Almost nobody in a building fire is killed by flames. The design question is not how to put it out — it is whether you can leave before the smoke arrives

After a fire in a hospital, a coaching centre or a commercial building, the questions asked in public are almost always the same: were there extinguishers, was there a fire clearance, did the alarm work. These are reasonable questions and they are rarely the ones that decided the outcome.

The outcome was mostly decided years earlier, on a drawing. How many staircases there are, how far any person is from one, whether that staircase is protected from smoke, and whether the fire can spread past the room it started in — those are architectural decisions, and they determine whether a building gives people enough time.

Because that is what fire safety in a building is: a race between the rate at which a space becomes unsurvivable and the rate at which people can leave it. Equipment matters at the margins. The floor plan sets the terms.

Smoke, not flame

The first correction to most people's mental picture is that fire kills mainly by what it produces rather than by heat.

Burning materials release carbon monoxide, which binds to haemoglobin far more readily than oxygen does and starves tissue of it. Many synthetic materials containing nitrogen also release hydrogen cyanide. Both are incapacitating well before they are lethal — and incapacitation is the fatal event, because a person who cannot think clearly or coordinate cannot escape. The great majority of deaths in building fires are attributed to smoke inhalation rather than burns.

Smoke also removes the ability to navigate. A corridor that was familiar becomes unrecognisable in dense smoke, and visibility can drop to arm's length. An escape route that depends on people finding their way through an unlit, smoke-filled space is not an escape route.

This reframes the design problem. The goal is not primarily to prevent people from being burned. It is to keep smoke out of the spaces people must travel through, for long enough.

The clock runs faster than it used to

Fire in a room does not grow steadily. It grows slowly, then reaches a point where the hot gas layer at the ceiling radiates enough heat to ignite everything in the room almost simultaneously. That transition is flashover, and after it the room is not survivable and the fire begins pushing out into the rest of the building.

The important change is how quickly that now happens. Fire research comparing rooms furnished as they would have been decades ago with rooms furnished in modern materials — polyurethane foam, synthetic fabrics, plastics, engineered board — finds dramatically shorter times to flashover, from tens of minutes down to a very few. The fuel load in an ordinary room has changed character: modern furnishings burn hotter, faster, and produce far more smoke.

The practical consequence is blunt. The available escape time in a contemporary interior can be a matter of a couple of minutes from ignition, and the fire brigade will not arrive inside that window. Whatever saves people has to already be in the building, built into its geometry.

What compartmentation is for

Buildings are designed as a set of boxes intended to hold fire in place for a rated period — an hour, two hours — rather than as open volumes. Walls, floors and doors with fire resistance ratings exist so that a fire starting in one unit stays in that unit while everyone else leaves.

This works only if the boxes are actually closed, which is where real buildings fail. Every pipe, cable tray, duct and conduit that passes through a fire-rated wall punctures it, and each penetration must be sealed with fire-stopping material that restores the rating. In practice, services are added and altered over a building's life by people who have no reason to know what the wall was rated for, and the compartment quietly stops being one. False ceilings and vertical service shafts are the classic route by which fire and smoke travel past every barrier on the drawing.

The staircase is the part that matters most. In a protected staircase the design intent is that it stays clear of smoke — through fire-rated enclosure, self-closing fire doors, and in taller buildings positive pressurisation, where fans hold the stair at slightly higher pressure than the floors so that smoke cannot enter when a door opens.

All of which is defeated by a wedge. A fire door held open for convenience — because it is heavy, or because the corridor is stuffy, or because trolleys pass through constantly — converts the protected staircase into a chimney. In a tall building the stack effect makes this worse, drawing smoke upward through any open vertical path with considerable force. The most expensive fire engineering in a building can be undone by a doorstop, and this is among the most common findings in post-fire investigations.

A building does not fail a fire test at the moment of the fire. It fails when a fire door is wedged, an exit is padlocked, a corridor becomes storage, or a compartment wall is drilled through for a cable.

Two ways out, and they must be genuinely two

Codes require more than one means of escape, and the requirement has a precise logic: a single route can be blocked by the fire itself. Two exits are only meaningful if they are independent — remote from each other, so that one incident cannot take both.

Several related rules follow from the same reasoning. Travel distance is limited, so no one is ever too far from a protected route. Exit width is set by how many people the space holds, since flow rate through a door is finite. Exits must discharge to open air, not into another part of the building. Doors on escape routes open in the direction of travel, because a crowd pressing against an inward-opening door cannot open it. And escape routes must remain clear, which is the requirement most often violated in daily use, since a corridor is convenient storage and a stair landing is convenient parking for equipment.

