遗传算法在结构损伤识别中的应用研究

结构损伤识别是结构系统在使用期间进行监测和维护的重要组成部分,而基于动态测试技术的结构损伤识别方法又是近来研究的热点。本文提出了一种基于改进遗传算法的结构损伤识别方法,主要改进包括:浮点编码、基于标准化几何分布排名的选择策略、最优保存策略、算术交叉算子、非均匀变异算子。在常规模态分析的基础上,以节点的残余力向量构造目标函数,提出了一种用于遗传搜索优化的新的目标函数形式。利用遗传算法重点进行了噪声条件下的结构损伤定位和定量研究,并用一平面桁架模型进行了数值模拟。结果表明:提出的方法不仅能够同时进行结构的损伤定位和定量计算,而且抗噪性能很好。最后,对方法应用中存在的一些问题进行了深入分析,给出了一些有益的结论。

Study on an Application of Genetic Algorithm to Damage Identification

Structural damage identification is a vital part of the monitoring and maintaining of structural systems during their service life.Damage identification of structures based on dynamic test and measurement techniques has been a hot topic in recent years.Damage identification is an inverse problem in structural engineering.There are three main questions in the structural damage identification: the presence,location and extent of the structural damage.An improved genetic algorithm(GA) for the structural damage identification is presented in the present paper.The improvements include the adoption of real-valued representation,normalized geometric ranking operator,elitist model, arithmetical crossover operator and non-uniform mutation operator.Based on the conventional modal analysis theory,a new formulation of the objective function for GA,formulated by the residual forces at the nodes,is presented.Firstly, the residual forces method is employed to localize the damage,which can accurately locate the structural damage using one of any order of the modal data without noise.Noise,however,can not be avoided during the dynamic measurement,which makes the identification of the structural damage completely impossible.Considering this problem,the localization and quantification of the extent of the structural damage with noise is further performed with GA implemented by the objective function formulated by the residual forces at the node which is based on the conventional modal analysis theory.Unlike the conventional optimization methods,such as the gradient methods,which are easily trapped into local minima for the damage detection problem,GA has the capacity of locating global minima for function optimization with multiple local minima.Most importantly,GA can treat the noise-polluted problems effectively and is especially applicable to damage detection.Finally,an illustrative example of a plane truss model is used to perform a numerical simulation.The results show that the presented methods can localize and quantify the structural damage simultaneously and has good performance to deal with noise.Some problems encountered in this method are discussed in detail and some concluding remarks are given.