foratvise
Get Started
Concentration and drying

Evaporation

Concentration of whole milk in a falling-film evaporator

Validated modelOEE bound 0.873 [0.748–0.932]20,000 Monte Carlo samples · seed 20261115 · computed November 16, 2026

In short

Evaporation can use at most 87 of every 100 planned hours. The rest goes to limits no plant can avoid.

  1. 87usable at most
  2. 13are a physical limit

The three limits that set it

Running time: up to 95 %

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

Speed: up to 93 %

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

Good product: up to 98 %

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

Pasteurized whole milk is concentrated under vacuum in a falling-film evaporator: the milk runs as a thin film down the inside of vertical tubes heated by condensing vapour, part of its water boils off, and the concentrate leaves at the solids content required by the spray dryer. The heating vapour may be live steam (single or multiple effects), vapour recompressed by a steam-driven thermocompressor (TVR) or vapour recompressed by an electrically driven fan or compressor (MVR). Energy results are given per kg of water removed.

Provenance of inputsComputed4Declared assumption13Verified quote21

2Results

Point value from the nominal parameters and 95% interval from 20,000 Monte Carlo draws of the uncertain parameters.
  • Water removed per kg of concentrate3.21 [3.23–4.03] kg/kg
  • Water removed per kg of milk fed0.763 [0.764–0.801] kg/kg
  • Water activity of the final concentrate (ideal solution of the dissolved solutes)0.967 [0.949–0.968]
  • Boiling-point elevation of the final concentrate0.731 [0.687–1.153] K
  • TheoreticalThermodynamic minimum: reversible work of separation per kg of water removed1.46 [1.34–1.80] kJ/kg
  • Carnot-type bound of mechanical vapour recompression for the declared film temperature difference, per kg of water22.6 [16.1–29.6] kJ/kg
  • Isentropic compression work of mechanical vapour recompression (CoolProp), per kg of water23.2 [16.4–30.5] kJ/kg
  • Best available techniqueEnergy of a mechanical vapour recompression evaporator (best available technique), per kg of water36.0 kJ/kg
  • Best documentedDocumented total energy of a single-stage MVR evaporator, per kg of water54.0 kJ/kg
  • TypicalDocumented total energy of TVR evaporators with the common number of stages (3 to 5), per kg of water396 [308–499] kJ/kg
  • Ideal heat of a single effect (latent heat, feed at the boiling point), per kg of water2,356 [2,332–2,379] kJ/kg
  • Ideal heat of an N-effect evaporator (steam economy N), per kg of water336.6 [333.1–339.8] kJ/kg
  • TheoreticalIdeal steam per kg of water, single effect1.0031 [1.0021–1.0041] kg/kg
  • TheoreticalIdeal steam per kg of water, double effect0.5035 [0.5026–0.5047] kg/kg
  • TheoreticalIdeal steam per kg of water, triple effect0.3370 [0.3361–0.3382] kg/kg
  • Ideal (isentropic) thermocompressor entrainment: kg of vapour recompressed per kg of motive steam20.7 [15.1–28.6] kg/kg
  • Ideal motive steam per kg of water, single effect with thermal vapour recompression0.0462 [0.0338–0.0625] kg/kg
  • Temperature lift implied by the BAT energy if it were ideal Carnot work5.09 [4.90–5.29] K
  • TheoreticalMaximum evaporation rate per unit of heating surface, clean film (U dT / h_fg)9.17 [4.03–16.03] kg/m²/h
  • Mean evaporation rate per unit of surface over a fouling-limited run8.55 [3.81–14.71] kg/m²/h
  • Run length until fouling reduces the duty to the declared floor (typical practice)21.4 [4.5–109.1] h
  • TheoreticalRun length that maximises A x P x Q (capped by the legal limit)7.37 [3.73–16.30] h
  • TheoreticalMaximum performance factor: mean over clean rate during a run at constant temperature difference0.933 [0.864–0.968]
  • TheoreticalMaximum availability (minimum cleaning after each optimal run)0.952 [0.907–0.978]
  • TypicalAvailability with the declared duty floor and a typical cleaning programme0.915 [0.659–0.963]
  • Share of product lost in the water-product mixing at start-up and shut-down0.0170 [0.0054–0.0462]
  • TheoreticalMaximum solids yield0.983 [0.954–0.995]
  • Share of concentrate outside the solids tolerance band6.33 × 10⁻⁵ [3.23 × 10⁻¹⁰–4.17 × 10⁻²]
  • TheoreticalMaximum OEE quality factor (solids control and start-up losses)0.983 [0.933–0.993]
  • TheoreticalPhysics-bounded OEE0.873 [0.748–0.932]

