Sensitivity Analysis for Global Parameter Identification. Application to Aerodynamic Coefficients

The parameter identification of grey-box nonlinear model structures remains a challenging task. Analytical models, constructed from physical laws, allow a complete description of the system behavior, with however a high complexity of the mathematical structure of the system. This paper presents an identification procedure adapted to nonlinear models and able to cope with the difficulties related to the model structure. This procedure, composed of several steps from model development to its validation, takes into account the identifiability of model parameters. It highlights the importance of performing sensitivity analysis of the output variables w.r.t. unknown parameters - input conditions and/or model parameters- to give insights on the choice of the experimental conditions able to provide sufficient information for the estimation step. These results are illustrated in the case of the identification of aerodynamic coefficients of a space probe based on free flight measurements. (C) 2018, IFAC (International Federation of Automatic Control) Hosting by Elsevier Ltd. All rights reserved.