RTPR fail for TAMU 500 bus system

[bjpalmer]

I ran a quick test of the RTPR code to see if it still works. The 14 bus test case is fine but the TAMU 500 bus test case fails immediately. This may have implications beyond RTPR. The error that is getting thrown is coming from the Cblock class setcoeffs method and is occurring because the esst1a model is somehow setting the Tf variable to zero between the time it is read in from the dyr file to the time the setcoeffs function is called in Esst1aModel::init. I think there may be a memory corruption problem in the dynamic simulation module.

[bjpalmer]

@abhyshr, @yliu250, @chenyousu, do you know how to run the TAMU 500 files in the RTPR directory as a standalone dynamic simulation? I'd like to use valgrind on it to see if there is a memory corruption issue.

[wperkins]

I think this is the same problem I reported with the TAMU 500 data files used to test the Python HADREC application (#198).

[bjpalmer]

I tried running dynamic simulation standalone code using the case9.raw file and the case9_GENSAL_ESST1A.dyr file. I used the input_9b3g.xml file and just replace the initial configuration and dyr files with these two files. This throws the same error running on 1 or 2 processors.

[abhyshr]

Need the error and the parameters in the dyr file to debug.

[bjpalmer]

The raw file and dyr file are both available in applications/data_sets directory. The input file was hacked from the input_9b3g.xml file (also available from the data_sets directory) by replacing the original raw and dyr file by the case9.raw and case9_GENSAL_ESST1A.dyr files. The error is an exception being thrown in the Cblock::setcoeffs method, which is called by the LeadLag::setparams method.

I made a mistake thinking setcoeffs was was being called in the init method with the value Tf. It is instead being called in the setparams method. The problem looks like the Tc or Tc1 variable is zero and this is triggering the exception. It doesn't look like a memory issue at this point, but more like a straightforward modeling problem.

[bjpalmer]

This is still an issue. It is a problem for running the dynamic simulation test files for individual devices.