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
  • TheoreticalDOE bandwidth study, concentration, thermodynamic minimum (set to zero for lack of chemical conversion; the separation work is not counted) · Benchmark0 kJ/kg
  • TheoreticalThermodynamic minimum: reversible work of separation per kg of water removed1.46 [1.34–1.80] kJ/kg
  • Best available techniqueEnergy of a mechanical vapour recompression evaporator (best available technique), per kg of water36.0 kJ/kg
  • Best documentedDocumented total energy of a single-stage MVR evaporator, per kg of water54.0 kJ/kg
  • Best documentedTotal energy of a 7-stage TVR evaporator (dairy industry comparison table) · Benchmark216 kJ/kg
  • TypicalDocumented total energy of TVR evaporators with the common number of stages (3 to 5), per kg of water396 [308–499] kJ/kg
  • TypicalTotal energy of a 3-stage TVR evaporator (dairy industry comparison table) · Benchmark504 kJ/kg
steam_single_effectkg/kg · lower is better
  • TheoreticalIdeal steam per kg of water, single effect1.0031 [1.0021–1.0041] kg/kg
steam_double_effectkg/kg · lower is better
  • TheoreticalIdeal steam per kg of water, double effect0.5035 [0.5026–0.5047] kg/kg
steam_triple_effectkg/kg · lower is better
  • TheoreticalIdeal steam per kg of water, triple effect0.3370 [0.3361–0.3382] kg/kg
Throughputkg/m²/h · higher is better
  • TheoreticalMaximum evaporation rate per unit of heating surface, clean film (U dT / h_fg)9.17 [4.03–16.03] kg/m²/h
run_lengthh · higher is better
  • TheoreticalRun length that maximises A x P x Q (capped by the legal limit)7.37 [3.73–16.30] h
Performancehigher is better
  • TheoreticalMaximum performance factor: mean over clean rate during a run at constant temperature difference0.933 [0.864–0.968]
Availabilityhigher is better
  • TheoreticalMaximum availability (minimum cleaning after each optimal run)0.952 [0.907–0.978]
  • TypicalAvailability with the declared duty floor and a typical cleaning programme0.915 [0.659–0.963]
Yieldhigher is better
  • TheoreticalMaximum solids yield0.983 [0.954–0.995]
Qualityhigher is better
  • TheoreticalMaximum OEE quality factor (solids control and start-up losses)0.983 [0.933–0.993]
OEEhigher is better
  • TheoreticalPhysics-bounded OEE0.873 [0.748–0.932]

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
    1.46 [1.34–1.80] kJ/kg → 36.0 kJ/kg (gap 34.5 kJ/kg)

    The theoretical minimum is the reversible work of separating water from the dissolved milk solutes. Milk is a dilute solution (water activity above 0.99, falling to about 0.97 at 50 % solids), so this work is only a few kJ per kg of water. Any evaporator spends far more for three physical reasons. First, a temperature difference across the film is needed to drive heat through the tube wall in finite area; with mechanical vapour recompression this lift is paid as compressor work, and the Carnot-type bound for the declared difference already exceeds the separation work more than ten times. Second, the boiling-point elevation must be overcome on top of it (it is the separation work itself, seen through Clausius-Clapeyron). Third, a real compressor is not isentropic and its motor is not loss-free, and fans, pumps, vacuum and the steam-heated finisher add consumption. The best available technique (MVR, about 10 kWh per tonne of water) sits close to these bounds.

  2. Best available techniqueBest documentedEnergy per unit of product
    36.0 kJ/kg → 54.0 kJ/kg (gap 18.0 kJ/kg)

    The comparison table of the BAT reference document gives 15 kWh per tonne for a single-stage MVR evaporator, against the 10 kWh per tonne reported for MVR in the technique description: the spread reflects compressor and fan efficiency, the temperature lift chosen at design (smaller lift needs more heating surface) and the auxiliaries counted.

  3. Best documentedTypicalEnergy per unit of product
    54.0 kJ/kg → 396 [308–499] kJ/kg (gap 342 kJ/kg)

    Most installed evaporators are steam-driven multi-effect trains with thermal vapour recompression of three to five stages. Each kilogram of live steam is reused once per effect plus the vapour entrained by the thermocompressor, but the latent heat of the vapour from the last effect is rejected to cooling water, and the thermocompressor is far from isentropic. This is why they need roughly six to fourteen times the final energy of an MVR evaporator; the ideal-economy outputs (heat of a single and of an N-effect evaporator) show how far even a perfect steam-driven train stays above MVR, because it delivers heat rather than work.

  4. TheoreticalBest documentedsteam_single_effect

    An ideal single effect with the feed already at its boiling point needs a little more than one kilogram of steam per kilogram of water, because the latent heat of the condensing steam at the higher temperature is smaller than that of the water boiling off. Documented single-stage evaporators need 1.1 to 1.4: the difference is the sensible heating of feed that enters below its boiling point, heat losses, venting of non-condensable gases and condensate sub-cooling.

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 0.70.
  • Loading time (ideal)
    1.000
  • Availability
    factor 0.952 [0.907–0.978]
    0.952 (−0.048)
  • Performance
    factor 0.933 [0.864–0.968]
    0.888 (−0.064)
  • Quality
    factor 0.983 [0.933–0.993]
    0.873 (−0.015)
  • Physics-bounded OEE
    0.873 [0.748–0.932]

6Full derivation

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

Scope. Pasteurized whole milk is concentrated under vacuum in a falling-film evaporator: it runs as a thin film down the inside of vertical tubes heated by condensing vapour, part of its water boils off, and the concentrate leaves at the solids content required by the spray dryer. The heating vapour may be live steam (one or several effects), vapour recompressed by a steam-driven thermocompressor (thermal vapour recompression, TVR) or vapour recompressed by an electric fan or compressor (mechanical vapour recompression, MVR). Energy results are expressed per kilogram of water removed. Symbols and their values, units, uncertainty distributions and sources are listed in the parameter table generated from config/physics/evaporation.yaml; numbers quoted in this text are the point values of that table.

