Classes of leak detection system

Classes 4-7 are designed to detect a leak after fuel has been released from the system¹⁸. For existing installations comprising single containment (wall or skin) tanks or pipework a leak may occur causing release of fuel to ground. The outcome of the risk assessment will indicate the likely impact of any such leak to ground and this guidance is intended to help to determine the most appropriate class of leak detection to be adopted. The class of leak detection system will be determined by its capability to detect a leak sufficiently early.

An effective leak detection system should provide early warning of a potential problem to enable prompt corrective action to be taken. Careful consideration should be given to the fail-safe features of any leak detection system, to ensure that safety and the surrounding environment remain properly safeguarded in the event of component failure or inadvertent disabling of the system.

Wetstock control systems under classes 6a and 6b are most likely to be carried out via remote monitoring either by a third party service provider or in-house company facility. These systems therefore rely on resources dedicated to leak detection, notification and subsequent protective actions and so reduce the dependency on filling station staff always being fully aware of, and competent in, the operation of on-site leak detection equipment. It should be emphasised however that the use of external parties does not remove the duties and responsibilities on the filling station operator for health, safety and environmental protection.

18 It should be noted that classes 4, 6 and 7 initially detect a 'loss' which may be due to a number of factors. The outcome of follow-up procedures (see 11.7), will then determine whether the loss is actually a leak. This should be noted where the term 'leak detection' is used in this section.

Table 11.2b Leak detection systems ClassDescriptionDetection capability Conditions of use/factors for consideration LeakDetection periodScope 4a

Approved1,3 ATG Dynamic reconciliation

96 l/d (4 l/h equivalent sudden leak)

24 hrs Detects leaks below liquid level in tanks. Leaks in offset fills and suction lines may be detected if the fuel delivery ticket volumes are manually entered into the ATG

−For A classification sites, only an approved class 4a system is acceptable. −Requires a reconciling ATG with interface to a pump controller. −Requires accurate tank calibration. The ATG may create its own tank chart to give more accurate tank volumes. −Ensure reconciliation functionality is activated. −Performance is enhanced if dispenser meters are calibrated to 'strike' (zero) at the time of ATG commissioning. −Requires tanks and lines to be confirmed tight at time of installation/commissioning. −Has to be able to lock out un-probed tanks (e.g. autogas). −Site staff training required on how to access and interpret reports. −The system should be configured to provide a threshold alarm 19,2 l/d over a period of 14 days. −In the event of a threshold alarm the site operator has to determine what proportion of the reported variance is normal and what could be due to a leak. −Suitable ATGs will quantify the causes of normal variance (not related to a physical leak), including delivery variance (by input of 'ticketed delivery') and temperature variance (by continuously calculating the effect of temperature change in the volume of stored fuel). The level of detail and analysis of the causes of variance provided by the ATG will depend on the manufacturer and model type used. −Where the residual variance exceeds the threshold of 19,2 l/d the site operator should commence an investigation accordingly. −Site operators should be adequately trained to understand and interpret warnings, alarms and reports in order to conduct an investigation in a timely manner. −In the absence of such adequately trained site staff, the site operator should engage the services of a suitably competent third party to remotely monitor all warnings and alarms and take appropriate investigative action.

48 l/d (2 l/h equivalent sudden leak)

7 days 19,2 l/d (0,8 l/h equivalent sudden leak)

14 days

ClassDescriptionDetection capability Conditions of use/factors for consideration LeakDetection periodScope 4b Approved1 ATG Statistical quiet period

18 l/d24 hrs

Detects leak in tank below liquid level

−Ensure this functionality is activated. −Designed for 24 hr sites where static tank testing would be disruptive to trading. −The system should notify the operator via an alarm when there have been insufficient inactive periods to perform the test. The operator should then schedule a static tank test in accordance with 4c. −Site staff training required on how to assess report results. −Individual approvals should be checked to confirm suitability for site configuration: −Throughput limitations. −Maximum tank capacity. −Tank manifolds. 4cApproved1 ATG Static tank test9 l/dUp to 6 hrs

Detects leak in tank below liquid level, during the period of the test

−Ensure this functionality is activated. −Where 4b (statistical quiet period) testing is not activated then a static test should be scheduled to run on a weekly basis. −The test should be run at a time when the tank contents level is at its normal maximum operating level. −Check the ATG manufacturer's minimum operating requirements, specific to the model of ATG, to perform a valid test. This will cover criteria such as: −Test duration (typically 6sixhrs). −Minimum tank contents (typically 40 % of capacity). −Waiting time after fuel delivery (typically six to eight hrs). −The ATG model along with the in-tank probe type, testing regime and operator procedures should be commensurate with the test leak rate specified. −Full records of all test results to be kept on site. −Individual approvals should be checked to confirm suitability for site configuration: −Maximum tank capacity. −Tank manifolds.

