foratvise
Get Started
Refrigeration and freezing

Refrigeration and freezing

Freezing of a food slab (lean white fish fillets on trays) in an air-blast freezer

Validated modelOEE bound 0.862 [0.812–0.913]20,000 Monte Carlo samples · seed 20261115 · computed November 16, 2026

In short

Refrigeration and freezing can use at most 86 of every 100 planned hours. The rest goes to limits no plant can avoid.

  1. 86usable at most
  2. 14are a physical limit

The three limits that set it

Running time: up to 86 %

The rest are stops imposed by physics or by regulation: cleaning, changeovers, defrosting.

Speed: up to 100 %

How close to its maximum speed it can run without damaging the product or the equipment.

Good product: up to 100 %

Some product falls outside specification even with the process well under control.

See the technical detail

Technical detail

Uncertainty intervals, physical models per station and sources, for anyone who wants to check each figure.

1Scenario

Chilled fillets of lean white fish, 50 mm thick on metal trays, enter an air-blast freezer at about 5 °C and are frozen with air at about −35 °C until, after thermal stabilisation, every point of the product is at −18 °C or lower (the legal definition of a quick-frozen food in the European Union). Heat is pumped to the ambient by a two-stage ammonia vapour-compression plant. Chilling is the special case in which the final temperature stays above the initial freezing point: the latent term and the frozen-layer conduction term vanish and the same balances apply.

Provenance of inputsDeclared assumption15Verified quote43

2Results

Point value from the nominal parameters and 95% interval from 20,000 Monte Carlo draws of the uncertain parameters.
  • Heat removed from the product (inlet to equilibrated final temperature)302 [290–325] kJ/kg
  • TheoreticalThermodynamic minimum work (exergy of the heat removed)32.2 [21.6–43.9] kJ/kg
  • Carnot work with all heat lifted from the evaporating temperature84.1 [68.5–105.4] kJ/kg
  • Best available techniqueCompressor work of the best available two-stage ammonia cycle (product load only)153 [124–215] kJ/kg
  • Best documentedLowest documented electricity use of deep freezing (vegetables)288 kJ/kg
  • TypicalCurrent typical on-site energy intensity of freezing (United States)1,363 kJ/kg
  • Equivalent reversible COP (heat removed over minimum work)9.36 [7.08–14.10]
  • COP of the best available two-stage ammonia cycle1.97 [1.44–2.45]
  • DOE 'thermodynamic minimum' of freezing vegetables, reproduced (sign: heat released by the product is negative)-124 kJ/kg
  • TheoreticalMinimum freezing time (Plank, surface at air temperature)0.340 [0.327–0.556] h
  • Best available techniqueFreezing time with the best available surface coefficient (impingement)0.680 [0.556–1.033] h
  • Freezing time predicted for a conventional air blast (model check)2.04 [1.63–3.27] h
  • TheoreticalMaximum throughput per unit of product held in the freezer (theoretical)2.94 [1.80–3.06] 1/h
  • Best available techniqueThroughput per unit of product held in the freezer (best available technique)1.47 [0.97–1.80] 1/h
  • Throughput per unit of product held in a conventional air blast0.490 [0.306–0.615] 1/h
  • Minimum dehydration (weight) loss during freezing0.00422 [0.00260–0.00692]
  • Minimum frost deposited on the coil per tonne of product4.22 [2.60–6.92] kg/tonne
  • TheoreticalMaximum yield (mass balance)0.9958 [0.9931–0.9974]
  • Share of units warmer than the legal temperature after stabilisation0.00135 [0.00025–0.00581]
  • TheoreticalMaximum OEE quality factor (temperature specification)0.9987 [0.9942–0.9997]
  • TheoreticalMaximum availability (defrost and cleaning)0.863 [0.813–0.915]
  • TheoreticalMaximum performance factor1.00
  • TheoreticalPhysics-bounded OEE0.862 [0.812–0.913]

3Ceilings per dimension

Each dimension shows the ceilings the model computes and, where units are comparable, documented benchmarks. Colour and marker shape identify the ceiling type throughout the Oracle.
  • Theoretical
  • Best available technique
  • Best documented
  • Typical
Energy per unit of productkJ/kg · lower is better
  • TheoreticalThermodynamic minimum work (exergy of the heat removed)32.2 [21.6–43.9] kJ/kg
  • Best available techniqueCompressor work of the best available two-stage ammonia cycle (product load only)153 [124–215] kJ/kg
  • Best available techniqueDOE bandwidth study, state-of-the-art on-site energy intensity of freezing (frozen fruit); whole freezing step including fans and losses · Benchmark649 kJ/kg
  • Best documentedLowest documented electricity use of deep freezing (vegetables)288 kJ/kg
  • TypicalDOE bandwidth study, current typical on-site energy intensity of freezing (frozen fruit) · Benchmark1,363 kJ/kg
  • TypicalCurrent typical on-site energy intensity of freezing (United States)1,363 kJ/kg
Cycle or residence timeh · lower is better
  • TheoreticalMinimum freezing time (Plank, surface at air temperature)0.340 [0.327–0.556] h
  • Best available techniqueFreezing time with the best available surface coefficient (impingement)0.680 [0.556–1.033] h
  • Best documentedMeasured freezing time (to −20 °C at the centre) of 50 mm haddock fillets on a metal tray, air blast 4 m/s, from 5 °C at −35 °C: 2 h 05 min · Benchmark2.08 h
Throughput1/h · higher is better
  • TheoreticalMaximum throughput per unit of product held in the freezer (theoretical)2.94 [1.80–3.06] 1/h
  • Best available techniqueThroughput per unit of product held in the freezer (best available technique)1.47 [0.97–1.80] 1/h
Yieldhigher is better
  • TheoreticalMaximum yield (mass balance)0.9958 [0.9931–0.9974]
Qualityhigher is better
  • TheoreticalMaximum OEE quality factor (temperature specification)0.9987 [0.9942–0.9997]
Availabilityhigher is better
  • TheoreticalMaximum availability (defrost and cleaning)0.863 [0.813–0.915]
Performancehigher is better
  • TheoreticalMaximum performance factor1.00
OEEhigher is better
  • TheoreticalPhysics-bounded OEE0.862 [0.812–0.913]

