Numerical methods for unsteady conjugate heat transfer

The aviation and energy industries are heavily dependent on gas turbines and devolve great efforts at reducing their impact. A key objective is enhancing turbine performance and efficiency, which requires advanced cooling systems. Computational fluid dynamics (CFD) is widely used to simulate turbomachinery flows, but detailed simulations remain prohibitively expensive for gas turbine design. Low-fidelity simulations, in which most of the flow complexity is modelled or averaged, remain the industry standard but result in inaccuracies. Because of that, additional experimental work and larger safety factors are required to ensure turbine integrity. Better predictions can be obtained by jointly simulating the fluid flow across the turbine and the heat conduction in its blades (called conjugate heat transfer or CHT), but a significant time-scale disparity between aerodynamic phenomena and unsteady solid conduction typically restricts its use. Methods have been developed that can accelerate the computation time of the flow field by up to 2 orders of magnitude by treating it in the frequency rather than time domain. The Favre-averaged non-linear harmonic (FNLH) method developed by Prof di Mare's research group is a prime example. Its formulation opens opportunities for treating the solid conduction problem in a similar approach to the fluid flow, which could discard the time-scale disparity issue and make unsteady CHT simulations viable for academic and industry use. This project focuses on extending FNLH to unsteady CHT problems and assess the academic and industrial viability of the developed frequency-domain method.