foratvise
Get Started
Filling and packaging line

Conveying and accumulation

Conveying and accumulation of bottles between the filler and the labeller

Validated modelOEE bound 1.00000 [0.99985–1.00000]20,000 Monte Carlo samples · seed 20261115 · computed November 16, 2026

In short

Conveying and accumulation can use at most 100 of every 100 planned hours. The rest goes to limits no plant can avoid.

  1. 100usable at most

The three limits that set it

Running time: up to 100 %

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

In a carbonated soft drink bottling line, filled and capped 0.5 L PET bottles leave the filler on lubricated plastic-chain conveyors, cross an accumulation table and are fed in single file to the labeller. The filler is the slowest machine of the line and the labeller runs faster (the V-graph of a bottling line). The accumulation table is the buffer: while the labeller is down, the filler keeps running into it; when the labeller restarts, its speed surplus empties it. Machine rates and the measured failure behaviour of the labeller are those of an industrial bottling plant; the conveyor geometry, the bottle and the conveyor speed are declared scenario values. The model bounds what physics allows the conveyor to do (stable transport, buffer capacity) and quantifies how much of the labeller's stopping time reaches the filler through a buffer of given size.

Provenance of inputsDeclared assumption8Verified quote7

2Results

Point value from the nominal parameters and 95% interval from 20,000 Monte Carlo draws of the uncertain parameters.
  • Largest acceleration that base friction can give a free-standing container (mu g)1.27 [0.75–1.81] m/s²
  • Inertial acceleration at which a container would tip (g D / 2h)3.54 [3.13–4.03] m/s²
  • Tipping limit over friction coefficient, D / (2 h mu): above 1 the container slides before it tips2.78 [1.91–4.86]
  • Share of containers that tip at the declared number of speed transfers (local friction above the tipping limit)2.66 × 10⁻¹⁴ [NaN × 10⁻Infinity–1.48 × 10⁻⁴]
  • TheoreticalMaximum OEE quality factor (containers that stay upright)1.00000 [0.99985–1.00000]
  • Conveyor speed needed to feed the labeller in single file with touching containers0.599 [0.573–0.626] m/s
  • Shortest time in which friction brings a container from rest to the single-file speed0.470 [0.330–0.803] s
  • Containers held by the accumulation table at densest (hexagonal) packing1,640 [771–3,813]
  • Time the full buffer keeps the filler running while the labeller is down204 [96–473] s
  • Time the labeller needs to empty a full buffer with its speed surplus (V-graph)1,406 [661–3,269] s
  • Mean number of containers in the buffer80.0 [53.6–86.4]
  • Share of time the filler is blocked by labeller stops that the buffer cannot absorb3.66 × 10⁻⁴ [1.59 × 10⁻⁶–3.30 × 10⁻³]
  • Blocking with the buffer over blocking without any buffer0.0134 [0.0001–0.1206]
  • TheoreticalThermodynamic minimum energy of horizontal transport0 kJ/kg
  • TypicalFriction dissipated by the belt sliding under accumulated containers, per kg of product (declared buffer and speed)0.0127 [0.0069–0.0242] kJ/kg
  • TheoreticalMaximum yield (conveying consumes no product)1.00
  • TheoreticalMaximum availability (no conveyor stop is forced by physics or regulation)1.00
  • Filler availability against labeller stops without a buffer (documented labeller MTBF and MTTR)0.973
  • Filler availability against labeller stops propagated through the declared buffer0.9996 [0.9967–1.0000]
  • TheoreticalPerformance factor (the conveyor carries the line rate with a speed surplus)1.00
  • TheoreticalPhysics-bounded OEE1.00000 [0.99985–1.00000]
  • OEE with the documented labeller stops propagated through the declared buffer0.9996 [0.9966–1.0000]

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
Qualityhigher is better
  • TheoreticalMaximum OEE quality factor (containers that stay upright)1.00000 [0.99985–1.00000]
Energy per unit of productkJ/kg · lower is better
  • TheoreticalThermodynamic minimum energy of horizontal transport0 kJ/kg
  • TypicalFriction dissipated by the belt sliding under accumulated containers, per kg of product (declared buffer and speed)0.0127 [0.0069–0.0242] kJ/kg
Yieldhigher is better
  • TheoreticalMaximum yield (conveying consumes no product)1.00
Availabilityhigher is better
  • TheoreticalMaximum availability (no conveyor stop is forced by physics or regulation)1.00
Performancehigher is better
  • TheoreticalPerformance factor (the conveyor carries the line rate with a speed surplus)1.00
OEEhigher is better
  • TheoreticalPhysics-bounded OEE1.00000 [0.99985–1.00000]

