基 于 多 可 信 度 代 理 模 型 的 尾 喷 管 优 化 设 计

The nozzle is one of the key components in the integrated design of the aircraft body/propulsion，which directly affects the thrust-drag matching and pitching moment matching and other characteristics of the aircraft. In order to improve the efficiency of the optimal design of the nozzle，a multi-objective optimization model based on multi-fidelity surrogate model is established. The CFD results based on coarse mesh and inviscid model are used as low-fidelity data，and the CFD results based on fine mesh and SST k-ω turbulence model are used as high-fidelity data，the Cokriging models are constructed with the maximization of thrust coefficient，lift coefficient and pitching moment coefficient as optimization objectives，and the Pareto solution set is obtained by combining NSGA-Ⅱ algorithm. The optimization results show that the three aforementioned objectives are improved by 2.94%，13.0% and 40.6% respectively with the error less than 0.5%，and the sensitivity analysis is completed；the optimized flow field wave system structure is more complex which changes the wall pressure distribution law；compared with Kriging model，Cokriging model has considerable prediction performance and saves 62% of the construction cost.