Research On Deadbeat Predictive Control Of Induction Motor For Vehicle

The operating environment of new energy vehicles is complex,which puts forward higher requirements for the vehicle motor control system,such as high dynamic response,low torque ripple,low current harmonic and low switching loss.Traditional direct torque control(DTC)of induction motor is difficult to meet diverse needs.Model predictive torque control(MPTC)predicts the effect of the control variables with good dynamic and steady-state performance and is easy to achieve multi-objective control,but there is still a large optimization space.On the basis of MPTC,deadbeat predictive control(DBPC)is introduced to predict the ideal control quantity according to the desired control effect,and control system performance is optimized.The main research work is as follows:In order to meet the rapid response needs of automotive induction motors,MPTC model and control system with discrete control set are established under the stationary two-phase coordinate system.In MPTC,the rotor flux with lagging response is not involved,the relevant physical quantities on the stator side are directly used,and there is no current loop,and the voltage vector is used to directly predict the torque and stator flux to achieve faster dynamic response and good steady-state performance.Simulate and compare MPTC with traditional DTC to illustrate their advantages and disadvantages in terms of control performance,weight coefficient,candidate set,and real-time verification of the microcontroller.In view of the limitations of MPTC,two improved strategies are proposed and compared:the optimal voltage vector duty cycle optimization strategy based on torque deadbeat control and the torque deadbeat model predictive control strategy.The simulation results show that through the improved strategies,the fixed duty cycle of the MPTC voltage vector can be adjusted,so the number of candidate vectors is indirectly increased,and the motor control performance is effectively improved.However,for the torque deadbeat model predictive control strategy,in addition to the optimization of motor control performance,the candidate set is also simplified,the number of vector traversals is reduced,and the weight coefficient of MPTC is eliminated,which is conducive to reducing the parameters to be tuned.The real-time verification results of the microcontroller show that the algorithm complexity of torque deadbeat model predictive control is effectively reduced.Considering that the MPTC based on torque deadbeat can only adjust the amplitude of the voltage vector,the stator flux deadbeat control is introduced to form three types of torque and flux deadbeat predictive control strategy: DBPC based on flux vector,DBPC based on motor load angle,and DBPC based on coordinate system components.Finally,the simulation results show that through the three control strategies,the ideal voltage vector of arbitrary amplitude and phase angle is predicted and output,and the motor control performance is greatly improved.The real-time verification results of the microcontroller show that the method based on coordinate system components performs better in real-time performance.Constraints cannot be applied on DBPC,which can lead to voltage overmodulation,and conventional overmodulation processing methods are not optimal.Therefore,based on the constrained optimization algorithm,two methods are proposed: the overmodulation processing problem is transformed into a constraint optimization problem,which is solved based on the interior point method;Or the DBPC of torque and flux is directly converted into a constraint optimization problem,the control range is strictly limited,and over-modulation is avoided.Considering the single step linear prediction of DBPC,the optimal performance of a single cycle is pursued at the expense of comprehensive performance,so the DBPC errors of multiple sampling cycles are comprehensively evaluated by cost function,and solved based on the sequence quadratic programming solver.The simulation results show that the problems of DBPC can be solved and the motor control performance is further improved by the DBPC based on constrained optimization algorithm.Compared with MPTC,it is not sensitive to the weight coefficient,and it is not easy to lose control when torque constraint is applied.