High Fidelity Modelling for High Altitude Long Endurance Solar Powered Aircraft

Kurzfassung

High Altitude Long Endurance (HALE) platforms are the aerial platforms capable of flying in the stratosphere for long periods of time. This master thesis presents aircraft system identification procedures geared towards such fixed wing platforms where aerodynamic forces and moments are parametrically modelled with so-called stability and control derivatives. The first part of the thesis addresses local System identification procedures intended for controller synthesis at low altitude flights whereas the second part of the thesis deals with a preliminary study on a new global system identification method. The local system identification procedure is based on the two step method, which offers flexibility regarding the aerodynamic structure. Therefore, it is suitable for the development of a system identification tool chain for various fixed wing platforms. Various system identification experiments have been conducted to collect flight test data. The parameters for the estimation of aerodynamic forces and moments are then found through an optimization procedure. Such parameters have been validated using a validation set from flight test data and their applicability for controller synthesis has been demonstrated. Global system identification typically requires the collection of flight test data at multiple points in the flight envelope and often, is combined with extensive Computational Fluid Dynamics (CFD) solutions as well as wind-tunnel experiments. Such an approach is time consuming and costly. This thesis presents a new method to overcome the limitations of the current methodology by applying a Parameter search on VLM-based (Vortex Lattice Method) dynamic simulations of aircraft System identification manoeuvres and correcting the estimated models with available flight test data. The current study shows improvements in fidelity with decrease in Root Mean Squared Error (RMSE) by factor 0.2 and 0.5 for x-axis and z-axis forces in body frame respectively, while reducing the effort for obtaining a model with similar fidelity.