小潜艇供电系统全桥DC-DC开关电源智能优化设计

Intelligent Optimum Design of Full-bridge DC-DC Switching Power Supply on Miniature Submarine's Power Supply System

小型潜艇水下潜行采用蓄电池供电,通过全电力供电系统推进。以蓄电池为直流电源,在推进系统正常运转的状况下,开关电源电压输出值保持恒定;传统的PID控制器很难对全桥DC-DC开关电源进行精确控制,因此,将粒子群算法(PSO)、BP神经网络与径向基函数(RBF)神经网络与传统PID控制相结合,提出带有RBF神经网络辨识的PSO-BP-PID控制方法;通过神经网络在线自学习对PID3个参数在线调整,最终实现系统恒电压输出控制;仿真结果得出:带有RBF神经网络辨识的PSO-BP-PID控制算法可以很大的改善系统控制效果,同时使系统具有更好的在线调整能力。

Miniature submarines, powered by an all--electric propulsion system, are battery powered under the water. While the propulsion system is running normally, the battery acting as DC power supply, the output voltage of switching power supply remains constant. It is very difficult to acuurately control full--bridge DC--DC switching retulator using the traditional PID controller. So, this paper will combine particle swarm optimization （PSO）. BP neural network, radial basis function （RBF） neural network, traditional PID and put for- ward the control method of PID algorithm based on neural network. Through online self--learning, the neural network adjusts the three parameters of the PID to get a eostant output voltage finally. Simulation results show that the PSO _ BP _ PID control algorithm with RBF neural network identification can greatly improve the control effect of the system, meanwhile, make the system online adjustment better.