beautypg.com

Franklin Fueling Systems T5 Tank Sentinel Programming Guide User Manual

Page 37

background image

33

Tank Leak Test Results

Leak test results are either

increase, passed, failed, or indeterminate.

Reasons Why Tank Leak Tests Fail

The tank leaks.
Temperature instability – temperature variations of the product within the tank after a delivery is the most common

source of interference and failed leak tests / false alarms (a failure to detect real leaks can also result). Look at the

hourly temperature data on the leak test report and retest if the variation is more than a few tenths of a degree.

Large changes in product temperature from the start to the end of the test.
Water Level changes from the start to the end of the test.
Evaporation and loss of product through the vent stack will look exactly like a leak. Evaporation can be a problem

during high seasonal temperatures and/or high winds, and when liquid level exposes the greatest surface area for

evaporation (half full tanks are worse). Seasonal variations in product composition, size of the tank (larger is worse),

and tank vent configuration are also evaporation factors.

Tank Deformation -- the tank changes shape after a large product delivery.
Tidal Action – in coastal communities, groundwater levels maybe be affected, and, as a result, may deform the tank.

During certain times of the day, the tide changes inconsistently.

Tank Cross-Talk – level changes in one tank causes a level change in an adjacent tank or compartment.
Product is being dispensed during a test.

Steps to Take When a Tank Leak Test Fails

Leak tests do occasionally fail and a single failed leak test should not be a cause for great concern. Remember, if the

confidence level of the test is 99%, there is a 1% chance that the system will give a false answer.

Caution

Don’t excavate/repair a tank because of a single failed leak test. Reference your site policy and

procedures plan.

If you fail a leak test, review the leak test report to determine if there is an obvious source of interference with the test (see

the Causes of Leak Test Failures section for an example of a Tank Leak Test Report). If such a source of interference is

identified, retest the tank as soon as possible. Retest the tank to validate or invalidate the first test result if no source of

interference can be identified. If necessary, run several tests.
If repeated tests indicate a leak and no obvious source of interference is found, then immediately have the tank

precision/pressure tested. If the precision test confirms the presence of a leak, then the tank owner must take

corrective action in accordance with federal, state and local regulations.

Warning

The owner of the tank is legally obligated to comply with reporting and procedural requirements

of federal, state and local regulations. These must be followed explicitly. Serious legal, health

and safety hazards could result from not taking immediate and proper action. For instances

when codes and regulations conflict with this manual, follow the regulations set by governing

agencies.

SCALD Tank Leak Tests

SCALD (Statistical Continuous Automatic Leak Detection) is optional on Franklin Fueling Systems tank gauges. It runs

24hrs a day and is used to perform 0.2 gph tests on tanks that are always in use.
SCALD works by collecting Quiet Intervals (QIs) in between dispensing. A QI is obtained when a thermally stable tank

is idle for 20 minutes with no dispensing, no deliveries and no other movement of the probe floats. Once four QIs are

collected, the Automatic Tank Gauge (ATG) will analyze the data and either Pass, Fail, Increase or discard that test

because the data is not statistically sound. Four QIs can be collected over a period of several days or weeks.

• If the result is a Pass, then a test result of Pass will be generated.
• If the result is Fail or Increase, then another test will be run to ensure that this is not a false alarm. Three Failed or

Increase tests in a row will produce a test result of Failed or Increase.

• If the data is not statistically sound, then the test is discarded. No new test result will be generated and SCALD

will continue to run. The fact that a result was discarded is recorded and can be seen in the Status number that

accompanies the next generated test result.

• On rare occasions certain conditions can arise that can prevent SCALD from getting enough QIs to complete a test.