There is a certain type of person who will tell you that they can judge a steak’s done-ness by pressing it with a finger. They’ll jab at it like it owes them money, declare it not ready yet and go back to watching the football.
Steel, in a fire, can be a bit like that steak, in that it can look fine, until it isn’t. When inadequately protected under fire conditions, it can be supporting as it was designed to one minute, and halfway to molten spaghetti a few later.
This is the uncomfortable part. Because while structural steel is often chosen for its reliability, its behaviour in a fire when not appropriately protected is nothing short of dramatic. If it could monologue, it would.
Structural protection is doing the heavy lifting, literally
Approved Document B requires structural elements to maintain their loadbearing capacity for the entire required fire resistance period. This is neither a choice nor a debate.
It does not indicate a preferred method for doing so – just that requirements must be met in order to facilitate safe escape. This choice sits within the fire strategy, which, broadly, relies on the steel staying below its critical temperature long enough for everything else to work. Compartmentation, evacuation, fire service intervention. Should the frame fail before those things can happen, the whole strategy unravels like a cheap cardigan in the tumbledryer.
This is not hypothetical. A number of factors can trigger progressive collapse far earlier than anyone expected, if the structural protection is not up to scratch.
But still, many design decisions rely on a collection of legacy assumptions that make the whole business feel deceptively stable. Until, of course, it isn’t.
The most dangerous assumptions in structural protection
These assumptions can show up anywhere. In designs, and on site:
1. That the test result is exactly the same as reality
The assumption:
If it satisfied the requirements of the relevant standard in a furnace test environment, it’ll behave the same in a real fire.
Where it comes from:
Fire testing is a very rigorous process, designed to be standardised, and reproducible. It gives us key information, like a time interval and limiting temperatures. These both have a key role to play.
Why it’s wrong:
Even for plasterboard partitions, test reports alone are not sufficient evidence. A third party verification (and a classification report) is what you should look for. But for structural protection, it’s even more complex. Steel heats unevenly, introducing internal stress. Junctions and interfaces are often a relative unknown. Services may need to penetrate structural members.
EN 1363-1 allows deformation, but it can’t predict it. A column in a single test may not always represent the on-site environment. If your assumption is ‘same test, same result’, you might be designing in additional, unknown risk. In these complex environments, additional, bespoke testing sheds a light.
2. That the load is the load
The assumption:
We haven’t just accounted for the load, we’ve designed for it.
Where it comes from:
Live load. Dead load. Snow load. And numbers for all of them. Your steel is sized accordingly. Sorted.
Why it’s wrong:
Fire introduces more than heat. It introduces change. Things move, deflect, and redistribute into places that they were not accounted to. A designed load is now something of a moving target.
EN 1993-1-2 provides a framework for reducing load factors under fire, but if the load during fire was underestimated (especially in partially protected assemblies), then you risk discovering that a system rated for 90-minutes of fire resistance buys you about 14 actual minutes before something important gives way.
3. That redundancy will save you
The assumption:
It’s fine, there’s enough redundancy in the structure to allow for local failure.
Where it comes from:
Engineers love robustness. If one element fails, others take the strain. That’s kind of the point.
Why it’s wrong:
Redundancy assumes failure happens slowly. It doesn’t. Under fire, stiffness is lost progressively. And then catastrophically. Buckling doesn’t care about your optimism, and neither does creep. Once one column’s gone, there’s no telling where the rest will go. You’re into cascade territory.
And because fire resistance is usually specified in terms of time, rather than in structural interaction, it’s entirely possible that you could have four members that were protected adequately in isolation, and still end up with collapse.
4. That connections are covered
The assumption:
If the member is protected, the connection must be fine too.
Where it comes from:
Yes, connections are fiddly. They get wrapped over, or boxed around, or just assumed to inherit the protection of the member itself. Out of sight, out of mind.
Why it’s wrong:
Research from testing has shown that connections are thermal hotspots. Bolts heat very quickly. Plate temperatures exceed member averages. The strength of a bolted or welded joint is only as good as its weakest link.
BS EN 1993-1-2 addresses connections, but a lot of specification clauses don’t go beyond acknowledging them. Which is fine, until a gusset plate lets go and the load path heads for the floor.
5. That perfect installation will happen
The assumption:
Once specified, the protection system will be installed as intended.
Where it comes from:
Everyone is acting in good faith. The detail’s been drawn. The boards or coatings are specified. Woohoo! Tick.
Why it’s wrong:
Site conditions and test conditions are not one on the same. On site, frames don’t align. Fixing centres vary. Gaps get taped over instead of sealed. And speaking of seals, well, fire stops are long forgotten. By the time anyone gets around to checking, the system is closed up and the photos are long gone.
This now matters more than ever. If the installation of your structural protection system can’t be verified, it might be carrying a lot of risk. The assumption that protection survives translation from spec to site is what’s landed many a building into remediation.
The plasterboard problem (in brief)
Framed encasement systems using plasterboard are everywhere. They’re cheap(er). Familiar. Visually clean. But they lean very heavily on many of these assumptions - especially when they’re detailed around connections, or installed without specialist oversight.
We’ve covered these in full elsewhere. For now: if your system is using plasterboard and making some assumptions, you might want to check again.
The only way is up
The world isn’t building five-storey concrete frame walk-ups anymore. Not exclusively, anyway. Modern buildings are structurally complex, interface-heavy, and watched by the Regulator like a hawk.
Designers can’t afford to keep specifying like they always have. Old assumptions can’t survive modern scrutiny. Every bolt, board, and bracket matters. Every junction, every abutment, every detail.
Promat does make specialist frameless board systems for structural protection. But, more importantly, we’ve got technical people who’ll pick up the phone and talk through your project challenges, and extensive testing to help bridge the gap between test and site conditions.
Fire doesn’t lie. Steel doesn’t either. But it does fail.
