Research On A Structural Topological Optimization With Displacement Constraints Based On ABAQUS

The structural topology optimization is an advanced technology based on theconventional optimization design, since its purpose is to make the mechanicalproperties and indexes are satisfied at structural design stage, therefore it hasradically changed the old design concept. At present，the complication and highflexibility of structures such as vehicle and engineering machinery etc have madetheir service environment more and more complicated . They are necessary and areof very important application value to investigate the structural topologyoptimization method and develop its application software based on existing largecommercial finite element analysis software.ABAQUS has an element library with various structural element models and amaterial model library, as well as the powerful FEA computational capabilities andthe capability dealing with highly nonlinear optimized problems. This thesis putsABAQUS as a platform and does the secondary development of FEA and structuraloptimization software, by using Python and Fortran languages. The chapter two inthis thesis introduces the main details and procedure of the secondary developmentbased on ABAQUS, and gives several examples that show the effectiveness and highefficiency of the secondary development.In addition, based on the mentioned above development work, for the structuretopological optimization problem with the objective function being the weight anddisplacement constraints, incorporating the method of independent, continuous,mapping(ICM), the technique of volume-preserving Heaviside density andsensitivity filtering, a set of efficient displacement approximation functions of thefinite element model structure with large dimensions has been built. Integrating themultiple displacement constraints measurement and the dual planning method, a newstructural topology optimization method with multiple displacement constraintsbased on Heaviside density and sensitivity filtering is proposed. Several typicalexamples verify that the proposed method is effective and may be used to obtain 0/1solutions.