Oct 28, 2026
Glass Furnace Linings Run 100-200 C Hotter Than Steel: Why the Backup Board Faces Alkali, Not Just Heat
A glass melting furnace sits 100 to 200 C above a steel reheating furnace at the hot face, and the backup lining feels every degree of it. This guide maps the real temperatures by zone (crown 1550 to
--- title: "Glass Furnace Linings Run 100–200 °C Hotter Than Steel: Why the Backup Board Faces Alkali, Not Just Heat" slug: nano-insulation-board-glass-furnace-guide summary: > A glass melting furnace sits 100 to 200 °C above a steel reheating furnace at the hot face, and the backup lining feels every degree of it. This guide maps the real temperatures by zone, the heat-loss ledger (crown 23.9%, regenerators 24.9%, uptake ports 20.6% of total loss in a measured container furnace), why glass is a harder job than steel (alkali vapour, 1650 °C crowns, arch expansion), the conductivity of a microporous board at 800–1200 °C against ceramic fibre and light brick, measured retrofits (22 mm dropped a 1300 °C melt's shell from 159 to 114 °C, −45% heat loss), a step-by-step thickness method, and the mistakes that destroy a glass-furnace install. ---
Most glass plants size the backup insulation on the wrong number. They take the flame temperature — 1550 to 1650 °C at the crown — and shop for a board that survives it. The board never sees that number. What it sees is the cold face of the refractory working layer, which in a glass furnace still runs 800 to 1100 °C at the spots that matter, and the atmosphere around it is loaded with sodium, boron, fluorine and sulphur vapour that quietly eats boards which would outlast a steel furnace.
That single correction changes the whole specification. On a glass furnace the thin microporous board you bolt or glue onto the cold face of the refractory is not mainly an energy saving — though it is a large one. It is the easiest lever you have on crown and regenerator heat loss, on operator-facing shell temperature, and on the part of the energy bill your auditors actually measure. The data below shows where the board pays back, where glass differs from steel, where the conductivity argument actually holds, and where a conventional install fails.
A glass furnace runs hotter than a steel mill — and the backup layer feels it
The numbers are not close. In a cross-fired regenerative container or float furnace:
| Zone | Hot-face temperature | Typical refractory | Backup-layer hot face (typical) | |---|---|---|---| | Melting tank sidewall (metal line) | 1500–1700 °C | Fused-cast AZS (33–41% ZrO₂) | 1000–1150 °C | | Crown (furnace roof) | 1550–1650 °C | Silica brick (≥95% SiO₂) | 900–1100 °C | | Breast / superstructure wall | 1450–1600 °C | Silica / high-alumina | 850–1050 °C | | Regenerator top checkers | 1300–1550 °C | High-alumina / magnesia | 800–1100 °C | | Regenerator bottom | 500–900 °C | Fireclay | 400–700 °C | | Glass bath (working end) | 1300–1450 °C | AZS / corundum | 800–1000 °C |
Sources: Highland Refractory brick-selection guide; RefractorySource glass-industry refractories; Lehigh University IMI-NFG lecture on industrial glass furnaces. The crown operates at 1570 °C normally and can reach 1600 °C in upset conditions (Glass Melting Furnaces — Design and Efficiency, 2006 industry handout).
Compare that to a reheating furnace wall, where the working lining cold face is usually 700–900 °C, or a coke oven combustion chamber at 1300–1400 °C. Glass is 100–200 °C hotter at the refractory, and because heat loss rises with temperature to the fourth power for radiation, that gap multiplies. A backup board that is comfortably within rating on a steel furnace can be at or above its limit on a glass crown.
The heat-loss ledger (where the energy actually goes)
Measured data on real furnaces is more useful than vendor averages. Two independent studies bracket the answer:
Measured container furnace (>100 t/d, Kerchensk combine, 1985, OSTI/GLCEA): heat
loss through the masonry split as crown of the flame space 23.9%, regenerators 24.9%, uptake ports 20.6%, furnace bottom 14.9%. The crown and the regenerator together are roughly half of all structural loss.
Japanese 1990s heat balance (39.7% thermal efficiency medium tank): ~19.8% of
heat input lost from the melting-tank walls, 5.1% from the regenerator, 17.9% from exhaust/flue/exchanger — total structural dissipation 42.8%, against only 25.8% carried by the glass and reaction heat (Wiley ACerS survey of soda–lime–silica compositions).
