RMG running get "xxx conversion: nan"

[shenghuiqin]

Topic

General area which your question is related to.

• Installation of RMG
• Running an RMG job
• Using RMG API
• Arkane (formerly CanTherm)
• Dependencies
• An error message

input file

#Data sources
database(
    thermoLibraries =['BurkeH2O2','FFCM1(-)','thermo_DFT_CCSDTF12_BAC','CBS_QB3_1dHR','DFT_QCI_thermo','primaryThermoLibrary'], # 'FFCM1(-)','primaryThermoLibrary', 'BurkeH2O2','DFT_QCI_thermo','CBS_QB3_1dHR'
    reactionLibraries = [('2005_Senosiain_OH_C2H2',False),('Glarborg/C3', False)], # 
    seedMechanisms = ['BurkeH2O2inN2','FFCM1(-)',], #
    kineticsDepositories = ['training'], 
    kineticsFamilies = ['default'],
    kineticsEstimator = 'rate rules',
)

#Constraints on generated species
generatedSpeciesConstraints(
    maximumRadicalElectrons = 2,
    allowed=['input species','seed mechanisms','reaction libraries'],
    maximumCarbonAtoms=9,
    maximumOxygenAtoms=4,
)

#List of species
species(
    label='C7H16',
    reactive=True,
    structure=SMILES("CCCCCCC"),
)
species(
    label='O2',
    reactive=True,
        structure=adjacencyList(
        """
        1 O u1 p2 {2,S}
        2 O u1 p2 {1,S}
        """),
)
species(
    label='N2',
    reactive=False,
    structure=SMILES("N#N")
)





#Reaction systems
simpleReactor(
    temperature=[(600,'K'),(1500,'K')],
    pressure=[(10.0,'bar'),(40.0,'bar')],
    nSims=8,
    initialMoleFractions={
        #"C7H10": 1,
        "O2":   0.10885, # phi=1 means 9.5 O2 per C7H10
        "N2":   0.881684808, # 8.1 times as much N2 as O2
        "C7H16":0.00946, #  
           
    },
    terminationTime = (1.0, 's'),
    terminationRateRatio = 0.01,
    terminationConversion={
                'C7H16': 0.99,
        },
)

simulator(
    atol=1e-16,
    rtol=1e-8,
)


model(
    toleranceKeepInEdge=0, # No pruning to start
    toleranceMoveToCore=0.4,
    toleranceInterruptSimulation=1,
    maxNumObjsPerIter=3, 
    terminateAtMaxObjects=True,
    maxNumSpecies=200, # first stage is until core reaches 100 species
    filterReactions=True, # should speed up model generation
)

model(
    toleranceMoveToCore=0.25,
    toleranceInterruptSimulation=1e8,
    toleranceKeepInEdge=0.01, # Pruning enabled for stage 2
    maximumEdgeSpecies=200000,
    minCoreSizeForPrune=100,
    minSpeciesExistIterationsForPrune=2,
    filterReactions=True,
    )

options(
    units='si',
    # saveRestartPeriod=(1,'hour'),
    generateOutputHTML=False,
    generatePlots=True,
    saveSimulationProfiles=True,
    saveEdgeSpecies=False,
)

Question

what happened, I get so many nan?
Usually, I won't get any nan with the similar input file, I want to know is this an error? something wrong with my input? or it's normal and I can ignore it,

Installation Information

Describe your installation method and system information if applicable.

• OS (include version if known): ubuntu
• Installation method: source install using conda env
• RMG version information:
  • RMG-Py: july 31
  • RMG-database: july 26

[mjohnson541]

So the nans appear in the first time step for the cases in this job that give conversion: nan, but in other cases the simulations run fine. I believe this is a result of a reaction with a bad rate and/or bad thermo that when added to the model under certain conditions breaks the solver residual calculation.

