Leak Detection for Gas and Liquid Pipelines by Transient Modeling

摘要

The leakage of hydrocarbon products from a pipelinernrepresents not only the loss of natural resources, but also is arnserious and dangerous environment pollution and potential firerndisaster. So quick awareness and accurately location of thernleak event are important to cut down the losses and avoid therndisasters.rnA leak detection method using transient modeling isrnintroduced in this paper. This method is suitable for both gasrnand liquid pipelines with comprehensive consideration of therntransient flow features of compressible flows and stochasticrnprocessing and noise filtering of the meter readings. Therncorrelations for diagnosing the leak location and amount arernderived based on the online real time observation and thernreadings of pressure, temperature, and flow rate at both endsrnof the pipeline. As an online real time system, great effortsrnhave been paid to the stochastic processing and noise filteringrnof the meter readings and the models to reduce the impact ofrnsignal noise. It is essential too for the robust real time pipelinernobserver to have the self study and adjustment abilities inrnresponse to the large varieties of pipeline configuration,rnpipeline operation conditions, and fluid properties.rnReal application cases are presented here to demonstraternthis leak detection method. For example, in the leak detectionrnof a crude oil pipeline of 34.5 km and Φ219mm, this methodrnlocated the leak at 16.6 km from the pipeline upstream endrnwhich is only 0.6 km away from the actual leak location.