基于PID控制器性能评价  

PID-based Controller Performance Assessment

作　　者：封子文[1] 李宏光[2] 陈兰朋[2] 成慧禹[3] 张天翔[4] 

机构地区：[1]中国石化塔河分公司,新疆库车842000 [2]北京化工大学信息科学与技术学院,北京100029 [3]中国石化洛阳分公司,河南洛阳471072 [4]中国石油大学(北京)地球物理信息学院,北京102249

出　　处：《石油化工自动化》2012年第3期30-33,53,共5页Automation in Petro-chemical Industry

基　　金：北京市重点学科基金资助项目(编号:XK100100435)

摘　　要：在化工过程中,90%以上的控制器类型为PID类型,对PID控制器进行性能评价具有重要的实际意义。PID控制器的性能评价多以理论计算的最小方差作为基准,但由于控制器受自身结构的限制,实际上根本无法达到这个数值。提出采用PID控制器能够达到的最小方差作为标准,来评价常规操作下的PID控制器性能,使得PID控制器性能评价结果具有实际意义。在求取PID可达最小方差过程中根据输入输出数据拟合ARMAX过程模型,采用SOS(Sums Of Squares)方法解决了非凸优化问题,保证了全局最优解。将该方法应用在TE(Tennessee-Eastman)过程仿真平台上,实验结果表明了其有效性和实用性。In chemical industry process, more than 90% of the controllers are PID types. It＇s of practical significance to evaluate the performances of PID controllers. Most existing controller performance assessments are based on theoretical minimum variance which can＇t be achieved due to the structural limitation of PID controllers. Achievable minimum variance of PID controller is introduced to assess the PID controller performance under normal operation and make the PID controller performance evaluation result more meaningful. The non-convex optimization problem is solved by fitting ARMAX process model based on the input/output data during getting the achievable minimum variance along with the application of SOS （Squares of Sums） method, and the optimized solution for the whole program is ensured. Test result of using this method on TE （Tennessee-Eastman） process emulation platform illustrates the feasibility and effectiveness.

关 键 词：PID控制器 最小方差 控制器性能评价 

分 类 号：TP273.2[自动化与计算机技术—检测技术与自动化装置]