Analysis And Optimization Of Aerodynamic Characteristics Of Variable Camber Airfoil Based On Neural Network And CFD

As people’s demand for a green living environment increases and fuel costs continue to rise,energy conservation,emission reduction,lift increase,drag reduction and noise reduction have become the core issues in the design of large civil aircraft.To achieve a leap in the performance of the aircraft,it must have good aerodynamic characteristics and power systems.Therefore,the design of the lift-enhancing device is particularly important for.In order to meet the needs of the aviation industry to increase lift,reduce resistance and reduce noise,flow control technology has attracted the attention of many researchers,and a variety of flow control techniques have been developed to improve the performance of high-lift devices,such as variable camber wing,trailing edge serration structure,blowing and flowing suction,etc.Compared with these flow control technologies,the variable camber wing is one of the most effective ways to improve aerodynamic and acoustic performance.In the future,low-noise and high-performance aircraft will be the focus and difficulty of research in the field of aviation.So,this paper researches the effect of variable camber airfoil on aerodynamic and acoustic performance,and carries out aerodynamic optimization design of variable camber airfoil.The optimized airfoil is calculated by using large eddy simulation combined with FW-H acoustic equation,and the mechanism of the improved of aerodynamic and acoustic performance of variable camber airfoil is analyzed.The major work and research findings of the thesis are as follows.(1)Firstly,the hybrid numerical simulation method of flow field/sound field in this paper is introduced.To provide a better initial solution for the flow field,the k-ω SST turbulence model is used for the constant calculation.To capture the flow field details in a refined manner,the large eddy simulation is used for the non-constant flow field calculation and the FW-H equation is used for the acoustic field.Secondly,the grid independence verification is carried out,this sets the stage for the following research and analysis.Third,the accuracy and reliability of the numerical calculation method in this paper are compared and verified based on the experimental data of NACA 0012 airfoil based on NASA.It has high reliability and accuracy and can be used for further research.(2)The variable camber airfoil model is established,and four parameters that control the variable camber of the airfoil are determined,namely the leading edge downward bending angle,deflection position of the leading edge,the trailing edge downward bending angle and the trailing edge deflection position.By changing one of the factors and fixing the other three factors,the effects of different factors on the aerodynamic performance of the airfoil at different angles of attacks is studied.The conclusion shows that the downward bending angle and deflection position have a great influence on the aerodynamic performance and critical angle of attack of the airfoil,and the leading edge downward bending angle can increase the angle of attack of critical and reduce the drag coefficient,which can effectively delay the arrival of the stall condition and increase the lift-to-drag ratio.The trailing edge variable camber can also effectively increase the lift coefficient of the airfoil in a certain angle of attack range,and then increase the lift-to-drag ratio and improve the aerodynamic performance.(3)In order to find the appropriate variable camber airfoil parameters,the aerodynamic optimization design of the variable camber airfoil was carried out.Firstly,the optimization algorithm used in this paper is introduced in detail.Secondly,within the design scope,the initial training database is obtained by CFD calculation.Then,based on the training database,an agent model of the objective function with respect to the design variables was developed using neural networks,and an iterative optimization strategy was constructed in which the wing optimization process was coupled with neural network prediction,and in each iteration,a deep neural network(DNN)was used to make predictions and a genetic algorithm was employed to maximize the lift coefficient and lift-to-drag ratio.The best predicted results are verified using Fluent calculations.If the DNN result is close to the Fluent result,stop the iterative process.Otherwise,insert the Fluent results into the database to update the DNN prediction model,continue to iterate,and finally obtain multiple sets of variable camber configurations based on the optimization strategy.The Pareto frontier curve is used to solve the conflict between the optimized multi-objectives,and the variable camber airfoil with the best aerodynamic performance is obtained.The optimization results show that when the angle of attack is less than 8°,the lift-drag ratio of the two-dimensional optimized airfoil is about 14 higher than that of the original airfoil.Third,in order to verify the 2D optimization results,the optimized 2D airfoil is stretched into 3D,and it is found that the aerodynamic performance trend of the 3D airfoil is basically the same as that of the 2D airfoil,which shows that the research results of the 2D airfoil are valuable.Finally,the acoustic calculation of the optimized airfoil is carried out and compared with the original airfoil.The results show that the noise of the optimized airfoil can be reduced by about 16～20d B in the middle and high frequency.The simultaneous variable curvature technique for the leading and trailing edges of the airfoil can effectively suppress the boundary layer flow separation,which in turn improves the aerodynamic and acoustic performance of the airfoil.In addition,the optimization strategy based on the combination of deep neural network and genetic algorithm has high computational efficiency and realizes the multi-objective optimization of NACA 0012 airfoil.The optimization results are ideal,which can provide reference for the aerodynamic and acoustic optimization design of airfoil.