In tall buildings, lifts are excluded from evacuation planning for good reasons — shafts carry smoke, power may fail, and a lift may open onto the fire floor — which is why high-rise design relies on protected stairs, refuge areas at intervals where people can wait in safety, and dedicated firefighting lifts with their own protected shaft and power supply.

People behave differently from how plans assume

Fire engineering has a well-studied and frequently ignored finding: the time between an alarm sounding and people actually beginning to move is often longer than the time it takes them to walk out.

That gap is pre-movement time, and it is spent doing entirely human things — deciding whether the alarm is real, since most alarms most people have heard were drills or faults; finishing a task; collecting belongings; looking for family or colleagues; seeking someone with authority to confirm. In unfamiliar buildings people also tend to leave by the route they came in, even when a nearer exit is signposted, because that is the route they know.

This has direct design and management consequences. Alarms that convey information reduce hesitation more than louder sirens do. Trained staff who direct people shorten pre-movement dramatically. Drills work primarily by removing the "is this real" delay. And exits that are visible and used routinely are used in an emergency, whereas an emergency exit kept locked and unused for years will not become familiar at the moment it is needed.

The pattern that recurs

Indian fire tragedies, in coaching centres, hospitals, restaurants and commercial buildings, tend to repeat a recognisable combination rather than presenting novel failures: a single staircase serving the whole building, often the only route out. Exits locked or obstructed, sometimes for security reasons. Additional floors or mezzanines added after approval, altering occupant numbers and travel distances the design never anticipated. Combustible interior finishes and decorative materials. Basements occupied for uses the design did not contemplate. And external cladding whose core is combustible — the mechanism that turned a single-flat fire at Grenfell Tower in London in 2017 into a whole-building catastrophe, using panel types that were widely installed in many countries.

India's National Building Code sets out fire and life safety requirements in detail, and state fire services issue clearances against them. The recurring failure is not usually an absence of rules but a gap between the approved drawing, the building as constructed and the building as it is used five years later — which is a question of enforcement, alteration control and maintenance rather than of design standards.

Why it matters for students and researchers

Fire safety sits exactly where architecture stops being an aesthetic discipline. It is one of the clearest cases in which a plan-level decision — where the stairs go, how the compartments are arranged, which spaces have two ways out — determines outcomes that no later intervention can recover.

The research questions are practical and locally underexplored. Evacuation behaviour in Indian building types and cultural contexts, since most modelling assumptions derive from studies elsewhere. Fire performance of the materials actually used in Indian interiors and facades. The dynamics of fire in informal and incrementally extended buildings, which no code anticipated and which house very large numbers of people. Egress design for hospitals, where a substantial proportion of occupants cannot self-evacuate and the strategy must be horizontal movement to an adjacent compartment rather than exit from the building. And the retrofitting question that applies to most of the existing stock: what can meaningfully be improved in a building that cannot be redesigned.

That span — building design and construction, design management and coordination, facilities management and related areas of architectural engineering — is the scope of the International Journal of Architectural Design and Planning (ISSN 2583-8903), a peer-reviewed hybrid open-access journal launched in 2023. For architecture and civil engineering students, fire is worth studying early, because it is the part of the brief where a drawing decision and a survival probability are the same decision.

Frequently asked questions

What actually kills people in building fires?

Predominantly smoke rather than flame. Carbon monoxide and, from some synthetic materials, hydrogen cyanide incapacitate people before they are lethal, and dense smoke removes the visibility needed to find an exit.

What is flashover?

The point at which the hot gas layer under a ceiling radiates enough heat to ignite everything in the room almost simultaneously. After flashover the room is not survivable and the fire begins spreading into the rest of the building.

Why is escape time shorter in modern buildings?

Because contemporary furnishings — polyurethane foam, synthetic fabrics and plastics — ignite faster, burn hotter and generate far more smoke than older materials. Testing shows times to flashover falling from tens of minutes to a very few.

Why does a wedged-open fire door matter so much?

Because the protected staircase depends on being sealed from smoke. An open fire door lets smoke into the escape route, and in tall buildings the stack effect then draws it upward, which can make the only escape path unusable.

Why do people delay before evacuating?

Because they first try to interpret the alarm, finish tasks, gather belongings or locate others. This pre-movement time often exceeds the actual walking time, which is why informative alarms, trained staff and regular drills matter more than louder sirens.

Why are lifts not used for evacuation?

Because lift shafts carry smoke, power supply may fail, and a lift may open onto the fire floor. High-rise designs instead use protected staircases, refuge areas and dedicated firefighting lifts with separate shafts and power.