Building and Environment 50 (2012) 190e201

Contents lists available at SciVerse ScienceDirect

Building and Environment

journal homepage: www.elsevier.com/locate/buildenv

Application of a developed adaptive model in the evaluation of thermal comfort in

ventilated kindergarten occupied spaces
Eusébio Z.E. Conceição a, *, João M.M. Gomes a, Nuno H. Antão a, Ma Manuela J.R. Lúcio b
a
Faculty of Sciences and Technology, University of Algarve, Gambelas Campus, 8005-139 Faro, Portugal
b
Vertical Grouping of Schools Professor Paula Nogueira, Rua da Comunidade Lusíada, 8700-000 Olhão, Portugal

a r t i c l e i n f o a b s t r a c t

Article history: In this work the application of a developed adaptive model in the evaluation of thermal comfort in
Received 19 June 2011 ventilated kindergarten occupied spaces is made. The study, conducted in real conditions, in mediter-
Received in revised form ranean environment for cold and warm thermal conditions is made.
27 September 2011
The adaptive model is developed in a Kindergarten, using subjective and experimental measurements.
Accepted 15 October 2011
Three different trained groups of subjects are used. Each one expressed its thermal subjective response,
in different classrooms, for cold and warm thermal conditions.
Keywords:
In the study, the classrooms are ventilated by natural airflow, while the indoor playground is venti-
Numerical simulation
Experimental results
lated by forced airflow. The natural airflow in the classrooms is measured using tracer gas decreasing
Building thermal behaviour concentration. The measurements are made for different window states, from close state to open state.
Adaptive comfort models In both warm and cold thermal conditions the simulation is made in real occupation and ventilation
Thermal comfort conditions. The natural ventilation is promoted, in both thermal conditions, using experimental
Air quality measurements. In cold thermal conditions the forced airflow is done from the greenhouse to the indoor
playground, while in warm thermal conditions the forced airflow is done in the opposite direction. In the
thermal comfort level evaluation the developed adaptive model is used, while in the air quality level
evaluation the carbon dioxide concentration released by the occupants is used.
Ó 2011 Elsevier Ltd. All rights reserved.

1. Introduction According to Ref. [2]; thermal comfort is a subjective response

and is defined as the ‘state of mind that expresses satisfaction with
Currently, buildings are large energy consumers. In most Euro- the existing environment’ Ref. [7]. This definition itself states that
pean countries the buildings contribute approximately 40% of total a specific value cannot be assigned to thermal comfort. People
energy consumption and could approach 50% when accounting adjust themselves to maintain and improve their wellbeing
energy costs with the construction materials [1]. The situation has through physiological, psychological and behavioural reactions to
not changed, however the improvement of construction technol- environmental stimuli Ref. [7]. Being so, the same thermal envi-
ogies and building regulations are increasingly being restrictive to ronment may be perceived differently by different occupants or
make buildings more efficient from an energy standpoint. Their different occupants may perceive the same thermal comfort level at
number has been increasing, using more equipment and having different thermal environments Ref. [23]. It is crucial to understand
higher standards of comfort than it ever had. building occupants’ perceptions of their indoor environment and
According to Ref. [26] the energy consumption in commercial their adaptive behaviour towards it Ref. [28].
and public buildings in the European Union in 1998, closely related Currently, on the area of thermal comfort research are two
to thermal comfort, was 56% (52% for space heating and 4% to possible approaches: they are static heat balance models of human
cooling). Thus, the energy consumption due to thermal comfort body based on laboratory studies and adaptive models based on
corresponds to a significant part of the energy bill of a building and field studies. The static heat balance most known is the lab-based
the operation costs due to energy consumption in the buildings Predicted Mean Vote (PMV) and Predicted Percentage of Dissatis-
have increased considerably, due to the rise of occupant’s expec- fied people (PPD) method developed by in Ref. [19]. This model is
tations for thermal comfort level [24]. described as a function of four environmental variables (air
temperature, mean radiant temperature, air relative humidity and
* Corresponding author. air velocity) and two personal variables (activity and clothing level
E-mail address: [email protected] (E.Z.E. Conceição). of the occupants). Physical and personal parameters take into

