Elevator shaft pressurization has recently received renewed attention as a means of smoke control in tall buildings. The basic idea is that a fan system floods the shaft with ambient air during a fire, thereby preventing smoke from entering the elevator shaft by creating positive pressure differences across all elevator doors. In the absence of fan pressurization, the driving forces of smoke movement, including the buoyancy of hot smoke and stack effect, can cause smoke flow through an elevator shaft to threaten life at locations remote from the fire.

The International Building Code1 (IBC) states in part:


708.14.2.1 Pressurization requirements. "Elevator hoistways shall be pressurized to maintain a minimum positive pressure of 0.10 inches of water (25 Pa) and a maximum positive pressure of 0.25 inches of water (62 Pa) with respect to adjacent occupied space on all floors.This pressure shall be measured with all elevator cars at the floor of recall and all hoistway doors on the floor of recall open..."


Similarly, for stairwell pressurization systems, Section 909.20.5specifies a range of +0.10 inches to+0.35 inches of water (+25 Pa to +88 Pa) across any (closed) stairwell door when used in conjunction with an automatic sprinkler system. In both systems, minimum pressure differences are imposed to prevent smoke from entering the shaft, whereas maximum values are specified to maintain proper door functioning.


The purpose of this paper is to bring attention to several phenomena that make strict adherence to the 2009 IBC code Section 708.14.2.1 difficult to achieve in elevator shaft pressurization systems in modern, well-sealed buildings. Alternative designs meant to meet the intent of the IBC are not addressed. For example, Ferreira and Klote2 suggest a zero net pressure smoke dilution system. The present authors have been studying such smoke control strategies using the CONTAM software developed at NIST.3 The following results represent an extension of previously published research to address the 2009 modifications to the IBC range of allowable pressure differences across both elevator and stairwell doors.4, 5 Only some of the primary results are presented in this article due to space limitations; additional details of the simulation approach are published elsewhere.4, 5, 6




Two 37-story buildings have been modeled in order to illustrate the system operation. Figs. 1a and 1b show upper floor plans for commercial and residential building models, respectively. Both buildings have additional exterior doors on the ground floors, as well as roofs with stairwell access doors.


The commercial model has an enclosed lobby surrounding the ground floor elevator and stairwell shafts with two lobby doors. For the purposes of this paper, the lobby doors are in the open position. The residential model has an additional garage level below ground.

Both sets of elevators and stairwells are enclosed by lobbies. For the purposes of this study, the garage lobby door is in the closed position. All elevator cars are in the Phase 1 position (on the ground floor with open doors) as required by Section 708.14.2.1 of the IBC. All interior building leakage parameters are provided in Table 1.7



Exterior-leakage-area-to-wall-area values were obtained by correlating the simulation results with experimental measurements of pure stack effect pressures in both a commercial bank building4 and a Korean residential building8 (i.e., the exterior building leakage area was adjusted until the ratio of the across-elevator-door-pressure difference to the theoretical-stack-effect pressure difference between the hoistway and the ambient matched those of the experimental measurements). Separate leakage values are used for the ground floor and the upper floors to match the experimentally measured pressure characteristics. All results are presented for cold day (10F [-12C]) conditions; however, the influence of the ambient temperature is predominantly on the required fan flow rates.


Shaft pressurization is achieved via fans pressurizing each of the elevator and stairwell shafts simultaneously with ambient air while all elevator cars are on the ground floor with open elevator doors (Phase 1 position) and all stairwell doors are closed. The elevator shaft fans are located on the roof, while the stairwell shafts are pressurized from the basement level. A heat transfer model was also derived, 6 which predicts the average shaft temperatures as functions of the temperature and flow rate of ambient supply air from the fans.



As changes in the fan flow rates result in changes in the average shaft temperatures, an iterative approach is required (e.g., increasing the flow rate of cold air into the shaft to achieve a desired pressure difference simultaneously decreases the average shaft temperature). In practice, for each simulation, the fan flow rates and average shaft temperatures are iterated until the minimum pressure difference across any set of doors (including the open ground floor elevator doors) is equal to +0.10 inches of water (25 Pa) for any elevator or stairwell door. The across-door-pressure differences that result from this process are presented in Figs. 2 and 3.



Fig. 2 presents pressure differences across both elevator and stairwell doors for both building models when the systems are calibrated with the exterior building doors opened (as is typically the case). Maximum pressures are only slightly violated for the commercial building elevator doors on the ground floor (Fig. 2a). A nearly vertical profile is observed for upper elevator door pressures due to the large amounts of air needed to overcome the multiple per floor, and relatively large, elevator door leakages, including the open ground floor elevator doors (fan flow rates are provided in the figure captions). Therefore, the elevator shaft temperature is near ambient and the stack-effect pressure gradient is minimized.



In contrast, the stairwells have much smaller leakages and fan speeds. Therefore, they have larger than ambient temperatures and exhibit stack-effect-related pressure gradients. The roof-level-stairwell-door pressure difference is very large due to the pressurized stairwell being in direct connection to the ambient pressure rather than the pressurized building interior. Also, substantially larger stairwell fan speeds are needed when operating in conjunction with elevator pressurization, indicating a strong interaction between the systems (not shown). 4, 5, 6


In contrast, the residential building model pressures are more complex (Fig. 2b). Much larger ground-floor pressure differences are required to produce the minimum +0.10 inches water (25 Pa) pressure differences within the shafts. This is primarily due to the existence of the garage level with its enclosed lobby and closed lobby door, which yields the minimum shaft pressure differences.



