Multi-Level Cascaded H-Bridge Inverter-Permanent Magnet Synchronous Motor Drive System Model Predictive Control

The multi-level cascaded H-bridge inverter-permanent magnet synchronous motor system has been widely used in the field of high-voltage and high-power AC speed regulation due to its advantages of high output power,high control precision and high inverter efficiency.At present,a classic control method is to combine the vector control method on the motor side with the modulation strategy on the multi-level inverter side.However,this method has inherent limitations such as complicated PI controller coefficient setting and continuous improvement of current inner loop dynamic performance.Therefore,it is urgent to explore an advanced control strategy.In this paper,the finite set model predictive control strategy is introduced into the multi-level cascaded H-bridge inverter-permanent magnet synchronous motor system.The discrete characteristics of the inverter are related to the motor control target,which has the advantages of no modulator and fast dynamic response.However,due to the large number of multi-level space voltage vectors,the multi-level model predictive control strategy has a large computational problem.Aiming at this problem,an improved multi-level model predictive current control strategy based on current error vector selection is proposed from the perspective of streamlined candidate voltage vector set.By analyzing the relationship between the current vector and the voltage vector increment,and based on the error between the predicted current vector and the reference current vector under the optimal control voltage of the previous control period,the candidate voltage vector set under steady state operation of the motor is established;And the hysteresis loop is introduced in the current loop.When the amplitude of the current error vector is large,the spatial constraint range of the candidate voltage vector is expanded to further optimize the candidate voltage vector set under the dynamic operation of the motor.For a cascaded H-bridge inverter of any level,this strategy can reduce the number of candidate voltage vectors involved to two in each control cycle,which avoids all voltage vectors participating in the prediction model and value function finding.Excellent calculation.Therefore,the computing burden is effectively reduced,and the motor system has good dynamic performance.The improved multi-level MPC strategy and the existing model predictive control strategy are compared experimentally on the multi-level cascaded H-bridge inverter-permanent magnet synchronous motor system platform.The experimental results show that the proposed strategy can effectively reduce the calculation time and improve the dynamic response speed of the current while maintaining excellent steady state performance.