Influence of the nodalisation on the results

2 posts / 0 new
Last post
M. Haag (Uni.Lu)
Influence of the nodalisation on the results

Dear ATHLET-Team,

I am using ATHLET 3.1A on a Windows machine.
I tried to do a “grid sensitivity analysis” for my model. Therefore, I increased the number of nodes in just one TFO until I’ll get a solution which doesn’t vary anymore in function of the number of nodes. Unfortunately, I couldn’t reach a stable solution. I attached a screenshot of the mass flux signal by varying the number of nodes (5, 20 and 40) of a 2m straight TFO.

Do you have any recommendations for the nodalisation to get the best physical results? Or how to do a 1-D grid sensitivity analysis?   

 

Best regards,

Michel

Athlet-Admin

Dear Michel,

Your problem and the attached figure are difficult to assess without any further information on the application, the boundary conditions etc..

Typically, the appropriate nodalization is case specific, so that a general recommendation cannot be given. Sometimes, the consideration of the l/d ratio helps to find a reasonable grid resolution.

For further investigation of your question, it would be helpful, if you could provide details of your application and the related dataset (if possible). You are welcome to send these information by email: philipp.schoeffel@grs.de . The forum messages are visible to everybody.

Best regards,

Philipp