NATURAL VENTILATION IN THE FAMU SCHOOL OF ARCHITECTURE BUILDING: INTENT MEETS REALITY
Walter Grondzik, PE
Florida A&M University
Tallahassee, FL 32307-4200
[see paper citation]
The Florida A&M University School of Architecture building has been celebrated in the architectural press as an exemplar of the fruits of energy-efficiency in design. Reports suggest that a key element of the design intent that led to the buildings trademark form was a concern for natural ventilation. The building is quite distinctive and stands out in a positive way among the more conventional buildings on the Florida A&M campus. In practice, however, the natural ventilation system is not now used and was never really successful in the eyes of the buildings users.
This paper looks at one aspect of the natural ventilation performance of the building -- the logic that triggers a changeover from button-up to natural ventilation mode. A hypothesis was developed that lack of concern for relative humidity may have doomed this system from the start. As an offshoot of a Vital Signs case study elective conducted during spring semester 1999, measurements were made in an effort to test this general hypothesis. This paper presents the results of those measurements and an analysis of their likely meaning.
The building housing the Florida A&M University (FAMU) School of Architecture has received substantial attention over the years, in part because of its well-publicized energy-efficiency features and in part because of its distinctive form (as shown in Figure 1). Although the form is not solely the result of energy considerations -- a philosophy of studio organization and hands-on experimentation also contributed to form-making decisions -- the building clearly speaks something about energy. The question is what? This paper will focus on one aspect of that broad question.
|Fig.1: East (or public) elevation of FAMU
School of Architecture building.
Many who are familiar with the building from afar derive their knowledge from an article published in Progressive Architecture -- An Energy Education. (1) As with most such two-dimensional photographic presentations, the building looks fine on paper and the energy flow arrows (magic arrows) seem reasonable. On site, the building leaks, pigeons home, and steel (lots of steel) corrodes. Many other features of the building are quite wonderful. What, however, about the energy-efficiency elements -- do they work? Well, in general, no. The natural ventilation strategies, which played a particularly prominent role in building design, have been disabled by the building users (students, faculty, and pigeons). Is this a result of user ignorance or user savvy (and by which of these three groups)?
A Vital Signs elective course conducted during Spring Semester 1999 provided an opportunity to explore these (and related) questions about the performance of this building. Although several post-occupancy evaluation studies of the architecture building have been completed, little to no information on measured performance (energy, daylighting, comfort) seems to exist. Further, the building is currently undergoing (from summer 1999 until who knows when) such extensive renovation and expansion that most of its original form and function will be lost forever. Opinions as to whether this is good or bad differ. Nevertheless, knowing this change was imminent, the Vital Signs course focused upon capturing a sense of existing building performance -- including daylighting, thermal chimney conditions, thermal nosebleeds, thermal comfort, and corrosion. As a sidebar to these student-initiated studies, the course instructor/author independently investigated the basis of operation of the natural ventilation system. The results of this small study are presented in this paper.
2. NATURAL VENTILATION SYSTEM
The FAMU School of Architecture natural ventilation system is reasonably simple for a building of its size and complexity. Thermal chimneys that constitute the South fašades of the several finger-like wings of the building (see Figure 2) provide the driving force behind the ventilation system. A signal from a computerized control system automatically opens p-panels (see Figure 3) on one side of a room (such as a studio space) and illuminates an indicator light encouraging occupants to open windows located on the side of the room opposite from the p-panels (see Figure 4).
|Fig. 2: South and West elevations of School of
South-facing walls/roofs are thermal chimneys.
|Fig. 3: P-panels on South interior wall
of a studio -- showing
relative size and location of these ventilation outlets and their
pneumatic actuator mechanism.
|Fig. 4: Windows on North exterior wall of a studio --
relative size and location of these ventilation inlets (roughly half
the window area is operable).
The automated panels are called p panels because of their pneumatic actuators (or so the story goes -- there are competing etymological theories). According to design intent, the windows would be dutifully opened and the open p-panels would vent the room into a dedicated vertical channel in the thermal chimney. The chimney would induce air circulation as a result of a stack effect, cool outside air would be pulled into the room, and the occupants would be comfortable. Figure 5 illustrates this intent in convincing magic arrow format.
|Fig. 5: Magic arrow diagram showing
design intent for
the natural ventilation system in the FAMU School of
Architecture building -- manually-operable windows act as
inlets, automatically-controlled p-panels act as outlets,
the South-facing Kalwall-clad thermal chimney induces air
flow via the stack effect, exhaust occurs via the distinctive
In reality the panels have all been sealed shut (yet another use for duct tape and bits and pieces of old study models), the windows are kept closed, and the occupants opt for air-conditioned comfort (opting not necessarily equaling receiving in this building). For years there has been speculation among students and faculty as to why this situation came to pass. Did the system ever really work? Was it ever given a chance to perform? Did occupants understand the system intent? Are people just spoiled by air-conditioning?
