基于代理模型的转子系统不平衡定量辨识  

Quantitative Identification of Unbalance in a Rotor System Based on Surrogate Model

作　　者：顾煜炯[1] 陈东超[1] 徐婧[1] 何成兵[1] 

机构地区：[1]华北电力大学国家火力发电工程技术研究中心,北京102206

出　　处：《动力工程学报》2015年第12期982-987,1011,共7页Journal of Chinese Society of Power Engineering

基　　金：国家自然科学基金资助项目(51075145);北京市自然科学基金资助项目(3132015);中央高校基本科研业务费专项资金资助项目(2014XS32)

摘　　要：提出了一种基于代理模型的转子系统不平衡定量辨识方法:采用拉丁超立方抽样和转子系统的有限元模型获取训练样本,采用粒子群算法优化的支持向量回归(PSO-SVR)构建代理模型,建立不平衡参数与各测点振动响应的对应关系;构建用于求解不平衡参数的目标函数,并借助建立的代理模型,采用PSO算法寻找满足目标函数的全局最优解,从而达到不平衡定量辨识的目的.并通过仿真算例来验证该方法的准确性和鲁棒性.结果表明:该方法能有效地辨识出转子系统中的不平衡参数,当有限元模型具有较小的误差时,仍能取得较准确的辨识结果,但尽量提高有限元模型的精度仍是不平衡参数辨识结果准确性的重要保障.A novel method for quantitative identification of unbalance in a rotor system was proposed based on surrogate model.Firstly,training samples were obtained based on Latin hypercube sampling（LHS）and finite element model of the rotor system.Then the support vector regression optimized by particle swarm optimization algorithm（PSO-SVR）was adopted to construct surrogate models so as to set up a relation between the unbalance parameters and corresponding vibration dynamic responses.To realize quantitative identification of the unbalance,objective functions were constructed to solve the unbalance parameters,of which the global searching ability was enhanced with the aid of above surrogate models and the particle swarm optimization algorithm.Results show that the method proposed can effectively identify unbalance parameters in the rotor system,even if small errors exist in the finite element model.However,an accurate finite element model is still an important guarantee for obtaining accurate identifying results.

关 键 词：转子系统 不平衡 定量辨识 PSO-SVR代理模型 

分 类 号：TK263.6[动力工程及工程热物理—动力机械及工程]