开关电源作为非线性时变动态系统,其数学建模研究一直是实现开关电源高性能控制的热点和难点问题.该文首先推导了时变负载情况下开关电源小信号模型;然后,利用RBF神经网络进行动态负载下开关电源的建模;最后,结合模型预测控制技术进行动负载情况下开关电源的非显式预测控制研究.仿真结果表明,基于RBF神经网络的预测控制技术可以大大降低由于负载和输入电压等变化对电源供电品质造成的影响,可以获得比数字PID控制和线性预测控制更好的控制性能.%Switching power converter as a nonlinear dynamic systems,mathematical modeling study has been a hot and difficult problems to achieve high-performance switching power supply control.In this paper,the small signal model of switching power supply under varying load conditions is derived at firstly.Then,the RBF neural network based modeling for dynamic load switching power supply is studied.Finally,non-explicit predictive control for switching power supply under dynamic load conditions is researched by combining the RBF neural network and nonlinear predictive control.The simulation results show that the RBF neural network based predictive control technology can greatly reduce the quality influence of the power supplying with load or input voltage varying,and can obtain higher control performance than the digital PID control and linear predictive control.
展开▼
