Research On Efficient Global Aerodynamic Optimization Design Algorithm Based On Co-Kriging Model

Aerodynamic shape design and optimization is one of the key problems in aircraft overall design.With the increasing refinement and complexity of the aerodynamic shape of high-performance aircraft,the traditional optimization design methods are facing an unbearable computational cost.As an approximate modeling method that can reduce a large number of high-reliability CFD analysis,agent model has been widely studied and applied.Aiming at the problem that the standard crow search algorithm is sensitive to parameters and difficult to obtain parameters,by studying the optimization principle of the algorithm,this paper analyzes the decisive role of parameters on the search direction of the algorithm in the process of algorithm optimization,and summarizes the influence of algorithm parameters on the search ability of the algorithm.A crow search algorithm based on parameter adaptive strategy is proposed,and the improved algorithm is applied to function testing and constrained engineering examples,it is found that the improved crow search algorithm achieves a better balance between robustness and convergence speed than the standard algorithm.Compared with other swarm intelligence algorithms,it is also verified that the improved algorithm has better optimization ability.Based on the Kriging model and combined with the improved crow search algorithm,an aerodynamic shape optimization design framework based on the surrogate model is established.The batch generation of airfoil mesh based on ICEM platform and batch calculation based on fluent platform are realized by writing rpl file and jou file.On this basis,the optimization framework constructed in this paper is applied to optimize the design of NACA 0012 airfoil.The lift-to-drag ratio of the optimized airfoil is 52.95% higher than that of NACA 0012 airfoil,and the feasibility of applying the optimization algorithm based on single surrogate model approximation to aerodynamic shape optimization design is verified.On the basis of single surrogate model modeling,by introducing low-precision data and taking co Kriging model as the basis of variable complexity surrogate model modeling,an efficient global aerodynamic optimization design algorithm framework is constructed by combining with the improved crow search algorithm,and the global optimization ability of the established optimization framework is verified by function test.Based on the flap design theory,the trailing edge flap is parameterized,and the designed flap is optimized by using the efficient global aerodynamic optimization design algorithm based on the Co-Kriging model.The results show that the stall characteristics of the optimized airfoil are greatly improved compared with NACA0012,the lift-to-drag ratio is increased by 80.42%,and the lift-to-drag ratio under design conditions is also increased by 53.49% compared with the initial airfoil with flap structure.By studying and improving the crow search algorithm and combining it with the application of surrogate model,this paper constructs the aerodynamic optimization design algorithm framework based on single surrogate model and the efficient global aerodynamic optimization calculation framework based on variable complexity surrogate model.It has been successfully applied to the aerodynamic shape optimization of traditional two-dimensional airfoil and airfoil with flap structure,and the aerodynamic shape with higher lift-to-drag ratio is obtained.