Wallace Morris, David Friedlander, H. T. Huynh, Xiao-Yen Wang
This study juxtaposes the accuracy of a commercial computational fluid dynamics (CFD) software, ANSYS Fluent, against two CFD research codes, EZ4D and Flux Reconstruction. The study uses three problems to make the comparison, two vortex transport problems and one Ringleb flow problem. Fluent was used to simulate all problems, however, data sets from each CFD research code were not available for all cases. The first vortex problem showed that quadrilateral meshes perform better in Fluent than triangular meshes when the flow is aligned normal to quadrilateral cells. In addition, it revealed that Fluent has more dissipation error but less phase error than EZ4D. Furthermore, it revealed that Fluent has less velocity error than EZ4D, but this error was measured using a metric that is biased towards phase error. Fluent and EZ4D had a similar order of accuracy for this problem. The second vortex problem showed that Fluent has more velocity error as well as density error and has a lower order of accuracy compared to Flux Reconstruction. The Ringleb problem showed that EZ4D was able to establish fully developed Ringleb flow on grids coarser than Fluent; however, once Fluent established fully developed Ringleb flow, it had entropy error approximately an order of magnitude less than EZ4D. In addition, the problem showed that the entropy error in the Flux Reconstruction simulation was approximately 2.5 to 3.5 orders of magnitude less than either EZ4D or Fluent.
Scholarly Commons Citation
Leake, C. (2016). Testing the Accuracy of Commercial Computational Fluid Dynamics Codes: Vortex Transport by Uniform Flow and Transonic Ringleb Flow Problems Using ANSYS Fluent. , (). Retrieved from https://commons.erau.edu/pr-honors-coe/5