华南理工大学学报(自然科学版) ›› 2020, Vol. 48 ›› Issue (4): 65-72,94.doi: 10.12141/j.issn.1000-565X.190117

• 机械工程 • 上一篇    下一篇

基于增强响应灵敏度法的分数阶系统参数识别

刘广 刘济科 吕中荣    

  1. 中山大学 航空航天学院 力学系,广东 广州 510275
  • 收稿日期:2019-03-21 修回日期:2019-07-27 出版日期:2020-04-25 发布日期:2020-04-01
  • 通信作者: 吕中荣(1975-),男,教授,博士生导师,主要从事非线性振动、流固耦合力学、参数识别、分数阶微分系 统、结构健康监测等的研究。 E-mail:lvzhr@mail.sysu.edu.cn
  • 作者简介:刘广(1992-),男,博士生,主要从事非线性振动、参数识别研究。E-mail:liug22@mail2.sysu.edu.cn
  • 基金资助:
    国家自然科学基金资助项目 (11972380); 广东省自然科学基金资助项目 (2018B030311001); 广东省科技计划项目 (2016A020223006); 中央高校基本科研业务费专项资金资助项目 (17lgjc42)

Parameter Identification of Fractional-Order System via Enhanced Response Sensitivity Approach 

LIU Guang LIU Jike Lü Zhongrong   

  1. Department of Applied Mechanics,School of Aeronautics and Astronautics,Sun Yat-Sen University,Guangzhou 510275,Guangdong,China
  • Received:2019-03-21 Revised:2019-07-27 Online:2020-04-25 Published:2020-04-01
  • Contact: 吕中荣(1975-),男,教授,博士生导师,主要从事非线性振动、流固耦合力学、参数识别、分数阶微分系 统、结构健康监测等的研究。 E-mail:lvzhr@mail.sysu.edu.cn
  • About author:刘广(1992-),男,博士生,主要从事非线性振动、参数识别研究。E-mail:liug22@mail2.sysu.edu.cn
  • Supported by:
    Supported by the National Natural Science Foundation of China (11972380),the Natural Science Foundation of Guangdong Province ( 2018B030311001 ) and the Science and Technology Planning Project of Guangdong Province(2016A020223006)

摘要: 与整数阶微分算子相比,分数阶算子可以更加准确地描述系统的某些特性,如遗传和记忆功能。为了更加准确和高效地求解分数阶系统,文中首先提出了一种基于Adams 离散和 Newmark-β 法的数值方法来获得系统的响应。对于分数阶系统的参数识别反问题,将其归结为一类非线性的优化问题,然后采用增强响应灵敏度法对识别方程进行求解。算例表明,文中提出的正问题计算格式可以快速、准确地得到系统的响应,增强响应灵敏度法能准确识别出系统的参数,并且对模拟的测量噪声不敏感。

关键词: 分数阶微分系统, Adams 离散, Newmark-β 法, 响应灵敏度分析, 参数识别

Abstract: Compared with the integer order differential operator,fractional order operator can describe some charac-teristics of the system more accurately,such as hereditary and memory properties. In order to find a more accurate and efficient way to solve the fractional order system,a new computational scheme designed to get the system re-sponse was put forward based on Adams-type scheme and the Newmark-β method. The inverse problem of parame-ter identification was considered as a nonlinear optimization problem. And the enhanced response sensitivity method was used to solve the identification equation. Numerical examples show that the calculation format of the direct problem proposed in the paper can quickly and accurately get response from the system,so the enhanced response sensitivity method can accurately identify the parameters of the fractional system and be insensitive to the measure-ment noise.

Key words: fractional-order differential system, Adams discrete, Newmark-β method, response sensitivity analy-sis, parameter identification

中图分类号: