Pipeline Leak Detection Using Particle Filters

While most of the available Leak Detection Systems (LDS) can detect pipeline leaks, leak localization is still an unresolved problem. The main reason for this problem is the limited number of sensors installed in long pipelines. Because of lack of measurements, precise leak location cannot be easily determined. This study uses particle filter as soft-sensor to estimate the states at intermediate locations with the available end-point measurements. The residuals between the estimated states and the real measurements at the intermediate locations are awn used for leak detection and localization. The proposed method can improve the leak localization accuracy by proper use of the intermediate pressure measurements. Simulation results show the efficacy of the proposed method. (C) 2015, IFAC (International Federation or Automatic Control) Hosting by Elsevier Ltd. All rights reserved.