Topological structural design using genetic algorithms
Abstract
A new approach to topological shape optimization is presented. The admissible volume for the design is specified, as are all loading conditions and required attachment points. This admissible volume is discretized and each element of the mesh is assigned a design variable which specifies a material property such as the modulus of elasticity. The initial mesh remains unaltered throughout the solution process. A genetic optimization algorithm is applied to determine the optimal material property for each element. A penalty function is employed to successively drive the material property closer to its upper or lower bound. Constraints concerning stress and/or displacement limits are computed via finite element analysis or other suitable structural analysis methods. The resulting optimization is an approximation to the optimum topology of the structural component. The efficiency and effectiveness of the method is studied through designed experiments as well as by the unbiased design over domains whose optimal solution is known. Design problems whose optimal solutions are not derivable by other means are also investigated.
Degree
Ph.D.
Advisors
Bernhard, Purdue University.
Subject Area
Mechanical engineering
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.