4Loss cascade

What separates each ceiling from the next one, from the physical limit to typical operation.
  1. TheoreticalBest available techniqueEnergy per unit of product
    32.2 [21.6–43.9] kJ/kg → 153 [124–215] kJ/kg (gap 121 kJ/kg)

    The theoretical work is the exergy of the heat removed: every kilojoule is pumped reversibly from the product temperature at which it is released up to the ambient temperature. A real plant pumps all of it from the evaporating temperature, which must be below the coldest air (−40 °C instead of the product's −1 °C plateau where most heat is released); it rejects at a condensing temperature above ambient; it throttles instead of expanding through a turbine; and its compressors are not isentropic. Each of these irreversibilities multiplies the work: the evaporating temperature alone explains about half of the gap, because most of the heat is latent heat released near −1 °C but lifted from −40 °C.

  2. Best available techniqueBest documentedEnergy per unit of product
    153 [124–215] kJ/kg → 288 kJ/kg (gap 135 kJ/kg)

    The best available cycle covers only the product load. A real freezer also removes the heat of its fans (25 to 30 percent of the refrigeration requirement even in a good design, according to FAO), heat leaking through the insulation, infiltrating moist air, defrost heat and the cooling of trays and belts; it operates part of the time at part load. The documented electricity use refers to deep freezing of vegetables (more water, hence about 10 % more heat per kilogram than lean fish), including auxiliaries.

  3. Best documentedTypicalEnergy per unit of product
    288 kJ/kg → 1,363 kJ/kg (gap 1,075 kJ/kg)

    Typical plants run evaporators colder than needed, condensing pressures higher than the climate allows, fans at full speed independent of load and residence times longer than required, and undershoot the product temperature far below −18 °C; the BAT reference document lists the adjustment of evaporating temperature, residence time and air flow as the largest savings.

  4. TheoreticalBest available techniqueCycle or residence time
    0.340 [0.327–0.556] h → 0.680 [0.556–1.033] h (gap 0.340 h)

    The theoretical freezing time assumes that the product surface is held at the air temperature (infinite surface coefficient), so only conduction through the growing frozen layer limits the rate. Even impingement jets leave a finite boundary-layer resistance P·a/h that adds to the conduction term R·a²/k.

  5. Best available techniqueBest documentedCycle or residence time

    In a conventional air blast at 4–5 m/s the surface resistance 1/h is several times the conduction resistance of a 50 mm slab, the tray adds contact resistance on one face, air warms as it crosses the load and the product is cooled from its initial temperature before freezing starts (a term Plank's equation neglects). The documented time is also measured to −20 °C at the centre.

  6. TheoreticalBest documentedYield

    The psychrometric minimum assumes that the air around the product is saturated at the air temperature and that all heat leaves the product surface by convection, carrying moisture at the lowest possible ratio of mass to heat (the slope of the ice saturation curve at the air temperature). Real air is unsaturated after it leaves the coil, the surface is warmer than the air, and moisture is lost from wet surfaces before freezing, so the measured loss is higher; drip and damage losses are technology.

  7. TheoreticalTypicalAvailability

    Defrost frequency follows the frost load (product dehydration plus infiltration) divided by the frost the coil can hold; a good design runs 8 h between defrosts, a poor one only 2 h. Cleaning is set by hygiene practice.

5Physics-bounded OEE

The highest OEE the operation can reach: availability, performance and quality limited only by unavoidable stops and physical capability.
Fraction of loading time that remains after each unavoidable loss. Axis starts at 0.75.
  • Loading time (ideal)
    1.000
  • Availability
    factor 0.863 [0.813–0.915]
    0.863 (−0.137)
  • Performance
    factor 1.00
    0.863 (−0.000)
  • Quality
    factor 0.9987 [0.9942–0.9997]
    0.862 (−0.001)
  • Physics-bounded OEE
    0.862 [0.812–0.913]

6Full derivation

Mass and energy balances, rate-limiting step, process capability and unavoidable stops, with every term explained.

