certification or training of all ATG system installers. The agency could
            run the certification program or could require a minimum amount of
            training by the manufacturers of the equipment. Most manufacturers of
            ATG systems already have some type of training program in place.
REFERENCES
             1. Maresca, J. W., N. L Chang, Jr., and P. J. Gleckler. January
               1988. A Leak Detection Performance Evaluation of Automatic
               Tank Gauging Systems and Product Line Leak Detectors at
               Retail Stations.  Vista Research Inc.

             2. Schwendeman, T. G. and H. K. Wilcox. 1987. Underground
               Storage Systems - Leak Detection and Monitoring.  Lewis
               Publishers. Chelsea, Michigan.

             3. U.S. Environmental Protection Agency. November 1988.
               Evaluation ofVolumetric Leak Detection Methods for
               Underground Fuel Storage Tanks. Vol.1. EPA/600/2-88/068a.
               Prepared for U.S.  EPA by Vista Research, Inc.

             4. U.S. Environmental Protection Agency. 1986. Under ground Tank
               Leak Detection Methods: A State-of-the-Art Review.
               EPA/600/2-86/001.  Prepared for U.S. EPA by IT Corporation.
94

-------
   Chapter VI
Vapor Monitoring

-------

-------
VAPOR  MONITORING
VI
SUMMARY
            Vapor monitoring is a relatively new release detection method that has
            been applied predominantly UST systems storing petroleum products.
            It has been used extensively in California, and in studies by the EPA it
            has been shown to detect small leaks quickly.

            Before vapor monitoring is selected for release detection, a thorough
            site assessment should be conducted to ensure that it is appropriate to
            use this method at the site.  Many of the problems that interfere with
            vapor monitoring can be addressed during the site assessment stage. If
            chosen as a release detection method, it should be noted that if a
            monitor indicates the presence of a high concentration of vapors, it does
            not necessarily mean there has been an UST release. Vapor monitoring
            results must be carefully interpreted to differentiate between spills,
            interferences, and releases.

            The discussion presented in this chapter covers many of the possible
            problems that may occur with vapor monitoring. This does not mean
            that all, or even most, of these problems will occur. Nor does it mean
            that all of the problems are of equal importance in tenns of frequency of
            occurrence or severity of impact to the effectiveness of vapor
            monitoring. Some problems, such as many of the environmental
            interferences mentioned, occur infrequently, while others have limited
            impact. Experienced vendors are well aware of these problems and how
            to deal with them. For example, an experienced vapor monitoring
            company knows better than to install a system in clay backfill or at a
            site where the monitor does not respond to the UST product. Release
            detection, however, is a growing industry, and new companies are being
            fonned with less experience. This chapter presents the full range of
            potential problems for educational purposes, not to imply that they will
            always occur.

BRIEF DESCRIPTION
           The two major components of a vapor monitoring system are the vapor
           monitoring well and the vapor monitoring device (sensor). Vapor
                                                                               95

-------
            monitoring wells are small-diameter wells (typically 2 to 4 inches)
            placed in the backfill material near the UST which are used for
            collection of vapor samples. A monitoring device can be permanently
            or temporarily placed in the vapor monitoring well to collect vapor
            samples. A few monitoring devices are simply buried in the UST
            backfill and do not requke a monitoring well.

            Vapor monitoring works according to the principles of volatilization
            (i.e., the change of a substance from a liquid to a gaseous state) and
            diffusion (spreading of a gas).  As a product leaks from an UST, some
            of the liquid volatilizes, and the liquid and vapor phases of the product
            spread throughout the surrounding soil. Vapor monitoring systems take
            advantage of this phenomenon and are designed to detect the volatile
            components of a stored substance. If the vapor sample that reaches a
            sensor is above some predetermined concentration, the monitor
            responds with an alarm.

            A typical vapor monitoring system is depicted in Figure 16.

            The successful implementation of a vapor monitoring system involves
            six different stages: (1) site assessment—an evaluation of the site is
            conducted to ensure vapor monitoring is an appropriate release
            detection method and to determine the site characteristics; (2) sensor
            selection—an appropriate type of vapor monitor is chosen; (3) network
            design—the proper placement (lateral and vertical) of vapor well(s) is
            planned; (4) construction and installation—vapor well(s) construction
            and installation are conducted; (5) operation and maintenance—the
            sensors are calibrated and monitoring begins; and (6) data
            interpretation—the monitoring results are evaluated and leak status of
            the UST is determined. The relationships among these  six stages are
            shown in Figure 17.

POTENTIAL PROBLEMS AND SOLUTIONS
            If vapor monitoring is installed and operated correctly, it can be an
            effective leak detection method.  The following sections discuss
            problems and solutions related to each of the six stages of the
            implementation of a vapor monitoring system. The order of discussion
            is not intended to prioritize the importance of the problems, rather it is
            intended to follow the order in which they would occur according to the
            flow chart in Figure 17. The discussion of problems and solutions is
            summarized in Table 7, which starts on page 99; the more serious
            concerns are marked by an asterisk.  A number of agency solutions are
96

-------
  Vapor
Monitprin
   W
                  Vapor
                Monitoring
                  Device
        Backfill
Native Soil
         Figure 16. Underground storage tank system
         with vapor monitoring wells.
                                                                       97

-------
                            Site Assessment
                                        Product volatility
                                        Backfill permeability
                                        Background contamination
                                        Other possible interferences
                     Sensor Selection Considerations
    Installation
    Device threshold and detection
      range
    Device specificity
                            Network Design
                                      • Design of UST system
                                         characterization
                                      • Design of well placement
                                         and depth
                       Construction & Installation
                                      • Placement of well casing,
                                         filterpack, bentonite seal,
                                         surface seal, protective casing
                                      • Well security
                                    i ' • Documentation
                        Operation & Maintenance
    Operation
I
Calibration of equipment
Setting alarm level
Maintenance
                             Interpretation
     Ana ysis
                   Leak
                   No Leak
                                        Differentiating between
                                         interferences and leaks
                                        Locating a leak
              Figure 17. General procedure for vapor monitoring
98

-------
Table 7.  Indicators and Solutions for Problems Encountered During Vapor Monitoring

Problem    	indicators	Tester Solutions	Agency Oversight Options,
Assure volatility of the
stored substance.
Assure LIST backfill is
permeable.
*Need to assess the level
of residual vapors at the site.
None.
Backfill is not sand or pea gravel.
False alarms.
Assure low soil temperature
won't interfere.
Assure backfill that is saturated
with water won't interfere.
Low temperatures (below 0°C) for
extended periods.
Standing water in vapor wells.
Assess possible interference
from methane contamination.
False alarm.
Choose proper monitoring
system; consult manufacturer to
verify that sensor responds to
product.  Add tracer compound.

Test backfill permeability by
conducting tracer test. Increase
monitoring well diameter; use
more wells or aspirated sensors.

Add spill and overfill protection
to the tank.  Determine back-
ground concentrations. Set
monitoring device threshold
above background. Add tracer
compound.  Aerate the soil to
reduce concentration

Install monitoring wells below
frost line. Increase number of
monitoring wells.

For wet climates, use portable
monitor in dry areas.
Do not use for saturated sites;
consider ground-water
monitoring.

Check site for methane. Choose
monitoring system not sensitive
to methane.
Check product vapor pressure
in chemical handbook; review
manufacturer's literature.
Verify response to tracer.

Review results of backfill
permeability test Review
monitoring system plans.
Review background
determination. Check
that threshold is above
background level.
Review monitoring system plans;
compare to expected soil
temperatures.

Review local water table data
to ensure monitor is above high
water table.

Require testing of water content
of soils.  Require monitor that
detects the presence of water.

Require testing for methane in
areas known to have problems.
                                                                                                                      Continued

-------
       Problem
                                  Indicators
                                    Tester Solutions
                                  Agency Oversight Options
o
o
Assess possible interference
from nearby UST sites.
       Assure device responds to
       the stored product.

       Assure device will respond to
       level of contamination.

       Identify UST configuration.
       Assure monitoring wells are
       placed properly.
False alarm at monitored site.
       Assure well screen is designed
       properly.
       Assure filter pack is properly
       designed.

       Assure well is sealed property.
       Property secure and mark
       monitoring well.
                                 Device does not react to product.


                                 Device shows erratic behavior
                                 or no response.

                                 None.


                                 Delayed detection or no detection.
       Assure proper well construction.     Well collapses.
                                 Well holds vacuum when purged
                                 with hand pump.
                                 Well holas vacuum when purged
                                 with hand pump.

                                 False alarms. Standing water or
                                 product in vapor wells.

                                 False alarms. Standing product
                                 in vapor wells.
Install background monitoring
well near site boundaries.

Use tracer compound.

Use devices as recommended by
manufacturers.

Use devices as recommended by
manufacturers.

Examine construction records.
Use metal detector.

Install at least one sensor per
tank in highly permeable backfill.
Install at least two sensors per
tank in less permeable backfill.
Install sensors within 2 feet of
bottom of tank.

Construct well  according to State
codes.

Use standard size No. 20 slots,
and maximize  length of well
screen.

Use appropriately sized filter
pack material.

Seal well properly with cement/
bentonite.

Mark and lock  well.
Check for nearby USTs that are
not monitored. Test vapor
concentrations near that site.
                                                                       Test device during site
                                                                       inspection.

                                                                       Check device during site
                                                                       inspection.

                                                                       Request site plan with
                                                                       monitoring plan.

                                                                       Review monitoring system
                                                                       plans—compare well
                                                                       placement with local
                                                                       regulations or manufacturer's
                                                                       specifications.
                                                                       Inspect well and documentation
                                                                       of well development.

                                                                       Review well designs (especially
                                                                       slot size or filter pack).
                                                                       Test with hand pump.
                                                                       Review well design for proper
                                                                       seal; check well box for damage.

                                                                       Inspect site to see that well
                                                                       is distinguished from fill pipe
                                                                       and well cover is locked.

-------
 Need to document well
 construction.
 Assure equipment is properly
 calibrated.
 Need to set the alarm level
 correctly.
Assure proper maintenance
for equipment.

'Interpret monitoring results
considering possible
interferences.
Troubleshooting is time-consuming.
False alarm or lack of detection.
False alarm or no detection.



False alarm or no detection.


False alarm.
Locate wells so that the leak
source can be identified.
 Provide well log or other
 construction data.
Conduct at least annual
calibrations using standard based
on lightest compound of stored
product.

Set at least 50% higher than
background.
Maintain per recommendations
of manufacturer.

Verify proper operation of
equipment. Take second
reading.

Check for spills, other contamin-
ation. Check trends in monitoring
records.

Check for possible spills.  Use
other methods to confirm.
Increase number of monitoring
wells. Use quantitative monitor-
ing device.
Require monitoring plans and
well drawings be kept on-site
or submitted to agency.

Require calibration by approved
contractor.

Inspect calibration records.

Review records of alarm level
settings—compare with initial
background levels.

Review maintenance records.
                                                                         Inspect monitoring records.

                                                                         Check to see if high readings
                                                                         correspond to deliveries or
                                                                         other possible spills.
                                                                         If high readings do not appear
                                                                         to be a spill, require additional
                                                                         testing with other monitoring
                                                                         methods.
* Indicates the most significant problems.

-------
             offered in the table for each problem, however they are suggestions and
             not all of them need be undertaken.
             Site Assessment
I Installation I  Assure volatility of the stored substance
 Operation
  Analysis
Volatility is the tendency of a product to change from a liquid to
a gas under standard temperature and pressure. As discussed
previously, vapor monitoring works according to the principles
of volatility and diffusion.  A stored substance must be
sufficiently volatile or vapor monitoring will not work. Vapor
monitors are not appropriate for UST systems that contain
non-volatile products. The volatility of a substance is measured
by its vapor pressure.  Table 8 lists the vapor pressures of some
common petroleum products.
             Product
                       Table 8
    Typical Vapor Pressures of Petroleum Products

                 Vapor
               Molecular
                Weight    True Vapor Pressure in psia at:
                @60°F     40°F       60°F        90°F
Gasoline
Gasoline
Gasoline
Jet naphtha
Jet kerosene
Distillate fuel No. 2
Residual oil No. 6
62
66
68
80
130
130
190
4.7
3.4
2.3
0.8
0.0041
0.0031
0.00002
6.9
5.2
3.5
1.3
0.0085
0.0074
0.00004
11.7
8.8
6.2
2.4
0.021
0.016
0.00013
                   The volatility of different products varies widely. For example,
                   gasoline, which contains lighter hydrocarbons, is more volatile
                   than diesel fuel, which is composed of heavier hydrocarbons.
                   Typically, vapor monitoring is an appropriate monitoring method
                   for most petroleum products.
 102

-------

Operation
Analysis
      Device manufacturers generally will provide a list of substances
      to which their vapor monitoring device will respond. To verify
      the manufacturer's claim, the monitoring device can be exposed
      to a known concentration of the stored substance and checked to
      see that the device gives an appropriate response.  Several
      California counties require inspectors to check a device's
      response to the stored product before it is installed.

      For less volatile products, a tracer compound may be combined
      with the stored product to satisfy the volatility requirement. A
      tracer compound is added solely for the purpose of providing a
      volatile component to the product. When a stored product
      containing a tracer is released, the tracer compound volatilizes
      more easily than the pure stored substance, making release
      detection by a vapor monitor easier. Freon or non-chlorinated
      compounds (e.g., Stoddard solvent) are often used for this
      purpose.  If a tracer is used, it should be established that the
      chosen monitoring device is sensitive  to the compound being used
      as a tracer, that the tracer and the stored substance can be mixed,
      and that the tracer will not interfere with the normal use of the
      stored substance.  Under some circumstances when using a tracer,
      a leak rate of as low as 0.0005 gal/h can be detected.

Assure UST backfill is permeable

      Vapor monitors work best in permeable materials. If the backfill
      surrounding a tank is not sufficiently permeable, vapors may have
      difficulty moving throughout the monitored area. Loosely
      packed, large-grained soils are more permeable than tightly
      packed, fine-grained soils. For example, gravel is more
      permeable than sand, which is more permeable than silt, which is
      more permeable than clay. Sand,  gravel, or other engineered
      backfills typically are recommended when vapor monitoring is to
      be used. However, many manufacturers of vapor monitoring
      devices have successfully operated their systems in clay
      materials.  Figure  18 illustrates, for different types of soils, the
      speed with which gasoline vapors would reach a sensor.
                                                                                   103

-------
                          6000
                                                                    Dry gravel backfill
                                                                    Dry sllty sand
                                                                    Native soil
                                                                    Dry gravel backfill
                                                                    Dry silly sand
                                                                    Native soil
                                                                    Moist sand backfill
                                                                    Wet silty sand
                                                                    Natve soil


                                                                    Moist sand backfill
                                                                    Moist sand
                                                                    Native soil
                                                                    Wet sand backfill
                                                                    Wet sand
                                                                    Native soil

                                                                    Wet sand backfill
                                                                    Wet clay
                                                                    Native soil
                                           10

                                           Number of Days
                                                                   30
                   Figure 18. The effect of soil conditions on vapor
                   concentrations at a well 8 feet deep located 6.4 feet from
                   the source of the leak.  Source:  U.S. EPA (February 25,
                   1988).
                   Backfill at new UST installations should be sand or pea gravel,
                   either of which ensures sufficient permeability.  However, at
                   existing sites the soil used to backfill an UST may not meet this
                   criterion (e.g., the backfill may be either native soils, clay, or silt).
                   Some jurisdictions in California check the permeability of the
                   backfill at existing sites by using a tracer test.  This is performed
                   after wells are installed by injecting a tracer in one well and
                   monitoring for its occurrence in another.
104

-------
[ Operation

I  Analysis
      If a site has soil of moderate or low permeability (e.g., silt or
      clay), vapor monitoring can still be used if some modifications
      are made to the monitoring wells or the monitoring well network,
      or if the appropriate type of monitoring device is selected.
      Depending on the soil, one or a combination of these solutions
      may be necessary. The monitoring wells can be modified by
      increasing the well diameter to approximately six inches. The
      monitoring well network can be designed to include more wells
      than typically would be required, which provides a greater
      number of vapor sampling points.  Typical monitoring well
      networks are discussed later in this chapter under the discussion
      titled "Assure monitoring wells are properly placed for effective
      vapor detection."

      A modification to enhance vapor movement in moderate to low
      permeability soils is to choose an aspirated vapor monitoring
      device. Vapor monitoring devices are available in both aspirated
      and passive forms. Aspirated devices use suction to create a low
      pressure area around the sensor, thus drawing the vapors through
      the surrounding media to the probe. A passive device waits for
      vapor to migrate to the sensor naturally.

Need to assess the level of residual background vapors at the site

      Vapor monitors may respond to vapors remaining from previous
      spills or leaks, falsely indicating a current leak. Use of a vapor
      monitoring system is not recommended, without further
      investigation, at sites with high background concentrations (e.g.,
      above 1,500 ppm for gasoline). A  new site with clean backfill
      typically has levels of contamination that fluctuate between
      0 and 500 ppm. The level of background contamination that
      renders a monitor unusable varies for different vapor monitoring
      devices. The manufacturer of a chosen sensor should be
      contacted to determine the level of background contamination that
      precludes the use of its sensor.

      To determine background concentrations, a temporary vapor well
      can be installed within the UST excavation area, and the chosen
      monitoring device can be used to get an initial reading. If the
      monitor indicates that background concentrations are high (e.g.,
      above 1,500 ppm for gasoline), further investigation should be
      undertaken to determine whether the concentrations are due to a
      current leak (levels above  4,000 ppm for gasoline may indicate
      this), a spill, or off-site interference.
                                                                       105

-------
I Operation [

|  Analysis |
                  When background contamination is due to a past release or
                  off-site interference, vapor monitoring is still appropriate if the
                  contamination levels are below the alarm threshold limit of the
                  chosen monitor (i.e., the level at which a monitor is set to react to
                  the presence of a substance as if there were a release, typically
                  about 500-2500 ppm). Some instruments have adjustable
                  threshold limits. If the background contamination levels exceed
                  the instrument's threshold limit, the site can be injected with air to
                  lower the level of contamination. This can be done by using an
                  air pump to inject low levels of air through temporary wells into
                  the soil. A vapor monitoring method should not be used when
                  background contamination levels cannot be reduced below the
                  instrument's threshold limit, unless a tracer compound is
                  introduced. The use of a tracer avoids the problem of background
                  contamination because the vapor monitor will react to the tracer
                  compound, not to the compounds that are contained in the
                  background contamination.

                  To prevent future background contamination, overfill protection
                  and spill containment should be installed when vapor monitoring
                  is used.

            Assure environmental conditions will not interfere

                  Temperature can be an inhibiting factor for proper vapor monitor
                  operation at UST site; the colder the temperature, the less volatile
                  a substance will be.  Figure 19 illustrates the difference in
                  volatilization rates for gasoline at different temperatures.
                  Generally, for approximately every 20-degree Fahrenheit increase
                  in temperature, the gasoline volatilization rate increases by about
                  one-third.
106

-------
 .s
 in
                       Number 01 Days
Figure 19. The effect of temperature on gasoline
volatilization rates.  Source:  U.S. EPA (February 25,
1988)
When monitoring wells extend below the frost line, temperature is
not a problem. Increasing the number of monitoring wells, s that
there are more sampling points to pick up the lower levels of
vapor, can compensate for continuous low temperatures.

If the backfill is saturated with water, because of a perched water
table, fluctuating water table, rainfall, etc., vapor monitoring
devices cannot be used. Saturated backfill conditions will inhibit
vapor movement. Figure 20 on the following page illustrates the
difference in the volatilization rate of gasoline for three different
soil moisture conditions.
                                                                107

-------
                         e
                         o
                         S
                         g
                         to
                         O
                            0,00012-
                            0.00010-
              0.00008-
              0.00006-
                         >  0.00004-
                            0,00002-
                                            10
                                                       —I—
                                                        20
                                                                   Dry gravel backfill
                                                      Moist sand backfill

                                                      Wei sand backfill
                                                                  30
                                              Number of Days
I Operation
    3T~
  Analysis
 108
      Figure 20. The effect of backfill moisture levels on
      gasoline volatilization rates.  Source: U.S. EPA
      (February 25, 1988)

      Additionally, if a vapor sensor is immersed in water, it will be
      rendered ineffective in most cases. Often a portable vapor
      monitor, rather than a permanent one, can be used in wet climates
      because the sensor is not actually installed in the vapor well.

Assess possible interference from methane contamination

      Some chemicals can interfere with the operation of a vapor
      monitor, causing it to react to their presence as if there were a
      release. Methane, which may migrate from near-by marshes,
      landfills, sewage lines or sewage treatment plants, is the most
      common chemical causing this reaction. Many vapor monitoring

-------
                   devices are able to distinguish among different volatile
                   compounds. If methane is present at a site, the chosen vapor
                   monitoring device should not be one that reacts to methane. This
                   information should be available from the manufacturer.
| Installation |  Assess possible interference from nearby active or abandoned UST sites
 Operation
     +
  Analysis
      If there is an active or abandoned UST site with a history of
      releases in proximity to the site being evaluated, it may cause
      interference with vapor monitoring. Vapors from a nearby site
      may travel to the monitored site causing the sensor to react as
      though there were a release from the monitored site when there is
      not. One way to differentiate between monitor alarms due to
      on-site releases and off-site interferences is to install a
      background well outside the excavation zone, on the side nearest
      the source of the potential interference. If the vapor levels
      increase in the background well while remaining relatively steady
      in the on-site wells, the monitor alarm is most likely due to
      outside interference.

      Another method of differentiating between releases and off-site
      interferences is to introduce a tracer to the monitored UST
      system. The vapor monitor will then react specifically to the
      tracer, thus ensuring that when the monitor alarms, it is reacting
      to the monitored UST.
             Sensor Selection
Operation
 Analysis
Assure vapor monitoring device responds to the stored product

      Specificity is the ability of a device to detect vapors from a
      particular substance being stored on site (e.g., hydrocarbons or a
      tracer). If the monitoring device does not react to the stored
      substance, it is totally ineffective. Equally important, of course, is
      a device's ability to avoid reacting to  substances for which the
      site is not being monitored (e.g., methane).