ClassDescriptionDetection capability Conditions of use/factors for consideration LeakDetection periodScope 5 Monitoring well with sensorsIndeterminate

Detects release once it has reached the well. Detects leak in tank below liquid level and anywhere in pipework

−Wells should be installed to a depth exceeding the water table range. −The monitoring well should be positioned around the installation to ensure any leakage can find a path to the well. −The type of sensor (liquid or vapour) should be appropriate for the prevailing ground conditions. −Liquid sensors should be hydrocarbon discriminating. 6a

Approved1 statistical inventory reconciliation (SIR) system with weekly analysis and enhanced 'real time' analysis 9 l/h14 daysDetects leak in tank below liquid level and anywhere in pipework

−Data have to be recorded and submitted on a daily basis. −Requires accurate data. −Site operators have a daily responsibility to notice and act upon extreme losses (or gains). −Real time analysis should involve −the recording of nozzle sales data for every transaction in conjunction with corresponding tank stock levels −diagnostic tools to pinpoint the source of the leak −Where dispensers are fitted with automatic temperature compensation or where Stage 2 vapour recovery is activated the site operator and SIR service provider should take into account a change in wetstock variance trend.

96 l/d 24 hrs 6b

Approved1 SIR system with weekly analysis (this replaces monthly analysis)

9 l/h14 days

Detects leaks below liquid level in tanks or pipework

−Data have to be recorded daily and processed monthly. −Requires accurate data. −Site operators have a daily responsibility to notice and act upon extreme losses (or gains). −Where dispensers are fitted with automatic temperature compensation or where Stage 2 vapour recovery is activated the site operator and SIR service provider should take into account a change in wetstock variance trend.

ClassDescriptionDetection capability Conditions of use/factors for consideration LeakDetection periodScope 6c

Daily inventory monitoring Involving statistical trend analysis2

See 'conditions of use/ factors for consideration' Detects leaks below the liquid level in tanks or pipework

−Requires suitably experienced staff on site or remotely based. −Requires accurate data. −The 'normal operating variance trend' and detectable leak rate should be determined and declared for each tank by the site operator −This method would require the use of spreadsheets and charting tools and the comparison of year on year data. 7a

Approved1 electronic pressure line leak detection (ELLD) on submersible turbine pump (STP) systems Three levels of test can be configured to perform testing at various l/h leak rates: 11,4 l/h (equivalent 280 l/d), 0,76 l/h (18 l/d) and 0,38 l/h (9 l/d) From the start of a line test, typically around 1 minute for gross testing (11,4 l/h) and 30 to 90 minutes for precision testing (0,76 and 0,38 l/h), subject to manufacturer specifications and site conditions Tests pressure system pipework for leaks each time dispensing stops. Automatically shuts down STP if a line test fails. Requires a 'passed test' to restore operation following a failed test.

−Requires sufficient STP idle time (periods of dispenser inactivity) for tests to run to completion. For 0,76 l/h and 0,38 l/h precision tests to be completed there need to be periods of between 30 minutes and 90 minutes of dispenser inactivity. Site operators need to be aware that on busy sites these tests may be less frequent. −Systems should provide a report detailing the recent precision test frequencies and results. This will help the site operator determine whether adequate precision testing is taking place under normal operating conditions. −Precision testing should be set to run continuously to ensure maximum completed tests. −During periods of continuous dispenser use precision testing will not take place, but gross resting (11,4 l/h) will occur more frequently, between dispenses. −Some systems may require lines to be independently confirmed as tight during system commissioning. −Under dispenser shear valves should be fitted. −Forms an integral part of the STP control system. −Systems can alert operators if insufficient time has been available for scheduled line tests, allowing the opportunity to create an idle period. −Some systems can also shut down the STP in the event of low fuel level or water presence, preventing STP damage. −Whilst ELLD tests lines from the STP check valve to the dispenser solenoid, consideration should be given to detecting leaks in the system outside this by the use of tank chamber and dispenser sump monitoring.

ClassDescriptionDetection capability Conditions of use/factors for consideration LeakDetection periodScope 7b Mechanical pressure line leak detection (MLLD)

11,4 l/h (equivalent 280 l/d) Typically less than 24 hrs

Tests pressure system pipework for leaks each time the pump starts, but does not shut down associated pump. Dispensing can continue at reduced flow of 11.4 l/m

−Requires lines to be confirmed as tight at time of installation/ commissioning. −Under dispenser shear valves have to be fitted. −Requires sufficient periods of pump idle time, which can make the system less effective on a busy site. −During periods of continuous dispenser use this leak detection system will not operate. −MLLD units should be checked regularly to ensure reliable operation in the event of a leak. −A leak of < 11,4 l/h could continue undetected indefinitely. −Detecting leaks quickly depends on the site operator being alert to slow flow conditions. Notes: 1. An approved system is one which has satisfactorily demonstrated its capability of achieving the specified standard of leak detection through accreditation by an independent third party to a recognised standard such as from CEN or EPA. Where no approved systems are available, the system functionality and performance standard should be verified by the supplier to the satisfaction of the regulatory authority. 2. Includes non-approved statistical based methods and methods using cumulative variance trend assessment usually performed by the site operator. 3. For Class 4a systems (ATG Dynamic reconciliation), where no approved systems are available, the ATG should have demonstrated its capability of achieving adequate level measurement accuracy through accreditation by an independent third party to a recognised level measurement standard such as from EN13352 Specification for the performance of automatic tank contents gauges or OIML R 85-3 Automatic level gauges for measuring the level of liquid in stationary storage tanks.

11.3 RISK ASSESSMENT BASED SITE CLASSIFICATION