| The permanent magnet synchronous motor(PMSM)drive field widely adopts finite control set model predictive torque control(FCS-MPTC)because the dynamic performance of FCS-MPTC is superior.However,in the application of complex operating conditions and high-performance requirements,FCS-MPTC still has three problems such as being sensitive to parameter mismatch,tedious work of adjusting weighting factor and poor steady-state performance.Therefore,a series of research are carried out:In the FCS-MPTC,the parameter mismatch will cause a prediction error in prediction process,which affects the control performance,and there is a hysteresis in the traditional prediction error compensation mechanism.It only updates the prediction error corresponding to one voltage vector every control cycle.Aiming at this problem,a model predictive torque control with online error compensation based on Adaptive Linear Neuron(Adaline)neural network is proposed.Firstly,the torque and stator flux linkage prediction error models are constructed and analyzed.Then,the Adaline neural network is constructed according to the incremental model of prediction error.Finally,an online compensation mechanism of prediction error based on the Adaline neural network is designed,which improves the prediction model of torque and stator flux linkage.The improved strategy can update the prediction error corresponding to all candidate voltage vectors and compensate online in every control cycle.The simulation and experimental results show that the proposed strategy has the strong parameter robustness.The work of adjusting the FCS-MPTC weighting factor is tedious.Aiming at this problem,a model predictive torque control based on complex power complex power deadbeat(CPD-MPTC)is proposed to avoid the tedious adjusting process.Firstly,according to the relationship between complex power and torque,stator flux linkage,the complex power prediction model and the calculation method of the reference complex power are obtained.The complex power error cost function is constructed.Then,according to the complex power deadbeat,a fast voltage vector optimization method is proposed,which divides the voltage vector plane to 6 sectors.It can reduce the evaluation range of the voltage vector,and the computational burden is saved.Simulation and experimental results show that the proposed strategy can ensure the control performance of torque and stator flux linkage by controlling the input complex power and achieve high-performance control of the motor with less computational burden.The steady-state performance of the FCS-MPTC is poor.Aiming at this problem,based on the CPD-MPTC,a CPD-MPTC with double vector algorithm is obtained to improve the steady-state performance.First,the voltage vector plane is further divided to achieve fast optimization of multiple candidate voltage vectors,and candidate double voltage vectors are obtained.Then,based on the geometry principle of plane,the vector plane is subdivided to obtain the criterion of the optimal double voltage vector combination.And the calculation method of the corresponding action time of each voltage vector is obtained too.The simulation and experimental results show that the proposed strategy can achieve better steady-state performance with low computational burden. |
