基于PNS算法的高超声速内转式进气道优化设计

Optimization Design of Hypersonic Inward Turning Inlet Based on SSPNS Algorithm

关键词:  SSPNS算法  内转式进气道  单目标优化  多目标优化

Key words:  SSPNS algorithm  Inward turning inlet  Single-objective optimization  Multi-objective optimization

摘要:

为解决内转式进气道优化设计中计算代价过大的问题，将单次扫描空间推进PNS方程(SSPNS)应用于高超声速内转式进气道单目标及多目标优化设计。计算来流马赫数4.5，收缩比5设计条件下的构型，SSPNS与NS计算结果中，流场结构基本一致，出口截面总压恢复系数、压升及马赫数符合得较好，SSPNS耗时不到NS的百分之一，验证了SSPNS的可靠性及高效性。以总压恢复系数最大为目标，使用多岛遗传算法和非线性序列二次规划(MIGA+NLPQL)的单次组合策略及指针自动控制优化策略(Pointer-2)分别进行单目标优化，得到的最优值为0.530及0.559，较初值0.464提高14%和20%。Pointer-2计算次数少、计算结果优，说明该优化模型中Pointer-2具有更高计算效率和更强的全局探索能力。以总压恢复系数和压升最大为目标，通过多目标遗传算法进行优化，得到Pareto前沿，并拟合出具有工程价值的总压恢复系数与压升非劣解关系式。

Abstract:

To reduce computational consumption in the optimization process of hypersonic inward turning inlet，single and multi-objective optimization designs were carried out based on single sweep parabolized Navier-Stokes(SSPNS) algorithm. The free-stream mach number of 4.5 and convergence ratio of 5 were set as the computational conditions. The flow field structure calculated by SSPNS algorithm was similar to that by NS algorithm，and the total pressure recovery，Mach number and pressure rise of outlet section showed good agreements，while the computing time by SSPNS was less than one percent of that by NS. The results verified the accuracy and efficiency of SSPNS algorithm. Single objective optimization aimed at the maximum total pressure recovery coefficient was implemented using a single combinational optimization strategy of MIGA and NLPQL，as well as Pointer automatic optimization strategy(Pointer-2). The optimized objective increased to 0.530 and 0.559 from the initial value 0.464，improved by 14% and 20%，respectively. Less iterations are needed using Pointer-2 and the results are better，which indicates that the Pointer-2 algorithm has higher computational efficiency and performs better in global exploration. Multi objective optimization aimed at the highest total pressure recovery coefficient and pressure rise was carried out based on multi-objective genetic algorithm. Both the Pareto front and the relationship formula between total pressure recovery and pressure rise of non-inferior solution proposed are helpful for further engineering application.