INDUSTRIAL VALIDATION AND VERIFICATION APPROACH FOR EXTERNAL FIBER OPTIC BASED LEAK DETECTION

摘要

External Leak Detection systems based on distributed fiber optic sensors (DFOS) offer the exciting potential to significantly reduce the overall amount of spilled product before a leak is detected and localized. Such systems are not well represented by industrial standards or recommended procedures and as a result most industrial attempts to validate the technology have been research-oriented and whilst these have contributed greatly to the knowledge base they have never been aimed at a full validation of the technology. Additionally, the lack of test facilities that can support the significant scale needed for validation (＞500 m straight line run) have led to a paucity of attempts to provide a baseline validation of such sensing technology leading to a lack of certainty over performance claims within the industrial user base and no robust method of testing such claims. With a significant customer base of deployed systems, OptaSense have developed a reproducible technology validation approach using full scale, full flow, representative leaks at the CTDUT test facility in Brazil. We have used these tests to validate 15 lpm leaks, detected and classified via their negative pressure pulse in ～10 seconds and larger 150 lpm leaks, detected by our four modes of leak detection in ～l minute. Valid automated detection of a negative pressure pulse (NPP) was observed down to 1 mm holes in the pipe - representing a leak rate of only 1.5 lpm. The use of the NPP is shown to be a compelling rapid detection method. However, care is needed in testing since the use of a valve opening to stimulate a pulse is shown to be significantly inferior to burst disks due to the increased valve-opening time giving rise to a reduced amplitude pressure pulse. The conventional external leak detection signals of Orifice Noise, Ground Strain and Temperature Change can all be shown to be replicated at the large-scale test facility by these means leading to the potential to establish a valid Probability of Detection for all approaches. With validation now possible, client verification on site has also been addressed with a two-step approach being developed that replicates the validation approach detailed above. Negative Pressure Pulses are used for stand-alone leak detection and can be safely stimulated via accessible valve sites and product release via a burst disc / valve and orifice combination. To stimulate the Multiple Mode behavior (excluding Negative Pressure Pulse) a controlled fluid release injection mechanism has been developed, which can be introduced at an appropriate offset from the pipeline (mirrored from fiber offset) at any desired location with the minimum of preparation. Ground probe deployment techniques have been designed to simulate a leak event at the appropriate location resulting in the similar external signals arising on the fiber. This paper presents the benefits of large-scale validation approaches to performance bound acoustic-based leak detection systems and presents established options for in-field verification on customer owned systems.