基于遗传算法的结构健康监测-结构工程专业论文.docx

摘 摘 要 目前，基于结构动力特性的损伤渗断正越来越成为研究热点。一般认为， 结构损伤的发生必然会导致结构动力参数(如刚度、频率、振型等)的改变，对 结构的反应f一般是动力反应)进行分析，提取信号中包含的损伤信息能对结构 的健康状况进行评估。其方法主要有两种，一是应用传统的信号处理方法，二 是与某个目标函数的最小化相联系，通过对其优化，提取损伤特征从而对结构 的损伤情况进行判断。遗传算法作为一种优化算法，由于具有并行性、自适应 性以及全域性等优点，自问世以来已被广泛应用于如函数优化、自动控制、机 器学习、神经网络等领域，在结构的损伤诊断方面也有所应用。本文将对遗传 算法在结构健康监测中的应用作进一步的研究。主要的研究内容有： I．对建立目标的方法进行了研究。基于广泛应用于结构健康监测的“残余 力向量法”进行进一步的研究。把“残余力向量法”的目标向量值转化为遗传 算法的搜索信息——目标函数值。 2．对遗传算法的改进进行了研究。针对目标函数决策变量多、精度要求高 等特点，本文在传统遗传算法的基础上采取了进一步的改进措施来提高遗传算 法的优化性能。诸如：实值编码方式、基于线性排名适应度分配、随机遍历抽 样的选择方法等。 3．对算法在结构健康监测中的应用进行了研究。对于应用各种不同的模 态、不同噪声、单一损伤和多处损伤情况下遗传算法的识别性能进行了研究。 最后，提出了损伤部位的多少与遗传算法性能之间的关系。 关键词：结构健康监测；损伤诊断；遗传算法 AbstractGreat Abstract Great attention has given to the structure health monitoring based on dynamic characteristics．We generally believe that the dynamic characteristic parameters(the stiffness，frequency,and mode)are changed when damage is occurred．Then we could use these parameters which involve damaged information to evaluate the health condition of a structure．The methods are mainly divided into two kinds：one is the traditional signal process method；another is minimized the value of an objective function which is established by the dynamic characteristic parameters． Genetic algorithm is a kind of optimized algorithm which was used in field of function optimizing，automatic controlling，machine leaning，NN and SO forth．An application of genetic algorithms to structural health monitoring has been studied， and the mainly points ate： 1．The method of establishing objective function has been studied．More research about residue force method which widely used in structure health monitoring has been done in this paper．A method to obtain objective value which used as searching information in GA throu曲residue force method has been brought out． 2．Some ways to ameliorate GA is researched in this paper．Since the objective function involve many decisive variables and more accurate solutions are needed． Some improvements are advanced to enhance the optimizing c