This paper is aimed at showing the performances obtained with an open-source CFD code for heat transfer predictions after the addiction of specific modules. The development steps to make this code suitable for such simulations are described in order to point out its potentiality as a customizable CFD tool, appropriate for both academic and industrial research. The C++ library, named OpenFOAM, offers specific class and polyhedral finite volume operators thought for continuum mechanics simulations as well as built-in solvers and utilities. To make it robust, fast and reliable for RANS heat transfer predictions it was indeed necessary to implement additional submodules. The package coded by the authors within the OpenFOAM environment includes a suitable algorithm for compressible steady-state analysis. A SIMPLE like algorithm was specifically developed to extend the operability field to a wider range of Mach numbers. A set of Low-Reynolds eddy-viscosity turbulence models, chosen amongst the best performing in wall bounded flows, were developed. In addition an algebraic anisotropic correction, to increase jets lateral spreading, and an automatic wall treatment, to obtain mesh independence, were added. The results presented cover several types of flows amongst the most typical for turbomachinery and combustor gas turbine cooling devices. Impinging jets were investigated as well as film and effusion cooling flows, both in single and multi-hole configuration. Numerical predictions for wall effectiveness and wall heat transfer coefficient were tested against standard literature and in-house set-up experimental results. The numerical predictions obtained proves to be in-line with the equivalent models of commercial CFD packages obtaining a general good agreement with the experimental results. Moreover during the tests OpenFOAM code has shown a good accuracy and robustness, as well as an high flexibility in the implementation of user-defined submodules.

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