Diagnostics and Measurements of Inltration and Ventilation

Systems in High-Rise Apartment Buildings

Helmut E. Feustel and Richard C. Diamond, Lawrence Berkeley National Laboratory
The provision of ventilation air for high-rise multifamily housing has plagued retrot practitioners and
researchers alike. How does one determine whether sufcient levels of outdoor air are being provided to
all apartments in a building? And how does one know whether the systems can be retrot to improve their
energy efciency without compromising air quality? We have been studying the air ows and ventilation
systems in high-rise buildings in Massachusetts and in California, and have seen all the horror stories of
poorly functioning systems that are neither efcient nor deliver satisfactory ventilation. Frequent problems
include the imbalance of supply and exhaust air, the lack of an unobstructed path for supply air, differences
in ventilation rates between upper and lower oors and a change in air ow due to seasonal variations in
temperature and wind. Based on our diagnostic tests of air ow and air leakage, which we use with our
multi-zone airow computer simulations, we have characterized some common problems and suggest
strategies to improve the performance of these systems.
Figure 1. Annual electricity consumption per oor for a
INTRODUCTION
12-story apartment building in Pittsburgh, Pennsylvania.
Consumption data have been normalized by mean values.
Because of the difculty in relying on inltration or natural
Error bars show one standard deviation above and below
ventilation to provide adequate air for occupants, multi-
the mean.
story residential buildings often have mechanical ventilation
systems to provide sufcient outside air for comfort and
health. The performance of these systems, however, is often
less than satisfactory, due to poor design, sporadic mainte-
nance, and interactions with both natural inltration and
occupant behavior. A review of the literature on air ow
and air leakage measurements in multifamily buildings in
North America is presented in Diamond et al., 1996.
Relying on either mechanical ventilation or inltration for
providing sufcient indoor air quality has an impact on the
energy consumption of the building. The effect of inltration
on energy use in a typical high-rise apartment building can
be seen directly in Figure 1.
0
0.5
1
1.5
2
0 1 2 3 4 5 6 7 8 9 10 11 12 13
floor
N
o
r
m
a
l
i
z
e
d