The Vital Signs case study opportunity led to development of a more specific question: what if the basis of operation and design for the system was flawed? This question led to a simple hypothesis: the natural ventilation system was intended to operate solely on the basis of outside air dry-bulb temperature -- ignoring (to its regret) ambient relative humidity.
3. TESTING THE HYPOTHESIS
A methodology was developed to test this hypothesis. Three data elements were required:  exterior air temperature;  exterior relative humidity; and  on-off status of the Open Windows indicator lights. The data for ambient air conditions were already being collected as part of the student case study efforts -- at an interval of 15 minutes over a multiple-week period. The simplest means of recording the status of the indicator light seemed to be to track the illuminance produced by the signal. An opaque box with a Hobo illuminance meter was taped to the wall surrounding one of the indicators and sealed so that the sensor read only the illuminance produced by the signal light. An interval of 6 minutes was selected for this measurement, as 15 minutes seemed too long a span between events and might result in lost data. This measurement arrangement (Figure 6) worked like a charm.
|Fig. 6: Open Windows for Natural
indicator light (upper left); Hobo-box covering
indicator light (upper right); Hobo light datalogger
in the box (bottom center).
Measurements of the Open Windows control status were taken over a 4-week period beginning in mid-March 1999. The number of data points collected was substantial; thus the results from a representative 1-week period are used herein. The primary difficulty encountered in data analysis involved the manipulation of over 1800 data points for each of 3 variables to correlate the 15-minute and 6-minute readings. A spreadsheet was used to produce readable and comparative plots of ventilation system operation. The results are intriguing.
Figure 7 shows Open Windows signal status versus exterior air temperature. Figure 8 shows Open Windows status versus relative humidity. Figure 9 shows Open Windows status versus time. Table 1 presents information on exterior conditions during the Open Windows states.
|Fig. 7: Open Windows signal status versus
dry-bulb temperature: period from 2 PM
March 19 through 2 PM March 23, 1999;
vertical bars indicate signal is on.
|Fig. 8: Open Windows signal status versus
relative humidity: period from 2 PM March 19
through 2 PM March 23, 1999; vertical bars
indicate signal is on.
|Fig. 9: Open Windows signal status versus
time: one-week period from 2 PM on March 18
though 2 am on March 26, 1999; vertical bars
represent illuminance readings in Hobo-box, with
a non-zero reading indicating signal is on.
The data shown in Figures 7 and 8 are not terribly informative as to the control logic that lies behind the illumination of the Open Windows for Natural Ventilation indicator. Figure 7 shows that occupants are instructed (by way of the signal) to open windows when exterior dry-bulb temperatures are between roughly 58 deg F and 75 deg F. Consistency in the use of temperature as the definitive triggering mechanism for the signal, however, is not at all evident from this plot.
The variation in relative humidity values that are seen in Figure 8 during the various Open Windows periods is so great as to suggest that relative humidity is an independent variable not connected to the ventilation strategy. Occupants are instructed to open windows when exterior relative humidity is as high as 90%. These data support the portion of the hypothesis that stated that relative humidity was not considered in the control sequence for the natural ventilation system.
The data in Figure 9 are not terribly useful in trying to establish the basis on which the Open Windows indicator light operates. There is no obvious cyclic pattern; although generally, but not always, there are two blocks of Open signals per day. March 21 was a Sunday, so it is possible there is a weekend schedule programmed into the control sequence. Beyond that, however, there is little evidence in this plot that a repetitive time function (such as a time clock) controls the Open Windows signal. Control based upon exterior air temperature seems the most promising possibility -- as the time-based patterns do not appear to suggest a totally random control function.
