Underwater gas pipeline leakage source localization by distributed fiber-optic sensing based on particle swarm optimization tuning of the support vector machine.

Abstract

Accurate underwater gas pipeline leak localization requires particular attention due to the sensitivity of environmental conditions. Experiments were performed to analyze the localization performance of a distributed optical fiber sensing system based on the hybrid Sagnac and Mach-Zehnder interferometer. The traditional null frequency location method does not easily allow accurate location of the leakage points. To improve the positioning accuracy, the particle swarm optimization algorithm (PSO) tuning of the support vector machine (SVM) was used to predict the leakage points based on gathered leakage data. The PSO is able to optimize the SVM parameters. For the 10km range chosen, the results show the PSO-SVM average absolute error of the leakage points predicted is 66m. The prediction accuracy of leakage points is 98.25% by PSO tuning of the SVM processing. For 20 leakage test data points, the average absolute error of leakage point location is 124.8m. The leakage position predicted by the PSO algorithm after optimization of the parameters is more accurate.