Leakage detection of natural gas pipeline based on neural networks and data fusion

It was important to detect the leaking of natural gas pipeline whether leakage happened, the comprehensive and comparative analysis of various methods of leak detection indicated that most common detection method which was difficult to identify and apply to the natural gas pipeline. In this paper the method was proposed based on RBF neural network and the data fusion of D-S evidence theory for detecting the pipeline leak. Extracted neural network´s input parameter through wavelet denoising, then put the parameter to neural network and calculated by multi-sensor data fusion algorithm so as to acquire leaking information.