摘要
对近 2 0年来国际上在基于动力学参数的结构健康诊断方面的成就与问题进行了系统的回顾和总结 ,扩展了目前流行的一些结构健康诊断方法 ,建立了一系列新的结构健康诊断公式。从中发现 :①结构的总体刚度矩阵 [K],模态特征值λj 和模态振型 [Φ]j 之间存在以下关系δλj=[Φ]Tj [δK][Φ]j 和δλj≤ 0。②对于损伤结构 ,[δΦ]=0和 {δλ}=0同时发生的情况一定不存在。也就是说 ,结构损伤必然会在结构的内禀动力特征 {Φ ,λ}上得到完整的反映。这一发现回答了人们长期以来的一项争论 ,即结构的内禀动力特征 {Φ ,λ}是否足以作为结构损伤识别和定位的充分条件 ,答案是肯定的。③通常应变类结构动力学参数比位移类结构动力学参数对结构损伤更敏感。④利用多层前向人工神经网络的自组织学习功能和广义遗传算法的全局优化功能 ,设计一个基于广义遗传算法和人工神经网络的混合型结构智能健康诊断演化系统是可行的 ,且该系统的知识库可实现自主进化。
The paper systematically reviews and summarizes the problems and achievements, which are found in dynamic parameter based structural health diagnosis in the past two decades. Some popular structural health diagnosis methods which are model based and model free are expanded and improved. A series of novel structural health diagnosis formula are proposed. This paper discovers that ① and (0, where, and are the global stiffness matrix, modal eigenvalue, and modal shape of structure, respectively.② With respect of damaged structure, the event of emerging=0 and=0 at the same time is impossible. So it is certain that structural damage can be expressed completely by the change of . This discovery provides answers for the first time a longtime for the debate in the world whether the structural inner dynamical feature is sufficient to be the basis of structural damage recognition and localization. And our answer is positive.③ Strain type dynamic parameters are usually more sensitive to structural damages than displacement type. ④It is feasible to design a mixed type dynamic parameter based structure intelligent health diagnosis system. This system is based on the integration of the generalized genetic algorithm and artificial neural network, which use self learning function of artificial neural network and the global optimization capability of the generalized genetic algorithm. What is more, the knowledge bank is capable to realize self revolution.
出处
《中国铁道科学》
EI
CAS
CSCD
北大核心
2002年第1期11-24,共14页
China Railway Science
基金
国家自然科学基金 ( 5 95 0 5 0 11
5 97780 39
5 9975 0 49)
航空科学基金 ( 95B5 10 62 )
863计划 ( 863 2 44 1
863 2 44 3)
清华大学基础研究基金 (JC2 0 0 0 0 0 3
JC2 0 0 10 0 4)
结构工程与振动教育部重点实验室
土木工程防灾国家重点实验室教
