Most railway fire approval is materials testing: small specimens, standard benches, tidy indices. Seats break the pattern. A seat is a shaped assembly of foam, fabric, interliner and shell whose fire behaviour depends on geometry as much as chemistry, so EN 16989 tests the complete seat, at full scale, in a furniture calorimeter arrangement - including, unusually, a test on a seat that has first been deliberately vandalised.
Why a furniture test exists at all
Vehicle fire scenarios for seats start with ignition of the seat itself - an arson or accident event with a substantial ignition source, not a match. Material-level tests (oxygen index, smoke density) cannot capture how a shaped cushion assembly burns once the cover is breached and the foam is involved. EN 16989 therefore follows the furniture calorimetry tradition: the complete seat is exposed to a defined burner ignition source under an exhaust hood, and its heat release is measured by oxygen consumption, the same principle as the cone calorimeter at room scale. The Motis range registers a dedicated configuration for this method (the OC-16989 code in the manufacturer's catalogue).
What is measured
The headline quantity is MARHE - the maximum average rate of heat emission - which summarises the seat's heat release curve into the single figure the requirement limits are written around. Alongside it, the test records smoke production quantities and the behaviour observations the method requires: flame spread over the seat, dripping and the progression of the fire through the assembly. Testing is performed on the seat as delivered and, in the vandalised condition the method defines, with the cover cut to expose the filling - because a slashed seat in service is a foreseeable condition, and a seat that only behaves when intact is not a safe seat.
How it connects to EN 45545-2
EN 45545-2's seat requirements combine this complete-seat test with material-level evidence for the seat's components. The hazard level of the vehicle programme sets the limits the MARHE and smoke results are judged against. In practice the two layers answer different questions: the materials bench says the ingredients are controlled; the seat test says the assembly - the thing passengers actually sit on - behaves when it burns.
What this means for seat manufacturers
Three practical consequences. First, design for the assembly: cover fabrics, interliners (fire barriers) and foam work as a system, and the cheapest way to pass is usually a barrier that keeps the foam out of the fire for the first minutes, not exotic foam chemistry alone. Second, test the vandalised condition early in development; it is the condition most likely to surprise. Third, treat the seat test as a programme event, not a bench test: full seats, a calorimetry hood facility and witnessed runs belong in the project plan and budget from the start.
Development testing: what to iterate before the witnessed run
Because a complete-seat test is a programme event, development teams de-risk it in layers. Material benches (oxygen index, smoke density on the component materials) screen the candidate fabrics, foams and interliners cheaply. Small assembly tests - cushion-scale mock-ups burned under a laboratory hood - reveal whether the barrier strategy works once seams, zips and geometry are real, because seat fires start at seams and edges more often than in open field areas. Only then does a complete seat go to the full EN 16989 arrangement. Teams that skip the middle layer tend to discover system problems - a seam that opens, a barrier that shrinks away from the cut edge in the vandalised condition - at the most expensive possible moment.
Reading the MARHE result like an engineer
MARHE compresses the seat's heat release history into its worst sustained rate, which makes it a good limit quantity but a lossy diagnostic. The full curve tells the development story: an early sharp peak usually means the cover system failed fast and the foam drove the fire; a late, broad rise suggests the barrier held and the assembly burned as its materials predicted. Smoke indices deserve the same treatment - a seat can pass on heat and fail the programme on smoke, particularly at the higher hazard levels. Specify the development target as the whole curve shape you want, not merely a MARHE number to beat, and keep the pre-test assemblies' curves in the same file as the witnessed result.
What a seat test costs in specimens, and why spares matter
Programme managers are regularly surprised by the specimen count. A complete-seat method consumes complete seats - as delivered and vandalised conditions, typically with repeats - and seats are among the most expensive specimens in fire testing, with production lead times to match. Build the seat order around the test programme from the start: development mock-ups, the witnessed specimens, and spares for the run that misbehaves (a thermocouple failure mid-test on the only seat is a schedule event, not a technicality). Seat suppliers who keep one production seat in reserve per programme sleep better than those who do not.
The vandalised condition, and what it teaches designers
The cut-cover condition deserves one more paragraph, because it changes design priorities in a way material tests never reveal. Once the cover is opened, the fire no longer meets the seat's showroom surface; it meets the barrier and the foam edge directly, and it meets them in a geometry - a slit with raised lips - that can funnel flame into the filling. Designs that survive this condition usually share two features: a fire barrier that wraps and overlaps at seams rather than merely lining flat panels, and foam whose behaviour in the first minutes is slowed by the barrier system rather than asked to perform alone. If your development programme tests only intact seats until the witnessed run, it has not yet tested the seat the standard is most interested in.
Where seat testing is heading
Seat fire requirements continue to tighten their connection between the furniture-scale result and the material evidence beneath it, and test houses report growing demand from programmes outside the traditional European market as EN 45545 spreads through international railway procurement. For manufacturers, the strategic conclusion is stable: design the seat as a fire system from the first sketch, keep the development curves, and treat the witnessed EN 16989 run as confirmation of work already done - never as the experiment itself. Programmes that internalise that order - materials, assemblies, then the complete seat - almost never fail in public; programmes that reverse it occasionally do.
For the surrounding railway programme, see the railway sector page for the material-level instruments, and the EN 45545-2 explainer for how requirement sets and hazard levels organise the whole approval.