Plastics extruders: thickness drifts with melt and recyclate

Film thickness drifts with melt temperature and recyclate share

What happens. The thickness profile of the film moves over the shift. The melt temperature changes with screw wear, ambient and throughput, and a changing share of recyclate shifts the viscosity as well.

What it costs. Scrap, and film that is made thicker than ordered so that it stays within tolerance. Over a year that is a lot of material.

Why the usual fix fails. The thickness gauge sits downstream, so its signal is late. Raising the nominal thickness removes complaints and costs resin on every metre.

How a fixed-price remote diagnosis checks it. From line data (screw speed, pressures, temperatures, take-off speed, gauge profile) and the recyclate share per lot I check how much of the drift follows the melt state and whether profile control on a computed melt state would hold the thickness tighter. The fixed-price remote diagnosis tells you what margin could be taken out.

Background: Computed measurements

Recycled content in packaging makes the melt vary by batch

What happens. The European Packaging and Packaging Waste Regulation requires a minimum share of recycled material in plastic packaging in the coming years. Recyclate comes with batch-to-batch differences in melt flow, contamination and moisture, so the line behaves differently from lot to lot.

What it costs. Scrap, unstable lines and supply risk if quality fluctuates. The more recyclate, the larger the swings.

Why the usual fix fails. Setting the line for each new batch by hand takes time and depends on the setter. Incoming testing tells you the batch, but not how to run it.

How a fixed-price remote diagnosis checks it. From melt pressure, torque and temperature records, lot data and quality results I check how much of the variation can be seen in the process signals and used to adapt the line. The fixed-price remote diagnosis tells you whether adaptive extrusion would help for your recyclate mix. See the background on the rules at industrie-fachwissen.de.

Background: Adaptive predictive control

Pipe wall thickness is measured downstream and corrected late

What happens. The wall thickness of the pipe is measured after cooling, far from the die. When it drifts, many metres have already been made.

What it costs. Over-thickness material as a margin against the minimum wall, and scrap when the wall is too thin.

Why the usual fix fails. Correcting from the downstream gauge is slow and tends to oscillate. A thicker wall by default is expensive.

How a fixed-price remote diagnosis checks it. From die pressures, melt temperature, haul-off speed and the gauge history I check whether the thickness can be predicted at the die and used for predictive control. The fixed-price remote diagnosis shows how much of the over-thickness could go.

Background: Predictive control of process plants

Starting up a new profile takes hours of trial

What happens. A new profile goes on the line and the setters work their way to a good result by trial: temperatures, speed, vacuum, cooling.

What it costs. Scrap, machine time and a dependency on the few setters who know how.

Why the usual fix fails. Settings from the last similar profile give a start but not the answer. Notes get lost between shifts.

How a fixed-price remote diagnosis checks it. From past start-ups (settings, times to a good profile, scrap) I check whether a model can suggest good start settings from the profile geometry and the material. The fixed-price remote diagnosis tells you how much start-up time that could save.

The film bubble becomes unstable at higher output

What happens. In blown film, pushing for more output makes the bubble unstable: it swings, flutters and sometimes breaks.

What it costs. Lost output, because the line is held below what the extruder could give, and breaks with their restarts.

Why the usual fix fails. Operators find a safe speed by experience and stay under it. Standard controllers on air volume and cooling do not see the dynamics of the bubble.

How a fixed-price remote diagnosis checks it. From bubble height, air volume, cooling and take-off data I check what drives the instability and whether a model of the bubble could support stabilising control. The fixed-price remote diagnosis tells you how much more output could be realistic.

Thickness profile bolts on cast film interact

What happens. The die bolts that shape the thickness profile influence their neighbours. Adjusting one bolt moves the thickness at several positions, and the profile control chases itself.

What it costs. Thickness bands, scrap at the edges and slow settling after changes.

Why the usual fix fails. Treating each bolt as its own loop ignores the interaction. Detuning the loops makes them stable and slow.

How a fixed-price remote diagnosis checks it. From the bolt positions and the measured profiles I identify how the bolts interact and check whether one multivariable control would settle faster. The fixed-price remote diagnosis tells you how much better the profile could become.

Background: Predictive control of process plants

The real melt temperature differs from the barrel reading

What happens. The temperature sensors in the barrel show one value, but the melt coming out is hotter or colder. Shear heating from screw speed and the material change what the melt really has.

What it costs. Degradation of the material, quality that varies, and a setpoint that is adjusted to results rather than to the real state.

Why the usual fix fails. A melt probe in the die measures at one point and wears or fouls. A hand-set offset is right only for one operating point.

How a fixed-price remote diagnosis checks it. From barrel temperatures, motor power, screw speed, pressure and any melt probe readings I check how well the melt temperature can be computed. The fixed-price remote diagnosis tells you whether the control can run on that value.

Background: Computed measurements

The sheet temperature profile in thermoforming is not measured

What happens. The heaters warm the sheet, but the temperature over the area is not measured. The edges, the centre and the sag differ.

What it costs. Uneven wall thickness, rejects and settings found by experience for each mould.

Why the usual fix fails. A pyrometer sees one spot. Setting each heater by hand does not account for ambient and sheet differences.

How a fixed-price remote diagnosis checks it. From heater settings, cycle times, sheet data and the wall thickness results I check whether the sheet temperature can be computed and the heater pattern optimised. The fixed-price remote diagnosis tells you what it would take.

Background: Computed measurements

Foam density varies with gas dosing and temperature

What happens. In foam extrusion, density follows the gas dosing, the melt temperature and the pressure. Small changes in one shift the density.

What it costs. Rejects, and material kept in stock because it is outside the density range.

Why the usual fix fails. Density is measured downstream, late. Loops for gas, temperature and pressure are tuned separately but affect each other.

How a fixed-price remote diagnosis checks it. From line data and the density results I check whether density can be estimated and gas, temperature and pressure controlled together. The fixed-price remote diagnosis tells you if coordinated control is worth building.

Preform heating drifts with the ambient temperature

What happens. In blow moulding, the preforms are heated before blowing. When the hall gets warmer or colder, the preform temperature moves, and so does the way the bottle blows out.

What it costs. Weak spots in the wall, rejects and more material to be safe.

Why the usual fix fails. Heater settings are readjusted by hand when someone sees thin spots. Surface temperature sensors see only part of the picture.

How a fixed-price remote diagnosis checks it. From oven settings, ambient data, preform temperature readings and the wall thickness results I check whether the preform temperature can be estimated and the heating adjusted. The fixed-price remote diagnosis tells you what to expect.

Background: Computed measurements

Thickness and orientation drift in film stretching lines

What happens. In stretching lines, the film thickness and the orientation of the molecules must stay within tight limits. Both change with temperature and stretch ratio, and the measurement is far from the actuators.

What it costs. Rejects and rolls with uneven properties.

Why the usual fix fails. Loops on single properties disturb the other one. Slow gauge signals limit how tight a loop can be.

How a fixed-price remote diagnosis checks it. From the line’s actuators, gauge profiles and quality results I check how the profile reacts and whether a profile control would hold it. The fixed-price remote diagnosis tells you how much tighter the tolerance could be.

Recyclate properties change with waste origin and supplier

What happens. Film made with recyclate behaves differently depending on where the waste comes from and who supplied it. Specks hurt printability and the barrier, and the thickness is corrected downstream after the fact.

What it costs. Thickness and quality excursions, with rejects that are found only at the end of the line.

Why the usual fix fails. Incoming tests on a sample say something about the lot but not about the melt in the extruder. More margin on thickness takes material.

How a fixed-price remote diagnosis checks it. From melt pressure, torque, temperature and supplier lot records I check whether the melt state can be estimated and used as a feedforward to the line. The fixed-price remote diagnosis tells you how much of the excursion can be anticipated.

Background: Computed measurements