1. Mass balance — maximum yield

Evaporation removes water and conserves solids. With xFx_F the total solids of the milk fed (1 minus its water content, 88.13 % in the USDA record of whole milk) and xCx_C the total solids of the concentrate (50 %, the lower end of the 50–60 % dry matter reported in the BAT reference document), the solids balance F xF=C xCF\,x_F = C\,x_C gives

F−CC=xCxF−1=3.21 kg water per kg concentrate,F−CF=1−xFxC=0.763 kg water per kg milk,\frac{F - C}{C} = \frac{x_C}{x_F} - 1 = 3.21 \ \text{kg water per kg concentrate}, \qquad \frac{F-C}{F} = 1 - \frac{x_F}{x_C} = 0.763 \ \text{kg water per kg milk},

where FF and CC are the masses of feed and concentrate.

Yield. No solids leave with the vapour in the ideal operation, so the steady-state solids yield is 1. Solids are lost in transients: at start-up the evaporator is full of water that the milk displaces, and at shut-down water pushes the last concentrate out. Each transition creates a water–product mixed zone that is a fraction κ\kappa of the hold-up volume; with a mean residence time of the product in the evaporator τ\tau (hold-up over flow) and a run length trunt_{run},

Ymax=1−2 κ τtrun,Y_{max} = 1 - \frac{2\,\kappa\,\tau}{t_{run}},

0.983 at the point values (τ\tau = 5 min, κ\kappa = 0.75, trunt_{run} = 7.4 h from Section 3). Entrainment of droplets into the vapour, leaks and rinsing losses depend on the installed technology.

2. Energy balance — minimum energy per kilogram of water

2.1 Thermodynamic minimum: reversible work of separation

The dissolved milk solutes (mainly lactose and the soluble salts) lower the chemical potential of the water. Separating a small mass of water dmwdm_w reversibly at the dead-state temperature T0T_0 and delivering it as pure liquid water requires δW=(RT0/Mw)ln⁡(1/aw) dmw\delta W = (R T_0/M_w)\ln(1/a_w)\,dm_w, where RR is the molar gas constant, MwM_w the molar mass of water and awa_w the water activity. For an ideal (Raoult) solution of nsn_s osmoles of dissolved solutes in a mass of water mwm_w,

aw=nwnw+ns=mwmw+a,a=Mw ns.a_w = \frac{n_w}{n_w + n_s} = \frac{m_w}{m_w + a}, \qquad a = M_w\,n_s .

Per kilogram of milk, nsn_s counts the lactose (5.05 % in the USDA record, molar mass 342.30 g/mol) and the soluble monovalent cations potassium and sodium (132 and 43 mg per 100 g), plus a declared share faf_a of osmotically active counter-ions (between 0 and 1 per cation); casein micelles, fat and colloidal calcium phosphate do not lower the water activity. With fa=0.5f_a = 0.5 this is 0.257 osmol per kg of water, and awa_w = 0.9954 for milk and 0.967 for the 50 % concentrate. Integrating from the water content of the milk mw,0m_{w,0} to that left in the concentrate mw,fm_{w,f} (both per kg of milk):

wsep=1mw,0−mw,f∫mw,fmw,0RT0Mw ln⁡ ⁣(1+amw)dmw=RT0Mw G(mw,0)−G(mw,f)mw,0−mw,f,w_{sep} = \frac{1}{m_{w,0} - m_{w,f}}\int_{m_{w,f}}^{m_{w,0}} \frac{R T_0}{M_w}\,\ln\!\left(1 + \frac{a}{m_w}\right) dm_w = \frac{R T_0}{M_w}\,\frac{G(m_{w,0}) - G(m_{w,f})}{m_{w,0} - m_{w,f}}, G(x)=x ln⁡ ⁣(1+ax)+a ln⁡(x+a),G(x) = x\,\ln\!\left(1 + \frac{a}{x}\right) + a\,\ln(x + a),

a closed form checked against numerical quadrature in the tests. The result is 1.46 kJ per kg of water: milk is a dilute solution, so separating its water is almost free. A real concentrate is somewhat less ideal (its activity is lower), which would raise the work slightly; the ideal value is therefore a lower bound, as a theoretical ceiling must be. The DOE bandwidth study sets the thermodynamic minimum of this step to zero because it counts only chemical conversions; the same study computes −RTln⁡aw-RT\ln a_w for sugar concentration, which is the method used here.

2.2 Boiling-point elevation

The concentrate boils at the temperature TsolT_{sol} at which the vapour pressure over the solution, aw psat(Tsol)a_w\,p_{sat}(T_{sol}), equals the pressure of the vapour space, psat(Tv)p_{sat}(T_v), where TvT_v is the saturation temperature of the vapour (set by the vacuum). The boiling-point elevation is

BPE=Tsol−Tv,psat(Tsol)=psat(Tv)aw,BPE = T_{sol} - T_v, \qquad p_{sat}(T_{sol}) = \frac{p_{sat}(T_v)}{a_w},

computed exactly with the IAPWS-95 saturation curve (CoolProp): 0.73 K for the final concentrate at TvT_v = 60 °C. Through Clausius–Clapeyron, BPE≈RT2ln⁡(1/aw)/(Mwhfg)BPE \approx R T^2 \ln(1/a_w)/(M_w h_{fg}), which links it directly to the separation work (checked in the tests).

2.3 Steam-driven evaporation: single effect and N effects

A single effect with the feed at its boiling point needs the latent heat hfgh_{fg} of the water boiled off, 2 356 kJ per kg at 60.7 °C. Steam condensing at the higher temperature Ts=Tv+BPE+ΔTT_s = T_v + BPE + \Delta T, where ΔT\Delta T is the temperature difference across the film, releases a slightly smaller latent heat, so the ideal steam consumption is

m˙sm˙w=hfg(Tv+BPE)N hfg(Tv+N(BPE+ΔT))\frac{\dot m_s}{\dot m_w} = \frac{h_{fg}(T_v + BPE)}{N\,h_{fg}(T_v + N(BPE + \Delta T))}

for NN effects in series, each reusing the vapour of the previous one as heating medium (ideal economy NN): 1.003, 0.504 and 0.337 kg of steam per kg of water for one, two and three effects. The documented requirements are 1.1–1.4, 0.6–0.7 and 0.4; the gap is the sensible heating of feed below its boiling point, heat losses, venting and condensate sub-cooling. With the largest number of effects reported (seven), the ideal heat is hfg/Nh_{fg}/N = 337 kJ per kg of water.

Thermal vapour recompression. In a thermocompressor, motive steam at pressure pmp_m expands and entrains vapour from the evaporator, compressing it to the pressure of the heating side. In the ideal (isentropic) limit, the work released by expanding the motive steam from pmp_m to the heating pressure, hm−h(pc,sm)h_m - h(p_c, s_m), equals the work of compressing the entrained vapour from psat(Tv)p_{sat}(T_v) to pc=psat(Tv+BPE+ΔT)p_c = p_{sat}(T_v + BPE + \Delta T). The entrainment ratio is the ratio of these two works, 20.7 kg of vapour per kg of motive steam at 8 bar, and the ideal motive-steam consumption of a single TVR effect is 0.046 kg per kg of water. Real thermocompressors entrain far less vapour, because the mixing of a supersonic jet with slow vapour is highly irreversible; this is why TVR evaporators, even with several stages, stay well above MVR.

2.4 Mechanical vapour recompression: Carnot-type bound and isentropic work

MVR uses the evaporator as a heat pump: vapour produced at the saturation temperature TvT_v is compressed until it condenses at Tc=Tsol+ΔTT_c = T_{sol} + \Delta T on the other side of the tube wall, where it delivers the heat that boils the next kilogram of water. The minimum work of a heat pump that takes heat at TvT_v and delivers hfgh_{fg} at TcT_c is the Carnot work

wCarnot=hfg Tc−TvTc=hfg BPE+ΔTTc,w_{Carnot} = h_{fg}\,\frac{T_c - T_v}{T_c} = h_{fg}\,\frac{BPE + \Delta T}{T_c},

averaged here over the concentration path (the BPE grows as the solids increase). With ΔT→0\Delta T \to 0 the Carnot work tends to the separation work of Section 2.1, because the BPE is the water activity seen through Clausius–Clapeyron (checked in the tests): the only irreducible cost is the separation, and everything else is the price of finite heat-transfer area. For ΔT\Delta T = 3 K the bound is 22.6 kJ per kg. The isentropic compression work of the vapour between the two saturation pressures, ws=cpT1[(pc/pv)(γ−1)/γ−1]w_s = c_p T_1[(p_c/p_v)^{(\gamma-1)/\gamma} - 1] with the IAPWS-95 heat capacities cpc_p and γ=cp/cv\gamma = c_p/c_v of the vapour (within 1.5 % of the full real-gas calculation of CoolProp), is 23.2 kJ per kg; it exceeds the Carnot bound because compression from saturated vapour produces superheat that is degraded on condensation.

If the reported MVR consumption (10 kWh per tonne, 36 kJ per kg) were ideal Carnot work, the lift would be T w/hfgT\,w/h_{fg} = 5.1 K; the declared film temperature differences (2–4 K) plus the BPE stay below it, so the scenario is consistent with the documented value.

2.5 The energy ceilings

Final energy per kilogram of water removed, as in the comparison table of the BAT reference document: theoretical (separation work) 1.46; best available technique (MVR evaporator, "approximately 10 kWh/t") 36; documented (single-stage MVR, 0.015 kWh/kg) 54; typical (TVR with three to five stages, the most common configurations, 0.084–0.140 kWh/kg) 396 kJ per kg. The DOE intensities for concentration of evaporated milk and powdered milk (172 Btu/lb current typical, 138 state of the art) are expressed per pound of a mix of products, not per kilogram of water, and are reported with compare: false. The statement in the BAT reference document that about 600 kWh are required to evaporate one tonne of water (2 160 kJ per kg) is of the same magnitude as, and about 8 % below, the ideal single-effect heat of Section 2.3.

3. Rate-limiting step — evaporation rate and run length

Film heat transfer. The rate at which a square metre of heating surface evaporates water is limited by the heat flux across the condensing film, the tube wall and the boiling milk film:

m˙clean′′=U0 ΔThfg,\dot m''_{clean} = \frac{U_0\,\Delta T}{h_{fg}},

where U0U_0 is the overall heat-transfer coefficient of the clean surface. With U0U_0 = 2 000 W/(m² K) and ΔT\Delta T = 3 K this is 9.2 kg per m² per hour. The small ΔT\Delta T that saves compressor work (Section 2.4) is paid with heating surface: area and energy trade against each other.

Fouling-limited run length. Milk proteins denature on the hot surface and minerals precipitate, adding a fouling resistance that grows during the run. With a linear growth rate R˙f\dot R_f,

U(t)=11/U0+R˙f t,tfoul=1/fmin−1U0 R˙f,U(t) = \frac{1}{1/U_0 + \dot R_f\,t}, \qquad t_{foul} = \frac{1/f_{min} - 1}{U_0\,\dot R_f},

where tfoult_{foul} is the time until the duty falls to the declared fraction fminf_{min} of the clean value, a common stopping rule (21.4 h at the point values; it defines the typical availability of Section 5). At constant ΔT\Delta T the mean rate over a run of length trunt_{run}, relative to the clean rate, is

Pmax=1trun∫0trunU(t)U0 dt=ln⁡(1+β)β,β=U0 R˙f trun,P_{max} = \frac{1}{t_{run}}\int_0^{t_{run}} \frac{U(t)}{U_0}\,dt = \frac{\ln(1 + \beta)}{\beta}, \qquad \beta = U_0\,\dot R_f\,t_{run},

the unavoidable performance loss of a fouling surface. A plant that raises ΔT\Delta T during the run keeps the rate instead and pays in energy.

Optimal run length. A longer run raises the availability (fewer cleanings) and spreads the start-up and shut-down losses over more product, but lowers the mean rate. The bound uses the run length that maximises

A P Q=truntrun+tCIP ln⁡(1+β)β (1−d)(1−2κτtrun),A\,P\,Q = \frac{t_{run}}{t_{run} + t_{CIP}}\,\frac{\ln(1+\beta)}{\beta}\,(1-d)\left(1 - \frac{2\kappa\tau}{t_{run}}\right),

found on a fine geometric grid for every Monte Carlo draw and capped by the 44 h legal limit of a continuous evaporator run. At the point values the optimum is trunt_{run} = 7.4 h, with PmaxP_{max} = 0.933; the legal limit binds only when fouling is slow.

4. Process capability — quality ceiling

The dryer accepts concentrate within a tolerance band ±δ\pm\delta around the solids target. With the solids under control distributed as x∼N(xC,σx2)x \sim \mathcal N(x_C, \sigma_x^2), the share outside the band, which is recirculated, is

d=2 Φ ⁣(−δσx),d = 2\,\Phi\!\left(-\frac{\delta}{\sigma_x}\right),

where Φ\Phi is the standard normal distribution function. The OEE quality factor includes the start-up and shut-down losses of Section 1: Qmax=(1−d) YmaxQ_{max} = (1 - d)\,Y_{max} = 0.983 at the optimal run length.

5. Unavoidable stops — maximum availability

Each run ends with cleaning-in-place. Its minimum duration tCIPt_{CIP} is taken through a parameter link from the cleaning model (output cip_time_min: the physical minimum of the kinetic and displacement phases plus declared hold times). That model is valid for recirculated pipe and plate circuits; the falling-film tubes are pipes, but the distribution of the liquid over the tube plate and the vapour side of the calandria are outside its validity, so the linked duration is a lower bound for the evaporator. With the optimal run length,

Amax=truntrun+tCIP=0.952.A_{max} = \frac{t_{run}}{t_{run} + t_{CIP}} = 0.952 .

The typical availability uses the declared duty floor and a typical cleaning programme of 1.5–3 h: Atyp=tfoul/(tfoul+tCIP,typ)A_{typ} = t_{foul}/(t_{foul} + t_{CIP,typ}) = 0.915.

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=0.873OEE_{max} = A_{max} \times P_{max} \times Q_{max} = 0.873

at the point values; the Monte Carlo interval over all declared uncertainties is in the output table. Unlike the pasteurizer, the evaporator has a physical performance loss: fouling lowers the rate continuously between cleanings.

7. Assumptions and validity

See the assumption and validity lists generated from the configuration file. The most important are: ideal solution of the dissolved solutes; energy compared as final energy per kilogram of water; linear fouling at constant temperature difference. The most influential inputs (sensitivity table) are the film temperature difference for the MVR bounds, the fouling rate and clean heat-transfer coefficient for run length, performance and availability, and the concentrate solids for the separation work.

References

  • 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 Savings Opportunities in U.S. Food and Beverage Manufacturing, 2017.
  • U.S. Food and Drug Administration. Grade "A" Pasteurized Milk Ordinance, 2023 Revision.
  • U.S. Department of Agriculture, Agricultural Research Service. USDA National Nutrient Database for Standard Reference, Release 28, food group 01 report, 2015.
  • Wagner, W., Pruß, A. (2002). The IAPWS formulation 1995 for the thermodynamic properties of ordinary water substance for general and scientific use. Journal of Physical and Chemical Reference Data, 31, 387–535. doi:10.1063/1.1461829
  • 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.
  • xw,Fx_{w,F}0.8813

    Water content of whole milk (mass fraction)

    Property · Fixed

    Verified quote
    Show verbatim quote
    88.13 61 254 3.15 3.25 0.67 4.80USDA National Nutrient Database for Standard Reference, Release 28 (2015): Food Group 01, Dairy and Egg Products (nutrient report) (opens in a new tab). U.S. Department of Agriculture, Agricultural Research Service, Nutrient Data Laboratory, 2015. p. 632. USDA SR28, food group 01 report (sr28fg01.pdf)
  • xCx_C0.5

    Total solids of the concentrate leaving the evaporator (mass fraction)

    Technology · Uniform 0.5 to 0.6

    Verified quote
    Show verbatim quote
    After the milk has been thickened from 11 % to 50–60 % dry matter in an evaporatorBest 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. 411. EUR 29978 EN; JRC118627; doi:10.2760/243911
  • xlacx_{lac}0.0505

    Lactose content of whole milk (mass fraction)

    Property · Fixed

    Verified quote
    Show verbatim quote
    4.80 0.0 5.05 0.00 0.00 0.00 5.05 0.00 0.00USDA National Nutrient Database for Standard Reference, Release 28 (2015): Food Group 01, Dairy and Egg Products (nutrient report) (opens in a new tab). U.S. Department of Agriculture, Agricultural Research Service, Nutrient Data Laboratory, 2015. p. 632. USDA SR28, food group 01 report (sr28fg01.pdf)
  • xKx_K132 mg/hg

    Potassium content of whole milk

    Property · Fixed

    Verified quote
    Show verbatim quote
    113 0.03 10 84 132 43 0.37USDA National Nutrient Database for Standard Reference, Release 28 (2015): Food Group 01, Dairy and Egg Products (nutrient report) (opens in a new tab). U.S. Department of Agriculture, Agricultural Research Service, Nutrient Data Laboratory, 2015. p. 632. USDA SR28, food group 01 report (sr28fg01.pdf)
  • xNax_{Na}43 mg/hg

    Sodium content of whole milk

    Property · Fixed

    Verified quote
    Show verbatim quote
    113 0.03 10 84 132 43 0.37USDA National Nutrient Database for Standard Reference, Release 28 (2015): Food Group 01, Dairy and Egg Products (nutrient report) (opens in a new tab). U.S. Department of Agriculture, Agricultural Research Service, Nutrient Data Laboratory, 2015. p. 632. USDA SR28, food group 01 report (sr28fg01.pdf)
  • MlacM_{lac}342.3 g/mol

    Molar mass of lactose (C12H22O11)

    Property · Fixed

    Computed

    12 x 12.011 + 22 x 1.008 + 11 x 15.999 = 342.30 g/mol (IUPAC conventional atomic weights).

  • MKM_K39.098 g/mol

    Molar mass of potassium

    Property · Fixed

    Computed

    IUPAC standard atomic weight of potassium.

  • MNaM_{Na}22.99 g/mol

    Molar mass of sodium

    Property · Fixed

    Computed

    IUPAC standard atomic weight of sodium.

  • faf_a0.5

    Osmotically active anions per soluble monovalent cation (0: cations only; 1: one monovalent counter-ion each)

    Scenario · Uniform 0 to 1

    Declared assumption

    Declared assumption bracketing the contribution of the counter-ions (chloride, citrate, phosphate) to the ideal-solution water activity; no open source for the osmolality of milk serum was verified. Calcium, magnesium and phosphate bound to the casein micelles are not counted. Replace with a cryoscopic measurement of the plant's milk (osmolality = freezing-point depression / 1.86 K kg/mol).

  • T0T_025 °C

    Dead-state (ambient) temperature for the reversible separation work

    Scenario · Uniform 15 to 35 °C

    Declared assumption
  • TvT_v60 °C

    Saturation temperature of the vapour in the evaporating space (vacuum level)

    Scenario · Uniform 50 to 70 °C

    Declared assumption

    Declared operating assumption for evaporation of milk under vacuum; no open source verified.

  • ΔT\Delta T3 K

    Temperature difference across the falling film (condensing vapour to boiling product)

    Scenario · Uniform 2 to 4 K

    Declared assumption

    Declared design assumption. The range is kept consistent with the reported MVR consumption: an ideal compressor with a larger lift would already need more work than the documented value (see the output 'implied_lift_bat').

  • NN7

    Largest number of effects operated in series

    Technology · Fixed

    Verified quote
    Show verbatim quote
    Up to seven stages can be operated in series, but three to five is more commonBest 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. 135. EUR 29978 EN; JRC118627; doi:10.2760/243911
  • pmp_m8 bar

    Pressure of the motive steam of the thermocompressor (saturated)

    Scenario · Uniform 6 to 12 bar

    Declared assumption

    Declared assumption (absolute pressure of the plant steam header).

  • eMVRe_{MVR}10 kWh/t

    Energy consumption of an MVR evaporator

    Benchmark · Fixed

    Verified quote
    Show verbatim quote
    It is reported that the energy consumption of a MVR evaporator is approximately 10 kWh/t of water evaporated, with negligible steam consumptionBest 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. 137. EUR 29978 EN; JRC118627; doi:10.2760/243911
  • eMVR,1e_{MVR,1}0.015 kWh/kg

    Total energy consumption of a single-stage MVR evaporator (comparison table)

    Benchmark · Fixed

    Verified quote
    Show verbatim quote
    MVR single-stage 0.015Best 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. 136. EUR 29978 EN; JRC118627; doi:10.2760/243911
  • eTVRe_{TVR}0.11 kWh/kg

    Total energy consumption of TVR evaporators with 3 to 5 stages (the most common configurations)

    Benchmark · Uniform 0.084 to 0.14 kWh/kg

    Verified quote
    Show verbatim quote
    TVR 4 stages 0.110Best 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. 136. EUR 29978 EN; JRC118627; doi:10.2760/243911
  • U0U_02,000 W/m²/K

    Overall heat-transfer coefficient of the clean falling film

    Scenario · Uniform 1,000 to 3,000 W/m²/K

    Declared assumption

    Declared assumption; no open source verified for milk falling-film evaporators.

