Failed check of initial species summing to 1

[sjlienge]

Dear PeleC community,

I would like to run the TGReact case in the RegTests to see how the TGV works in PeleC.
For that I have converted the LiDryer.cti to .yaml format and generated the cantera/profiles.dat file.
When I now want to run the executable I receive the following error:

amrex::Abort::0::Error:: Failed check of initial species summing to 1 !!!

I have tried a few things, e.g. setting all mass fractions defined in the prob.H (line 69-80)

#ifdef USE_FUEGO_EOS
      massfrac[H2_ID] =
        prob_parm.d_input[prob_parm.input_H2_init + m * prob_parm.nvars] * f0 +
        prob_parm.d_input[prob_parm.input_H2_init + mp1 * prob_parm.nvars] * f1;
      massfrac[O2_ID] =
        prob_parm.d_input[prob_parm.input_O2_init + m * prob_parm.nvars] * f0 +
        prob_parm.d_input[prob_parm.input_O2_init + mp1 * prob_parm.nvars] * f1;
      massfrac[N2_ID] =
        prob_parm.d_input[prob_parm.input_N2_init + m * prob_parm.nvars] * f0 +
        prob_parm.d_input[prob_parm.input_N2_init + mp1 * prob_parm.nvars] * f1;
      eos.PYT2RE(p, massfrac, T, rho, eint);
#endif

to

#ifdef USE_FUEGO_EOS
      massfrac[N2_ID] = 1.0;
      eos.PYT2RE(p, massfrac, T, rho, eint);
#endif

I think the problem is very trivial but I dont see it.
Thank you in advance already.

[marchdf]

Did you set the other mass fractions to 0 when you set N2 to 1? You need to make sure they are all initialized to something.

[sjlienge]

massfrac is initialized as:
amrex::Real massfrac[NUM_SPECIES] = {0.0};
I have tried a very blunt:

      massfrac[H2O2_ID] = 0.0;
      massfrac[HO2_ID] = 0.0;
      massfrac[OH_ID] = 0.0;
      massfrac[O_ID] = 0.0;
      massfrac[H_ID] = 0.0;
      massfrac[H2O_ID] = 0.0;
      massfrac[O2_ID] = 0.0;
      massfrac[H2_ID] = 0.0;
      massfrac[N2_ID] = 
        1.0 - ( massfrac[H2_ID] + massfrac[O2_ID] + massfrac[H2O_ID] + massfrac[H_ID] 
              + massfrac[O_ID] + massfrac[OH_ID] + massfrac[HO2_ID] + massfrac[H2O2_ID] );

But that also does not seem to work :(

[baperry2]

Are you compiling with USE_EXPLICIT_REACT = TRUE so Eos_model = Fuego? If not, that would explain the error you are seeing.

This case is untested and appears to have gotten rather dusty though, so you are likely to encounter some things that need to be modified to get it to run (e.g., it still uses integers rather than strings to specify pelec.chem_integrator).

[baperry2]

See #793 for the full changes required to get this case to work.

[sjlienge]

Thank you so much for looking into this so quickly! I wanted to reproduce the validation from the wiki. The case is running now and I transferred the residual changes from the issue to this setup. I will still be careful with it.