TABLE 1: TIME, TEMPERATURE, AND RELATIVE HUMIDITY FOR 1 WEEK OF OPEN WINDOWS SIGNALS
|Date/Time: Open||DBT (F)||RH (%)||Date/Time: Open||DBT (F)||RH (%)|
|03/19/99 19:55:06.0||73.2||43.5||03/22 continued|
|03/19/99 20:10:06.0||72.5||44.5||03/22/99 17:55:06.0||70.4||23.7|
|03/19/99 20:25:06.0||71.8||45.7||03/22/99 18:10:06.0||70.4||23.7|
|03/20/99 08:55:06.0||58.7||83.7||03/22/99 19:40:06.0||66.3||40.3|
|03/20/99 09:10:06.0||59.4||80.0||03/22/99 19:55:06.0||65.6||42.4|
|03/20/99 09:40:06.0||62.2||73.5||03/23/99 10:55:06.0||60.1||56.9|
|03/20/99 18:40:06.0||71.8||59.2||03/23/99 11:25:06.0||62.9||55.0|
|03/20/99 18:55:06.0||70.4||62.0||03/23/99 11:40:06.0||64.9||50.2|
|03/20/99 19:10:06.0||69.7||65.9||03/23/99 11:55:06.0||65.6||45.5|
|03/20/99 19:25:06.0||68.3||71.1||03/23/99 12:10:06.0||67.0||38.8|
|03/20/99 19:40:06.0||68.3||74.1||03/23/99 12:25:06.0||68.3||36.1|
|03/20/99 20:10:06.0||67.0||80.5||Maximum value =||73.1||91.8|
|03/20/99 20:25:06.0||66.3||82.4||Minimum value =||58.0||23.7|
|03/20/99 21:10:06.0||65.6||87.0||Direction of Change||On||Off|
|03/20/99 21:25:06.0||64.9||88.4||and Control "Pairs"||Temp||Temp|
|03/22/99 13:25:06.0||64.9||24.4||Trigger Type||HumR*||Enthl**|
|03/22/99 17:10:06.0||70.4||23.7||* = humidity ratio||lb/lb|
|03/22/99 17:25:06.0||70.4||23.7||** = enthalpy||Btu/lb|
Table 1 isolates those times during the one-week period from March 19 to March 25 when the Open Windows indicator light is on. Specific dates, times, dry-bulb temperatures (DBT), and relative humidities (RH) are arrayed in the table in an attempt to get a better handle on the driving force behind the Open Windows signal. The temperatures encompassed by Open signals during this week range from 58 deg F to 75 deg F. This is not an unreasonable control range; 58 deg is a bit cool for Tallahassee tastes, but 75 deg is a likely maximum inlet temperature for ventilation in a building with moderate internal loads. The range of relative humidities encompassed by Open signals, however, is not reasonable. To ventilate this building with air at more than 60% relative humidity would cause discomfort (and ruined drawings and models) rather than promoting comfort. Humidity ratio and enthalpy as possible control triggers are also examined in Table 1.
4. POTENTIAL FOR NATURAL VENTILATION
An interesting question beyond that involving the control scenario for natural ventilation in this particular building is whether ventilation is really a viable strategy in a commercial building in North Florida. Someone (who may remain anonymous) once wrote: if it werent for the humidity, Miami would be a thermally comfortable place. Well, fine. But there is humidity -- in Miami and in Tallahassee.
Review of a Weathermaker analysis shows that only a few hours during a typical year fall within the bounds of a set of conditions (assumed to be 65 to 75 deg F with a maximum relative humidity of 60%) that would be amenable for natural ventilation of an institutional building. Such conditions occur during only 320 hours of an 8760-hour year -- or 4% of the time. This does not seem a great potential. The thermal chimney, on the other hand, was found (in a student case study) to be generating world-class temperature differentials to support stack-effect ventilation.
Control signals for the School of Architecture natural ventilation system do tell occupants to use ventilation as a cooling strategy without regard to exterior relative humidity. During the period of this study, for example, occupants were told to open windows for ventilation when relative humidity was as high as 90%. This would surely doom the use of such a system for comfort conditioning (when other options are available).
The exact mechanism triggering the Open Windows signals is not obvious. Temperature seems the best guess, although if this is the case the system is out of calibration and inconsistent. Sophisticated triggers (such as humidity ratio or enthalpy) do not look likely. Commissioning the controls for this system might have enabled it to operate in a more rational fashion -- and perhaps have given it a better chance for success.
(1) Anon: An Energy Education: Florida A&M Architecture School, Tallahassee, Progressive Architecture, 66(4), April 1985, pp. 74-77.
(2) Sustainable Buildings Industries Council: Weathermaker, part of the Energy-10 software package, Washington, DC: 1999 (Version 1.3).
This is a slightly modified version (with color photos and graphs) of a paper presented at the 2000 annual conference of the American Solar Energy Society (ASES) -- Grondzik, W.T.: "Natural Ventilation in the FAMU School of Architecture Building: Intent Meets Reality," Proc. 2000 Conference of the Amer. Solar Energy Society (Madison, WI), Amer. Solar Energy Soc., Boulder, CO, 2000, CD-ROM not paginated.
This page was last updated 27 May 2001.