      To determine which vapor monitoring device is suitable for
      different stored products, literature provided by the device
      manufacturer should be consulted. Most vapor sensor
      manufacturers list the types of stored products that their device
      will effectively monitor.
                                                                                    109

-------
 Operation"]
    *
 Analysis
Assure selected device responds to the level of contamination

      For a device to be effective at a site that has high background
      concentrations, it must be able to record and monitor a high level
      of vapors. Such a monitor is necessary because the background
      level at a contaminated site is significantly higher than zero, and
      the monitor must be able to evaluate and record levels higher than
      the threshold. If a device does not have a suitable threshold level
      or a broad enough detection range (i.e., the lowest and highest
      levels of vapor a monitor will detect), it will not accurately reflect
      changes in concentrations at a site, thus making the identification
      of a release difficult if not impossible.

      The appropriate monitoring device to use when there are high
      background levels is one that is responsive to a high level of
      vapors.  In any case, the level of background contamination and
      the desired range of detection should be considered before
      choosing a monitoring device.  Specific information about a
      device's threshold and detection ranges should be available
      through the manufacturer.
              Network Design
[Installation']  Identify system configuration to prevent damage
 	Jr'"	
 Operation
     *
  Analysis
      The number and location of USTs and associated piping must be
      identified to help determine where the monitoring devices should
      be installed to ensure efficient leak detection and to avoid
      damaging the UST system during installation of the monitoring
      wells. Information about the UST system should be available
      from past installation records and construction plans.

      Information obtained from available records should be reviewed
      with the site owner and operator to determine if any  alterations
      not indicated in the records were made to the UST system.
      Additionally, a metal detector or ground penetration radar can be
      used to determine the general location of tanks and piping.
 110

-------
  T
Operation I
Analysis
Assure monitoring wells are placed for effective vapor detection

      Improper location of vapor sensors will increase the amount of
      time before a release is detected and may, in extreme cases, allow
      a leak to go undetected. Detection time is a function of the
      distance between monitoring wells and a leak.  The placement
      and depth of monitoring wells are affected by the configuration of
      the UST system, the mobility and volatility of the stored product,
      the permeability of the backfill and surrounding soil, and the type
      of vapor monitoring device.

      The federal UST regulation requires that vapor sensors be placed
      in wells that are installed in the UST excavation area (backfill).
      In addition, vapor wells are normally placed as close to the tank
      as is technically feasible; sometimes this is accomplished by
      installing the wells at a slant instead of vertically.

      Although travel times vary, a rule of thumb is that hydrocarbon
      vapors will migrate 15 feet in about 15 to 20 days in unsaturated
      sand or gravel backfill.  This seems to be a reasonable assumption
      based upon sandbox experiments and computer models (U.S.
      EPA, February 1989). State and local agencies have adopted a
      variety of network design requirements for vapor monitoring (see
      Table 9).

      Although the requirements are diverse, they tend to require wells
      to be separated by no more than 20 to 35 feet. These
      requirements seem reasonable  based on EPA research and field
      experience indicating that a design that includes at least one well
      every 40 feet should be sufficient for gasoline tanks in a clean dry
      backfill. In general,  this translates to one well for a single tank.
      If the backfill is not highly permeable (e.g., it is native fill
      material) or the migration of vapors  is impeded by other factors, it
      is recommended that the number of sensors be increased by a
      factor of two.

      The ideal depth of a vapor well, as indicated by industry
      recommendations, would be a depth at least equal to that of the
      base of the tank, and preferably one to two feet deeper. Even if
      vapor monitoring wells are installed at this ideal depth, few wells
      will be deeper than 10-15 feet. Research indicates that diffusion
      (spreading of vapor in many directions) is the dominant type of
                                                                                  111

-------
                                                    Table 9
                Typical State Network Design Requirements for Vapor Monitoring
State
Vapor Well Location Requirements
Vapor Well Depth Requirements
California

Santa Clara County
City of Torrance


City of Vemon

Delaware


Maine
Iowa
General: every 35 feet of long dimension (if
        passive monitoring device, need more
        wells).
At station: 1 per tank
          1 piping
          1 pump island

For 1 tank: one at each end
More than 1 tank:  every 20 feet

Design for 15 feet diameter of influence

1 per tank within 5 feet of tank


According to manufacturer's specifications.
At a minimum:
  -  1 within 5 feet of each dispenser
  -  1 at each piping joint
  -  no piping run > 15 feet from well
  -  1 at each end of tank

  -  1 at longitudinal ends of a single tank
  -  if cluster of tanks where tanks ^ 10 feet apart,
    at least 4 wells, 1 on each side of cluster
  -  all wells > 1 feet from nearest tanks
  -  all wells within excavation zone
At least to bottom of tank
N/A
N/A

2 feet below tank bottom or to ground water,
whichever is less

Manufacturer's specifications
                                                                    2 feet below bottom of tank
Source:  Reference 15.
112

-------
  vapor movement over time, but because most leaks occur in the
  bottom of the UST, a deeper monitoring well may reduce the leak
  detection time. Deeper wells, however, are not mandatory for
  effective leak detection. A rule of thumb is that the less
  permeable the backfill is, the closer the vapor well should be to
  the recommended ideal depth. Figure 21 gives a comparison of
  the time it takes for gasoline vapor to reach wells at various
  depths and distances from a release.
       6000-
       7000-
SENSOR PROXIMITY
DEPTH  TO LEAK

Shallow  Near
(2 ft)    (2.2 ft)
                                             Deep
      Intermediate
      (6.4 tt)
                                             Shallow  Intermediate
                                              «)   (6.4)
                                             Deep   Distant
                                             (Btt)   (13.1ft)
                                             Shallow  Distant
                                             (2 ft)   (13.1ft)
                        Number of Days
Figure 21.  The effect of vapor sensor placement on leak
detection time.  Source: U.S. EPA (February 25,  1988)
Whether an aspirated or passive sensor is chosen also affects well
depth and placement. For most soils, an aspirated system will
detect vapors more quickly than a passive system at a given
depth, since the aspirated system draws the vapor to the sensor.
However, if the soil is very permeable (e.g., gravel), an aspirated
sensor system and a passive sensor system will perform similarly
                                                                  113

-------
                  at the same depth. Some counties in California require slant wells
                  that extend beneath the tank if a passive vapor monitoring device
                  is to be used.

                  Figure 22 shows a typical configuration for well placement at a
                  gas station. This configuration includes wells within the tank
                  backfill area, and wells located so that sensors will detect releases
                  from piping at the end of pump islands in the pipeline backfill
                  area. A background well, although not necessary for effective
                  leak detection, can be installed at a location upgradient from the
                  pipelines and the tanks to evaluate surrounding soil vapor levels.
             Construction and Installation
| Installation^  Assure proper monitoring well construction
 Operation
  Analysis
Improper construction of monitoring wells can render vapor
monitoring ineffective (e.g., surface runoff could enter the well,
the well casing could collapse, etc.). Construction of vapor
monitoring wells and the installation of vapor monitoring devices
should be done by a qualified contractor who is aware of any
specific state requirements or any industry codes that may affect
construction and installation.  Figure 23 shows a cross-section of
a typical monitoring well.

There has been discussion about monitoring well construction for
vapor monitoring wells, especially when ground water in the area
is deep and the wells are installed in the UST backfill. Santa
Clara County in California has thousands of vapor monitoring
wells that are modified in construction. In many cases a modified
well structure, as shown in Figure 24, may be perfectly
acceptable.

A typical vapor  well is less than 6 inches in diameter. The casing
may be polyvinyl chloride  (PVC), cast iron, galvanized steel,
polyethylene, polypropylene, fluorocarbon resins, Teflon®, or
stainless steel. In choosing the type of casing, the local site
conditions should be considered.  For instance, galvanized steel
casings deteriorate in corrosive environments.  The most
commonly selected materials for a monitoring well are PVC or
stainless steel. Both of these materials meet the structural needs
of the vast majority of vapor monitoring wells.
 114

-------
                                          Pump Island
                                 Vapor Well In
                            Product Line Backfill
                                          Tank Backfill Area
                         Product Lines
          Background Vapor Well
                             Vapor Well
                           In Tank Backfill
                                                                   TANK
                                                                   TANK
                                                                   TANK
                                                            Official Storage
Figure 22.  Sketch of typical underground storage tank site.
                                                                                      115

-------
    Existing Grade
Watertight Manhole
   Threaded Cap
   Concrete
  Annular Seal
   Casing
   Gravel Pack
   Porous Backfill
                                                             Well Screen
                                                            Well Plug
Figure 23. Typical vapor monitoring well cross section.
116

-------
 Existing
  Grade
Water-tight manhole,
security box,
or concrete cap
 Native Soil
 Tank Backfill
                  Well Cap
                  Seal of Less
                  Permeable  ~
                  Material than
                  Backfill
                  Well Casing
                  Perforated
                    Casing
                  Well Plug
Figure 24. Modified vapor monitoring well cross section.
                                                                          117

-------
jlnstaljiiiitip'f^l  Assure the well screen is designed for proper influx of vapors
 Operation
     I
  Analysis
I Installation I

I Operation I
  Analysis
[installation]
   '  *     .
 Operation
  Analysis
      The casing has a slotted or perforated section (called the well
      screen) that allows for influx of vapor to the monitoring device. If
      the well screen has perforations that are too large it may become
      clogged with surrounding soil particles, thus blocking the influx
      of vapors. Gathering vapor samples is similarly inhibited if the
      perforations are too small or only cover a short length. A typical
      well screen would have the standard size #20 slots. The well
      screen section usually begins from 2 to 5 feet below the ground
      surface and extends to the base of the casing. In general, the well
      screen extends over as much length as is possible.

Assure filter pack is designed to prevent clogging of the well screen

      The well screen area should be surrounded by a filter pack that
      allows for passage of vapors while preventing passage of
      fine-grained soil particles that could clog the well screen.  If the
      filter pack material is of too small a size, it may block the passage
      of vapors and clog the well screen; if it is too large, soil particles
      may migrate through the filter pack and clog the screen.
      Typically, several inches of filter pack are placed in the bottom of
      the borehole before the well casing is installed. The filter pack
      should extend 1-2 feet above the well screen. Materials other
      than carefully graded gravels that are acceptable for filter pack
      include clean quartz sand, silica, and glass beads.

See that well is sealed to eliminate introduction of contaminants

      The area outside the casing, above the well  screen, should be
      sealed (annular seal) to prevent contaminants such as infiltrating
      surface water or other liquids from entering the well that may
      interfere with monitoring or reach the ground water.  A
      cement-bentonite mixture, bentonite chips,  or antishrink cement
      mixtures are normally used for this purpose. The annular seal
      usually extends for 1 or 2 feet above the filter pack. A concrete
      seal is placed above the annular seal up to the ground surface to
      provide additional protection to the well casing from
      contamination and physical damage. Ideally, the interface of the
      bentonite seal and the cement seal should be located below the
      frost line to protect the well from damage due to frost heaving. A
118

-------
 Operation
  Analysis
    JL
 Operation
    JL
  Analysis
      protective steel casing of a diameter larger than the monitoring
      well can be placed in the cement seal to provide additional
      protection to the well.

Secure and mark vapor monitoring well properly

      If a well is not secured and marked properly, it may be damaged
      accidentally causing interference with its integrity. Monitoring
      wells may be mistaken for other pipes, such as a tank fill pipe, in
      which case the well may be inadvertently filled with product.  To
      prevent tampering, the well must be secured with a threaded cap,
      covered and locked. The well should be visibly marked to
      prevent accidental damage, and service stations should protect
      monitoring wells with a traffic box to prevent vehicle damage.

Document well construction properly

      To aid in future identification of well problems and to prove that
      the monitoring wells comply with state codes, the design and
      construction of each monitoring well should be documented on a
      well completion diagram.  This diagram indicates well design
      specifications, including the type  and depth of filter pack, annular
      seal, concrete seal, well diameter, and well screen design. In
      addition, drilling and boring logs should be completed indicating
      the depth of the well and the type of backfill in which it was
      placed. This documentation should be useful to state or local
      agencies  to determine whether correct procedures relating to
      design and installation were followed.
             Operation and Maintenance
| Installation]  Calibrate equipment properly to detect vapors from stored product

  Analysis
      The most important step for successful operation of a vapor
      monitor is the initial calibration of the device, which  should be
      performed by a professional. Calibration consists of exposing the
      monitor to a pure gas standard to ensure that the monitor correctly
      responds to vapors. If a device is not calibrated correctly, it is
      likely to give either false positive or false negative results. False
      negatives occur when the device is calibrated to indicate low
      concentrations of vapor when a high concentration is present.
      False positives occur in the opposite situation.
                                                                                     119

-------
                  Monitoring devices should be calibrated at least annually,
                  specifically to the substance stored in the monitored UST, and
                  preferably to the lightest component of that substance. The
                  lightest component is that which volatilizes the easiest. Santa
                  Clara County in California requires that devices used at sites
                  containing petroleum products be calibrated annually using
                  certified 1000 ppm isobutane (the lightest component of
                  gasoline). Some areas require that  tank owners use an approved
                  contractor or agency to perform an annual calibration.
 Installation   Need to set the alarm level to avoid false alarms
 Operation:
    "
 Analysis
      An alarm level is the level of concentration of vapors in the soil
      that triggers an audible or visual alarm.  The alarm level is
      determined by the professional after die device is installed. If an
      alarm level is set too low, the alarm will be triggered by small
      spills, interferences, or normal fluctuations and become a
      nuisance; if set too high, it will not be triggered by a real release.
      Some monitors do not have alarms, in which case the operator
      should be aware of the vapor reading level that indicates a
      possible leak.

      To select an appropriate alarm level, a background reading is
      needed to determine the site's current condition. The alarm level
      should be set to a value at least 50 percent higher than the
      background concentration (EMSL-Las Vegas). In general, for
      gasoline, vapor levels of 3,000-4,000 ppm with an increasing
      trend will be indicative of a leak. However, this level will vary
      from site to site and for different brands of monitoring devices.

Assure adequate maintenance for vapor monitoring equipment

      Typically, maintenance of vapor monitoring systems includes
      cleaning, calibration, and operational checks. For manual
      systems, maintenance may consist of recharging the electrical
      component by plugging it into an electrical source or changing the
      batteries, and keeping the device clean.  In addition, some
      systems may require periodic replacement of a filament or a lamp.

      Automatic systems are often self-checking. Self-checking
      systems verify the integrity of the sensor outputs, inputs, power
      supply, alarms, and displays, thus ensuring every aspect of the
      system is operational at all times. If an automatic system is not
      self-checking, periodic calibration and checks of the electrical
120

-------
                   system should be performed. The power source used for
                   automatic systems should be periodically checked to ensure that
                   the supply has not been cut off. The final maintenance item is
                   simply the replacement of paper for the device's recorder.

                   For any type of vapor monitoring system, major maintenance
                   items should be performed by a qualified professional following
                   the manufacturer's instructions.
             Interpretation
I installation [   Consider normal fluctuations and any other circumstances
 Operation
    x
Interpretation of monitoring results is a critical step in vapor
monitoring. Interpretation is hindered by normal vapor level
fluctuations of 100-400 ppm over the course of a week. An alarm
or a high reading does not necessarily indicate a leak; it may be
the result of a spill, or simply be the result of background or
other interference.

When an alarm level is recorded, the operator should first verify
the vapor monitoring system's integrity. This entails determining
that the system is working properly and that it is calibrated
correctly.  If after this initial check the monitoring device is found
to be operating properly, an additional reading should be taken to
verify the results. Should further monitoring confirm the
preliminary results, any potential interfering factors (excessive
rains, a spill from a product delivery, a leak from a nearby tank,
etc.) should be evaluated next.

If no interfering factors are found and monitoring results have
been confirmed, the monitoring alarm can be attributed to a tank
release of some sort. Several owners have reported that an alarm
always followed product deliveries before spill and overfill
equipment was added. When there is not an obvious spill, one
way of differentiating between leaks and spills is by looking at
past monitoring records. Records typically apply only to the
numerical information given by quantitative monitoring devices;
differentiating between a spill and leak is more difficult when
using a qualitative device which simply indicates the presence of
vapor.  Figure 25 show graphically the difference between a leak
and a spill.
                                                                                    121

-------
                 .0
                 I
                 CD
                 O
                 o
                 O
                 *_
                 o
                 O.
                 CO
                                         Time

                                        A Leak
                 c
                 o
                 CD
                 O
                 C
                 o
                 O
                 l_
                 8.
                                         Time

                                        A Spill
                  Figure 25. Interpretation of vapor monitoring results.
                  A leak is normally indicated by a gradual increase in vapor
                  concentration that eventually reaches a high level plateau. Spills,
                  on the other hand, are depicted by a sharp increase in vapor
                  concentration, followed by a gradual decrease. If a manual vapor
                  monitor is being used, readings taken on a daily basis following
                  the recorded high vapor level (also the high point on the graph),
                  should follow one of the trends indicated in the figure.
122

-------
[ Installation]  Place monitoring wells so that leak location is identifiable
    ±
[ Operation
    IT"
I
Generally, vapor monitoring is non-specific; a vapor monitor
cannot locate the source of the vapor to which it is reacting.
Because of this, confirmation of a release location (i.e., the
release is attributed to the tank, piping, or interference) should
always be done before corrective action begins to avoid
unnecessary work.

Specificity of vapor monitoring can be greatly improved by
increasing the number of monitoring wells and using a
quantitative monitoring device. With these improvements, it is
reasonable to assume that the well at which the highest vapor
levels are recorded is also nearest the source of the release. To
increase specificity, monitoring wells should be adjacent to
individual tanks, not between two or more tanks.
APPROACHES TO ENSURING EFFECTIVE VAPOR MONITORING
            Chapter 1 provides a general description of the four types of oversight
            that can be used. The following sections discuss how these approaches
            may be applied specifically to vapor monitoring systems.
             Site Inspections
            The site could be visited prior to installation of vapor monitoring to
            ensure that the method is appropriate for the specific site. Particular
            attention should be paid to the background concentrations found in the
            UST backfill, the type of UST backfill, and the volatility of the product
            stored in the UST. Site inspections could also take place during
            installation to confirm the proper location and installation of monitoring
            wells. Finally, periodic visits to ensure that vapor monitoring
            instruments are properly calibrated would be beneficial.
             Data Review
            Before vapor monitoring is implemented, the implementing agency
            could require that a pre-installation site assessment report and that the
            proposed monitoring well network design be submitted for review.
            Whenever a vapor monitor indicates a suspected release monitoring
            records and their determined interpretation could be submitted for
                                                                                123

-------
           review. Because this may result in numerous submissions, records
           should only be required for releases that require reporting, not simply
           for any vapor monitoring alarm.
            Guidance and Training
            Guidance developed for owners/operators should emphasize proper
            calibration of equipment and the correct manner of interpreting results.
            Network design guidance would be useful for both owners/operators
            and installers.
            Approval and Certification
            One approach an implementing agency can use is to require certification
            of all vapor monitoring system installers, either through agency
            programs or those performed by manufacturers. Some vapor monitoring
            equipment manufacturers already train installers, and the implementing
            agency could review these programs. Another option for implementing
            agencies is to allow installation of only the types of vapor monitoring
            devices that meet specified requirements.  This would involve setting up
            an approval program to which vendors could apply.
REFERENCES
            1.  Eklund, B. and W. Crow. March 1987. Survey of Vendors of
                External Petroleum Leak Monitoring Devices for Use With
                Underground Storage Tanks. Report for J. Jeffrey van Ee,
                Environmental Monitoring Systems Laboratory, U.S. EPA.

            2.  Geonomics, Inc. [not dated]. Soil Vapor Monitoring for Fuel Leak
                Detection.

            3.  Hanselka, Reinhard and Paul M. Allen. September 1985.
                "Aspirated Vapor Sensing for Leak Detection." In: Sensors
                Magazine.

            4.  Kaman Tempo, March 1988. Evaluation of U-Tube Underground
                Tank Systems for Soil Vapor Testing.

            5.  Kaman Tempo. March 1988. Fuel Vapor Background
                Concentration Measurement and Tracer Testing in Underground
                Storage TankBackfill,
124

-------
 6.   Levine Fricke, Inc. February 11,1985. Capabilities of Soil Sentry
     Underground Tank Leak Detection System Under Field Test
     Conditions.  Report for Genelco, Inc.

 7.   Weston, R. F.  August 1986. Underground Storage Tank Leakage
     Prevention, Detection and Correction.

 8.   Weston, R. F.  May 14,1987. Summary of Available Monitoring
     Technologies for Underground Storage Tanks.

 9.   U.S. EPA. [not dated]. Development of Procedures to Assess the
     Performance of External Leak Detection Devices. Report by
     Radian Corporation for Environmental Monitoring Systems
     Laboratory, U.S. EPA.

10.   U.S. EPA. [not dated]. Proposed - Guidance Document for
     External Monitoring of Underground Storage Tanks. Report by
     Dennis Weber and Klaus Stetzenbach for Environmental
     Monitoring Systems Laboratory, U.S. EPA.

11.   U.S. EPA. [not dated]. Soil Gas Sensing for Detection and
     Mapping of Volatile Organics.  Report by Radian Corporation, and
     Environmental Monitoring Systems Laboratory for Environmental
     Research Center, U.S. EPA.

12.   U.S. EPA. March 1987. Soil-Gas Measurement for Detection of
     Subsurface Organic Contamination.  Prepared for Environmental
     Monitoring Systems Laboratory, U.S. EPA.

13.   U.S. EPA. June 1987. Processes Affecting Subsurface Transport
     of Leaking Underground Tank Fluids. Report for Environmental
     Monitoring Systems Laboratory, U.S. EPA.