E
n
e
r
g
y

U
s
e
The gure plots annual energy consumption per oor in a
12-story apartment building in Pittsburgh, Pennsylvania. The
energy consumption on the lower oors is 28% higher than
the mean, and decreases with height until the next-to-the-
In this study we report on the ventilation air ows we have
highest oor where the consumption is 32% lower than the
measured and modeled in three high-rise apartments and
mean. Energy consumption on the top oor is higher due
discuss implications for both energy efciency as well as
to conduction losses through the roof. The reason for this
occupant health and comfort.
variation in energy use is the inltration of outside air due
to stack effectbecause of pressure differentials caused by
inside-outside temperature differences. Air in the building
BUILDING DESCRIPTIONS
rises up the vertical shafts (stairs and elevators) and draws
in colder outdoor air at the base of the building. Apartments
We have focused our attention on three high-rise apartments on lower oors get a greater burden of outdoor air which
buildings, all built around the same period and of similar imposes an energy penalty in winter. The upper units get
size. One of the buildings is in Chelsea, Massachusetts and warm air from below, but the lack of outdoor air to these
units poses an indoor air quality penalty. the other two buildings are in Oakland, California.
Diagnostics and Measurements of Inltration and Ventilation Systems - 1.95
ing is eleven stories and is of poured concrete bearing- Margolis Apartments
wall construction. The apartments are heated by baseboard
heaters supplied by hot water from boilers on the roof. The The Margolis Apartments is a modern, 150-unit high-rise
units have large sliding windows. Supply air is provided apartment building for the elderly and handicapped, located
on each oor by individual units with heating coils and is in Chelsea, Massachusetts, in the greater Boston metropoli-
designed to enter from the hallways under the apartment tan area. The building was designed and built in 19731974
doors. The ventilation air is exhausted through the kitchen and is typical of high-rise construction from that period. The
and bathroomregisters. Four exhaust fans on the roof provide building has thirteen stories and is of steel-frame construc-
the exhaust air ows. tion. The individual apartments have electric-resistance heat-
ers in each room, and double-pane windows and sliding
Westlake West was built in 1977 and has 13 oors. The
balcony doors.
building is of pre-cast concrete panels that were assembled
on site. The apartments are heated by individual fan-coil The building has a mechanical ventilation system, with
recirculating units supplied by hot water from boilers kitchen and bathroom exhaust fans in each apartment vented
assisted by solar collectors on the roof. There is a single air into separate vertical shafts which have additional exhaust
supply fan for the building that delivers air to each oor fans located on the roof. The supply air system for the
through a register in the hallway. The air enters the individual building is provided by a fan and heating unit on the roof
apartments through the 1/2 gap under the apartment doors. that connects to a vertical shaft which has supply registers
The air is heated in winter and is at ambient temperature in the main hallway on each of the oors. Supply air enters
the rest of the year. The apartments have exhaust registers the apartments by a slot under the front door of each unit.
in the kitchen and bathroom, which are driven by 36 roof-
mounted exhaust fans that operate 24 hours per day. The The building is exposed on all sides to the wind. Weather
apartments have large sliding windows in the bedrooms and data from an airport located within 5 km indicate a mean
sliding doors onto the balconies. annual wind speed of 6 m/s with up to 26 m/s wind speeds
in winter. The prevailing winds are from the northwest in
winter and from the southwest from Spring to Fall.
Measurements & Analysis
In December, 1993, the building underwent extensive retro- The measurements and analysis that we are reporting here
ts. New double-pane windows with low-e glass replaced consist of four parts: 1) air leakage measurements of the
the old windows throughout the building. A computerized apartments measured pre- and post-retrot, 2) air ow mea-
energy management system was installed that allowed for surements of the apartments pre-retrot, 3) pressures and
tracking and controlling of the thermostats in the individual ows between the apartments and the circulation areas and
apartments. Efcient light bulbs were installed in the individ- 4) computer simulations of the air ows in the building (only
ual apartments and in the parking areas. A new sprinkler at Margolis) under different weather conditions.
system was installed throughout the building. Improvements
to the abandoned ventilation system were completed a
Air Leakage Measurements at Margolis
year later.
We measured the air leakage in nine apartments, before and
Prior to the window retrot, drafts were a major complaint
after the new windows were installed. The average pre-
expressed by the tenants. Since the retrot, there have been
retrot total effective leakage area for the one-bedroom
according to building managementfewer complaints
apartments was 225 cm
2
and 256 cm
2
for the two-bedroom
about window drafts. There was mention of the windows
apartments. The post-retrot total effective leakage area for
being hard to open for some of the residents, both from the
