How to Size an Atmosphere Furnace: Work Backwards from Annual Tonnage

Capacity is the part of furnace selection where conversations break down. The buyer says "we need three thousand tonnes a year", the supplier says "our machine can do that", both believe agreement has been reached — and the gap shows up after installation.

Nobody is lying. The word "tonne" simply hides assumptions that were never stated: finished tonnes or charged tonnes, twenty-four hours a day or twenty, three hundred days a year or three hundred and thirty, what bulk density, and whether saggars and trays count toward the weight. Each one shifts the result substantially.

The right approach is not to ask "how big a furnace do you need" but to break annual output into numbers you can compute, then work backwards to the chamber.

Step one: turn annual tonnage into hourly throughput

Fix four base figures first:

• Annual output as finished product, in tonnes per year
• Effective working days, after maintenance, holidays, commissioning and changeovers. Many plans assume 330 days and run around 300; use 300 at the planning stage
• Operating hours per day, 20–24 for continuous equipment, actual shifts for batch
• Yield, the good-part rate of the sintering step including scrap and rework

The formula is plain:

Hourly throughput (kg/h) = annual tonnes × 1000 ÷ effective days ÷ daily hours ÷ yield

A worked example, every figure assumed, used only to show the method: 3,000 tonnes a year, 300 days, 20 hours, 95% yield gives about 526 kg/h. That number is the starting point for sizing.

Step two: convert finished weight into charged weight

This is the step most often skipped and the one that produces the largest error.

Finished weight is not what goes into the furnace. Three layers sit in between:

• Moisture and binder: granulated powder carries a few percent moisture, and binder burns off during debinding, so the charge is heavier than the product
• Saggars, trays and crucibles: these travel through the furnace, consuming weight and volume without contributing to output
• Fill factor: material in a tray is not solid. Void space between particles means bulk density is far below true density

So compute the actual load per tray:

Load per tray (kg) = tray volume × bulk density × fill factor

Bulk density must be measured, never substituted with true density. For the same LFP powder, different particle size and granulation can shift bulk density by thirty percent. Fill factor depends on loading method and how well the material is settled, and usually needs a meaningful discount.

Get the charge weight wrong and every dimension downstream is wrong. There is no shortcut here: run one trial load with the real material.

Step three: calculate cycles and rhythm

With load per tray and hourly throughput, the tray rate follows:

Trays per hour = hourly throughput ÷ load per tray

Then look at process time. A full sintering cycle includes heating, soaking and cooling. Soak time comes from the process and cannot be compressed. Batch equipment adds loading, evacuation or purge, cooling and unloading.

• Batch: cycle = load + heat + soak + cool + unload. Daily output = load per cycle × cycles per day
• Continuous: hourly output is set by roller speed or pusher interval, equal to trays passing per hour multiplied by load per tray

One counter-intuitive point: the longer the soak, the smaller the advantage of continuous operation. Continuous saves the waiting around heat-up and cool-down; if the soak itself runs for fifteen hours, the difference between the two approaches shrinks.

Step four: derive the chamber from the rhythm

Now the dimensions are calculated, not guessed.

Batch: effective chamber volume ≥ trays per cycle × tray volume ÷ chamber fill ratio. The chamber fill ratio accounts for gas channels, element placement and clearance between trays; you cannot simply pack the chamber solid.

Continuous: chamber length is the sum of zone lengths. Each zone's length follows from residence time and travel speed:

Zone length (m) = residence time (h) × travel speed (m/h)

Travel speed comes from the rhythm: trays required per hour multiplied by tray spacing. The chain therefore reads output → rhythm → speed → zone length → total length.

Width and height follow tray dimensions and gas flow organisation. The atmosphere has to pass through the load, not just around it.

The five places the arithmetic usually breaks

1. Using finished tonnes as charge weight. Moisture, binder and furniture are ignored, actual charge is higher than assumed, and the furnace comes out undersized.
2. Using true density instead of bulk density. Real tray density can be half of true density or less; this single mistake can inflate the chamber absurdly.
3. Assuming ideal uptime. Equipment needs maintenance, changeovers and troubleshooting. Plans built on 365 days rarely land.
4. Ignoring changeover loss. For high-mix, low-volume production, every change requires re-stabilising atmosphere and temperature. That time is not in the formula but it is in the calendar.
5. Confusing margin with safety. Leaving 15–20% headroom is normal practice, but more is not better: a large margin means running at low load for years, which hurts both energy consumption and atmosphere stability.

Check the result in reverse

Once you have dimensions, run the calculation backwards: with this chamber and rhythm, what is the hourly output? If it does not match step one, one of your assumptions was optimistic.

Check three more things at the same time:

• Can the gas supply keep up? Larger chamber means more gas exchange; source, piping and flow meter range must be sized together
• Is the power sufficient? Heating rate and load together determine power demand. ZKR atmosphere furnaces span 100–2000 kW, configured per project
• Can uniformity be held? Longer and wider chambers make uniformity harder. Ask about measuring points and the uniformity specification at proposal stage, not at acceptance

Frequently asked questions

How big a roller hearth kiln do I need for 5,000 tonnes of LFP a year?

That cannot be answered directly — three numbers are missing: bulk density, saggar size and soak time. Supply those and run the four steps; the size follows. Any proposal that quotes dimensions without asking for those three is guessing.

Can batch furnaces match continuous output?

Yes, by running several in parallel. But that means multiple atmosphere control systems and multiple loading stations, so labour and maintenance rise. A single batch unit is limited by cycle time and that limit cannot be compressed.

How much headroom is appropriate?

15–20% is the norm. If you know you will double output within two years, do not build the margin into one machine — leave room in the building and the gas supply, and size the equipment for current demand.

The capacity is settled, so why do quotations still differ so much?

Because capacity only fixes chamber size. Price also depends on zone count, heating element material, atmosphere type, control system, refractory grade and automation level. Two furnaces of the same size with different specifications can differ by a factor of two, and that is normal.