Exit capacity

Exit capacity is the number of people that an exit, stair, door or corridor can pass in a given time, or the number of people for whom it is designed. It is usually expressed per metre of width, for example as persons per metre per minute or persons per second per metre of effective width. Together with walking speed, exit capacity determines the evacuation time of a building. When many people reach an exit at once, the time they need to pass it, the flow time, usually decides how long the evacuation takes.

Measured flows reach about one person per second through a typical office stair 1,120 mm wide, and up to about 1.3 persons per second per metre of effective width through doors and corridors. Building regulations translate this into simple rules. Examples are a fixed number of persons per metre of door or stair width in the Netherlands, or units of exit width in the United States. Performance-based design calculates flow times directly and adds them to the other parts of the required safe egress time.

Flow, speed and crowd density

The movement of a crowd through an exit is described by three quantities: speed, density and flow. Proulx gives the basic relation: the flow equals the speed times the density times the width of the walkway, with the speed measured along the slope for stairs. Speed depends on density. When there is less than about 0.5 person per m², people walk at about 1.25 m/s, their normal unrestricted speed. At higher densities they slow down, and at 4 or 5 persons per m², comparable to a crowded lift, movement stops.

Flow is therefore small at both low and very high densities and reaches a peak at an intermediate density of around 2 persons per m². Nelson and Mowrer use a linear relation between speed and density for densities between 0.54 and 3.8 persons per m². Below that range people walk at their own pace, and above it nobody moves until the crowd has thinned out. In their method the flow per metre of width reaches its maximum at a density of 1.9 persons per m². The book Introduction to Fire Safety Engineering shows the same pattern with data from the London Transport Board of 1958. The capacity of an escape route rises up to a density of about 4.5 persons per m² and then drops sharply as congestion sets in.

Specific flow against crowd density for corridors and doorways and for a steep and a shallow stair. Based on Nelson and Mowrer (2002), pp. 3-370 to 3-371.

Flow per unit of width is called specific flow. Nelson and Mowrer give maximum values of 1.3 persons per second per metre of effective width for corridors, aisles, ramps and doorways. For stairs the maximum ranges from 0.94 for steep stairs (190 mm risers, 254 mm treads) to 1.16 for shallow stairs (165 mm risers, 330 mm treads). Observations of total evacuations of office buildings gave an optimum on stairs of 1.18 persons per second per metre of effective width, at 2.0 persons per m² and 0.5 m/s along the stair. On a typical stair 1,120 mm wide this meant one person per second, each person occupying slightly less than two treads and descending one storey every 15 seconds.

For design, Proulx refers to the levels of service of Fruin, which range from A, with little chance of congestion, to F, with much congestion. For emergency movement in limited space, levels C, D and E are recommended. These correspond to flows of about 0.5 to 1.4 persons per second per metre of effective width on level walkways and 0.4 to 0.9 on stairs. Ramps with slopes up to about 5 % do not reduce walking speed. An upward slope of 10 % reduces it by about 10 %, and an unusually steep upward slope of 20 % by about 25 %.

Effective width

People walking through an exit route keep a small clearance from walls, handrails and other obstacles, to allow for body sway and to keep their balance. The useful or effective width is therefore the clear width minus a boundary layer on each side. The boundary layers are 15 cm at stair walls and doors, 20 cm at corridor walls, 9 cm at handrails, 10 cm at obstacles and 46 cm in wide concourses. Next to theatre chairs and stadium benches the boundary layer is zero.

Effective width of a door, a corridor and a stair: the clear width minus a boundary layer on each side. Based on Nelson and Mowrer (2002), pp. 3-368 to 3-369.

For stairs this means that the effective width is 300 mm less than the nominal width credited in building codes. A stair with a nominal width of 1,120 mm has an effective width of about 812 mm. The boundary layer concept explains why capacity is not simply proportional to the measured width. In New Zealand, C/VM2 (second edition, 2025) uses the same boundary layer values to calculate the effective width of doors, stairs and corridors.