the one-bedroom apartments was 230 cm
2
and 248 cm
2
for
latching mechanism and the effort needed to lift the double-
the two-bedroom apartments.
hung sash. No problems with condensation on the windows
were reported since the retrot. The northwest-facing units
By way of comparison, Kelley et al. (Kelly 1992) measured
(weather side) continue to be the hardest units to maintain
the air leakage pre-and post-retrot in a high-rise apartment
thermal comfort. Also the second oor units (above the open
in Revere, Massachusetts, a few kilometers north of the
parking areas) continue to be a problem in cold weather.
Margolis apartment. They found an average pre-retrot leak-
age for 17 of the apartments of 904 m
3
/h at 50 Pascals, and
Westlake East and Westlake West a post-retrot leakage of 763 m
3
/h at 50 Pascals, a reduction
of 15%.
Westlake East was built in 1968 under the HUD202 program
of housing for the elderly and handicapped, and has 200 With no difference in leakage areas before and after the
retrot we saw no change in the air ows, with an average units, 158 studio and 42 one-bedroomapartments. The build-
1.96 - Feustel and Diamond
at 50 Pascals of 1175 m
3
/h pre-retrot and 1185 m
3
/h post- the building supply on, but with the local exhaust fan off,
ranged from 50 to 170 cfm, with a mean value of 92 cfm, retrot, a difference that was not statistically signicant. In
viewof this lack of reduction in inltration we were surprised signicantly higher than the design value, (see Figure 2),
2) air ow at the bathroom exhaust register was smaller than that tenants who had previously complained of drafts were
now satised. One explanation is that tenants were pre- the kitchen exhaust ows, and ranged from 40 to 86 cfm,
with a mean value of 53 cfm, 3) With the addition of the viously experiencing down drafts at the window due to cold
surface temperatures, which no longer occur because of the local exhaust fan operating, kitchen exhaust ows reach
values between 170 to 200 cfm (mean 188 cfm) and, new double-pane, low-e windows.
4) the air ow at the bathroom register with the local bath-
We also note that these measurements, both pre- and post- room fan operating (together with the roof top exhaust fan)
retrot, were made in very windy conditionsbeyond the produced 110 to 140 cfm (mean 122 cfm).
limits allowed for standard blower-door tests. While this
problem is not uncommon in low-rise buildings, it is an
Under normal operating conditions, i.e., the local bathroom
even bigger problem in high-rise buildings, where wind
and kitchen exhaust off, the total exhaust ow in the apart-
speeds are often much higher than for buildings at ground
ments would be between 100 and 260 cfm (mean 145
level. Furthermore, the measurement technique being used
cfm). The mean air ow supplied to the apartments from
is based on a reference pressure describing the pressure eld
the corridor was measured at 22 cfm, so on average, under
around the building. In large buildings, it is very difcult
these weather conditions, the apartment would be drawing
to nd a pressure point which acts as the reference pressure
in an additional 120 cfm of outside air through the exterior
for the apartment being investigated. There is also the possi-
wall and windows. This over-ventilation suggests the need
bility that the measurement technique itself, i.e., depressur-
for lowering the roof exhaust fan ow rates.
ization with a blower door, temporarily seals the windows
and distorts the ndings.
Temperatures, Pressures and Flows at
Margolis
Air Flow Measurements at Margolis
We measured the temperature of the supply air at the hallway
Ventilation rates were measured using tracer gas in two
registers for oors 213. They were all in the range of
apartments in various congurations of exhaust ventilation.
2830C (8386F). These temperatures were higher than
With no exhaust ventilation we found typical rates to be
the setpoint in the energy management control system
about 0.2 air changes per hour (ACH). We also measured
(EMCS) for the air supply, which is surprising, but in fact
the leakage from one apartment to another, using tracer
it serves as a more efcient strategy by providing air heated
gases, and found little leakage between apartmentsless
with the gas system than the individual electric units in the
than 4% of the total leakage was to adjacent apartments
apartments and, it avoids cold drafts along the oor.
(Diamond 1996). This was not altogether surprising given
the concrete construction of the building.
Figure 2. Kitchen exhaust air ow with the local exhaust
These ventilation rates are below the recommended 0.35
fans off. Measurements were made at the same exhaust shaft
ACHgiven in ASHRAEStandard 62. Operation of the build-
at oors 5, 10, 11 and 12 at Margolis.
ing supply system and the exhaust systems increased the
ventilation rate to 0.44 ACH. If the mechanical ventilation
systems were operating at their designed ows, the apartment
ventilation rates might well meet the ASHRAE standard
without excess ventilation.
We measured the exhaust air ow from the kitchen hoods
and the bathroom vents using a hot-wire anemometer. The
lter area of the kitchen hood was divided into 5 sub-areas
and an average velocity for each area was determined. From
the air velocity, the ows were then calculated. The air
velocity was also measured at several locations in the bath-
room exhaust register.
The exhaust ows of the seven apartments investigated
0
20
40
60
80
100
120
140
160
180
1 2 3 4 5 6 7 8 9 10 11
Floor
E
x
h
a
u
s
t

