The model selection must match the task to accurately capture relevant target quantities. A balance between minimum requirements and excessive complexity is essential. While more complex models increase accuracy, they also entail more effort and potential uncertainties due to additional parameters. Ultimately, model selection is a compromise between effort and accuracy, as highly complex models are often less robust due to numerous uncertain input parameters. -> Refer to the Dlubal technical article.
It is crucial to distinguish between validation and verification. Validation demonstrates that the correct equations are being solved, meaning the simulation solves the task accurately enough with the chosen model. In CFD applications, there is a distinction between "program verification" and "solution verification." Program verification aims to show that a software program calculates correctly under its conditions and solves the equations properly. Solution verification ensures the calculation is internally consistent, achieving a stable solution where expected effects occur and are no longer significantly model-dependent. While program verification is performed by the software manufacturer, solution verification is always done by the user.