Calculating flow time

The time needed for a group of people to pass a point in an exit route follows directly from the flow rate. Nelson and Mowrer write it as the number of people divided by the calculated flow, which is the specific flow times the effective width. The book uses the same relation in step 6 of its hand calculation, with the flow capacity per metre of width, the width and the number of people.

$$t_{\text{flow}}=\frac{P}{F_s{W}_e}$$

Symbol	Description	Unit
$t_{\text{flow}}$	flow time: time for the whole group to pass the exit	s
$P$	number of people assigned to the exit	persons
$F_s$	specific flow, the flow per metre of effective width	persons/(s·m)
$W_e$	effective width of the exit (clear width minus boundary layers)	m

The relation is the core of the hydraulic evacuation model. It assumes that all people start at the same moment, that the flow is not interrupted by individual decisions, and that nearly everyone can keep up with the group. Where a route narrows or two flows merge, the outgoing flow cannot exceed the maximum specific flow times the effective width of the next element. If more people arrive, a queue forms. In a time-based analysis, the evacuation time of a space is the larger of the walking time to the exit and the flow time through it. The flow time then becomes part of the required safe egress time.

The book gives a worked example for 300 people leaving a space by four stairs: two of 2.4 m and two of 1.6 m wide. It uses the Dutch flow capacity of 45 persons per metre per minute for stairs, a walking speed of 1 m/s and an equal split of 75 people per stair. A stair 1.6 m wide then passes 1.6 × 45 = 72 persons per minute, or 1.2 persons per second, so the narrow stairs govern. The flow time is 75 / 1.2 = 63 s, rounded up. The first person needs 5 s to reach the stair and the last person 10 s from the stair to the exit, so the evacuation time is 5 + 63 + 10 = 78 s.

Hand calculation of evacuation time for 300 people and four stairs: the 1.6 m stairs govern, giving 5 + 63 + 10 = 78 s. Based on Van Rossum (2027), §7.5.

The same stair can be checked with the values of Nelson and Mowrer. With a boundary layer of 0.15 m on each side, the effective width is 1.3 m. With maximum specific flows of 0.94 to 1.16 persons per second per metre for stairs, the flow is 1.2 to 1.5 persons per second. The flow time for 75 people then lies between 50 and 62 s, close to the book's result. Both methods assume ideal conditions. Nelson and Mowrer stress that real evacuations are usually slower than the modelled time, for example because some exits are overloaded while others are hardly used.

Exit capacity in regulations

In the Netherlands, building regulations have set requirements for the width of escape doors since 1992. After the limit for the flow capacity of escape doors was questioned, TU Delft carried out laboratory experiments in 2009. In 13 of the 16 experiments the average measured capacity of a doorway was higher than 135 persons per minute. Lower values were found with an open door, with a wide opening of 275 cm, and with a group that included people who could not move unaided. With 5 % of such people, the capacity was 10 % below 135 persons per minute, the lowest result.

The results were included in the Dutch building decree (Bbl), which expresses flow capacity in persons per metre of clear width. The most common value is 90 persons per metre for a doorway. A stair has 45 persons per metre when it bridges a height of more than 1 m and 90 persons per metre for 1 m or less. Other passages, including sliding doors, have 135 persons per metre. The book lists these rules as the simplest way to determine evacuation time. More detailed options are a calculation of holding and flow capacity, a hand calculation and evacuation simulation.

In New Zealand, C/VM2 calculates the flow through a door with the same equation as Nelson and Mowrer, with a density of 1.9 persons per m². For a standard door leaf with a self-closer, the flow is capped at 50 persons per minute. This corresponds to a door 0.95 m wide with an effective width of 0.65 m. The cap reflects the way a self-closing door leaf gets in the way of the people passing through it. If the door is held open or has no self-closer, the full flow from the equation applies, and an automatic sliding door counts with its full effective width once it is open. A double door with self-closers on both leaves can pass at most 100 persons per minute.

Flow through a doorway against effective width in C/VM2, with the cap of 50 persons per minute for a self-closing door leaf. Based on Ministry of Business, Innovation and Employment (2025b), pp. 33–34.

C/VM2 also limits how people may be assumed to spread over the exits. A main entrance must be designed for half of the occupants, and without a main entrance no exit may be given more than half of the occupants. Doors on escape routes used by more than 50 occupants must open in the direction of escape.

