Hi! Our team is executing simulation runs through a python code for doing parametric analysis. We are doing parametric simulations for a range of inputs, weather file, wall-type, roof-type and others. We are using python scripts to trigger a .bat file for executing the simulation. Broadly speaking, this is what we are doing:
- The automation routine (in python) modifies a functional .inp (for the above stated parametric simulations).
- A python script updates a bat file with the location of the modified inp and applicable weather file. This script also executes the bat file.
- The bat file executes the modified inp to generate simulation output (.sim and other files) The automation routine is generating and executing several hundred inps in a single pass.
Interestingly some of the sim files are smaller in size. On reviewing these sim files it was found that the sim files were truncated. There seems to be no pattern to these truncated sim files. It is puzzling that one inp file will execute error-free for different parametrics, but not for some. For example, a file will throw a truncated sim for some locations but not for all. To test this, we did the following:
- Ran the inp manually for the given location. It ran error free and provided a complete .sim.
- Reviewed other files had run error free and generated complete sim for the same location.
- Looked for pattern, say every nth file to see if there were some hardware issues (memory overruns, etc), but no discernible pattern was found. Since file sizes are different, we understand that this is not a fool proof way of testing it.
Please see attached screenshot. 
.
On the left is the automated run on the modified inp and on the right is the same inp executed by browsing to file on eQuest.
Any suggestions on how to overcome this problem are welcome. Thanks in advance to all the community members here!
