Rationale of Comparison in a Single CFD code

Abstract

An important component of any computational method research project must always be a comprehensive assessment of all new findings to ensure that meaningful and reliable conclusions can be obtained. This was also attempted in previous collaborative EU-funded turbulence modelling research projects such as Haase et al. (2009) and Schwamborn and Strelets (2012) in which a variety of novel ideas and models dedicated to solving specific issues of hybrid RANS-LES modelling were developed. While different partners were able to demonstrate considerable improvements with their developed method in their respective CFD codes, cross-code comparison and thus a more general assessment was often more difficult to achieve despite carefully selected validation test cases. Even when potential setup uncertainties such as boundary conditions, grid resolution and user input parameters were eliminated, differences of the numerical infrastructure still persisted and obscured clear conclusions.