4Loss cascade

What separates each ceiling from the next one, from the physical limit to typical operation.
  1. TheoreticalTypicalEnergy per unit of product
    0 kJ/kg → 0.0127 [0.0069–0.0242] kJ/kg (gap 0.0127 kJ/kg)

    Horizontal transport lifts nothing and changes nothing in the container, so its thermodynamic minimum is zero. Energy is dissipated wherever the belt slides under a bottle that is held back: on an accumulation table the chain keeps running under every stationary bottle and dissipates mu m g v per bottle, which with the mean buffer level of the scenario is still a small fraction of a joule per kilogram of product. The electricity of a real conveyor line is dominated by chain friction on the wear strips, gearmotor losses and the conveyors that run empty or under a full table for long periods; lubrication lowers the friction term but too little friction lets bottles tip and fall.

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 0.95.
  • Loading time (ideal)
    1.000
  • Availability
    factor 1.00
    1.000 (−0.000)
  • Performance
    factor 1.00
    1.000 (−0.000)
  • Quality
    factor 1.00000 [0.99985–1.00000]
    1.000 (−0.000)
  • Physics-bounded OEE
    1.00000 [0.99985–1.00000]

6Full derivation

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

Scope. In a carbonated soft drink bottling line, filled and capped 0.5 L PET bottles leave the filler on lubricated plastic-chain conveyors, cross an accumulation table and are fed in single file to the labeller. The filler is the slowest machine and the labeller runs faster (the V-graph of a bottling line). The accumulation table is the buffer that decouples the two machines. Machine rates and the measured failure behaviour of the labeller come from an industrial bottling plant (Bär et al., 2021, Table A1); conveyor geometry, bottle and conveyor speed are declared. Symbols and their values, units, uncertainty distributions and sources are listed in the parameter table generated from config/physics/conveying.yaml. Numbers quoted here are point values; the output table gives the Monte Carlo intervals.

1. Mass balance — maximum yield

Conveying neither creates nor consumes product: every bottle that enters leaves, so Ymax=1Y_{max} = 1. A bottle that tips and falls is a nonconforming unit (section 4), not a mass loss of conforming product.

2. Energy balance — minimum energy per kg

Horizontal transport lifts nothing and changes nothing in the bottle: its final state equals its initial state, so the thermodynamic minimum is zero. Energy is dissipated wherever the belt slides under a bottle that is held back. On an accumulation table the chain keeps running under every stationary bottle and dissipates the friction power μ m g vc\mu\,m\,g\,v_c per bottle, where μ\mu is the coefficient of friction, mm the mass of the full bottle, gg the gravity acceleration and vcv_c the belt speed. With the mean buffer level xˉ\bar x of section 5 and the filler output n˙F(1−PN)\dot n_F(1 - P_N), the dissipation per kg of product is

eacc=μ g vc xˉn˙F (1−PN)=0.013 kJ/kg.e_{acc} = \frac{\mu\,g\,v_c\,\bar x}{\dot n_F\,(1 - P_N)} = 0.013\ \text{kJ/kg}.

It is small because the buffer is nearly empty most of the time (80 bottles on average). The electricity of a real conveyor line is dominated by chain friction on the wear strips, gearmotor losses and conveyors that run empty or under a full table for long periods.

3. Rate-limiting step — single-file speed and stable acceleration

Single-file speed. To feed the labeller with touching bottles of diameter DD the conveyor must run at least at

vsf=n˙L D=0.60 m/s,v_{sf} = \dot n_L\,D = 0.60\ \text{m/s},

with n˙L\dot n_L = 33 197 bottles per hour, the measured set performance of the labeller. Conveyors run faster than this, so they do not set the cycle of the line: Pmax=1P_{max} = 1.

