Aeroservoelastic Model Modification and Uncertainty Quantification with Bayesian Posteriori Estimation

Abstract

A half model of a scaled aircraft is designed and tested in a wind tunnel. Based on the step-sinusoidal swept frequency test, the aeroservoelastic frequency response function is measured and compared with the predictions of a theoretical model. According to the discrepancies, the key uncertainty factors such as actuator model uncertainty, aileron aerodynamic uncertainty and structural damping uncertainty are investigated. Therefore, the original theoretical model is modified and updated with these model uncertainties considered, respectively. Motivated by the still existing discrepancies of experimental data and comprehensive updated theoretical outcomes, we take the above aeroservoelastic model uncertainty as an aleatory probabilistic one. Hence, the probability of each model uncertainty is quantified by Bayesian Posteriori Estimation theory. In addition, the uncertainty probabilities of aeroservoelastic response and aeroservoelastic stability boundary are predicted. The aeroservoelastic uncertainty quantification methodology with model uncertainty is validated by comparing the probabilistic predictions of the theoretical model set with experimental data. Taking model uncertainty into consideration, the present method is able to predict the probabilistic aeroservoelastic stability bounds and response as well. Results indicate that the updated comprehensive model is the best theoretical model with the highest posterior probability. In addition, the unsteady aerodynamics due to aileron deflection has the most significant effect on aeroservoelastic response.