CE marked · ISO 9001:2015 · 18-month warranty

· products · By

Twin Screw Granulator Cost per Tonne of Saleable Compound

Cost per tonne of saleable compound, not of output: reject rate, re-pelletizing passes and filler loading, with every assumption labelled.

TENGLONG TL-TS twin screw granulator — product photograph from the TL-TS range

Summary: Most cost-per-tonne arithmetic divides total operating cost by tonnes produced, and that denominator flatters any line making material nobody accepts. A twin screw granulator is usually bought for three things a single screw cannot deliver — filler loading, volatile removal and colour consistency — and each of those changes the denominator rather than the numerator. This page builds the model the other way round: cost per tonne of compound that passes incoming inspection, with the reject rate, the re-pelletizing passes and the filler credit carried as separate terms. The arithmetic is illustrative and every assumption is labelled, so you can substitute your own resin price, loading rate and reject rate and see whether the higher machine price still loses.

TENGLONG TL-TS twin screw granulator — product photograph from the TL-TS range
TENGLONG TL-TS twin screw granulator: one of the configurations in the published TL-TS range.

Start with the denominator, not the machine price

The number a purchasing manager compares is cost per kilogram of pellets. It is easy to compute, easy to defend in a meeting, and wrong for a compounding line, because the pellets that fail incoming inspection are still sitting in that denominator. They were melted, filtered, cut and dried with the same electricity and the same operator attention as the good ones. Dividing by total output quietly charges the customer for them.

Two lines can turn out identical tonnage and differ by a fifth in saleable tonnage. On a 600 kg/h platform that gap is worth more per year than the price difference between the two machines, which is exactly why the comparison keeps going the wrong way in spreadsheets that never see a reject bin.

So the model on this page has three layers. Layer one is conversion cost per tonne: energy, wear parts, labour and capital recovery, the familiar arithmetic. Layer two is yield cost per saleable tonne: the reject rate, the cost of putting rejects through a second pass, and the discount on material sold down a grade. Layer three is formulation value: how much filler the dispersion actually allows you to load, and what that displaces. Layers one and two are arithmetic you can close in a spreadsheet. Layer three is the only layer where the machine choice shows up as more than a rounding difference, and it is the layer most quotations leave out.

Layer one is the small number, and that is the whole point

Conversion cost per tonne is total hourly cost divided by hourly output in kilograms, times a thousand. Energy dominates it, and energy for granulation sits in a published band of roughly 0.26–0.50 kWh/kg depending on polymer and moisture. At a US$0.10/kWh tariff that band is worth US$26–50 per tonne between the best and worst case, and a single tariff change or a wetter feedstock moves the number further than the gap between two well-built machines.

Read that carefully, because it cuts against the usual sales argument in both directions. A twin screw does not win on layer one. Co-rotating screws with a 40:1 L/D and gravimetric dosing cost more to buy and usually draw slightly more per tonne than a well-set single screw granulator on the same polymer. If conversion cost were the whole story, the cheaper machine would always win, and in practice it does not. The reason is in the next two layers.

For the full item-by-item breakdown of layer one, including how the capital recovery term collapses when annual tonnage falls, see what a plastic granulator costs to run. This page will not repeat it, because repeating it is how cost models get mistaken for decisions.

Layer two: the yield cost that never reaches the spreadsheet

A reject is not a loss of the full tonne. It is a loss of the difference between what that tonne cost to make and what it can be sold for after it has been through the line again. That difference has four parts, and only the first one is obvious: the second pass consumes the same energy and labour per kilogram; the polymer takes another thermal history, which costs intrinsic viscosity in PET and causes colour drift in most thermoplastics; the regrind has to be blended back at a ratio the customer will accept; and the tonnage that could have been made during those hours was not made.

Written as a term, yield cost per saleable tonne is the reject rate divided by one minus the reject rate, multiplied by the cost of a second pass plus the downgrade loss. An 8% reject rate is not an 8% cost. It is roughly 8.7% of the good tonnage burdened with a full extra pass, which on a 600 kg/h line is where a five-figure annual number hides inside a sentence like ‘we get a few off-spec at start-up’.

