search

ufz / ogs5

OpenGeoSys THM/C simulator version 5
https://www.opengeosys.org/ogs-5
Other
39 stars 97 forks source link

No advective heat transport using coupling of HeatConduction and DeformationFlow #162

Closed ThieJan closed 3 years ago

ThieJan commented 3 years ago

Testing the coupling enabled in https://github.com/ufz/ogs5/pull/159 with the Benchmark of P. Vogel from chapter 10.1 in the Benchmark book (Kolditz et al 18) I observed some deviations in the solution especially concerning the temperature profile. Here are the corresponding input file sets for the staggered coupling and the monolithic HM coupled with the T-process: T-HM.zip T_H_M.ZIP

The H-T-M staggered version is the one used by Peter which fits the analytical solution quite well. Looks like the advective heat transport is not included correctly.

ThieJan commented 3 years ago

After some debugging I added the initialization of a coupling pointer for this process coupling combination. https://github.com/ufz/ogs5/pull/163

However, the results still deviate. Debugging the T-HM example I noticed that in CFiniteElementStd::CalcAdvection() the loop runs over 27(=3x3x3) Gaussian points while the data contains valid velocities only for the first 8 points. I suspect this contribites to the deviations.

wenqing commented 3 years ago

Yes, With T-HM, the adevction computation in HM is not tested. I will take a look.