Simulating full-coverage film cooling remains an elusive task for aerodynamicists given the small scale of the holes relative to the duct size where the holes are applied. Source term models were developed to simulate the effect through a perforated surface; however, the documented approaches failed to adequately describe how source term locations within the computational domain were selected. This paper presents a continuous ‘checker-board’ surface function that enables a distributed selection of cells where the source terms are applied; furthermore, the source term strengths applied to cells within a given hole are weighted. A 3:1 aspect ratio S-duct with an 1.5 area ratio exhaust diffuser, and 4% porosity applied to the upstream convex bend was evaluated. Steady-RANS obtained with the realizable k-ε model and source terms derived based on the approach of Andreini et al. (2014) had good pressure distribution, outlet velocity, and coolant mass flow agreement with respect to experiment when the hole diameter was resolved with two nodes. Reducing the computational domain element count by 75% and simulating hole diameters 2.8-times larger with 4% surface porosity gave back pressure and outlet distortion coefficients within grid uncertainty of the finest grid solution; however, local-convex-surface-averaged quantities showed grid dependency.

This content is only available via PDF.
You do not currently have access to this content.