1、AN-04-1 1-3 Smoke Control and High-Rise Office Buildings with Operable Windows: Two Case Studies Jeffrey A. Maddox, P.E. ABSTRACT This paper describes the smoke control aspects of two naturally ventilated high-rise ofice buildings with operable windows. One utilizes its operable windows in conjuncti
2、on with an HVAC system to provide tenants with fresh air: The other utilizes a series of automatically controlled windows and vents in conjunction with tenant-controlled windows to provide ventilation. Both buildings, located in San Francisco, were designed under the Uniform Building Code. INTRODUCT
3、ION Due to the moderate San Francisco climate, naturally ventilated high-rise office buildings are becoming popular again in San Francisco. This paper describes the smoke control aspects of two such buildings: one that houses the headquar- ters of a private company, and a second that houses various
4、government agencies (Federal Building). The headquarters building utilizes its operable windows in conjunction with an HVAC system to provide tenants with fresh air. The San Fran- cisco Federal Building utilizes a series of automatically controlled windows and vents in conjunction with tenant- contr
5、olled windows to provide the primary means of ventila- tion for the building. The smoke control systems of both build- ings were designed under the Uniform Building Code with similar goals but significantly different results. OFFICE BUILDING The Gap World Headquarters building is a 15-story office b
6、uilding with an atrium lobby in downtown San Francisco. The building faces the San Francisco Bay on the east while the other three sides face an urban setting. The upper floors receive the primary ventilation through standard HVAC units, which move air through shafts to each level. Supply air is pro
7、vided through an underfloor plenum. Return air is through an above-ceiling plenum. The operable windows consist of3- foot (0.9 m) wide by 2-foot (0.6 m) tall windows that are hinged at the bottom. The windows open a maximum of 4 inches (100 mm). The implicit goals of the system included life safey a
8、nd property protection. The design team evaluated several smoke control options, including a passive smoke control approach, the exhaust method with each floor as a zone, the pressurization approach, Figure I Company headquarters. J.A. Maddox is a senior consultant with Rolf Jensen the requirements
9、to pressurize the stairs and the floor could be balanced to achieve both goals without exceeding the 30-pound (133.5 N) door force. CONTAM 96 To evaluate floor-to-floor and stair pressure differentials, Dr. James Milke of the University of Maryland, acting as the third party peer reviewer, used the
10、CONTAM 96 and W model and the fan capacities found in Table 1. ASH RAE Transactions: Symposia 647 Table 1. Gap Smoke Control Fan Capacities Smoke Control Zone Ground Floor NE Smoke Zone Exhaust Smoke Zone Supply Adjacent Zone Supply 14,000 cfm O cfm O cfm I sw I 17.000cfm I O cfm I o cfin I SE Atriu
11、m - 13,000 cfm o cfm O cfm 95,000 cfm 95,000 cfm O cfm I Parking Levels I ocfm I O cfm I 94.000cfm2 I Floors 2-5 Floor 6 74,000 cfm O cfm 74,000 cfm 74,000 cfm O cfm r74,ooo cfin - 5th noor L44,OOO cfm - 7th floor Floors 7-15 I 44,OOOcfm I o cfm I 44.000 cfm I Floor 8 and 16 Elev. Machine Rooms O cf
12、m 2,000 cfm O cfm In the first iteration, exterior pressure coefficients were determined using theoretical wind behavior and winds from four directions, including the predominant wind direction (the northwest) and the east, north, and west wind directions. Initial modeling showed that the required p
13、ressure differen- tials could be achieved with any wind configuration and any open window configuration. The various scenarios modeled included all windows opened, all windows closed, and the worst-case variable scenario where the windows were opened only on the windward side of the fire floor and o
14、nly opened on the three remaining sides of the floors above and below. As the design progressed, a wind tunnel test was performed on a model of the building to more accurately deter- mine the pressure coefficients on all four sides of the building at various elevations. These pressure coefficients w
15、ere signif- icantly different from the theoretical coefficients in that they showed some turbulence around the building under high wind scenarios. This turbulence caused some pressure coefficients to be changed from positive to negative and to vary between 0.2 to 0.5. Using these new pressure coeffi
16、cients, the building was again modeled with CONTAM. The model showed that adequate pressures could be achieved in all cases when the windows were fully closed or when all of the windows were fully opened. The model showed a 20% failure rate under conditions when only some of the windows were opened.
