Aerodynamic design and optimisation of a maritime surveillance UAV

Abstract

The aerodynamic design, development, analysis and optimisation of an Unmanned Aerial Vehicle (UAV) for the purposes of surveillance of South African oceans, is presented in this research work. Low order aerodynamic methods, such as the Vortex Lattice Method (VLM) were used to define the initial sizes and to also further develop the UAV design before starting the Computational Fluid Dynamic (CFD) analysis. This was followed by detailed 3D modelling of the aerodynamic surfaces using Solidworks. CFD software, ANSYS Fluent was then used to analyse and determine the aerodynamic coefficients and performance in more detail. Before the detailed CFD work began, a study was performed to compare the results of the Spalart-Allmaras and the Transitional SST turbulence models. The results of the study showed that the Spalart-Allmaras model, which does not model laminar flow, showed a 42% larger viscous drag than the Transitional SST turbulence model, which models laminar flow, and the transition from the laminar to turbulent regimes. For the Reynolds numbers present in the current study, the viscous drag can be a significant portion of the overall drag, thus, the Transitional SST Turbulence model was selected to better capture the transition from laminar to turbulent flow, and to better predict the viscous drag. This matches the results of other studies such as (Chen, et al., 2020). The full CFD analysis was then carried out, the results of which were used as a baseline, and as insight to possible areas of aerodynamic improvement. An updated design was realised, using the insight from the initial analysis to change the aerofoil profile used for the wing to one which maximises the extent of the laminar flow regime, reducing drag. A winglet design was also incorporated in order to reduce the extent of the observed vortices off the wingtips, increasing aerodynamic efficiency. Beyond this, the updated design also incorporated design envelopes within the fuselage for more modular sub-systems, which also resulted in a shift of the Centre of Gravity (COG), and thus a shift in the location of the wing as well. This updated design realised a performance improvement of 13.37% in the endurance over the initial design. As a final step, ANSYS Fluent Discrete Adjoint Optimisation was used in an attempt to further reduce the drag and improve aerodynamic performance. The drawback of the ANSYS Fluent Adjoint Optimiser is that it does not currently have support for the Transitional SST Turbulence model used for the rest of the simulations. Thus, the fully laminar assumption was made for the adjoint calculations. Due to this restriction, only key areas of the UAV, that involved mostly laminar flow, were selected for the optimisation, keeping other areas constant. The decision was also made to only focus on a reduction in drag for the optimisation, as the lift of the UAV was already designed with cruising flight in mind, and an increase in lift would require a change in cruising speed, and thus a change in overall performance. The results were successful in reducing the drag by 0.25% using only small, laminar areas of the UAV. However, the optimisation process had an adverse effect on the lift, reducing it by 2.16% in an area not selected for optimisation, namely, the wing, which resulted in an overall loss in performance. This overall loss in performance is believed to be a result of a combination of the laminar assumption used for the optimisation, and the choice to only optimise discrete areas of the UAV, instead of the UAV as a whole. This resulted in the design iteration previous to the adjoint optimisation, namely, the improved design being selected as the final aerodynamic design. The improved design not only met but exceeded all the defined requirements.