  • R˙f\dot R_f1.00 × 10⁻⁵ m²·K/W/h

    Growth rate of the fouling resistance on the heating surface

    Scenario · Lognormal, σ 0.7

    Declared assumption

    Declared assumption (linear fouling); no open source verified. Milk fouling depends on product, pre-heat treatment and temperature; replace with the plant's measured decline of the overall coefficient.

  • fminf_{min}0.7

    Fraction of the clean duty at which a run is stopped for cleaning

    Scenario · Uniform 0.6 to 0.8

    Declared assumption
  • tlegalt_{legal}44 h

    Legal maximum length of a continuous evaporator run before cleaning

    Regulatory · Fixed

    Verified quote
    Show verbatim quote
    evaporators shall be cleaned at the end of a continuous operation, not to exceed forty-four (44) hoursGrade "A" Pasteurized Milk Ordinance, 2023 Revision (opens in a new tab). U.S. Department of Health and Human Services, Public Health Service, Food and Drug Administration, 2023. p. 99. PMO 2023 Revision
  • tCIPt_{CIP}0.374294 h

    Minimum duration of the cleaning-in-place of the evaporator: physical minimum of the kinetic and displacement phases plus declared hold times (output cip_time_min of the cleaning model)

    Property · Triangular 0.242184, 0.374294, 0.536161 h

    Computed

    Output `cip_time_min` of the cleaning model: 22.46 min [14.53, 32.17].

  • tCIP,typt_{CIP,typ}2 h

    Duration of a typical cleaning-in-place programme of the evaporator

    Scenario · Uniform 1.5 to 3 h

    Declared assumption

    Declared assumption: rinse, alkaline wash, rinse, acid wash, rinse with operating margins, used for the typical availability. Replace with the validated programme of the plant.

  • τ\tau5 min

    Mean residence time of the product in the evaporator (hold-up volume over flow)

    Scenario · Uniform 2 to 10 min

    Declared assumption
  • κ\kappa0.75

    Share of the hold-up volume lost as a water-product mixed zone at each start-up and shut-down

    Scenario · Uniform 0.5 to 1

    Declared assumption
  • σx\sigma_x0.005

    Standard deviation of the concentrate solids under control (mass fraction)

    Scenario · Uniform 0.003 to 0.01

    Declared assumption

    Declared process-capability assumption; replace with the plant's measured value (density or refractometer).

  • δ\delta0.02

    Half-width of the solids tolerance band accepted by the dryer

    Scenario · Fixed

    Declared assumption

8Benchmarks

Documented values from open sources used to check the ceilings; no documented value may beat the theoretical limit.
BenchmarkCeilingValueProvenance
Steam requirement of single-stage evaporators, lower end (food and drink sector)
Show verbatim quote
Steam requirements for single-stage evaporators range from 1.1 tonnes to 1.2 tonnes of steam per tonne of evaporated waterBest 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. 787. EUR 29978 EN; JRC118627; doi:10.2760/243911
Best documented1.1 Verified quote
Steam requirement of single-stage evaporators, lower end (multistage evaporation technique)
Show verbatim quote
Steam requirements for single-stage evaporators are 1.2–1.4 t/t of evaporated waterBest 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. 135. EUR 29978 EN; JRC118627; doi:10.2760/243911
Typical1.2 Verified quote
Steam requirement of double-effect evaporators, lower end
Show verbatim quote
In the case of double or third effect, the steam requirement lowers respectively to 0.6–0.7 tonnes and 0.4 tonnes of steam per tonne of evaporated waterBest 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. 787. EUR 29978 EN; JRC118627; doi:10.2760/243911
Best documented0.6 Verified quote
Steam requirement of triple-effect evaporators
Show verbatim quote
In the case of double or third effect, the steam requirement lowers respectively to 0.6–0.7 tonnes and 0.4 tonnes of steam per tonne of evaporated waterBest 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. 787. EUR 29978 EN; JRC118627; doi:10.2760/243911
Best documented0.4 Verified quote
Total energy of a 7-stage TVR evaporator (dairy industry comparison table)
Show verbatim quote
TVR 7 stages 0.060Best 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. 136. EUR 29978 EN; JRC118627; doi:10.2760/243911
Best documented0.06 kWh/kgVerified quote
Total energy of a 3-stage TVR evaporator (dairy industry comparison table)
Show verbatim quote
TVR 3 stages 0.140Best 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. 136. EUR 29978 EN; JRC118627; doi:10.2760/243911
Typical0.14 kWh/kgVerified quote
Energy stated as required to evaporate one tonne of water (spray-drying section; compare with the ideal single-effect heat)
Show verbatim quote
Taking into account that about 600 kWh energy is required to evaporate 1 tonne of waterBest 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. 412. EUR 29978 EN; JRC118627; doi:10.2760/243911
—600 kWh/tVerified quote
DOE bandwidth study, concentration of canned evaporated milk and powdered dry milk, current typical (per lb of product, a mix of evaporated milk and powder: not comparable with a value per kg of water removed)
Show verbatim quote
Concentration 2,488 172 138 138 0Bandwidth 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. 57. DOE/EE-1571; OSTI ID 1512366
Typical172 Btu/lbVerified quote
DOE bandwidth study, concentration, state of the art (same basis as the current typical value)
Show verbatim quote
Concentration 2,488 172 138 138 0Bandwidth 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. 57. DOE/EE-1571; OSTI ID 1512366
Best available technique138 Btu/lbVerified quote
DOE bandwidth study, concentration, thermodynamic minimum (set to zero for lack of chemical conversion; the separation work is not counted)
Show verbatim quote
Concentration 2,488 172 138 138 0Bandwidth 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. 57. DOE/EE-1571; OSTI ID 1512366
Theoretical0 Btu/lbVerified quote

