结构优化设计的组合优化策略

INTEGRATED OPTIMIZATION STRATEGY IN STRUCTURE OPTIMIZATION DESIGN

智能优化算法和基于灵敏度分析的梯度方法在解决结构优化设计问题时均有不足,因此提出将两者结合起来的组合优化策略,旨在利用智能优化算法的全局搜索能力和梯度方法的快速局部搜索能力,达到优化质量和优化效率间的平衡。为减少计算所需的结构重分析次数,使用响应面法建立结构响应的响应面模型(response surface model,RSM)。使用Camel测试函数证明组合优化策略的可靠性和有效性。最后,在多学科优化设计集成平台iSIGHT上集成Nastran,采用两种不同的组合优化策略完成工字型截面悬臂梁的截面尺寸优化。结果显示,组合优化策略能高效且尽可能地保证解的全局最优性。

Both intelligent optimization algorithms and gradient methods based on sensitivity analysis had difficulties in solving structure optimization design, thereby integrated optimization strategy based on combination of them was proposed so as to make full use of global exploratory ability of intelligent optimization algorithms and fast local exploratory ability of gradient methods, in order to gain the balance between optimization quality and efficiency. Response surface methodology was used to the construction of Response Surface Model（RSM） of structure response in order to reduce the number of structure analysis. Camel test function testified the reliability and efficiency of integrated optimization strategy. In the end, integrating Nastran into iSIGHT which was a multidiseiplinary design optimization platform and using two different integrated optimization strategies carried out the section sizing design of a cantilever with I section. The result shows that integrated optimization strategy can effectively assure the global optimum to a great extent.