粒子群优化算法及其在结构动力修改中的应用研究

【作者】 史志俊；

【导师】 孙木楠；

【作者基本信息】 南京理工大学 ， 工程力学， 2005， 硕士

【摘要】 近年来新结构和新材料的使用,在有效地减轻了结构质量的同时,却降低了结构的刚度,从而使结构动力学问题显得尤为突出。结构模型修改已经逐步演化为一个多学科的研究课题。针对这一情况,本文首先介绍了一种国际上最新提出的高效的进化算法——粒子群优化算法的基本原理、实现方法以及目前的主要应用情况。而后,将粒子群优化算法运用到结构模型修改这一领域之中,对一个实际的五层框架结构的动力模型进行修正,得出了修改后的结构刚度。并用修改后模型计算出的模态数据与实验测得的模态数据的相似度ER和MAC来衡量模型修改的准确性,最后的计算结果ER与MAC和基于灵敏度分析、神经网络、和遗传算法的模型修改方法的计算结果进行了比较。对比结果表明,在大多数情况下,用粒子群优化算法计算所得到的修正结果要比其它算法所得到的计算结果更加接近于实际测得结果。因此,用粒子群优化算法进行模型修改是可行的。更多还原

【Abstract】 Recent years, the use of new structures and new materials has given rise to some structural dynamic problems because the rigidity of structures is reduced although it effectively improves the reduction of the structures’ masses. Because of this, the structural model updating has become a multidisciplinary problem in the dynamics field. In this paper, a newly emerged efficient evolutionary computation technique -the Particle swarm optimization algorithm is investigated, including the principal theory, methods to realization and current application. Then the Particle Swarm optimization algorithm is applied to the structural model updating of a five-floor structure, from which the updated rigidity of the structure is obtained. The precision of model updating is measured by the similarities (ER and MAC) between experimentally measured mode data and the mode data computed by the updated model. In addition, the results computed by means of Particle swarm optimization algorithm are compared with the results computed by other methods such as Sensitivity analysis, neural network and Genetic algorithm etc. Conclusions can be drawn from the comparison that the Particle swarm optimization algorithm gives more appropriate results than those of other methods so updating structural model by PSO algorithm is valid.更多还原