Leak Detection in Gas Pipelines - a Knowledge Based Approach

Leak detection systems are a fundamental part of the management system associated with any long distance pipeline. Such systems are necessary to avoid or reduce environmental and economical losses in case a leak develops in the pipeline. There are several solutions, with different efficiency and cost-effectiveness, for this problem. One of the most appealing is the use of software based leak detection systems (SBLDS). These systems use already available data from the Supervisory Control and Data Acquisition System to detect and locate leakages in real-time. This solution is appealing due to its cost-effectiveness and real-time operation. Nevertheless, this technique relies heavily on the sensing equipment already installed on the pipeline. Noise, drift, low sampling rates, low sensor resolution, and other aspects have a negative influence on system performance. Most of the schemata devised for SBLDS uses a lot of statistical analysis and filtering to cope with these problems. But a lot of useful information is not taken into account in the development of these systems. This information is available as operational knowledge, pipe reliability and risk analysis information. Such background knowledge can be used to improve a SBLDS performance. This article presents a software based leak detection method and proposes an implementation. The method is an extension of dynamic model based leak detection with operational and risk analysis knowledge, using expert system techniques.