Wavelet Neural Network Based Intelligent System for Oil Pipeline Defect Characterization

Abstract

Wavelet neural network is a new kind of network which fuses advantages of wavelet transform and neural computing. It utilizes the good localize character of the wavelet transformation and combines the self learning function of the neural network. It has the ability of strong adaptive learning and function approach. Wavelet neural network has the simple implementation process and fast convergence rate, therefore it can be used to detect the defect of oil pipe. This paper presents a wavelet neural network approach for detection and characterization of defects from magnetic flux leakage signal.