Glass Europe furnace model (2023): wall losses alone are 0.88 GJ/t on a new
furnace, 1.15 GJ/t average, and 1.44 GJ/t at end of life — out of total specific consumption of 4.18 / 4.48 / 4.85 GJ/t (LHV). Structural loss grows ~1% per year of campaign as linings degrade.
One more figure that should change how you insulate: in a large furnace the un-insulated surface is about 10% of the area but produces about 35% of the total losses; in a smaller furnace 12% of area produces over 40% (2006 design-and-efficiency handout). The gaps, the uncooled patch, the door that lost its board — those dominate. A continuous microporous backup that closes those gaps is worth more than its average conductivity suggests.
Why glass is a harder insulation job than steel
Three things make a glass furnace unlike the steel and coke furnaces we cover elsewhere:
Alkali and batch vapour. Sodium sulphate, borates, fluorides and sulphur
volatilise from the melt and condense on cooler surfaces. This is why regenerator top checkers corrode and why a bare microporous board can chalk, soften or react if it is directly exposed. The working refractory is selected for alkali resistance; the backup must be sealed away from the vapour, not left open to it.
Higher sustained temperature. A 1550–1650 °C crown means the backup hot face
runs 900–1100 °C for the whole campaign. That is above the continuous rating of many "1000 °C class" boards and right at the edge of it.
Arch expansion. The crown is a sprung arch that grows as it heats. A rigid board
glued to a moving silica arch without allowance for movement cracks at the joints and lets hot gas track behind it.
None of these is a reason to avoid microporous board. They are reasons to specify it correctly — sealed, correctly graded, and jointed for movement.
Conductivity at real temperature, not at 25 °C
Every board datasheet prints a glorious number at room temperature. It is the least useful figure in the catalogue. Here is a typical microporous board next to the materials it replaces, at the temperatures a glass-furnace backup actually sees (manufacturer datasheets: Microtherm/Promat-class silica microporous; ceramic-fibre blanket per Firebird PCW and industry averages; light insulating firebrick per Chinese refractory makers):
| Material | 25 °C | 800 °C | 1000 °C | 1200 °C | |---|---|---|---|---| | Microporous board (1000 / 1200 °C class) | ~0.023 | ~0.046 | ~0.055 | ~0.070 | | Ceramic-fibre blanket (RCF/PCW) | ~0.030 | ~0.15–0.19 | ~0.22 | (above rating) | | Light insulating firebrick | ~0.10 | ~0.18–0.25 | ~0.30 | ~0.40 |
At 800 °C the microporous board is roughly 3–4× better than ceramic fibre and 4–6× better than light brick; at 1000 °C the gap widens. This is the whole commercial case: at equal thermal resistance the board is a fraction of the thickness, which on a crown or in a confined regenerator chamber is the difference between fitting the insulation and not.
Compare the two board grades in our SLM-1000 vs SLM-1200 comparison guide. For glass-furnace crowns and upper regenerator walls the high-strength nano insulation boards in the 1200 °C class are the safer default, because the backup hot face routinely sits in the 1000–1100 °C band.
What measured retrofits actually delivered
These are vendor- and industry-published numbers; treat them as班次-scale evidence, not a guarantee for your furnace:
Glass melt at 1300 °C, 22 mm microporous backup: shell temperature dropped
159 → 114 °C, bottom 184 → 130 °C, total heat loss −45% (microporousboard.com application note).
Insulation effect by zone (same source): tank bottom −72%, tank walls −87%,
breast walls −79%, regenerators −80%, crown −49% heat loss. Melting rate rose 15–20% for the same energy, average glass temperature +30–40 °C.
Glass-fibre melter breast wall: microporous board + light brick vs all-light-brick
design cut the insulation thickness −40% while improving performance (Zhengzhou Jinshan refractory case, 2026). Lower profile means less structural load on the breast wall — a direct benefit on a tall furnace.
Same board chemistry on a 1150 °C petrochemical cracking furnace (Tianjin
Petrochemical, Gute Energy): +30 mm board took the shell from 95–110 °C to 65–70 °C, saving ~¥300,000/year in fuel. Not glass, but the same physics and the same board — useful as a second data point on payback.
The crown number (only −49%) is the honest one: the crown is the hottest, most expansive surface, so a board helps less there than on a vertical wall — but the crown is also where shell temperature and refractory life matter most, so it is still worth doing.
Sizing the backup layer for a glass furnace
Resist the urge to copy a steel-furnace thickness. Work it zone by zone:
Find the backup hot-face temperature, not the flame temperature. Use the
refractory cold-face temperature for the zone (the table in the second section is a starting point; a thermocouple in an existing furnace is better). This number sets your board grade.
Pick the grade by hot-face temperature with margin. If the backup hot face is
1000–1100 °C, specify the 1200 °C class so you are not running at the limit for a 10-year campaign. Below ~900 °C the 1000 °C class is fine. Our thickness sizing method walks the conduction calc; the datasheet guide explains which numbers on a sheet are real.
Set the shell-temperature or heat-flux target. Glass plants usually target a
safe, touchable shell (often <90–120 °C on accessible surfaces) and a defined wall loss in GJ/t. Solve thickness backwards from that, not from a rule of thumb.
Add a vapour barrier on the hot side. On crowns and regenerator walls, the board
goes on the cold face of the refractory and is wrapped or faced so alkali vapour cannot reach the silica core. A glass-fibre-cloth wrap has been reported intact after ten years of continuous service (microporousboard.com).
Detail the joints for expansion. On the crown, stagger and gap the boards for
arch growth; do not butt them tight. See our joint-gap control note — a hot joint is a thermal bridge that undoes the board.
The mistakes that ruin a glass-furnace install
Specifying on flame temperature. A 1650 °C crown does not need a 1650 °C board;
it needs a board rated for the ~1000–1100 °C backup hot face. Overspecifying wastes money, underspecifying (using an 1100 °C board at 1100 °C backup) burns it out in years.
Leaving the board exposed to vapour. Uncovered microporous silica in a glass
superstructure chalks and degrades. Seal it.
Butt-jointing a moving arch. Crown expansion opens tight joints; gas tracks
behind the board and you lose the insulation exactly where you needed it.
Wetting the board during install. Microporous board is ruined by water — its
conductivity jumps and does not recover. Store dry and keep cutting dust controlled (silica dust is respirable; see our storage and cutting guide). A performance verification step after install catches a board that was soaked.
Supplier checks before you buy
Glass furnaces are unforgiving, so the purchase matters more than the price. Before ordering, run the supplier audit checklist and confirm: declared grade and real continuous-use temperature (with a test report, not a brochure claim), alkali-resistance or sealing provision for the vapour zone, density and compressive strength for the load (crown boards carry less than bottom boards), and lot-to-lot conductivity data.
FAQ
Can one board grade serve the whole furnace? Rarely. Crown and upper regenerator want the 1200 °C class; lower regenerator and bottom are fine on 1000 °C class. Grade by zone.
Is the payback real or theoretical? Published retrofits show shell drops of 45 °C to 45% heat loss and melting-rate gains of 15–20%; on a 1150 °C furnace the same board chemistry paid back in roughly a year of fuel savings. Your furnace, your fuel price, your own number — build it from the sizing method above.
Does insulation hurt crown life? Done right it helps: lower shell temperature and less thermal swing reduce refractory thermal load. Done wrong (tight joints, vapour exposure) it can trap heat against the silica and accelerate wear.
Microporous board vs ceramic fibre on a glass furnace? At 800–1000 °C the board is 3–5× lower conductivity at a fraction of the thickness, and it does not slump or compact the way fibre does at temperature. For thin, high-performance backup it is the stronger choice; fibre still wins where you need a compressible, easily shaped fill.
Get a zone-by-zone spec
A zone-by-zone spec also stops you from overspending on a 1200 °C board where a 1000 °C board would do, and from underspending where it would burn out before the campaign ends. Most glass plants that insulate only the easy walls and leave the crown and regenerator exposed are leaving 30–40% of their structural heat loss untouched.
If you send us your furnace type, the refractory you run, and the shell-temperature or wall-loss target, we will return a graded, zone-by-zone thickness spec with the board grade called out per location. Use the RFQ template to give us the numbers that matter, or see how the board sits in a full anchored ladle back-up assembly and other high-temperature linings we support.
