结构损伤识别中的模型缩聚问题

On the Model Reduction Techniques in Structural Damage Identification

将逐级近似模型缩聚法用于工程结构损伤识别问题中。以一个平面桁架结构为算例,比较了残余力向量法在各级缩聚模型下的损伤识别结果。结果表明:无论对哪种缩聚方法,并非缩聚模型的级数越高损伤识别效果越好,采用逐级近似的GUYAN法时第二级缩聚模型识别精度最高,采用改进GUYAN递推缩聚法时第一级缩聚模型识别精度最好。分析了出现这种结果的深层次原因,指出了在模型缩聚逐级近似过程中存在的对结构损伤识别有利和不利的2种趋势,给出了良好的模型缩聚方法应该满足的2个条件。

Succession-level approximate reduction techniques are used in structural damage detection in this paper. Using the succession-level approximate reduction technique, results of the residual force algorithm under different level reduced models are discussed through a plane truss structure. It is shown that the result for the second level reduced model using the successive-level Guyan method and the result for the first level reduced model by the successive-level improved Guyan technique are the most accurate. Accounting for that the accuracy of structural damage detection is not better when the level of the structural model reduction increases, the favorable trend and the adverse aspect are pointed out in damage identification using the succession-level approximate reduction technique. In the end, the proposals are provided for the applicable model reduction method in structural damage identification and the trends are expected for the future work.