Stable acceleration. A free-standing bottle is accelerated only by the friction at its base, so its acceleration cannot exceed

aslide=μ g=1.27 m/s2a_{slide} = \mu\,g = 1.27\ \text{m/s}^2

with μ\mu = 0.13, the coefficient of friction measured for bottles on a conveyor track with a dry lubricant (US 9 562 209 B2). The shortest time to bring a bottle from rest to the single-file speed is tramp=vsf/(μg)t_{ramp} = v_{sf}/(\mu g) = 0.47 s: a belt that accelerates faster only slides under the bottles.

4. Process capability — quality ceiling (stability)

A bottle of base diameter DD whose centre of gravity is at height hh is a rigid body. When the belt accelerates under it, the friction force μmg\mu m g acts at the base and the inertial force at the centre of gravity; the bottle tips about its base edge when the overturning moment μmg h\mu m g\,h exceeds the restoring moment mg D/2m g\,D/2 of its weight, i.e. when

μ>D2h.\mu > \frac{D}{2h} .

The corresponding inertial acceleration is atip=gD/(2h)a_{tip} = g D/(2h) = 3.54 m/s². Because friction limits the acceleration to μg\mu g, a bottle slides before it can tip whenever μ<D/(2h)\mu < D/(2h); the stability margin is

D2h μ=2.78.\frac{D}{2h\,\mu} = 2.78 .

The lubricant documents keep friction at a proper value rather than the minimum, because for bottles an unacceptably low coefficient of friction "may result in an excessive number of tipped and fallen bottles" (US 7 745 381 B2). In this model low friction can never tip a free bottle, so that mechanism must come from contact forces the model leaves out (guide rails, curves, pushing neighbours). In the model, tipping can occur only where the local friction is high (a dry spot, a spill of sugar syrup). With the local friction scattered as N(μ,σμ2)\mathcal N(\mu, \sigma_\mu^2) and ntn_t speed transfers between filler and labeller,

Qmax=[1−Φ ⁣(μ−D/(2h)σμ)]nt=1−3×10−14,Q_{max} = \left[1 - \Phi\!\left(\frac{\mu - D/(2h)}{\sigma_\mu}\right)\right]^{n_t} = 1 - 3\times10^{-14},

with Φ\Phi the standard normal distribution function. For a full bottle on a lubricated track the stability ceiling does not limit quality; it would for empty PET bottles (high centre of gravity) or high-friction tracks, which are outside the validity of the model.

5. Buffer theory — what accumulation does for the line

Buffer capacity. The densest arrangement of equal circles in a plane is the hexagonal packing, which covers π/(23)\pi/(2\sqrt3) = 90.7 % of the area. A table of length LbL_b and width WbW_b therefore holds at most

N=π23 LbWbπD2/4=1 640 bottlesN = \frac{\pi}{2\sqrt3}\,\frac{L_b W_b}{\pi D^2/4} = 1\,640\ \text{bottles}

(LbL_b = 10 m, WbW_b = 0.6 m, DD = 65 mm), an upper bound: edge effects and the single-file lanes reduce it.

Two-stage line with a buffer. The filler (rate a=n˙Fa = \dot n_F = 28 998 per hour while not blocked) feeds the buffer; the labeller empties it at a+ba + b when it runs, where b=n˙L−n˙Fb = \dot n_L - \dot n_F is its speed surplus (the V-graph: 33 197 against 28 998 per hour). The labeller fails at rate p=1/MTBFp = 1/MTBF and is repaired at rate r=1/MTTRr = 1/MTTR, with exponential times as in the plant simulation of the source (measured MTBF = 1 423 s, MTTR = 40 s). The filler is the bottleneck and is treated as reliable here: its own stops belong to the filling model and only lower the buffer level, so the blocking computed below is an upper bound. Material is treated as a continuous flow of level x∈[0,N]x \in [0, N].

Let fU(x)f_U(x) and fD(x)f_D(x) be the stationary densities of the level while the labeller is up and down, P0P_0 the probability of an empty buffer with the labeller up (the labeller then runs at the filler rate) and PNP_N the probability of a full buffer with the labeller down (the filler is blocked). While the labeller is down the level rises at aa; while it is up it falls at bb. In the stationary state the probability flux across every level is zero,

a fD(x)=b fU(x),a\,f_D(x) = b\,f_U(x),

and the balance of the up state, −bfU′=−pfU+rfD-b f_U' = -p f_U + r f_D, then gives

fU(x)=K eλx,λ=pb−ra.f_U(x) = K\,e^{\lambda x}, \qquad \lambda = \frac{p}{b} - \frac{r}{a} .