14.   U.S. EPA. October 20,1987. Underground Fuel Storage in
     Barnstable County, Massachusetts. Report for Region I Office,
     U.S. EPA.
15.  U.S. EPA. February 1988. Interim Report: Summary of Several
    State Underground Storage Tank Regulations. Report for
    Environmental Research Center, U.S. EPA.

16.  U.S. EPA. February 25,1988. Research for Abatement of Leaks
    From Underground Storage Tanks Containing Hazardous
                                                                   125

-------
                Substances. Report by Camp Dresser & McKee, Inc., for Phil
                Durgin, Environmental Monitoring Systems Laboratory, U.S. EPA.

            17.  U.S. EPA. February 29,1988. Background Hydrocarbon Vapor
                Concentration Study for Underground Fuel Storage Tanks. Report
                for Phil Durgin, Environmental Monitoring Systems Laboratory,
                U.S. EPA.

            18.  U.S. EPA. March 24,1988. Standard Practice for Evaluating
                Performance of Under ground Star age Tank External Leak!Release
                Detection Components and Systems. Report by Radian
                Corporation for Environmental Monitoring Systems Laboratory,
                U.S. EPA.

            19.  U.S. EPA. August 1988. Leak Lookout - Using External Leak
                Detectors to Prevent Petroleum Contamination from Underground
                Storage Tanks. Prepared by John D. Kotler for the Office of
                Underground Storage Tanks,  U.S. EPA.

            20.  U.S. EPA. February 1989. Soil Vapor Monitoring for Fuel Tank
                Leak Detection - Data Compiled for Thirteen Case Studies.
                Prepared by On-Site Technologies for Environmental Monitoring
                Systems Laboratory, U.S. EPA.
126

-------
      Chapter VII
Ground-water Monitoring

-------

-------
GROUND-WATER MONITORING
VII
SUMMARY
            The application of ground-water monitoring as an UST release detection
            method involves the use of one or more permanent observation wells
            that are placed close to the tank and are checked periodically for the
            presence of free product on the water table surface. When properly
            designed and installed, ground-water monitoring systems can result in
            effective detection of releases from UST systems.

            Factors that have the greatest impact on the proper operation of a
            ground-water monitoring system are those associated with
            environmental conditions of the site (e.g., depth to ground water, range
            of ground-water table fluctuation), characteristics of the UST system
            (number and size of tanks, type of stored product), and the presence of
            other subsurface  structures. These site-specific characteristics will
            determine the design and complexity of the monitoring well system.

            In general, ground-water monitoring is most effective at sites where the
            water table is within or very near to the excavation zone of the tank. The
            method is also more effective at UST sites where no residual product is
            present in the subsurface materials due to prior releases.

            The discussion presented in this chapter covers a range of problems that
            may occur with ground-water monitoring.  This does not mean that all,
            or even most, of these problems will occur at the same time.
            Furthermore, the problems addressed below are not necessarily of equal
            importance, in terms of the frequency of their occurrence or in the
            severity of their impact on the effectiveness of ground-water
            monitoring.

            Professionals experienced in ground-water monitoring will know how
           to identify and correct these problems.  For example, a qualified firm
           would not install a ground-water monitoring system at a site where the
           water table is 50 feet below the ground surface. Release detection,
           however, is a growing industry  and new companies with little
           experience in ground-water monitoring are opening up.
                                                                               127

-------
BRIEF DESCRIPTION
            A ground-water monitoring system consists of two main components:
            the monitoring well and the free product detection device. Monitoring
            wells are constructed of small-diameter (2-4 inches) well casing
            extending from the ground surface to several feet below the lowest
            water table level.  The portion of the well casing that is perforated (or
            slotted) is referred to as the well screen and extends from the bottom of
            the well to several feet above the water table surface, allowing liquids to
            enter the well. Figure 26 shows the components of a typical
            ground-water monitoring system.

            When a leak occurs from an UST, the released product will migrate
            downward through the backfill material and underlying soil.  When
            liquid-phase (free) products less dense than water encounter the
            ground-water table, they will float and spread out horizontally on the
            surface of the water table.  Monitoring wells properly installed next to a
            leaking tank will intercept the free product layer that accumulates on the
            water table surface.

            The presence of free product can be measured by sensing devices which
            can be either permanently installed in the monitoring well or manually
            inserted in the well to take a discrete measurement. Devices which are
            permanently installed can be operated automatically on a continuous
            basis.

            Although EPA is requiring only that free product be detected when
            ground-water monitoring is used as a release detection method, free
            product monitoring wells also can be used for sampling and analysis of
            dissolved product. Several states have been conducting analysis of
            dissolved product in ground water as a requirement for release detection
            monitoring.

            The process of designing, constructing, and installing a ground-water
            monitoring system can be separated into six phases:

                    1.    Site assessment
                    2.    Selection of a monitoring device
                    3.    Design of the monitoring well network
                    4.    Construction and installation of monitoring wells
                    5.    Operation and maintenance of the monitoring system
                    6.    Interpretation of the monitoring results
 128

-------
                 MONITORING
                  WELL
                  PAVEMENT
                                    BACKFILL
WATER TABLE
  SURFACE
t
                                              STORAGE
                                               TANK
                                      FREE PRODUCT LAYER
                               PRODUCT/WATER CONTACT
                  WELL SCREEN
                                     PERIMETER OF
                                        TANK   	
                                      EXCAVATION
     Figure 26. Monitoring wells installed in the excavation zone
     will quickly detect a release when the ground-water table is
     within the tank excavation.  Source:  U.S. EPA 1987
                                                                129

-------
            Figure 27 shows the relationship of each of these phases and identifies
            the important components of each phase.

POTENTIAL PROBLEMS AND SOLUTIONS
            Successful operation of a ground-water monitoring system depends
            upon a number of factors. In general, operational problems will be due
            either to the monitoring well or to the product monitoring device.
            Operation of a monitoring well can be affected by a variety of site
            factors.  Once identified,  such site problems can sometimes be
            overcome by careful system design.  Problems related to the monitoring
            device typically are easier to overcome than problems related to the
            monitoring well; another  device better suited to the site-specific
            conditions usually can be selected.

            Table 10 is a summary of the problems that may be encountered during
            the installation and operation of a ground-water monitoring system and
            ways to identify these problems. More than one solution may be offered
            to agency personnel for a particular problem and not all of the solutions
            need to be undertaken. The problems identified in Table 10 are
            presented in the order that the method process is implemented; the order
            does not indicate  any prioritization or measure of importance of the
            problems. The most serious concerns are identified in the table by an
            asterisk.
             Site Assessment
            The characteristics of the site environment and of the stored product that
            may affect proper operation of a ground-water monitoring system are
            discussed below.

            Assure that depth to ground water is less than 20 feet

                   Ground-water monitoring is a very effective release detection
                   method for UST sites where the surface of the ground-water
                   table is within the tank excavation zone (see Figure 26). This
                   occurs at approximately 25 percent of UST sites based on results
                   of EPA's national survey data (U.S. EPA 1986a).  Product
                   released from a tank will be quickly intercepted by monitoring
                   wells placed in the backfill of the excavation.
130

-------
                        Site Assessment
                                    LIST system layout
                                    Hydrogeologica! characterization
                                    Other onsite/offsite factors
                                    Exploratory borings
                        Sensor Selection
Installation
                                I
   •  Continuous or intermittent
   •  Permanent or manual
                        Network Design
                                  • Number and location of wells
                                  • Design specifications for wells
                                \ ' • Drilling method
                    Construction & Installation
                                    Borehole drilling
                                    Installation of well, filterpack
                                      and seals
                                    Development of well
                                    Securing the well
                                    Documentation of well installation
Operation
                    Operation & Maintenance
I
Calibration of equipment
Maintenance of well and sensor
                         Interpretation
 Analysis
1
                  Leak
                  No Leak
                                    Differentiating between
                                     interferences and leaks
                                    Locating a leak
      Figure 27. General procedure for ground-water monitoring
                                                               131

-------
to
        Table 10.  Indicators and  Solutions for Problems Encountered With Ground-Water Monitoring
        Problem
Indicators
Tester Solutions
Agency Oversight Options
        Assuring that ground water is
        less than 20 feet or greater
        than 3 feet deep.
        Assuring that hydraulic
        conductivity is greater than
        0.01 cm/sec.
        Assuring that water table
        fluctuation does not exceed
        well screen.
        Responding to steep ground
        water gradient.
        Assuring that stored product
        is soluble in water.
        Assuring that specific gravity
        of stored product is less than
        water.
         Determining presence of
         background contamination.
        Assuring proper selection of a
        sensor.
None.
None.
Well is dry. Water table is at top
of well screen.
 Ground surface slopes steeply.
 Product mixes with water and is
 not observable as a separate
 liquid phase.
 Product is observable as a
 separate liquid phase and will
 not float on water surface.
 Stains are observed on ground
 surface.  Contamination is found
 during performance of site
 assessment

 Device is not sensitive to the
 stored product. Environment
 conditions adversely affect
 sensor operation.
Take water level measurements.
Perform exploratory borings.
Use another release detection
methods.

Inspect boring log. Perform slug
test.  Replace material with
appropriate backfill. Select
another release detection method.

Replace well with one that is
properly screened.  Select
another release detection method.
Place wells downgradient of
tanks.
Obtain data on product solubility
from chemical handbooks or from
manufacturer.  Select another
release detection method.

Obtain data on product solubility
from chemical handbooks or from
manufacturer.  Select another
release detection method.

Cleanup residual product
Select another release detection
method.
 Select another sensing device.
 Select another release detection
 method.
Review water level data. Oversee
 performance of exploratory
 borings.
Review boring logs. Review slug
test data.
Inspect well boring and
construction logs.  Evaluate
long-term water level data
taken in same aquifer.

Review water level data from
a minimum of three wells.
Review topographic map of site.
Conduct site visit to observe
topography.

Review solubility data.  Review
data from chemical/petroleum
supplier.
 Review chemical data on
 specific gravity of product.
 Review site assessment data
 collected by owner/operator.
 Conduct a site visit.
 Review site assessment data
 collected by owner/operator.
 Review information on sensor
 from manufacturer.

-------
       Assuring adequate number of
       monitoring wells present.

       Assuring proper well
       placement.

       Determining if conduits are
       present near the tank field.
       * Assuring adequate
       maintenance of sensors.
Site-specific (see Table 7).
Site-specific.
Product or vapors are observed in
utility trench or basement.
Sensor float is hung up on the
well casing. Occurrence of false
alarms.
       *Assure diameter of well casing   Sensing device will not fit into
       Is not too small.                  the well.
       * Assure well screen slot size is    Flow rate of water into well is
       not too small.                     restricted. Well is dry.
       * Assure well screen slot size is
       not too large.
       * Assure well was properly
       developed.
Filter pack is collecting in the well.
Flow rate of water into well is
restricted. Well is dry.
       'Ensuring proper interpretation    Recording of false alarms.
       of environmental Influences.
       Review site assessment data.
       Check maintenance schedule for sensors.
Install additional wells.
Install additional wells in
proper location (s).

Identify and mark location of
utility lines. Install wells
between tank field and subsurface
conduits.

Visually check sensor on a frequent
basis.
Select another sensing device.
Install a new well of larger
diameter.

Redevelop well to increase flow
rate of water into well. Install
a new well that is properly
designed.

Bail out well to remove
particulate material. Install a
new well that is properly designed.

Redevelop well to increase flow
rate of water into well. Install
a new well if flow rate cannot be
increased.

Check sensor to make sure that
it is operating property. Conduct
a site assessment to determine if
any background contamination is
present.
Review site assessment data.
Conduct a site visit.

Review site assessment data.
Conduct a site visit.

Review site plans and utility
maps.  Conduct a site visit.
Have owner/operator keep a
record of sensor checks. Conduct
a site visit to confirm that sensor
is operating properly.

Review well construction
log.  Conduct a site visit to
observe well design.

Review well construction
log.  Conduct a site visit to
observe well design.
Review well construction
log.  Conduct a site visit to
observe well design.

Review well construction log
to determine the method used
to develop well and the length
of time it was conducted.
U)
       * Indicates the most significant problems.

-------
                   Ground-water monitoring is not an allowable release detection
                   method when the water table surface is greater than 20 feet
                   below the ground surface (BGS). Although monitoring wells
                   installed under this situation do not necessarily present an
                   operational problem, a leak at a site with a deep water table
                   could go undetected for months until the product migrates down
                   to the water table.  Restricting use of ground-water monitoring
                   to sites where the water table is less than 20 feet deep will
                   minimize the potential for widespread environmental
                   contamination and ensure relatively rapid detection of a release.

                   Ground-water monitoring is also not recommended for use at
                   UST sites where the water table is less than 2 to 3 feet BGS.
                   Monitoring wells used for free product detection cannot be
                   properly constructed (with a surface grout seal) when the water
                   table is very shallow. Free product will be excluded from the
                   well screen if the surface seal extends below the water table
                   surface (see Figure 28).  Monitoring wells that are not properly
                   sealed may be susceptible to contamination from surface spills
                   and runoff, which may result in the false reporting of a tank
                   release (see Figure 29).

                   Depth to the water table can be determined at existing
                   monitoring wells by taking water level measurements.  Well
                   boring and completion logs can be inspected to determine where
                   the top and bottom of the well screen are located. An
                   improperly constructed well should not be used for free product
                   detection or for sampling of dissolved product.

                   At new UST installations, the depth to the ground-water table
                   can be determined by taking water level measurements at wells
                   located close to the site. Another alternative is to conduct
                   exploratory borings on-site. If water is encountered at a depth
                   less than 3 feet BGS or is not  observed down to a depth of
                   20 feet BGS, ground-water monitoring would not be an
                   appropriate release detection method.

            Determine hydraulic conductivity of backfill or native soil

                   Monitoring wells may be placed in either the excavation backfill
                   or in the soil surrounding the tank excavation. If the backfill
                   material or the soils situated between the UST system and the
134

-------
  GROUND SURFACE
 IXXXXXXXXXXXXXXXXX>
                                            FREE PRODUCT LAYE
WATER TABLE
  SURFACE
           Figure 28. The well seal will prevent interception of free
           product when the water table surface is very shallow.
           Source: U.S. EPA 1987
                                                                 135

-------
                                                  Spilled Petroleum Product or
                                                  Contaminated Water
                                               Free Product Layer
                                               Product/Water Contact
                Figure 29. Monitoring well that does not have a proper
                surface seal placed above the filter pack will be
                susceptible to contamination from surface runoff.
                Source:  U.S. EPA 1987
136

-------
       monitoring well have a low hydraulic conductivity (less than
       0.01 cm/sec), movement of product released from an UST will
       be restricted. Product released from an UST either will not
       reach the monitoring well or will reach the well only after a long
       period of time. Even though backfill or native soils with a low
       hydraulic conductivity may limit the amount of product that can
       be released from a tank (because the migration rate of the
       product away from the tank is very slow), a slow leak over a
       long time period may result in relatively high concentrations of
       dissolved product in the ground water.

       Hydraulic conductivity is a measure of the rate of flow of a fluid
       in a porous medium and is a property of both the soil and of the
       stored product. In general, the hydraulic conductivity of a soil
       increases with an increase in soil porosity and grain size, and a
       decrease in fluid viscosity. The relative hydraulic conductivities
       of the major soil classes are: gravel > sand > silt > clay (see
       Figure 30). Materials that are  considered to be suitable for
       backfill include clean, graded sands and gravels. Silts and clays
       are much less permeable  than sand and gravel and are generally
       not appropriate for use as backfill.

       Ground-water monitoring is most effective when the hydraulic
       conductivity of the backfill or soils situated between the tank
       and the well is greater than 0.01 cm/sec. The hydraulic
       conductivity of materials can be estimated from information
       provided on boring logs for wells installed onsite or near to the
       site. If materials consist of well-sorted sand or coarser, as is
       required by national installation codes for new tanks, the
       hydraulic conductivity is  most likely greater than 0.01 cm/sec. If
       there is some uncertainty about the hydraulic conductivity of the
       medium, it can be estimated by conducting an in situ test, called
       a slug test,  in the well.

       The hydraulic conductivity of soils and backfill at existing sites
       can be improved only by  replacement with materials having a
       greater permeability. A more cost effective solution may be the
       selection of another release detection method.

Need to evaluate the range of ground-water table fluctuation

       Floating product will not  be detected in a monitoring well if the
       surface of the ground-water table falls below the bottom, or
       extends above the top, of the well screen.  When the ground-
                                                                        137

-------
                    CLAY
                 GLACIAL TILL
                             SILT
                                     SAND
                                                 GRAVEL
            -10      -8      -6      -4      -2               2
          10      10      10      10      10       1      10
                              K (CM/SEC)
                Figure 30. Range of hydraulic conductivities (K) for
                the major soil classes.  Source: Dragun 1987
138

-------
       water level is below the bottom of the well screen, water will not
       be encountered in the well (the well is dry).  When the water
       table surface is higher than the top of the well screen, product
       floating on the ground-water table will not be able to enter the
       monitoring well. Free product would enter the well once the
       water surface falls below the top of the well screen.

       The surface of the ground-water table fluctuates in response to
       seasonal variations in ground-water recharge and other
       influences such as ground water extraction. The location and
       design of a monitoring system must consider the range of water
       table fluctuation at a site (see Figure 31). Wells should be
       screened over the entire interval of ground-water levels. The
       range of fluctuation can be determined based on long-term water
       table level measurements taken in the same aquifer. The high
       and low water  table levels can sometimes be determined from
       soil characteristics  such as color observed during drilling of the
       well borehole.  To help the regulator identify problems with
       water table fluctuations, the owner/operator should obtain and
       record water level measurements  on a monthly basis.

Need to assess the ground-water flow gradient

       Monitoring wells placed upgradient of the tank field may not
       intercept a free product plume at UST sites where the
       ground-water table surface has a relatively steep slope. Product
       released from an UST under these conditions will migrate down
       through the soil to the ground-water table. When product
       reaches the water table it will generally migrate laterally in the
       direction of ground-water flow. Monitoring wells placed either
       upgradient of the UST system or perpendicular to the direction
       of ground-water flow may not intercept a product release under
       these conditions.

       A rough estimate of the direction  and gradient of ground-water
       flow can be made from observation of the ground surface
       gradient. For example, when there is a steep decline of the
       ground surface, ground water will most likely flow in the
       direction of the decline (downgradient).  Ground water will also
       tend to flow towards surface waters, which are discharge points
       for ground water. The ground surface gradient can be
       determined from onsite observations and from information
       presented on topographic maps, which are available from the
       state or federal geological survey.
                                                                       139

-------
       FREE PRODUCT LAYER
                                                    GROUND SURFACE
                                                 RANGE OF WATER TABLE
                                                    FLUCTUATION
                Figure 31. The well screen is placed to extend over the
                entire range of water table fluctuation.
                Source:  U.S. EPA 1987
140

-------
        The installation of monitoring wells on all four sides of the tank
        field is usually adequate to ensure detection of a release at sites
        located on a relatively steep gradient or if there is some
        uncertainty about the ground-water flow gradient.

 Stored product must not be soluble in water

        Products that are highly soluble in water will not be detectable
        as a separate liquid phase.  When a soluble product is released
        from a tank and reaches the ground-water table, it will mix with
        the water and disperse throughout the aquifer. Therefore,
        ground-water monitoring for free product detection cannot be
        used when the stored product is highly soluble in water. Most
       petroleum products are not highly soluble in water. Though
        gasoline and diesel fuel are composed of many individual
        chemicals that are soluble to some extent in water, these
       products will be observable as a separate liquid layer.

       The type of product stored in each tank should be recorded on
       the UST notification  form.  This form does not require reporting
       information on the solubility of each stored product, but this
       information can be found in standard chemical handbooks. The
       manufacturer of the product also should be able to provide this
       type of data.

Specific gravity of the stored product must be less than that for water

       Products that are denser (heavier) than water (i.e., specific
       gravity is greater than 1.0) will not float on the water surface.
       Instead, these products will migrate down through the
       unsaturated and saturated (aquifer) zones until an impermeable
       zone is encountered.  The free product will accumulate at the
       interface of an impermeable zone, such as a clay layer below the
       aquifer.

       Products that are denser than water include the halogenated
       hydrocarbons (e.g., trichloroethene, dichloroethane) and coal tar.
       Ground-water monitoring is not an effective release detection
       method under these conditions since no product will enter the
       well. Furthermore, the presence of these types of products would
       be difficult to detect with most of the currently available free
       product sensors.  As with product solubility, information on
                                                                       141

-------
                  density can be obtained from either chemical handbooks or from
                  the manufacturer of the product.

           Need to determine the presence of background contamination onsite

                  Ground-water monitoring is most effective for use as a release
                  detection method when there is not any background
                  contamination present due to prior spills or releases. Residual
                  product present in soils above the water table (unsaturated) can
                  reach ground water by vertical migration due to gravity forces
                   and from infiltration of rainfall and surface runoff.  Also, if the
                  water table rises up to an area where residual product is located,
                   free product may accumulate on the water table surface. The
                   observation of free product in a monitoring well due to residual
                  product will result in a false indication of a leak.

                   The site owner/operator should be interviewed to determine if
                   any prior releases or large spills have occurred at the site or near
                   to the site. Confirmation of background contamination can be
                   made by conducting a preliminary site assessment involving
                   investigation of the subsurface using field analysis techniques
                   (e.g., soil vapor surveys). If residual free product is discovered,
                   it should be removed prior to initiating ground-water
                   monitoring.
             Sensor Selection
            Assure selection of appropriate sensor

                   Selection of an improper monitoring device may result in either
                   a release not being detected or going undetected for a long time.
                   For example, a monitoring device that is not sensitive to the
                   stored product will not detect a release, regardless of the
                   thickness of the product layer. A particular device may be
                   adversely affected by some environmental conditions such as
                   very low temperatures. The problem can be resolved by
                   selecting a device that is appropriate for the site-specific
                   conditions. Currently, however, free product sensing devices are
                   not available for all types of stored products. Furthermore, as
                   discussed in the previous section of this chapter, "Site
                   Assessment," ground-water monitoring is not an appropriate
                   release detection method for all stored products.
142

-------
A variety of devices are available for measurement of free
product. Devices may be permanently installed in the well for
automatic, continuous measurements of product thickness.
Manual devices range from grab samplers used for collection of
a liquid sample for visual inspection of free product, to devices
that can be inserted into the well to electronically indicate the
presence of free product.  None of these devices measure the
leak rate  from an UST system, however; they only indicate the
presence of free product.