Scope. Chilled fillets of lean white fish, 50 mm thick on metal trays, frozen in an air-blast freezer from about 5 °C until every point of the product is at −18 °C or lower after thermal stabilisation, which is the legal definition of a quick-frozen food in the European Union (Directive 89/108/EEC). Heat is pumped to the ambient by a two-stage ammonia plant. Chilling is the special case in which the final temperature stays above the initial freezing point TfT_f: the latent part of the enthalpy and the frozen-layer term of the freezing time vanish, and every balance below applies unchanged. Symbols, values, units, uncertainty distributions and sources are listed in the parameter table generated from config/physics/refrigeration.yaml.

1. Mass balance — maximum yield

Freezing changes the state of the water inside the product but not the amount of product, except for the water that leaves the surface as vapour (sublimation from ice, evaporation from a wet surface before the surface freezes). That vapour is carried by the air to the evaporator coil, where it is deposited as frost. Drip, pieces broken off and product stuck to trays are losses of the installed technology and are not counted here.

The physical lower bound of the dehydration follows from the analogy between heat and mass transfer across the same boundary layer. With a convective heat-transfer coefficient hch_c, the Chilton–Colburn analogy gives the mass-transfer coefficient

hm=hcρa cp,a Le2/3,h_m = \frac{h_c}{\rho_a\, c_{p,a}\, Le^{2/3}},

where ρa\rho_a and cp,ac_{p,a} are the density and specific heat of the air (their product is the heat stored per cubic metre of air and per kelvin), and Le=αa/DvaLe = \alpha_a / D_{va} is the Lewis number, the ratio of the thermal diffusivity of air to the diffusivity of water vapour in air. The local convective heat flux and vapour flux are

qc′′=hc (Ts−Ta),m′′=hm (ρv(Ts)−ρv,∞),q_c'' = h_c\,(T_s - T_a), \qquad m'' = h_m\,\big(\rho_{v}(T_s) - \rho_{v,\infty}\big),

with TsT_s the surface temperature, TaT_a the air temperature, ρv(Ts)\rho_v(T_s) the vapour density in equilibrium with the surface (saturation over ice or water) and ρv,∞\rho_{v,\infty} the vapour density of the bulk air. Dividing,

m′′qc′′=1ρacp,aLe2/3  ρv(Ts)−ρv,∞Ts−Ta.\frac{m''}{q_c''} = \frac{1}{\rho_a c_{p,a} Le^{2/3}}\;\frac{\rho_{v}(T_s) - \rho_{v,\infty}}{T_s - T_a}.

The driving force is smallest when the bulk air is saturated at its own temperature, ρv,∞=ρv,sat(Ta)\rho_{v,\infty} = \rho_{v,sat}(T_a) (air can carry no more vapour than that without forming fog). The second factor is then the slope of the chord of the saturation curve between TaT_a and TsT_s; because the saturation curve is convex, every chord is steeper than the tangent at its lower end. Therefore, whatever the surface temperature history,

m′′qc′′  ≥  1ρacp,aLe2/3dρv,satdT∣Ta.\frac{m''}{q_c''} \;\ge\; \frac{1}{\rho_a c_{p,a} Le^{2/3}}\left.\frac{d\rho_{v,sat}}{dT}\right|_{T_a}.

Integrating over the whole freezing, the convected heat is the share fcf_c of the total heat removed qq (section 2) that leaves the surface by convection rather than by radiation. Radiation carries no moisture; it is linearised as hr=4εσTˉ3h_r = 4\varepsilon\sigma \bar T^3 with surface emissivity ε\varepsilon, Stefan–Boltzmann constant σ\sigma and mean absolute temperature Tˉ\bar T between the freezing point and the air, so that fc=hc/(hc+hr)f_c = h_c/(h_c + h_r). The minimum relative weight loss and the yield ceiling are

Δmin=fc qρacp,aLe2/3dρv,satdT∣Ta,Ymax=1−Δmin.\Delta_{min} = \frac{f_c\, q}{\rho_a c_{p,a} Le^{2/3}}\left.\frac{d\rho_{v,sat}}{dT}\right|_{T_a}, \qquad Y_{max} = 1 - \Delta_{min}.

The saturation vapour pressure over ice, pw,sat(T)p_{w,sat}(T), comes from the psychrometric formulation in CoolProp; the vapour density is ρv=pwMw/(RT)\rho_v = p_w M_w/(R T), with MwM_w the molar mass of water and RR the molar gas constant, and its derivative is evaluated numerically at TaT_a. The bound does not depend on the freezer design or the freezing time: it only states the lowest ratio of water to heat that convective cooling can achieve. It vanishes for contact (plate) freezers, where heat leaves through metal and not through air — consistent with the zero loss that FAO reports for blocks frozen in contact freezers. Air-blast losses documented by FAO (1.0 % for a tray of fillets, 1.2 % for individually frozen haddock) lie above the bound because real air leaving the coil is not saturated, the surface is warmer than the air and wet surfaces lose water before they freeze.

The same water is the minimum frost load on the coil: Δmin\Delta_{min} kilograms of frost per kilogram of product, reported per tonne.