In addition, and perhaps even more importantly, the large elevator shaft fans leak air into the building on all floors through the elevator doors. This is observed to produce very large pressure differences of approximately 0.65 in water (162 Pa) across the residential doors on all upper floors (i.e., greater than 70 lbf {310 N}of force on a typical door). These forces are directed from the corridor towards the inside of the residences.


Such large forces could result in either difficulty in opening doors or in injuries resulting from rapid door openings. These forces are also sensitive to the enclosed garage-level lobby door leakage area. If the garage doors open directly to the ambient (i.e., no lobby), the minimum pressures across elevator doors move to the ground floor, and the residential door forces are reduced by more than 50% (not shown). The authors are not aware of any other published study that has examined the effects of elevator pressurization on residential doors.



Although other sections of the IBC address allowable residential door forces, there is no direct mention of this possible interaction pertaining to elevator shaft pressurization in the section.


Another factor found to be important for elevator shaft pressurization is the position of the exterior building doors. Pressurization systems could certainly be required to operate when exterior doors are closed (at night, on cold days, etc.). Although, in practice, elevator pressurization systems are calibrated with the exterior building doors propped open, strict adherence to current IBC language makes no allowance for improper performance if the exterior doors are closed. Fig. 3 presents the (hypothetical) requirements of a system calibrated with the building exterior doors closed for the commercial building model. Greater than 5 inches of water pressure (1.3 kPa) differences are observed on all upper floors (and approximately 50 inches of water {13 kPa} for the roof-level stairwell doors - not shown).


The explanation for this is as follows. Air is forced into the shaft from the roof, and some is lost along the way through the closed elevator doors and into the building interior. However, a relatively large flow rate is needed to achieve the +0.10 inches of water (25 Pa) pressure difference across the first-floor open elevator doors due to their much larger leakage areas (this ground-floor pressure difference is also highly sensitive to changes in the fan speeds).


With the exterior doors closed, this air flowing into the first floor from the shaft has no direct path to escape the building and acts to pressurize the first floor. Therefore, the second floor interior building pressure is much less than on the first floor. However, the pressure within the shaft only varies hydrostatically and is only slightly lower at the second floor In this case, the across-elevator-door-pressure difference is increased substantially on the second floor (as well as on all remaining floors).


Fan flow rate requirements are greater than five times larger than when calibrated with the exterior building doors open. Such large flow rates can cause serious problems on their own because even stairwell pressurization (only) systems can cause doors to slam shut and create difficulty opening stairwell doors during testing. However, the most serious issue is the very large pressure differences observed on all upper floors across both the stairwell and elevator doors (Fig. 3).


Such pressures result in forces in excess of 500 lbf (2000 N) acting on the doors and would certainly prohibit proper door functioning. These large pressure differences are not a direct function of the building height or the stack effect. They are dictated by the number of elevator cars and their associated leakages as the flow rate needed to produce a desired pressure difference is only a function of the leakage area. These pressure issues (described above) are directly related to the well-sealed nature of the ground floor when exterior doors are closed.




The authors have explored many system configurations and have yet to find any that strictly satisfy the pressure limitations of Section 708.14.2.1 of the IBC under all operating conditions and/or positions of the exterior doors (including the use of louvers, vents and/or changes in the fan location). 4 However, at this point it is clear that one primary source of problems with strict adherence to the IBC code language is related to the position of the exterior building door (more directly, to cases where the shaft air has no escape route to the outside ambient and pressurizes the building).


One sensible addition to the IBC language would, therefore, be to require an open flow path to the ambient from any floor to which the elevator cars may be recalled, and perhaps garage level floors as well. This could be accomplished by requiring either automatically opening louvers or doorways during system activation. This would greatly alleviate problems with open-floor-plan buildings such as the commercial building model (Fig. 2a), but not necessarily for more complicated buildings (Fig. 2b).


Richard Miller and Donald Beasleyare with Clemson University.



  1. International Building Code, International Code Council, Washington, DC, 2009.
  2. Ferreira, M. and Klote, J., Rethinking the `Smokeproof' Enclosure, Consulting -Specifying Engineer, January/February, 2011, pp. 28-38.
  3. Walton, G. and Dols, S., CONTAM User Guide and Program Documentation, NISTIR 7251, National Institute of Standards, Gaithersburg, MD, 2010.
  4. Miller, R.S. and Beasley, D.E., On Stairwell and Elevator Shaft Pressurization for Smoke Control in Tall Buildings, Building and Environment, 44, 1306-1317, 2009.
  5. Miller, R.S. and Beasley, D.E., Smoke Control by Pressurization in Stairwells and Elevator Shafts, The Singapore Engineer, 6-11, February 2009.
  6. Bowers, D., Ellison, J. Beasely, D.E., and Miller, R.S., Numerical Study of Elevator and Stairwell Shaft Pressurization Systems Using Detailed Building Models, Proceedings of the Eighth International Conference on Performance-Based Codes and Fire Safety Design Methods, Society of Fire Protection Engineers, Bethesda, MD, 2010.
  7. Klote, J.H. and Milke, J.A., Principles of Smoke Management, ASHRAE Inc., Atlanta, GA, 2002.
  8. Jo, J., Lim, J., Song, S., Yeo, M., and Kim, K., Characteristics of Pressure Distribution and Solution to the Problems Caused by Stack Effect in High-Rise Residential Buildings, Building and Environment, 42, 262-277, 2007.