Skip to content

0D IDT CH₄

Created Sep 20, 2026 Last modified Sep 20, 2026

This model predicts the zero-dimensional (0-D) homogeneous ignition delay time (IDT) of a methane (CH₄) / air mixture. The fuel is fixed as methane; there is no mixture or SMILES input. For a given initial thermodynamic state and dilution, the model returns how long the well-mixed charge takes to ignite.

Ignition delay here is the induction time of a single, spatially uniform control volume: no spray, no walls as a heat sink, and no 3-D engine geometry. The charge is closed and adiabatic at constant volume. Its initial state is P0P_0, T0T_0, ϕ0\phi_0, and EGR.

0D homogeneous reactor A single well-mixed control volume with no spatial gradients, closed and insulated, carrying its initial conditions: P0 in bar, T0 in kelvin, equivalence ratio phi0, and EGR in percent. INITIAL STATE P0 (bar) T0 (K) 0 EGR (%) adiabatic wall 0D homogeneous reactor single zone · no spatial gradients · V = const
0-D homogeneous reactor — single zone, no spatial gradients, V=constV = \mathrm{const}.

Evaluating a detailed kinetic mechanism at every (P,T,ϕ,EGR)(P, T, \phi, \mathrm{EGR}) query is too slow for interactive sweeps. Scimlate uses a trained ignition-delay ML model instead: the four inputs below map to a fixed feature order, the model returns ignition delay in seconds.

Inference data flow: physical inputs are converted to model features, the surrogate predicts delay, and Scimlate reports milliseconds.InputsP, T, φ, EGRbar, K, —, %FeaturesP, T, φ, EGR/100EGR → fractionSurrogatetrained modelτ in secondsIDT1000 × τmilliseconds
Physical inputs → feature vector → surrogate prediction → ignition delay in ms.
  • Input: pressure PP (bar), temperature TT (K), equivalence ratio ϕ\phi, and exhaust-gas recirculation (EGR) in percent.

    Single point (importable Scimlate input bundle):

    {
    "format": "scimlate-inputs",
    "version": 1,
    "model_type": "zero-d-ignition-delay-time-ch4",
    "parameters": {
    "pressure": "20",
    "temperature": "1000",
    "phi": "1.0",
    "EGR": "0"
    }
    }

    Multi-sweep (start:end:step on each axis inside parameters; inclusive ranges expand as a Cartesian product — 3×5×4×2=1203 \times 5 \times 4 \times 2 = 120 evaluations):

    {
    "format": "scimlate-inputs",
    "version": 1,
    "model_type": "zero-d-ignition-delay-time-ch4",
    "parameters": {
    "pressure": "20:30:5",
    "temperature": "900:1100:50",
    "phi": "1.0:1.6:0.2",
    "EGR": "0:15:15"
    }
    }
  • Output: ignition delay in milliseconds (ignition_delay), with in_range when a row falls outside the training envelope.

QuantitySymbolCLI keyUnit
PressureP0P_0pressurebar
TemperatureT0T_0temperatureK
Equivalence ratioϕ0\phi_0phi—
Exhaust-gas recirculationEGR0\mathrm{EGR}_0EGR%
Ignition delayτ\tauignition_delayms

The training envelope is a soft box: values outside it warn and set in_range to false, but inference still runs. Predictions outside this box are less reliable.

InputSymbolUnitWorking range
PressurePPbar[10, 50][10,\, 50]
TemperatureTTK[700, 1500][700,\, 1500]
Equivalence ratioϕ\phi—[0.5, 2.0][0.5,\, 2.0]
Exhaust-gas recirculationEGR%[0, 30][0,\, 30]

Structured Arrhenius sweeps in the training set were denser at 900900–1175 K1175\,\mathrm{K}; the claimed envelope is the full 700700–1500 K1500\,\mathrm{K} interval.