2. Energy balance — minimum energy per unit of product

Heat removed. The enthalpy of the product as a function of temperature, h(T)h(T), is taken from the FAO table for white fish (datum −40 °C, kcal/kg, converted with 4.184 kJ/kcal). It already contains the sensible heat of the unfrozen product, the latent heat of the water that freezes progressively over a range of temperatures below TfT_f, and the sensible heat of the frozen product. Between tabulated temperatures hh is linear. With TinT_{in} the inlet temperature and TendT_{end} the mean equilibrium temperature at the outlet (section 4),

q=sh [h(Tin)−h(Tend)],q = s_h\,\big[h(T_{in}) - h(T_{end})\big],

where shs_h is a multiplier for species and seasonal variation.

Thermodynamic minimum. A kilojoule of heat released by the product at temperature TT below the ambient T0T_0 can be pumped to the ambient by a reversible (Carnot) machine with the work (T0/T−1)(T_0/T - 1) kJ. Since the product releases its heat over a range of temperatures, the minimum work is the exergy of the heat removed,

wmin=∫TendTin(T0T−1)dhdT dT=∑ici[T0ln⁡Tb,iTa,i−(Tb,i−Ta,i)],w_{min} = \int_{T_{end}}^{T_{in}} \left(\frac{T_0}{T} - 1\right) \frac{dh}{dT}\,dT = \sum_i c_i\left[T_0 \ln\frac{T_{b,i}}{T_{a,i}} - (T_{b,i} - T_{a,i})\right],

where the sum runs over the linear segments of the enthalpy table, cic_i is the slope dh/dTdh/dT of segment ii (its apparent specific heat, which is very large in the band just below TfT_f where most of the water freezes) and [Ta,i,Tb,i][T_{a,i}, T_{b,i}] is the part of that segment between TendT_{end} and TinT_{in}, in kelvin. The ratio q/wminq/w_{min} is the coefficient of performance of an ideal machine that follows the product temperature.

Sign convention and the DOE "thermodynamic minimum". The US DOE bandwidth study reports a negative thermodynamic minimum for frozen products (for general vegetables −124.4 kJ/kg, −53.5 Btu/lb). Its appendix computes it as ΔG=−17(1.26a+0.84)−26(3.35a+0.48)\Delta G = -17(1.26a + 0.84) - 26(3.35a + 0.48) kJ/kg, i.e. −[(TR−TF) cp,fresh+(TF−Tfz) cp,frozen]-[(T_R - T_F)\,c_{p,fresh} + (T_F - T_{fz})\,c_{p,frozen}] with room temperature TRT_R = 25 °C, freezing temperature TFT_F = −1 °C, freezer temperature TfzT_{fz} = −18 °C and Siebel specific heats for moisture content aa. The model reproduces it exactly from the quoted inputs (output doe_tm_reproduced). Two features separate it from the quantity computed here:

  1. Sign. The DOE study books energy released by the material as negative ("the process of freezing is exothermic"). Here the energy that the process must be supplied with is positive: the product releases qq, but that heat sits below the ambient temperature and can only be rejected by spending work.
  2. Content. The DOE value contains only the sensible heat above and below the freezing point; the latent heat is excluded because the Gibbs energy of a phase change at equilibrium is zero. It is therefore neither the enthalpy released (which is about three times larger because of the latent heat) nor the work of a refrigerator.

In the DOE framework, cooling to room temperature is free and heat is assumed perfectly recovered, so the cost of moving heat below the ambient is not counted. The minimum work of a refrigeration process cannot be negative: by the second law, removing heat from a body colder than the surroundings requires positive work. The DOE number is kept in the benchmark list with compare: false, and the theoretical ceiling of this model is wmin>0w_{min} > 0.

Best available technique. The heat is lifted from an evaporating temperature Te=Ta−ΔTeT_e = T_a - \Delta T_e (the refrigerant must be colder than the air, which must be colder than the product) to a condensing temperature Tc=T0+ΔTcT_c = T_0 + \Delta T_c. Because the pressure ratio of ammonia between about −40 °C and +35 °C is close to 20, industrial plants use two compression stages. The cycle computed with CoolProp is: (1) saturated vapour leaves the evaporator at pep_e; (2) the low-stage compressor raises it to the intermediate pressure pi=pepcp_i = \sqrt{p_e p_c} with isentropic efficiency ηis\eta_{is}, h2=h1+(h2s−h1)/ηish_2 = h_1 + (h_{2s} - h_1)/\eta_{is}; in an open flash intercooler the discharge gas is desuperheated by evaporating liquid, the liquid from the condenser (5, saturated at pcp_c) is flashed, saturated vapour (3) goes to the high stage and saturated liquid (7) at pip_i is throttled to the evaporator; (4) the high stage compresses to pcp_c with the same efficiency. The energy balance of the flash vessel gives the high-stage flow per kilogram of low-stage flow, m˙H/m˙L=(h2−h7)/(h3−h5)\dot m_H/\dot m_L = (h_2 - h_7)/(h_3 - h_5), and

COP=h1−h7(h2−h1)+m˙Hm˙L(h4−h3),wBAT=qCOP.COP = \frac{h_1 - h_7}{(h_2 - h_1) + \frac{\dot m_H}{\dot m_L}(h_4 - h_3)}, \qquad w_{BAT} = \frac{q}{COP}.

The output work_carnot_evaporator, q (T0−Te)/Teq\,(T_0 - T_e)/T_e, isolates the cost of lifting all the heat from the evaporating temperature instead of from the product temperature at which it is released; nearly half of the gap between wminw_{min} and wBATw_{BAT} is due to that alone.

The energy ceilings. Theoretical: wminw_{min}. Best available technique: wBATw_{BAT} (compressor work for the product load only). Documented: the lower end of the electricity use of deep freezing of vegetables in the BAT reference document, 80 kWh per tonne. Typical: the DOE current typical on-site energy intensity of freezing, 586 Btu/lb. The DOE state-of-the-art value (279 Btu/lb) and the upper BAT reference value (280 kWh/t) are listed as benchmarks. The documented values cover the whole freezing step of vegetables (fans, heat leaks, infiltration, defrost); vegetables hold more water than lean fish and need about 10 % more heat per kilogram, so the comparison is conservative.

3. Rate-limiting step — minimum freezing time and maximum throughput

Most of the heat is latent heat released at a freezing front that moves from the surface into the product. Plank's quasi-steady model treats the frozen layer of thickness xx and the surface film as two resistances in series between the front, at the initial freezing point TfT_f, and the air at TaT_a; the heat flux through them must carry away the latent heat released by the advancing front:

q′′=Tf−Ta1/h+x/k=ρLdxdt,q'' = \frac{T_f - T_a}{1/h + x/k} = \rho L \frac{dx}{dt},

where hh is the surface heat-transfer coefficient, kk the thermal conductivity of the frozen layer, ρ\rho the density of the product and LL the heat to be removed per kilogram between the initial freezing point and the final temperature (from the same enthalpy table, L=sh[h(Tf)−h(Tend)]L = s_h[h(T_f) - h(T_{end})]). For a slab of thickness aa cooled from both faces the front meets the centre at x=a/2x = a/2; separating variables and integrating,

tF=ρLTf−Ta(P ah+R a2k),P=12,R=18.t_F = \frac{\rho L}{T_f - T_a}\left(\frac{P\,a}{h} + \frac{R\,a^2}{k}\right), \qquad P = \frac{1}{2},\quad R = \frac{1}{8}.

The first term is the time spent pushing heat through the surface film; the second, the time spent conducting it through the frozen layer. Plank's model neglects the sensible heat removed before freezing starts, so it underestimates the real time.

  • Theoretical minimum (h→∞h \to \infty): the surface is held at the air temperature and only conduction through the frozen layer limits the rate, tmin=ρLRa2/(k(Tf−Ta))t_{min} = \rho L R a^2 / (k (T_f - T_a)). No freezer using air at TaT_a can freeze the slab faster.
  • Best available technique: impingement jets, with the surface coefficient hBATh_{BAT}.
  • Model check: with the coefficient of a conventional air blast at 4–5 m/s, the predicted time agrees with the FAO measured time of 2 h 05 min for 50 mm haddock fillets on trays at −35 °C (documented ceiling).

A freezer holds a mass MM of product; in steady operation it delivers M/tFM / t_F per hour. The maximum throughput per unit of product held is therefore 1/tmin1/t_{min} (theoretical) and 1/tF,BAT1/t_{F,BAT} (best available technique).

4. Process capability — quality ceiling

After thermal stabilisation each unit leaves at an equilibrium temperature that varies with its position in the freezer, its thickness and the loading; it is modelled as T∼N(Tend,σT2)T \sim \mathcal N(T_{end}, \sigma_T^2). The legal limit is TspecT_{spec} = −18 °C at all points; the operator sets the mean zz standard deviations below it, Tend=Tspec−z σTT_{end} = T_{spec} - z\,\sigma_T. The share of units warmer than the limit is

d=Φ(−z).d = \Phi(-z).

The margin is paid in energy and time (a lower TendT_{end} raises qq, wminw_{min} and LL). The quality ceiling is Qmax=1−dQ_{max} = 1 - d. The minimum dehydration of section 1 lowers the mass of product that is in specification; it is a material-yield loss, reported as YmaxY_{max}, and does not enter the OEE (as the give-away of the drying and filling models).

5. Unavoidable stops — maximum availability

The water lost by the product and the water carried by infiltrating air are deposited as frost on the evaporator coil; frost bridges the fins, reduces air flow and heat transfer, and has to be melted. The frost load has a physical floor, the minimum dehydration of section 1 times the throughput, so defrosting can never be eliminated in an air-blast freezer; how often it is needed depends on the frost the coil can hold. FAO states that a good design operates for at least 8 h between defrosts (a poor one only 2 h). With a defrost of duration tdt_d after every running interval tdeft_{def}, and a cleaning of duration tclt_{cl} every tcyct_{cyc},

Amax=(1−tcltcyc)tdeftdef+td.A_{max} = \left(1 - \frac{t_{cl}}{t_{cyc}}\right)\frac{t_{def}}{t_{def} + t_d}.

6. Physics-bounded OEE

Scope of the bound. The theoretical label of OEEmaxOEE_{max} means physics-bounded under the declared scenario: the ceiling follows from the physics of the operation for the declared product, equipment and schedule, and it is a bound only conditional on those declarations. It follows the common definition of all models: AmaxA_{max} counts only stops required by regulation or forced by a physical mechanism, at their physical minimum duration, plus the start-up they force; PmaxP_{max} is the ratio of the physics-limited ideal cycle to the cycle of the capable process; QmaxQ_{max} is the share of conforming units, and mass losses of conforming product are reported only as yield.

