On the numerical calculation of stagnation point heat transfer using the k-epsilon model of turbulence

Numerical simulations of turbulent heat transfer near a stagnation point are performed in many studies. It will be shown that the k-epsilon model for these flows gives unrealistic results. Using the Boussinesq hypothesis for the eddy viscosity results in positive source terms in the equation for the kinetic energy of turbulent fluctuations (k) for both accelerating and decelerating flows, which is contradictory to the exact equation for k derived from the Navier-Stokes equations. The k-epsilon model gives a sharp peak of k just before stagnation and an overestimated heat transfer in the case of an impinging jet on a flat plate. Some investigators altered the k-epsilon model in such a way that the results looked better. A better alternative is to abandon the k-epsilon model for these types of flows and allow anisotropy in the stagnation region into the model.