| SUMMARY This Technical Note presents the results of measurements of the air leakage of twelve large office buildings. Various different construction types have been tested, with four of the buildings being naturally ventilated and the remainder air-conditioned. Ideally air-conditioned buildings should have minimal air infiltration and naturally ventilated buildings should have air infiltration under occupant control. The results of this research indicate that for relatively large office buildings in the United Kingdom there are openings left in the structure which are on average equivalent to 5 square metres (54 square feet). The average normalised leakage of these office buildings was 21.80 m3.hr-' .m-2 at an envelope test pressure of 50 Pascals. The naturally ventilated buildings tended to be tighter than air-conditioned buildings and pre 1990 buildings tended to be tighter than post 1990 buildings. For comparison the average air leakage of large American office buildings is 8.25 m3.hr-1.m-a2n d large Canadian office buildings, post retrofit where applicable, is 6.76 m3. hr-.m' -2. The average air leakage of UK factory/warehouse buildings is 35.86 m3.hr-l .m-2 and the average air leakage of Swedish factory/warehouse buildings is 4.37 m3.hr-'.m-2. BSRIA experience with buildings with an airtightness specification, has idicated that it is possible for UK office buildings to attain an air leakage less than 3.0 m3.hr-'.m-2,s uperstores less than 3.0 m3.hr-'.m-2,a rchival storage units less than 2.0 m3.hr-' .m-2 and cold and chill stores less than 1 .O m3. hr.' .m-'. Whole building predicted infiltration rates exceed 1 air change per hour in nine of the buildings for wind speeds of 12.6 m.s' ('Strong breeze' - Beaufort force 6) and above. In two buildings the air change rate is nearly three per hour for these wind speeds. Individual offices located on the windward side of such buildings will experience 'fresh' air change rates many times higher than those predicted for the entire building. BSRIA recommends that all new buildings should have a maximum air leakage of 10 m3.hr-'.m-2 and air-conditioned or low energy buildings should have a maximum air leakage of 5 m3 .hr-l .m-2 at a test pressure of 50 Pascals. Office Air Leakage of Contents CONTENTS 1 INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 2 MEASUREMENTPROTOCOL . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 2.1 DESCRIPTION OF TEST PROCEDURE . . . . . . . . . . . . . . . . . . 2 2.2 DATA PROCESSING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 2.3 SMOKE TESTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 2.4 ON-SITE REQUIREMENTS . . . . . . . . . . . . . . . . . . . . . . . . . 4 3 DESCRIPTION OF THE BUILDINGS TESTED . . . . . . . . . . . . . . . . . . 5 4 RESULTS OF THE BUILDING TESTS . . . . . . . . . . . . . . . . . . . . . . . 13 5 PREDICTION OF VENTILATION RATES . . . . . . . . . . . . . . . . . . . . . 15 5.1 WIND DATA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 5.2 BUOYANCY PRESSURE . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 5.3 INPUTDATA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 5.4 RESULTS OF PREDICTIONS . . . . . . . . . . . . . . . . . . . . . . . . 16 6 PREDICTION OF VENTILATION HEAT LOSS . . . . . . . . . . . . . . . . . 19 7 AIR LEAKAGE PATHS IN NEW OFFICE BUILDINGS . . . . . . . . . . . . 20 8 CONCLUSIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 9 REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 LIST OF TABLES 1 Results of the Measurements carried out on the Office Buildings . . . . . . . . 14 2 Predicted ventilation rates during typical spring temperatures . . . . . . . . . . . 18 LIST OF FIGURES 1 Normalised air leakage of offices . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 2 Predicted air change rates: based on actual leakage . . . . . . . . . . . . . . . . . 8 3 Average predicted ACR’s for different air tightness targets . . . . . . . . . . . . 9 4 Comparison of two 20. Wrn3 office buildings . . . . . . . . . . . . . . . . . . . . 10 BSRIA Introduction Air BLueiatkdmagees in Office l INTRODUCTION This report presents the results of measurements of the air leakage of twelve large office buildings. Various different construction types have been tested, with four of the buildings being naturally ventilated and the remainder air-conditioned. The measurement protocol is presented in Section 2 and the details of the buildings tested are presented in Section 3. The results of the air leakage tests are given in Section 4. The resultant ventilation rate has been predicted for all of the office buildings using a BSFUA air infiltration calculation model. The air infiltration rates have been calculated for all buildings tested at eight wind speeds, at three wind directions and for three internaVexterna1 temperature differences and the results are presented in Section 5. The variation in air leakage characteristics of office buildings varies from marginally acceptable to worse than standard UK factory buildings. Therefore, for two similarly sized office buildings the predictive model was run for all anticipated wind speeds in increments of 1 m s ' and all expected external temperatures in 1 deg C intervals to provide the ventilation heat loss of each building. The energy penalties for poorly constructed buildings in terms of airtightness are presented in Section 6. BSRIA 1 Measurement Protocol Air Leakage in Office 2 MEASUREMENT PROTOCOL The Ventilation & Special Projects Section designed the "Fan Rover" to assess the ventilation heat loss in factories and warehouses [l1. This same equipment was used to measure the air leakage characteristics of large office buildings in this project. This section describes the general method used, which is in accordance with the draft IS0 standard (June 1990) for determination of building airtightness by the fan pressurisation method. 