论文部分内容阅读
本文研究了采用多目标优化算法NSGA-Ⅱ与前馈神经网络的耦合优化方法的收敛性问题。提出了一种改进的拥挤距离和coarse-to-fine的迭代策略,有效地解决了原耦合方法不收敛的问题。在此基础上,提出了基于数值模拟和耦合优化方法的多目标气动设计框架,并对转子37进行了三目标优化。优化结果表明,多目标优化结果的气动性能优于单目标优化结果。同时,多目标优化可以提供多种可选择方案,具有很高的实用性。
This paper studies the convergence problem of the coupled optimization method using multi-objective optimization algorithm NSGA-Ⅱ and feedforward neural networks. An improved congestion distance and coarse-to-fine iterative strategy is proposed to effectively solve the problem that the original coupling method does not converge. On this basis, a multi-objective aerodynamic design framework based on numerical simulation and coupling optimization is proposed, and the rotor 37 is optimized for three-objective. The optimization results show that the aerodynamic performance of the multi-objective optimization results is better than the single-objective optimization results. At the same time, multi-objective optimization can provide a variety of options, with high practicality.