Estimation of steady and unsteady aerodynamic parameters at high angles of attack

The steady and unsteady aerodynamic parameters of the N-350 light trainer aircraft are estimated from simulated response data for angles of attack ({dollar}alpha{dollar}) up to 40{dollar}spcirc{dollar} and sideslip angles ({dollar}beta{dollar}) to {dollar}pm{dollar}15{dollar}spcirc{dollar}. Nonlinear aerodynamic models of forces and moments are identified by applying the Estimation-Before-Modeling (EBM) technique to simulated flight data for high angle of attack maneuvers including stalls and post-stall trajectories. For 0{dollar}spcirc le beta le{dollar} 10{dollar}spcirc{dollar}, the steady aerodynamic coefficients and significant stability and control parameters that are estimated match very well with the true values at most angles of attack, and the estimated unsteady aerodynamic parameters are in good agreement with the trend of the true model.; Unsteady aerodynamics at high angles of attack are formulated using a nonlinear indicial function representation with a time lag, and simulated indicial functions related to {dollar}alpha{dollar} and {dollar}beta{dollar} are incorporated into a six degree of freedom (6DOF) simulation. With these unique indicial functions, unsteady aerodynamic effects such as hysteresis can be simulated, and the results are quite reasonable compared to some experimental results for longitudinal motion. The modified identification algorithm has better ability to predict dynamic responses than does the estimator without indicial functions.; To check the estimation accuracy of the EBM technique and the program coding before applying this approach to the high angle of attack regime, parameters are estimated in the linear aerodynamic regime. Comparing these results with true model values and the outputs of maximum likelihood estimation, the EBM technique is generally superior to maximum likelihood estimation for the cases considered in this regime. An original approach for parameter sensitivity analysis is developed in the linear aerodynamic regime. The results show that parameter estimation accuracy is closely related to the magnitude of the parameter sensitivity function.; Other novel aspects of the modified estimation technique include the use of Maximum a Posteriori (MAP) identification for smoothing/filtering flight data. States, forces, and moments are satisfactorily estimated using a forward smoother and a backward information filter based on a variational technique which minimizes the MAP cost function.; The flight data at high angles of attack are partitioned into small subspaces of the {dollar}alpha{dollar}/{dollar}beta{dollar} domain after filtering and smoothing. In each subspace, low order aerodynamic models are used, and parameters are identified by a Stepwise-Multiple-Linear-Regression (SMLR) analysis.