A
ir

F
lo
w

[
c
f
m
]
showed the following characteristics: 1) air ow at the
kitchen exhaust register with both the roof exhaust fan and
Diagnostics and Measurements of Inltration and Ventilation Systems - 1.97
We also measured the post-retrot supply air ows at the using the multizone air ow model COMIS, a simulation
tool, developed at Lawrence Berkeley Laboratory, which hallway registers and they were all within a range of
9001300 m
3
/hr (530760 cfm) per oor, with the average calculates air ows based on mass balance calculations for
individual zones (Feustel, 1990). matching the design specication for the supply air ow.
The air velocity in the elevator shaft was measured at the In order to limit the amount of input needed for the simula-
top of the shaft at the oor of the penthouse elevator room. tion model, each apartment was modeled as one zone, assum-
The air velocities ranged between 0.7 and 1.5 m/s with both ing the internal doors to be open. In order to account for
cabs running (regardless of direction) suggesting the air ow the stack effect and the inter-zonal ows between the oors,
is determined more by wind and stack effect than by the all 13 oors were modeled.
movement of the cabs. The air ow at the top of the elevator
shaft during the rst measurement was out of the shaft. The results show, that with wind blowing perpendicular to
the windward side and no stack effect present, air moves The direction reversed later in the day, i.e., down the shaft,
when the wind shifted direction from the northwest to the from the windward side facade through the corridors into
the leeward side apartments. Under the previous conditions northeast.
with no ventilation system present, only a small portion of
Inside the building, the air velocity from the elevator shaft the inltration air is exhausted through the vertical shafts
into the corridor ranged from 2 m/s at the 13th oor down of the exhaust system. Dampers at the apartment level and
to 0.7 m/s at the 3rd oor. The temperature in the elevator on top of each of the shafts restrict the exhaust ow.
was 19
o
C (66F) when the outside temperature was 7C
(45F). The air velocity at the trash chute at the 13th oor, When the building is operating without the mechanical venti-
with the door open, was 4 m/s, upwards, another indicator lation system, the air mass ow distribution for windward
of the stack effect in the building. side apartments on different oors follows a predictable
pattern .With increasing wind speed, the distribution of inl-
The pressures from the stairwells to the hallway follow the tration becomes more pronounced, showing a minimum at
expected pattern of positive pressures to the outside above the third oor and a maximum at the 11th oor.
the neutral pressure level (roughly the midpoint of the build-
ing) and negative pressures below, with the proles of both With a larger inside/outside temperature difference of 20
the north and south towers being similar. The pressure range oC and zero wind speed, the air ow for the windward
from 4 to 8 Pascals is relatively small, due to the rela- apartments decreases with height above ground from 83 m3/
tively mild temperatures outside during the measurement h (50 cfm) on the second oor to zero at the level of the
7C (45F) and the low wind speeds (Figure 3). 11th oor (Figure 4). With increasing wind speed the air
Ventilation Simulations at Margolis
Figure 4. Mass air ow into apartments at different wind
speeds and an inside/outside temperature difference of 20 K
Based on the measured air leakage data from the building
for the windward apartments with the mechanical ventilation
we conducted extensive air ow modeling of the apartments
system off.
Figure 3. Pressure differences between the stair towers and
the hallways at Margolis.
-4
-2
0
2
4
6
8
2 3 4 5 6 7 8 9 10 11 12 13
Story [-]
P
r
e
s
s
u
r
e

D
i
f
f
e
r
e
n
t
i
a
l

[
P
a
]
North Stair Tower
South Stair Tower
2 4 6 8 10 12 14
Story [-]
0
100
200
300
400
M
a
s
s

