计算机集成制造系统 ›› 2019, Vol. 25 ›› Issue (第3): 576-585.DOI: 10.13196/j.cims.2019.03.005

• 当期目次 • 上一篇    下一篇

多源未知载荷下主元回归的多点振动响应预测

王成1,詹威1,李海波1+,赖雄鸣2,缑锦1,张惠臻1   

  1. 1.华侨大学计算机科学与技术学院
    2.华侨大学机电及自动化学院
  • 出版日期:2019-03-31 发布日期:2019-03-31
  • 基金资助:
    国家自然科学基金项目(51305142,61502181);中国博士后科学基金会项目(2014M552429);福建省高校青年自然基金重点资助项目(JZ160409);泉州市科技计划资助项目(2018Z008,2018C110R,2017G019)。

Multipoint vibration response prediction method based on principal component regression under multi-source unknown load conditions

  • Online:2019-03-31 Published:2019-03-31
  • Supported by:
    Project supported by the National Natural Science Foundation,China(No.51305142,61502181),the China Postdoctoral Science Foundation,China(No.2014M552429),the Natural Foundation Key Program for Young Scholars in the Universities of Fujian Province,China(No.JZ160409),and the Science and Technology Planning Project of Quanzhou City,China(No.2018Z008,2018C110R,2017G019).

摘要: 为解决线性时不变系统在不相关多源未知载荷激励下的多点频域振动响应预测问题,对该问题进行了形式化描述并建立了数学模型。从训练集、问题的输入、问题的输出等方面比较了其与不相关多源载荷激励已知下振动响应预测问题的区别,并提出一种基于主元回归的不相关多源未知载荷条件下的线性时不变系统频域振动响应预测方法。通过声振联合激励的圆柱壳结构的多点振动响应预测实验表明,采用主元回归方法的3db误差和均方根误差明显低于多元一次线性回归方法。

关键词: 多点振动响应预测, 主元回归, 不相关多源载荷, 线性时不变系统, 频域

Abstract: To solve the multipoint vibration response prediction problem in frequency domain for linear systems under unknown uncorrelated multi-source unknown load excitations,the formal description and its mathematical model were established.The difference between this problem and multipoint vibration response prediction problem in frequency domain for linear time-invariance systems under unknown uncorrelated multi-source known load excitations was compared in aspect of training set,input,output,etc.A principal component regression was proposed for this problem.The multipoint vibration response prediction experimental results in the cylindrical shell structure under vibration and acoustic combined excitations showed that the 3db and Root Mean Square Error (RMSE) of the vibration response prediction using the principal component regression method were significantly lower than the multivariate linear regression method.

Key words: multipoint vibration response prediction, principal component regression, multiple uncorrelated loads, linear time-invariance system, frequency domain

中图分类号: