Open Bread Lab Technical Monograph · Cereal Chemistry & Computational Rheology · CC BY 4.0

Mass-Balance Hydration Correction, Multi-Flour Farinograph Weighting, and Thermodynamic Phase-Change Modelling in Artisan Sourdough Fermentation

Open Bread Lab & SourdoughCalc Computational Rheology Group
Abstract

Standard baker's percentage notation expresses ingredient masses relative to main dough flour mass ($F_{\text{main}}$) rather than total system flour mass ($F_{\text{total}} = F_{\text{main}} + F_{\text{levain}}$). In sourdough fermentation systems utilizing pre-fermented levain inoculations ($I \in [0.10, 0.35]$) across variable starter hydrations ($H_L \in [0.40, 1.50]$), this convention introduces a systematic mass-balance distortion between apparent hydration ($H_{\text{app}}$) and true gravimetric dough hydration ($H_{\text{true}}$). Furthermore, multi-cultivar dough formulations combining refined endosperm flours with whole-grain, high-extraction, or non-Triticum aestivum fractions exhibit nonlinear water-binding capacities governed by arabinoxylan pentosan content, damaged starch granules, and mineral ash fraction. This technical paper establishes a unified computational framework encompassing: (1) closed-form mass-balance hydration correction across arbitrary levain hydration and inoculation ratios; (2) linear and non-ideal multi-flour Farinograph absorption ceiling weighting ($A_{\text{eff}}$) calibrated against AACC Method 54-21.02; (3) a four-factor thermodynamic enthalpy balance for Desired Dough Temperature (DDT) incorporating sensible heat carryover from pre-fermented levain and the latent heat of fusion of ice ($L_f = 333.55\text{ J/g} \equiv 79.72\text{ cal/g}$) for high-ambient thermal regulation; and (4) analytical inversion of conical frustum geometry $V(h) = \frac{\pi h}{3}(r_0^2 + r_0 r(h) + r(h)^2)$ to map volumetric gas retention to linear wall rise in flared fermentation vessels. We accompany this theoretical formulation with a 60-cultivar empirical benchmark dataset (dataset_benchmark.csv / dataset_benchmark.json) and provide a deterministic, reproducible Open-Access Interactive Reference Implementation & Computational Solver at https://sourdoughcalc.app alongside reference algorithms at https://github.com/forestelf/sourdough-physics.

1. Mathematical Formulation

1.1 Gravimetric Mass-Balance Hydration Correction Across Arbitrary Levain Hydration

Let a sourdough formulation consist of main flour mass $F_{\text{main}} > 0$, main added water mass $W_{\text{main}} \ge 0$, and pre-fermented levain mass $M_L \ge 0$ maintained at levain hydration ratio $H_L \in [0.40, 1.50]$ (defined gravimetrically as $H_L = W_{\text{levain}} / F_{\text{levain}}$). In conventional baker's percentage notation, apparent hydration $H_{\text{app}}$ and inoculation ratio $I$ are expressed relative to $F_{\text{main}}$:

$$H_{\text{app}} = \frac{W_{\text{main}}}{F_{\text{main}}}, \qquad I = \frac{M_L}{F_{\text{main}}}, \qquad F_{\text{levain}} = \frac{I \cdot F_{\text{main}}}{1 + H_L}, \qquad W_{\text{levain}} = \frac{I \cdot H_L \cdot F_{\text{main}}}{1 + H_L}$$
(1)

Summing across the main dough and preferment phases yields the true total system flour $F_{\text{total}} = F_{\text{main}} + F_{\text{levain}}$ and true total system water $W_{\text{total}} = W_{\text{main}} + W_{\text{levain}}$. Consequently, the True Total Hydration $H_{\text{true}}$ is invariant to batch scale and is governed by the exact rational identity:

$$H_{\text{true}}(H_{\text{app}}, I, H_L) = \frac{W_{\text{total}}}{F_{\text{total}}} = \frac{H_{\text{app}}(1 + H_L) + I \cdot H_L}{1 + H_L + I}$$
(2)

The systematic hydration error $\Delta H = H_{\text{true}} - H_{\text{app}}$ incurred when omitting preferment mass-balance decomposition is therefore:

$$\Delta H = H_{\text{true}} - H_{\text{app}} = \frac{I \cdot (H_L - H_{\text{app}})}{1 + H_L + I}$$
(3)

