Extrusion Line Start-Up and Shutdown: The Sequence That Saves Scrap

Why start-up is where the scrap is made

Ask a plant where its scrap comes from and the answer is usually "defects". Look at the numbers and most of it comes from two places: the first hour of a run and the last twenty minutes before a changeover.

The reason is thermal. An extruder is a several-hundred-kilogram mass of steel that has to reach and hold a temperature profile within a few degrees, and the barrel does not reach equilibrium when the controller display says it has. The display reads a thermocouple in a well a few millimetres from the heater; the screw and the melt channel behind it are still catching up.

Heating: the soak time nobody allows

Heat soak time scales with the square of the barrel diameter, because it is a conduction problem through steel. A small machine settles quickly; a large one does not.

Table 1 — Heat-up and soak times by screw diameter. The soak column is the one that gets cut when the shift starts late.

Two practices shorten the effective soak. First, set the barrel zones to temperature in the order they will be needed — feed zone first, then compression, then metering and die — so the whole barrel is not fighting for heat at once. Second, switch on the die and adaptor heaters earlier than the barrel, because the die is the largest mass and the slowest to stabilise.

What goes wrong without a soak is subtle. The melt in contact with the screw flights is hotter than the melt in the channel centre, because the steel is still absorbing heat. The result is a viscosity profile that varies across the flow, which shows up as die lines, inconsistent gloss and dimensions that drift for the first hour. None of it looks like a temperature problem at the time.

Alignment before the first pellet

While the machine is cold and empty is the only convenient time to check alignment. Once it is running and hot, everything has moved and the measurement is meaningless.

Table 2 — Alignment checks. All of them are cheap on a cold machine and expensive to discover from the product.

The die-to-barrel check is the one that gets skipped, because the die was on the machine last week and "nothing changed". Something always changes — a die cleaned and reassembled, a different adaptor, a new screen pack housing. Ten minutes with a dial indicator saves an hour of chasing wall variation.

Bringing the line up

The start-up sequence matters because each step establishes a condition the next one depends on. Rushing to full line speed before the die is thermally stable produces scrap that has to be pulled back through the whole downstream anyway.

Table 3 — The nine-step start-up. Step 8 is where most of the scrap is generated, because the line is ramped faster than the cooling can follow.

The single most useful habit is to ramp in steps and check dimensions at each one. A line brought up from 2 m/min to 12 m/min in one move will produce 200 m of out-of-tolerance product before anyone notices, because the cooling length that worked at 2 m/min is too short at 12. Ramping in 20% steps with a measurement at each costs ten minutes and saves the 200 m.

Reaching steady state

Steady state is not when the product looks right. It is when the variables that affect the product have stopped moving. Five indicators tell you.

Table 4 — Steady state indicators. Head pressure is the earliest and most sensitive; dimensions are the last to settle.

In practice a profile line reaches dimensional steady state 30–60 minutes after the last speed change. On heavy sections it takes longer, because the thermal mass of the product itself is part of the system — a 10 mm rod needs its whole length in the tank to be at equilibrium before the cutter sees consistent material.

Shutdown: purge, park or hold

How a line is shut down determines how it starts up next time, and the difference is measured in hours.

Table 5 — Shutdown options. Holding at temperature is cheapest for short breaks and the most expensive mistake for a weekend.

PVC is the material that punishes the wrong choice hardest. Rigid PVC at 190 °C degrades within about 20–30 minutes of standing still, and the degradation is autocatalytic — once it starts it accelerates. A PVC line left at temperature over lunch without screw rotation will need the die stripped. Hold with the screw turning slowly, or purge.

Polyolefins are forgiving by comparison and can be held at temperature overnight. But even PP will oxidise over a long hold, and the first material out will be discoloured and low in molecular weight. Purge a barrel volume before starting production.

Restart after a weekend

A cold restart is a different procedure from a warm one, and treating it as the same is the reason Monday morning produces so much scrap.

Bring the machine up slowly. Heating a cold barrel at full power creates a large temperature gradient between the heater band and the inner wall, and on a worn barrel that gradient can crack the liner. Most controllers have a slow-start or soft-start mode; use it. Then follow the full soak from Table 1 — a machine that has been cold for two days needs the full time, not a discounted version of it.

Check the die bolts and the adaptor flange before feeding. Thermal cycling loosens them, and a leaking adaptor flange discovered at production temperature means shutting down again. Torque them cold, then re-check hot.

The checklist

A start-up checklist is worth more than any single process improvement, because it converts knowledge held by two experienced operators into a procedure anyone can follow. Keep it to one page and make it specific to the machine, not generic.

Table 6 — A one-page start-up and shutdown checklist. Logging the soak start and end time alone removes most of the first-hour variation.

The logging matters as much as the checklist. A plant that records soak time, purge volume and the time to steady state for every run can see which machines and which operators are losing an hour, and can fix it. Without the record, every shift believes it is doing the same thing.