基于模型预测控制的电动汽车直流充电桩功率器件开路故障诊断方法

An Open-circuit Fault Diagnosis Method for EV DC Charging Pile Power Device Based on Model Predictive Control

针对电动汽车直流充电桩故障中较为常见的功率器件开路故障,提出了一种基于模型预测控制的电动汽车充电桩实时故障诊断方法。采用模型预测电流控制方法替代传统PWM整流器电流环中的PI控制器。在此基础上,将功率器件开路故障情况下整流器实际应用的开关状态与通过模型预测控制得到的最优开关状态相比较,建立故障特征信号,从而实现充电桩功率器件开路故障诊断与定位。由于不同功率器件故障特征信号之间互不影响,因此该方法对于单管、双管及多管同时故障均有显著效果。仿真结果验证了该充电桩故障诊断方法的正确性和有效性。

Aiming at solving the open-circuit faults of power devices which is common in DC charging pile of electric vehicle,this paper proposes a real-time fault diagnosis method for charging pile of electric vehicle based on model predictive control.The conventional PI controller in the current loop of the PWM rectifier is substituted with model predictive current control.On this basis,the actual switching state of the rectifier with open-circuit fault of the power device is compared with the optimal switching state obtained through model predictive control.Then,the fault characteristic signal is established to realize the open circuit fault diagnosis of charging pile power devices.Since the fault characteristic signals of different power devices are mutually independent,the proposed method can locate the faults of single,double and multi-tube of power devices.The simulation results verify the correctness and effectiveness of the proposed fault diagnosis method for charging pile.