In this thesis, a new credibility assessment framework is developed for computational wind engineering (CWE) simulations. The framework is mainly developed for testing code implementation correctness and estimation of the discretization uncertainty for eddy-resolving, and unsteady simulations. The framework is composed of two main milestones.
First, a modular and flexible procedure for code verification is developed with the ability of testing black box codes. The code verification procedure focuses on the consistency of the code implementation and convergence of field variables. The procedure for code verification consists of analytical benchmarks, either exact or manufactured, with increasing complexity to test the implementation of each term in the Navier-Stokes equation.
Second, the credibility assessment framework has a guideline for the quantification of discretization error/uncertainty. More precisely, guidelines are defined for solution verification.
The discretization error/uncertainty estimation is based on Richardson Extrapolation approach. A solution biased uncertainty estimator is used to account for using unstructured grids, non-uniform refinement, and non-asymptotic solutions. The newly developed framework has a new definition for the measurement of grid size, handling simulation data with anomalous behavior, and for the safety factor definition in the uncertainty quantification of the discretization error. The assessment methodology is suited to both well- and ill-behaved sequences of simulations.
The performance of the assessment methodology is checked with a glimpse on validation with experimental data. Finally, it can be concluded that the developed verification methodology is highly qualified to judge the quality of CWE simulations. Moreover, the generality and modularity of the framework makes it applicable to any software environment regardless of the discretization scheme. Consequently, the methodology encourages further research on the identification of the reliability of CWE simulations.
«
In this thesis, a new credibility assessment framework is developed for computational wind engineering (CWE) simulations. The framework is mainly developed for testing code implementation correctness and estimation of the discretization uncertainty for eddy-resolving, and unsteady simulations. The framework is composed of two main milestones.
First, a modular and flexible procedure for code verification is developed with the ability of testing black box codes. The code verification procedure...
»