The engineering question is which defects a given platform can hold down without an operator watching. Twin screw compounding holds colour and dispersion because the mixing elements do the work in the screw, not in the die. That is why the reject rate on a recipe with 20–30% filler is usually dominated by start-up and shutdown transitions rather than by steady-state drift — and why the saving shows up as fewer transitions, not as faster running.

Layer three: filler loading is the term that pays for the machine

Here is the mechanism in one sentence: better dispersion means the same mechanical properties can be reached at a higher filler load, and every percentage point of filler displaces a percentage point of resin.

Take an illustrative set of assumptions and label them as assumptions, because they are not our data and they are probably not yours. Resin at US$1,100 per tonne, calcium carbonate at US$180 per tonne, a specification that currently fixes loading at 20%, and a dispersion quality that would allow 24% without losing tensile strength or impact. Four points of substitution is 40 kg of resin replaced by 40 kg of filler in every tonne, a raw-material saving of about US$37 per tonne of compound. On 600 kg/h over 4,000 hours that is a little over US$88,000 a year, against a machine price difference in the low tens of thousands.

Two honest qualifications. First, that credit only exists if the customer’s specification permits the higher load or if the product is your own; where the recipe is fixed by a tier-one buyer, the credit is exactly zero and the twin screw has to justify itself on layers one and two alone. Second, the credit is a commercial negotiation as much as an engineering result — somebody has to accept the change, and impact strength at low temperature is usually where the argument happens. Treat the number as the size of the prize, not as a guarantee that you collect it.

This is also the reason a masterbatch granulator and a filled-compound line are specified differently even though both are twin screw machines: one is judged on dispersion of a small addition, the other on how much addition the screw can carry.

A worked comparison at 600 kg/h, with three denominators

The table below uses one output figure, one resin price and one power tariff across all three columns, and varies only the things the machine choice actually changes. Every figure is illustrative arithmetic, not a quotation.

Reading the table without fooling yourself

Column four is the only one that should be compared, and it says something uncomfortable for both sides of the argument. Between the first two rows the twin screw wins by four dollars a tonne — real, but not the reason anyone buys one. The reject rate does the work there, and a single screw line run by an operator who knows the material can close most of that gap. Between rows two and three the loading rate does the work, and the gap widens by an order of magnitude.

Which gives a clean decision rule. If your formulation is fixed and your reject rate is already low, buy on layer one and stop reading. If your formulation is yours to set and your filler load is below what the specification would allow, the machine is not a cost, it is the instrument that unlocks the saving — and the quotation should be argued on that basis, not on price per kilogram of pellets.

Before any of this is real, the reject rate and the achievable loading have to be measured on your material rather than assumed. That is what the factory test run is for: four hours on your own feed, with output, melt pressure, melt temperature and kWh logged every minute, and the extrudate sampled for dispersion. The twelve-platform comparison is the right page for shortlisting, and single screw versus twin screw covers the mechanism behind the mixing claim.

What this model deliberately leaves out

It leaves out your maintenance practice, which moves wear cost more than the machine choice does. It leaves out the tariff structure, because a demand charge or a night rate can invert the energy comparison. It leaves out changeover time on a line running six colours a week, which is often the largest single term in real compounding economics and has nothing to do with cost per tonne at steady state. And it leaves out the possibility that your reject rate is high for a reason no machine fixes — wet feed, contaminated regrind, or a die that has been polished one time too many.

Use the three-layer structure as a way of asking better questions rather than as an answer. The layers are ordered by how much they can move the result, and the order is the useful part.

Illustrative arithmetic at 600 kg/h, resin at US$1,100/t, filler at US$180/t and power at US$0.10/kWh. Assumptions, not quotations; substitute your own figures.
Line and recipe Conversion cost / t Yield cost / t Filler credit / t Cost per saleable tonne
TL-SJ single screw, 20% filler, 8% reject US$88 US$31 — US$119
TL-TS twin screw, 20% filler, 3% reject US$104 US$11 — US$115
TL-TS twin screw, 24% filler, 3% reject US$104 US$11 −US$37 US$78

Frequently asked questions

Q1: Does a twin screw granulator cost more per tonne to run than a single screw?