Any of the automatic sensors can be permanently installed in a
monitoring well to take continuous measurements.

• Differential float devices operate using a system of two floats;
  one float reacts only to liquids with a density similar to water,
  and the other float responds to liquids lighter than water (most
  hydrocarbons).

• Product soluble devices are coated with or constructed from a
  material that degrades when exposed to hydrocarbons
  resulting in a change of pressure (i.e., an air leak) or a change
  in resistance (for electrical resistivity devices).

• Thermal conductivity devices measure heat loss when the
  floating probe comes into contact with a non-polar liquid.
  Thermal conductivity is the most commonly used continuous
  device.

Some of the manual devices that can be used include the
following:

• Grab sampling devices, such as a bailer or bucket, obtain a
  ground-water sample from the well which is visually
  inspected for the presence of free product (i.e., a sheen on the
  surface of the water) or electronically analyzed on-site.

• Chemical sensitive pastes are attached to a weighted tape
  measure that is lowered into the well and which react (by a
  change in color) when hydrocarbons are present.

• Interface probes detect polar versus nonpolar substances using
  properties of thermal conductance and reflection/refraction of
  infrared light.
                                                                 143

-------
             Network Design
            The monitoring well network is designed after an initial assessment of
            the tank field layout and local hydrogeological conditions. This phase of
            the process involves determining the number and location of monitoring
            wells and their approximate depth. The well network design must be
            based on site-specific conditions and should be developed by a qualified
            professional. For common system configurations, however, a generic
            design recommended by the manufacturer is probably sufficient.
            Problems which may be encountered due to improper well network
            design are addressed below.

            Ensure an adequate number of monitoring wells

                   If the number of monitoring wells at  an UST site is inadequate, a
                   release from a tank may not be intercepted and, thus, may not be
                   detected.  For example, if a monitoring well cannot be placed in
                   the backfill and is situated in soils which are highly fractured, a
                   release from the tank may not be intercepted by the well (see
                   Figure 32).

                   In general, the use of a single monitoring well may be adequate
                   for UST sites with only one tank. However, one well may not
                   always be reliable in detecting free product.  If the site consists
                   of multiple tanks, more than one monitoring well should be
                   provided. The exact number of wells should be based on the site
                   hydrogeology and the UST system configuration. A general rule
                   of thumb recommended by professionals is the placement of at
                   least one monitoring well on each side or comer of the tank
                   field.

                   A number of states and localities currently have specific
                   requirements for design of a ground-water monitoring well
                   network. Table  11 summarizes the network design criteria
                   included in several regulations. Although the regulations each
                   have a different approach, most require that single tanks have
                   one to two wells and that typical multiple tank systems have
                   three to four wells (or more for a larger tank field).

                   Identification of this problem is difficult and requires knowledge
                   of the site hydrogeological characteristics.  To assist the
                   regulator in evaluating this type of problem, the owner/ operator
144

-------
                                                           Free Product Layer
                                               Free Product
                                               Layer
Product-Filled
Fractures
        V
             Water-Filled
             Fractures
                                                                Water-Filled
                                                                  Cavities
                         Note: No Free Product in Well
Fractured Rock
                                                     Karstic Limestone
            Figure 32. Free product will preferentially flow
            through fractures and cavitites; wells that do not
            intercept these structures will not detect product.
            Source: U.S. EPA 1987
                                                                       145

-------
                                      Table 11

    Typical Network Design Requirements for Ground-Water Monitoring
Fremont, California
Torrance, California
Iowa
Maine
Nebraska
Delaware
Vernon, California

Florida

Maryland

South Carolina

Wisconsin
Single tank:  1 downgradient well within 10 feet of the excavation
perimeter

Multiple tanks:  1 well placed every 35 feet on the longest dimen-
sion of the excavation with a minimum of two wells

Single tank:  1 downgradient well. If the ground-water gradient is
not known, 2 wells on opposite sides of the tank

Multiple tanks:  To be evaluated

Single tank:  1 well at each longitudinal end of the tank

Multiple tanks:  4 wells placed on each side of the tank field
Wells must be placed within 1 to 20 feet of the nearest tank

Ground  water < 15 feet: No fewer than 2 wells at either end
of the tank

Ground water > 15 feet: No fewer than 4 wells for each tank or for
multiple tanks located in the same excavation, one well at both
ends of each tank and at each end of the excavation

Ground water > 15 feet: 1 well per tank

Ground water < 15 feet: 2 wells per tank

New installations: 4 wells placed around tank excavation fields

Existing installations: 3 wells placed in the excavation around
the tank(s)

Wells downgradient of tank(s) being monitored

4 wells placed in the excavation around the tank(s)

2 wells placed at opposite corners of the tank field

Minimum of 2 wells placed every 30  feet

3 wells required only for new UST installation
 Source: Reference 11.

 146

-------
       could be required to provide a site plan of the UST facility
       showing the layout of the UST system (number and size of
       tanks) and the locations of existing monitoring wells. Other
       useful information includes the ground-water flow gradient and
       geological characteristics of the site.

Assure proper placement of wells relative to the tank field

       Improper placement of monitoring wells may result in delayed
       or missed detection of a product release. For example,
       monitoring wells placed too far away from the tank field may
       not detect a release until a large volume of product has leaked
       from the tank. Monitoring wells should be placed as close to the
       tank excavation as possible. Care needs to be taken when
       selecting a monitoring well site to ensure that the installation of
       the well will not interfere with any subsurface structures such as
       utilities or UST system piping. In some cases, the installation of
       a well upgradient of the UST system might protect owners from
       false alarms at sites where there are other (offsite) petroleum
       UST systems located near the site being monitored (see
       Figure 33).

Need to check for subsurface conduits near the tank field

       Buried fill material or subsurface utility conduits (e.g., trenches
       constructed for telephone, power, gas, sewer, and water lines)
       may act as preferred pathways for free product migration.  Free
       product generally will flow more easily through these open
       subsurface structures than through surrounding soils because the
       material is more permeable and offers less resistance to flow.
       Therefore, free product may not reach the ground-water table
       and will be channeled away from monitoring wells (see
       Figure 34).

       This problem may be indicated when product or odors are
       observed in utility conduits but floating product is not observed
       in the monitoring well. At UST sites where there are subsurface
       conduits present, monitoring wells should be placed between the
       tank field and the conduit to ensure that product released from
       the tank is intercepted by the well before it reaches the conduit.

       The locations of buried conduits can be found by examining site
       construction maps and maps of all utilities present in the area.
       The utility companies should be notified, prior to conducting
                                                                       147

-------
CO
             SERVICE
             STATION
             NO. 2
                                               MONrrORING
                                                  WELL
                                           DISSOLVED AND
                                           FREE PRODUCT
                                              PLUME
                                           DIRECTION OF
                                          GROUND-WATER
                                             FLOW
                    Lack of upgradient well implicates Station No. 1
                                                                         TANK
BUILDING
                               Figure 33. Off-site sources of contamination should be
                               considered when designing the monitoring well
                               network. Source: U.S. EPA 1987

-------
   DIRECTION OF
GROUND-WATER FLOW
                           PRODUCT IN
                          UTILITY TRENCH
                                                MONITORING
                                                   WELL
X
                                                           TANK
                      SERVICE
                      STATION
                                                             BUILDING
             Figure 34.  Subsurface utility conduits will act as a
             preferential pathways for free product migration.
             Source: U.S. EPA 1987

-------
                    well installation, that drilling will be performed at the site. To
                    help troubleshoot any operational problems that may occur over
                    the life of the monitoring system, the location of all subsurface
                    conduits should be indicated accurately on a current site plan.
             Construction and Installation
             This section addresses problems that may occur as a result of poor or
             improper design, construction, or installation of a monitoring well.  It is
             important to keep in mind that the design and construction requirements
             for a monitoring well used for free product detection are not the same as
             for a drinking water well or for a monitoring well constructed solely for
             sampling and analysis of dissolved product.  The proper design and
             construction of a monitoring well is crucial to effective detection of free
             product and these tasks should be performed by an experienced
             hydrogeologist.

             Assure proper design of the monitoring well

                    Improper design of a monitoring well may result in a release
                    either not being detected at all or going undetected for months,
                    resulting in widespread environmental contamination. Figure 35
                    shows the construction of a typical monitoring well used for free
                    product detection. Design factors that are considered to be the
                    most important to proper operation of a monitoring well are
                    addressed below.

                    Before beginning design of a monitoring well,  any specific state
                    or local construction requirements should be identified. Some
                    states, such as Florida  and California, have developed specific
                    criteria that must be met for length of well screen, screen slot
                    size, filter pack specifications, etc.

                    Well casing and screen material: Most materials used for
                    ground-water monitoring wells are compatible with petroleum
                    hydrocarbon products. However, some materials, such as steel,
                    may deteriorate in highly corrosive environments. This could
                    result in the collapse of a well casing or infiltration of paniculate
                    material through holes created by corrosion.

                    A variety of materials are available for construction of the well
                    casing and screen, including fluorocarbon resins, cast iron,
150

-------
                                      GROUND SURFACE
      WELL CAP
     VENT HOLE
  BENTONITE OR
  CEMENT GROUT
WATER TABLE
  WELL SCREEN
  BOTTOM
  CASING PLUG
        \
PROTECTIVE STEEL

CASING
                                              ANNULAR SEAL
                                               FILTER PACK
              Figure 35. Components of a typical ground-water
              monitoring well installed in a borehole.
              Source:  U.S. EPA 1987
                                                              151

-------
                   galvanized steel, polyethylene, polypropylene, Teflon , stainless
                   steel, and polyvinyl chloride (PVC). The selection of an
                   appropriate material should be based on environmental
                   conditions, structural requirements, and compatibility with the
                   contaminants of interest.

                   Typically, the material of choice for monitoring well
                   construction is PVC or stainless steel.  Both of these materials
                   meet the structural needs of a monitoring well. Some types of
                   PVC will deteriorate when in continual contact with petroleum
                   hydrocarbons. However, the use of an impervious PVC material
                   is acceptable and is common for construction of monitoring
                   wells. The type of material used for construction of the well
                   screen and casing should be indicated on a well completion log
                   that is kept onsite for inspection by the regulator.

                   Well diameter: The inner diameter of the well casing and
                   screen typically ranges from 2 to 4 inches. Though smaller
                   diameter casing is available, diameters less than 2 inches are not
                   recommended. A monitoring well that is constructed of very
                   small diameter casing could limit the type of hydrocarbon
                   monitor selected for use.  Monitoring devices are also more
                   likely to get hung up on small diameter casing. Monitoring
                   wells larger than 4 inches in diameter can be installed; larger
                   diameter wells could later be used as extraction wells if
                   remediation of ground water is required in the future due to a
                   release incident.

                   The diameter of well casing and screen used should be
                   documented on a well completion log.  Regulators can refer to
                   this log to determine the design of the monitoring well.

                   Screen slot size: The size of the well screen slots is very
                   important to the proper operation of a monitoring well. If the
                   slot size is too large, soil particles will be allowed to pass into
                   the monitoring well, which may eventually become filled with
                   soil. If the slot size is too small, it may prevent the flow of
                   product into the monitoring well.

                   The slot size should be wide enough to permit the flow of
                   ground water and free product into the well, but not so wide as
                   to allow the passage of filter pack or fine-grained soils into the
                   well. Therefore, the slot size should be determined based on the
                   type and texture of surrounding soils and should  be  selected by
152

-------
       an experienced hydrogeologist. Premanufactured well screens
       are available in slot sizes typically ranging from 0.008 to 0.120
       inches.

       This problem can be evaluated by requiring the slot size of the
       well screen to be indicated on the well completion log.  If the
       slot size of an existing well is inappropriate and causing an
       operational problem, the well should be properly closed and
       another well installed with the correct slot size.

       Length of well screen:  The total length of the well screen will
       depend on the depth to the water table and the range of water
       table fluctuation levels.  The screen must extend over the entire
       interval of the high and low water table levels as discussed in the
       section of this chapter titled "Site Assessment."

Ensure proper installation of the monitoring well

       The process of well installation involves borehole drilling and
       placement of well in borehole; installation of filter pack,
       bentonite seal, surface seal, and protective casing; well
       development; well security; and documentation of well
       construction and installation.

       Borehole drilling generally does not need to be conducted for
       monitoring wells  installed in the excavation field of new UST
       facilities. For these sites the well may be located during tank
       installation and the backfill placed around it. Operational
       problems associated with borehole drilling for ground-water
       monitoring wells  installed outside of the excavation zone
       (typically for well installation at existing UST facilities) will not
       be discussed in this document. Though selection of an
       inappropriate drilling method could affect the hydraulic
       conductivity of the subsurface materials, this situation is not a
       common problem. A discussion on the types of available
       drilling methods is provided in U.S. EPA 1986.

       The use of an inappropriate filter pack may allow the
       introduction of sediment into the well, eventually causing it to
       become filled or restricting the flow of fluid into the monitoring
       well.  Typically, several inches of filter pack are placed in the
       bottom of the borehole before the well is installed. The filter
                                                                         153

-------
                    pack, which may consist of clean sand, silica, or glass beads, is
                    placed in the borehole to extend 1 to 2 feet above the top of the
                    well, screen.

                    The annular space (the space between the wall of the borehole
                    and the outer well casing) above the filter pack is then sealed to
                    prevent migration of contaminants to the well screen. Materials
                    such as a cement-bentonite mixture, bentonite chips, or
                    antishrink cement mixtures can be used as the sealant in
                    unsaturated zones. This seal should extend at least 2 feet above
                    the filter pack.

                    The area above the annular seal should be filled to the surface of
                    the ground with a cement-bentonite seal to prevent migration of
                    liquids from the ground surface and to protect the well casing
                    from structural damage.  Ideally, the interface of the annular seal
                    and the seal above it should be located below the frost line to
                    protect the well from damage due to frost heaving.  A well
                    casing (referred to as a protective casing) with a diameter larger
                    than that of the monitoring well (4 to 8 inches) can be placed in
                    the cement seal around the monitoring well to provide additional
                    protection from physical damage. This protective outer casing is
                    typically made of black steel or PVC.

                    Monitoring wells installed in the backfill of new UST sites can
                    be constructed using the backfill as the filter pack if the material
                    meets the requirements of an appropriate filter pack (see
                    Figure 36).  These wells still should be constructed with a
                    surface grout to prevent infiltration of contaminated runoff from
                    migrating vertically down to the filter pack and to help stabilize
                    the well casing and screen.

                    Flow of liquids into the well may be prevented or severely
                    reduced if a monitoring well is not properly developed. The
                    process of well development involves the surging (mixing) and
                    removal of ground water from the well casing. The purpose of
                    this procedure is to dislodge material trapped in the well screen,
                    to remove sediment in the well introduced during the installation
                    process, and to restore the natural flow rate of ground water
                    (hydraulic conductivity).
154

-------
CEMENT PAD
                                                                               BACKFILL
                                                                               MATERIAL
                    Figure 36. Components of a monitoring well
                    constructed with the backfill material used as the filter
                    pack. Source: U.S. EPA 1987

-------
                    A variety of techniques can be used for well development. The
                    well can be surged using surge blocks, bailers, or pumps. Well
                    development is continued until ground water in the well is
                    relatively free of turbidity (cloudiness).

                    If well development is not continued for an adequate length of
                    time, the result will be a reduction in the flow of fluid into the
                    well. This problem can be evaluated by inspecting the well
                    completion log to determine how long the well was developed
                    during installation. The problem can sometimes be corrected by
                    redeveloping the well until the recharge rate of the well is
                    improved.

             Need to secure the monitoring well

                    A monitoring well that has not been properly secured is
                    susceptible to tampering or accidental contamination. For
                    example, product could accidentally be delivered to the
                    monitoring well instead of the tank fill pipe. To secure a
                    monitoring well, a threaded or flanged cap is placed at the top of
                    the well to prevent the introduction of any foreign matter into
                    the well. The well cap should be locked in place to prevent
                    tampering. The well should be unlocked only when entrance to
                    the well is requked. The equipment used to secure a well should
                    be documented on the well completion log. Monitoring wells
                    should also be clearly marked (e.g., by color coding or labelling)
                    to distinguish them from tank fill pipes. The security  of a well
                    can be easily checked by inspection of the well head.

             Need to document well construction and installation

                    The first step that should be taken when a problem with
                    operation of a monitoring well occurs is to  obtain a copy of the
                    boring log and well construction diagram for each well. It is
                    much more difficult to troubleshoot operational problems if
                    these documents are not provided. The boring log describes the
                    soil types and texture of different geologic strata encountered
                    and thek interval depth, drilling method(s)  used, depth to ground
                    water, etc.  The well construction diagram depicts the well
                    design specifications, including the type and depth of filter pack,
                    annular seal, cement seal, well diameter, and well screen design.
                    This information should be documented on the well boring and
                    completion logs for each monitoring well.
156

-------
 Operation and Maintenance
This section describes the most common operator-induced problems that
will affect proper performance of ground-water monitoring. The most
common types of human errors are described below.

Need for proper maintenance of sensors

       Ground-water sensor devices that are improperly maintained can
       falsely indicate that a release has occurred or fail to detect a
       release.  Inadequate maintenance of sensor devices can result in
       a build-up of ice or algae, which is a common problem with
       continuous monitoring devices. Thermal conductivity devices
       can get caught on the float.  Ice also is a maintenance problem
       with product-soluble devices. Maintenance generally is not
       considered a problem with intermittent devices. To prevent
       maintenance problems with monitoring devices, inspections of
    •  the sensor should be conducted on a regular and frequent basis
       to ensure that it is operating properly.

       Calibration of manual ground-water monitoring devices is
       typically not necessary. Automatic devices should be calibrated
       if this is recommended by the manufacturer, according to their
       specifications.

Ensure the integrity of sensor coatings

       Electrical resistivity sensors and hydrocarbon-soluble devices
       use hydrocarbon sensitive coatings that degrade when exposed
       to hydrocarbon products. These coatings may also biodegrade
       over long periods of time. After the sensor has been exposed to
       hydrocarbon product, it must be replaced. If the sensor is not
       replaced, it will fail to detect a future release.
                                                                      157

-------
             Interpretation
            Ensure proper interpretation of environmental influences

                   False alarms in continuous monitoring systems are often caused
                   by environmental influences.  The most common error is failing
                   to identify changes in electrical resistance that are due to
                   equipment shorts or power surges.  This is primarily a problem
                   that affects thennal conductivity devices and can result in a false
                   negative measurement.

            Ensure proper interpretation of monitoring results

                   Another potential source of errors is the failure of the operator to
                   determine that the source of an alarm or positive result is from
                   an offsite source of contamination or from the accidental
                   introduction of product into the monitoring well (e.g., surface
                   spills).  False alarms can also occur when the local water table
                   rises and contacts residual product from previous spills (see the
                   discussion in the section of this chapter titled "Site Assessment,"
                   addressing background contamination). This is more of a
                   problem with intermittent devices because these devices rely on
                   the results obtained from a discrete sample and not on data
                   trends.

                   Malfunction of monitoring devices, in particular thermal
                   conductivity devices, may result in false positives. Another
                   potential source of errors is the long (up to 10 hours) response
                   times exhibited by electrical resistivity sensors. Although this is
                   not documented as a common error, an operator unaware of this
                   time constraint may prematurely decide that a resistivity test
                   indicates no leak. The response and lag times for all other
                   continuously operating devices mentioned in this manual are less
                   than 30 minutes.

                   The interpretation of monitoring results from ground-water
                   monitoring wells does  not require a high degree of technical
                   expertise. Manual systems require inspection of a liquid sample
                   obtained from the well, or the use of a sensor which will either
                   electronically indicate the presence of free product (i.e., a
                   change in thermal conductance) or will change color.
                   Automated detection systems require the greatest degree of
158

-------
                  interpretation. For systems with an alarm, the operator must
                  determine if a triggered alarm is legitimate or a false alarm.
                  Systems using a continuously operating strip chart recorder may
                  require interpretation of the data trends indicated on the chart.
REFERENCES
            1. American Petroleum Institute. 1986. Observation Wells as
               Release Monitoring Techniques. Prepared by Weston for API
               Marketing Operations and Engineering Water Subcommittee.

            2. Dragun, J. 1988. The Soil Chemistry of Hazardous Materials.
               Hazardous Materials Control Research Institute, Silver Spring,
               Maryland.

            3. Electric Power Research Institute.  1987.  Summary of Available
               Monitoring Technologies for Underground Storage Tanks -
               Draft. Prepared by Roy F. Weston, Inc.

            4.  U.S. Department of Energy. 1985.  Procedures for the
               Collection and Preservation of Ground Water and Surface
               Water Samples and for the Installation of Monitoring Wells,
               2nd Edition. Prepared by Bendix Field Engineering Corporation
               for the U.S. Department of Energy, Nuclear Energy Programs,
               Division of Remedial Action Projects.

            5. U.S. EPA. 1985. September 1985.  Practical Guide for
               Ground-Water Sampling. Prepared by the Illinois State Water
               Survey for the Robert S. Kerr Environmental Laboratory.
               EPA/600/2-85/104.

            6. U.S. EPA. 1986.  Septmeber 1986. RCRA Ground-Water
               Monitoring TEGD. Office of Solid Waste and Emergency
               Response, OSWER-9950.1.

