Application of Neural Networks Based Method for Estimation of Aerodynamic Derivatives

A Feed Forward Neural Networks (FFNNs) based modified delta (MD) method is recommended for estimating aerodynamic derivatives of an aero-elastic aircraft. The FFNNs is trained using differential variation of aircraft motion and control variables and coefficients, as the network inputs and outputs respectively. The FFNNs training is carried out using Lavenberg-Marquardt Back Propagation algorithm. The trained neural network is then presented with a suitably modified input file and the corresponding predicted output file of aerodynamic coefficients is obtained. An appropriate interpretation and manipulation of such input-output files yields the estimates of the aerodynamic derivatives. The method is applied on the simulated flight data of two configurations of an aero-elastic aircraft for the parameter estimation. The FFNN based technique is also applied to validate the estimated aerodynamic derivatives. The results suggest that the FFNN based MD method can advantageous be used for estimation of the aero-elastic aircraft derivatives.