Grey-box modelling of a non-linear aerodynamic system using genetic algorithms

This investigation presents a grey-box modelling approach for an experimental non-linear aerodynamic test rig, a twin-rotor multi-input-multi-output system (TRMS), using genetic algorithms (GAs). The dynamic equations of the system in terms of two degrees of freedom are developed using Newtonian method. The measurable parameters of the system are then measured and the rest are estimated using physical knowledge of the system. In order to improve the accuracy of the model, the estimated parameters are retuned using a GA-based optimization approach. For the sake of fast convergence, the estimated parameters are utilized as the initial populations of the GA. The performances of the white- and grey-box models are compared with respect to each other to validate the enhanced performance of the grey-box approach. The process is suitable for precise modelling of other non-linear systems to be used for the development of sophisticated controllers. However, it is noted that the model parameters are optimally found offline based on the real data recorded from the TRMS and when being used in control they will be static.