摘要
提出了一种基于代理模型的裂纹识别方法,利用初始样本构造Kriging代理模型,建立裂纹模型参数与结构响应的关系,来代替结构的原有结构参数与动力响应关系,最大程度地减少了反演优化迭代过程中反复网格剖分和冗繁的有限元计算次数。使用最优设计加点准则进行代理模型修正,以改进初始代理模型的准确性。为了识别连续体结构上的裂纹模型参数,采用随机粒子群优化方法搜索代理模型多极值域下的全局最优解。数值算例对具有裂纹的悬臂梁和板结构进行了裂纹识别。结果表明,该方法能有效地识别裂纹参数,并且具有良好的抗噪性能。此外,讨论了初始样本数量对裂纹识别效率及识别结果的影响。
A method of crack identification is presented based on the Kriging surrogate model. The initial samples are used to construct the initial Kriging model establishing the relationship between the crack parameters and the corresponding structural dynamic responses instead of bewteen the dynamical input and output and thereby avoiding either the re-meshing process at ev- ery iterative step of optimization or the time-consuming finite element calculation. To improve the accuracy of the surrogate model, an optimal point-adding process is carried out to reduce the computational cost. For identifying crack parameters based on the constructed Kriging model, a robust stochastic particle swarm optimization （SPSO） algorithm is applied to enhance the global searching ability. Numerical studies for a cantilever beam and a plate having a respective crack are performed. The effec- tiveness and noise immunity of this method are demonstrated by the identification results. In addition, the effects of initial sampling size on the identification efficiency and the precision of the identification results are also investigated
出处
《振动工程学报》
EI
CSCD
北大核心
2013年第6期879-885,共7页
Journal of Vibration Engineering
基金
国家973计划资助项目(2011CB706504)