On conversion cost alone, usually yes: the co-rotating screws, gravimetric dosing and larger drive draw slightly more per tonne on the same polymer. The comparison only turns when yield and filler loading are counted. On a fixed recipe with a low reject rate, the single screw is often the cheaper tonne, and buying a twin screw for that duty is hard to justify.

Q2: How do I estimate a reject rate before I have the machine?

You cannot measure it from a catalogue, so measure something adjacent instead: the reject rate on your current line, split into start-up transitions and steady-state drift. Transitions scale with colour and recipe changes; steady-state drift scales with dispersion quality. That split tells you which platform can actually reduce your number, and the factory test run then confirms it.

Q3: Is a second pelletizing pass really that expensive?

It costs the full energy and labour of a pass, plus a thermal history. In PET the second pass costs intrinsic viscosity, which limits how much regrind can be blended back before the customer notices. In most thermoplastics it costs colour. Add the tonnage you did not make while re-running material, and an 8% reject rate behaves like 8.7% of good output carrying an extra pass.

Q4: Can I raise filler loading on an existing twin screw line?

Sometimes, and the limit is usually dispersion quality rather than screw geometry. Raising load changes melt viscosity, which changes die pressure and cut quality, so the experiment is a loading ladder with mechanical testing at each step, not a single jump. If the line was specified for a lower load, check the drive torque margin first.

Q5: What if my customer’s specification fixes the formulation?

Then the filler credit is zero and the twin screw must justify itself on conversion and yield cost alone. That is a legitimate reason to buy a simpler machine. The honest question becomes whether the fixed recipe has a colour or dispersion tolerance the cheaper platform cannot hold, because if it does, the reject rate decides the purchase and the loading argument is irrelevant.

Engineer’s note

Every figure in that table is illustrative arithmetic with the assumptions printed beside it — not a measured result from a customer line and not a quotation. Reject rate and achievable filler load are properties of your material and your specification, and neither can be read off a catalogue. We write the confirmed numbers onto the quotation after a factory test run on your own feed, because that is the only place the denominator becomes real.

Engineering note: the numbers on this page are our own published reference ranges for the machine family, not a specification guarantee for any individual unit. Final output, specific energy and pellet quality are set by the incoming material, its moisture and contamination level, and the screen pack fitted. We write the confirmed figures onto the quotation after a factory test run on your own material.

Who wrote this

Zhou Gong (周工) is a senior plastic machinery engineer at TENGLONG Plastic Machinery in Linyi, Shandong. The factory was registered as Linyi Hedong District Tenglong Plastic Machinery Factory on 7 June 2016 and grew out of a screw and barrel workshop, so his work is rooted in the parts that decide whether a granulator holds output: screw geometry, nitriding case depth, barrel alignment and melt filtration. He prepares the technical specification for customer quotations, reviews the four-hour factory test data before shipment, and runs commissioning support on site. Screws are nitrided in-house to a 0.5–0.7 mm case at HV 900–1000 and straightened to 0.03 mm runout over the full length; barrel, screw, gearbox and frame are aligned to 0.05 mm before the test run.

Figures quoted here are engineering reference ranges published on our own product pages, not guaranteed values for any specific machine. Real output, specific energy and pellet quality depend on the incoming material, its moisture and contamination level, and the screen pack fitted. Final values are written after a factory test run on the customer’s own material, and the PLC-logged test report ships with the machine. For a quotation that reflects your material, send the feed form, contamination level and target output and a configuration comes back within one working day.

Send your formulation, current filler load, reject rate, target output and power tariff to the quotation request form and the three-layer arithmetic comes back filled in with your numbers, alongside the configuration it implies.

Related reading

العربيةAR