Oil Tube Defect Detection Based on Multi-sensor Data Fusion with Self-adaptive Genetic Fuzzy Neural Network ⋆

An oil tube defect detection method based on self-adaptive genetic fuzzy neural network is proposed, and it got the original information by multi-group vortex sensors and leakage magnetic sensors. We made multi-scale wavelet transform and frequency analysis to multi-channels original data and extracted multiattribute parameters from time domain and frequency domain, and then we selected the key attribute parameters that have bigger correlativity with the defect pattern of oil tube among of multi-attribute parameters. The fitness function, genetic operator and encoded mode of self-adaptive genetic algorithm are improved. The oil tube defect pattern is clustering four classes that are crack, etch pits, eccentric wear and unbroken by self-adaptive genetic algorithm. The fuzzy neural network structure is constructed and optimized, and the Levenberg-Marquart optimizing algorithm is used to train the fuzzy neural network. The self-adaptive genetic fuzzy neural network is adopted to make the multi-sensor data fusion to detect the defect pattern of oil tube and those key attribute parameters are used to as input of network. The experimental results show that this method is feasible and effective.