Improved differential evolutionary algorithm for nonlinear identification of a novel vibration-assisted swing cutting system

作     者：Lu, Mingming Wang, Hao Zhao, Dongpo Lin, Jieqiong Gu, Yan Yi, Allen 

作者机构：Changchun Univ Technol Sch Mechatron Engn Key Lab Micro Nano & Ultraprecis Mfg Jilin Prov Changchun 130012 Jilin Peoples R China Ohio State Univ Dept Ind Welding & Syst Engn Columbus OH 43210 USA 

出 版 物：《INTERNATIONAL JOURNAL OF ADAPTIVE CONTROL AND SIGNAL PROCESSING》 (国际自适应控制与信号处理杂志)

年 卷 期：2019年第33卷第7期

页      面：1066-1078页

核心收录：

学科分类：0808[工学-电气工程] 08[工学] 0811[工学-控制科学与工程] 

基　　金：Ministry of Science and Technology State Key Support Program [2016YFE0105100] Micro-Nano and Ultra-Precision Key Laboratory of Jilin Province [20140622008JC] Science and Technology Development Projects of Jilin Province [20180101034JC, 20180201052GX] Education Department Scientific Research Planning Project of Jilin Provincial [JJKH20181038KJ] 

主　　题：Hammerstein-Wiener model improved differential evolutionary algorithm system identification test functions vibration-assisted swing cutting 

摘      要：Vibration-assisted swing cutting (VASC) is a new precision machining technology. VASC not only inherits the characteristics of EVC intermittent cutting but also alleviates the problem of EVC residual height. However, system identification is key if you want to achieve precise control. In order to solve this problem, a new improved differential evolutionary (IDE) algorithm is proposed to identify and optimize the Hammerstein-Wiener model parameter in VASC system. IDE algorithm is applied to transform the identification problem of the model into the optimization problem in the parameter space, and the optimal solution of the parameter of the model in the parameter space is obtained. Meanwhile, the IDE algorithm and the conventional five differential evolutionary algorithms perform performance comparison tests on six different test functions. The test results show that the IDE algorithm is strengthening the global search capability, accelerate the convergence rate to the global optimal solution, and indicate that the IDE algorithm can be effectively applied to the parameter optimization of Hammerstein-winner model. Based on the input and output data collected from the experiment, the accuracy of the identification model can be up to 98%, which prove the superiority of the proposed IDE algorithm for system identification.