基于W-SPSO算法的高阶时滞系统频域辨识方法  

Frequency Domain Identification Method for High-order Time-delay Systems Based on W-SPSO Algorithm

作　　者：邹军 王亚刚[1] 李菲菲[1] Zou Jun;WANG Ya-gang;LI Fei-fei(School of Optical-electrical and Computer Engineering,University of Shanghai for Science and Technology,Shanghai 200093,China)

机构地区：[1]上海理工大学光电信息与计算机工程学院,上海200093

出　　处：《控制工程》2021年第9期1856-1860,共5页Control Engineering of China

基　　金：国家自然科学基金资助项目(61074087,61703277);国家自然科学基金青年基金资助项目(11502145)。

摘　　要：工业生产中常会出现高阶时滞系统,其控制器设计较为复杂,而常见的思路是先将其转换为低阶系统。对此,一些学者通过选取一定数量的特征频率点,根据幅频特性进行频域辨识,以二阶纯滞后模型拟合高阶系统,通过对比一阶时滞和二阶不加时滞模型,验证了该模型的可行性与准确性,然而由于选点的随机性以及数量的局限性,与原系统的拟合度难以保证。基于W-SPSO算法寻优能力的改进方法,通过引入合适的适应度函数,对选取的响应点进行优化,从而避免上述缺陷。仿真结果表明,改进后的Nyquist曲线与原对象拟合度更佳。High-order time-delay systems often appear in industrial production, and the design of their controller is complicated. The common idea is to convert the high-order time-delay system into a low-order one. In this regard, some scholars select a certain number of characteristic frequency points, perform frequency domain identification according to the amplitude-frequency characteristics, and fit the high-order system with a second-order pure lag model. By comparing the first-order time-delay model and the second-order model without delay, the feasibility and accuracy of the model are verified. However, due to the randomness of the selected points and the limitation of the number, it is difficult to guarantee the fitting degree with the original system. The improved method based on the optimization ability of W-SPSO algorithm optimizes the selected response points by introducing an appropriate fitness function, hence, avoiding the above-mentioned defects. The simulation results show that the improved Nyquist curve fits better with the original object.

关 键 词：高阶时滞 幅频特性 频域辨识 二阶纯滞后模型 拟合度 W-SPSO 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]