            7. U.S. EPA. 1986a. May 1,1986.  Underground Motor Fuel
               Storage Tanks: A National Survey. Vol.  1, Technical Report.
               Prepared for the Office of Pesticides and Toxic Substances. May
               1, 1986. EPA-560/5-86-013.
                                                                                159

-------
            8. U.S. EPA. 1988.  February  1988. Free-Product Release
               Detection for Underground Storage Tank Systems, Vol. 1.
               Capabilities and Limitations of Wells for Detecting and
               Monitoring Product Releases. Prepared by Geraghty and Miller,
               Inc. for the U.S. EPA Office of Underground Storage Tanks.

            9. U.S. EPA. 1988a. February 1988. Free-Product Release
               Detection for Underground Storage Tank Systems, Vol. 2.  The
               Effectiveness of Petroleum Tank Release Detection Monitoring
               with Wells in Florida. Prepared by Geraghty and Miller, Inc. for
               the U.S. EPA Office of Underground Storage Tanks.

            10.  U.S. EPA. 1988b. July 1988. Common Human Errors in
               Release Detection Usage. Prepared by Camp Dresser & McKee
               Inc. for the Office of Underground Storage Tanks.

            11.  U.S. EPA. 1988c. February 1980. Interim Report: Summary of
               Several State Underground Storage Regulations.  Prepared by
               the Environmental Research Center, University of Nevada-Las
               Vegas for the Environmental Monitoring Systems Laboratory.
160

-------
       Chapter VIII
  Secondary Containment
with Interstitial Monitoring

-------

-------
SECONDARY  CONTAINMENT
WITH INTERSTITIAL MONITORING
VIII
 SUMMARY
            The use of secondary containment with interstitial monitoring as an
            UST release detection method for petroleum storage tanks involves a
            barrier outside the primary tank with a release detection device between
            the inner and outer barriers. The space between the barriers is called the
            interstitial space. The outer wall or liner contains the leak long enough
            for it to be detected by the monitoring system.  This method is required
            in several states and is considered to be the most protective of the
            environment because leaks are generally detected before they can
            contaminate the environment.

            The factor that has the greatest impact on the proper functioning of
            secondary containment is installation.  The factors that have the greatest
            impact on the interstitial monitoring systems are system installation and
            operation and maintenance factors.

            The discussion presented in this chapter covers a range of possible
            problems that may occur with secondary containment with interstitial
            monitoring. This does not mean that all, or even most, of these
            problems will occur at the same time.  Nor does it mean that all of the
            problems are of equal importance, in terms of frequency of occurrence
            or severity of impact. Some problems, such as false alarms caused by
            curing of fiberglass tanks, seldom occur, while other problems,  such as
            residual contamination, may have limited impact. Experienced
            installers are well aware of these problems and how to deal with them.
            For example, an experienced installer would use waterproof electrical
            junction boxes to prevent detection system failure during precipitation.
            Release detection, however, is a growing industry, and new companies
            are being formed with less experience. This chapter presents a range of
            potential problems for educational purposes, not to imply that they will
            always occur.
                                                                              161

-------
BRIEF DESCRIPTION
             An interstitial monitoring system is intended to detect any leak from the
             tank under normal operating conditions and not to measure the leak rate.
             Secondary containment with interstitial monitoring for tanks consists of
             two components. The first is an outer barrier, which directs the leak
             toward the interstitial monitoring system. The purpose of the outer
             barrier is to retain the leak for a sufficient time so that it can be detected.
             The barrier is not requked to contain the leak from a petroleum tank so
             that it does not contaminate the environment (as is requked for
             hazardous substance tanks). The second component is the interstitial
             monitoring system, which detects the leak and alerts the operator.

             The outer barrier can be either an outer shell (on a double-walled tank),
             a synthetic liner around the tank (a tank jacket), or a liner in the
             excavation of the tank system that is between the soil and the backfill
             material (Figures 37 through 39). The outer shell of a double-walled
             tank is generally made of the same materials as the tank, e.g., steel or
             fiberglass reinforced plastic (FRP), while a liner can be made of various
             synthetic materials, such as high-density polyethylene, polyester
             elastomers, epichlorohydrin, and polyurethane.

             The outer barrier either completely surrounds the tank (the fully or
             completely enclosed design) or covers only the bottom half of the tank
             (partially enclosed design). In the completely enclosed (double-walled
             or jacketed) design, a leak from any part of the tank is trapped  in the
             interstitial space between the inner and outer barriers and is detected. In
             the partially enclosed design, leaks from the bottom half of the tank
             would be contained for detection by the outer shell, but leaks originating
             above the outer shell could potentially enter the backfill and avoid
             detection by an interstitial monitoring system.

             Excavation liner systems also may  be either fully or partially enclosed.
             A fully enclosed liner system would have a liner section across the top
             of the tank that would be sealed to the sides of the excavation liner. A
             partially enclosed design might not have the top liner, and its excavation
             liner might not reach the top of the tank.

             Concrete vaults are also used for secondary containment but are not
             commonly used for petroleum products because vaults are more costly
             to construct. Other references are available on concrete vaults, which
             are not discussed further in this chapter.
162

-------
                                              /o\
                             Inner Tank Wall
                                                                  • Outer Wall


                                                                   Inner Wall
Sampling
Standpipe

   or

Electronic
Detection
"Double-Walled Steel Tank
                                         Interstitial Space
                       •Double-Walled FRP Tank
 A cross section of a double-walled tank is shown in Figure 40
          Figure 37.  Two double-walled tank configurations.
          Source: U.S. EPA (January 1989)
                                                                         163

-------
                                                                                Extrusion Welded Seams
    Steel Tank
        \	/?Jips
f -'^  '^V/>'*'  *   :  x$p$s§8

  £<*'': -:f -  ,>1:^^^S
                                      Manway
           J

            x
                   ' >- /•<•>."W/w-V-A
                      j\«/..s.%.\.%.%.%.%,

                      fi's'-:ft'%Ws^s:
       .   '  '       "• !.%«"••*• •%«%"%«%«*.Bli»^i
              '  v   K>;V/.->; w-v/xx.

    "•            x   V^V-""j« i   •• ••S»%1
                       ^w-ftw^ft
                       ^?iSS5>^^
'tttiftfif^fi'f^fififffififf
"%"%"v.%"%V%V%V'.V".V%V%VO'
•'f'fifififififSfififJfif,
  •••i^i^V-^v-W^'^''
  ^fff'ftffflf'f'f-f'f'i
• . -£•?•?••*•?••?••*•• '•• '•'"••*•
f'f;f'f'f'fiJ''m''f'f'f'f
i«<..%.%.S.%."..%.-..%.S"%'

^ftft&ftftftftf^-"

:$ft$ftftftftftftf«

SW#W-£££ftxv
. Vtftftftftftftftf-
'fiflfiftfiftfifffr
•i^V'V'V-i^V--'-'1- •
f**'if'f;-''f'J':m'ff','':
,.•..%.%.%.%.%.«..%.• -
'f*f*f*f*fmjt*fr

##;£&;&»
•^V-^W^v  '
                                                                                                    Premolded

                                                                                                    End Cap
                                s-.tftftfW•%•*.*
                                 ,WW"


                                 ?•%"
            Interstitial Space
                                                 Jacket
Figure 38. Jacketed tank. Source:  U.S. EPA (August 23,1986)

-------
Monitoring Point Suitable
for Leak Detection and
Withdrawal of Accumulated
Water
                                                                                               Collar to Connect
                                                                                               Pipe Trench to
                                                                                               Tank Liner
       Pavement      Liner Turnback
                                            Trench Top Liner
Backfill
Tank
Excavation
Liner
Some Partially
Enclosed Liners
Would End Here
                                                       Pipe Leak
                                                       Detection
                                                       Monitoring
                                                       Point
                                 Slotted Pipe for
                                 Leak Detection and
                                 Withdrawal of Water
                                 Native Soil
                                                                  Interstitial. Space
              Figure 39.  Tank with excavation liner.  Source: U.S. EPA (August 22,1986)

-------
             Interstitial monitoring systems operate to detect leaks based on the
             following mechanisms:

             •   Electrical conductivity
             •   Pressure sensing
             •   Fluid sensing
             •   Hydrostatic monitoring
             •   Manual detection
             •   Vapor monitoring

             The applicability of the monitors to the types of secondary containment
             systems is summarized in Table 12.

             Vapor monitoring is discussed in Chapter 6 of this handbook.
             Therefore, it will not be discussed in this chapter except where the
             problems are specific to interstitial monitoring systems.

             The electrical conductivity monitor depends upon the leaked product
             changing the resistance of sensing wires that are placed in the interstitial
             space of tank systems. The leaked product completes an electrical
             circuit and allows current to flow, thus activating an alarm.  Designs are
             available for both electrically conductive (e.g., water or a water-based
             detergent solution) and for non-conductive products (e.g., like
             petroleum).  This type of system is used in secondary containment
             systems that utilize an excavation liner, a jacket, or a double-walled
             tank.

             Pressure sensing systems are used only in the interstitial space of
             double-walled tanks.  The space is either put under a vacuum or
             pressurized, and leaks are identified by the detection of pressure
             changes that occur when either the inner or outer tank shell develops a
             hole or crack.

             Fluid sensing systems are used in the interstitial space of double-walled
             tanks to detect a leak into the normally dry space of either the product in
             the tank or of ground water. One system uses an optical principle in
             which the change in the reflectance of a mirror is detected when the
             product or ground water leaks into the space and covers the mirror.

             The hydrostatic method (Figure 40) is used in double-walled tanks and
             is based on detection of the change in level of a fluid that completely
             fills the interstitial space.  When a breech in the inner wall occurs, the
166

-------
                                     Table 12
      Applicability of Leak Detection Methods to Secondary Containment Systems
Containment
Full double-
walled tank
Tank jacket
Partial double-
walled tank
Fully enclosed
excavation liner
Partially enclosed
excavation liner
Electrical Pressure
conductivity sensing

X X
X

X X

X

X
Fluid Hydrostatic
sensing monitoring

X X
X

X

X

X
Manual Vapor
methods monitoring

X X
X X

X X

X X

X X
Source: U.S. EPA.
                                                                    167

-------
o\
CO
      Concrete
      Traffic Pad
.Optional Reservoir
Liquid Level Sensor
                        •Stable Reservoir
                        Liquid Level
                           Stored
                           Product
                       Double-Wall
                       Underground Tank
Normal Conditions

The reservoir liquid level will
be stable if both the inner and
outer tank are tight.

The optional reservoir sensor
will activate an alarm if the
reservoir drains or overfills.
                    Outer Wall Breach

                    If the groundwater is below
                    the tank top, the monitor fluid
                    drains into the ground causing
                    the reservoir to drain.
                         Reservoir
                         Drains
                           Inner
                           Wall
                           Breach
                   Inner Wall Breach

                   Monitor fluid drains into the
                   primary tank causing the
                   reservoir to drain. No petroleum
                   product escapes from the
                   primary tank to pollute the site.
'S^"',""""™'	'"•r
'  jfitoundwater  ^

   - Res&ilwflf ;'  "^
                                                                         If the groundwater is over the
                                                                         tank top, the reservoir will
                                                                         overfill with groundwater and
                                                                         activate the high level alarm
                                                                         on the reservoir sensor.
    Figure 40. Hydrostatic monitoring system.  Source:  Owens Corning Fiberglass Corporation

-------
            fluid leaks into the tank, and the fluid level in the space is lowered.  If
            the breech occurs in the outer wall, the fluid will leak into the
            surrounding soil, and the fluid level in the space will decrease, or
            ground water will leak into the interstitial space and cause it to overfill.
            Alarms are set so that either a decrease or an increase in fluid level will
            be detected.

            The manual detection method is simply the use of a pole with a cloth or
            a petroleum-detecting paste on one end.  When the pole is inserted in a
            pipe that extends into the interstitial space, any leak into the space can
            be detected from visual observation of the cloth or from a change in
            color of the paste. This method is used in both double-walled tanks and
            in liner systems of secondary containment.

            To be successful, most methods rely on the leaked product being
            dkected by the containment to the position of the sensing device so that
            detection can be accomplished.  That is, the containment is usually
            sloped toward the sensor in such a way that leaks entering the interstitial
            space will be detected. Some electrical conductivity methods, however,
            rely on a continuous sensing wke that is capable of detecting leaks
            along the sensor's length and, thus, do not require a sloped containment.

            The successful implementation of secondary containment with
            interstitial monitoring involves the following two phases: (1) instal-
            lation, including site assessment, an evaluation of the site environmental
            conditions to assist in the determination of what type of secondary
            containment and interstitial monitoring system is appropriate; and
            (2) operation and maintenance.  The relationship between these phases
            is shown in Figure 41 along with some of the important factors that
            should be considered for successful implementation. The problems
            associated with the use of a secondary containment system with
            interstitial monitoring are discussed below in reference to these two
            phases.  The order of discussion does not necessarily reflect the priority
            of the problems discussed.

POTENTIAL PROBLEMS AND SOLUTIONS
            The main problems in the implementation of an interstitial monitoring
            system result from inadequate attention to installation, such as an
            inappropriate choice of containment for the site conditions or faulty
            installation of the monitoring system or containment, and lack of
            attention to the system during operation. The major problems and some
                                                                                  169

-------
   Insta
                       Installation
ation
Assess hydrogeological conditions
Evaluate residual contamination
Select partial or complete
  containment system
Choose interstitial monitor
  appropriate for site
Prepare site
Use qualified professionals
Select proper backfill
Follow installation codes
  and manufacturer's
  recommendations
                        Operation
   Operation
    Ana ysis
                       • Set threshold level
                       • Calibrate periodically
                       • Maintain and conduct routine
                           checks per manufacturer's
                           recommendations
                        Analysis
                                      Leak
                                      No Leak
                         Confirm leak detection events
                           against spills, overfills, and
                           environmental conditions
                           that may cause false alarms
     Figure 41.  General procedure for secondary containment with
     interstitial monitoring
170

-------
            solutions are discussed below, including regulatory approaches to
            oversee the implementation effectively.  A number of Agency solutions
            are offered for each problem, but not all of them need be undertaken. A
            summary of the problems and solutions is given in Table 13.  An
            asterisk identifies the most serious concerns with containment and
            monitoring systems.
            Installation
Operation
Analysis
Assuring proper installation of secondary containment

   The proper installation of secondary containment is important to the
   success of interstitial monitoring.  If the containment fails to collect
   the leaked product, the monitoring system may be unable to detect
   the leak before it reaches soil and ground water.

   For excavation liners, rocks, tree roots, or debris in the excavation
   may cause damage to the liners, which could cause leaks to go
   undetected.  All rocks, tree roots, debris and other protruding objects
   should be removed prior to the installation of these secondary
   containment systems.

   If improper backfill is used, two problems might arise. FRP
   double-walled tanks might not be supported sufficiently so that,
   when the tank is filled, the settling of the tank under the load of the
   fuel might cause cracks in the outer wall and produce leaks that
   escape detection. With excavation liner systems, if the backfill is
   not of sufficient permeability, leak detection may be delayed due to
   slow migration of the leaked product to the monitoring system.
   Both problems can be avoided by using sand or pea gravel as the
   backfill material.  These materials both provide adequate support to
   double-walled tanks and are of sufficient permeability so that leak
   detection will not be delayed. To prevent excessive settling of
   double-walled FRP tanks, additional precautions should be taken,
   such as compaction of the backfill under the tank and the use of a
   filter fabric in the excavation to prevent the backfill from migrating
   in shallow ground-water or tidal areas.

   Surface water run-on and precipitation can cause the backfill to
   become saturated.  If this  occurs, a vapor monitoring system will not
   function because the vapor cannot migrate to the monitoring system.
   Manual leak detection methods may also be adversely affected if the
   water is sufficient to float leaked product away from the monitoring
                                                                                  171

-------
Table 13. Indicators and Solutions for Problems with Secondary Containment with Interstitial Monitoring
Problem
Indicators
Tester Solutions
Agency Oversight Options
Account for shallow ground
water.
Ground water leaks into partial
containment causing failure of
vapor monitoring and manual
methods.
  Use double-walled tank, fully
  enclosed liner or jacket, or
  monitoring method not sensitive
  to water.
  Require site assessment to
  establish ground water depth
  and fluctuation.
*Prevent faulty electrical
Installation.
False alarm or inoperative system.
  Use waterproof, corrosion-proof
  junction boxes; follow
  manufacturer's installation
  procedures and local codes; use
  manufacturer's representative to
  supervise.
  Certify installers; review
  installation plans; inspect
  installation.
*Prevent faulty containment
Installation.
 Leaks escape detection.  Sudden
 catastrophic product loss.
  Install per local and national
  codes; use representative of
  manufacturer to supervise.
  Review plans; inspect; certify
  installers; require supervision
  of installation by manu-
  facturer's representative.
 Prevent interference from
 surface water run-on or
 precipitation.
 Water leaks into containment and
 causes failure of vapor monitoring
 or manual methods.
  Cover containment with sloped,
  impervious synthetic membrane,
  or use method unaffected by
  water in containment.
  Review manufacturer's
  literature to ensure sensor
  will respond or require test
  with the product.
 Assure proper operation and
 maintenance.
 System failure or false alarm.
  Electrical conductivity. Replace
  sensor wires after detection.
                                                                     Pressure systems. Check
                                                                     tightness of plumbing connections
                                                                     on routine basis.
  Review plans; train staff on
  monitoring systems; inspect;
  require demonstration of
  system and system tests
  before operation.

-------
                                                                       Fluid sensing systems. Set
                                                                       threshold to distinguish product
                                                                       from water.
                                                                       Hydrostatics monitoring systems.
                                                                       Add fluid in hot weather; add anti-
                                                                       freeze in cold weather.
                                                                       Manual detection. Use consistent
                                                                       procedure.
                                                                       Vapor monitoring. Waft for tank to
                                                                       cure so that emission of gases in
                                                                       interstitial space ceases. Set
                                                                       threshold above gas concentration.
Prevent false alarm due to
residual (background)
contamination.
False alarms.
Use system that can distinguish
leak from background contamina-
tion. For site conditions where
background would interfere with
detection, use double-walled tank
or other completely sealed
secondary containment.
                                                                        Require site assessment; review
                                                                        plans; require system tests to
                                                                        show that leaks can be
                                                                        distinguished from background
                                                                        contamination.
* Most frequent causes of failure.

-------
 Operation
 Analysis
   location. To avoid this problem, a synthetic liner should be placed
   over the tank and its backfill to prevent water from entering the
   backfill.  Another solution would be to use release detection
   methods that function adequately in water, such as electrical
   conductivity or ground-water monitoring methods.

   Improper sealing of the seams of liners can result in leaks escaping
   the containment systems without being detected. The installation of
   liner systems should be conducted only by professional installers
   who are experienced in liner installation.

   The installation of secondary containment should be supervised by a
   representative of the manufacturer or by an experienced professional
   who has been trained for the task in order to minimize the potential
   for damage to the containment and the consequent inability of the
   monitoring system to detect the leak.  Some states, for example,
   California and Rhode Island, requke that the installation be
   supervised by the manufacturer or by a representative of the
   manufacturer.  Other regulatory agencies, such as Dade County,
   Florida, require inspection of secondary containment systems during
   installation. Dade County also requkes prior agency review of the
   secondary containment design and installation plan. The City of
   San Jose requkes integrity testing of all secondary containment
   systems before they are accepted for use.

   In general, state  and local installation codes are in effect that, when
   followed, will promote proper installation and that will help
   minimize the possibility of damage to the secondary containment
   system. State and local installation codes must be followed,  and the
   reference section at the end of this chapter lists other national codes
   which may be used for guidance.

Accounting for shallow ground water

   Shallow ground water may adversely affect vapor monitoring and
   manual methods of interstitial monitoring. The problems related to
   vapor monitoring in shallow ground-water areas were discussed in
   Chapter 6. Vapor monitoring will not function properly in saturated
   soils because the movement of vapor is slowed or prevented.
   Manual methods might not detect a leak under high ground-water
   conditions if enough water is present to float the product away from
   the test location.
174

-------
Operation
    *
 Analysis
    A site assessment to determine the depth to ground water and the
    ground-water fluctuation should be made to assist in the choice of
    an appropriate secondary containment system. If the ground water
    at any time is expected to reach the level of the containment, a fully
    enclosed containment system could be used to prevent the ground
    water from transporting the leaked product away from the detection
    system. The use of the fully sealed containment would also allow
    vapor monitoring to be used at sites where vapor monitoring
    normally would not be effective. Pressure sensors, fluid sensors,
    and hydrostatic methods are used in double-walled tank systems
    and, therefore, should not be adversely affected by ground water as
    long as the containment does not leak.

    At sites where shallow ground water exists, the forces on both
    double-walled tanks and excavation liner systems caused by the
    ground water may cause shifting of the containment and consequent
    damage that would allow leaks to escape and go undetected.  These
    sites should be dewatered before installation of the containment, and
    the containment should be properly anchored to prevent shifting.

Preventing false alarms from background contamination

    Background contamination of soils and ground water can cause false
    alarms in interstitial monitoring systems that are not completely
    sealed. When the site has been previously contaminated or when a
    threat of contamination exists from neighboring properties, it may
    be necessary to use a monitoring method that can distinguish a leak
    from background contamination or to use a double-walled tank, a
    tank jacket, or an excavation liner that is completely sealed. This
    precaution will ensure that detected leaks are only from the tank
    system of concern and did not originate elsewhere. The background
    contamination problem and its solutions are discussed in more detail
    in the ground-water monitoring and vapor monitoring chapters.
           Assuring selection of proper interstitial monitoring system
                                                           —
Operation

Analysis
   Interstitial monitoring systems that are inappropriate for the stored
   product will not allow detection of leaks. An understanding of the
   monitoring method and its ability to detect the specific product is
   necessary.  For example, electrical conductivity systems function
   differently for conductive products than for non-conductive products
   such as petroleum. For petroleum products, one common design
                                                                                  175

-------
 Operation
 Analysis
   (Figure 42) is based upon the ability of petroleum products to
   complete an electrical ckcuit by degrading a coating that separates
   two metal conductors. Another design (Figure 43) allows the
   petroleum product to penetrate a braided cover on an electrical
   cable, which causes a conductive polymer jacket to swell and
   contact conductive metal wires, thus completing an electrical circuit.
   A non-petroleum electrical conductivity monitoring system used in
   an UST containing petroleum might not detect a leak. The coatings
   and conductive polymers of these systems are formulated to be
   specific to non-petroleum products and will not allow detection of
   petroleum.