2.1 DESCRIPTION OF TEST PROCEDURE The air leakage characteristics of the building were determined using an air pressurisation technique. This technique requires air to be supplied to the building at a variety of air flow rates, along with the measurements of the resulting pressure differential across the building. This pressure differential and measured air flow rate can be related by the equation:- Q = k,(dp)" Where: . Q is the air flow rate supplied to the building m3.d dp is the pressure differential across the building Pa the iks air leakage coefficient m3.s-l.Pa-" n is an exponent normally between 0.5 and 1 .O. BSRIA developed this pressurisation technique to assess the air leakage of large buildings using the "Fan Rover". This equipment consists of a mobile fan unit mounted on a trailer, and driven using the rear power take-off of a Land Rover, thus alleviating the need for any intrusion into a building's electrical system. The unit is designed to supply up to 30 m3.s1, and has a lower measurable air flow rate of 5 m3.d using the standard built-in flow grid. A special flow grid consisting of two tubes across the unit incorporates total pressure holes spaced at Log-Chebycheff intervals. The unit was calibrated at BSRIA using standard anemometric and tracer gas techniques. The pressure differential across the Flow Grid and the building are both measured using Furness Controls Type FC014 Micromanometers regularly calibrated by the manufacturers, augmented by regular calibration checks by BSRIA. Lower flow rates are measured by attaching a Wilson flow grid to the rear of the fan unit. Throughout the test periods, air temperatures were measured using type UU thermistor probes connected to a Grant Squirrel data logger. All air temperatures were recorded at 2 minute intervals. The accuracy of these probes is better than f/-0.2"C. The internal temperature was averaged for the period of each test to provide a mean internal air temperature. Similarly, the external temperature was averaged throughout the period of each test. 2 BSRIA Air Buildings Office in ProtocMole a surement In addition, the local wind speed and direction was measured at a height of 10 m throughout the period of the tests, and recorded at intervals of 2 minutes and averaged for each test period. The wind speed and direction were measured with a Porton anemometer and windvane mounted on the top of a Clark telescopic mast which was secured to the rear of the Land Rover. If the local wind speed exceeds force 3 or 4, then it is not satisfactory to carry out the air leakage test. All buildings were tested with all external doors and windows closed and with all internal doors wedged open. The test procedure was carried out with all mechanical ventilation openings sealed with polythene sheet and self-adhesive tape. Air inlets and outlets only were sealed and not smoke extract fans, grilles or flues. 2.2 DATA PROCESSING The results of measurements directly associated with air leakage tests were initially verified 'on-site'. This consists of converting the pressure difference across the flow grid into an air flow rate using the required calibration. The pressure difference across the building versus the measured air flow rate were plotted on a lap-top computer, and the slope determined. This also provides confirmation that
the relationship between these parameters was generally as expected and that a
door or window had not opened during the tests. Provided that the envelope
surface area was known, provisional results were determined on-site.
The data was further processed in the laboratory for two factors. The corrections
applied to the air flow rates were firstly an air density correction, determined from
the air temperature and barometric pressure of the flow grid. The second
correction applied was for the change in air temperature as air enters the building,
which results in a change in the air volume. For example, with pressurisation the
outside air passes though the apparatus into the building, and mixes with the inside
air. If the indoor air temperature is higher, the volume increases, and the volume
flow out of the building envelope is slightly greater than the measured air flow
rate. A regression analysis was carried out on the pressure differential across the
building, and the corrected air flow rate to calculate values of 'k' and 'n'. The
correlation coefficient was also calculated to indicate the 'fit' of the data to the
calculated relationship. Using the calculated relationship, the air flow rate
required to pressurise the building to 50 Pa was determined and included with the
data presentation in the building report. The air flow rate required to pressurise
the building to 50 Pascals was normalised with respect to the surface area (S) of
the building, thus yielding values for Qso/S (m3.hr-*.m-2).
2.3 SMOKE TESTS
It is quite helpful to carry out smoke tests to determine the location of the air
leakage paths. Not all buildings in this research programme were subject to a
smoke test but whole building smoke tests can be undertaken for unoccupied
buildings and partial smoke tests undertaken for occupied buildings.
BSRIA 3
Measurement Protocol Air Leakage in Office Buildings
Smoke generators utilising food grade polyglycols were used to identify air
leakage paths. For whole building smoke tests five generators are used to fill up
the entire building and the Fan Rover operated to displace the smoke through the
air leakage paths in the structure. For occupied buildings, the Fan Rover would
be operated to provide a suitable positive pressure within the building and smoke
directed at particular candidate air leakage paths. Video recordings are usually
taken of the smoke egress from outside the building and demonstrate quite clearly
problem details.
