Significant Difference in Results Using My Own Mesh in Benchmark Example
Hi everyone,
I’m running the provided benchmark example, and I also tried to generate my own mesh for the same case. However, I’m seeing a significant difference in the results when using my mesh compared to the provided one. I’ve reviewed my process multiple times but can’t figure out where the issue might be.
As shown below, the only difference between the two runs is the input mesh (vtu file). My custom vtu file is attached. Can anyone help me identify what might be causing this discrepancy?
Thank you in advance!
Weixin Image_20241014223536.zip (30.9 KB)
3D_basin_1.zip (60.4 KB)
Hi Shuoshuo,
I think you tried to simulate the Beier Sandbox Benchmark?
According to your results, something with the soil temperature looks incorrect. Do you specify all temperatures in Kelvin and only convert it to °C in the Postprocessing?
One possible problem could be the mesh size in z-direction with respect to the time step, they are not independent of each other. If the time step is small, a fine mesh resolution in z-direction is needed. Please try to run the simulation with your mesh and a tilmestep of 120s (currently is 60s) - maybe this fix the unreasonable solution. For timestep of 60s I recommend a mesh resolution of 0.5m in z-direction.
Best regards,
Max