9Unavoidable stops

Stops that physics or regulation impose; they set the maximum availability.
  • Cleaning at the end of each continuous run (at most every 44 h)Verified quote

    Milk proteins and minerals foul the heating surface, so the duty falls during a run; hygiene regulation also caps a continuous evaporator run at 44 hours. For the bound the run ends at the length that maximises A x P x Q (at most 44 h), followed by cleaning-in-place of minimum duration (cleaning model).

    Show verbatim quote
    evaporators shall be cleaned at the end of a continuous operation, not to exceed forty-four (44) hoursGrade "A" Pasteurized Milk Ordinance, 2023 Revision (opens in a new tab). U.S. Department of Health and Human Services, Public Health Service, Food and Drug Administration, 2023. p. 99. PMO 2023 Revision

10Assumptions

  1. The minimum cleaning duration comes from the cleaning model, whose validity covers recirculated pipe and plate circuits; the falling-film tubes are pipes, but the distribution of the cleaning liquid over the tube plate and the vapour side of the calandria are outside that validity, so the linked duration is a lower bound for the evaporator. The bound uses the run length that maximises A x P x Q (fouling decline, cleaning and start-up losses together), capped by the legal 44 h; the declared duty floor and the typical cleaning duration give the typical availability.
  2. The dissolved solutes (lactose and the soluble salts) form an ideal solution: a_w = n_w / (n_w + n_s) (Raoult). Fat, casein micelles and colloidal salts do not lower the water activity. The real activity of a concentrated milk is somewhat lower, so the separation work is a lower bound, as a theoretical ceiling must be.
  3. The boiling-point elevation is obtained exactly from the water activity: the solution boils where a_w p_sat(T) equals the vapour-space pressure (IAPWS-95 saturation pressure, CoolProp).
  4. Energy is compared as final energy per kg of water removed, as in the BAT reference document (Table 2.41): heat for steam-driven evaporators, electricity for MVR. Primary-energy factors are outside the boundary.
  5. Fouling adds a resistance that grows linearly in time; the temperature difference is held constant, so the duty falls as U(t) = 1/(1/U0 + R_f t). A plant that raises the temperature difference instead keeps the rate and pays in energy.
  6. Start-up and shut-down losses are the water-product mixed zones, a declared share of the hold-up volume at each transition; concentrate outside the solids band is recirculated and counted as a quality loss.

11Range of validity

  • Whole milk concentrated to 50–60 % solids under vacuum (vapour saturation 50–70 °C) in falling-film evaporators.
  • For fruit juices replace the composition parameters (sugars dominate the water activity) and the regulatory run length.

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.
  • U0U_0Overall heat-transfer coefficient of the clean falling film
    elasticity -0.066Spearman ρ -0.36
  • R˙f\dot R_fGrowth rate of the fouling resistance on the heating surface
    elasticity -0.066Spearman ρ -0.85
  • tCIPt_{CIP}Minimum duration of the cleaning-in-place of the evaporator: physical minimum of the kinetic and displacement phases plus declared hold times (output cip_time_min of the cleaning model)
    elasticity -0.048Spearman ρ -0.13
  • τ\tauMean residence time of the product in the evaporator (hold-up volume over flow)
    elasticity -0.017Spearman ρ -0.13
  • κ\kappaShare of the hold-up volume lost as a water-product mixed zone at each start-up and shut-down
    elasticity -0.017Spearman ρ -0.07
  • δ\deltaHalf-width of the solids tolerance band accepted by the dryer
    elasticity +0.001Spearman ρ —
  • σx\sigma_xStandard deviation of the concentrate solids under control (mass fraction)
    elasticity -0.001Spearman ρ -0.21
  • tlegalt_{legal}Legal maximum length of a continuous evaporator run before cleaning
    elasticity 0.000Spearman ρ —
  • Output rises with the parameter
  • Output falls with the parameter

13Validation checks

A model is published only when every check passes.
  • Dimensional analysis30 outputs converted to their declared unitsPassed
  • Every cited value verified against its source21 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.

15References

  1. 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. 411. EUR 29978 EN; JRC118627; doi:10.2760/243911
  2. 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. 57. DOE/EE-1571; OSTI ID 1512366
  3. Grade "A" Pasteurized Milk Ordinance, 2023 Revision (opens in a new tab). U.S. Department of Health and Human Services, Public Health Service, Food and Drug Administration, 2023. p. 99. PMO 2023 Revision
  4. USDA National Nutrient Database for Standard Reference, Release 28 (2015): Food Group 01, Dairy and Egg Products (nutrient report) (opens in a new tab). U.S. Department of Agriculture, Agricultural Research Service, Nutrient Data Laboratory, 2015. p. 632. USDA SR28, food group 01 report (sr28fg01.pdf)
  5. doi:10.1021/ie4033999 — CoolProp: IAPWS-95 water properties (saturation, enthalpy, entropy).
  6. doi:10.1063/1.1461829 — IAPWS-95 formulation for water.

Processes that use this operation