Author(s)

M. El-Sayed, M. Edghill & J. Housner

Abstract

For large-scale structural optimization problems the time and number of finite element analyses required to achieve the optimum solution can be prohibitive. For these structures, with a limited number of design variables the ability of determining a performance parameter over the full range of the design variable variation with a small number of finite element analyses can be very valuable. One of the main problems in performing finite element analysis based on the need of the optimization routine is that the full potential of each analysis to achieve a complete picture of the design space is never realized. By combining the information of several well-spaced analyses outside the optimization process, the full shape of the relevant performance parameters over the specifi

Keywords