Flight trajectory modelling and flight path prediction of an unmanned aerial vehicle for the New Zealand Defence Force and the Defence Technology Agency.

Abstract

Volcanic ash has the potential to significantly damage aircraft components, leading to engine failure on a catastrophic level. Between 1953 and 2009, there were 129 reported incidents of aircraft encountering volcanic ash. During the Eyjafjallajökull eruption in 2010, the International Civil Aviation Organisation released recommended guidelines for all aircraft to take all actions to avoid volcanic ash. The NZDF provides aerial safety analysis and key information for civil aviation in and around New Zealand airspace, requiring key data around volcanic ash quantities in the atmosphere and its composition.

The motivation for this thesis was to gain understanding and develop the tools required to analyse the capability of a low-cost UAV used by the NZDF to sample high-altitude volcanic ash during a natural disaster. This thesis addresses aerodynamic analysis, flight path modelling, flight simulation, and key environmental considerations. This research built on work already conducted by the DTA, around design and build of an initial prototype, with some initial wind tunnel testing to measure aerodynamic coefficients and predict flight performance. This was completed with support from the University of Auckland.

The flight model was separated into two distinct phases, the ascent phase, and the descent phase. The ascent phase described a high-altitude weather balloon with the UAV as a payload. This is launched and ascends uncontrolled through the atmosphere until it reaches the altitude to release the UAV. Once the UAV is released, the descent phase begins. The descent phase describes flight of the UAV in one of three scenarios investigated. The first scenario was one where the UAV is released at a position determined by the results of the ascent phase and does not use any additional motorised thrust. The second scenario adds thrust to the first scenario and maintains the same release position. The third scenario increases the distance at which the UAV is released. This long-range scenario releases the UAV at a distance that is possible but unlikely, to test the UAV’s capable range. Motor thrust is also included in this scenario.

The results of the simulation were positive and supported the accuracy of the model. The results were split into the same ascent and descent phases and described not only the result of the simulations, but also a statistical analysis of the repeated simulations. The average final position during the ascent phase over 200 simulations was 24.77km North-East of the launch position at 30,000ft. This took an average of 33 minutes and 12 seconds to reach. During scenario one of the descent phase, the average landing distance from the target location was 588m, with a flight time of 19 minutes and 13 seconds. Scenario two had an average landing position of 598m, with an average flight time of 17 minutes and 49 seconds. Finally, scenario three had an average landing position of 14m, with an average flight time of 98 minutes and 11 seconds. With a goal of the UAV reaching within 1km of the target location, it can be concluded with good confidence that this can be achieved.

Following this, validation was required to justify the result and prove the accuracy of the model. The average error between the simulation and the flight data during turning manoeuvres was 7.6%. Furthermore, the error in maintaining head was small. This is important to understand the number of iterations required to achieve an accurate answer. The convergence study concluded that for the ascent phase, it is recommended to complete 100 iterations, and 200 iterations for the descent phase. The subsequent error in the ascent phase is below 0.04%, and the descent phase below 5%, while also minimising the computational expense of running numerous simulations.

The primary motivation of this research was to provide the NZDF and the DTA with a means to determine whether the use of a UAV is viable to collect data during natural disasters such as volcanic eruptions. The results and their validation have provided a strong foundation for the NZDF to develop and assessment process for using a UAV for high-altitude ash sampling. Another key benefit of this research was to combine complex aerodynamics with easy-to-use modelling techniques, to make a flight prediction model readily accessible, providing current and accurate results. This provides benefit to not only the NZDF, but also the wider aviation industry. Assessing the results and conclusions of this thesis, it can be stated with good confidence that the motivation of this research has been achieved with good accuracy.