2.4 ON-SITE REQUIREMENTS
The following outlines the requirements to complete the test programme on-site.
1. All mechanical ventilation openings are required to be sealed with
polythene sheet and self-adhesive tape. Smoke extract fans grilles, etc.
should not be sealed.
2. All exterior doors and windows need to be kept closed during the actual
pressurisation tests.
3. Sufficient access to a door is required for the Fan Rover. The overall
length, including ductwork is 15.5 m. A distance of a least 20 metres is
therefore required to manoeuvre the facility into the required position and
be normal to the access door. It is preferable to attach the unit to a roller
shutter door or a double door, although we have the facility to attach the
unit to a single door. The transformation to a single door means that the
maximum flow rate is reduced by approximately 10%.
4. The actual test is usually completed in well under one hour. During this
period no personnel should enter or exit the building. It is preferable if the
building is unoccupied.
5. There should be no moveable objects near the fan unit inside the building,
since they would be displaced by the air flow.
6. The integrity of the structure should be complete for the tests.
7. Envelope areas for the building need to be calculated. These are the areas
of the roof and walls wherever the air sealed surface has been defined.
4
inL ea k agAei r TBesutiel dd intgh e of
3 DESCRIPTION OF THE BUILDINGS TESTED
Building Number 1
Approximate date of construction 1970
Flat roof
Prefabricated panels fitted together using 'Clasp' construction
Single glazed windows, some openable
2 storey
Height of Building
Volume of Building
Footprint Area
Envelope Area
Building Number 2
Approximate construction date
Refurbished
Pitched roof (49")
Solid walls, outer skin brick
Leaded windows, openable
1, 2 and 3 storey
Height of Building
Volume of Building
Footprint Area
Envelope Area
Building Number 3
Approximate date of construction 1991
6.10 m
1,951 m3
308 m2
881.5 m'
Elizabethan (I)
Victorian
8.5 m
14,109 m3
1,486 m'
5,131 m'
Flat roof with pitched mansard (45")
Steel frame
Brick outer cladding wall
Pre-cast panels for inner wall
Metal framed double glazed windows, some openable
Full height central atrium
5 and 6 storey
Height of Building
Volume of Building
Footprint Area
Envelope Area
18.85 m
39,149.2 m3
2,268 m*
8,932.5 m*
@ BSRIA 5
Description of the Building Tested Air Leakage in Office Buildings
Building Number 4
Approximate date of construction 1985 and 1990
Pitched roof between ground and first floor
Flat roof with pitched mansard (40")
Steel frame
Brick outer cladding wall
Blockwork and plasterboard inner cladding wall
Metal framed double glazed windows, fixed
1 and 2 storey
Height of Building
Volume of Building
Footprint Area
Envelope Area
7.0 m
14,855.5 m3
1,650.7 m*
4,457.0 m2
Building Number 5
Approximate date of construction 1963
Flat roof
Pre-cast panels on frame
Metal openable windows with secondary glazing
5 storey with 7 storey central core between wings
Height of Building
Volume of Building
Footprint Area
Envelope Area
14.5 m
16,571.6 m3
1,111 m2
4,508.3 m*
Building Number 6
Approximate date of construction 1991
Pitched roof (17")
Steel frame
Brick - block cavity wall
Metal framed triple glazed windows, with encapsulated blinds
3 storey
Height of Building
Volume of Building
Footprint Area
Envelope Area
10.28 m
10,589.8 m3
1,031 m'
2,689.4 m:
6
Air Leakage in Office Buildings Figures
Figure 1 Normalised air leakage of offices
I RCR14 tested
-.., .,,,.ilated BSRIA tested
mechanically ventilated offices
m naturally ventilated offices
I meehanlcal ventilation unsealed
trickle ventilators ope'n
Non-BSRIA data m Arnericun large offices
EZZ4 Canadian large offices: before retrofit
50
Y c
4
U P
to
c m u o w
Bdtlsh Brltlsh Amerlcon Conadlan
@ BSRIA 7
Figures Air Leakage in Office Buildings
Figure 2 Predicted air change rates: based on actual leakage
l+ mean WINTER 1 I+ mean SPRING I
1.
0.
Wnd Spnd (m.s-')
1.
1.
0.
8
Air Leakage in Office Buildings Figures
Figure 3 Average predicted ACR's for different air tightness targets
1 .o
a
E
c
O
W
E
V
0
z L
0.2 r I I
Actual Leakage
1 0 m3hhr-'m-*
5 m3hr m -1 -2
Offfae Numbw
BSRIA 9
Figures Air Leakage in Office Buildings
Figure 4 Comparison of two 20,000m3 ofice-buildings
"1 t
L
I
r Weather Data
t occurs ,o- 10%
x u rs 1 0-50%
rccurs 5040%
--"rs 80-90%
rs at
least one 1 hour W.U.
