论文部分内容阅读
通过验证实例分析,气动设计中精细化优化模型对设计结果收敛精度的提高有很大帮助,但同时也带来优化算法搜索困难的问题,并且由于不同类型设计变量之间的相互耦合干扰使优化难以收敛到全局最优解。于是提出基于响应均值灵敏度的概念对大规模的设计变量进行重要性分组的策略,依据设计变量分组情况应用系统分解思想对多变量设计问题进行分层协同优化来降低系统的复杂度,这既保证了精细化设计的要求,又缓解了优化算法对大规模问题搜索困难的问题。与传统气动优化方法相比,基于系统分解的分层协同优化算例有较大寻优效率和性能提升,证明了该方法的可行性。
By validating the case analysis, the refinement optimization model in aerodynamic design is very helpful to improve the convergence accuracy of the design results, but it also brings about the problem of search difficulty of the optimization algorithm, and the optimal coupling of the different types of design variables Difficult to converge to the global optimal solution. Therefore, the strategy of grouping importance of large-scale design variables based on the concept of response-average sensitivity is proposed. According to the grouping of design variables, the idea of system decomposition is applied to reduce the complexity of the system by using hierarchical decomposition and multi-variable design to reduce the complexity of the system. The requirements of fine design, but also ease the optimization algorithm for large-scale problem search difficult problem. Compared with the traditional aerodynamic optimization method, the hierarchical optimization algorithm based on system decomposition has greater optimization efficiency and performance improvement, and proves the feasibility of the method.