F
l
o
w

[
k
g
/
h
]
v = 0 m/s
v = 4 m/s
v = 8 m/s
v = 12 m/s
At Tin - Tout = 20K
(0.35 ach)
1.98 - Feustel and Diamond
ow curves show a more balanced air ow distribution until Figure 6. Mass air ow into windward apartments at differ-
ent wind speeds, with no inside/outside temperature differ- the velocity driven air ows override the stack effect. As
the pressures forcing the air ow are additive, the air ows ence and the mechanical ventilation system on.
for any given wind speed are higher if stack pressure is
present.
The air ows for the leeward side are shown in Figure
5. With increasing wind speed the air ow entering the
apartments through the outside wall becomes smaller. The
zero wind speed curve is the same for the windward side
and leeward side. The top oors do not experience any
inltration. Higher wind speeds cause higher negative pres-
sures on the facade, which lower the level for the neutral
pressure. At wind speeds of 12 m/s no inltration occurs at
the apartments facing the leeward side.
Air ows into the apartments are slightly higher when the
ventilation system is in operation. Figure 15 shows the air
2 4 6 8 10 12 14
Story [-]
0
100
200
300
400
M
a
s
s

F
l
o
w

[
k
g
/
h
]
v = 0 m/s
v = 4 m/s
v = 8 m/s
v = 12 m/s
At Tin - Tout = 0K
(0.35 ach)
ows entering the apartments located on the windward side
through the facade for different wind speeds when no stack
effect is present. At low wind conditions, inltration is
almost independent of the height above ground. With higher
Figure 7. Ventilation air ow, supply and exhaust, for Wes-
wind speeds, we see that the inltration ows follow the
tlake East.
wind pressure prole.
Pressures and Airows at Westlake East
The supply air ows, measured at the hallway registers using
a ow-hood, ranged from 66620 m3/hr (39365 cfm) per
oor, with the average, 330 m3/hr (195 cfm). This is below
our rough estimate of the design specication for the supply
air ow (Figure 7). There was a clear difference between
Figure 5. Mass air ow into apartments at different wind
speeds and an inside/outside temperature difference of 20 K,
for the leeward apartments with the mechanical ventilation
system off.
0
100
200
300
400
500
600
700
800
900
1 2 3 4 5 6 7 8 9 10 11
Floor
A
i
r

F
l
o
w

[
m
3
/
h
r
]
Supply Air Flow
Exhaust Air Flow
the high ows on the upper oors (oor 7 and above) and
the lowows on the lower oors (oor 5 and below), indicat-
ing that these systems are not operating correctly.
The exhaust air ows from the corridor were measured at
different oors and the ows ranged from 350490 m3/hr
(205290 cfm), averaging 410 m3/hr (240 cfm). Because
the exhaust ows on the lower oors were often higher than
the supply ows, there was not much possibility for supply
2 4 6 8 10 12 14
Story [-]
0
100
200
300
400
M
a
s
s

