Abstract: Hypersonic aerothermal simulation is conducted in a blunt flow, and its grid dependency and iteration convergence are investigated by means of a computational fluid dynamic (CFD) method. The grid effect is crucial to aerothermal simulation. The study demonstrates that heat transfer prediction is very sensitive to grid spacing in the normal direction near the wall. The value of heat transfer exhibits dramatic differences of several times, or even an order of magnitude depending on grid spacing. It is shown that a computational mesh which results in accurate surfacepressure does not guarantee accurate heattransfer values. The authors hold the view that the great temperature gradient near the wall results in grid sensitivity for heat transfer simulation. Aeroheating computation is converged significantly slower than the convergence of pressure and flow field. The difference of the value of heat transfer with the last convergence value is as high as 20% when the pressure and flow field is fully converged. Therefore, in aerothermal numerical simulation it is not appropriate to take the falling orders of equation residual and pressure convergence as heat transfer convergence criteria. In order to get the convergence solution it is necessary to monitor the convergence of heat transfer data directly.

Key words: computational fluid dynamics, aerothermal, numerical simulation, grid, convergence, hypersonic