Study Of Fault Detection For Rectifier Devices

With the rapid development of power electronics and continual emerging of novel power electronic devices, power electronic equipments have been increasingly applied to all aspects of industry and living.The requirement of its maintainability is also more and more important. Because there are many devices in the main circuits of high voltage and strong current systems, one will take much time and make many efforts to deal with it using normal fault diagnosis methods.Aiming at the own characteristics of fault, new technologies such as artificial intelligence, neural network, fuzzy algebra, wavelet analysis, etc have been successfully used to fault diagnosis of power electronics. In addition, study about fault prediction of electronic equipment is less, while it is propitious to determine the location of fault in advance and shorten the downtime of power electronic circuit, which can improve efficiency and reduce losses.In this paper, fault conditions of three-phase rectifier circuit are analyzed and fault diagnosis method based on residuals principle is proposed; simultaneously, the internal structure of switching device is analyzed and the variation rule of device aging to damage is found, based on which the prediction of switching devices failure is achieved. Major work and conclusions are as follows:(1) The stuck-open faults in three-phase controlled bridge rectifier is showed by example, fault conditions in the main circuit are analyzed. Various fault type are possibly considered in simulation analysis, possibly accorded with the actual work, fault waveforms are analyzed to obtain various fault conditions, based on which fault classification can be achieved.a. The fault type which can be categorized mainly includes four categories, the first category is 0°control angle; the second category is 0°～60°control angle; the third category is 60°control angle; the fourth category is control angle greater than 60°.b. Under a certain control angle, when the fault with same type occurs in different element, the fault voltage waveform only moves along time axis, while the wave shape is invariant.c. Fault voltage waveform changes with the change of conduction angle, which will cause trouble in fault feature extraction. Therefore, the influence of conduction angle change should be minimized in fault diagnosis.(2) Fault diagnosis of rectifiers based on residual method is realized and encoding method is proposed. Fault device can be located by querying encoding.(3) In harsh working conditions, the turn on-threshold voltage of switching device drifts, utilizing the characteristic and signal processing methods, fault prediction and fault localization of switching devices are realized.