F
l
o
w

[
k
g
/
h
]
v = 0 m/s
v = 4 m/s
v = 8 m/s
v = 12 m/s
At Tin - Tout = 20K
(0.35 ach)
air to reach the apartments. Most of the air supplied at one
end of the hallway was exhausted directly out the other end.
Diagnostics and Measurements of Inltration and Ventilation Systems - 1.99
In the East building we observed that a recent weatherstrip- where the occupants block this gap, with towels, rugs, or
other objects. The reasons offered by the residents for block- ping of the apartment doors blocked the path for the supply
air. One of the residents noted that we used to get plenty ing this gap were to block the hallway light to prevent
noise, I dont know, we just do it, to keep out the of airnowwe get nothing. It was evident fromthe discus-
sions with residents that very few were aware of the door cockroachesalthough they arent a problemanymore and
to keep out the cooking smells. slot as the source of ventilation air. Residents indicated
that they keep their windows open, often year round. The
weatherstripping was installed when tenants complained of
The exhaust air ows from the corridors were measured at
cold drafts from the hallwaywhich were likely due to the
different oors, and the ows ranged from 66 to 119 cfm,
unheated supply air after the supply-air heaters failed.
averaging 97 cfm. Little leakage was noticed between the
corridors and the stairwells because of the good weatherstrip-
ping on the stairwell doors. Leakage to the elevator shaft
Pressures and airows at Westlake West
was not measured, but we observed that the openings from
the elevator shaft to the penthouse were minimal. We did
The supply air ows, measured at the hallway registers using
observe signicant airow into the hallway on the upper
a ow-hood, ranged from 320870 m3/hr (190510 cfm)
oors.
per oor, with the average, 570 m3/hr (340 cfm), below our
estimate of the design specication for the supply air ow
Air Leakage Measurements at Westlake East (Figure 8).
& Westlake West
If we assume 16 apartments per oor, and the ASHRAE 62
We made air leakage measurements using a single blower standard for ventilation of 15 cfm per person, then we would
door on four apartments, two in each building. Because the need 480 cfm per hallway (assuming double occupancy) if
buildings were concrete construction, we assumed that there all the supply air were reaching the apartments. We know,
would be little air owbetween apartments. The wind speeds however, that this is not the case, due in part to the exhaust
were very lowduring the morning measurements in the West air from the garbage chute rooms and the air ow to the
building, and while the wind picked up in the afternoon, we elevators. And while the average supply air ow (340 cfm)
were still able to make reasonable measurements in the East was below the ASHRAE standard, several of the oors had
building as well. much higher ow rates that would meet this standard.
East Building: The two East building apartments are 49 The 1/2 undercut beneath the apartment doors appears suf-
m2 and 34 m2 (490 and 340 ft2) in oor area, with volumes cient for the supply air ow, but we noted several cases
of 117 m3 and 82 m3 (3930 and 2730 ft3). The two apart-
ments (#1015 and #503) had air ows at 50 Pascals pressure
difference of 445 and 416 m3/hr (260 and 245 cfm) respec-
Figure 8. Ventilation air ow, supply and exhaust, for
tively. This is roughly 4 air changes per hour at 50 Pascals,
Westlake West
and represents tight construction. This was not surprising
given that this was a poured concrete building. With the
sliding windows taped at the joints, the leakage was reduced
313%. The exhaust registers when untaped added another
3550%.
West Building: The two West building apartments are
roughly twice the oor area of the East building apartments,
both 52 m2 (520 ft2) in oor area, with a volume of 125
m3 (4165 ft3). The two apartments (#826 and #1134) had
air ows at 50 Pascals pressure difference of 1089 and 1038
m3/hr (640 and 610 cfm) respectively, which corresponds
roughly to 8 air changes per hour at 50 Pascals, which is
relatively tight construction. With the sliding windows and
balcony doors taped at the joints, the leakage was reduced
roughly 10%. The exhaust registers when untaped added
0
100
200
300
400
500
600
700
800
900
1 2 3 4 5 6 7 8 9 10 11 12 13
Floor
A
i
r