OEEmax=Amax×Pmax×Qmax,Pmax=1.OEE_{max} = A_{max} \times P_{max} \times Q_{max}, \qquad P_{max} = 1.

Pmax=1P_{max} = 1 refers to the rate of the installed air-blast freezer (its freezing time, section 3): a freezer cannot run faster than its freezing time allows without failing the temperature specification, which is counted in QmaxQ_{max}. The physical limit of the rate is not hidden in PP: the conduction-limited freezing time (infinite surface coefficient) gives the theoretical throughput per unit of product held, capacity_per_holdup_max, reported in the rate dimension together with the best available (impingement) and air-blast values; against that limit the installed freezer reaches about one sixth of the theoretical rate. The freezer is cleaned by hand and with foam, not in place, so the cleaning model does not apply and the cleaning duration of section 5 is declared. The point values and Monte Carlo intervals over the declared uncertainty of every input are reported in the output table.

7. Assumptions and validity

See the assumption and validity lists generated from the configuration file. Thermal properties of the product (enthalpy curve, frozen conductivity, density) are FAO values for lean white fish; they can later be replaced by a composition-based property model. The most influential inputs (sensitivity table) are the ambient temperature and the compressor efficiency for energy, the frozen conductivity and the surface coefficient for time, and the defrost and cleaning durations for availability.

References

  • Food and Agriculture Organization of the United Nations. Freezing and refrigerated storage in fisheries, FAO Fisheries Technical Paper 340, 1994 (chapters 4, 5, 9 and 19).
  • Council Directive 89/108/EEC on quick-frozen foodstuffs for human consumption, consolidated text of 11.12.2008.
  • European Commission, Joint Research Centre. Best Available Techniques (BAT) Reference Document for the Food, Drink and Milk Industries, 2019.
  • U.S. Department of Energy. Bandwidth Study on Energy Use and Potential Energy Saving Opportunities in U.S. Food and Beverage Manufacturing, 2017.
  • Chilton, T. H., Colburn, A. P. (1934). Mass transfer (absorption) coefficients — prediction from data on heat transfer and fluid friction. Industrial & Engineering Chemistry, 26, 1183–1187. doi:10.1021/ie50299a012
  • Pham, Q. T. (1986). Simplified equation for predicting the freezing time of foodstuffs. International Journal of Food Science & Technology, 21, 209–219. doi:10.1111/j.1365-2621.1986.tb00442.x
  • Bell, I. H., Wronski, J., Quoilin, S., Lemort, V. (2014). Pure and pseudo-pure fluid thermophysical property evaluation and the open-source thermophysical property library CoolProp. Industrial & Engineering Chemistry Research, 53, 2498–2508. doi:10.1021/ie4033999

7Parameters and provenance

Every input with its value, unit, role, uncertainty distribution and provenance. Open a quote to read the passage of the source that states the value.