The boundaries close the solution: the empty atom leaves at the failure rate, bfU(0)=pP0b f_U(0) = p P_0, and the full atom is fed by the rising level and left at the repair rate, afD(N)=rPNa f_D(N) = r P_N, so

K=pP0b,PN=pr eλNP0,P0[1+preλN+(1+ba)pb eλN−1λ]=1.K = \frac{p P_0}{b}, \qquad P_N = \frac{p}{r}\,e^{\lambda N} P_0, \qquad P_0\left[1 + \frac{p}{r}e^{\lambda N} + \left(1 + \frac{b}{a}\right)\frac{p}{b}\,\frac{e^{\lambda N} - 1}{\lambda}\right] = 1 .

PNP_N is the share of time the filler is blocked by labeller stops that outlast the buffer. Two limits check it: without a buffer (N→0N \to 0) PN=p/(p+r)P_N = p/(p + r), the down-time share of the labeller, i.e. every labeller stop reaches the filler; with λ<0\lambda < 0 (the labeller's surplus drains faster than its stops fill) PNP_N falls exponentially with NN. The mean level is xˉ=(1+b/a)K∫0Nxeλxdx+NPN\bar x = (1 + b/a)K\int_0^N x e^{\lambda x}dx + N P_N. The tests check PNP_N and xˉ\bar x against an independent event simulation of the same line.

Scenario values. With the documented rates and failure behaviour, λ=−2.50×10−3\lambda = -2.50\times10^{-3} per bottle and λN=−4.1\lambda N = -4.1. The buffer time (how long a full buffer keeps the filler running) is N/n˙FN/\dot n_F = 204 s, five times the mean repair time; the time the labeller needs to empty a full buffer with its surplus is N/bN/b = 1 406 s, about one mean time between failures. Then

PN=3.7×10−4,Ascen=1−PN=0.99963,P_N = 3.7\times10^{-4}, \qquad A_{scen} = 1 - P_N = 0.99963,

against Aunbuf=MTBF/(MTBF+MTTR)A_{unbuf} = MTBF/(MTBF + MTTR) = 0.9727 without any buffer: the table absorbs 98.7 % of the labeller's stopping time. The buffer theory states the role of conveying in a line: conveyors are interlinked so that a buffer absorbs the downtime of individual machines, and only when the buffer time is exceeded does the stop propagate (Bär et al., 2021).

6. Unavoidable stops and physics-bounded OEE

No conveyor stop is required by regulation or forced by a physical mechanism, and no cleaning-in-place applies to conveyors, so the cleaning model is not used:

Amax=1,Pmax=1,OEEmax=Amax×Pmax×Qmax=1−3×10−14.A_{max} = 1, \qquad P_{max} = 1, \qquad OEE_{max} = A_{max}\times P_{max}\times Q_{max} = 1 - 3\times10^{-14} .

Scope of the bound. The theoretical label of OEEmaxOEE_{max} means physics-bounded under the declared scenario. Conveying does not bound the OEE of the line by physics: the conveyor carries the line rate with a speed surplus, and a full bottle on a lubricated track slides before it can tip. What conveying contributes is the buffer. The propagated labeller stops are the labeller's losses seen at the filler, so they enter only availability_scenario (0.99963) and oee_scenario (0.99963).

No double counting. The labeller's own stops are reported by the labelling model (availability_documented); this model reports the share of them that reaches the filler through a buffer of given size. A line calculation must use either the labeller's own downtime (the line without buffers) or the propagated share computed here, never both. In the physics bound, the labelling model forces no stop that a buffer could absorb (its AmaxA_{max} is 1), so the physics-bounded availability of conveying is 1 and the buffer matters only for the declared and documented stops.

7. Assumptions and validity

See the assumption and validity lists generated from the configuration file. The most influential inputs (sensitivity table) are the buffer length and width, the bottle diameter and the documented failure behaviour for the scenario availability, and the friction coefficient and its local scatter for stability.

