摘要
针对红外加热系统的复杂动态特性和非线性问题,提出一种改进的基于RBF神经网络的无模型自适应控制方法。对红外加热过程的传热机制进行分析,建立红外加热过程输入功率与温度之间的全格式动态线性化数据模型;针对MFAC控制器参数难以整定的问题,基于参数灵敏度分析与参数关联,提出一种基于参数敏感性分析和参数内部关联的参数估计方法,并通过该方法设置MFAC控制器的参数;进一步,利用RBF神经网络任意逼近非线性函数的能力,在线辨识无模型自适应控制器的伪偏导值,以实现红外加热系统温度的精准控制。最后,通过实验和仿真验证该方法的可行性与优越性。结果表明:RBF-MFAC控制下的系统温度控制精度为±0.3℃,调节时间为60 s;MFAC的温度控制精度为±1℃,调节时间为75 s;PID的温度控制精度为±1.2℃,调节时间为96 s。这说明与原MFAC和PID算法相比,改进后的RBF-MFAC算法具有更快的响应速度、更小的稳态误差。
Aiming at the complex dynamic characteristics and nonlinear problems of infrared heating system,an improved model-free adaptive control method based on RBF neural network was proposed.The heat transfer mechanism of the infrared heating process was analyzed,and a full-format dynamic linearized data model between the input power and the temperature of the infrared heating process was established.To address the problem of difficult parameter tuning of the MFAC controller,a parameter estimation method was proposed based on parameter sensitivity analysis and internal parameter correlation,and this method was used to set the parameters of the MFAC controller.Furthermore,the RBF neural network was introduced,which was equipped with the capability of arbitrarily approximating nonlinear functions,and it was used to identify the pseudo-bias of the model-free adaptive controller online,so as to achieve the accurate temperature control of the infrared heating system.Finally,the feasibility and superiority of the method were verified by experiments and simulations.The results show that under RBF-MFAC control,the system temperature control accuracy is±0.3℃,with an adjustment time of 60 s;MFAC has a temperature control accuracy of±1℃,with an adjustment time of 75 s;PID has a temperature control accuracy of±1.2℃,with an adjustment time of 96 s.This indicates that compared with the original MFAC and PID algorithms,the improved RBF-MFAC algorithm has faster response speed and smaller steady-state error.
作者
李喜龙
徐智浩
杨晓京
LI Xilong;XU Zhihao;YANG Xiaojing(Faculty of Mechanical and Electrical Engineering,Kunming University of Science and Technology,Kunming Yunnan 650500,China;Institute of Intelligent Manufacturing,Guangdong Academy of Sciences,Guangzhou Guangdong 510070,China)
出处
《机床与液压》
北大核心
2025年第24期110-115,共6页
Machine Tool & Hydraulics
基金
国家重点研发计划(2022YFF0607800)
广东省基础与应用基础研究基金(2022A1515011749)
江门市科技计划项目(2023780200050009211)
广东省科学院青年人才专项(2023GDASQNRC-0204)
广东省科学院发展专项资金项目(2024GDASZH-2024010102)。
