More and more, the on-line (or real-time) pipeline software market is being driven by the need for model based leak detection systems. Unfortunately, the instrumentation requirements of such leak detection systems are not easily determined by pipeline operating companies. Consequently, the design of pipeline instrumentation seldom takes leak detection into account. In particular, intuition leads one to believe that the more pressure instrumentation that is installed on a pipeline, the more sensitive leak detection will become. This article will show that in some cases intuition can be a poor guide. In a recent modeling study, leak sensitivity improvements from additional pressure measurements installed on a natural gas pipeline were analyzed. Several cases were considered where key parameters affecting the sensitivity of the leak detection software were varied. The resulting minimum detectable leak sizes, times required to detect the leak, and the leak locating and sizing accuracies were determined. The results of these cases and some general guidelines for pipeline instrumentation are presented.
The HBJ (Hazira-Bijiapur-Jagdishpur) pipeline system carries natural gas from the South Bassein field in western off-shore India to the interior of India. The system contains 1700 kilometers of pipe with diameters ranging from 36 inches to 12 inches. Four compressor stations along the pipeline provide enough energy to meet high pressure requirements at pipeline delivery points. The HBJ system will provide a vital function in the overall development of India's natural gas industry. The gas from the pipeline will provide a feedstock to six fertilizer plants and fuel to several power generating plants. One of the primary concerns of GAIL (Gas Authority of India Limited) is the safe operation of the pipeline, in particular leak detection sensitivity. GAIL, the pipeline owner and operator, commissioned a study by SSI (Scientific Software Intercomp) to evaluate the relationship between the pipeline's instrumentation and the sensitivity'of the leak detection system. In particular, possible leak sensitivity improvements from additional pressure measurements were analyzed. The performance of a leak detection system is highly dependent on the pipeline's instrumentation quantity and quality. A great deal of experimentation and analysis is required to accurately determine the effect of specific instrumentation characteristics on the sensitivity of a leak detection system. (Ref. 1) The best way to determine these effects is to conduct a simulation study using a SCADA/pipeline simulator and the leak detection system itself. This was the approach taken by SSI. enabling the reader to understand the procedures and results of the study. Four parameters are discussed which form the performance criteria for measuring the sensitivity of a leak detection system. Then the procedures used to conduct the leak detection sensitivity study are given. The pipeline configurations and operating conditions are provided. The test environment used to conduct the leak simulations is described and illustrated. Subsequently, the leak detection study results are presented and discussed. These results consist of the effect of additional pressure measurements on minimum detectable leak size, detection time, leak location and size estimates.