基于数据挖掘的IMC-PID在线自整定算法的实现

Online Self-Tuning of IMC-PID Controllers Based on Data Mining

提出了利用过程历史数据自动进行数据挖掘的PID参数在线自整定算法。算法以PID回路的动态响应特性为依据,通过给定ε-不敏感损失函数、辨识信任度函数,从可行数据集中选取有效数据集,以此作为回路参数自整定的有效数据。为确保PID控制尽可能达到最佳性能和鲁棒性,提出了基于对象组进行IMC-PID参数整定的方法。算法已应用于多个生产装置上,实际的投运结果表明,这种算法具有简便易用,推广能力强等特点,是PID参数整定算法中一种切实可行的算法。

An algorithm of automatic on-line tuning of PID parameters is designed,which uses the process historical data for data mining.This algorithm obtains the feasible data group on the basis of dynamic characteristic of PID control loop.Throughout the defined ε-insensitive loss function and identification confidence function,valid data group are selected from operative data group,as the valid data for model identification and PID parameters on-line tuning.To ensure the PID control to achieve the best control performance in the different condition of running period and satisfy the robustness,a method of obtaining the identification object group based on the valid data group is presented.These objects could be used as the IMC-PID parameter tuning.It does not require staff participation,and the obtained PID parameters can be directly applied to the control loop without further adjustment.This algorithm has been successfully applied to many production plants,and achieved satisfactory results.The actual results show that this algorithm,as a practical algorithm of PID parameter tuning,has obvious advantages of simple use and convenient promotion.