| The low frequency oscillation(LFO)over-voltage problem in vehicle-grid coupling system problem has occurred frequently,seriously affecting the safe and stable operation of the railway.Therefore,it is urgent to carry out the study on its mechanism and suppression measures.Since the control strategy of vehicle rectifiers is the key influencing factor,considering that the model predictive control(MPC)has significant advantages in converter control,a model predictive control based on optimal solution is designed to achieve better control effect of electric multiple units(EMUs).Then,an improved MPC based on extended state observer(ESO)is implemented to further optimize the system’s robustness.And simulational and experimental platforms are built to verify the enhancement in control performance and LFO suppression capability of the proposed methods.Firstly,based on the reduced-order principle and the parameters of multi-conductor transmission line,the reduced-order simulation model of traction network is built.Based on the equivalent circuit of the EMUs’ traction drive system,the mathematical model of the single-phase rectifier of EMUs in the dq frmme is deduced.Secondly,the RTLAB integrated simulation platform and dSPACE semi-physical experimental platform are established,respectively based on the simulation model of vehicle-grid coupling system and traditional dq decoupling control,laying the foundation for the comparison analysis with the proposed algorithms below.Then,to improve the control performance of EMUs,the current predictive model of traction line-side converter(LSC)is deduced based on the mathematical model of the singlephase rectifier in the dq frame,and the two-step prediction is adopted to eliminate delay error.The performance function is constructed by the error of current predictive values with their reference values and the variation value of control voltages.The optimal control voltages for the next sampling interval can be got through finding the the variation value of control voltages by differentially solving the performance function.Finally,combined with sinusoidal pulse width modulation(SPWM),the switch sequence for control output is obtained.The proposed algorithm was evaluated by simulations and experiments.Finally,to solve the problem that the control accurancy of MPC will be significantly decreased when the system has serious model mismatch or parameter errors,a robust predictive control combining with extended state observer is designed to improve the antiinterference capability.Firstly,based on the current prediction model of EMUs’ rectifier,the total system disturbance caused by the circuit parameter error and the unmodeled term of the system is extended into a new system variable.Then,a linear ESO is construced.The total disturbance of the system is estimated to provide real-time disturbance compensation for the control voltage calculated by the predictive control algorithm.Finally,based on the bandwidth and transfer function method,the stability of the system is analyzed to achieve the pole configuration,and the appropriate gains of observer are chosen.The proposed algorithm was evaluated by simulations and experiments. |