In the United States, codes have long used units of exit width. The New York City code of 1938 allowed 30 persons per unit of exit width, and the code of 1968 allowed 60. That change effectively halved the required capacity of new buildings. The International Building Code of 2003, as compared in the World Trade Center investigation, allows 0.005 m of exit width per person in fully sprinklered buildings. The New York code, the International Building Code and the Life Safety Code all allow the capacity of each stair to be doubled when two stairs are located in separated areas of refuge. The required capacity is based on the occupant load of the area served, so capacity rules are always linked to occupant density assumptions.

Those assumptions can differ widely from reality. Studies in Canadian office buildings found that each occupant had on average about 25 m² of gross rentable area, whereas codes traditionally assume 9.3 m² per person. Code estimates of the number of occupants may therefore be wrong by a factor of more than two.

Experience from tall buildings

In tall buildings, exit capacity is a matter of time rather than of a single floor. NIST noted that the egress capacity required by building codes follows from calculations for a single floor, independent of the height of the building. It does not consider the time needed for a full building evacuation. In the World Trade Center towers on 11 September 2001, the buildings were only a third to half occupied. The flow of people had slowed considerably some 20 minutes before each tower collapsed, which showed that the stair capacity was adequate that morning.

NIST estimated that with 20,000 occupants per tower, evacuation by the stairs would have taken just over three hours. About 14,000 people might then have died, because the stairs could not have carried that many people in the available time. A fourth stair, which the New York code required because of the assembly spaces at the top of each tower, would probably have reduced this shortfall. During the bombing of 1993 the stairs met the code requirements for exit capacity, but they did not allow all occupants to leave quickly. Stair doors were held open by the crowd, which increased the airflow up the stairs caused by the stack effect.

Limits of capacity calculations

Capacity values describe ideal flows. Nelson and Mowrer distinguish the modelled evacuation time from the actual time and introduce an apparent evacuation efficiency. It accounts for delays caused by wardens directing people, stopping and restarting at merging points, people delaying their own start, and unbalanced use of the exits. Drills confirm the last point. In 1972, a total evacuation of a 15-storey office building by four stairs took between 6.6 and 9.3 minutes per stair, mainly because the demand on the stairs was unequal.

The way people are divided over the exits is uncertain. The book notes that both an equal split and a split in proportion to capacity are very optimistic assumptions. People prefer the route by which they entered the building and routes that look attractive, so some exits may be little used. Capacity also falls when the group includes people who cannot walk unaided. C/VM2 therefore requires the egress analysis to cover the whole length of the escape route, so that the flow is not restricted at some point closer to the final exit.

See also

• Travel distance
• Walking speed
• Protected escape route
• Evacuation modelling

References

• Averill, J. D., Mileti, D. S., Peacock, R. D., Kuligowski, E. D., Groner, N., Proulx, G., Reneke, P. A., & Nelson, H. E. (2005). Occupant behavior, egress, and emergency communications (NIST NCSTAR 1-7). National Institute of Standards and Technology.
• Ferreira, M. J., & Strege, S. M. (2005). Smoke management systems (NIST NCSTAR 1-4D). National Institute of Standards and Technology.
• Ministry of Business, Innovation and Employment. (2025b). Verification Method C/VM2: Framework for fire safety design (2nd ed.).
• National Institute of Standards and Technology. (2005). Final report on the collapse of the World Trade Center towers (NIST NCSTAR 1).
• Nelson, H. E., & Mowrer, F. W. (2002). Emergency movement. In P. J. DiNenno, D. Drysdale, C. L. Beyler, W. D. Walton, R. L. P. Custer, J. R. Hall Jr., & J. M. Watts Jr. (Eds.), SFPE handbook of fire protection engineering (3rd ed., pp. 3-367–3-380). National Fire Protection Association; Society of Fire Protection Engineers.
• Proulx, G. (2002). Movement of people: The evacuation timing. In P. J. DiNenno, D. Drysdale, C. L. Beyler, W. D. Walton, R. L. P. Custer, J. R. Hall Jr., & J. M. Watts Jr. (Eds.), SFPE handbook of fire protection engineering (3rd ed., pp. 3-342–3-366). National Fire Protection Association; Society of Fire Protection Engineers.
• Van Rossum, E. (2027). Introduction to Fire Safety Engineering. Fire Safety Academy.