Fault Diagnosis of Traction Machine for Lifts Based on Wavelet Packet Algorithm and RBF Neural Network

Considering about the fault features of traction machine for lifts, the basic characteristics of faults types are analyzed. By detecting vibration signals from vibration sensors, uses wavelet packet to decompose fault signal, extracts the signal characteristics of 8 frequency components from the low-frequency to high frequency in the third layer. The 8 obtained eigenvalues as the fault signals are extracted into radial basis function (RBF) artificial neural network. Since Particle Swarm Optimization (PSO) algorithm can improve the efficiency in finding the optimal weights for the RBF neural network, we use the RBF neural network optimized by PSO algorithm to set up the fault diagnosis model. The experimental result shows that the proposed technique is succeeded in diagnosing and locating faults effectively.