Research On Fault Diagnosis Of Current Sensor In Induction Motor Drives

Industrial robots are widely applied into different fields owing to some merits such as efficiency, strongly universal etc. Frequency controlled speed regulating system(FCSRS), as its actuator, plays an important role in robots normal running when keeping reliable performance. Fault occurred would lead to performance degradation, system breakdown, and cause huge economic loss. Therefore, research on improving the stability and fault diagnosis of FCSRS makes a big difference in boosting reliability of industrial robots. Generally, FCSRS adopts vector control to achieve high performance, which needs signal measurement components to provide accurate steady feedback signal. Therefore, to develop a drive system that has the capability to detect fault and tolerate fault could decrease maintenance cost, improve operation safety and bring about economic benefits.With fault diagnosis of FCSRS fed by voltage source inverter as research background, the dissertation studied fault diagnosis aimed at the fault of current sensor in virtue of vector control and extended Kalman Filter algorithm, and further analyzed fault types of current sensor, diagnosis and fault-tolerance. The model of current observer was built through exploiting the characteristics of extended Kalman Filter algorithm like insensitive to model uncertainty, parameters variation. Meanwhile, an ideal motor model was derived from mathematic equations. After combined the two model, control structure possessed an capability to detect fault of current sensor. The proposed scheme could detect the open-circuit fault by computing the residuals between measured variables and estimated variables. Offset fault and gain-change fault could be identified by a slide window effectively.Based on analysis on current distortion degree and performance decrease under three special fault types, the dissertation researched selections of fault feature variables and setting of threshold. The proposed diagnosis algorithm could identify fault types easily and provide corresponding control configure scheme. Its effectiveness and practicability had been validated further by numerous simulation and physical experiment tests. Besides, it was robust to external disturbances like variable load and speed.