Addressing the high-precision modeling requirements of unsteady aerodynamics during complex aircraft maneuvers, a method for modeling non-steady aerodynamic forces based on an adaptive genetic algorithm (AGA) optimized long short-term memory (LSTM) neural network was proposed. Computational fluid dynamics (CFD) simulations were conducted to capture maneuver flight data during rapid turns at varying bank angles and rolling and looping maneuvers at different Mach numbers. An AGA-LSTM model was developed using this data to predict aerodynamic coefficients under non-steady conditions. Specifically, predictions for the aerodynamic coefficients during a 60° bank angle rapid turn maneuver were made, demonstrating accurate estimation of lift coefficient, drag coefficient, and pitch moment coefficient that closely matched CFD simulation results. To further validate the proposed model’s accuracy, predictions were compared with CFD simulation data and a traditional LSTM neural network model for Envelopment maneuvers. The results indicate that the AGA-LSTM neural network model provides closer predictions to simulation data compared to traditional LSTM models, thus offering improved prediction accuracy.
TobakM. On the use of the indicial-function concept in the analysis of unsteady motions of wings and wing-tail combinations:NACA-TR-1188[R].Washington,D.C.:The Superintendent of Documents, U. S. Government Printing Officé,1954.
[3]
TobakM, PearsonW E. A study of nonlinear longitudinal dynamic stability[M].Washington,D.C.:National Aeronautics and Space Administration,1964.
[4]
GomanM, KhrabrovA.State-space representation of aerodynamic characteristics of an aircraft at high angles of attack[J].Journal of Aircraft,1994,31(5):1109-1115.
[5]
AbramovN, KhrabrovA, KolinkoK,et al.Simple wings unsteady aerodynamics at high angles of attack-experimental and modeling results[C]//24th Atmospheric Flight Mechanics Conference.Virginia:AIAA,1999:4013.
LyuY X, CaoY Y, ZhangW G,et al.Dynamic surface control design of post-stall maneuver under unsteady aerodynamics[J].Aerospace Science and Technology,2018,80:269-280.
BagherzadehS A.Nonlinear aircraft system identification using artificial neural networks enhanced by empirical mode decomposition[J].Aerospace Science and Technology,2018,75:155-171.
SrinivasM, PatnaikL M.Adaptive probabilities of crossover and mutation in genetic algorithms[J].IEEE Transactions on Systems,Man,and Cybernetics,1994,24(4):656-667.
SpalartP R, JouW H, StreletsM,et al. Comments on the feasibility of LES for wings and on a hybrid RANS/LES approach[C]//Advances in DNS/LES:Direct numerical simulation and large eddy simulation. Seattle:Boeing Commercial Airplane Group,1997:1-12.
[15]
MenterF.Zonal two equation k-ω turbulence mo-dels for aerodynamic flows[C]//23rd Fluid Dynamics,Plasmadynamics,and Lasers Conference.Virginia:AIAA,1993:2906.
[16]
LeeD H, KimC J, HeoM J,et al.Development of real-time maneuver library generation technique for implementing tactical maneuvers of fixed-wing aircraft[J].International Journal of Aerospace Engineering,2020,2020(1):7025374.