Carbon and material efficiency in sheet metal plants

A fabricated steel bracket that leaves a plant at 1.2 kg carries roughly 2.5 kg of CO2 in the steel before a single machine has run. The same bracket's share of cutting, bending, welding and coating energy is about 0.6 kg. Any carbon programme that starts with the machines is therefore working on 19% of the problem while the other 81% sits in the purchase order and the scrap skip.

That ordering is not an opinion about priorities. It follows from the mass balance. Steel is bought at 2.1 kgCO2e per kg and arrives as a sheet. Between the sheet and the shipped part, 20–35% of the mass becomes offcut. Every kilogram of offcut carries the full 2.1 kgCO2e it cost to make, and it is sold for remelting at a fraction of that, so the cost of the loss stays inside the delivered part.

This page sets out that balance with reproducible numbers, the monthly record that proves it, and the point at which a yield improvement is worth more than any machine change. The plant, quality and sustainability hub covers the rest of how the plants are run.

Carbon and material efficiency: which number decides it?

Kilograms of CO2e per delivered kilogram, calculated from mass bought, mass shipped and energy metered. That single ratio lets a buyer compare two suppliers, two nest layouts or two years without arguing about which machine is more efficient.

The arithmetic is short. If yield is Y, then every delivered kilogram needs 1/Y kilograms of bought steel. Embedded carbon per delivered kilogram is that figure multiplied by the mill factor, plus process energy per kilogram.

At 65% yield and 2.1 kgCO2e per kg of steel, bought steel per delivered kg is 1/0.65 = 1.538 kg, and embedded carbon is 1.538 × 2.1 = 3.23 kgCO2e. At 80% yield the same calculation gives 1.25 kg and 2.63 kgCO2e. That is a 0.6 kg reduction per delivered kilogram, or 19%, from a nesting change that costs nothing but engineering time.

Run the same sensitivity on the machines and the answer is different. Cutting electricity per delivered kilogram is 0.36 kgCO2e at the factors below. Halving it saves 0.18 kg, which is about four points of nest yield. Halving a 22 kW laser's wall draw is not on offer; four points of yield usually is.

The material and energy balance for one delivered tonne

The figures below are per delivered tonne of fabricated CRCA parts at 2 mm average thickness, with a 70% nest yield, one weld per part and a powder coat. Energy factors are stated so the reader can substitute their own.

Two rows in that table surprise people. The first is that cutting is the largest energy band even though it is small against steel, because a laser holds 22–25 kW of wall draw for the whole nest. The second is that the offcut row has no credit. Scrap sold to a reroller is genuinely recycled, but the emissions that made the sheet are not refunded to the part that was shipped. Netting them out flatters the figure and hides the yield problem the customer is asking about.

The process bands total 645 kgCO2e, or 18% of the footprint. That is the entire scope for improvement inside the plant, and the largest single item in it is the laser.

Why welding is not the number people expect

A MIG arc at 25 V and 160 A puts 4 kW into the joint. At 0.5 m/min travel that is 0.133 kWh per metre of weld. A 3 mm fillet deposits about 35 g of metal per metre, so weld energy lands near 3.8 kWh per kilogram of deposited metal, or 2.7 kgCO2e per kilogram of weld metal at the grid factor.

A 22 kg genset base frame with 1.4 kg of weld metal therefore carries about 3.8 kgCO2e in its welding energy, against 46 kgCO2e in its steel. Doubling weld efficiency saves 1.9 kg. Lifting the nest yield on the same frame from 70% to 75% saves 4.4 kg. The arithmetic keeps pointing the same way.

Where the process numbers come from

Cutting energy is metered, not estimated. A 12 kW source with chiller, extraction and drives draws 22–25 kW at the wall, so a 4-hour cutting shift is 88–100 kWh regardless of how much of it is piercing. Divide by kilograms cut in that shift and the per-kilogram figure falls out. That is why the same machine shows 0.4 kWh/kg on a dense nest of small parts and 0.7 kWh/kg on a nest of large outlines with long pierce lists.

The record: what to measure, how often, and what it prevents

The balance is only useful if it closes. That means five lines, taken from documents that already exist, every month.

Kilograms bought by grade. From the mill invoice and the goods-received note. CRCA, stainless, aluminium and galvanised go in separate lines because their factors differ by a factor of four.

Kilograms dispatched. From the packing list, by part number, net of packing.

Kilograms of scrap by grade. From the weighbridge ticket on the scrap skip. This is the line most plants do not have at part-number level, and without it a yield loss on one job hides inside a good month on another.

kWh by meter. Sub-meter the laser, the press shop, the brake line, the weld cells, the compressor and the coating line. One plant-wide meter cannot answer the question a customer asks.

Coating consumables. Powder kilograms bought, gas or LPG metered to the oven, and pretreatment chemical additions logged per tank.

