PIPELINE ACOUSTIC PRE-WARNING SYSTEM AND METHOD OF SIGNAL ANALYSIS

This paper describes a new acoustic pre-warning system for pipelines, aimed at preventing third party damage by monitoring the pipeline acoustic signals. Many environmental factors, such as the by-passing of vehicles and pedestrians, could introduce background noise into long distance transmission of pipeline acoustic signals. As a result, normal pipeline acoustic pre-warning system is disturbed to identify abnormal events. In this work, statistical methods were applied to signal analysis in order to extract feature parameters of different events. Then the optimal feature subset was obtained by gene arithmetic to differentiate hazardous events between normal events effectively. PetroChina has applied the new pre-warning system to their long distance transmission pipelines and the system operates well.