The furnace that is itself a measuring instrument

In a melting furnace the lining protects the steel. It should last, withstand attack and let as little heat as possible escape. When it slowly recedes, the energy bill notices first.

In a fire test furnace the lining is something else: it is part of the measuring instrument. Such a furnace tests how long a component withstands a fire — a door, a façade, a glazing unit. The specimen is clamped into a test frame, then the furnace runs the standard temperature-time curve: a normative curve that reproduces the course of a room fire. After thirty, sixty or ninety minutes it is settled whether the component reaches its class.

For that result to be worth anything, the curve has to be reproducible. And it depends not only on the burners but on the wall behind them: how fast the furnace heats up, how much heat the lining swallows, how evenly the radiation reaches the specimen. Change the lining and you change the measurement.

What we found

The job covered two furnaces at a manufacturer of fire-resistant glass — one small and one large test furnace. Both were worn in the same places, and that is no coincidence: they are the places where the geometry is interrupted.

• The viewing windows. Through them the specimen is observed during the test. Every opening is a disturbance in the wall build-up: the reveal stands at an angle in the radiation field, and its edges take the full temperature jump at every test.
• The pillars. In the small furnace over the full height, in the large one the upper sections. A pillar carries load and at the same time stands free in the firing chamber — it is loaded from more sides than a wall.
• The lower fields of the large furnace, left and right. That is where radiation bears on the same surface for the longest.

Cast, not bricked

Viewing windows and pillars were broken out and recast with a two-component special castable. The reason is dimensional accuracy: a reveal that determines the viewing angle and the seal can be produced more precisely cast than bricked. And a monolithic reveal has no joints for the temperature jump to attack.

The insulating layer and the anchors were completely renewed in the process. This is the point where test furnaces are most often economised on — and the reason the same spot is back on the schedule two years later. Renew only the visible layer and you set fresh castable on old, embrittled insulation.

Lightweight brick instead of dense castable

In the lower fields we laid insulating firebrick of group 26. That looks like a step backwards from casting, but it is the right choice: insulating brick has a low bulk density and therefore stores little heat.

And that is exactly what matters in a test furnace. It should follow the standard curve, not heat up against its own mass. Every kilogram of lining that has to be heated along makes the curve more sluggish. A dense, heavy material would last longer — and would falsify the measurement.

The wall build-up behind it

• Ceramic fibre blanket, 1260 °C, 13 millimetres — takes up the expansion.
• Microporous boards, 25 millimetres — the actual insulation, at very low thickness.
• Calcium silicate boards, 25 millimetres — load-bearing and dimensionally stable towards the steel.

This staggering is no accident. It keeps the shell cool without building up thickness — and in a test furnace every centimetre of wall is a centimetre less test space.

The test frames

Added to this was the refurbishment of the two frames that hold the specimen. They received ceramic fibre paper in two layers and a needle-reinforced castable — steel needles of heat-resistant grade, two to three per cent. The needles hold the castable together when the frame is heated abruptly during the test. Without them, a component that thin cracks on the first cycle. The work was done lying down: for the lining, the frames have to rest horizontally at floor level, otherwise the castable runs out of the formwork.

The frame of the job

Five fitters on day shift, both furnaces in parallel, around five to six days. The lead time was three to four weeks — not because of the installation, but because of the material: special castable and insulating brick of this grade are not held in stock.

What finally decides service life

Not the installation, but the first heat-up. Freshly cast refractory concrete contains water. Heat too fast and it expands as steam many times over, bursting the lining from inside — faster in the thin sections of a test frame than anywhere else. That is why every lining needs a dry-out schedule: a curve with holding times, matched to material and component thickness. On this job we supplied the schedule; the operator carried out the dry-out himself — the usual route when the plant has its own control system.

• The lining belongs in the measuring chain. In the same inspection routine as thermocouples and burner control, not in maintenance whenever convenient.
• Heavier is not better here. Line a test furnace with dense material and you buy service life with sluggishness. The standard curve does not forgive that.
• Renew the insulation behind the hot face as well. Fresh castable on old insulation brings the same spot back onto the schedule in two years.
• Agree the dry-out beforehand. Who dries, to which curve, with which measurement — that is the difference between a lining that holds and one that cracks on the first test.

Related articles