Result Validation through code Verification and Mesh Sensitivity Test

Abstract

Listen

Translate article

An investigation on the validation of the results from George and Kay, (2017) and the reproduction of results from Srynarayana et al. (2008) was conducted with the aid of COMSOL MultiPhysics through Code Verification and Mesh Sensitivity Test for code validation and correctness of the results. Mesh dependence test was conducted to examined temperature distribution and velocity field at the point (0,2) at the core of the plume and plotted as a function of time for different mesh sizes. The results confirmed that both velocity field and temperature distribution at the specified point converged as mesh size decreases. For code verification the results showed that values from the individual terms in the continuity equation vary slightly with a maximum variation of ≈ ±0.010 when taken their sum. Interpolated values showed that the Y-Momentum equation were resolved properly and positive on both sides of the equation: though, with a maximum variation of ≈ ±0.003. Fluid flow in the upward direction is very key for a rising plume because water at the plume’s core at that point is expected to be undiluted and positively buoyant. Results from the X-Momentum equation showed both positive and negative values with a maximum variation of ≈ ±0.011. The values from the temperature equation showed both negative and positive values but approximately zero in all. In reproducing the results from Srinarayana et al. (2008) with COMSOL MultiPhysics. The overall behaviour showed that results appear similar. A flapping behaviour was also observed which seems to have occurred earlier within the simulation time. However, the Comsol Multiphysics used is reliable as this could reproduce results from other CFD code (Gerris) while the validated results from George and Kay, (2016) are also reliable and independent of mesh size ≤ 0.05.