17、 The majority of these failures were associated with the worst-case window scenario previously described. Although it can be argued that building codes inherently allow for some risk of failure, the City of San Francisco required that the system provide adequate pressures under any and all windiopen
18、 window scenarios. The modeling effort was refined to describe a maximum number of windows opened on each face of the building so that adequate pressures were always achieved. As a result, some floors were permitted to have only two operable windows on each face of the building, while on other floor
19、s, the number was not limited. Floor 2 Maximum Number of Op en in g s Square Feet Per Floor 2 2.88 3 4 I 5 I 10 I 14.4 2 2.88 11 15.84 6 7 I 8 5 7.2 5 7.2 5 9 10 7.2 6 8.64 6 8.64 12 13 I 11 I 6 I 8.64 I 7 10.08 7 10.08 15 I 14 I 7 I 10.08 I 7 10.08 Acceptance Test The building has undergone smoke c
20、ontrol acceptance testing over a period of approximately two years. In addition ASHRAE Transactions: Symposia to confirming the sequence and operation of the system and its components, the testing included hundreds of hours of pres- sure differential testing. Pressure differentials associated with e
21、very zone were tested under no wind and moderate wind scenarios. Moderate winds ranged anywhere from 5 to 20 miles an hour (2.2 and 8.6 ms) as recorded at the start of each days testing. Tests were performed with windows fully closed, fully opened, and in the worst-case position. All of these tests
22、demonstrated that adequate pressure differentials were achieved. The City of San Francisco, however, required that either the system be tested at or near the design wind speed of 29 mph (12.56 ds) or that the CONTAM model be rerun utilizing pressure coefficients measured at the building in conjuncti
23、on with the measured pressure differentials and wind speed at the time of the test. After waiting many months for a windy day, the testing team was able to measure pressure differentials under wind scenarios ranging between 20 and 30 miles per hour (8.6 and 13 4s). Pressure differentials were record
24、ed as real-time wind data were obtained from a local weather station website, measured a few blocks from the building. The wind measurements via the website were compared to wind speed measurements taken at the building at the same time, on a mid-height exterior deck and on the roof of the building.
25、 It is interesting to note that even when a 29-mile-an- hour (12.96 ms) or greater wind was measured at the weather station, the bilding experienced a much smaller wind, typi- cally 10 to 15 miles per hour (4.5 to 6.7 m/s) or less. Even over a period of three or four windy days, it was not possible
26、to measure pressure differentials of every floor at or above the 29-miles-an-hour (12.96 msec) wind condition. However, the numerous measurements showed a trend where the range in pressure differentials was much less than expected under varying wind conditions. The CONTAM model had AWS * LIVE Curren
27、t Conditions Latest Observation From - in San Francisco, CA Temperature (OF) 57.0 54.8 60.3 0.8t Humidity (%) 90.4 87.6 / 93.9 0.6 * Wind (mph) Daily Rain (“) 0.02 0.04 “/h at 00:08 0.00 W 5.0 W 13.6 Pressure (“Hg) 30.26 30.24 30.31 0.01+ Wind Chili: Monthly Rain: Dew Point: 56.4 OF 1.31 “ 54.3 OF M
28、etric I Full Observation I Real-Time I Ma0 Loc ation I Satellite, Historic Observations: 7-r 1 -Week I I-Month I YA Records Forecast for KPIX-TV in San Francisco. CA Click here for the KPIX-TV Extended Forecast IW InstaCams I Cite I ist I ZiD Code Search. I Yatches ti Warn ings I I Pollen Counts I S
29、tormooedia I KPIX WeatherNet Classroom I -1 I Jive Java Map I Extremes I - I live DsQlay Hl et I I How to JO in I Wireless WeatherNet WeatberNet Hom e I National Sites I National Figure 4 Real-time weather report from PIX website. ASHRAE Transactions: Symposia 649 lmflr above-dosed) flr bebw closed
30、L I Courtesy of Dr. James Milke. Figure 5 Range of tested pressure differentials under varying wind speed. predicted changes to pressure differentials as much as 50% from a non-wind to a high wind condition. The actual measurements showed that this variation in pressure differen- tial was less signi
31、ficant, on the order of 20% maximum. With this information, the City of San Francisco accepted the results of the smoke control system testing. FEDERAL BUILDING The new San Francisco Federal Building is an 18-story office building whose upper floors receive their primary venti- lation through operab
32、le exterior windows and vents. The lower five-story portion of the building includes a more typical HVAC design, including HVAC units that supply and return air through a series of shafts. The size, aspect ratio, and construc- tion of the upper floors allow for the wind-driven ventilation system. Op
33、erable vents on the windward side (northwest side) allow wind in, while the leeward windows and vents allow air out. This airflow is intended to move through the mostly open office space under high wind and low wind conditions. A series of relatively large vents are located at the perimeter windows
