Research On Bi-directional Evolutionary Structural Optimization Method For Continuum Structures

Topology optimization is the recent focus of structural optimization. The bi-directional evolutionary optimization (BESO) method has the advantages of the easy combination with finite element analysis program and its application in large-scale problems. The simplicity of the algorithm has made it an increasingly powerful design tool in topology optimization for continuum structures, especially in the material arrangements in structural systems. For the problems with stress constraint, an algorithm was proposed by adding a mass of elements to the structure when needed during the iteration. The ability to find the better topology and shape is increased and the problem of converging prematurely to partial optimum solutions is avoided. For the problems with stress and displacement constraints, under the consideration of the stress constraints, the addition of elements to structure when displacement constrains were not satisfied and otherwise the removal of elements form structure were mainly performed to satisfy the stress and displacement constraints gradually. For the problem of singular structures in optimization process, the singularity are overcome by classifying the singular forms of their modalities and deleting the singular element or returning to the nonsingular structure to ensure the continuity of optimization by adjusting the optimization parameters. A new optimum design module was secondarily developed by PCL language (PATRAN Command Language) for topology optimization of 2-D continuum structures. The BESO topology optimization module has the functions of building FEA model easily, real-time visualization of optimization progress, and convenience to examining optimization results depending on the strong pre- and post-processing ability of PATRAN. The results for 2-D plates and beams are illustrated and the effectiveness and usefulness of the algorithms are showed.