E ' 5M
Requirement
+50 kW
+25 kW
-25 .kW
-50 kW
-75 kW
-100 kW
-125 kW
-150 kW
-175 kW
-200 kW
.. .
Wlnd Sprd
10
~.
Q BSRIA
Air Leakage in Office Buildings Descriotion of the Building Tested
Building Number 7
Approximate date of construction 1986
Flat roof
Steel frame
Brick outer cladding
Blockwork inner wall
Metal framed windows double glazed
6 storey
Height of Building
Volume of Building
Footprint Area
Envelope Area
20.0 m
15,360 m3
814 m2
3,328 m*
Building Number 8
Approximate date of construction 1989
Flat roof
Steel frame
Curtain walling on front protruding wings, brick outer cladding on remainder
Central full height atrium
Metal framed double glazed windows, all fixed
5 storey, diminishing floor area with height on front wings
Height of Building
Volume of Building
Footprint Area
Envelope Area
18.30 m
21,008 m3
1,375 m2
4,782.9 m2
Building Number 9
Approximate date of construction 1991
Central sine crt oioonf short pstayglo ed a
Steel frame
Brick - block cavity wall
Full height central atrium
Metal framed windows, double glazed
1 and 3 storey
Height of Building
Volume of Building
Footprint Area
Envelope Area
11.35 m
44,335 m3
3,087 m2
8,810 m2
BSRIA 11
Description of the Building Tested Air Leakage in Office Buildings
Building Number 10
Approximate date of construction 1990
Pitched roof (30")
Steel frame
Brick - blockwork cavity wall
Metal framed double glazed windows, some openable
Height of Building
Volume of Building
Footprint Area
Envelope Area
10.1 m
10,356.8 m3
1,047 m*
2,785.5 m*
Building Number 11
Approximate date of construction 1992
Multi- blocks each with own section of pitched roof (22")
Steel frame
Brick - block cavity wall with insulation fixed to inner skin
Double glazed metal windows openable
Central atrium at entrance
2 and 3 storey
Height of Building
Volume of Building
Footprint Area
Envelope Area
13.60 m
20,379.3 m3
2,291 m2
5,504.4 m2
Building Number 12
Approximate date of construction 1992
Pitched roof (23")
Steel and concrete framed building
Outer cladding decorative blockwork/brick/curtatam walling
Inner cladding blockwork
Full height atrium at entrance
Metal double glazed windows, some openable
2 and 3 storey
Height of Building
Volume of Building
Footprint Area
Envelope Area
10.0 m
17,576.8 m3
1,853 m*
4,723.9 m2
12 @ BSRIA
inAL iera k eage OfficeR Beusuilld tisn gs of the TBeusitled di ngs
4 RESULTS OF THE BUILDING TESTS
Various researchers nationally and internationally present the results of air
pressurisation tests in different ways. The most commonly accepted method is to
express the air leakage of the structure as the air flow rate required to pressurise
the building to 50 Pascals (Qso) divided by the envelope area (S). The results for
the twelve office buildings tested in this project are presented as buildings 1 to 12
in the chart presented in Figure 1. The naturally ventilated buildings are denoted
by solid blue bars and air-conditioned buildings or mechanically ventilated
buildings by solid red bars. The orange 'hats' are the difference between the
mechanical ventilation openings sealed and unsealed. The average air leakage for
these twelve buildings is 2 1.80 m3 .hr-'.m-2. The naturally ventilated buildings
(17.40 m3.hr-' .m-2) tend to be tighter than air-conditioned buildings
(23.98 m3.hr" .m-2). Pre 1990 buildings (17.81 m3.hr-1.m-2t)e nd to be tighter than
post 1990 buildings (24.62 m3.h r-'.m -2). The overriding conclusion is that an
office building can have an air leakage anywhere between 10 and 40 m3h -'.m -2.
Other useful data on airtightness of office buildings are included in the chart. The
data for buildings A, B and C are for fully glazed small air-conditioned buildings.
The naturally ventilated building D was fitted with full window width trickle
ventilators and the yellow 'hat' is the increased air leakage with all the trickle
ventilators open. The yellow section, therefore, denotes the user control over the
base air leakage rate. Ideally, the size of the blue and yellow sections should be
reversed, such that the base air leakage is low and the user control high. Building
E was a new air-conditioned building which had an airtightness specification target
and accordingly tested for conformance. Office Building E is comparable to an
average Swedish factory building.
Other international data is included in the chart for large American 12] and
Canadian [31 office buildings and all average below 10 m3.hr-' .m-2. Five of the
Canadian buildings [41 were subjected to retrofits and the reduction in air leakage
was apparent in four of the buildings as indicated by the green hatching compared
with the original purple hatching. The retrofit had no impact on Canadian
building number 3.
