Abstract
Towards complex architecture and light-weight structures, conventional wind load estimation techniques fail. Hence, Computational Wind Engineering (CWE) plays a crucial role. CWE can be a very helpful tool in all design stages from schematic to detailed design stage. With the current advances in CWE, numerical wind tunnel simulations has a great potential towards improving the structural design quality through better understanding of the wind loads on structures. However, the quality of simulations is still questionable. Despite the increasing attention given to the quantification of error and uncertainty in CFD, the techniques that have been developed for general fluid engineering problems to assess the quality of CFD simulations are still marginally used in CWE (Jrg Franke, 2010). This paper is part of a project aiming at defining a framework to assess the predictive capability of wind load computation using CWE with error estimation. The framework consists of three main work packages: Code Verification, Solution Verification, and Validation. Overall, the generic definition of the framework is stepwise exemplified with the open-source code KRATOSMultiphysics. In this paper, Code Verification is the main concern. The Method of Manufactured solutions is used to verify the Variational Multiscale (VMS) element in KratosCFD incompressible Navier-Stokes solver. Taylor Green vortex is the basis for the verification test. The Taylor-Green vortex is a well studied test problem for large eddy simulation (LES) subgrid scale models. Moreover, Taylor-Green vortex is modified to have more extensive testing for the code. Finally, a second order convergence rate is observed which verifies the tested code functionality, then numerical errors are to be quantified.