   In addition, if the stored product is changed over the life of the tank
   an evaluation of whether to change the monitoring system also
   should be made to ensure that the system is appropriate for the new
   product. As discussed in the chapter on vapor monitoring, the vapor
   monitor's ability to detect  a leak depends upon the volatility of the
   product. If the product were changed from gasoline to used oil
   during the tank's lifetime,  the vapor monitor might not be able to
   detect leakage of the new product due to its lower volatility. All
   monitoring systems should be evaluated for use for the specific
   stored substance by checking the information available from the
   manufacturer and by asking the manufacturer to verify adequate
   response to the stored product, if necessary.

Assuring proper installation of monitoring system

   One major problem of all non-manual interstitial monitoring
   systems is improper electrical installation. If junction boxes are not
   waterproof and corrosion proof, short circuits may occur, which will
   cause system failure and inability to detect leaks. Waterproof
   junction boxes are more expensive than non-waterproof ones and,
   therefore, are sometimes left out of the as-built system by
   installation contractors to save money. Another common problem
   caused by "short cuts" taken by installation contractors is the use of
   two-conductor electrical cable in place of three-conductor cable.
   This substitution can cause a false alarm in some systems or may
   cause a diagnostic trouble alarm in other systems to alert the
   operator that a fault is present. Some systems require that bridge
   resistors be used to span.the unused sensor channels on the system's
   electrical control board. When these bridge resistors are omitted, a
   false alarm also may be given. It is important, therefore, that the
   manufacturer's recommended procedures be  followed during
   installation, that local electrical codes be adhered to, and that short
176

-------
 Metal Conductors —.
                                Coating Degradabte
                                by Hydrocarbons
Figure 42.  Cross section of electrical conductivity sensor
using degradable coating.  Source: DETEX Systems, Inc.
            Sensor Wires
  Continuity Wire
                       Conductive
                       Polymer
                       Layer
                                         Overbraid
Figure 43. Cross section of electrical conductivity sensor
using polymer jacket that swells. Source: Raychem
Corporation
                                                              177

-------
                cuts that attempt to ckcumvent these recommendations and codes
                not be used.

                When vapor monitoring systems are used in double-walled FRP
                tanks, the tanks must be allowed to cure a sufficient length of time
                so that outgassing of organic vapors from the tank interior walls will
                not cause false alarms. This outgassing should occur at the factory,
                but it occasionally does occur in the field. In the field, the primary
                solution is to set the threshold setting of the vapor monitor above the
                level of the gas being produced and then to check the threshold
                setting periodically to see if it can be lowered as the tank continues
                to cure. These periodic checks must be made or the potential will
                exist for small leaks to go undetected due to the higher threshold
                settings.
             Operation
[installation)  Preventing false alarms and undetected releases
  Analysis
False alarms and undetected releases may be caused by operational
factors. These factors and some solutions are discussed below for
the specific type of system in which they may occur.

Electrical conductivity systems.  False alarms will result if
electrical conductivity systems are not replaced after exposure to a
leak. The degradable coating system sensing wires must be replaced
after exposure to leaks; otherwise, the system will give a false alarm
because the coating has been degraded. The sensor cable in a
conductive polymer system also must be replaced after exposure
because the conductive polymer has closed the electrical circuit by
swelling and the cable can not be cleaned sufficiently to avoid a
false alarm if reused.

Pressurized systems. Losses of pressure or vacuum in these
systems will cause an alarm even if the leak that results in the
pressure loss is not caused by leaking product. The pipe fittings,
vacuum tubing, and vacuum or pressure gauges used in these
systems can become loose due to vibration from traffic or due to
physical contact with vehicles or with personnel. When a leak is
detected, the first action should be to check all plumbing
connections to the  interstitial space, to retighten any loose fittings,
and to  re-establish  the pressure or vacuum level. If the plumbing is
tight, then it can be assumed that a leak in the inner or outer tank
 178

-------
 wall has occurred because environmental conditions, such as
 temperature differences between the tank and the delivered product
 or the effect of barometric pressure, have not been found to cause
 false alarms.

 Fluid sensing systems. Optical devices, which are a class of fluid
 sensing systems, allow the leaked product to deposit on a mirror.
 The reflectivity of the mirror is decreased by the deposits, and the
 amount of light reflected from a light source, thereby, can be
 detected and can cause an alarm to be actuated.  Any deposit on the
 mirror, including condensation water, will cause the mirror's
 reflectivity to decrease. The system's threshold should be set so that
 condensation water can be distinguished from leaked product.

 Hydrostatic monitoring systems. Temperature extremes may
 cause false alarms or system failure in hydrostatic monitoring
 systems. Because hydrostatic systems use fluids to detect leaks,
 precautions must be taken in hot and cold weather to ensure that
 evaporation does not cause false alarms and that freezing
 temperatures do not disable the system and prevent it from detecting
 leaks.  During cold weather, an antifreeze solution that is compatible
 with the tank, the secondary containment, and the interstitial
 monitoring system should be used to prevent freezing of the fluid.
 Additional fluid must be added periodically to the interstitial space
 to compensate for evaporation during hot weather.

 Manual detection. Manual methods may not detect a leak if
 improper operating procedures are used. Manual methods consist of
 using a pole with either a cloth or a petroleum-detecting paste to
 indicate that a leak has occurred. If the pole is not always inserted
 in the same manner into the access hole to the interstitial space, the
pole may not sample leakage from the lowest point in the
containment and,  therefore, may miss leaks. Although this problem
would probably not cause large leaks to go undetected for long, a
small leak might not be detected within the required monthly
monitoring period. Specific spots should be designated as test
positions, these positions should be the lowest points in the
containment to which all leaks will drain, and the process of
inserting the pole  should be done as consistently as possible from
one test to the next. That is, the pole should not be inclined and
should be inserted until it touches the bottom of the containment.
                                                                   179

-------
               Vapor Monitoring. The primary solution is to set the threshold
               setting of the vapor monitor above the level of the gas being
               produced and then to check the threshold setting periodically to see
               if it can be lowered as the tank continues to cure.

APPROACHES TO ENSURING EFFECTIVENESS	
             Installation
            Several options are available to regulators to allow them to check for
            installation errors in secondary containment and interstitial monitoring
            systems.  For example, regulators could require site assessments and
            then review plans for secondary containment systems before installation
            to ensure that proper systems are chosen for shallow ground water and
            contaminated sites. A manufacturer's representative could supervise the
            installation or could certify installers. Inspection of the installations
            during construction and integrity testing are other alternatives.

            The regulatory agency could review monitoring system plans and
            specifications before installation to check the selection of the
            monitoring system for specific site conditions, such as existing
            contamination.
             Operation
            Manufacturers of monitoring equipment could be requested to provide
            training on the equipment for regulatory personnel, as is done for tank
            tightness testing in Rhode Island. Regulators could also request that
            manufacturers submit videos on the use of thek equipment, as is done in
            Massachusetts. In this way, agency personnel would become familiar
            with the equipment and would be better able to inspect monitoring
            systems and recognize both installation and operational problems that
            would affect monitoring.

            Regulatory agency personnel could conduct periodic inspections of
            monitoring systems during operation, as is requked by the city of
            Austin, Texas.  This approach allows calibrations and threshold values
 180

-------
            to be checked and maintenance and leak detection records to be
            examined to ensure that proper attention is being given to the system.

            Regulatory agencies could also require demonstration of the monitoring
            system and a demonstration of the system tests on-site before operation
            and periodically thereafter to promote regular system checks by owners
            and operators and to ensure that the monitoring system is both
            responsive to any leaks and can distinguish existing contamination from
            leaks.
REFERENCES
            1. U.S. EPA.  January 1986. Under ground Tank Leak Detection
               Methods: A State-of-the-Art Review. EPA 600/2-36-001.
               Hazardous Waste Engineering Research Laboratory, U.S. EPA.

            2. Baker/TSA, Inc. June 1987. Final Report: Study to Evaluate Cost
               and Effectiveness of Secondary Containment Methods for
               Underground Storage Tanks. Prepared for the Utility Solid Waste
               Activities Group.

            3. Radian Corp. September 7, 1984.  Secondary Containment for
               Underground Petroleum Products Storage Systems at Retail Outlets
               Final Report. Prepared for American Petroleum Institute, Ground
               Water Task Force.

            4. Government Institutes, Inc. 1987.  Underground Storage Tank
               Management, 2nd Edition.

            5. Petroleum Marketers Association of America.  November 1986.
               Underground Storage Tank Leakage Prevention, Detection, and
               Correction.

            6.  U.S. EPA. August 22,1986. Guidance Manual for the Storage and
               Treatment of Hazardous Waste in Tank Systems.
               EPA/530-SW-84-004.  Policy Dkective No. 9483.00-1A. Report by
               Fred C. Hart Associates, Inc., for Waste Treatment Branch, Waste
               Management Division, Office of Solid Waste, U.S. EPA.

            7. National Sanitation Foundation. 1983.  " Flexible Membrane
              Liners," NSF Standard 54.
                                                                               181

-------
             8. Steel Tank Institute.  1988.  "Standard for Dual Wall Underground
               Steel Storage Tanks," Standard F 841-88.

             9. Underwriters Laboratories, Inc.  1987. "Standard for
               Glass-Fiber-Reinforced Plastic Underground Storage Tanks for
               Petroleum Products," UL 1316.

            10. Petroleum Equipment Institute. 1987. "Recommended Practices
               for Installation of Underground Liquid Storage Systems," PEI/RP
               100.
182

-------
   Chapter IX
  Piping Release
Detection Methods

-------

-------
PIPING  RELEASE DETECTION METHODS  IX
SUMMARY
           Information collected by the U.S. Environmental Protection Agency
           demonstrates that about 25% of the underground storage tank systems
           in the United States are leaking.  Piping and loose fittings are
           responsible for the majority of these leaks, and piping is responsible for
           most of the large, catastrophic releases. Thus, an important part of any
           release detection program is the use of equipment to prevent or
           minimize the releases from piping.

           There are a number of piping release detection methods available,
           representing a wide variety of approaches. Each method has advantages
           that make it appealing under certain conditions.

           •   Flow restrictors provide nearly continuous release detection for a
           small capital investment. They are readily available and require little
           owner/operator involvement.

               Flow shutoff devices also provide nearly continuous leak detection
           and, because they are automated, require little effort from the on-site
           staff.  As their name implies, these devices completely stop a leak when
           it is detected.

           •   Line tightness tests require no permanent equipment and, therefore,
           no capital investment. Performed infrequently, line tightness tests
           interfere little with the daily operation of an UST. Line tests can
           often be perforated conveniently  as part of a tank tightness  test.

           •   Interstitial monitoring within secondary containment minimizes the
           environmental damage while providing sensitive release detection.
           Interstitial monitors can also be coupled with automatic sensors,
           shutoffs, and alarms.

           •   External monitoring of underground piping using ground-water or
           vapor monitoring can easily be integrated into external monitoring
           systems for USTs.
                                                                             183

-------
            Available studies demonstrate that flow restrictors, flow shutoffs, and
            line tightness tests are capable of meeting the required performance
            standards when designed and operated properly.

            The discussion in this chapter focuses primarily on the first three types
            of piping release detection:  flow restrictors, flow shutoffs, and line
            tightness tests. Interstitial, vapor, and ground-water monitoring for lines
            are essentially the same as for tanks, and the discussions in the chapters
            covering these methods are applicable to piping. The aspects of those
            release detection methods that apply only to underground piping are
            included in the sections below.

            The discussion presented in this chapter covers a range of possible
            problems that may occur with each piping release detection method.
            This does not mean that all, or even most, of these problems will occur
            at the same time or at the same site. Nor does it mean that all of the
            problems are of equal importance, in terms of frequency of occurrence
            or severity of impact to the effectiveness of the release detection
            method. Some problems, such as use of incorrect threshold value,
            happen less often, and other problems, such as tampering, are relatively
            easy to fix. Experienced testers, vendors, and installers are well aware
            of the problems and how to deal with them. For example, an
            experienced tester recognizes a vapor pocket in the line and knows the
            methods to use to try to remove the vapor pocket.  Release detection,
            however, is a growing industry, and new companies are being formed
            with less experience. This chapter presents a range of potential
            problems for educational purposes, not to imply that they will always
            occur.
BRIEF DESCRIPTION
            To understand the workings of the piping release detection technologies,
            it is important to understand how the different types of UST piping
            systems work. This section presents descriptions of common piping
            systems followed by descriptions of piping release detection
            technologies.
184

-------
  Piping Systems
 Figure 44, which follows on the next page, has been prepared to
 illustrate the following discussion of UST piping systems.

 Pressurized lines

     In a pressurized piping system, a submerged centrifugal pump
     located near the bottom of the tank moves the stored product from
     the tank to the point of end use (e.g., a dispenser at a service
     station). The delivery piping lines extend from the pump discharge
     point to the dispenser.  The product is essentially "pushed" from
     the tank, typically at positive pressures of 28 to 32 pounds per
     square inch (psi), although some piping systems  are pressurized up
     to 60 psi. Very large releases can occur very quickly if a hole or
     break occurs in a pressurized UST pipeline because the pump
     continues to push product through the line and through the hole or
     break. The higher the operating pressure of a line the higher the
     leak rate when a hole is formed. Pressurized systems generally are
     chosen for high-volume sites because the product can be delivered
     very quickly.

Suction lines

     Typical suction systems use a positive displacement pump at or
     near the point of end use to draw the product from the tank to the
     pump. The pump creates a lower pressure at the pump end of the
     pipe, thereby allowing atmospheric pressure to push the product
     along the pipe to the delivery point. Typical suction lines in the
     U.S. operate at a vacuum of 3 to 5 psi. When the pump is shut off
     or a hole or break develops, suction is interrupted, and the product
     flows backwards through the pipe, away from the dispenser and
     towards the tank. One or more  check valves in the pipe close when
    product begins to flow backwards through the pipe. Product is
    held in the pipe between the check valve and the point of end use
    or between check valves if more than one is present.  Product in the
    pipe between the tank and a check valve drains back into the tank.

    Suction systems are characterized as "European" or "American"
    systems. In the European system, the check valve is located
    immediately below the pump. When the pump is turned off or
                                                                     185

-------
                                                           Vent Pipes
Product Dispensers
Submerged Pump
Assembly (Inside Tank)
                                                                        Line Leak
                                                                        Detectors
                  Figure 44.  Typical retail gasoline station.
    186

-------
      lot®
Operation
Analysis
                 there is a line failure, suction is broken and most of the product
                 drains directly back into the tank. In the American system, the
                 check valve is located near the top of the tank, where it is often
                 called an angle check, or at the bottom of the suction line within
                 the tank, where it is called a foot valve. When there is a line
                 failure, product cannot drain into the tank and is released to the
                 environment. Although the total release is small, it can occur each
                 time product is dispensed over a long period, resulting in a
                 significant cumulative effect.

                 Although suction piping is environmentally safer than pressurized
                 systems, it has some limitations, including:

                 •  Potential to vapor lock at high altitudes and high ambient
                 temperatures;

                 •  Tank location restricted to within 50 feet of the pumps for
                 proper operation;

                 •  Slower delivery of product from the UST to the point of end use
                 than with pressurized systems; and

                 •  Larger diameter (higher cost) pipe than required for pressurized
                 systems.
            Piping Release Detection Technologies
Figure 45 is a flow chart of the process used for establishing release
detection systems for underground piping.

Automatic flow restrictors

     Currently, most UST systems with pressurized piping delivery
     systems use a device that restricts the flow of product from the
     pump to the point of end use in the event of a leak.  These devices
     are installed only on pressurized lines and do not entirely shut off
     the flow of product.  Flow restrictors are self-contained mechanical
     devices installed directly in the pipe using special fittings or in the
     pump housing.  The device has a diaphragm or piston that is
     activated by the pressure in the pump delivery system. Each time
     the pressure in the piping system drops below a preset threshold,
     typically 1 to 2  psig, a test of the system is performed.  No leak test
     is conducted if the system pressure remains above the threshold.
                                                                                  187

-------
                        Assess Site and Needs
                                      Whether to combine with tank
                                        release detection system
                                      Automated vs. manual
                                      Suction vs. pressurized piping
  Installation
                           Select Methods
                                      Flow restrictor or flow shutoff
                                      Tightness test, vapor monitoring,
                                        ground-water monitoring, or
                                        interstitial monitoring
                              Installation
                                       Proper procedures
                                       Trained and experienced
                                        installers
                       Operation & Maintenance
   Operation
I
Proper procedures
Routine inspection
 and maintenance
    Ana ysis
                            Interpretation
                  Leak
                  No Leak
          Figure 45. General procedure for piping release detection
188

-------
Operation
    *
Analysis
                After the dispenser is turned on, product flows through the line at
                1.5 to 3 gal/h. If there is no leak in the line, the line pressure
                reaches about 10 psig in about 2 seconds, and the diaphragm or
                piston opens completely, allowing full product flow and
                pressurization of the line. If there is a leak of at least 3 gal/h, the
                line pressure will not reach 10 psi, and flow is restricted to 3 gal/h.
                Leaks smaller than 3 gal/h are indicated if more than 2 seconds are
                required to fully pressurize the line. Reference No. 5 cited at the
                end of this chapter contains a more complete description of the
                workings of a flow restrictor.
            Automatic flow shutoff devices
 Automatic shutoff devices are relatively new piping release
 detection devices and have had limited actual use at operating UST
 sites.  Shutoff devices can be used only on pressurized lines.
 Several types of shutoffs are available, but all rely on detecting
 changes in line pressure. There are essentially two groups of
 shutoffs: those that monitor pressure increases and those that
 monitor pressure decreases. These devices respond to a suspected
 leak by completely shutting off the flow of product. All shutoff
 devices are permanent installations. The degree of automation can
 vary.  Some systems are run entirely by personal computers and,
 once a leak is suspected and the piping has been shut down, cannot
 be overridden by the on-site staff.

 One group of shutoff devices checks for leaks by monitoring line
 pressure decrease over time. A pressurized line will not be able to
 maintain a constant pressure in a static situation if a leak is present.
 Some  shutoff devices monitor the decay of line pressure over
 5-minute intervals and shut off the line if the rate of decay exceeds
 predetermined values (e.g., from 16 psi to 6 psi in 5 minutes). At
 least one shutoff device measures the time it takes for pressure to
 decay  from one predetermined value to another (e.g., time to go
 from 10 psi to 5 psi).  Most of these shutoff devices require more
 than one test indicating a leak before shutting off the line.  Tests  of
 the line are not run while the product dispenser is on, and most of
 the shutoff devices require a minimum amount of time between
 dispensings to run a test. Another type of shutoff system combines
 a pressure decay test with the flow restrictor described above.

Another shutoff device monitors the rate of pressure increase in  a
piping system once the pumps are activated. A leak in a
                                                                                  189

-------
                pressurized line will cause the line to pressurize at a slower rate
                than usual. For any given length of piping between the pump and
                the dispenser, the amount of time it should take for the length of
                piping to become fully pressurized can be calculated and
                programmed into the detection device. Should the pressure not rise
                quickly enough, a leak will be indicated, and the system will be
                shut down.
            Line tightness testing
Operation
 Analysis
A variety of line tightness tests are available. The following
descriptions are specific to pressurized lines. The tests may be
performed on suction lines using variations on these procedures.
Tightness tests on suction lines are typically performed at about 7
psi positive pressure (not vacuum).

No single pressure value is recommended for a line tightness test.
The method must be capable of detecting a leak of 0.1 gal/h at 1.5
times the line operating pressure. However, the actual line test
may be done at any pressure as long as the detectable leak rate is
mathematically equivalent to the federal performance standard.
This means that the evaluation to demonstrate that a line tightness
test meets the performance standards can be conducted at any line
pressure and then converted to a value equivalent to 1.5 times the
typical line operating pressure.

In a direct volumetric line tightness test, a hand pump or the
dispenser and the submerged pump is used to pressurize the piping
leading back to the pump in the tank. Under one approach to line
tightness testing, if pressure decreases in the piping system,
product is added to the piping system to bring the pressure back to
the level at the beginning of the test. The amount of product added
over time is recorded to estimate the leak rate. Another approach
to volumetric line testing observes the volume of product lost over
time in a tube above ground that is connected to the pressurized
piping, and no attempt is made to maintain constant line pressure.
An alternative approach is to pressurize the line pressure using a
pressure gauge on the hand pump or temporarily installed on the
dispenser. Some method of converting pressure change over time
to a leak rate is necessary. The conversion method will be specific
to the type of test and will be supplied by the manufacturer.
190

-------
Operation
Analysis
     In an indirect line tightness test, the piping is tested as a component
     of a full system test. First, a tightness test is performed of the
     entire UST system, as described in Chapter 4. An overfill test
     method must be used in order to include the piping.  If no leak is
     indicated by the test results, then both the tank and the lines are
     assumed to be nonleaking. If the total system test indicates a leak,
     the tank is isolated from the piping and tested by itself using the
     same test procedure. If the results of the tank tightness test
     indicate that the tank is tight, then the leak is presumably from the
     piping. If the tank is found to be leaking, then the condition of the
     piping is unknown, and the lines must be tested directly. The
     indirect approach is not a practical approach to conducting a line
     tightness test if the line is the only part of the UST system of
     concern at the time of the test.

     A relatively new type of line tightness test is the helium gas test, in
     which helium gas is injected into an empty product line. While the
     line is pressurized, a tester holding a portable helium detector
     walks over the piping route to detect the presence of helium rising
     from the ground. This method not only indicates a possible leak, it
     helps to locate where along a run of piping the leak is occurring.

