首页|修正-联合正则化的冲击载荷识别与响应重构

修正-联合正则化的冲击载荷识别与响应重构

扫码查看
针对传统结构响应重构中正则化方法对冲击载荷峰值识别精度低、非加载区识别结果振荡且识别精度易受噪声干扰等问题,提出基于修正-联合正则化的冲击载荷识别与结构响应重构方法。基于状态空间模型,推导冲击载荷及结构响应的重构方程。对测量响应降噪,利用降噪后响应与识别响应的差值修正L2 正则化解。联合L1 正则化解的稀疏性优势,在保证冲击载荷非加载区域识别稳定的同时,获得更高精度的峰值识别结果,实现结构动态响应的重构。通过数值和实验案例验证了所提方法的有效性,对比了传递矩阵法和粒子滤波法的响应重构效果。结果表明,所提方法具有良好的抗噪性,能够较准确地识别冲击载荷,有效地重构结构动态响应。
Impact load identification and response reconstruction based on updating-combination regularization
An impact load identification and structural response reconstruction method based on the updating-combination regularization was proposed aiming at the problems of low accuracy in identifying peak impact loads,oscillation in identifying non loading areas,and susceptibility to noise interference in traditional regularization methods for structural response reconstruction.The reconstruction equations for the impact load and structure response were derived based on the state space model.The difference between the denoised response and the identification response was used to update the L2 regularization solution.Then higher accuracy peak identification results were obtained combining with the L1 regularization solution that had sparsity advantage while ensuring the stability of the identification of impact load in unloaded region,which realized the reconstruction of structural dynamic responses.The proposed method was verified through numerical and experimental cases,and the effect of response reconstruction based on the transfer matrix method and the particle filter method was compared.Results show that the proposed method has good anti-noise performance.The method can accurately recognize the impact load,and effectively reconstruct the dynamic response of the structure.

response reconstructionimpact loadregularizationtransfer matrixparticle filter

殷红、石咏荷、彭珍瑞、王增辉

展开 >

兰州交通大学机电工程学院,甘肃兰州 730070

西安交通大学机械工程学院,陕西西安 710049

响应重构 冲击载荷 正则化 传递矩阵 粒子滤波

国家自然科学基金甘肃省"创新之星"项目

621610182022CXZX-569

2024

浙江大学学报(工学版)
浙江大学

浙江大学学报(工学版)

CSTPCD北大核心
影响因子:0.625
ISSN:1008-973X
年,卷(期):2024.58(5)
  • 34