The results of the measurements are also presented in Table 1. The first two data
columns present the air flow relationship between pressure and flow rate. Q= and
Qso are the flow rates required to pressurize the building to the subscripted
pressure in Pascals and N,, and N,, are the equivalent air change rates for the flow
rates required to pressurise the building to the subscripted pressure difference
(Pascals). The predicted average air change rate is also provided for an average
wind speed of 4.38 m.s-', integrated over three wind directions with an external
temperature of 1 1°C. The normalised air leakages of the building (Q25/S and
Qso/S) are provided for test pressures of 25 and 50 Pascals, respectively. The last
column indicates the area of holes left in the structure, which are given in square
feet since most people can visualise this unit more easily. It is not difficult to
conclude that 90 sq. ft. of holes left in an office building is unacceptable by any
standard.
BSRIA 13
Results of the Buildings Tested Air Leakage in Ofice Buildings
Table 1 Results of the Measurements carried out on the Office Buildings
o!
tv)
t- 9
$1
W
\o
N
m c?
W
c?
t-
0
09
m
t-
9
-
W
W
do!
N
md-
\4
-
m
0
Y
S 4"
-
m
0
L?
3
9
N
14 BSRIA
Air Leakage in Office Buildings Prediction of Ventilation Rates
5 PREDICTION OF VENTILATION RATES
The natural ventilation rates have been predicted using the BSRIA designed and
developed "CRKFLO" computer program ['l, which calculates the flow rates
between components, taking into account inside/outside temperature differences
and wind effects on the building. The parameters which influence the natural
ventilation performance of the building, wind effects and temperature difference,
are described in this section, which also includes the results of the predictions.
5.1 WIND DATA
The first element in the prediction of natural ventilation rates is to establish the
wind pressure coefficients for all the surfaces of the building structure. Each
building was split up into twenty six external nodes of influence. The first twenty
nodes characterize the external walls and nodes twenty one to twenty-six represent
the roof sections. These are the node numbers for which a wind pressure
coefficient needs to be assigned. Wind pressure coefficients are the fraction of the
wind velocity pressure at building height. The data was established from an
amalgam of data but principally from 16]. The wind velocity was varied between
zero and 15.7 m.s-' in seven non equal steps. For these wind speeds various wind
directions have also been incorporated into the predictions and an average taken of
the three basic directions, ie 0", '45", 90". The mean wind speed factors were
derived using the Deaves and Harris model equation 17], which requires a
Roughness Category for the terrain upwind of the building. Roughness Categories
for the buildings varied according to location between 3, 4 and 5, which equate to
open country, urban and city centre respectively. This Roughness Category
establishes the aerodynamic roughness and zero-plane displacement values which
are then used in the Deaves and Harris iterative model to calculate the basic
friction velocity and gradient height. This does not hold true where the building
height is less than the zero plane displacement, which does occur in city centre
environments, in which case the CIBSE wind gradient data [*l has been used. Thus
the basic mean wind speed measured over standard meteorological terrain at a
height of ten metres can be translated to the friction velocity and gradient height
upwind of the tested building, which are then used to derive the wind velocity at
building height. The wind velocity at building height and the outside air
temperature (density correction) are then used to calculate the velocity pressure.
The wind pressure acting on a particular section of the building is simply the
product of the velocity pressure and the pressure coefficient.
5.2 BUOYANCY PRESSURE
Under normal circumstances the buoyancy pressures (stack effect) acting between
inside and outside of a building would be a simple function of the absolute
temperature difference between inside and outside of the building. Without wind
forces, and inside air temperatures greater than external air temperatures, air
would infiltrate the lower sections of the building and exfiltrate at higher levels.
BSRIA 15
Prediction of Ventilation Rates Air Leakage in Office Buildings
The buoyancy pressures have been calculated as a function of height of the
building and used as a correction to the wind pressures calculated previously. The
design internal temperature has been assumed to be 21°C for Spring and Winter
conditions and 28°C for Summer conditions. The outside air temperature was set
to -4°C for the Winter design condition, 11°C for the Spring condition and 24°C
for the Summer design condition. For average wind speeds and above, the
dominant pressures tend to be wind effects.
5.3 INPUT DATA
A special front end software package has been developed in order to provide
comparability of results with relative ease and uses subroutines to undertake
sensitivity studies more easily. The following parameters were then input into the
model for each building:
Building volume
Height of building
Roughness category
Designated wind speeds
Internal temperatures
External temperatures
Envelope area
Measured air leakage characteristics of the building
The largest unknown in this study is the distribution of the air leakage paths.
However the percentage air leakage through the roof sections does not in fact alter
the overall air change rate particularly significantly. Wind direction is, however,
quite a dominant factor.
