Aerodynamic optimisation of aerofoils for martian rotorcraft using direct numerical simulations

Martian conditions present unique challenges for rotorcraft design due to the planet's lower atmospheric density and lower speed of sound compared to Earth. Consequently, Martian rotorcraft blades operate in a low-Reynolds-number compressible flow regime, a relatively unexplored domain atypical for terrestrial helicopters. Non-conventional aerofoils with sharp edges and flat surfaces have demonstrated improved performance under these conditions, inducing earlier separation and unsteady vortex roll-up on the suction side. Several studies have performed optimisation using Reynolds-Averaged Navier-Stokes (RANS) and Unsteady RANS (URANS) solvers. However, given the unsteady, vortex-dominated, and transitional nature of the flow, RANS/URANS solvers have limited predictive capability. This thesis overcomes these limitations by employing high-order accurate Direct Numerical Simulations (DNS) via PyFR to optimise aerofoils for Martian rotorcraft. First, DNS results are compared against experimental data, contributing to the understanding of Martian aerodynamics and demonstrating the utility of DNS in accurately resolving Martian flow over triangular aerofoils. Next, DNS shape optimisation of triangular aerofoils maximising lift and minimising drag is performed, revealing optimised aerofoils that achieve up to a 48% increase in lift or a 28% reduction in drag compared to a reference aerofoil previously studied experimentally. Flow characteristics of optimal aerofoils are analysed to elucidate flow physics that yield optimal performance. The sensitivity of these optimal aerofoils to free-stream perturbations is then explored by injecting eddies upstream of the aerofoils. Findings show that such disturbances lead to a breakdown of coherent vortices and a reduction in aerodynamic forces. This motivates an optimisation that includes free-stream eddies, revealing that optimal aerofoil shapes are sensitive to free-stream conditions. The work in this thesis constitutes the first use of DNS for aerodynamic shape optimisation.