基于T-S型模糊神经网络的内模自适应PID算法研究  

Research on internal model adaptive PID algorithm based on T-S fuzzy neural network

作　　者：张皓 高瑜翔[1,2] 唐军 杜鑫昌[1,2] 刘海波 ZHANG Hao;GAO Yuxiang;TANG Jun;DU Xinchang;LIU Haibo(College of Communication Engineering, Chengdu University of Information Technology;Meteorological Information and Signal Processing Key Laboratory of Sichuan Education Institutes, Sichuan Chengdu 610225, China;College of Electronic Information and Artificial Intelligence, Yibin Vocational And Technical College, Sichuan Yibin 644000, China,)

机构地区：[1]成都信息工程大学通信工程学院 [2]气象信息与信号处理四川省高校重点实验室,四川成都610225 [3]宜宾职业技术学院电子信息与人工智能学院,四川宜宾644000

出　　处：《工业仪表与自动化装置》2021年第6期118-124,共7页Industrial Instrumentation & Automation

基　　金：四川省教育厅高校创新团队项目(15TD0022)。

摘　　要：针对内模PID算法在控制具有大滞后特性的温度模型时,因内模参数无法根据内部模型的失配程度进行自适应调整,使得系统超调量增大,调节时间变长,控制响应不及时,控制品质下降等问题。该文提出了基于T-S型模糊神经网络的内模自适应PID算法,大大降低上述问题对系统控制性能指标的影响。Matlab仿真结果表明:在内部模型与被控模型的参数出现正向失配或负向失配时,该文算法控制响应最快、超调量最小,控制精度最高,具有最好的控制性能指标。When the internal model PID algorithm is controlling a temperature model with large hysteresis characteristics,because the internal model parameters cannot be adjusted adaptively according to the degree of mismatch of the internal model,the system overshoot increases,the adjustment time becomes longer,and the control response is not timely,control the quality degradation and other issues.This paper proposes an internal model adaptive PID algorithm based on the T-S fuzzy neural network,which greatly reduces the impact of the above problems on the system control performance indicators.The Matlab simulation results show that when the parameters of the internal model and the controlled model are positively mismatched or negatively mismatched,the algorithm in this paper has the fastest control response,the smallest overshoot,the highest control accuracy,and the best control performance index.

关 键 词：内模PID 大滞后 温度模型 T-S型模糊神经网络 

分 类 号：TP273.3[自动化与计算机技术—检测技术与自动化装置] TP273.4[自动化与计算机技术—控制科学与工程]