Research On Model Predictive Control Of AC Drive System Fed By Indirect Matrix Converter

Indirect Matrix Converter(IMC)can realize sinusoidal input/output currents,an adjustable input power factor,energy bidirectional flow and high power density,which is very suitable for AC motor drive system,and it can also be extended to feed multi-motor drives.In the AC drive system fed by indirect matrix converter,the performance requirements of both the grid and the motor should be considered.In this paper,the model predictive control(MPC)method with the simple structure and idea is used to study the control of single motor or dual motor drive system to improve the performance.Traditional single vector model predictive control is used in single motor drive system fed by IMC.Complex multi-step commutation is needed in the commutation of bi-directional switchs,which will lead to distortion of current waveform and decrease of system reliability.The error between output voltage and reference voltage is larger when the system runs at medium and low speed,which will cause the fluctuation of motor torque and flux linkage.In this paper,two model predictive control methods based on dual-vector are proposed.Among them,the fixed duty ratio dual-vector MPC method can realize zero-current commutation,shorten commutation time,improve the safety and reliability of the system.The consideration and evaluation of the commutation process can make the predictive model more accurate.The variable duty ratio dual-vector MPC method,through theoretical deduction,the optimal operating time and corresponding switching state combination are determined which improve system performance,especially in low and medium speed operation.In the dual motor drive system fed by IMC,since the large number of switching states and the complexity of predictive model,the large computational complexity of MPC algorithm.A simplified model predictive voltage control strategy(MPVC)is proposed,which employs input currents and output voltages as control objectives,decreases the number of switching states and optimizes algorithm program.The purpose of reducing computational complexity is realized,thus the control frequency and switching frequency of the system are increased to improve system performance.The control method proposed in this paper is tested on the prototype of drive system fed by IMC.The results show that in the single motor drive system,the two-vector MPC method has smaller fluctuation of motor torque and less distortion of the grid current and the motor current than the single-vector MPC method.In the dual motor drive system,the whole calculation period by using the conventional MPFC is 2.5 times than the simplified MPVC,and the system performance of the simplified MPVC method is improved significantly.