34、near the ceiling on each floor, while a series of small trickle vents are located near the floor. These upper and lower vents are automatically controlled by the building manage- ment system to achieve the desired temperature and ventila- tion rates. The tenants of the building can also open operabl
35、e windows located on either side ofthe building. These windows can open a maximum of 4 inches (1 O0 mm). The southeast vents and windows open onto the space that is partially enclosed by an exterior scrim. The perforated metal scrim acts to enhance the flow of air to the building and also acts as a
36、sunscreen on the southern exposure. Portions of the exterior scrim are also operable. The scrim optimizes ventilation rates while retaining its sunscreen properties. The ventilation aspects of this design were evaluated by the mechanical engi- neer and Lawrence Berkeley National Laboratorys Commer-
37、cial Building Systems Group utilizing the Comis model. Figure 6 New San Francisco Federal Building floor plan. Pressure coefficients were obtained from wind tunnel tests of the building. The building includes one internal stair, which is pressurized in the smoke control mode, and two external stairs
38、. NFPA 1 O 1 was the applicable code for exiting while the smoke control system and other aspects followed the Uniform Build- ing Code. Smoke Control Approach The design firm in conjunction with the U.S. General Services Administration investigated the possibilities of the pressurization approach an
39、d a passive smoke control approach. The primary goal of the system was to provide adequate life safety for the occupants. Due to the lack of typi- cal HVAC systems, the pressurization approach would have required dedicated smoke control fans and shafts, adding a significant cost. The installation of
40、 shafts and fans would have been counterproductive to the overall energy efficiency goals and architectural goals of the building. Instead, a passive smoke control approach was chosen. In the smoke control mode, the exterior vents, which are controlled by the BMS system, and any doors to barriers wo
41、uld be closed, forming individual compartments on a floor-by-floor basis. Each floor is further compartmented with two-hour fire and smoke barri- 650 ASHRAE Transactions: Symposia ers to form a two-hour lobby, for exiting and other building code compliance reasons. The central two-hour lobby compart
42、ment creates a safe zone for exiting, which comple- ments the passive smoke control approach. The lobby has no operable windows. The internal pressurized stair is accessed from the two-hour lobby. Stair pressurization calculations demonstrated that adequate pressures can be obtained with winter stac
43、k effect, without excessive door forces. The tenant-operated windows are not automatically oper- ated in the smoke control mode. These windows open directly to the exterior on the north side of the building and to the exte- rior scrim on the south side. The operable panels of the south side scrim op
44、en in the smoke control mode, reducing the chim- ney effect of the scrim itself. While this would allow smoke to flow out either the northwest or southeast faces ofthe building, this also creates the chance of re-entrainment through the open windows. The Comis model provided data that re-entrainment
45、 might be possible under low wind scenarios. Under moderate or high wind, the study concluded that smoke would be pushed through and out of the building, without significant re-entrain- ment. Timed egress calculations were performed to analyze the possible effects of the re-entrainment on the overal
46、l level of life safety. Because of the close proximity of the two-hour compartment walls, the egress times to a point of safety are estimated to be unusually short, on the order of half a minute. Given the small-scale size of the window openings, the risk to life safety was deemed acceptable. It is
47、acknowledged that the risk of damage to the building and its contents due to potential smoke spread is somewhat increased because of these open windows. This design explicitly relies on the sprinkler system to address this risk. This design is also more dependent on the proper closing of smoke barri
48、er doors, compared to pressur- ization systems. CONCLUSION A building with operable windows can utilize all of the currently available smoke control schemes, given the appro- priate building geometry and flexibility in the design of the mechanical systems and architectural elements, such as stair- w
49、ays and windows. The size of the fans required for the pres- surization approach need not be significantly above the normal HVAC needs as long as the size of the operable window is rela- tively small. Depending upon the required smoke interface height, the exhaust method could be a good alternative. Passive smoke control methods work well with compart- mented buildings when the exiting systems are designed to compliment the compartments. Some level of additional control interface is likely to either operate the windows or modulate fans. The use of CONTAM or other wind-based models i
