关键词: 超燃冲压发动机;变结构尾喷管;试验设计方法;代理模型;多目标优化 

Abstract: Most of time, scramjet has to work on off-design conditions, making it cannot work efficiently all the time. Considering this, a multi-objective optimization design of a geometry-variable nozzle was done here, based on design of experiment (DOE) and surrogate model. Thrust coefficient, lift coefficient and moment coefficient were selected as objective function to form multi-objective optimization of cubic-curve-nozzle. Using multi-objective genetic algorithm, Pareto solutions were found on the design-condition. Then for different Mach numbers and angles of attack, optimal geometry-variable nozzle, of which the undersurface can rotate, was obtained. The correctness of theoretical analysis was confirmed by simulations and several conclusions were achieved. Firstly, the nozzle can work efficiently through a wide envelop of Mach numbers and angles of attack by employing the optimal geometry-variable nozzle. Secondly, the relationship between nozzle performance and geometry parameters, Mach numbers and angles of attack is so complex that changing arbitrary parameter will have great influence on nozzle performance. Thirdly, based on DOE, using surrogate model to replace numerical simulation, the design time can be reduced and the optimization process can be greatly simplified. 

Key words: Scramjet;Geometry-variable nozzle; Design of experiment; Surrogate model; Multi-objective optimization