基于虚拟材料的ESO检测算法研究及其在薄壁结构中的应用
发布时间：2018-10-11 18:44
【摘要】：渐进结构优化算法(Evolutionary structural optimization,ESO)是基于有限元方法的结构拓扑优化算法研究热点之一。ESO采用不断删除设计域中的低效单元使结构逐渐趋于最优的朴素思想,方法易于实现。自创立以来,ESO在算法的理论研究和工程应用方面获得了广泛的成果。ESO属于启发式算法,Rozvany以Tie-beam算例来质疑ESO方法的算法有效性。就此,目前已有研究者提出多种Tie-beam的解答算法,并以此改进ESO算法的不足。其中基于虚拟材料的试探检测算法通过检测传力路径完整性,较好地解决了ESO方法在Tie-beam问题中的缺陷。本文的研究进一步表明,检测算法对ESO优化进程中绝大多数的迭代步的检测是没有必要的。因此,本文将原虚拟材料的试探检测过程拆分为两个步骤:通过预判定选出优化进程中结构刚度或强度发生异常变化的迭代步,然后只对这些可能出现结构连接性破坏的异常迭代进行试探检测,从而大幅提高了检测算法的效率。原虚拟材料的试探检测算法是就平面问题的求解而设计的。本文将算法扩展到板壳问题,推导了相应的检测判断准则,设计了算法流程。算例表明,本文扩展的基于虚拟材料的ESO检测算法可成功弥补原ESO方法在求解某些薄壁结构时失效的不足。本文将扩展的虚拟材料检测算法应用于某型号垃圾集装箱箱体薄壁结构加强筋的布置问题,成功获得可用于生产实际的箱体薄壁结构最优拓扑。本文提出的ESO改进算法成功防止了原ESO方法在薄壁结构的拓扑优化中可能发生的失效,而且不影响ESO方法本身的寻优能力,对ESO算法研究有一定理论意义,同时具有工程实用价值。
[Abstract]:Progressive structural optimization algorithm (Evolutionary structural optimization,ESO) is one of the research hotspots of structural topology optimization algorithm based on finite element method (FEM). ESO adopts the simple idea of deleting inefficient elements in the design domain to make the structure become more and more optimal, and the method is easy to implement. Since its creation, ESO has obtained extensive achievements in the theoretical research and engineering application of the algorithm. ESO is a heuristic algorithm, and Rozvany uses the Tie-beam example to question the validity of the ESO algorithm. At present, researchers have put forward a variety of Tie-beam solution algorithms, and improve the shortcomings of the ESO algorithm. The testing algorithm based on virtual material solves the defect of ESO method in Tie-beam problem by detecting the integrity of the transmission path. The research in this paper further shows that the detection algorithm is not necessary for the detection of most iterative steps in the ESO optimization process. Therefore, in this paper, the testing process of the original virtual material is divided into two steps: the iterative step of abnormal structural stiffness or strength change in the optimization process is selected by pre-determination. Then only these anomalous iterations which may occur structural connectivity breakage are probed and detected, thus greatly improving the efficiency of the detection algorithm. The testing algorithm of the original virtual material is designed to solve the plane problem. In this paper, the algorithm is extended to the plate and shell problem, the corresponding detection criteria are deduced, and the algorithm flow is designed. An example shows that the extended ESO detection algorithm based on virtual materials can successfully compensate for the failure of the original ESO method in solving some thin-walled structures. In this paper, the extended virtual material detection algorithm is applied to the arrangement of stiffeners in a certain type of waste container thin-walled structure, and the optimal topology of the thin-walled structure can be obtained successfully. The improved ESO algorithm proposed in this paper successfully prevents the failure of the original ESO method in the topology optimization of thin-walled structures, and does not affect the optimization ability of the ESO method itself, which is of theoretical significance to the study of the ESO algorithm. At the same time, it has engineering practical value.
【学位授予单位】：重庆大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TB30
[Abstract]:Progressive structural optimization algorithm (Evolutionary structural optimization,ESO) is one of the research hotspots of structural topology optimization algorithm based on finite element method (FEM). ESO adopts the simple idea of deleting inefficient elements in the design domain to make the structure become more and more optimal, and the method is easy to implement. Since its creation, ESO has obtained extensive achievements in the theoretical research and engineering application of the algorithm. ESO is a heuristic algorithm, and Rozvany uses the Tie-beam example to question the validity of the ESO algorithm. At present, researchers have put forward a variety of Tie-beam solution algorithms, and improve the shortcomings of the ESO algorithm. The testing algorithm based on virtual material solves the defect of ESO method in Tie-beam problem by detecting the integrity of the transmission path. The research in this paper further shows that the detection algorithm is not necessary for the detection of most iterative steps in the ESO optimization process. Therefore, in this paper, the testing process of the original virtual material is divided into two steps: the iterative step of abnormal structural stiffness or strength change in the optimization process is selected by pre-determination. Then only these anomalous iterations which may occur structural connectivity breakage are probed and detected, thus greatly improving the efficiency of the detection algorithm. The testing algorithm of the original virtual material is designed to solve the plane problem. In this paper, the algorithm is extended to the plate and shell problem, the corresponding detection criteria are deduced, and the algorithm flow is designed. An example shows that the extended ESO detection algorithm based on virtual materials can successfully compensate for the failure of the original ESO method in solving some thin-walled structures. In this paper, the extended virtual material detection algorithm is applied to the arrangement of stiffeners in a certain type of waste container thin-walled structure, and the optimal topology of the thin-walled structure can be obtained successfully. The improved ESO algorithm proposed in this paper successfully prevents the failure of the original ESO method in the topology optimization of thin-walled structures, and does not affect the optimization ability of the ESO method itself, which is of theoretical significance to the study of the ESO algorithm. At the same time, it has engineering practical value.
【学位授予单位】：重庆大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TB30
【相似文献】
相关期刊论文 前10条
1 卢东滨;陈平;白天;柴朋军;付朝军;;复合材料薄壁结构老化性能研究[J];玻璃钢/复合材料;2014年06期
2 殷文;;轻质高强薄壁结构混凝土的收缩变形研究[J];混凝土与水泥制品;2013年07期
3 程奂，
本文编号：2264855
本文链接：https://www.wllwen.com/kejilunwen/cailiaohuaxuelunwen/2264855.html