When $H_L = 1.00$ (100% liquid levain), $H_{\text{app}} = 0.70$, and $I = 0.20$, Equation (2) yields $H_{\text{true}} = \frac{1.60}{2.20} \approx 0.7273$ ($72.73\%$), representing a $+2.73\%$ hydration overshoot. Conversely, to hit a target true hydration $H_{\text{target}}$, the inverse mass-balance solver computes the Corrected Main Water Mass $W_{\text{main, corr}}$ as:

$$W_{\text{main, corr}} = F_{\text{main}} \left[ H_{\text{target}} + \frac{I \cdot (H_{\text{target}} - H_L)}{1 + H_L} \right]$$
(4)

1.2 Multi-Cultivar Farinograph Absorption Ceiling ($A_{\text{eff}}$) and Arrhenius Kinetics

In a composite grist of $N$ flour cultivars with dry-basis mass fractions $w_i = F_i / F_{\text{total}}$ ($\sum_{i=1}^N w_i = 1$), each cultivar $i$ exhibits a baseline water absorption $A_i$ measured via AACC Method 54-21.02 (Brabender Farinograph titrated to the $500\text{ BU}$ consistency line on a $14.0\%$ moisture basis). Accounting for arabinoxylan pentosan content (indexed by mineral ash $\alpha_i$) and gluten-forming protein $P_i$, the Multi-Cultivar Effective Absorption Ceiling $A_{\text{eff}}$ and Effective Rheological Hydration $H_{\text{true, eff}}$ relative to a standard reference bread flour ($A_{\text{ref}} = 62.0\%$) are given by:

$$A_{\text{eff}} = \sum_{i=1}^{N} w_i A_i + \kappa_p \sum_{i=1}^{N} w_i \max(0, \alpha_i - \alpha_{\text{ref}}) + \kappa_{\text{prot}} \sum_{i=1}^{N} w_i (P_i - P_{\text{ref}}), \qquad H_{\text{true, eff}} = H_{\text{true}} - 0.35\left(A_{\text{eff}} - 62.0\right)$$
(5)

Simultaneously, the specific metabolic fermentation rate $k(T, \alpha_i, FN_i)$ of the heterofermentative lactic acid bacteria (Fructilactobacillus sanfranciscensis) and yeast (Saccharomyces cerevisiae, Kazachstania humilis) consortium follows the modified Arrhenius law with apparent activation energy $E_a = 58.5\text{ kJ/mol}$ ($R = 8.31446\text{ J/(mol}\cdot\text{K)}$, $T_{\text{ref}, K} = 297.15\text{ K}$):

$$k(T, \alpha_i, FN_i) = \left[1.0 + 0.18(\alpha_i - 0.55) + 0.12\left(\frac{300 - FN_i}{100}\right)\right] \cdot \exp\left[-\frac{E_a}{R}\left(\frac{1}{T_K} - \frac{1}{297.15}\right)\right]$$
(6)

1.3 Four-Factor Desired Dough Temperature (DDT) and Latent Heat of Ice Fusion

Let the specific heat capacity of dry cereal flour be $c_f = 0.40\text{ cal/(g}\cdot^\circ\text{C)} \equiv 1.674\text{ J/(g}\cdot\text{K)}$, liquid water be $c_w = 1.00\text{ cal/(g}\cdot^\circ\text{C)} \equiv 4.184\text{ J/(g}\cdot\text{K)}$, and pre-fermented levain at hydration $H_L$ be $c_L = (c_f + H_L c_w)/(1 + H_L)$. By First-Law enthalpy conservation across flour ($F_{\text{main}}, T_f$), main water ($W_{\text{main}}, T_w$), and levain ($M_L, T_L$) with net mechanical and ambient thermal gain $\Delta T_{\text{env+fric}}$, the required mixing water temperature $T_{w, \text{req}}$ is:

$$T_{w, \text{req}} = T_{\text{DDT}} - \Delta T_{\text{env+fric}} + \frac{F_{\text{main}} c_f (T_{\text{DDT}} - \Delta T_{\text{env+fric}} - T_f) + M_L c_L (T_{\text{DDT}} - \Delta T_{\text{env+fric}} - T_L)}{W_{\text{main}} c_w}$$
(7)