Interstitial monitoring within secondary containment

     Another method of detecting leaks from underground piping is to
     place a monitor in the interstitial space between the piping and an
     outer barrier. The containment is a barrier between the piping
     containing the product and the environment. If a hole forms in the
     piping and the product leaks into the interstitial space, the barrier
     will direct the release towards the monitor, which detects the
     release. The types of barriers and interstitial monitors for piping
     are essentially the same as those for tanks, and the details are given
     in Chapter 8. A summary of the information applicable to
     underground piping is provided below.

     When secondary containment is used for piping, care must be
     taken that the containment extends the full length of piping, from
     the connection to the tank directly to the dispenser. Containing
     either end of a piping run is the most difficult portion and is often
     neglected as a result.

     One common method of secondarily containing underground
    piping is the use of trench liners (see Figure 39). The trench that is
     dug to install the piping can be lined with a flexible membrane
                                                                                   191

-------
                liner that is impervious to the stored product. The backfill and
                piping then are placed within the lined trench. The liners often are
                thermoplastic or polymeric sheets, typically 50 mm thick. Instead
                of flexible liners, rigid U-shaped pieces of plastic may be used to
                line the bottom and sides of the piping trench. When liners are
                used, the trench may be sloped away from the tank excavation to
                help differentiate between tank leaks and pipe leaks.

                The interstitial space to be monitored is the backfill between the
                trench liner and the piping. The simplest monitor consists of a
                sump at the lowest point of the piping system to "collect" the liquid
                from any leaks.  This sump can be monitored by visual inspection,
                a dipstick, hydrocarbon sensors such as those used in ground-water
                monitoring, or vapor monitors in the airspace of the sump.  A
                single monitor at the sump may indicate that a leak has occurred
                but does not help to locate the leak along a run of piping.
                Interstitial monitors placed at intervals along the run of piping can
                help to identify the location of the leak so that less piping must be
                dug up.

                When vapor monitoring is used, a typical well may be used that
                extends to the bottom of the trench; such a well will be shorter than
                that used for tank monitoring. Another approach is to use a
                horizontal slotted tube at or below the level of the piping rather
                than the conventional vertical well; these horizontal wells may be
                up to 10 feet long.

                Another form of secondary containment for piping is double-
                walled piping. The primary, or inner, piping that carries the
                product is contained within an outer pipe of larger diameter.  The
                inner and outer piping both may be made of fiberglass-reinforced
                plastic (FRP) or the inner pipe may be made of galvanized steel
                and the outer pipe of FRP. In a few specialized cases, both pipes
                may be made of steel. Care must be taken that the product in the
                lines is compatible with the FRP. For service stations, the
                diameters of the inner and outer piping often are 2 and 3 inches,
                respectively.

                A monitor is placed in the space between the inner and outer pipe.
                Double-walled piping is often sloped to a containment structure or
                observation well that can be monitored for the presence of
                hydrocarbon liquids or vapors. Small sumps may be placed
                periodically along the run of piping and monitored for liquids or
                vapors.
192

-------
Operation"]
             External monitoring

                  Both ground-water and vapor monitoring can be used to monitor
  Analysis I        f°r releases from underground piping. The descriptions provided
1	1        in Chapters 6 and 7 for tanks are applicable to piping.  For both
                  methods, wells a couple of inches in diameter are installed at
                  intervals along the run of piping. For ground-water monitoring,
                  the wells extend below the ground water, and a sensor detects the
                  presence of free product floating on the water. For vapor
                  monitoring, any leaked product will evaporate and diffuse through
                  the soil, and the vapor monitor will detect its presence.  It is
                  theoretically possible to connect piping monitoring wells to shutoff
                  devices, so that whenever a predetermined hydrocarbon liquid or
                  vapor level is detected, the delivery of product is halted. At least
                  one automated piping shutoff system has been designed to
                  incorporate a vapor monitoring system.

 POTENTIAL PROBLEMS AND SOLUTIONS
           This section presents a discussion of problems that have been
           encountered with piping release detection methods. Some of the
           problems and solutions are similar to those for tank release detection
           methods. To avoid repetition, this chapter includes only problems
           unique to the piping release detection methods; for additional potential
           problems, see the discussions in Chapters 4 and 6 through 8. The
           problems for each release detection method are presented generally in
           the order of importance.  Table 14 presents a summary of the indicators
           and solutions to common problems as well  as possible approaches that
           implementing agencies can use to prevent or overcome the problems
           with piping release detection methods.  A number of agency solutions
           are offered for each problem, but not all of them need be undertaken.
           The most serious concerns are indicated by an asterisk.  Table 14 and
           the discussion below are presented in the order of the flow chart
           (Figure 45 on page 188), not in order of importance. The most serious
           concerns have been indicated in Table 14 by an asterisk.
            Automatic Flow Restrlctors and Shutoff Devices
           In addition to meeting the regulatory requirement for continuous
           monitoring of large line leaks, automatic shutoff devices may also be
           used to meet the regulatory requirement for less frequent monitoring for
                                                                                193

-------
Table 14.  Indicators and Solutions for Problems Encountered During Piping Release Detection
Problem
Indicators
Tester Solutions
Agency Oversight Options
• Flow Restrictors and Shutoffs

Need to prevent tampering.           None.
*Need to identify and remove
vapor in the line.
Assure that necessary line
conditions for a valid test are
reached.

*Need for properly functioning
check valves.

• Line Tightness Testing

Assure sufficient waiting
time between filling line
and beginning test.
Assure that sufficient test data
are collected.

*Need to recognize and remove
vapor pockets.
Slow product delivery.
None.
False alarms.
Erratic readings. Readings that
increase or decrease, then level off.
None.
Excess product from line after
test.
Lock restrictors or shut off
control panel. Educate staff.

Increase pressure to absorb
vapor. Flush line with product
at high rate.

Select device based on
knowledge of device design
and UST operations.

Replace, clean, or repair
valves and retest.
 Wait at least 3 hours between
 filling lines and starting data
 collection.
 Collect data for at least 1 hour.
 Empty line, valve off blind ends,
 and retest. Increase pressure to
 absorb vapor.  Flush line with
 product at high rate.
 Inspect and perform test
 to see if device functions.

 None.
 Review plans prior to
 installation.
 Observe test. Check that
 valves are actually replaced
 or repaired.
 Review test reports for
 reasonable waiting times
 and data trends. Observe
 test.

 Observe test.  Review test
 results.

 Observe test.  Review test
 results.

-------
   Number of tests to be
   conducted must be fixed.
   High percentage of lines declared
   light."
   Assure use of proper protocol
   and correct threshold.
   High percentage of tanks declared
   light.'
   *Need for properly functioning
   check valve.
  False alarms.
     Interstitial Monitoring Within Secondary Containment
   *Assure correct installation of
   containment.
Product observed outside
containment.
   • External Monitoring (Ground-Water and Vapor)

   *Assure that monitoring wells      None.
   are properly placed.
Develop clear protocol with
specific number of tests and
follow it.
Criteria for determining "tight"
or "leaking" must be clear.
Threshold value for declaring
leak should be smaller than
minimum detectable leak rate
by a factor of at least 2.

Replace, clean, or repair valves
and retest.
Follow manufacturer's
specifications, with special
attention to seams and joints.
Pressure-test double-walled pipes.
                                     No more than 40 feet between
                                     wells. Conduct site assessment
                                     before installation.
Approve multiple-testing
strategies.  Observe tests.
Receive test results, and
track pass/fail ratios for
companies and methods.

Review test results and
calculations to see if
they agree with protocol.
Keep track of pass/fail
ratios for companies and
methods.

Observe test. Check that
valves  are actually replaced
or repaired.
Observe installation.
Certify/license
installers.  Review pressure
test results.
                                   Review site plans and
                                   monitoring plan before
                                   installation.
   * Indicates the most significant problems.
VO
U)

-------
             smaller leaks.  Shutoffs may be used in place of line tightness testing,
             ground-water monitoring, or vapor monitoring if they are sensitive
             enough to meet the performance standard for line tightness testing.
I Installation |   Need to prevent tampering
  Analysis
Even when there is no leak in the line, flow restrictors can cause
slight delays in the delivery of product after the dispenser is turned
on. When a leak occurs in a line with a flow restrictor, the delay of
delivery is even longer, causing the rate of delivery to be much
lower than usual.  When a shutoff device detects a leak, product
flow stops altogether. Such conditions can cause customers to
complain about poor service. At production facilities, the delayed
delivery of product from an UST may slow down production. In
reaction to the slow-down, production facility sometimes
deactivate the line leak detectors by removing them completely,
thus defeating the purpose of their operation.

A seal that can be installed on flow restrictors to indicate that
tampering has occurred has recently become available. Shutoff
devices usually are automated, and the control box can be
programmed and locked so that  only management personnel with
keys can override the shutoff function. Training on-site staff in the
importance of piping leak detection and the proper response to
warning signals also helps to overcome the problem of tampering.

One approach that owner/operator representatives or implementing
agency personnel can use to combat the problem of tampering is to
perform periodic checks on the line to see how the restrictor or
shutoff performs. Proper operation can be checked by simulating a
leak in the system and monitoring the line pressure. A defective
device will not restrict or shut off the flow of product. Some
shutoff devices have an automatic test mode that will perform this
type of test.
| Installation]  Need to identify and remove vapor in the line
  Analysis
At high altitude or high temperature, liquids volatilize more
quickly. As a result, product vapors may form in the piping. These
vapors can significantly increase the amount of time required for
the product to reach operating pressure because additional product
and time are needed to compress the vapor pocket. Such delays
may be interpreted falsely by the device as leaks, and product flow
will be restricted or shut off. If additional time is spent
196

-------
                  pressurizing the line, the vapors may eventually be absorbed back
                  into the liquid.
 | installation |   Line conditions need to be right for a valid test
  Analysis
As described above, most flow restrictors and shutoff devices have
minimum requirements for line conditions that must be met to
conduct a valid test, such as a period of at least 5 minutes in which
the line is not being used or a decrease in line pressure below a
threshold level (e.g., 2 psi).  At a busy service station, there may
not be enough time between dispensings to conduct a test with
some types of restrictors or shutoff devices. A fully pressurized
line without a leak or without large thermal changes may not drop
as low as 2 psi for days, even with no withdrawals occurring in the
line.

The level of use of an UST system and the design of the restrictors
or shutoffs should be considered together when selecting the line
leak detection device. If a device requires a minimum amount of
time to conduct a test, the typical time between dispensings at the
UST should be determined before selecting that device.  For
systems that require the line pressure to fall below 2 psi to conduct
a test, if there is  a leak in the line, the pressure will decrease after a
dispensing so that a test can be performed. High pressure can be
maintained in a line for long periods of time only if there is no
leak.
I installation]  Need for properly functioning check valves
                  As described above, there are usually check valves in a line that
                  prevent product from draining backwards towards the tank any
                  further than the valve. During a line tightness test, the pressure is
                  being maintained between the check valve and the hand pump or
                  dispenser. If the check valve does not close tightly, it may allow
                  product to seep through the valve and drain to the tank. For line
                  leak detection methods such as flow restrictors and shutoffs that
                  rely on pressurizing the line, this loss of product (and
                  corresponding loss of line pressure) due to a bad check valve
                  would falsely indicate a leak.

                  If the flow restrictor or shutoff indicates a leak but there are no
                  other indications that the line is leaking, it is possible to service the
                  check valve(s) in the line and retest the line. If a check valve is
                  faulty, it must be replaced. Sometimes, dirt or foreign particles
                                                                                     197

-------
                  become trapped in the check valve, preventing it from sealing
                  completely. In this case, cleaning the valve should be sufficient.  If
                  the line tests tight after the replacement or repair, then no
                  additional work is needed.
              Line Tightness Testing
I  Analysis
             This section discusses the problems of tightness tests that are performed
             only on the line. When lines are checked for leaks as part of tank
             tightness testing, the problems are the same as those discussed in
             Chapter 4 on tank tightness testing.
| Installation]  Need to allow sufficient time between filling and testing
The pressure in a pipeline is a function of the temperature,
coefficient of thermal expansion of the product, and the
compressibility of the product. For line tightness testing,
temperature is the single most important variable. As the
temperature of the product increases or decreases, the volume of
the product also increases or decreases, thus changing the line
pressure.  Shrinking of product as it cools may imitate a leak
because the line pressure is decreasing, and swelling of product as
it warms may mask aleak because of the increased line pressure.
When product moves from the tank into a line, there may be a
temperature gradient between the product and the surrounding
backfill.  If the product remains in the line, as it does during most
line tests, its temperature changes towards that of the backfill.  The
extent and rate of this change vary with the material of piping
construction, the backfill material, and the product in the line.  For
a 2-inch steel pipe in gravel backfill, if the temperature differential
between the product and the backfill is 5 to 15 degrees Centigrade,
the temperature in the product may take 3 hours to stabilize.
Temperature changes of 0.1 to 0.5 degrees Centigrade can cause a
5 to 10 psi change in pressure. The maximum changes in product
temperature occur immediately after product has been delivered to
the tank, when the differential between product and backfill
temperatures is the greatest.

The solution to this problem with line tightness testing is to wait at
least 3 hours after filling the line with product before beginning
data collection for the test. This time period allows the
temperature of the product to stabilize. An increasing or
decreasing trend in the data that eventually levels off indicates that
 198

-------
                 the temperature continued to change during part of the test but it
                 eventually stabilized and that there is no leak. Determination of
                 leak status should use only those data obtained after the readings
                 have leveled off.
Installation   Ensure that sufficient test data are collected
                 As discussed in Chapter 4 on tank tightness testing, insufficient test
                 data may not allow important trends in the data to be identified
                 and, thus, problems or leaks may be missed. In addition, if the test
                 does not last long enough, small leaks may be missed.

                 As a rule, obtaining more data increases the probability of correctly
                 identifying the presence of a leak. For line tightness tests, data
                 should be collected for at least one hour.
 Analysis
Installation]  Need to recognize and remove vapor pockets

                 Another factor that may affect the results of a line tightness test is
                 the presence of vapor pockets. When a line is completely filled
                 with product, vapor may become trapped in some areas, such as
                 deadend piping, bends in piping, or vertical stubs. This vapor
                 expands and contracts in response to temperature and pressure
                 changes more quickly and to a greater degree than the product in
                 the lines. Changes in vapor pocket size affect the line pressure,
                 thus masking or imitating a leak. Any test conducted with a vapor
                 pocket in the line is invalid.  For further discussion of vapor
                 pockets, see Chapter 4.

                 One approach to determining if a vapor pocket is present is to
                 measure the amount of product that drains from the line when the
                 pressure is released after the test. If no vapor pocket is present,
                 only a small amount of product will drain out as the walls of the
                 piping relax under the reduced pressure. If a vapor pocket was
                 present, a larger amount of product (at least 0.05 gal) will drain out
                 because the vapor pocket was compressed significantly under
                 pressure and expands when the pressure is released, pushing more
                 product from the line.

                 There are several approaches to removing vapor in a line  that can
                 be tried.  Sometimes increasing the line pressure will recondense
                 the evaporated product.  Sometimes flushing the lines with high
                 velocity product using the dispenser will remove vapors.
                 Sometimes pulling a vacuum on the line will remove the vapors,
                                                                                   199

-------
                   although if there are any slight cracks in the line or if the pump is
                   exposed, air will be drawn back into the line.  These approaches
                   will not work if the air is trapped in piping stubs. In that case, it
                   may be possible to uncover those portions of the piping and install
                   valves that shut off the blind end from the main run of piping.
                   Whatever method is used to try and remove the vapor, the line
                   should be retested after the 3-hour wait between filling and data
                   collection. It is not always possible to remove trapped vapor from
                   piping and conduct a valid tightness test.
I Installation [  Need for properly functioning check valve
!...   i_,                              	
                  The problems with bad check valves for line tightness testing are
                  the same as for flow restrictors and shutoff devices. The preceding
                  section contains a discussion of these problems (page 197).
  Analysis [
 | Installation")  Number of tests must be fixed in the protocol
     JL                                          	
  Operation |
                 When the results of a line tightness test indicate that the line is
                 leaking but the leak rate is only slightly above the threshold value
                 for declaring a leak, some testers repeat tests on the line until the
                 results of one test indicate that the line is tight. As discussed in
                 Chapter 4 on tank tightness testing, this approach is invalid and
                 reduces the probability of detecting a leak.  A multiple-testing
                 strategy is a valid approach to line tightness testing, but all of the
                 data from all of the tests must be used in the analysis unless the
                 protocol specifically excludes them (e.g., vapor pockets, bad check
                 valves). The number of tests to be performed and how the data are
                 analyzed must be explicitly defined in the line testing protocol, and
                 no deviations should be allowed from the protocol.
| Installation]  Proper protocol and correct threshold must be used
I  Operation |
                 As discussed in Chapter 4 on tank tightness testing, lack of a
                 well-defined data analysis protocol and clear criterion for declaring
                 a leak allows testers to make subjective decisions, leading to
                 unclear or false determinations of the status of the line.  A reliable
                 data-analysis protocol will have clear and detailed instructions on
                 how to convert raw data on pressure or volume changes to an
                 estimated leak rate. The protocol should specify how to determine
                 which data to use; when, If ever, it is permissible to discard data;
200

-------
                what conversion factors to use; how to determine the conversion
                factors; and what mathematical computations are needed.

                To determine if the piping is leaking, the estimated volumetric leak
                rate must be compared to a threshold value. This threshold value
                must be predetermined as part of the test design and its use must be
                well defined in the test protocol.  Discretion on the part of the
                tester in determining the leak status should not be allowed.

                As discussed further in Chapter 4, in order for a test method to
                perform well in detecting small leaks, the threshold value must be
                smaller by a factor of 2 or more than the smallest leak to be
                detected. The federal regulation requires a line tightness test
                method to have a minimum detectable leak rate of 0.1 gal/h. For a
                test method to meet this requirement, its threshold must be less
                than 0.1 gal/h.  The most commonly used threshold for line
                tightness testing is 0.025 gal/h.
            Interstitial Monitoring Within Secondary Containment
Operation
 Analysis
The problems and solutions specific to interstitial monitoring for piping
are discussed below. Additional information on interstitial monitoring
for tanks that may be applicable to piping is included in Chapter 8.

Ensure correct installation of secondary containment

    As discussed on pages 191 and 192, there are two types of
    secondary containment: trench liners and double-walled piping.
    Incorrect installation is a problem for both types.

    Incorrect installation of the liner is the most important potential
    problem with trench liners. Piping trenches are very narrow and
    long, and piping usually joins to a building or dispenser. To cover
    a very narrow trench and difficult areas such as near buildings or
    dispensers usually requires piecing together smaller pieces of liner.
    Seams are the most vulnerable to leakage, and trench liners can
    have many seams. A trained and experienced professional is
    necessary to ensure that the liner is designed for as few seams as
    possible for the site and that the liner is installed correctly.
                                                                                  201

-------
                 Incorrect installation also is a problem with double-walled piping.
                 Joining segments of piping so that the joint is tight is more difficult
                 for double-walled than for single-walled piping. As part of the
                 installation procedure, the inner pipe should be tested before the
                 outer pipe is installed and tested. The inner pipe should be tested
                 at 50 psi or 1.5 times the working pressure of the system. The
                 outer piping should be tested at 5 psi. In addition, double-walled
                 piping sometimes left "pen" (single-walled) where it joins the tank
                 or dispenser. Sumps may need to be placed at these points if
                 complete containment cannot be installed. Trained and
                 experienced personnel should be used to install double-walled
                 piping.
             External Monitoring
 Operation
  Analysis
The problems and solutions specific to using ground-water or vapor
monitoring as release detection for piping are discussed below. The
problems with external monitoring of piping that are the same for
external monitoring of tanks are discussed in Chapters 6 and 7.

Assure that monitoring wells are properly placed

     The area that a monitoring well network must cover for piping is
     very large because piping runs can be very long, and leaks can
     occur in any portion of the line. Detection is, in part, a function of
     the distance between the monitoring wells and a leak. Because of
     the large area covered by piping systems, monitoring well
     networks are sometimes designed with too few wells, to reduce the
     cost. If monitoring wells are placed too far from each other or
     from the pipe, the amount of time before a leak is detected may
     increase or, in extreme cases, a leak may go undetected.

     Table 15 presents a summary of the requirements for various states
     on the placement of vapor and ground-water monitoring wells with
     regard to piping.  Although the requirements are diverse, in
     general, wells must be separated by no more than 20 to 35 feet.
     These requirements are reasonable based on-EPArresearch
     indicating that a design that includes at least one well every 40 feet
     should be sufficient for gasoline tanks in a clean, dry backfill. If
     the backfill is not highly permeable (e.g., it is native fill material)
     or the migration of liquid product or vapors is impeded by other
     factors, the number of sensors should be increased by a factor of
     two.
202

-------
                        Table 15

   Typical State Network Design Requirements For
   Vapor and Ground-Water Monitoring of Piping
Maine
 Vapor
According to manufacturer's
specifications
At a minimum: 1 well
at each piping joint
No piping run > 15 feet from well
Santa Clara County, California
 Vapor, aspirated systems
South Carolina
 Ground water

Vernon, California
 Vapor
General:  1 well every 35 feet
At station: 1 well for each set of piping
1 well at each pump island

Minimum of 2 wells every 30 feet
Design network for 15-foot diameter of
influence
Source:  U.S. EPA
The Federal regulation requkes that ground-water monitoring wells
for tanks be placed as close as possible to the tank and that vapor
monitoring wells be installed in the backfill. These placement
criteria should be followed for piping leak detection as well.