Stoichiometric methane/air at P0=20 barP_0 = 20\,\mathrm{bar}, T0=1000 KT_0 = 1000\,\mathrm{K}, and no EGR:

QuantityValue
P0P_02020 bar
T0T_010001000 K
ϕ0\phi_01.01.0
EGR0\mathrm{EGR}_000 %
Predicted τ\tau7.45407.4540 ms

Raising temperature to 1100 K1100\,\mathrm{K} at the same pressure, ϕ\phi, and EGR shortens the delay to 1.2300 ms1.2300\,\mathrm{ms}, as expected for an Arrhenius induction time.

A two-pressure, two-temperature, two-EGR sweep at ϕ0=1\phi_0 = 1 produces 2×2×1×2=82 \times 2 \times 1 \times 2 = 8 predictions (Cartesian product). Each row is one (P0,T0,ϕ0,EGR0)(P_0, T_0, \phi_0, \mathrm{EGR}_0) evaluation of the same forward map.

Scimlate bills inference from the model’s cartesian_weighted usage contract, not from analytics counters or output format. Sweep syntax is scalar_csv_range: a scalar is one value, 10,20,30 is a list, and 10:40:5 is an inclusive start:end:step range (10:15 defaults to step 11).

NN is the Cartesian product of the four sweep lengths. A dimension is active only when its expanded length is greater than one:

N=nP nT nϕ nEGRN = n_P\, n_T\, n_\phi\, n_{\mathrm{EGR}} k=∣{p:np>1}∣,W=∑p activewp(W=1 if k=0)k = \bigl\lvert \{ p : n_p > 1 \} \bigr\rvert, \quad W = \sum_{p\,\mathrm{active}} w_p \quad (W = 1 \text{ if } k = 0) m(k)=2max⁡(k−1, 0),unit=0.01 W m(k),credits=unit×Nm(k) = 2^{\max(k-1,\,0)}, \quad \mathrm{unit} = 0.01\, W\, m(k), \quad \mathrm{credits} = \mathrm{unit} \times N
Sweep axisWeight ww
Pressure44
Temperature33
Equivalence ratio22
EGR11

Missing weights default to 11. With multi_sweep_factor: 2, a one-axis sweep has m=1m = 1, two axes m=2m = 2, three axes m=4m = 4, and four axes m=8m = 8. After a successful run Scimlate uses the run’s prediction count to settle NN. The rows below match common requests: a lone state; 700:1500:50; 10,30,50 × 0,15,30; the 8-point worked sweep; and 20:30:5 × 900:1100:100 × 1.0,1.5.

RunActivekkWWmmNNUnitCredits
Single point—001111110.010.010.010.01
TT: 700:1500:50TT11331117170.030.030.510.51
PP × EGR arraysPP, EGR225522990.100.100.900.90
Worked 8-point sweepPP, TT, EGR338844880.320.322.562.56
PP×TT ranges × ϕ\phi pairPP, TT, ϕ\phi33994418180.360.366.486.48

No per-prediction interval is emitted. The table is a global residual summary on the 43030-point training set used to fit the model.

MetricValue
Median absolute relative error0.99%
MAPE1.40%
90th-percentile absolute relative error3.12%
MAE(log⁡10τ\log_{10} \tau)0.006
Maximum absolute relative error≈16%

These residuals describe hold-in training-set agreement, not a withheld test split or a propagated kinetic-mechanism uncertainty.

  • Model implementation: carnotai/src/mechanical/fuel_chemistry/scripts/0d_idt_ch4_fwd (CarnotAI repository).
  • ONNX artifact: carnotai/src/mechanical/fuel_chemistry/models/0d_idt_ch4_fwd/0d_idt_ch4_fwd.onnx.
  • Credit contract: CarnotAI SCHEMA.md (usage) and wiki/model_usage.md (cartesian_weighted).
  • Display symbols and Arrhenius chart: CarnotAI wiki/entities/0d_idt_ch4.md.
  • Training and residual packaging: EulerTrain methane SIDT benchmark (43030 clean points).