基于神经网络的电动汽车动态无线充电功率控制

Dynamic Wireless Charging Power Control of Electric Vehicle Based on Neural Network

为实现多初级绕组并联电动汽车的动态无线充电,建立双发单收无线电能传输系统模型,得到动态模式下的传输特性与磁场分布。为保证输出功率的稳定调节,增加基于副边控制的升压型变换器。设计了反向传播(BP)神经网络控制器,通过控制高频开关管占空比的变化,实现输出电流的平稳控制。仿真分析表明,所设计的BP神经网络控制器在控制速度、超调量、控制稳定性等方面均优于传统PID控制,同时具有较好的泛化能力。

To realize dynamic wireless charging of multi primary winding parallel electric vehicle, a doubletransmitting and single-receiving wireless power transmission system model was established, to get the transmission characteristics and distribution of magnetic field in dynamic mode. To ensure the stable regulation of output power, a boost converter based on sub side control was added. A Back Propagation （BP） neural network controller was designed to control the output current smoothly by controlling the change of duty cycle of the high frequency switch tube. The simulation analysis shows that the control speed, overshoot and control stability of the designed BP neural network＇controller are better than the traditional PID controller, and has better generalization ability.