基于Simulink仿真的电子加速器控制方法研究  

Research on Control Method of Electron Accelerator Based on Simulink Simulation

作　　者：岳宏伟 李中平[2,3] 周有为 曹树春 任洁茹[1] 张子民 赵永涛 YUE Hongwei;LI Zhongping;ZHOU Youwei;CAO Shuchun;REN Jieru;ZHANG Zimin;ZHAO Yongtao(MOE Key Laboratory for Nonequilibrium Synthesis and Modulation of Condensed Matter,School of Physics,Xi’an Jiaotong University,Xi’an 710049,China;Institute of Modern Physics,Chinese Academy of Sciences,Lanzhou 730000,China;University of Chinese Academy of Sciences,Beijing 100049,China)

机构地区：[1]西安交通大学物理学院物质非平衡合成与调控教育部重点实验室,陕西西安710049 [2]中国科学院近代物理研究所,甘肃兰州730000 [3]中国科学院大学,北京100049

出　　处：《原子能科学技术》2025年第1期197-204,共8页Atomic Energy Science and Technology

基　　金：国家重点基础研究发展计划(2019YFA0404900);国家自然科学基金(U1532263,11705141)。

摘　　要：电子加速器广泛用于材料改性、消毒灭菌和污水处理等领域。然而,在实际应用中束流强度的控制存在无法快速、准确调整的问题,大大降低了生产效率。为了解决电子加速器中束流变化的非线性、时变性和不稳定性问题,本文采用了PID算法、模糊PID算法和模糊神经网络PID算法,并利用MATLAB中的Simulink构建了相应的控制系统仿真模型,对这3种算法的性能进行了比较。通过对比PID算法、模糊PID算法和模糊神经网络PID算法的仿真结果可知:模糊PID算法在稳定时间、超调量、加入扰动后稳定时间方面分别降低了59.6%、48.9%、64.9%;模糊神经网络PID算法在稳定时间、超调量、加入扰动后稳定时间方面分别降低了77.9%、79.6%、87.1%。模糊PID算法和模糊神经网络PID算法有望提高电子加速器束流控制的精度和效率。Electron accelerators are widely used in material modification,disinfection and sterilization,sewage treatment and other fields.However,in practical applications,the control of electron accelerator beam intensity can't be adjusted quickly and accurately,which greatly reduces the efficiency and quality of production and processing.This paper aims to solve the problems of nonlinearity,timevarying and instability in the beam control process of electron accelerator.To achieve this,the PID algorithm,fuzzy PID algorithm,and fuzzy neural network PID algorithm were employed.The basic principles of each algorithm were first introduced.Then,a mathematical simulation model for beam current control was constructed based on the processing of experimental data from electron beam emission experiments and the theoretical formulas related to electron accelerator beam emission.The three algorithms were subsequently applied to this mathematical simulation model within MATLAB's Simulink environment.Finally,simulations were conducted in Simulink,with the desired beam current set to 100 mA and the simulation time to 40 seconds.A 5%step response(5 mA)was introduced at 25seconds as a disturbance.The performance of each algorithm was then compared and analyzed in terms of stabilization time,overshoot,and post-disturbance recovery time.The results show that compared with the PID algorithm,the performance of the fuzzy PID algorithm and the fuzzy neural network PID algorithm is significantly improved.Specifically,the system stabilization time of the fuzzy PID algorithm is reduced by 59.6%,the overshoot is reduced by 48.9%,and the post-disturbance recovery time is reduced by 64.9%.The fuzzy neural network PID algorithm improves these indicators more significantly,the stabilization time is reduced by 77.9%,the overshoot is reduced by 79.6%,and the post-disturbance recovery time is reduced by 87.1%.Based on these results,it is concluded that the fuzzy PID algorithm and the fuzzy neural network PID algorithm can improve the performance of the e

关 键 词：电子加速器 PID算法 模糊PID算法 模糊神经网络PID算法 SIMULINK 

分 类 号：TL506[核科学技术—核技术及应用]