I've made a PR that diagnoses nan presence as a DASxError, which allows RMG to handle this more properly, but the fundamental issue is the reaction/species that's causing this. It will at least have entered before you start seeing the nans, but may have entered a few iterations earlier since the nan production seems to be dependent on the temperature and pressure.

[shenghuiqin]

on another input file of heptane
I

• changed
  terminationRateRatio = 0.01, terminationConversion={ 'C7H16': 0.99, },
  to
  terminationRateRatio = 0.01,
  terminationConversion={ 'O2': 0.5, },
• added pressure dependence,

and i got almost all (210/212) conversion: nun

input.py

database(thermoLibraries =['BurkeH2O2','FFCM1(-)','thermo_DFT_CCSDTF12_BAC','CBS_QB3_1dHR','DFT_QCI_thermo','primaryThermoLibrary'], # 'FFCM1(-)','primaryThermoLibrary', 'BurkeH2O2','DFT_QCI_thermo','CBS_QB3_1dHR'
reactionLibraries = [('2005_Senosiain_OH_C2H2',False),('Glarborg/C3', False)], #
seedMechanisms = ['BurkeH2O2inN2','FFCM1(-)',], #
kineticsDepositories = ['training'],
kineticsFamilies = ['default'],
kineticsEstimator = 'rate rules',
)

#Constraints on generated species
generatedSpeciesConstraints(
maximumRadicalElectrons = 2,
allowed=['input species','seed mechanisms','reaction libraries'],
maximumCarbonAtoms=9,
maximumOxygenAtoms=4,
#allowSingletO2 = True,
)

#List of species
species(
label='C7H16',
reactive=True,
structure=SMILES("CCCCCCC"),
)
species(
label='O2',
reactive=True,
structure=adjacencyList(
"""
1 O u1 p2 {2,S}
2 O u1 p2 {1,S}
"""),
)
species(
label='N2',
reactive=False,
structure=SMILES("N#N")
)

#Reaction systems
simpleReactor(
temperature=[(600,'K'),(1500,'K')],
pressure=[(10.0,'bar'),(40.0,'bar')],
nSims=8,
initialMoleFractions={
#"C7H10": 1,
"O2": 0.10885, # phi=1 means 9.5 O2 per C7H10
"N2": 0.881684808, # 8.1 times as much N2 as O2
"C7H16":0.00946, #

},
terminationTime = (10.0, 's'),
terminationRateRatio = 0.001,
terminationConversion={
            'O2': 0.5,
    },

)

simulator(
atol=1e-16,
rtol=1e-8,
)

model(
toleranceKeepInEdge=0,
toleranceMoveToCore=0.4,
toleranceInterruptSimulation=1,
maxNumObjsPerIter=3,
terminateAtMaxObjects=True,
maxNumSpecies=200,
filterReactions=True,
)

model(
toleranceMoveToCore=0.3,
toleranceInterruptSimulation=1e8,
toleranceKeepInEdge=0.01,
maximumEdgeSpecies=100000,
minCoreSizeForPrune=100,
minSpeciesExistIterationsForPrune=2,
filterReactions=True,
)

options(
units='si',
# saveRestartPeriod=(1,'hour'),
generateOutputHTML=False,
generatePlots=True,
saveSimulationProfiles=True,
saveEdgeSpecies=False,
)
pressureDependence(
method='modified strong collision',
maximumGrainSize=(0.5,'kcal/mol'),
minimumNumberOfGrains=250,
temperatures=(300,2000,'K',8),
pressures=(0.01,100,'bar',5),
interpolation=('Chebyshev', 6, 4),

)

[shenghuiqin]

I substitute 'FFCM1(-)' with 'Klippenstein_Glarborg2016' on my input file, and no "conversion: nan" shows

[mjohnson541]

In that case it may be that the FFCM1(-) thermochemistry isn't compatible with the reaction libraries you're using.

[shenghuiqin]

yeah, I need to be more careful when selecting those libraries.
thank you for your help.
should I close this issue?