5.4 RESULTS OF PREDICTIONS
The results of the predicted ventilation rates are presented in Figure 2. All twelve
of the buildings are presented on the figure with building number twelve on the
top left. The three curves present the predicted increase in ventilation rate with
wind speed for the three temperature differentials used in the model and signified
Winter, Spring and Summer. The orange histograms for each building present the
percentage occurrence of wind speeds 19] at the location of the building (in practice
the data applies to the nearest Meteorological Office anemograph station to the
building). Building 12 is located in an exposed coastal region of South-East
England and thus high wind speeds are more common than, for instance,
building 4 which was located in Oxfordshire. Tall buildings and exposed locations
lead to higher incidence of high wind speeds and therefore greater ventilation heat
loss. It is therefore more critical to avoid high fabric air leakage rates in such
circumstances. The five poorest buildings (4, 6, 7, 11 and 12) all exhibit high
whole building air change rates at high wind speeds. It should be borne in mind
that whilst the average air change rate appears low, these are the whole building
air infiltration rates, which means that for a given wind direction, all of the
16 BSRIA
of Ventilation Rates
infiltration load will be centred on the windward faces of the building and more
particularly those offices which experience large air leakage paths. The air change
rate in individual offices in such locations will be many times higher than the
predicted values for the whole building. Equally 1 .O air change per hour in a
building of 20,000 m3 is still a significant infiltration flow rate at 5.6 m3.s-'.
Table 2 presents the predicted whole building air-change rates for all the buildings
tested for the range of expected wind speeds in the Spring temperature condition
of 10°C external temperature and 21°C internal temperature. The predicted
ventilation rates are based on the actual measured air leakage of the building and
also with air tightness specifications of 10 m3.hr-'. m-2a nd 5 m3h -'.m -'. The
benefits of reduced fabric air leakage particularly under high wind speeds is
indicated by the reduced predicted ventilation rates. Whole building predicted
infiltration rates exceed 1 air change per hour in nine of the buildings for wind
speeds of 12.6 m. S' ('Strong breeze' - Beaufort force 6) and above. For building
number 4, the air change rate is nearly three per hour for these wind speeds.
Individual offices located on the windward side of the building coupled with large
air leakage paths will experience 'fresh' air change rates many times higher than
those predicted for the entire building.
Ventilation requirements for the occupants are only of significance in naturally
ventilated buildings. Three of the four naturally ventilated buildings were quite
good in terms of airtightness and satisfied the ventilation requirements of the
occupants for average wind speeds and external temperature, based on 10 litres per
second per person. Building 11 exceeded these ventilation requirements by a
factor of five under the same weather conditions. It would be more appropriate
from a design point of view to provide minimum ventilation under more severe
weather conditions and leave the occupants to control the ventilation under less
adverse conditions.
Figure 3 presents the predicted air change rates for different airtightness targets
for mean wind speeds and Spring temperatures. The target for all buildings
should be the top of the orange section at 10.0 m3.hr-' .m-2. The target for airconditioned
or low energy buildings should be the top of the green section at
5 .O m3.hr-' .m-2. If no specification is imposed on the building then the fabric air
tightness is expected to be anywhere between 10 and 40.0 m3.hr-1.m-2T. he red
bars present the actual air leakage of the buildings tested. The standard to which a
building which is actually built is therefore quite variable. Building number 9
with an air leakage of 19.24 m3.h -.'m -2i s around the average for UK office
buildings and the predicted whole building ventilation rate is 0.45 air changes per
hour under average conditions. However, most of the air leakage paths in this
building, as observed by smoke egress, were from the roof and roof/wall joints.
This section of the building formed the supply plenum for the air-conditioning
system on the First floor of the building. The location of the air leakage paths can
therefore have a seriously detrimental effect on the performance of the HVAC
system.
BSRIA 17
Prediction of Ventilation Rates Air Leakage in Office Buildings
Table 2 Predicted ventilation rates during typical spring temperatures.