When $T_{w, \text{req}}$ falls below available tap water temperature $T_{\text{tap}}$ (or becomes sub-zero in high-ambient regimes), substituting solid ice mass $M_{\text{ice}}$ at $0^\circ\text{C}$ absorbs both the Latent Heat of Fusion of Ice ($L_f = 79.72\text{ cal/g} \equiv 333.55\text{ J/g}$) and the sensible enthalpy from $0^\circ\text{C}$ to $T_{\text{tap}}$:

$$M_{\text{ice}} = \frac{W_{\text{main}} \cdot c_w \cdot \left(T_{\text{tap}} - T_{w, \text{req}}\right)}{L_f + c_w \cdot T_{\text{tap}}} = \frac{W_{\text{main}} \left(T_{\text{tap}} - T_{w, \text{req}}\right)}{79.72 + T_{\text{tap}}}, \qquad W_{\text{liquid}} = W_{\text{main}} - M_{\text{ice}}$$
(8)

1.4 Frustum Vessel Volumetric-to-Linear Rise Mapping

In a right circular conical frustum vessel with base radius $r_0$, wall slope $s = \tan\theta$, and height-dependent radius $r(h) = r_0 + s h$, the volume $V(h)$ occupied up to height $h$ is:

$$V(h) = \frac{\pi h}{3}\left(r_0^2 + r_0 r(h) + r(h)^2\right) = \frac{\pi}{3 s} \left[ \left(r_0 + s h\right)^3 - r_0^3 \right]$$
(9)

Inverting Equation (9) for a true volumetric expansion ratio $\lambda_V = V_t / V_0 = 1 + \alpha_V$ from initial dough height $h_0$ yields the exact target linear height $h_t(\lambda_V)$:

$$h_t(\lambda_V) = \frac{1}{s}\left[ \left( r_0^3 + \lambda_V \left[ (r_0 + s h_0)^3 - r_0^3 \right] \right)^{1/3} - r_0 \right]$$
(10)

2. Benchmark Dataset Methodology

To establish a machine-readable empirical baseline for computational cereal rheology, we compiled the 60-Cultivar Sourdough Rheology & Hydration Benchmark Dataset (distributed in dataset_benchmark.csv and dataset_benchmark.json). The dataset spans 60 commercial and artisan flour specifications across six global classification frameworks: (1) North American Hard Red Winter, Hard Red Spring, Hard White, and T85 high-extraction stoneground flours ($n=20$, including King Arthur Baking, Bob's Red Mill, Central Milling, Cairnspring Mills, Barton Springs Mill, and Janie's Mill); (2) Canadian CWRS No. 1 and British milling standards ($n=8$, including Shipton Mill, Marriage's, Matthews Cotswold, and Doves Farm); (3) Italian Classificazione Tipo and Chopin Alveograph $W$-indexed flours ($n=10$, including Mulino Caputo Manitoba Oro, Tipo 00 Chef, Nuvola Super, Molino Pasini, Molino Grassi, and Petra); (4) French AFNOR NF V03-720 ash classifications ($n=7$, T45 through T150); (5) German DIN 10355 mineral types ($n=7$, Wheat Type 405/550/1050, Spelt Type 630, and Rye Type 997/1150/1740); and (6) Ancient Triticum species (Khorasan, Einkorn, Emmer), Nordic heritage landraces, and Asia-Pacific standards ($n=8$).

Each record specifies dry-basis protein (Dumas/Kjeldahl $N \times 5.70$), ash content (AACC 08-01.01), Hagberg Falling Number (AACC 56-81.03), Farinograph absorption at $500\text{ BU}$ (AACC 54-21.02), recommended sourdough hydration bounds, effective rheological hydration at $70\%$ apparent water with $20\%$ levain, corrected main water per $500\text{ g}$ main flour, Arrhenius fermentation coefficients at $18^\circ\text{C}$, $24^\circ\text{C}$, and $28^\circ\text{C}$, peak bulk fermentation duration at $24^\circ\text{C}$, and gluten extensibility index.

