In English

Development of a Test Suite for Verification & Validation of OpenFOAM

Simon Persson
Göteborg : Chalmers tekniska högskola, 2017. Diploma work - Department of Applied Mechanics, Chalmers University of Technology, Göteborg, Sweden, ISSN 1652-8557; 2017:59, 2017.
[Examensarbete på avancerad nivå]

Computational fluid dynamics is an important tool which can be used to simulate various properties of a flow. CFD simulations are used within the nuclear power industry to aid in the evaluation of thermal loads within a given system. In order to verify and validate some aspects of the open source solver OpenFOAM a test suite was developed. The main goal was to build a series of automated tests while at the same time enable future additions of new tests. The verification process focused on the evaluation of certain numerical schemes used for both incompressible and compressible solvers in OpenFOAM. In order to evaluate the results of each case two different test criteria where used. The difference between simulated results and exact solutions to a specific problem was used to calculate a discretization error for the whole computational domain. The discretization error can then be used together with a refinement of the mesh to calculate an observed order of accuracy. In order to obtain exact analytic equations to the governing equations the method of manufactured solution was implemented. A solution for the equations was manufactured and a resulting source term was implemented in the simulation. The results from the simulation should then correspond to the manufactured solution. It was found that the results for the order of accuracy could be used as a test criteria when evaluating numerical schemes and solvers. A validation test case was also implemented into the test suite. The simulation was based upon an experiment previously used as a benchmark for various CFD simulations. It was found that a coarse mesh with a limited amount of cells could be used for evaluation of thermal mixing. In order to determine if a test is accurate enough a validation metric evaluated from the difference between the results from simulation and the experimental data. The metric was defined from the discretization error also used in the verification cases. The error was evaluated at the points where experimental data was available.

Nyckelord: Verification, Validation, OpenFOAM, CFD, MMS, Manufactured Solution

Publikationen registrerades 2017-06-29. Den ändrades senast 2017-06-29

CPL ID: 250253

Detta är en tjänst från Chalmers bibliotek