Difference between revisions of "Project-Komtek-2020-Adhika-Satyadharma"
(Created page with " == '''Abstract:''' == The Grid Convergence Index (GCI) is a widely used method to analyze the discretization error of a Computational Fluid Dynamics (CFD) result. However, t...") |
|||
| Line 1: | Line 1: | ||
| − | + | == '''On the Effect of Iterative and Round-Off Errors to the Grid Convergence Index Calculation''' == | |
| − | == ''' | + | == Abstract == |
The Grid Convergence Index (GCI) is a widely used method to analyze the discretization error of a Computational Fluid Dynamics (CFD) result. However, this method is very sensitive to noise/error within the data. This error, which consists of iterative error and round-off error, needs to be minimized to ensure that it does not have a considerable effect in the GCI calculation. This research is trying to calculate the limit of both of these errors analytically, by modeling the effect of the error to the GCI calculation itself. The model is tested on a wide range of possible values in the range of 1 < r12, r23, f2/f1, f3/f1 < 2 and the iterative and round-off error within the range of 0.0001% to 10%. The preliminary results shows that both the round-off error and iterative error must be 3 order of magnitudes lower than the desired discretization error. | The Grid Convergence Index (GCI) is a widely used method to analyze the discretization error of a Computational Fluid Dynamics (CFD) result. However, this method is very sensitive to noise/error within the data. This error, which consists of iterative error and round-off error, needs to be minimized to ensure that it does not have a considerable effect in the GCI calculation. This research is trying to calculate the limit of both of these errors analytically, by modeling the effect of the error to the GCI calculation itself. The model is tested on a wide range of possible values in the range of 1 < r12, r23, f2/f1, f3/f1 < 2 and the iterative and round-off error within the range of 0.0001% to 10%. The preliminary results shows that both the round-off error and iterative error must be 3 order of magnitudes lower than the desired discretization error. | ||
Latest revision as of 08:19, 6 April 2020
On the Effect of Iterative and Round-Off Errors to the Grid Convergence Index Calculation
Abstract
The Grid Convergence Index (GCI) is a widely used method to analyze the discretization error of a Computational Fluid Dynamics (CFD) result. However, this method is very sensitive to noise/error within the data. This error, which consists of iterative error and round-off error, needs to be minimized to ensure that it does not have a considerable effect in the GCI calculation. This research is trying to calculate the limit of both of these errors analytically, by modeling the effect of the error to the GCI calculation itself. The model is tested on a wide range of possible values in the range of 1 < r12, r23, f2/f1, f3/f1 < 2 and the iterative and round-off error within the range of 0.0001% to 10%. The preliminary results shows that both the round-off error and iterative error must be 3 order of magnitudes lower than the desired discretization error.
