基于自适应RBF神经网络的升压电路控制方法

Boost Circuit Control Method Based on Adaptive RBF Neural Network

摘要: 能量色散X射线荧光光谱(Energy Dispersive X-ray Fluorescence Spectroscopy, EDXRF)系统对高压直流电源的精度、稳定性和动态响应要求极高。作为高压电源系统的重要组成部分，Boost升压电路的控制器性能直接影响系统的输出强度与稳定性。针对传统PID控制在Boost电路控制中存在的超调和稳态误差等问题，本文提出了一种基于粒子群优化(Particle Swarm Optimization, PSO)的自适应径向基函数(Adaptive Radial Basis Function, RBF)神经网络的电压控制方法。该方法通过RBF网络实现PID参数的非线性映射，并采用改进的PSO算法进行全局参数优化。仿真结果表明，与常规PID控制方法相比，该控制策略在超调抑制、稳态误差和动态响应速度方面均具有明显优势，验证了所提方法在Boost升压电路电压控制中的有效性。

Abstract: The Energy Dispersive X-ray Fluorescence Spectroscopy (EDXRF) system requires extremely high precision, stability, and dynamic response from high-voltage DC power sources. As an important component of the high-voltage power supply system, the controller performance of the Boost boost circuit directly affects the output strength and stability of the system. This paper proposes a voltage control method based on Particle Swarm Optimization (PSO) and Adaptive Radial Basis Function (RBF) neural network to address the problems of overshoot and steady-state error in traditional PID control in Boost circuit control. This method achieves nonlinear mapping of PID parameters through RBF network and uses an improved PSO algorithm for global parameter optimization. The simulation results show that compared with conventional PID control methods, this control strategy has significant advantages in overshoot suppression, steady-state error, and dynamic response speed, verifying the effectiveness of the proposed method in voltage control of Boost boost circuits.