For additional information on monitoring well design and
installation, see Chapters 6 and 7.
                                                               203

-------
ENSURING EFFECTIVE RELEASE DETECTION FOR PIPING
             Site Inspections
            Site inspections could provide useful oversight for several of the piping
            release detection methods. Proper installation of secondary containment
            is crucial to successful interstitial monitoring, so observation during
            installation could be a powerful tool. Development and use of a
            checklist of important elements of proper installation during the
            inspection would increase the usefulness of the site visit.  The City of
            San Jose, California, has developed several secondary containment
            inspection techniques. In one technique, a lined piping trench is filled
            with water, the water level immediately after filling is marked, and the
            water level 24 hours later is measured;  1/4 inch of water loss due to
            evaporation is assumed.  If returning the next day is infeasible, then
            paper can be placed under the seams of the trench liner before it is filled
            with water. A leak will mark the paper within minutes; and the effects
            of evaporation are avoided. For double-walled pipes, the City of San
            Jose performs either a hydrostatic or pneumatic pressure test. If a
            pneumatic test is performed, soapy water is applied to all pipe
            connection during the test. A leak will be indicated by bubbles.

            If selected as the release detection method, line tightness tests are
            required annually or every 3 years, so the number of tests that would be
            conducted within a jurisdiction each year is relatively small.
            Implementing agency personnel could be onsite for some of these tests
            to ensure that the proper waiting time and test duration are observed and
            that no vapor pockets are present.

            For sites where external monitoring is planned, a visit to the  site before
            installation to ensure that conditions are, in fact, appropriate  is an
            option. A checklist of pertinent features, collection of soil samples, and
            measurement of the depth to ground water might be considered while on
            site.
            Because tampering is the primary problem with automatic flow
            restrictors, random inspections by agency personnel to check for
            tampering would be effective. At a minimum, the seal on the restrictor
            should be checked. It is also possible to simulate a leak in the part of
            the line under the dispenser and observe the response of the restrictor.
            The records of monitoring results and repair and maintenance could be
            checked whenever a site visit is made.
204

-------
 Data Review
 The implementing agency could requke that all line tightness test results
 be submitted to the agency for review and recalculation. See Chapter 4
 for discussion of the types of data review and compilation that could be
 useful. In particular, the length of the waiting and test times should be
 checked.

 The agency could also require submittal of the results of external and
 interstitial monitoring. Because these methods are performed each
 month, the volume of data may be overwhehning. In this case, only the
 results of tests from every other month or every six months could be
 requked.
 Guidance and Training
 Education of owner/operator staff on the importance of piping leak
 detection might help prevent some problems such as tampering.
 Education in how the methods work might help owner/operators
 provide meaningful oversight during installation of equipment or
 operation of a method. Review of manufacturers' in-house training
 programs is another possible oversight mechanism.
 Approval and Certification
For external and interstitial monitoring the design of the release
detection system is particularly important to its success. For these
methods, it may be appropriate to requke submittal of the plans for
review and approval prior to any installation.  The plans should include
a site map and pertinent hydrogeological data, manufacturer's
information, and the installer's recommendations.

Because proper installation is so important to interstitial monitoring
with secondary containment, certification or licensing of installers could
be requked. Either the implementing agency could run the program or
they could requke that installers have a minimum amount of training by
third parties, such as manufacturers.  Similarly, line tightness testers
could be licensed or certified to conduct tests to ensure that they
understand the important aspects of the testing and analysis protocol.
Line tightness testing is often performed in conjunction with tank
                                                                      205

-------
            tightness testing. If there is an existing licensing program for tank
            testers, it would be relatively easy to include line testing.
REFERENCES
            1.   U.S. Environmental Protection Agency. July 1988. Common
                Human Errors in Release Detection Usage.  Prepared for U.S.
                EPA by Camp Dresser & McKee, Inc.

            2.   Maresca, J.W., J. L. Chang, Jr., and P. J. Gleckler. January 1988.
                A Leak Detection Performance Evaluation of Automatic Tank
                Gauging Systems and Product Line Leak Detectors at Retail
                Stations. Vista Research, Inc.

            3.   Maresca, J. W., and J. S. Farlow. December 17,1987.  Pipeline
                Leak Detection Modeling and Analysis Results.  Presentation to the
                U.S. EPA's Office of Underground Storage Tanks.

            4.   National Fke Protection Association. 1987. NFPA329:
                Under ground Leakage of Flammable and Combustible Liquids.
                Batterymarch Park, Quincy, Massachusetts.

            5.   Schwendeman, T.G. and H. K. Wilcox. 1987.  Underground
                Storage Systems - Leak Detection and Monitoring. Lewis
                Publishers, Chelsea, Michigan.
206

-------
  SUBJECT INDEX & APPENDICES
Subject Index
Appendix A—List of Figures
Appendix B—List of Tables
207
212
215

-------

-------
 SUBJECT INDEX
 Air pockets (see vapor pockets)
 Alarm levels
    automatic tank gauging 82
    vapor monitoring 106, 110,120
 Allowable methods 5,7
 "American" piping systems 185-186
 Annular seal (see well seals)
 Approval of methods 11-12, 30-31,44,
 74-75, 93-94, 124, 205-206
 Aspirated vapor monitoring systems 105,
 113

 B
 Backfill
    ground-water monitoring 134-137
    interstitial monitoring 171-172
    tank tightness testing 60
    vapor monitoring 103-105, 113
 Background contamination
    ground-water monitoring 142
    interstitial monitoring 175
    vapor monitoring 105-106,110
 Bailer 143
 Barrier (see liners and double walls)
 Bentonite seal (see well seals)
 Blended fuels (ATG) 91-92
 Borehole (see installation, ground-water
 monitoring)
 Bucket (see bailer)
 Bungs (see fittings, loose)
Calculations, performed correctly
   automatic tank gauging 92
   tank tightness test 73
Calibration chart (see tank chart)
 Calibration
    ground-water sensor 157
    vapor sensor 119-120
 Cement seal (see well seals)
 Certification of methods or personnel (see
 approval)
 Check valves 185, 197-198
 Chemical-sensitive pastes 143
 Coefficient of thermal expansion
    automatic tank gauging 86
    line tightness test 198
    tightness test 50,67-68,72
 Condensation (see evaporation)
 Conduits (ATG) 78, 83, 86
 Constant product level 69-70
 Creepage 17, 37

 D
 Data review 10-11,30,44,74,93,123-124,
 205
 Detection criteria (see threshold)
 Delivery of product
    automatic tank gauging 86-87
    tank tightness testing 58, 60
 Diffusion 96, 111
 Dip stick
    piping 192
    tanks 169,179
Dissolved product 128
Double walls
    piping 192,202
    tanks 162-163
Drop tubes 57
                                                                                 207

-------
E
Electrical conductivity interstitial monitors
166,175-176,178
"European" piping systems 185-186
Evaporation
    automatic tank gauging 88
    inventory 20
    tank tightness testing 70
Excavation liner (see liners)

F
Filter pack
    ground-water monitoring 153-154
    vapor monitoring 118
Fittings, loose 52,56
Fluid-sensing interstitial monitors 166
Free (floating) product 128,179
Flow rate (see leak rate)
Flow resttictor 183,187-189
Flow shutoff 183,189-190
Fractured rock 144-145

G
Gaskets (see fittings, loose)
Gauge stick (pole) 14,17,
Gauging procedure
    inventory control 16-17
    manual tank gauging 34,37
Gravel pack (see filter pack)
Ground-water depth
    ground-water monitoring 130-137,
139-141
    tightness testing 64-66
    vapor monitoring 107-108
    interstitial monitoring 174-175
Ground-water flow gradient 139-141,
147-150
Ground water influences
    automatic tank gauging 90
    tank tightness testing 64-66
Ground-water monitoring sensors 142-143
Guidance 11,30,44,74,93,124,205
H
Height-to-volume conversion factor
   tank tightness testing 50,67,72
   automatic tank gauging 91
Helium gas 191
Hydraulic conductivity 134-137
Hydrostatic interstitial monitors 166-167,
179
Installation
    automatic tank gauging 78-81, 83, 86
    ground-water monitoring 150-156
    interstitial monitoring 161,176-178,180
    secondary containment 161,171-174,
201-202
    vapor monitoring 111-114
Interpretation of results
    automatic tank gauging 89
    ground-water monitoring 158-159
    inventory control 29-30
    line tightness tests 200-201
    manual tank gauging 42-43
    tank tightness testing 50, 71-73
    vapor monitoring 110, 121-123
Interstitial monitors, types
    tanks 166-169
    piping 191-192
Interstitial space 161-162
Inventory mode (ATG) 78, 82
Inventory reconciliation (see reconciliation)
Joints (see fittings, loose)
Junction boxes 176

K
K (see hydraulic conductivity)
 Leak detect mode (ATG) 78, 81
 Leak rate 50, 71-73, 89, 200-201
 208

-------
 Line tightness testing 190-191,198-201
 Liners
     piping (see trench liners)
     tanks 161,165

 M
 Manifolded tanks tightness testing 56
 Manual tank gauging
     monthly standard 42
     weekly standard 42
 Manways (see fittings, loose)
 Moisture
     interstitial monitoring 171,174
     vapor monitoring 107-108
 Methane 108-109
 Multiple-testing strategy 70-71, 88-89

 N
 Network design for monitoring wells
    ground water, tanks 144-150
    piping 202-203
    vapor, tanks 105,109-114,123
 Nonvolumetric tank tightness tests 48

 o
 Operation and maintenance
    ground-water monitoring 157
    interstitial monitoring 161,180-181
    vapor monitoring 119-121
 Optical interstitial monitors 166,179
 Outgassing  178
 Overfill tightness tests 48-50
Passive vapor monitoring systems 105,
113-114
Permeability of soil (see backfill)
Piping, abandoned 56-57
Porosity (see backfill)
Poor access 57
Pressure-sensing interstitial monitors 166,
178-179
Pressurized piping 185
 Product-finding paste
     interstitial monitoring 169
     inventory control 17
     manual tank gauging 37
 Product-soluble devices
     ground-water monitoring 143
     interstitial monitoring 176-177
 Protocol
     automatic tank gauging 88-89
     line tightness test 200
     tank tightness testing 50, 70-73
 Pump meter 25
 Purging (see well development)
 Purpose of handbook 1-2

 R
 Reconciliation
    inventory control 190-193
    manual tank gauging 42-43
 Residual vapors (see background
 contamination)
 Restrictions on methods 7
 Sampling frequency
    automatic tank gauging 81-82
    tank tightness testing 68
 Seals (see well seals)
 Seams of liners
    piping 201
    tanks 174
 Secondary containment (see also liners and
 double walls)
    piping 191-192
    tanks 162-165
 Site assessment
    ground-water monitoring 127-128, 130,
    interstitial monitoring 174-175
    vapor monitoring 95
Site inspection 7,10, 30,43, 73, 93,123,
 174-175, 204
Slot size (see well screen)
Soil moisture (see moisture)
                                                                                      209

-------
Soil porosity (see backfill)
Solubility of product 141
Specific gravity of product 141-142
Specificity of sensor
    ground-water monitoring 142-143
    interstitial monitoring 175-176
    vapor monitoring 109
Spill and overfill protection for vapor
monitoring 106
Sticking (see gauging)
Stock control (see reconciliation)
Striker plate 17,37
Subsurface conduits 147,149-150
Suction piping 185,187
Surface grout (see well seals)
Surface seal (see well seals)

T
Tampering
    ground-water wells 156
    piping 196
    vapor wells 119
Tank chart
    inventory control 13-14, 21-24
    manual tank gauging 39-42
Tank deformation
    automatic tank gauging 86-87
    tank tightness testing 60-61
Tank end deflection (see tank deformation)
Tank jacket 162,164
Temperature
    automatic flow restrictors 196
    automatic tank gauging 86-88
    ground-water monitoring 157
    interstitial monitoring 179
    inventory control 20
    line tightness testing 198-199
    manual tank gauging 38
    tank tightness testing 58-62, 68-69
    vapor monitoring 106-107
Test length and frequency
    automatic tank gauging 81,82
    line leak detectors 187-188,197
    line tightness testing 199
    manual tank gauging 33,38
    tank tightness testing 68
Threshold value
    automatic tank gauging 89
    tank tightness testing 50,72
    line tightness testing 200-201
Tilted tanks
    inventory control 21
    tank tightness testing 63
Tracer compounds 103, 106, 109
Training (see guidance)
Trench liners 191-192, 201

U
Underfill tightness testing 48-51, 81
Use of handbook 3-12

V
Vapor pockets
    flow restrictors/shutoffs 196-197
    line tightness testing 199-200
    tank tightness testing 62-64
Vaults 162
Volatility of product
    interstitial monitoring 176
    vapor monitoring 96,102-103,106-108
Volumetric tank tightness testing 49-51

w
Waiting times
    automatic tank gauging 86-87
    line tightness testing 198-201
    tank tightness testing 58-62
Water in tank
    automatic tank gauging 82,90
    inventory control 17, 20
    manual tank gauging 38
 210

-------
 Water-finding paste 17, 20,
 Water sensor (ATG) 82,90
 Water table (see ground-water depth)
 Wells, ground-water
    casing 150-152
    depth 153
    development 154, 156
    diameter 152
    documentation 156
    number (see network design)
    placement (see network design)
    screen 139-140,150-152
    seals 134-136
    security 156
Wells, vapor
    casing 118-119
    depth 111-114
    diameter 96,105,114
    documentation 119
    number (see network design)
    placement (see network design)
    screen 118
    seals 118-119
    security 119
Wiring
    automatic tank gauging 78, 83, 86
    interstitial monitoring 176
                                                                                   211

-------
 LIST OF FIGURES
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6
Figure 7
Figure 8
Figure 9
Figure 10
Figure 11

Figure 12
Figure 13

Figure 14
Figure 15
Figure 16
Figure 17
Figure 18

Figure 19
Development of a state or local leak detection program            4
General procedure for inventory control                        15
Sample calibration chart                                   22-23
Sample inventory control daily reconciliation form               27
Sample inventory control monthly reconciliation form            28
General procedure for manual tank gauging                     35
Sample calibration chart                                   40-41
General procedure for conducting a volumetric tank test          49
Comparison of partially filled and overfilled tanks               51
How temperature changes can be mistaken for a leak             59
How structural deformation of the tank
can be mistaken for a leak                                    61
Location of vapor pockets in an overfilled tank                  63
Effect of ground water on the rate of flow
through a hole in an underground tank                         65
Schematic of automatic tank gauging system                   79
General procedures for ATG systems                           80
Underground storage tank system with vapor monitoring wells    97
General procedures for vapor monitoring                       98
The effect of soil conditions on vapor concentrations
at a deep vapor well                                        104
The effect of temperature on gasoline volatilization rates        107
212

-------
 Figure 20   The effect of backfill moisture levels
             on gasoline volatilization rates                                 108
 Figure 21   The effect of vapor sensor placement on leak detection time      113
 Figure 22   Typical underground storage tank site                          115
 Figure 23   Typical vapor monitoring well cross section                    116
 Figure 24   Modified vapor monitoring well cross section                   117
 Figure 25   Interpretation of vapor monitoring results                       122
 Figure 26   Typical ground-water monitoring system                       129
 Figure 27   General procedure for ground-water monitoring                 131
 Figure 28   Well seal will prevent interception of free product
             when water table is low                                       135
 Figure 29   Well without proper surface seal may be contaminated by
             surface runoff                                                136
 Figure 30   Range of hydraulic conductivities (K)
            for the major soil classes                                      138
 Figure 31   The well screen is placed to extend over the entire range
            of water table fluctuation                                      140
 Figure 32   Free product will preferentially flow through
            fractures and cavities                                         145
 Figure 33   Off-site sources of contamination should be considered
            when designing the monitoring well network                   148
Figure 34   Subsurface utility conduits will act as preferential
            pathways for free product migration                            149
Figure 35    Components of a typical ground-water monitoring
            well installed in a borehole                                    151
Figure 36    Components of a monitoring well using backfill material
            as the filter pack                                             155
                                                                           213

-------
Figure 37   Two double-walled tank configurations
Figure 38   Jacketed tank
Figure 39   Tank with excavation liner
Figure 40   Hydrostatic monitoring system
Figure 41   General procedure for secondary containment
            with interstitial monitoring
Figure 42   Cross section of electrical conductivity sensor
            using degradable coating
Figure 43   Cross section of electrical conductivity sensor
            using polymer jacket that swells
Figure 44   Typical retail gasoline station
Figure 45   General procedure for piping release detection
163
164
165
168

170

177

177
186
188
214

-------
 LIST OF TABLES
 Table 1


 Table 2


 Table 3


 Table 4

 Table 5


 Table 6


 Table 7


 Table 8

 Table 9


 Table 10


 Table 11


 Table 12


 Table 13


Table 14


Table 15
 Effect of Site Conditions on Success of Release
 Detection Methods for Tanks and Piping                      8-9

 Indicators and Solutions for Problems Encountered
 During Inventory Control                                 18-19

 Indicators and Solutions for Problems Encountered
 During Manual Tank Gauging                                36

 Monthly and Weekly Manual Tank Gauging Standards           42

 Indicators and Solutions for Problems Encountered
 During Tank Tightness Testing                            53-55

 Indicators and Solutions for Problems Encountered
 with Automatic Tank Gauging Systems                     84-85

 Indicators and Solutions for Problems Encountered
 During Vapor Monitoring                               99-101

 Typical Vapor Pressures of Petroleum Products                102

 Typical State Network Design Requirements
 for Vapor Monitoring                                      112

 Indicators and Solutions for Problems Encountered
 with Ground-Water Monitoring                         132-133

 Typical Network Design Requirements
 for Ground-Water Monitoring                              146

 Applicability of Leak Detection Methods
 to Secondary Containment Systems                          167

 Indicators and Solutions for Problems with
 Secondary Containment with Interstitial Monitoring            173

 Indicators and Solutions for Problems Encountered with
 Piping Release Detection                               194-195

Typical State Network Design Requirements
for Vapor and Ground-Water Monitoring of Piping            203
                                                                     215

-------

-------
                                                                             July 1992

     ,                          * ffcwtor: 'A Swamry o/Leak&mGftm Mett&ds/dr
 Mf^^^tme^gnmd Storage Totd® $y*tem$ describes & leak detection method not Included.
 In tli§ o^ginal publication,

 Statistical Inventory Reconciliation

 Will I be in compliance?
 Statistical inventory reconciliation (SIR), when performed according to the vendor's specifica-
 tions, meets Federal leak detection requirements for new and existing underground storage tanks
 (USTs) and piping as follows. SIR with a 0.2 gallon per hour leak detection capability meets the
 Federal requirements for monthly monitoring for the life of the tank and piping.  SIR with a 0.1
 gallon per hour leak detection capability meets the Federal requirements as an equivalent to tank
r-tightnessJesting..-SIR could, in some-cases, meet-the-Eederal requirements-for line tightness.
 testing as well. (For additional leak detection requirements for piping, see the sections on leak
 detection for piping.)  You should find out if there are State or local limitations on the use of SIR
 or requirements that are different from those presented below.

 How does it work?
 Statistical inventory reconciliation analyzes inventory, delivery, and dispensing data collected
 over a period of time to determine whether or not a tank system is leaking.

      • Each operating day, you measure the product level using a gauge stick or other tank level
       monitor. You also keep complete records of all withdrawals from the UST and all
       deliveries to the UST. After data have been collected for the period of time required by
       the SIR vendor, you provide the data to the SIR vendor.

      • The SIR vendor uses sophisticated computer software to conduct a statistical analysis of
       the data to determine whether or not your UST may be leaking. The SIR vendor provides
       you with a test report of the analysis results.

 What are the regulatory requirements?
--rrs^To be;allowable as-monthly-monitoring,a-SIR-method must-be-able-to detect-a-leak at-—=--
       least as small as 0.2 gallons per hour and meet the Federal regulatory requirements
       regarding probabilities of detection and false alarm. Data must be submitted monthly.

      • To be allowable as an equivalent to tank tightness testing, a SIR method must be able to
       detect a leak at least as small as 0.1 gallons per hour and meet the Federal regulatory
       requirements regarding probabilities of detection and false alarm.

      • The individual SIR method must have been evaluated with a test procedure to certify that
       it can detect leaks at the required level and with the appropriate probabilities of detection
       and false alarm.

      • If the monthly test report is inconclusive, you must take the steps necessary to find out
       conclusively whether your tank is leaking.

      • You must keep on file both the test reports and the documentation that the SIR method
       used is certified as valid for your UST system.
                                                               d8) Printed on Recycled Paper

-------
Win it work at my site?
     • Generally, few product or site restrictions apply to the use of SIR.

     • A SIR method may be used on tanks with up to 1.5 times the volume at which that
       method was evaluated. If you are considering using a SIR method for tanks greater than
       18,000 gallons, discuss its applicability with the vendor.

     • Water around a tank may hide a hole in the tank or distort the data to be analyzed by
       temporarily preventing a leak. To detect a leak in this situation, you must check for water
       at least once a month.

What other information do I need?
     • Data, including product level measurements, dispensing data, and delivery data, should
       all be carefully collected according to the SIR vendor's specifications.  Poor data
       collection may produce inconclusive results and non-compliance.

     • The SIR vendor will generally provide forms for recording data, a calibrated chart
       converting liquid level to volume, and detailed instructions on conducting measurements.

     • Statistical inventory reconciliation should not be confused with other release detection
       methods that also rely on periodic reconciliation of inventory, withdrawal, or delivery
       data. Unlike manual tank gauging, automatic tank gauging systems, or inventory control,
       SIR uses a sophisticated statistical analysis of data to detect releases. This analysis can
       only be done by competent vendors of certified SIR methods*

     • You should "shop around," ask questions, get recommendations, and select a method and
       company that meet the needs of your site.

How much does it cost?
     • There are no installation costs. Equipment costs are minimal, although you should ensure
       that dispensing meters are in calibration and that your gauge stick or other tank level
       monitor is in good condition.  Annual costs for the service may depend on your data
       quality, how you provide data to the vendor (paper, diskette, or modem) and the number
       of tanks and sites you have.

     • Here are possible costs for a typical station with three tanks:

       - Used as a monthly monitoring method (with 0.2 gallon per hour leak detection
          capability), SIR could cost $840 to $ 1200 yearly; or

       - Used as the equivalent to tank tightness testing (with 0.1 gallon per hour leak detection
         capability), SIR could cost $225 to $540, to test three tanks one time. If piping is also
         tested using SIR, additional costs would be added.

-------