Learning Forum

[Ashen Perera]

Hi,

 

so I am modeling a shell and tube heat exchanger , with near wall inflation of 2.4 mm first cell height and a global mesh with element size of 1.8 mm initially. So basically for 1.8 mm and 1.6 mm the residual error drops below 1e-3 easily. But when making the global mesh further refined to around 1.2,1.4 mm the residual finds it hard to drop below 1e-3. But the results converge in the monitor plots to decent values comparable to literature work. The mesh metrics are also good , anything I need to think about to get the residual tolerance better 

[AG]

 It's good to hear that your results are converging and comparable to literature despite the residuals not dropping below 1e-3 for finer meshes. When you refine the mesh, it's possible that the solver requires more iterations to converge due to the increased number of elements and the complexity of the solution. Since your results are converging and are comparable to literature, and your mesh metrics are good, it suggests that your simulation is behaving as expected. Running simulation for longer time will sort out this issue. 

However additionally check for following details : 1. Investigate if there are any regions in the domain where the mesh might be too coarse or too fine, which could affect the residuals. 2. Make sure you are using proper volume mesh growth rate (<1.1) in all cases and accordingly 1st cell height is also changed. 3. Even though you mentioned the mesh metrics are good, it's worth re-evaluating the mesh quality in the refined areas. Ensure that there are no highly skewed elements or small faces that could be affecting the solution.

Remember, achieving a balance between accuracy and computational cost is key. It's not uncommon for finer meshes to require more iterations and computational power to reach similar levels of convergence as coarser meshes.