8Benchmarks

Documented values from open sources used to check the ceilings; no documented value may beat the theoretical limit.
BenchmarkCeilingValueProvenance
DOE bandwidth study, 'thermodynamic minimum' of freezing general vegetables. Not comparable: it is the negative sensible heat released by the product (latent heat excluded), not the work a refrigerator needs; see the derivation, section 2.
Show verbatim quote
General vegetables (to produce frozen vegetables) 0.9 3.50 1.97 -124.4 [-53.5]Bandwidth Study on Energy Use and Potential Energy Savings Opportunities in U.S. Food and Beverage Manufacturing (opens in a new tab). U.S. Department of Energy, Office of Energy Efficiency and Renewable Energy, Advanced Manufacturing Office, 2017. p. 56. DOE/EE-1571; OSTI ID 1512366
Theoretical-124.4 kJ/kgVerified quote
Electricity use of deep freezing of vegetables, lower end of the reported range
Show verbatim quote
During this process, energy is consumed at a rate of 80–280 kWhe/t of frozen vegetables.Best Available Techniques (BAT) Reference Document for the Food, Drink and Milk Industries (opens in a new tab). European Commission, Joint Research Centre; Publications Office of the European Union, 2019. p. 450. EUR 29978 EN; JRC118627; doi:10.2760/243911
Best documented80 kW·h/tonneVerified quote
Electricity use of deep freezing of vegetables, upper end of the reported range
Show verbatim quote
During this process, energy is consumed at a rate of 80–280 kWhe/t of frozen vegetables.Best Available Techniques (BAT) Reference Document for the Food, Drink and Milk Industries (opens in a new tab). European Commission, Joint Research Centre; Publications Office of the European Union, 2019. p. 450. EUR 29978 EN; JRC118627; doi:10.2760/243911
Typical280 kW·h/tonneVerified quote
DOE bandwidth study, state-of-the-art on-site energy intensity of freezing (frozen fruit); whole freezing step including fans and losses
Show verbatim quote
Frozen fruit Freezing 1,520 586 279 279 -55Bandwidth Study on Energy Use and Potential Energy Savings Opportunities in U.S. Food and Beverage Manufacturing (opens in a new tab). U.S. Department of Energy, Office of Energy Efficiency and Renewable Energy, Advanced Manufacturing Office, 2017. p. 40. DOE/EE-1571; OSTI ID 1512366
Best available technique279 Btu/lbVerified quote
DOE bandwidth study, current typical on-site energy intensity of freezing (frozen fruit)
Show verbatim quote
Frozen fruit Freezing 1,520 586 279 279 -55Bandwidth Study on Energy Use and Potential Energy Savings Opportunities in U.S. Food and Beverage Manufacturing (opens in a new tab). U.S. Department of Energy, Office of Energy Efficiency and Renewable Energy, Advanced Manufacturing Office, 2017. p. 40. DOE/EE-1571; OSTI ID 1512366
Typical586 Btu/lbVerified quote
Measured freezing time (to −20 °C at the centre) of 50 mm haddock fillets on a metal tray, air blast 4 m/s, from 5 °C at −35 °C: 2 h 05 min
Show verbatim quote
Haddock fillets 50 mm thick on metal tray Air blast 4 m/s 5 -35 2 05Freezing and refrigerated storage in fisheries (FAO Fisheries Technical Paper 340), chapter 5: Freezing time (opens in a new tab). Food and Agriculture Organization of the United Nations, 1994. FAO Fisheries Technical Paper 340; ISBN 92-5-103579-2
Best documented2.0833 hVerified quote
Weight lost by a tray of fillets during air-blast freezing (part of the heat leaves through the tray, which carries no moisture). A loss, not a yield fraction: it is consistent with the bound (1.0 % against the minimum dehydration of yield_max) but not tested by the yield gate
Show verbatim quote
Tray of fillets Air blast 1.0Freezing and refrigerated storage in fisheries (FAO Fisheries Technical Paper 340), chapter 9: Weight loss from fish during freezing and cold storage (opens in a new tab). Food and Agriculture Organization of the United Nations, 1994. FAO Fisheries Technical Paper 340; ISBN 92-5-103579-2
Best documented1 %Verified quote
Weight lost by individually quick-frozen haddock in an air-blast freezer (a loss, not a yield fraction; not tested by the yield gate)
Show verbatim quote
IQF haddock Air blast 1.2Freezing and refrigerated storage in fisheries (FAO Fisheries Technical Paper 340), chapter 9: Weight loss from fish during freezing and cold storage (opens in a new tab). Food and Agriculture Organization of the United Nations, 1994. FAO Fisheries Technical Paper 340; ISBN 92-5-103579-2
Typical1.2 %Verified quote
Weight lost by blocks frozen in a contact (plate) freezer: no convective exchange with air, hence no dehydration (different freezer type)
Show verbatim quote
Blocks of fish Contact freezer metal to fish contact 0Freezing and refrigerated storage in fisheries (FAO Fisheries Technical Paper 340), chapter 9: Weight loss from fish during freezing and cold storage (opens in a new tab). Food and Agriculture Organization of the United Nations, 1994. FAO Fisheries Technical Paper 340; ISBN 92-5-103579-2
Theoretical0 %Verified quote
Defrost interval of a poorly designed air-blast freezer (hours between defrosts, not an availability fraction)
Show verbatim quote
A good freezer design should be able to operate for at least 8h before a defrost is required but a poor design may require defrosting every 2h.Freezing and refrigerated storage in fisheries (FAO Fisheries Technical Paper 340), chapter 4: Freezers (opens in a new tab). Food and Agriculture Organization of the United Nations, 1994. FAO Fisheries Technical Paper 340; ISBN 92-5-103579-2
Typical2 hVerified quote

9Unavoidable stops

Stops that physics or regulation impose; they set the maximum availability.
  • Defrost of the evaporator coilVerified quote

    Water lost by the product and water carried by infiltrating air are deposited as frost on the coil; frost bridges the fins, reduces air flow and heat transfer, and must be melted periodically. The minimum frost load from the product is the psychrometric minimum dehydration computed by the model.

    Show verbatim quote
    Moisture lost from fish during freezing and from air infiltrating into the cooler will eventually be deposited as frost on the cooler surface.Freezing and refrigerated storage in fisheries (FAO Fisheries Technical Paper 340), chapter 4: Freezers (opens in a new tab). Food and Agriculture Organization of the United Nations, 1994. FAO Fisheries Technical Paper 340; ISBN 92-5-103579-2
  • Cleaning and sanitation of the freezerDeclared assumption

    Food-contact belts, trays and air paths accumulate product debris and must be cleaned; the interval and duration are declared assumptions.