F
l
o
w
,

[
m
3
/
h
r
]
Supply Air Flow
Exhaust Air Flow
another 10%. The large access panel to the plumbing chase
added another 10%.
1.100 - Feustel and Diamond
On the exhaust side, studies have shown that when apartment
CONCLUSIONS
occupants have local control over bathroom and kitchen
exhaust, they use them less than one hour per day, if at all
In any study of a building as complex as a high-rise apart-
(Shapiro-Baruch, 1993), which makes it difcult to size the
ment it is important to validate the ndings using as many
supply ventilation system. Continuous exhaust ventilation,
techniques as possible. In the case of the Margolis and
however, presents the possibility of over ventilation and
Westlake Apartments we have been fortunate to have differ-
unnecessary use of energy.
ent data sources: leakage measurements, pressure tests and
air inltration measurements which have all been used to
Efforts to improve the energy efciency of high-rise apart-
validate the model. Because comparisons between the model
ment buildings have been frustrated because of the lack of
and measurement data agree well in several areas, such
knowledge on air ows for individual apartments. Ventila-
as similar directions and magnitude of pressure differences
tion rates for individual apartments vary greatly due to
across apartment doors and stairwell doors, we have con-
height, orientation, and wind speed and outdoor temperature.
dence in the simulation results.
Any recommendations for reducing air leakage will have to
take these variables into account, so that efforts to tighten the
shell for energy efciency do not create health and comfort Based on our analysis of the air ow simulations at Margolis
we see that the ventilation to the individual units varies problems for the residents.
considerably. With the mechanical ventilation system disa-
bled (pre-retrot case), units at the lower level of the building
ACKNOWLEDGMENTS
had adequate ventilation only on days with high temperature
differences, while units on higher oors had no ventilation
We would like to thank Bob Nason and the staff of the
at all. Units facing the windward side were over-ventilated
Chelsea Housing Authority, John Jordan and the staff of
when the building experienced wind directions between west
Westlake Christian Terrace, Larry Bravo and John Snell,
and north. At the same time, leeward side apartments would
EUA/ CCS & Citizens Conservation Corporation, Michaela
not experience any fresh airair ows would enter the
Abraham from Oak Ridge National Laboratory, Richard
apartments from the corridor and exit through the exhaust
Karney and Michael Myers, US DOE. Lastly, we would
shafts and the cracks in the facade. With the mechanical
like to thank the residents of the Margolis and Westlake
ventilation system operating, we found wide variation in the
Apartments for their cooperation in allowing us to take the
air ows to the individual apartments.
measurements in their homes.
A fundamental issue here is the design question of how to
This works was supported by the Assistant Secretary for
best supply ventilation to individual apartments in a high-
Energy Efciency and Renewable Energy, Ofce of Build-
rise building. Using the corridor as the supply route has
ing Technology, Building Systems Division, of the U.S.
several challenges, including the control of the temperature
Depar t ment of Ener gy, under cont ract DE-AC03-
of the supply air, the temperature of the corridor, the opening
76SF00098.
from the corridor to the apartment, and the balance between
supply and apartment exhaust.
REFERENCES
A major conclusion from our measurements and simulations
Diamond, R.C., H.E. Feustel, and D.J. Dickerhoff. 1996
is that each apartment has to be supplied with ventilation
Ventilation and Inltration in High-Rise Apartment Build-
air directly. Pressure drops of the system have to be high
ings, Lawrence Berkeley Laboratory Report, LBL-38103,
enough to overcome natural forces to be able to ensure an
Berkeley, California.
even distribution of ventilation air. If ventilation air is sup-
plied directly to the individual apartments, the apartments
Feustel, H.E. and A. Raynor-Hooson. 1990. Fundamentals
should be uncoupled from the rest of the building by tight
of the Multizone Air Flow Model COMIS, Technical Note
apartment doors. This condition not only decreases the
TN 29, Air Inltration and Ventilation Centre, Bracknell,
impact of natural forces on the distribution of ventilation
UK.
air, but also reduces the disturbance to tenants of odors or
noise from other apartments. In winter, supply air has to be
preheated to avoid unpleasant cold drafts. Supply air pro- Kelly, Mark E., McQuail, John E., and OBrien, Robert.
1992. Case Study of Ventilation Improvements in a Multi- vided by vents in the envelope should either be preheated
by heating elements in the vent itself, or be supplied adjacent family Building, in the Proceedings of the 1992 ACEEE
Summer Study, vol. 2, American Council for an Energy- to heating sources. Ducted supply air should be preheated
in the central unit. Efcient Economy, Washington DC, 1992.
Diagnostics and Measurements of Inltration and Ventilation Systems - 1.101
Shapiro-Baruch, Ian. 1993. Evaluation of Ventilation in and Development Authority, Albany, New York, Report
93-5. Multifamily Dwellings, New York State Energy Research
1.102 - Feustel and Diamond