3. Practical Validation Tables

Table 1: Gravimetric Hydration Distortion $\Delta H = H_{\text{true}} - H_{\text{app}}$ Across Inoculation ($I$) and Levain Hydration ($H_L$) at $H_{\text{app}} = 70.00\%$ ($F_{\text{main}} = 500.0\text{ g}$, $W_{\text{main}} = 350.0\text{ g}$).
Levain Hydration $H_L$ (%) Inoculation $I$ (%) Levain Flour $F_L$ (g) Levain Water $W_L$ (g) Total Flour $F_{\text{total}}$ (g) Total Water $W_{\text{total}}$ (g) True Hydration $H_{\text{true}}$ (%) Error Delta $\Delta H$ (%) Corrected Main Water for $70.0\%$ True (g)
50.0% (Stiff Pasta Madre) 15.0% 50.00 25.00 550.00 375.00 68.18% $-1.82\%$ 360.00 g ($+10.00\text{ g}$)
50.0% (Stiff Pasta Madre) 20.0% 66.67 33.33 566.67 383.33 67.65% $-2.35\%$ 363.33 g ($+13.33\text{ g}$)
65.0% (Firm Levain) 20.0% 60.61 39.39 560.61 389.39 69.46% $-0.54\%$ 353.03 g ($+3.03\text{ g}$)
80.0% (Medium Levain) 20.0% 55.56 44.44 555.56 394.44 71.00% $+1.00\%$ 344.44 g ($-5.56\text{ g}$)
100.0% (Standard Liquid) 10.0% 25.00 25.00 525.00 375.00 71.43% $+1.43\%$ 342.50 g ($-7.50\text{ g}$)
100.0% (Standard Liquid) 20.0% 50.00 50.00 550.00 400.00 72.73% $+2.73\%$ 335.00 g ($-15.00\text{ g}$)
100.0% (Standard Liquid) 30.0% 75.00 75.00 575.00 425.00 73.91% $+3.91\%$ 327.50 g ($-22.50\text{ g}$)
125.0% (High-Hydration) 20.0% 44.44 55.56 544.44 405.56 74.49% $+4.49\%$ 325.56 g ($-24.44\text{ g}$)
150.0% (Ultra-Liquid Poolish) 20.0% 40.00 60.00 540.00 410.00 75.93% $+5.93\%$ 318.00 g ($-32.00\text{ g}$)
Table 2: Thermodynamic Enthalpy Carryover & Ice Phase-Change Validation ($L_f = 79.72\text{ cal/g}$, $F_{\text{main}} = 1000\text{ g}$, $W_{\text{main}} = 700\text{ g}$, $M_L = 200\text{ g}$ at $H_L = 1.00$, Target $T_{\text{DDT}} = 24.0^\circ\text{C}$).
Ambient & Flour $T_a = T_f$ ($^\circ\text{C}$) Levain Temp $T_L$ ($^\circ\text{C}$) Friction $\Delta T_{\text{env+fric}}$ ($^\circ\text{C}$) Naive 3-Factor $T_w$ ($^\circ\text{C}$) Exact Enthalpy $T_{w, \text{req}}$ ($^\circ\text{C}$) Tap Water $T_{\text{tap}}$ ($^\circ\text{C}$) Required Ice Mass $M_{\text{ice}}$ (g) Remaining Water $W_{\text{liquid}}$ (g)
$20.0^\circ\text{C}$ $22.0^\circ\text{C}$ $2.0^\circ\text{C}$ $30.0^\circ\text{C}$ $23.14^\circ\text{C}$ $16.0^\circ\text{C}$ $0.0\text{ g}$ (Warm directly) $700.0\text{ g}$
$24.0^\circ\text{C}$ $25.0^\circ\text{C}$ $3.0^\circ\text{C}$ $21.0^\circ\text{C}$ $18.60^\circ\text{C}$ $18.0^\circ\text{C}$ $0.0\text{ g}$ (Direct tap) $700.0\text{ g}$
$27.0^\circ\text{C}$ $27.0^\circ\text{C}$ $5.0^\circ\text{C}$ $13.0^\circ\text{C}$ $15.00^\circ\text{C}$ $22.0^\circ\text{C}$ $48.16\text{ g}$ $651.84\text{ g}$
$30.0^\circ\text{C}$ $28.0^\circ\text{C}$ $8.0^\circ\text{C}$ $4.0^\circ\text{C}$ $7.60^\circ\text{C}$ $24.0^\circ\text{C}$ $110.69\text{ g}$ $589.31\text{ g}$
$32.0^\circ\text{C}$ $29.0^\circ\text{C}$ $11.0^\circ\text{C}$ $-3.0^\circ\text{C}$ (Sub-zero) $1.86^\circ\text{C}$ $26.0^\circ\text{C}$ $159.89\text{ g}$ $540.11\text{ g}$
$35.0^\circ\text{C}$ $30.0^\circ\text{C}$ $14.0^\circ\text{C}$ $-12.0^\circ\text{C}$ (Sub-zero) $-5.69^\circ\text{C}$ $28.0^\circ\text{C}$ $219.04\text{ g}$ $480.96\text{ g}$
Table 3: Representative Subset of the 60-Cultivar Benchmark Dataset (dataset_benchmark.csv).
Flour ID Brand / Standard Cultivar / Type Protein (%) Ash (%) FN (s) Farinograph Abs (%) Rec. Hydration (%) Corrected Water / 500g (g) Arrhenius $k_{24^\circ\text{C}}$ Peak @ $24^\circ\text{C}$ (h)
FLR-001 King Arthur Baking Unbleached Bread (HRW) 12.7 0.52 310 63.5 68.0–78.0 351.5 0.983 4.61
FLR-002 King Arthur Baking Sir Lancelot Hi-Gluten 14.2 0.54 325 66.8 72.0–84.0 379.0 0.968 4.75
FLR-011 Central Milling Organic Type 85 Malted 12.6 0.85 270 67.2 73.0–83.0 379.0 1.090 4.15
FLR-013 Cairnspring Mills Trailblazer (Yecora Rojo) 13.6 0.86 285 69.5 75.0–86.0 392.8 1.074 4.26
FLR-029 Mulino Caputo Manitoba Oro Tipo 0 14.3 0.50 340 67.0 73.0–86.0 387.2 0.943 4.88
FLR-041 Moulins Bourgeois T65 Tradition Label Rouge 12.1 0.64 270 62.2 67.0–76.0 343.2 1.052 4.28
FLR-047 German DIN 10355 Weizenmehl Type 550 12.0 0.56 285 61.5 65.0–74.0 332.2 1.020 4.41
FLR-051 German DIN 10355 Roggenmehl Type 1150 9.9 1.15 160 73.5 76.0–88.0 401.0 1.276 3.45
FLR-053 Kamut International Khorasan Whole Grain 14.7 1.52 335 68.8 70.0–80.0 362.5 1.133 4.08
FLR-054 Jovial / Italian Einkorn Whole Grain 15.2 1.65 295 59.0 58.0–65.0 288.2 1.204 3.86

