Strong stabilization servo controller with optimization of performance criteria
Strong stabilization servo controller with optimization of performance criteria
复制标题
DOI：
10.1016/j.isatra.2011.03.005
复制
发表时间：
2011-07-01
期刊：
影响因子：
7.3
通讯作者：
Chowdhury, Amor
中科院分区：
文献类型：
--
作者：
Sarjas, Andrej;Svecko, Rajko;Chowdhury, Amor
采用极点配置技术和H-无穷大指标综合一种简单的鲁棒控制器，用于无刷直流电机和无刷直流电机伺服机构的控制。该方法包括基于所选择的特征多项式、使用Manabe标准多项式形式和参数解来求解多项式方程。参数解被直接引入到伺服控制器的结构中。在所选参数解的基础上，通过不确定性模型和对平行于(无穷大)的范数的评估来评估闭环系统的鲁棒性。设计过程和优化采用了一种遗传算法差分进化-DE算法。DE优化方法在整个优化过程中根据谱平方多项式和Siljak的绝对稳定性检验来确定次优解。用李帕托夫稳定性条件检验了所设计控制器在优化过程中的稳定性。这两种方法都利用了Siljak检验和Lipatov条件，根据多项式的系数来检验系统的稳健性和稳定性，这对于闭环控制的自动化设计和DE等优化算法的应用都是非常方便的。(C)2011年《国际行政程序法》。爱思唯尔有限公司出版。保留所有权利。
Synthesis of a simple robust controller with a pole placement technique and a H-infinity metrics is the method used for control of a servo mechanism with BLDC and BDC electric motors. The method includes solving a polynomial equation on the basis of the chosen characteristic polynomial using the Manabe standard polynomial form and parametric solutions. Parametric solutions are introduced directly into the structure of the servo controller. On the basis of the chosen parametric solutions the robustness of a closed-loop system is assessed through uncertainty models and assessment of the norm parallel to.parallel to(infinity). The design procedure and the optimization are performed with a genetic algorithm differential evolution - DE. The DE optimization method determines a suboptimal solution throughout the optimization on the basis of a spectrally square polynomial and Siljak's absolute stability test. The stability of the designed controller during the optimization is being checked with Lipatov's stability condition. Both utilized approaches: Siljak's test and Lipatov's condition, check the robustness and stability characteristics on the basis of the polynomial's coefficients, and are very convenient for automated design of closed-loop control and for application in optimization algorithms such as DE. (C) 2011 ISA. Published by Elsevier Ltd. All rights reserved.