大滞后系统的Smith在线辨识预估控制的研究

Research on large time-delay system based on online identification of Smith predictive control

工业过程中普遍存在大时滞对象,为解决大滞后复杂系统因无法建立精确数学模型而难于控制的问题,将史密斯(Smith)预估控制原理和在线辨识方法结合起来,在Smith预估控制系统中,用系统辨识的模型代替传统预估补偿模型,根据最小二乘辨识算法辨识模型的各个参数,提出了Smith在线辨识预估控制算法;针对二阶加纯滞后对象在滞后时间有/无建模误差进行了仿真研究。研究结果表明,Smith在线辨识预估控制的性能指标和鲁棒性有很大的改善。

The large time-delay is often existent in industrial processes. In order to solve the problem of difficult to control without accurate system model during the long time-delay control, an online identification of Smith predictive controller for a system with the long time-delay was proposed. Combined the Smith predictive control with online identification, the tradition system model in Smith predictive control was replaced by system identification model which utilized reeursive least-square to identify the various parameters of the model. The results of simulation for the second-order plus pure delay model objects with or without error in the lag time show that the online identification of Smith predictive control has the best performance indexes and better robustness.