Taken monthly, those five lines produce a per-tonne figure that moves when something changes. Taken annually from the accounts, they produce a number that moves when prices change, which is not the same thing.

Two smaller records support the monthly close. The first is the nest report from the laser: parts per sheet, yield percentage and remnant size, saved with the program revision. The second is the scrap segregation log, because a skip that mixes CRCA with galvanised and stainless cannot be sold as a clean grade and the credit disappears.

Closing the balance to within 3%

The mass balance closes when kilograms bought, minus kilograms dispatched, minus kilograms of scrap, equals the change in work in progress. When it does not close, the gap has a signature and the signature names the cause.

A gap of 2–5% that appears every month and follows no pattern is uncounted scrap: slugs from a press tool, trimmed strip ends, first-article parts. A gap that appears once, in the month a new part number starts, is setup loss, and it belongs to that job rather than the month. A gap that grows with the number of grades on the floor is a segregation error, usually galvanised sheet weighed as CRCA. A gap that appears only when the laser runs a new nest is a remnant booked as a finished part.

Energy does not balance because it is metered. The check there is per-piece consumption against the previous revision. A jump of more than 15% on an unchanged program means a nozzle, lens or assist-gas setting has drifted, which is a cutting-cost problem that happens to show up in the carbon record.

Recalculating when the part changes

A part revision changes the balance, and the record has to follow. If a bracket moves from 2 mm to 3 mm CRCA, mass per part rises by 50%, the nest yield usually falls because the flat pattern grows faster than the sheet, and cutting time per part rises with the perimeter rather than the area. Recompute all three rather than scaling the old figure by mass.

The same applies when a part moves from laser to a press tool. Cutting energy per part drops sharply at volume, but the blank layout changes, so the yield can improve or worsen. Laser cutting versus stamping covers the volume crossover; the carbon consequence follows the same crossover with a smaller amplitude.

What we measure, and what the numbers look like

Ganesh manufactures under ISO 9001:2015 and ISO 14001:2015 across five units in Mumbai (Vasai) and Nagpur, which means the environmental monitoring that ISO 14001 clause 9.1 requires is already a documented routine rather than a project. The treatment plant has its own daily log, and the seven-tank pretreatment line records chemical additions and rinse conductivity per shift, because those records also control coating adhesion.

On the material side, the yield figure comes from the nest report rather than from a plant-wide average. A part with a 92% nest and a part with a 61% nest both feed the same monthly total, and only the per-part report tells you which is which. That is the number we quote when a customer asks for the carbon per piece, and it is the number that changes when a drawing revision lets us re-nest.

The machine data that feeds the energy side is on the laser cutting capability page and the paint shop page: a 12 kW fibre laser with a 40 ft × 16 ft bed, a 200 T 5-axis brake with a 4,450 mm bed, three CLOOS robotic MIG cells, the seven-tank pretreatment line and the 18 ft × 14 ft powder booth. Those are the assets that draw the 15% of the footprint that is not steel.

We do not publish a plant-level carbon intensity per kilogram for an arbitrary part, because it depends on the part. What we can do is give the balance for a specific part family with the calculation attached. That is more useful to a buyer and considerably harder to get wrong.

Improving yield before improving machines

The order of work is fixed by the arithmetic. Yield first, then scrap segregation, then process energy, then the tariff.

Re-nest at the current sheet size. A 1,250 × 2,500 mm sheet is not always the cheapest format. A 1,500 × 3,000 mm sheet often lifts yield by 3–6 points on large flat parts because the remnant fraction falls. How nesting reduces cutting cost has the geometry.

Design the part to nest. Two parts that share a common edge can be cut as one blank and split, or designed with matching profiles so the offcut between them disappears. This is a drawing change, not a machine change, and it is the cheapest carbon reduction available.

Keep grades separate. A segregated CRCA skip sells as a known grade. A mixed skip sells as mixed. The emissions are identical; the credit is not.

Cut at the right thickness. A part re-specified from 3 mm to 2 mm because the load case allows it removes a third of its steel. That is a design decision with a mass consequence, and it dwarfs any process setting.

Then look at the machines. Idle time on a 22 kW laser for an hour costs 15.6 kWh, which is 11 kgCO2e. It matters at plant level and it is the last item on the list for a reason.

What to ask a fabricator for

Five questions produce a defensible answer, and a supplier who cannot answer them is estimating rather than measuring.

1. The nest yield for this part number, from the current program revision.
2. The scrap split by grade and the destination of each stream.
3. Metered kWh per 1,000 parts for the route you would run.
4. Coating consumables per square metre coated, and the oven energy per cycle.
5. The mill and grade of the steel, with the mill's published cradle-to-gate figure.

Ask for the calculation, not the total. A per-tonne figure with no mass balance behind it cannot be checked, and a buyer who cannot check it cannot defend it. When you have the drawing and the annual quantity, send it for review and we will return the balance for that part family with the yield, the route and the energy per piece shown line by line.