10Assumptions

  1. Performance: P = 1 refers to the freezing time of the installed air-blast freezer; the conduction limit (infinite surface coefficient) is reported in the rate dimension (capacity_per_holdup_max), not hidden in P. Availability: coil defrost (frost from product dehydration and infiltration) and the sanitation of the freezer are counted; the freezer is cleaned by hand and with foam, not in place, so the cleaning model does not apply and the cleaning duration is declared. Two-stage ammonia cycles and impingement freezing are documented technologies; their bat values are computed with declared performance parameters (approach temperatures, isentropic efficiency, surface coefficient).
  2. The tabulated enthalpy of lean white fish (FAO) represents the product; its differences are scaled by ±5 % for species and season. FAO tables are in kilocalories; the conversion uses 4.184 kJ/kcal (the thermochemical calorie, 0.07 % below the International Table calorie).
  3. The thermodynamic minimum work is the exergy of the heat removed with the ambient as the only heat sink; enthalpy is linear between tabulated temperatures.
  4. Freezing time follows Plank's quasi-steady model for an infinite slab cooled from both faces; it neglects pre-cooling above the freezing point and is therefore a lower bound of the real time.
  5. The best available cycle is a two-stage ammonia cycle with an open flash intercooler at the geometric-mean pressure, saturated states at evaporator and condenser outlets and equal isentropic efficiency in both stages; fans, heat leaks and defrost heat are excluded from it.
  6. Minimum dehydration: the air is saturated over ice at the air temperature, all heat leaves through the product surface (no tray contact), and the convective share of the surface heat flux carries moisture according to the Chilton–Colburn analogy.
  7. The equilibrium temperature of the product units is normally distributed; units warmer than −18 °C after stabilisation are off-specification. The performance factor is 1 at the conduction-limited rate of the installed freezer.

11Range of validity

  • Lean, high-moisture foods frozen from the chilled state in moving air (fish fillets, meat cuts, vegetables of comparable water content); slabs whose thickness is small compared with their length and width.
  • Contact (plate) and cryogenic freezers change the surface coefficient and remove the dehydration term; the energy balance and the exergy minimum still apply.

12Sensitivity

Elasticity: relative change of the output for a relative change of each parameter. Spearman rank correlation from the Monte Carlo draws is shown alongside.
  • tcyct_{cyc}Production time between two full cleanings of the freezer
    elasticity +0.091Spearman ρ —
  • tclt_{cl}Duration of the cleaning, drying and cool-down sequence
    elasticity -0.091Spearman ρ -0.85
  • tdeft_{def}Running time between two defrosts of the evaporator coil (good design)
    elasticity +0.059Spearman ρ —
  • tdt_{d}Duration of one coil defrost including recovery of the air temperature
    elasticity -0.059Spearman ρ -0.52
  • zzMargin of the mean equilibrium temperature below the legal limit, in standard deviations
    elasticity +0.014Spearman ρ +0.04
  • Output rises with the parameter
  • Output falls with the parameter

13Validation checks

A model is published only when every check passes.
  • Dimensional analysis23 outputs converted to their declared unitsPassed
  • Every cited value verified against its source43 verified, 0 failedPassed
  • No documented value beats the theoretical limitPassed
  • Ceilings in physical orderPassed
  • Fractions within 0 and 1Passed

14Causes of loss in this operation

Probability that each cause is documented as a relevant loss of a line, pooled over all segments.

Evidence synthesis in progress

Cause probabilities will appear here once the included studies have been extracted and the Bayesian model has passed its convergence checks: each cause with its probability, 95% credible interval, number of studies and units, and evidence grade.

Pipeline so far: 33,254 records identified, 16,957 unique, 16,954 screened, 1,469 passed screening.

15References

  1. Freezing and refrigerated storage in fisheries (FAO Fisheries Technical Paper 340), chapter 4: Freezers (opens in a new tab). Food and Agriculture Organization of the United Nations, 1994. FAO Fisheries Technical Paper 340; ISBN 92-5-103579-2
  2. Freezing and refrigerated storage in fisheries (FAO Fisheries Technical Paper 340), chapter 5: Freezing time (opens in a new tab). Food and Agriculture Organization of the United Nations, 1994. FAO Fisheries Technical Paper 340; ISBN 92-5-103579-2
  3. Freezing and refrigerated storage in fisheries (FAO Fisheries Technical Paper 340), chapter 9: Weight loss from fish during freezing and cold storage (opens in a new tab). Food and Agriculture Organization of the United Nations, 1994. FAO Fisheries Technical Paper 340; ISBN 92-5-103579-2
  4. Freezing and refrigerated storage in fisheries (FAO Fisheries Technical Paper 340), chapter 19: Some related facts and figures (opens in a new tab). Food and Agriculture Organization of the United Nations, 1994. FAO Fisheries Technical Paper 340; ISBN 92-5-103579-2
  5. Council Directive 89/108/EEC of 21 December 1988 on the approximation of the laws of the Member States relating to quick-frozen foodstuffs for human consumption (consolidated text 11.12.2008) (opens in a new tab). Publications Office of the European Union, 2008. p. 4. CELEX 01989L0108-20081211; OJ L 40, 11.2.1989, p. 34
  6. Best Available Techniques (BAT) Reference Document for the Food, Drink and Milk Industries (opens in a new tab). European Commission, Joint Research Centre; Publications Office of the European Union, 2019. p. 450. EUR 29978 EN; JRC118627; doi:10.2760/243911
  7. Bandwidth Study on Energy Use and Potential Energy Savings Opportunities in U.S. Food and Beverage Manufacturing (opens in a new tab). U.S. Department of Energy, Office of Energy Efficiency and Renewable Energy, Advanced Manufacturing Office, 2017. p. 40. DOE/EE-1571; OSTI ID 1512366
  8. doi:10.1021/ie4033999 — Thermophysical property library (ammonia, water, humid air).
  9. doi:10.1021/ie50299a012 — Analogy between heat and mass transfer (minimum dehydration).
  10. doi:10.1111/j.1365-2621.1986.tb00442.x — Freezing-time prediction from Plank-type equations.