Turkey Cooking Time Calculator & Brisket Stall Simulator

The model is derived here from conservation of energy in a symmetric plane slab of half-thickness $L$. Sixteen cell-centered control volumes solve $\rho c_p\,\partial T/\partial t=k\,\partial^2T/\partial x^2$ with an insulated symmetry plane. Implicit midpoint integration uses cooking steps no longer than 450 s or 6% of the diffusion time, reduced further when convection exceeds 10 W/(m² K). The first 4 s of cooking and rest use 2 s backward-Euler startup steps. Steps near removal are reduced to 10 s, resting steps are at most 30 s while the core is warming, and may grow to 90 s after the core has started cooling, and the target crossing is located by 16 interval bisections before the boundary switches to room conditions. A quadratic symmetry-plane fit estimates the core temperature. Terminal steps end the selected rest or cooking horizon exactly. Instantaneous surface readings are reconstructed for the retained display frames; every integration step checks the evaporation-pressure limit independently. The cached display history preserves removal, the full-step sampled peak and worst energy residual; animation interpolates the cache. Integration, slow-heating duration and peak detection use the full time-step history.

At each exposed face, $q_{in}=h(T_a-T_s)+\epsilon\sigma(T_{a,K}^4-T_{s,K}^4)-h_{fg}\dot m''$. A Lewis-analogy scenario model uses $\dot m''=f_w(h/c_{p,a})\,0.621945\ln[(P-p_v)/(P-p_{sat}(T_s))]$, clipped to nonnegative evaporation and the remaining supply per step. $c_{p,a}=1006$ J/(kg K). The effective wetted fraction includes unmodeled surface transport resistance; it is not a calibrated meat correlation. Saturation pressure and the liquid-water thermodynamic wet-bulb balance are reused from the catalog's wet-bulb calculator (ASHRAE 2017 / PsychroLib 2.5). Wet bulb excludes radiation and is not the meat stall temperature. A defensive 98% pressure cap keeps trial boundary iterations finite. If that cap is active in a converged step or end state while positive wetness and liquid supply remain, integration stops and cooking-time, peak and carryover predictions are withheld. This is a conservative validity guard; boiling and drying are not extrapolated.

Initial evaporatable supply per face area is $f_m\rho L$. Effective wetness is constant until the last 15% of supply, then decreases linearly to zero. The solver ends a step at a requested cooking-wrap time and latches an ideal tight foil state only while still cooking. That state is preserved throughout rest; a wrap scheduled after removal is not applied and raises a warning. At the core target, the boundary switches to the room; the same field continues through the selected rest. An independent no-evaporation run uses its own removal and rest.

Choi–Okos (1986) component density and conductivity equations and protein/fat heat capacities, as presented by ASHRAE 2018 ch. 19, are evaluated at 40 °C and frozen for the run. Density uses reciprocal mass-fraction mixing, heat capacity uses mass fractions and conductivity uses volume fractions. Remaining solids are protein-like. Water heat capacity is approximated as 4.18 kJ/(kg K), supported by the slope of liquid enthalpy in NISTIR 5078 Table 1; this avoids the 4.1289 coefficient in the current online ASHRAE water-heat-capacity transcription. Latent heat is approximated by $(2500.9-2.36T_s)$ kJ/kg, within about 0.4% of NIST's 20–100 °C values.

An independent adaptive SciPy BDF method-of-lines benchmark additionally checks rapid-cook carryover and removal time. Its 48-, 96- and 192-cell results converge to 10.73224 K carryover for a 3 cm dry section cooked at 180 °C to 74 °C and then rested at 22 °C. The benchmark code and convergence output are attached; these are numerical reference predictions, not meat measurements.

Checks compare two dry heating times with a separately evaluated 12-mode analytical plane-wall series (CSJMU, attached p. 6), and vapor pressures with PsychroLib's published ASHRAE values and NIST data. These validate the numerical heat and psychrometric components, not absolute cooking times for a particular meat.

The section labels near-balance when evaporation consumes at least 90% of incoming surface heat. This is a visualization cue; the separately reported slow-heating duration uses its stated core-temperature criterion.

Initially uniform, fully thawed meat; a homogeneous symmetric section heated equally on both broad faces. For an effectively one-sided region, enter twice the actual heat path. No overall weight-to-thickness rule is assumed. The two presets change cooker/removal temperatures only; thickness and composition stay as entered.

Air and enclosing walls have equal, constant temperature. There is no direct flame, pan contact, fat cap, cavity, bone, shrinkage, dripping, fat rendering or explicit internal moisture diffusion. Initial material properties, geometry and sensible heat capacity remain fixed even as the abstract moisture reservoir depletes. Available supply is not the total water fraction or a prediction of total cooking loss. Wetness, its final 15% taper, the 20% supply default and $h=10$ W/(m² K) are illustrative scenario choices, not fitted experimental constants.

Resting is uncut on a rack, with $h=8$ W/(m² K), room relative humidity 50%, and the same emissivity; a cooking wrap stays on. Condensation and the sensible energy carried with lost liquid are omitted. Do not interpret the section picture as turkey anatomy.

Use for comparative planning over effective thicknesses 3–14 cm, cooker settings 100–180 °C, fully thawed initial temperatures 0–20 °C, and pressure 80–110 kPa. High starting temperatures are sensitivity studies; keep raw meat refrigerated. The target must lie below the cooker and at least 1 K below the local boiling point. Values outside those bounds or unusual combinations need a different model. An unwrapped, still-wet surface activating the 98% vapor-pressure guard is excluded; all cooking-time and carryover predictions are withheld with a clear input message. A target not reached within the window produces no removal/carryover prediction. Heat-transfer coefficients, thickness and surface-moisture behavior dominate uncertainty; there is no defensible universal percentage error bar or promise of serving time.

The slow-heating metric means core rise below 2 K/h between 55 and 85 °C while still cooking. It can be zero. Peak and carryover cover only the selected rest, and can miss a later peak. The no-evaporation line ends after its own rest; it is not a prediction for a real foil package with different contact/radiation behavior. The diagram color scale is fixed at 0–110 °C and clips hotter surface colors.

Food safety — US guidance checked 26 October 2026: Doneness is not proof of safety. USDA consumer minimums must be measured with a food thermometer before removal: poultry 165 °F (73.9 °C); whole beef roasts 145 °F (62.8 °C) plus at least 3 minutes rest. Ground beef requires 160 °F (71.1 °C) and is excluded. Turkey roasting guidance specifies an oven no lower than 325 °F. Check the thickest breast and innermost thigh and wing; do not rely on this single simulated cold point or on predicted carryover. Follow safe handling and prompt refrigeration guidance.

FSIS Revised Appendix A, December 2021, is commercial process guidance. Its p. 25 identifies humidity as a critical operating parameter for its time-temperature tables, with specific exceptions and scientific gaps. This simulator does not implement those lethality schedules or establish compliance. Check current official guidance and the rules in force for a real commercial process. Not for frozen or stuffed birds, injection, ground meat, pathogen inactivation, hot holding, cooling or storage.