Current Sensor Fault Diagnosis For Position-sensorless Permanent Magnet Synchronous Motor Drive System

Permanent magnet synchronous motor(PMSM)has many significant advantages,comparing with asynchronous inductor motor,such as high efficiency,low losses,high power factor,etc.,PMSM has been widely used in aerospace,industrial production,rail transportation and other fields.In order to realize the closed-loop control of the PMSM,it is necessary to install a mechanical encoder and at least two current sensors to provide feedback signals for the control system.The sensors are kinds of susceptive elements in the motor drive system,which are susceptible to failures due to factors such as electromagnetic interference and sudden load changes.The sensors failure will cause the lack or deviation of the measurement signal,which will deteriorate the control performance of the system,and even cause the secondary failure of the motor drive system.In order to eliminate the adverse effects caused by the failure of the mechanical encoder,the sensorless control for PMSM is researched by academia all over the word.In order to enhance the system reliability under the sensorless control,enable the control system to make a timely and targeted respond when different types of current sensor faults occurred,and provide accurate information for maintenance,the fault diagnosis and accurate type identification of current sensor faults are mandatory.In this dissertation,the three-level T-type inverter-fed permanent magnet synchronous motor drive system is taken as the research object,focusing on the space vector pulse width modulation(SVPWM)of the T-type inverter,the sensorless control method for PMSM and the current sensor fault diagnosis algorithms.The specific research contents are summarized as follows:Firstly,the topology of the three-level T-type inverter and the current flow path under different output states are analyzed.The modulation principle and specific implementation of the SVPWM are summarized.In order to balance dc-link capacitor voltages,the algorithm adjusting the action time of positive and negative small vectors is adopted.The PMSM model under synchronous rotating coordinate system is established,and the vector control model of PMSM is built.An extended state observer is designed to track the speed and rotor position of the position-sensorless PMSM.Secondly,the mathematical models of four typical current sensor faults are established,and two types of current sensor fault diagnosis methods are proposed based on the waveform characteristics and function characteristics of the output current signal from fault sensor.(1)Based on the waveform characteristics of the output signal of the fault current sensor,a fault diagnosis algorithm based on the difference operator is proposed,which realizes the online localization and type identification of the fault current sensor.By periodically reconstructing the current residual,the robustness of the algorithm is enhanced and the influence of parameter mismatch and current disturbances are reduced.The fault type identification is realized by calculating the second difference and the cycle integral of current signal.(2)Based on the function characteristics of the output signal of the fault current sensor,the fault localization and type identification algorithm are proposed for the current sensor based on the differentiation operato.The influence of the speed on the diagnosis accuracy is weakened by changing the reconstruction method of the current residual.The rapid localization for broken-line fault and stuck fault of current sensor is realized by using the differentiation operator and current residual together.Finally,based on the dSPACE MicroLABbox controller,an experimental platform of PMSM drive system fed by three-level T-type inverter is built to verify the effectiveness and feasibility of the above control method and diagnosis algorithm.