摘要
考虑到基于2范数的正则化算法存在对结构识别结果过度光滑的效果,提出了基于模态振型与L1正则化的损伤识别方法。以一2D简支梁有限元模型为数值算例,比较了使用不同振型数以及不同损伤程度对损伤识别效果的影响。数值模拟结果表明,对于多损伤工况,当损伤结构的振型数和无损结构的振型数乘积数大于6时,能较好地进行损伤定位,并能对损伤程度给出定性的描述。
Considering that the regularization algorithm based on norm 2 has over-smooth effect on the structure recognition results,a damage detection method based on modal vibration and L1 regularization was proposed.Taking the finite element model of a 2D simply-supported beam as a numerical example,the effect of using different numbers of vibration modes and that of using different damage degrees on the damage recognition effect were compared.Simulation results show that,for multi-damage cases,when the product of the number of vibration modes of the damaged structure multiplied by that of the undamaged structure is larger than 6,the presented method could detect the structural damage location preferably and give a qualitative description for the damage degree.
作者
麻胜兰
刘昱昊
张铮
吴琛
MA Shenglan;LIU Yuhao;ZHANG Zheng;WU Chen(School of Civil Engineering, Fujian University of Technology, Fuzhou 350118, China)
出处
《福建工程学院学报》
CAS
2020年第1期40-45,共6页
Journal of Fujian University of Technology
基金
国家自然科学基金项目(51808119)
福州市科技局项目(2018X0066)
关键词
损伤检测
有限元方法
模态分析
稀疏恢复
L1正则化
damage detection
finite-element method
model analysis
sparse recovery
L1 regularization
作者简介
第一作者:麻胜兰(1986-),女,福建龙岩人,讲师,博士,研究方向:结构健康监测与诊断加固。