Office Wind Speed II
Number
Building
Leakage
m3h-lm-2 2.34 1.042.3 8 9.78 6.9152 .6 15.7
1 Actual
0.13 0.01.12 1 5 0.33 0.46 0.74 0.59
10 0.20 0.39 0.02.46 0 0.84 1.08 1.35
0.24 0.02.03 9 0.60 0.84 1.08 1.35
2 Actual 0.26 0.31 0.51
5 0.11 0.13 0.33 0.22 0.61 0.407.7 5
10 0.26 0.02.34 4 0.66 0.94 1.21 1.51
0.78 1.10 1.42 1.76
3 Actual 0.25 0.26 0.30 0.44 0.61 0.78
5 0.13 0.01.11 3 0.18 0.33 0.205.4 0
10 0.22 0.02.12 5 0.37 0.65 0.05.08 0
0.97
4 Actual 0.53 0.65 1.04 1.60 2.27 2.92 3.63
10 0.17 0.21
5 0.11 0.0.91 7 0.27 0.38 0.49 0.60
0.34 0.53 0.75 0.96 1.20
5 Actual 0.29 0.31 0.40 0.56 0.73 0.93 1.15
10 0.19 0.20 0.27 0.37 0.55
5 0.10 0.01.01 3 0.18 0.30 0.02.43 7
0.62 0.76
6 Actual 0.41 0.48 0.59 0.80 1.07 1.36 1.68
10 0.18 0.21 0.25 0.34 0.46 0.58 0.72
5 0.09 0.12 0.01.01 7 0.22 0.28 0.35
7 Actual 0.87 0.88 1.02 1.27 1.66 2.10 2.60
10 0.22 0.22 0.25 0.32 0.42 0.53 0.65
5 0.11 0.11 0.13 0.16 0.21 0.26 0.33
8 Actual 0.25 0.28 0.32 0.42 0.58 0.74 0.92
10 0.21 0.23 0.27 0.35 0.49 0.62 0.77
5 0.11 0.12 0.14 0.18 0.24 0.31 0.39
9 Actual 0.29 0.34 0.45 0.66 0.91 1.16 1.44
10 0.15 0.18 0.23 0.34 0.47 0.61 0.75
5 0.07 0.09 0.12 0.17 0.23 0.30 0.37
10 Actual 0.31 0.35 0.45 0.68 0.93 1.19 1.47
10 0.18 0.20 0.26 0.39 0.54 0.69 0.85
5 0.09 0.10 0.13 0.19 0.27 0.34 0.42
11 Actual 0.69 0.79 0.95 1.44 2.07 2.68 .3.36
10 0.21 0.24 0.29 0.44 0.63 0.82 1.03
5 0.11 0.12 0.15 0.22 0.32 0.41 0.52
12 Actual 0.67 0.74 0.87 1.25 1.68 2.19 2.73
10 0.18 0.20 0.24 0.34 0.45 0.59 0.74
5 0.09 0.10 0.12 0.17 0.22 0.29 0.36
~~ ~~
18 BSRIA
Air Leakage in Office Buildings Prediction of Ventilation Heat Loss
6 PREDICTION OF VENTILATION HEAT LOSS
Two office buildings of similar size (20,000 m3) were selected for more detailed
predictive evaluation. Building 8 with a relatively low fabric air leakage rate of
12.0 m3.hr1.m-w2 as compared with Building 11 which exhibited a relatively high
fabric air leakage rate of 32.63 m3h -'.m -2. The internal temperature of the
buildings was fixed at 21°C and the infiltration model was run for all three wind
directions, for all expected external temperatures in 1 .O deg C intervals and for all
expected wind speeds in 1.0 m.s-' intervals. For this predictive exercise, both
buildings were relocated to Kew for the climatic data set. The results are
presented in Figure 4.
The top left hand graph presents the predicted air change rates for all expected
temperatures and wind speeds for building number 8. The yellow shading presents
the relative occurrence of such weather conditions. The yellow dots present
occurrences outside the ninety percent of the time contour.
The bottom left hand graph presents the data for building number 11 and shows
that high air change rates (greater than 1 .O) cut through the section of high
climatic occurrence. This is precisely what should be avoided. High fabric air
leakage will mean that high and uncontrollable air change rates will occur for very
significant proportions of the time.
These two buildings were selected because they were almost the same internal
volume. An energy comparison could therefore be usefully determined as depicted
by the right hand set of graphs in Figure 4. A ventilation heating load capacity of
100 kW (solid brown line), including people, lighting, small power and solar gain
would be adequate for nearly all climatic occurrences for building number, 8. A
ventilation heating load capacity of nearly 300 kW (dashed red line) would be
required for building 11. The infiltration load per annum has been calculated and
printed in red in the lower right hand corner of each graph. For building 8 the
infiltration load was 861 GJ per annum, whilst for building 11 this load rises to
2439 GJ per annum. The difference between these two buildings is therefore
1578 GJ per annum. This is not however a direct heating requirement since all
internal gains need to be taken in to account but also this data assumes 24 hours a
day, 365 days a year conditioning. However the difference in ventilation heat loss
is not jus, a large energy wastage of several thousand pounds sterling per annum
(based on gas heating costs) but there will be (and are) staff complaints in
building l1 leading to wrangling between consultants, architects and builders, plus
the additional expense of remedial action to correct the deficiencies. Excessive air
leakage in occupied buildings is very expensive to correct.
@ BSRIA 19
Air Leakage Paths in New Office Buildings Air Leakage in Office Buildings
7 AIR LEAKAGE PATHS IN NEW OFFICE BUILDINGS
In comon with UK building practice, office buildings tend to be unique
structures and as such the locations of the main air leakage paths vary
considerably. It is entirely feasible and cost effective to specify new buildings
with an air tightness specification which should be met before handover. Section 2
is a typical method statement for carrying out air tightness tests.
