新能源列车并联三电平Boost直流变换器模型预测均流研究OA

The high-voltage DC power supply system in rail vehicles is one of the important architectures of new energy trams.Utilizing DC converters to transform unstable input power sources into stable DC power ensures the stable operation of the rail vehicles.Parallel connection of high-power converters can achieve dual optimization of inductor volume and converter efficiency,in order to achieve load current sharing of multi module DC converters and improve system stability,an effective current sharing control strategy is required.In response to the situation where there are deviations between components that make it difficult for parallel modules to achieve current sharing,based on the analysis of traditional droop control,it is found that when multiple operate in parallel,the load regulation rate under traditional droop control is relatively large.Therefore,by taking two modules in parallel as an example,based on the classic current sharing method,a three-level Boost DC converter multi module parallel current sharing control method based on model predictive control improvement droop control was proposed.By predicting the output current and output voltage of each converter,the optimal droop coefficient was predicted to achieve balanced current output from each three-level Boost DC converter to the load.By using the MATLAB/Simulink software platform,simulation was conducted on the multi-module parallel current sharing control of a three-level Boost DC converter using improved traditional droop control and model predictive control under load disturbance.The simulation results verified that the improved droop control has a better current sharing effect on the output current of the parallel module in the presence of bias in the boost inductance.From the simulation results,it can be seen that for traditional droop control systems,the uneven current of the system is about 10%before load disturbance,and when the load jumps,the uneven current of the output current is about 15%.While using the droop control improved by model predict control the maximum non-currrent sharing degree is 6%.In order to verify the effectiveness of the simulation,this paper conducted a semi-physical experiment using the RT-LAB platform,and the experimental results were basically consistent with the simulations.The conclusion of this article is that the improved droop control using model predictive control has good current sharing and load adjustment effects.