Acoustic leak detection
To provide guidance for use during inspection of hydrocarbon gas detection systems and for assessment of hydrocarbon leak detection aspects of safety cases. The guidance in this SPC is in addition to the current assessment guidance and APOSC. Generic guidance on aspects of fire and gas detection systems is available in International Standard ISO 13702.
1 Addressees should consider the contents of this SPC when undertaking inspections of gas detection systems, the assessment of safety cases and proposals for the implementation of Acoustic Leak Detection (ALD). Inspection focal points may wish to send this to their duty holders for information. Feedback on issues raised in this SPC as a result of inspections would be welcome.
2 ALD has been installed and is being considered for installation on platforms and rigs in the Southern North Sea. It has been shown to have some advantages and benefits over both fixed point and open path gas detection. However, performance data for ALD is incomplete and the technique has some limitations, which should be considered carefully before proceeding with its implementation.
3 ALD has been installed offshore on normally unattended installations (NUIs), manned platform complexes, open pipe bridges and open areas during combined operations in the Northern and Southern North Sea. The technology is becoming more widespread across the sector and, where sufficient operational confidence has been achieved of ALD has been fully integrated into alarm/emergency shutdown (ESD) systems. Research is continuing in order to provide a comprehensive design basis for ALD.
4 Fixed point catalytic / IR or open path gas detection, is intended primarily for the detection of accumulations of flammable gas in enclosed areas. It has limitations in open or semi-enclosed well ventilated areas. The detection process relies upon:
- The released gas reaching the detector: and
- The gas concentration at the detector being sufficient to trigger an alarm.
5 Field trials have indicated that gas plumes in well ventilated area may not be detected by fixed point, or open path detectors. Analysis of hydrocarbon release incidents shows that conventional gas detectors have 65% detection efficiency. The potential therefore exists for gas from undetected leaks to disperse to adjacent areas, accumulate in ventilation dead spots in the modules, or for the release to continue undetected for some time (eg on NUIs or in high noise areas).
6 ALD detects the leak not the gas plume. It does not require the released gas to disperse to the detector head but detects the ultrasonic noise generated at the leak source itself (any gas release whose source pressure is above approximately 2 bars will generate such noise) ie it is independent of ventilation rates. Therefore it has the potential to provide early warning of a hydrocarbon release and to fill in gaps in the effectiveness of traditional gas detection systems. Some limitations of ALD are described below.
Regulations and guidance
7 Safety cases should contain a description of the design philosophy and overall operation of gas detection, alarm and ESD systems in order to assist in the assessment of its contribution to risk reduction. Performance standards for gas detection systems should be based upon the operating environment, taking into account the level of confinement, ventilation and alarm/ESD actions required.
8 The Offshore Installations (Safety Case) Regulations 2005 (SCR05) (SI 2005/3117) Schedule 2 para 11 requires "A description of arrangements for protecting persons on the installation from toxic gas……."
9 The Offshore Installations (Prevention of Fire and Explosion, and Emergency Response) Regulations 1995 (PFEER) Reg. 5(2)(c) requires:
"The establishment of appropriate standards of performance to be attained by anything provided by measures for -
i) ensuring effective evacuation escape recovery and rescue to avoid or minimise a major accident..."
10 In this respect a performance standard should be developed for the ALD system in relation to reliability, availability and the ability to detect a leak of a given minimum size and pressure.
11 PFEER Reg. 10 requires that:
"the duty holder shall take appropriate measures -
With a view to detecting fire and other events which may require emergency response, including the provision of means for -
i) detecting and recording accumulations of flammable or toxic gases
ii) identifying leakages of flammable liquids, and
With a view to enabling information regarding such incidents to be conveyed forthwith to places from which control action can be instigated"
12 Where there is a need to detect accumulations of gas (eg in enclosed areas) ALD alone will not be sufficient or appropriate and other additional means of detection should be identified.
13 PFEER Reg. 19(2)(b) requires that there is a suitable written scheme for the systematic examination by a competent and independent person of plant…provided:
"for detecting and recording accumulations of flammable gases..."
14 SCR05 (Reg. 2) also defines safety-critical elements (SCEs) and requires that a suitable written scheme be developed for ensuring they will be suitable and remain in good repair and condition. ALD systems will usually be defined as SCEs.
Principle of operation
15 In the past onshore process industries have identified gas leaks by attaching microphone probes to process pipework. This method is limited by noise interference transmitted through the pipework system and flow generated noise. At present offshore, the main interest is in the external detection of noise generated at the leak source.
16 The effectiveness of ALD has been improved by the use of non-contact transducers (microphones), which detect ultrasonic noise at frequencies above 24 kHz. Frequencies below approximately 24 kHz are removed by a high pass filter. This development has effectively eliminated audible background noise interference, which in the past had made the technique unreliable.
