Star-CCM
This page contains resources for running and trouble shooting Star CCM+ on our MATLAB sim queue. For a detailed overview of Computational Fluid Dynamics (CFD), its theory and best practices, see the CFD section of the Aero Handbook.
Sim Computer Connection (RDP)
As of the F26 design cycle, the aero department of GTMS relies on two CPU's operated out of our own shop space for running CFD simulations. This is a big alternative to using a cloud-based computing cluster like PACE that requires the use of slurm and server based file transferring. Computing clusters are often faster, but more expensive and (in the case of the school's computing cluster) unreliable. The school system works off of research tokens, and without them the cluster can't be accessed. Using our own computers with a POD STAR-CCM+ license allows for unlimited access to both running sims and analyzing them locally. There should also be work towards integrating these local CPU's with a dedicated GPU to reduce solver time.
The two CPU's are affectionately named "Murray" (like Gordon) and "Newey" (like Adrian). Both computers have remote desktop connection (rdp) enabled from IT. To connect to either computer, use your GT credentials and the following names:
Newey: me06w1710lab07.ad.gatech.edu
Murray: me03w1710grd04.ad.gatech.edu
There are two methods for connecting to the Sim PCs:
Jump Desktop
Jump Desktop is straight forward to use and will drop you into the last person who used the computer (assuming they didn't log out of their Windows account), allowing anyone to see the Sim Q run. Unlike Remote Desktop Connection, the Georgia Tech VPN is not necessary. Generally, Jump Desktop is preferred.
Use the following credentials to log into Jump Desktop:
email ID: gtms@gatech.edu
password: 070-082-007
Remove Desktop Connection (RDP)
RDP is a bit more robust than Jump and allows for multiple users at once on a single computer. Use your personal GT login for RDP. You must be connected to the Georgia Tech Global Protect VPN. When using RDP, if the Sim Q is running on your account, be sure to lock your account rather than signing out; signing out will kill the Sim Q.
Monitoring Runs (Connect to Server)
Our sim queue utilizes command-based execution of meshes and solutions to eliminate the use of the GUI and thus improve runtime efficiency. However, some runs do require constant monitoring as it moves through each iteration. You can achieve this through the server function. Our computers are not server-based, in that the only "server" we connect to is the host computer itself. To do this, go to File -> Connect to Server. The default host and port should be sufficient to find the currently running .sim file. Hit Scan -> Ok and you should see the active .sim file in the GUI.
To Disconnect, do not exit the program normally. Instead, got to File -> Disconnect. This will safely close the GUI without risking halting the solution or mesh.
MATLAB Sim Queue
How to Run the Queue
Various Scripts
add.m
add FXXYYY- adds already-prepared runs
aeroload.m
- outputs forces from a run, rather than coefficients
arcNums.m
- uploads run numbers to SharePoint
arcPeri.m
- uploads run Periview files to SharePoint
arcRuns.m
- uploads run folder to SharePoint
cleanRuns.m
comp.m
comp FXXYYYY FZZ####- similar to compareRuns.m, but for two runs
compareRuns.m
- compares any number of runs through several parameters, at both part and vehicle level
copyJavaFile.m
- copyJavaFile(original directory, target directory, filename.java)
- Copies a java file from one directory to another
cq.m
cq- clears the active queue
- returns user SharePoint directory
loadPeri.m
- In an effort to move away from SharePoint, Peri files are now scattered throughout sim folders on Argo. Loading these files from Argo is slow and tedious, so this script will search for the Peri files of a given run(s) and download them to C:\periview on your local computer. If Periview is not already installed, it will install it. It can also run Periview directly from Matlab. it will give the option to delete old Peri files to avoid unwanted files.
makeRun.m
- generates run files from baseline run
makeTemp.m
mq.m
mq- displays mesh queue
q.m
q- displays solve and mesh queue
runMeshQ.m
- runs mesh queue
runQ.m
- runs queue with the runs in the queue
- runs can still be added to queue once it is initialized
saveInit.m
status.m
status
- continuously displays solve and mesh queue
- discontinued by hitting
ctrl+c(this command is true for MATLAB in general) - automatically run when from runQ.m command
getNumCores.m
- Uses
np.matto track total number of cores and current amount of available cores - Grabs percentage of available cores based on requested action (mesh, solve, postpro, etc.)
- Helper function used in
runQ.m,runMeshQ.m,runPeriTemp.m
Common Errors
File Related Errors
If the scripts error due to not being able to find a file, ensure the account which is being used has the user's GT OneDrive mounted to the file explorer. getSharePointDir.m uses the system variable "OneDriveCommercial" , which will return no value should your GT OneDrive not be accessible. Additionally, getSharePointDir.m only looks two levels deep within your OneDrive for the Documents - GT Motorsports folder from SharePoint, so if this folder is nested within more than two folders from your OneDrive, it will return no result and the script will error.
The scripts assume the following paths:
from Sharepoint:
Periview: ...\Documents - GT Motorsports\Aero\CFD\Periview
CFD Numbers: ...\Documents - GT Motorsports\Aero\CFD\CFD Numbers
CFD Archive: ...\Documents - GT Motorsports\Aero\CFD\CFD Archive
Meshing Errors
Overlapping Surfaces
The mesh may fail if two surfaces overlap each other or intersect each other.
Sim Queue Errors
unrecognized function or variable 'bslsimtype'
This error appears when creating a new run with a baseline not in the CFD folder. This can be because the sim is not loaded into the computer being used in the correct folder, or it does not exist. To fix this, ensure that the sim folder, with the .sim file, is stored locally on the CFD folder (not the CFD folder in SharePoint).
Changing Thing in Star
Radiator/Fan Setup
Post-Processing
Post-processing CFD, or examining the simulation's results, can be broken up into two categories, qualitative and quantitative, as discussed in the aero handbook. For quantitative data, such as overall and by-part Cl, use comp.m or compareRuns.m, as described above. For qualitative data, see the Periview section below.
Periview
Periview is our qualitative post-processing tool, used to look at various plots from data outputted by Star. While you can download the software locally following these instructions, it is also possible to run it from SharePoint, assuming SharePoint is mounted to your computer's file explorer.
