37    LEAK TIGHTNESS TESTING OF UNDERGROUND TANK
SYSTEMS
Richard K. Henry, Environmental Analyst
Lawrence Livermore National Laboratory
Livermore, California 94550
INTRODUCTION
One hundred fifty underground tanks and sumps used to contain petroleum products and potentially contaminated process wastewater have given Lawrence Livermore National Laboratory an
unusual opportunity to experiment with a variety of leak test methods and compare applicability of
each method. Five types of commercially available Precision Test leak test equipment have been used:
Associated Environmental Systems, Hunter Leak Lokator, Horner Ezy-Check, Heath Petro-Tite, and
Tank Auditor.
TANK SYSTEMS
Two types of underground tank systems are present at LLNL: tanks and associated piping for the
storage of fuel for boilers, emergency generators, and vehicles; and tanks or sumps and associated
piping for the retention of potentially contaminated wastewater generated by metal finishing and wet
chemistry operations.
The fuel storage tank (Figure 1) is a relatively closed system. The tank is generally constructed of
steel, and tank connections typically include the fill pipe, vent line, return line, and a manway. A leak
in the fuel storage system is most likely to occur in one of these connections or in the piping itself as a
result of improper installation or mechanical failure.
The wastewater retention systems (Figure 2) at LLNL vary considerably in design and in construction materials. Tanks are made of steel, polyethylene, steel-clad concrete, concrete, or fiberglass. The
systems involve open-top sumps, drains, complicated and extensive piping, lines for sewer discharge,
and pump out lines. Because of greater lengths and complexity of piping, and because the containers
are often open at the top, wastewater systems are not only more difficult to test for leaks but are also
more leak-prone than the fuel storage systems.
All tanks and sumps were tested if they: 1) do not have secondary containment; 2) are empty or in-
use; 3) are intended to contain hazardous or potentially hazardous liquid substances; and 4) are not
dedicated exclusively to emergency use.
TANK TESTING METHODS
A program of annual tank testing, using the Precision Test, was initiated in 1985 to determine the
tightness of all underground tank systems at LLNL. The definition of a Precision Test, as stated in
National Fire Protection Association Code 329, is "any test that takes into consideration the temperature coefficient of expansion of the product being tested as related to any temperature change during
the test and is capable of detecting a loss of 0.05 gal (190 mL) per hour" and "should account for all
the variables which will affect the determination of the leak rate." These variables include not only
temperature changes of the product during the test but also pressure changes, tank end deflection,
evaporation, vapor pockets in the tank or piping, and operator error or misuse of the test equipment.
The tests performed on the tanks and sumps at LLNL use volumetric leak detection methods. They
identify a leak or determine leak rate based on the measurement of properties associated with a change
in volume. Any factor involved in testing a tank that may mask a change in volume must be strictly
accounted for during the actual test. The most important of these is temperature. Temperature effects
are determined by the material's coefficient of thermal expansion. Gasoline has a relatively large
coefficient of thermal expansion of .0007 per degree Fahrenheit. A reduction in temperature of 0.05 F
reduces the volume of gasoline in a 1500 gallon tank by 0.05 gallons. During a tank test this would
register as a loss equal to the presently allowed limit of uncertainty per hour. The actual average