Some ultrasonic noise interference will remain arising from PSVs, pneumatic actuators, generator cooling fans, leaks in instrument air systems etc. Adjustment of an attenuation setting on the detector allows the remaining background interference to be "backed-off". The detector will then respond only to noise whose sound power level is above the background level.
In order to determine the level of the background interference a "mapping ultrasonic sound level meter" must be used to characterise the area before installation of ALD. The required attenuation is determined through this mapping process. It is essential that the person carrying out this mapping process is fully aware of the principles of operation of ALD to ensure that the attenuation is not set too low (unwanted alarms may result) or too high (potential leaks may be missed).
A gas leak will normally be long term and continuous (ie over several minutes), so short timescale ultrasonic noise can reasonably be assumed to be spurious. Short timescale, intermittent ultrasonic noise might be caused by, for example PSVs (assuming they do not stick open, or the pressure excursion is sufficient to hold them open for longer than a few seconds), control valves, or "pinging" of flapping valve tags. Such interference can be eliminated by setting a time delay to ensure that only continuous noise will cause an alarm. The delay can be set on the detector head or on the control panel.
17 Operational issues are related to optimising the time delay for the detector location and workforce education concerning operation of ALD.
18 ALD responds to the source of the hydrocarbon release, not the effect of the release ie dispersed gas. There is therefore potential for very rapid alarm and response (however, time delay settings will have an influence on this).
19 ALD can detect ultrasonic noise generated from gas releases in open, well ventilated areas where traditional gas detection may not be wholly effective or independent of ventilation.
20 A single detector can cover a relatively large area of process plant or pipework - up to 10 m radius. Depending upon the criticality of the alarm or trip actions required, two detectors should be considered at a single location to provide for voting.
21 Current operating experience suggests that the transducers are robust and do not suffer from significant calibration drift. They appear to be able to withstand significant contamination and require little maintenance.
22 Current developments include EX enclosure rating, analogue output and a self test facility.
23 ALD does not give an indication of gas concentration. It cannot detect accumulations of gas in enclosed areas and provides only an alarm/trip. Also, following cessation of a leak, ALD would not give an indication of gas still present within the area. Hence it should be used alongside conventional gas detection in such circumstances.
24 ALD is susceptible to interference from continuous and short timescale background ultrasonic noise. The sensitivity must therefore be attenuated in terms of detection threshold and response time in order to prevent spurious alarms. The effect of this de-sensitisation on detector effectiveness is subject to ongoing research.
25 ALD will detect both toxic and flammable gas releases provided the release source is capable of generating ultrasonic noise ie the source is under pressure. ALD will not respond to low pressure sources of flammable or toxic gas such as vaporisation of de-pressurised liquids eg pools of drilling mud, condensate or stabilised crude.
26 It is not possible to state at the present time how ALD responds to high pressure or flashing liquid releases such as live crude, condensate, two phase or, hydraulic fluid releases. This is subject to ongoing research.
27 The philosophy of integration of ALD into existing gas detection systems, alarm and trip systems is not clear and no consensus exists at present.
28 ALD can offer significant benefits as an addition to existing gas detection systems or, in situations where traditional gas detection may be considered to be unreliable. However, the replacement of existing gas detection equipment with ALD should not be considered where accumulations of gas can occur.
29 Where accumulations cannot occur, installation of ALD as an alternative to hydrocarbon gas detection should be considered only after a suitable and sufficient assessment of the risks has been performed. Where there is an existing hydrocarbon gas detection system, it is recommended that ALD should be considered only as an addition, not a replacement.
30 In situations where traditional gas detection has not already been installed due to high ventilation rates (eg on an open pipe bridge or on the open well ventilated deck of an NUI) ALD can provide an alternative and effective means of leak detection.
31 The effects of integration of ALD alarm into existing alarm/trip systems should be fully assessed. ADL systems currently available can be integrated into conventional fire and gas control systems without difficulty.
32 ALD should only be installed following a full and thorough evaluation of background ultrasonic noise interference. This should include background noise mapping of the area by suitably trained personnel using appropriate equipment.
33 ALD should not be considered for liquid, two phase or condensate releases, at present as its performance has not yet been evaluated.
34 The design criteria for setting the maximum range has not been fully explained. They should not be used with a (long) maximum range, which implies a narrow margin above background noise level.
35 In addition, a hand held ultrasonic leak detector has proved very effective in detecting incipient and small scale leaks on process equipment eg weeping flanges, gland packing and small scale compression type fittings. The detection of such leaks as part of normal plant maintenance and equipment integrity assurance is strongly recommended.
Further Information can be obtained from: OSD 3.2 Fires, Explosion and Risk Assessment 0151 951 3144.