Research On A Topology Optimization Method With Random Dynamic Response Constraints And Its Sotfware Development

With the social development and progress, the modern structures tend to belighter and more complex, and these structures are subjected to diverse and seriousload sources of dynamic excitation during their service life, which lead to newchallenges to optimization theories and its application technologies. The structuraltopological optimization is the most challenging and promising research topic instructure optimization, it can provide an efficient design method and procedure forstructural conceptual designs.The purpose of structural optimizations is to shorten the structural design period,improve design quality and applies an advanced theory into engineering practices.Therefore, it is very meaningful to develop a convenient and powerful structuraloptimization software system. Because ABAQUS is of the powerful capabilities of thepre- and post-process for structural finite element models and the reliable andauthoritative capabilities for structural analyses, this thesis does a secondarydevelopment on commercial finite element analysis software ABAQUS in order toinvergating the structural topological optimization algorithm and developing itsapplication software. The following three aspects are the main work of the secondarydevelopment in this thesis: 1) to build a structural model information interlinkingsoftware module by using the Python language. The module may obtain the allstructural model information required by the optimization design computation byadopting a structural model information produced by using the ABAQUS processmodule; 2) to make a set of sensitivity analysis programs and an optimization solvingprogram in the Fortran language; 3) to make a batch ordering program for structuraloptimizations in order to realizing the automatic cycle iteration in the wholeoptimization process.Because the dynamic topology optimization problem is more difficult than thestatic topology optimization problem, and is a challenge problem in the structuralproblem, there are few researches in the dynamic topological optimization designaspect. Based on the idea of the ICM method (Independent, Continuous and Mapping)and the evolutionary structural optimization method, topology optimization problemwith the smallest structural weight and only dynamic response constraints on thespecified locations is investigated under white-noise force excitation. In order tocontrol the change quantity of topologic design variables, new dynamic response constraints limits are formed and introduced into the optimization model at thebeginning step of each subloop iteration. At the same time, a reasonable model of therational approximation of material properties (RAMP), which is used in thepseudo-density method, is introduced and the lowest density elements is removed toavoid the occurrence of localized mode in this thesis.Moreover, the element deletion and adding criterion and a set of structuraloptimization strategies are given to automatically adjust the design space. Someelements with artificial material property are inserted around the cavities andboundaries of the structure optimized to get a non singular structure and realize thefunction of element restoring. At last, the dual programming method is used to solvethe optimization problem. The mathematical programming method with a fewerheuristic parameters can be applied to large-scale problems that need a lot of finiteelements in its fixed design domain, and the optimum topology without any gray areascan be obtained. The several topology optimization examples are solved todemonstrate the method is corrective and effective.