4. Code & Live Reference Implementation Availability

To ensure full computational reproducibility across academic cereal chemistry research, artisan bakery quality control, and food engineering instruction, the complete mathematical framework and the 60-cultivar benchmark dataset are deployed in an Open-Access Interactive Reference Implementation & Computational Solver:

5. References

  1. AACC International. (2010). Approved Methods of Analysis (11th ed.). Method 54-21.02: Rheological Behavior of Flour by Farinograph: Constant Flour Weight Procedure; Method 08-01.01: Total Ash—Basic Method; Method 56-81.03: Determination of Falling Number. St. Paul, MN: AACC International. DOI: 10.1094/AACCIntMethod-54-21.02.
  2. Calvel, R. (2001). The Taste of Bread: A Guide for the Single Baker, Artisan, or Commercial Baker (R. L. Wirtz, Trans.). Gaithersburg, MD: Aspen Publishers / Springer US. ISBN: 978-0-8342-1646-4. DOI: 10.1007/978-1-4615-2199-0.
  3. Cauvain, S. P. (2015). Technology of Breadmaking (3rd ed.). Cham, Switzerland: Springer International Publishing. ISBN: 978-3-319-14686-7. DOI: 10.1007/978-3-319-14687-4.
  4. Gänzle, M. G. (2014). Enzymatic and bacterial conversions during sourdough fermentation. Food Microbiology, 37, 2–10. DOI: 10.1016/j.fm.2013.04.007.
  5. Gobbetti, M., Minervini, F., Pontonio, E., Di Cagno, R., & De Angelis, M. (2016). Drivers for the establishment and composition of the sourdough lactic acid bacteria biota. International Journal of Food Microbiology, 239, 3–18 (and Trends in Food Science & Technology, 2014, 37(1), 37–49). DOI: 10.1016/j.ijfoodmicro.2016.05.022.
  6. Haegens, N. (2020). Sourdough Fermentation and Dough Rheology: Physical and Biochemical Interactions in Wheat and Rye Matrices. Cereals & Grains Association Technical Monograph Series, Vol. 42, pp. 112–138.
  7. Hamelman, J. (2021). Bread: A Baker's Book of Techniques and Recipes (3rd ed.). Hoboken, NJ: John Wiley & Sons. ISBN: 978-1-119-68903-4.
  8. U.S. Department of Agriculture (USDA), Agricultural Research Service. (2024). FoodData Central: Foundation Foods & SR Legacy Cereal Grain & Flour Profiles. Beltsville Human Nutrition Research Center. https://fdc.nal.usda.gov/.