[pak_sargon]

Hello everyone, Im doing a research on the hydrogen-diesel dual fuel for compression ignition engine for my undergraduate thesis using Converge simulation. The model was fixed properly and correctly until no issue addressed in the diagnosis. For hydrogen injection, i use inflow boundary condition with the real geometry added to the head using CAD and diesel using spray physical modeling function provided in the software.



However, the case setup shows two warning which are:
1) There may be an alignment issue between boundary 2 and boundary 3. Normal direction of triangles connected to boundary 3 should be perpendicular to direction of motion.
2) There is no dirichlet bc at either inflow and outflow for pressure at all.





For the sake of checking that the simulation can run perfectly, the embedding and AMR was disable to save the computational load. The timeline of my simulation goes like this:
hydrogen injection start:-137 degree, end: -105 degree)

diesel spray start: -9 degree, end 20 angle degree).



At crack angle of -87, the simulation failed with an error that says:
"(Rank 0) FATAL_ERROR: [ 681] polygon vertex is out of box"






What is the cause of this problem?

[usamaubk]

for the first case the problem may lie with the orientation of normals, you can try to fix them by going into Transform-normal and try to fix all normals or maybe run diagnosis dock and see problem triangles and try to fix their orientation.. for 2nd what is your BC at outflow?

[pak_sargon]

From the diagnosis dock, there is no error at all except when I include the error of small angle, there will be (2) error. Other than that, no error. For second case, I didnt include any outflow as I only interested in the compression and expansion engine stroke. So no outflow out of the cylinder. The inflow that I mentioned before was the inlet of the hydrogen injector, since it is gas so cannot use the spray physical model.

[Xiao Ren]

Hi Fauzan,



From your description, the crash is possibly due to the alignment issue, as shown by the 1st warning:
1) There may be an alignment issue between boundary 2 and boundary 3. Normal direction of triangles connected to boundary 3 should be perpendicular to direction of motion.


To fix this alignment issue, you can go to Studio and use the Geometry tools to recreate the triangles of bound 3, which should be aligned perpendicularly to direction of motion. The detailed procedures can be found from our training slides: Preparing the Piston and Valves for Motion, which can be download from https://hub.convergecfd.com/downloads/.


This issue can be quickly checked by changing check_grid_motion_flag from 0 to 1 in inputs.in. It will only check the grid but not solving the hydrodynamics.


If further assistance is needed, you can send an email to support@convergecfd.com (US); supportEU@convergecfd.com (EU) or support.in@convergecfd.com (India).

[pak_sargon]

Thank you Xiao, we managed to resolve the issue. Yeah it was something related to the geometry not the case setup. Turns out the direction of the compression stroke was wrong (move sideways instead of up and down), so thats why it failed. We not really sure as we already specify the direction to be (0, 1, 0) in the compression ratio setting but it didnt save everytime we run the simulation, and move in (0, 0 , 1) like I mentioned earlier.



So, we decided to rotate the cylinder to align with z axis and it solved our problem!