References

  • Bär, R. M., Zeilmann, M., Nophut, C., Kleinert, J., Beyer, K., Voigt, T. (2021). Simulation of energy and media demand of beverage bottling plants by automatic model generation. Sustainability, 13, 10089. doi:10.3390/su131810089
  • United States Patent 9 562 209 B2 (2017). Dry lubricant for conveying containers (Ecolab USA Inc.).
  • United States Patent 7 745 381 B2 (2010). Lubricant for conveying containers (Ecolab Inc.).

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
Availability of the labelling machine itself (97.3 %, measured); it equals the filler availability against labeller stops only when there is no buffer (availability_unbuffered)
Show verbatim quote
Labelling machine 33,197 30 1423 40 97.3Simulation of Energy and Media Demand of Beverage Bottling Plants by Automatic Model Generation (Bär, R. M.; Zeilmann, M.; Nophut, C.; Kleinert, J.; Beyer, K.; Voigt, T.), Sustainability 13, 10089 (2021), CC BY 4.0 (opens in a new tab). MDPI (copy distributed by a co-author's institution), 2021. p. 18. doi:10.3390/su131810089
Best documented97.3 %Verified quote

9Unavoidable stops

Stops that physics or regulation impose; they set the maximum availability.

10Assumptions

  1. Bottles are rigid bodies standing on a flat track; the only horizontal force on a free bottle is base friction (Coulomb, coefficient mu). A bottle tips only if the local friction exceeds D / (2h); bottles pushed by neighbours or guide rails are outside the model.
  2. Buffer capacity is the densest (hexagonal) packing of equal circles on the table area, an upper bound; edge effects and the single-file lanes are neglected.
  3. Buffer theory: continuous flow; the labeller fails and is repaired with exponential times (as in the source plant simulation); the filler is the bottleneck and is treated as reliable in this two-stage model: its own stops belong to the filling model and only lower the buffer level, so the blocking computed here is an upper bound.
  4. Availability: no conveyor stop is forced by physics or regulation, so availability_max = 1. The propagated stops of the labeller are the labeller's losses seen at the filler: in a line calculation use either the labeller's own downtime (no buffer) or the propagated share computed here, never both.
  5. No cleaning-in-place applies to conveyors, so the cleaning model is not used. Performance is 1: the conveyors run faster than the line rate and do not set the cycle.

11Range of validity

  • Free-standing rigid or filled containers on horizontal mass and single-file conveyors; empty PET bottles (high centre of gravity, light weight), inclines and air conveyors are outside the model.
  • One buffer between a reliable bottleneck and one faster downstream machine with exponential failures; several buffers in series need a decomposition method.

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.
  • DDBase (standing) diameter of the 0.5 L PET bottle
    elasticity 0.000Spearman ρ +0.09
  • hhHeight of the centre of gravity of the full bottle above its base
    elasticity 0.000Spearman ρ -0.22
  • σμ\sigma_\muLocal scatter of the friction coefficient along the track (dry spots
    elasticity 0.000Spearman ρ -0.87
  • μ\muCoefficient of friction between the bottles and the lubricated conveyor track
    elasticity 0.000Spearman ρ -0.24
  • Output rises with the parameter
  • Output falls with the parameter

13Validation checks

A model is published only when every check passes.
  • Dimensional analysis22 outputs converted to their declared unitsPassed
  • Every cited value verified against its source7 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. Simulation of Energy and Media Demand of Beverage Bottling Plants by Automatic Model Generation (Bär, R. M.; Zeilmann, M.; Nophut, C.; Kleinert, J.; Beyer, K.; Voigt, T.), Sustainability 13, 10089 (2021), CC BY 4.0 (opens in a new tab). MDPI (copy distributed by a co-author's institution), 2021. p. 18. doi:10.3390/su131810089
  2. US 9 562 209 B2, Dry lubricant for conveying containers (Ecolab USA Inc.), granted 2017-02-07 (opens in a new tab). United States Patent and Trademark Office (full text as published by Google Patents), 2017. US9562209B2
  3. us7745381b2_lubricant
  4. doi:10.3390/su131810089 — Bottling-plant simulation with measured set performances, MTBF and MTTR of every machine (Table A1).

Processes that use this operation