BSRIA has conducted airtightness and smoke tests on many buildings with air
leakage specifications of 10 m3.hr-'. m-2 or lower. This section provides some
general guidance on minimisation of air leakage in office type buildings.
An airtight surface should be defined for the building and this is usually the inner
surfaces of the structure so that ventilation of cavities and roof spaces is
preserved.
Most modem windows are rarely a source of serious leakage. However the cills
inadequately covering exposed cavities can be a significant air leakage path.
Frames usually have a good mastic seal to the structure, but this should be
checked, especially where recesses are involved. Large glazing systems are
usually quite satisfactory but frame to frame and frame to structure details are
sometimes deficient.
Perimeter heating pipes quite often run through a ventilated mullion cavity and
these penetrations should be sealed. Equally all electrical, service and drain
penetrations through the structure or into a cavity should be sealed. Particular
cognisance should be taken of riser shafts and penetrations into and out of the
shaft, which tend to terminate at ventilated roof level.
The major sources of air leakage however tend to be at roof level and roof to wall
joints, especially where they tend to be hidden behind suspended ceilings. Offices
with a concrete slab roof with or without a plant room above tend to be more
airtight and easier to seal. Profiled metal decking type roof sections do leak,
sometimes quite badly, and should be treated with an additional mastic seal on all
end lap and longitudinal joints. Insulation filled plastic bags, tape sealed together
should be avoided since they can rarely be adequately tape sealed at the perimeter
of the structure. The life expectancy of the tape seal can be short since they can
be subject to large temperature variations and mechanical damage can be easily
sustained. Wall to roof section joints are typical major air leakage paths and these
details should be designed with air 'sealing in mind. Changes in roof section also
add to potential leakage through additional joints. It is also worth checking where
the other end of Bison beams end up!
Lift shafts are quite normally vented to atmosphere, which in turn means that the
lift shaft should be air sealed and good lift doors fitted with adequate seals.
Architects have a major role to play in designing out problem details, for instance,
multiple steel beam joints at inner surface walls, multi level roofs (at air seal
20 BSRIA
LeinaA kia rg e Office LBeuaA ikldiarign eg s nePw a itnh s Office Buildings
surface) with high joint lengths, etc. Quality assurance by the builder on site is
singularly the most important factor. Sealing techniques which can be easily seen
and therefore easily quality assured should be preferred and specified. Subcontractors
should also be aware of the airtightness requirements. For instance,
well jointed blockwork and brickwork does not leak sufficiently to be of concern.
Poorly jointed blockwork and brickwork can, and does, leak as do blocks left or
knocked out for services. The use of unsurfaced treated mineral wool or glass
fibre insulation to provide an air seal is wholly unsatisfactory and should be
avoided.
Most of the above is covered by good building practice but ewelope integrity is
all too often overlooked during the design and build process.
@ BSRIA 21
Conclusions Air Leakage in Office Buildings
8 CONCLUSIONS
Ideally air-conditioned buildings should have minimal air infiltration and naturally
ventilated buildings should have air infiltration under occupant control.
The results of this research indicate that for relatively large office buildings in the
United Kingdom there are openings left in the structure which are on average
equivalent to 5 square metres (54 square feet).
The average normalised leakage of these office buildings was 21.80 m3.hr-' .m-2 at
an envelope test pressure of 50 Pascals. The naturally ventilated buildings tended
to be tighter than air-conditioned buildings and pre-1990 buildings tended to be
tighter than post 1990 buildings. The average air leakage of large American office
buildings is 8.25 m3.hr-' .m-' and large Canadian office buildings, post retrofit
where applicable, is 6.76 m3.h -'.m -2. The average air leakage of UK
factory/warehouse buildings [l1 is 35.86 m3.hr-' .m-2 and the average air leakage of
Swedish factory/warehouse buildings [lo1 is 4.37 m3.hr-' .m-2.
Whole building predicted infiltration rates exceed 1 air change per hour in nine of
the buildings for wind speeds of 12.6 m. S-' ('Strong breeze' - Beaufort force 6)
and above. In two buildings the air change rate is nearly three per hour for these
wind speeds. Individual offices located on the windward side of such buildings
will experience 'fresh' air change rates many times higher than those predicted for
the entire building.
It is quite clear that the air leakage of some UK office buildings is unacceptably
high and leads to high predicted ventilation rates at all but low wind speeds.
BSRIA experience with buildings with an airtightness specification, has indicated
that it is possible for UK office buildings to attain an air leakage less than
3.0 m3.hr-'.m-2,s uperstores less than 3.0 m3.hr-'.m-2,a rchival storage units less
than 2.0 m3.hr-'.m-2 and cold and chill stores less than 1.0 m3.hr-'.m-2.
BSRIA recommends that all new buildings should have a maximum air leakage of
10 m3.hr-' .m-' and air-conditioned or low energy buildings should have a
maximum air leakage of 5 m3.hr-' .m-2 at a test pressure of 50 Pascals. |