多质量弹性扭转系统的模型辨识  

The Model Identification of the Multi-mass Elastic Torsion System

作　　者：张永宏[1] 王琦[1] 李远禄[1] 井晨志 

机构地区：[1]南京信息工程大学信息与控制学院,江苏南京210044

出　　处：《信息与控制》2015年第1期62-66,共5页Information and Control

基　　金：国家自然科学基金资助项目(61271395);中国博士后科学基金资助项目(2013M531385)

摘　　要：为了分析多质量弹性扭转系统的特性,需要建立系统模型.考虑到待辨识的系统具有一定的弹性以及分数阶系统的优良特性,尝试将其辨识为分数阶系统.在时域中,对分数阶系统进行辨识时,确定系统的阶数不是很方便,考虑到粒子群算法的随机搜索功能,最终,利用粒子群算法来辨识多质量弹性扭转系统,获得了系统的分数阶模型.作为对比,还对系统进行了整数阶系统辨识,从性能指标来看,分数阶系统优于整数阶系统.为进一步对系统进行特性分析和控制奠定了基础.In order to analyze the characteristics of the multi-mass elastic torsion system, an estimation of the model is necessary. Hence, a fractional system identification considering the elasticity of the system and combining the advantages of the fractional system is implemented. Obtaining the fractional orders in time domain is not convenient. Hence, the fractional order model is obtained by considering the stochastic search pattern of the particle swarm optimization （PSO）, and the PSO algorithm is used to identify the multi-mass elastic torsion system. In contrast, the integer system identification based on PSO is implemented as well. According to these results, the performance indicator of fractional system is better than that of the integer system, laying a foundation for analysis and control of the system.

关 键 词：弹性扭转系统 分数阶系统 系统辨识 粒子群算法 

分 类 号：TP273.2[自动化与计算机技术—检测技术与自动化装置]