基于模态修正策略和稀疏正则化的损伤识别

Damage identification using modal changes correction strategy and sparse regularization

结构损伤识别问题中,剪切层结构简化带来的模型误差会对识别结果带来不可预料的影响。为了尽可能地减少模型误差的影响,提出了模态改变修正策略。并结合损伤位置的稀疏性,引入稀疏正则化以解决损伤识别反问题的不适定性。此外,为了使稀疏正则化引起的额外计算成本尽可能小,提出了一种解耦的损伤识别目标函数,并用交替优化方法进行求解。文章通过识别结构健康监测标准模型成功验证了所提方法的正确性,并且展示出模态改变修正策略的显著优点:仅需要得知剪切层结构的层数即可利用测量的模态数据进行损伤识别,适用于难以获取实际质量与刚度的真实剪切层建筑的损伤识别。

For structural damage identification,model errors brought by the simplification of shear structure will often have bad effects on the result.To alleviate the errors,the modal measurement changes correction strategy is proposed.And because of the sparsity of damage itself,sparse regulariza‐tion is invoked to solve the ill-posedness of damage identification that is an inverse problem.More‐over,in order to reduce the extra computational cost caused by sparse regularization,a new objective function is established accordingly,and the alternate optimization method is used to get the solution.The accuracy of the identification method present here are verified by identifying the SHM Benchmark problem.Additionally,a significant advantage of the modal measurement changes correction strategy has been shown.That is,only the number of floors is required to detect damage,which extremely suitable for the damage identification of real multi-story buildings whose actual mass and stiffness are hard to obtain.