Research Of The Fault Diagnosis For The Low Voltage Electrical Connectors

In communication systems, various kinds of electrical connectors are the most vulnerable components. When faults happen in a connector, the signal that is passing through it will be changed and thus the whole system will be affected. Also, the faults in electrical connectors take the main responsibility for the low quality of communication line and bit error rate. So it is necessary to research the fault diagnosis technology for electrical connectors.In this paper basic electrical contact theory is explained at first. It briefly demonstrates the cause of electrical contact fault and theories about it. Then wavelet theory and artificial neural network are introduced. It explains the basic concepts and mathematical characteristics of wavelet transform and neural networks. A fault diagnosis algorithm is proposed. To diagnose the fault in electrical connectors, this algorithm I proposed mainly uses wavelet transform to analyze the signal that passed through the fault connectors. And this algorithm was tested and verified in different electronic circuits. Considering that real signals are often mixed with noise signal, in order to improve the performance of fault diagnosis the algorithm of wavelet de-noising is modified with singular value decomposition. Also, artificial neural network is applied to diagnose the fault in a larger scale circuit. And the fault diagnosis using artificial neural networks is simulated in computer. Though the simulation of artificial neural networks is relatively simple, the result of simulation proves that using artificial neural network to diagnose fault in electrical connectors is feasible.