Define advecting pulse test problem. #463
Conversation
First attempt. With a22=0 and a32=0, it successfully reduces to SSP-RK2 scheme. However, with a22=1 and 0<a32<=0.5 (IMEX PD-ARS), the RadShock and Coupling test failed.
the Asymptotic Marshak test. - Fixed a problem in implementing IMEX PD-ARS: Egas and Fgas should update by 0.4 dt Src instead of dt Src in the first stage. - Add RadhydroPulse test problem
for more information, see https://pre-commit.ci
for more information, see https://pre-commit.ci
for more information, see https://pre-commit.ci
- Add the O(beta tau) terms to the matter-radiation exchange - Add the RadhydroUniformAdvecting test for this new physics - Modified boundary condition in RadhydroShock problem - Passed all tests
for more information, see https://pre-commit.ci
for more information, see https://pre-commit.ci
![github-actions[bot]](https://avatars.githubusercontent.com/in/15368?s=60&v=4){kind=small}
|
Almost ready for merging after Add-O(beta-tau)-terms is accepted. Before merging, I need to think about a more robust way to validate the result. Currently, I compare the solutions in advecting and non-advecting cases visually and confirm they are consistent. However, to make it more convincing, we need to compare the shifted advecting result with the non-advecting result in one figure. @BenWibking What's your idea on how to achieve this? Is it possible to run a simulation with v=0 via |
|
/azp run |
|
Azure Pipelines could not run because the pipeline triggers exclude this branch/path. |
As I think about it, this is probably not necessary. Radiation advection is strictly tested in the RadhydroUniformAdvecting test, so there is no need to validate it here. Since there is no analytic solution to this problem (as we are aware of), we can just keep it as it is. So, ready to merge after |
…astro/quokka into advecting_pulse_test_new
Yes, this is possible. You can create a new |
|
I think this PR is merging into the wrong branch... it says: "chongchonghe wants to merge 4 commits into Add-O(beta-tau)-terms from advecting_pulse_test_new" Do you mean to merge this into |
OK, same as the other one. I'll fix it. |
|
It looks like there's a timeout failure on moth: https://dev.azure.com/quokka-code/quokka-CI/_build/results?buildId=7786&view=logs&j=0410d313-00ef-550d-37c8-facbf7a1cb10&t=62d0f0fc-52cd-5dfc-847e-751a25f0b5e6 and a possible hardware failure on avatargpue: https://dev.azure.com/quokka-code/quokka-CI/_build/results?buildId=7783&view=logs&j=655dacad-9e02-5f72-e0cf-043ebfb7cecc&t=040ddf27-e411-58f7-ab97-dfe1d2a20f3a |
|
/azp run |
|
Azure Pipelines successfully started running 4 pipeline(s). |
|
/azp run |
|
Azure Pipelines successfully started running 4 pipeline(s). |
…astro/quokka into advecting_pulse_test_new
|
/azp run |
|
Azure Pipelines successfully started running 4 pipeline(s). |
…astro/quokka into advecting_pulse_test_new
…astro/quokka into advecting_pulse_test_new
|
/azp run |
|
Azure Pipelines successfully started running 4 pipeline(s). |
|
/azp run |
|
Azure Pipelines successfully started running 4 pipeline(s). |
…astro/quokka into advecting_pulse_test_new
|
OK. The RadhydroPulse test takes way too long to run in 3D on GPU, so I have to turn it off. I keep it on only when compiled with 1D. I'll test it with 3D GPU on gadi manually and ensure it gives the right result. /azp run |
|
/azp run |
|
Azure Pipelines successfully started running 4 pipeline(s). |
|
It looks like there's an instability present in the non-advecting velocity (see below), but that can be fixed in a separate PR. |
Description
Defined the advecting pulse test problem (Krumholz+2007). This test gives exactly the same result (visually) as in CASTRO II paper and consistent result between non-advecting and advecting case, demonstrating the code's ability to resolve diffusion limit as well as capture radiation advection in this limit.
Notes:
Checklist
/azp run.