Robust identification of nonlinear aerodynamic model structure

A robust nonlinear identification technique based on Minimum Model Error (MME) optimal estimation, is used to identify nonlinear aerodynamic model structure. Data was obtained from digital simulations of longitudinal flight of a McDonnel Douglas F4C type aircraft with a standard linear model plus a nonlinear pitching moment derivative. The base model used in the algorithm knew the linear portion of the truth perfectly, but knew nothing of the nonlinear portion (e.g., where the nonlinearities were buried, what forms they took, let alone the values of any coefficients and exponents involved, etc…). The algorithm detected that some nonlinearity was present, determined where it was, determined the functional form, and very accurately parameterized the functional form. The test was performed twice, once with clean data and once with nominally white gaussian noise added to both input and output measurements. In both cases, the identification was excellent. The main objective was to identify the nonlinear pitching moment (C_m) aerodynamic coefficient. Pertinent identification results for the records analyzed including time history comparisons between the simulated flight data and computer-generated responses using the resulting model estimates are presented.