Low-complexity Model Predictive Control For Linear Induction Machines In Urban Transit

Due to the longitudinal end effect(LEE),the mutual inductance of linear induction machine(LIM)is variable with its working conditions,which deteriorates the operating performance of LIM.In order to improve the performance of LIM,the model predictive control(MPC)is applied to LIM,and it can consider the influence of LEE by combining with the equivalent circuit of LIM.In order to improve its practicability,the MPC is improved by reducing its complexity in this paper,so that it can meet with the requirement of LIM drive systems,and the main contributions are listed below.(1)In the urban transportation applications,the field oriented control(FOC)and direct thrust control(DTC)are the common used control methods for LIM.However,these control methods cannot consider the variable mutual inductance caused by LEE so that it may cause the parameter mismatch,which may bring the imperfect control performance.So,in order to compensate the LEE,both model predictive current control(MPCC)and model predictive thrust control(MPTC)are applied to LIM,and it can avoid the papameter mismatch by combining with the equivalent circuit of LIM.Compared with FOC,the results verify that the current tracking errors of MPCC can be reduced under the same switching frequency,and less influenced by the working conditions of LIM.On the other hand,compared with DTC,the MPTC can bring both flux and thrust ripples reduction.(2)Because the MPC has the disadvantage of heavy computational burde,the complexity of MPC is reduced in this paper by judging the sector of reference voltage vector(RVV).With the help of proposed search method,the online calculation time of MPCC with different modulation strategies is almost same to that of FOC,which can effectively reduce the hardware cost.And the simplified multistep finite control set MPC can remove more than 99.00% of candidate voltage vectors in advance by combining with the branch and bound method.The simplified method for MPTC with arbitrary two voltage vectors can directly find the optimal voltage vector combination,which is more effective than the previous simplified method.Besides,the simplified method can be further applied to the case of with current limit,which can avoid the overcurrent in dynamic process and ensure the safe operation of LIM.Therefore,the simplified MPC is low complexity in practical applications,which is more popular in LIM drive systems.(3)In order to further improve the performance of MPCC,the MPCC is combined with the voltage vector modulation strategies in this paper,and the current tracking errors can be reduced by incaresing the accuracy of voltage vector modulation methods.The performance of MPCC with different modulation strategies are compared and analyzed in both simulation and experimeants,and the results indicates that the current tracking errors and THD can be reduced by improvinvg the modulation accuracy,but it increases the switching frequency.In order to increase the switching loss of inverter as little as possible,the performance of continuous and finite control set MPC is improved in this paper by increasing the predictive steps.Finally,both simulation and experiments demonstrate that the continuous control set MPC can reduce the current tracking errors by increasing its predictive steps when the weighting factors are same,but different predietive steps have little influence to its switching frequency.On the other hand,when the switching frequency is controlled as almost same,the finite control set MPC can further reduce the current THD by increasing its predictive steps.(4)Becaue the performance of MPTC mostly relies on the selection of weighting factor,two weighting factor less MPTC methods are proposed by rewriting the cost function in this paper.Both simulation and experiments verify that the proposed MPTC methods produce less flux and thrust ripples,and effectively avoid the complex process of tunning weighting factor,which can improve the convenience of MPTC.Besides,in order to further reduce its flux and thrust ripples,the MPTC is combined with arbitrary two voltage vectors in this paper,and it removes the restrictions on the voltage vectors combinations.Both simulation and experiments demonstrated that MPTC with aribtary two voltage vectors can reduce both flux and thrust ripples,compared with the modulatrion methods of fixed voltage vector combination.And different working conditions have little influence to its flux and thrust ripples,switching frequency,and current THDs.Compared with the traditional control strategies,the MPC adopts the online optimization methods so that it can achieve the optimal control performance.Besides,in order to improve its feasibility,the computational burden of MPC can be reduced by the simplified method.Finally,both simulation and experiments demonstrate that the MPC can further improve the operating performance of LIM,which can find greatly potential application in urban transportation.