基于商函数的动态载荷识别最优正则化参数选取方法

发布时间：2018-05-21 20:24

  本文选题：载荷识别 + 不适定问题　； 参考：《东北石油大学学报》2016年02期

【摘要】：为解决载荷识别反问题,研究选取最优正则化参数商函数方法。利用Tikhonov正则化方法的最优化问题的最小二乘解,定义以正则化参数为自变量的商函数;根据不同的正则化参数,使用二次规划理论,求解Tikhonov正则化方法的最优化问题的最优解;基于不同最优解对应商函数的不同特点,将最优正则化参数的商函数方法,与广义交叉检验(Generalized Cross-Validation,GCV)准则所得载荷识别结果进行比较。数值仿真及试验验证结果表明,商函数方法对于共振区及非共振区下载荷识别问题具有较好的合理性和适用性。
[Abstract]:In order to solve the inverse problem of load identification, the optimal regularization parameter quotient function method is studied. Using the least square solution of the optimization problem of Tikhonov regularization method, the quotient function with regularization parameters as independent variable is defined, and the optimal solution of the optimization problem of Tikhonov regularization method is solved by using quadratic programming theory according to different regularization parameters. Based on the different characteristics of quotient functions corresponding to different optimal solutions, the quotient function method of optimal regularization parameters is compared with the load identification results obtained by generalized cross-validation GCVC criterion. The numerical simulation and experimental results show that the quotient function method is reasonable and applicable to the identification of the download load in the resonant and non-resonant regions.
【作者单位】： 哈尔滨工业大学航天学院;北京宇航系统工程研究所;
【基金】：国家自然科学基金项目(11372084)
【分类号】：O347.1
，

本文编号：1920671