0360-1323/$ e see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.buildenv.2011.10.013
 E.Z.E. Conceição et al. / Building and Environment 50 (2012) 190e201 191

consideration of the behavioural adaptation to some extent. patterns and expectations levels are changing continuously and
However, these parameters completely ignore the psychological rapidly Ref. [24].
dimensions of adaptation, social and cultural aspects of an occu- To overcome the above discrepancies, it is believed that only the
pant, which are otherwise so prominent in any naturally ventilated combination of both heat balance model and adaptive model will
buildings. These parameters are important because of strong rela- eventually be able to account for both the thermal and non-thermal
tionship between occupant’s behavioural adaptation and environ- influences on occupant’s response in real buildings Ref. [27]. In
ment Ref. [24]. order to determine a model that combines these two approaches, in
The adaptive theory assumes that people consciously or recent years, some researchers have been developing a theoretical
unconsciously respond to a given thermal environment to which adaptive model of thermal comfort based on the “Black Box” theory,
they are exposed in order to restore their own thermal comfort taking into account factors such as culture, climate, social,
Ref. [22]. The recent development of the adaptive theory of psychological and behavioural adaptations, which have an impact
thermal comfort explains occupants’ thermal comfort in different on the senses to detect thermal comfort. The model is called the
environmental contexts, particularly in naturally conditioned Adaptive Predicted Mean Vote (aPMV) model Ref. [27]. The appli-
buildings, from the point of view of the adaptive approach cation of this method on experimental studies was used by the
Ref. [8]. Occupants in a given environment can achieve thermal following authors Refs. [13,24,25,27].
comfort through adjustments to their personal environmental For acceptable thermal comfort conditions the [21] defines three
conditions in the form of taking on/off clothing, opening/closing comfort categories (A, B and C), establishing limits for PMV and PPD
windows, shifting the sun shadow and switching on/off fans, indexes. In the category C, used in this work, the limits of the
heating or air-conditioners, etc. (de Dear and Brager, 1998) [18]. suggested PMV values change between 0.7 and
0.7 (PPD ¼ 15%)
The adaptive model reveals that the thermal comfort is a func- and consequently the limits of the suggested aPMV values change
tion related to the outdoor air temperature. Ref. [4] was one of the between 0.7 and
0.7.
firsts to propose the adaptive control algorithm. Other researchers The air exchange rate inside an occupied compartment can be
have also presented several empirical equations for the indoor calculated using different recommendations and methodologies
thermal comfort temperature based on the different surveys of presented in national (D.L. n 79/2006 of April 4th) and interna-
free-running buildings. tional [3] standards. Both standards recommend, in order to guar-
The majority of adaptive models are applied to define the antee acceptable air quality, that the internal carbon dioxide
neutral temperature as a function of outdoor Refs. [16,17,20,23]; or concentration should be lower than 1800 mg/m3. The Portuguese
outdoor and indoor Ref. [16] temperatures. normalization, presented in D.L. n 79/2006 of April 4th, defines an
It is argued that the adaptive thermal comfort model does not airflow rate of 35 m3/h per occupant (9.7 l/s).
explicitly includes as the inputs the environmental parameters like The measurement with tracer gas is a very interesting technique
air temperature, air relative humidity, mean radiant temperature for assessing, locally or globally, the effect of a ventilation system in
and air velocity and the personal variables such as clothing insu- a given space. A good tracer gas should not be toxic or flammable and
lation level and metabolic rate, which have well known impact on nor should it have adverse effects on health at the used concentra-
human thermal sensation. As the adaptive approach to thermal tions. It should be colourless, odourless and should not be absorb-
comfort is based on the thermal comfort conducted in the field, able by existing accessories or surfaces inside the compartment. It
where researchers usually collect data on the thermal environment must have a density approaching that of air or the pollutant that
and the simultaneous thermal response of subjects, all the adaptive simulates. The idea underlying this technique is to trace or follow the
models include these parameters implicitly. movement of the air in a space through a gas known as a tracer.
The fundamental assumption of the adaptive approach is Several techniques for the use of tracer gases can be used. The
expressed by the adaptive principle: if changes produce discomfort, choice of one of them has to do with the specific features of the
people react in ways which tend to restore their comfort Ref. [22]. space, where there are measurements, and the present flow
The surveys reveal there is a strong relationship between occu- induced ventilation system. These techniques can determine the air
pant’s behavioural adaptation and environment. That is, it can’t be exchange rate, the airflow rate, the age of the air, the efficiency,
ignored the psychological dimensions of adaptation, social and among other quantities. Several studies have been conducted
cultural aspects of an occupant, because the adaptation is within the air quality analysis using the method of tracer gases.
geographical influenced and highly influenced by the local climatic Several examples can be found in Sandberg [29], Sandberg and
conditions and socio-cultural background. Sjöberg [30], Breum [31], Crawford and Neill [32], Niemelä et al.
Field studies in naturally ventilated buildings have shown that [33], Niemelä et al. [34], Roulet et al. [35], Breum and Skotte [36],
the PMV predicts warmer thermal sensations than those that the Roulet [37], Breum [38] and Roulet and Cretton [39].
occupants actually feel in that kind of buildings Ref. [7], that is, Numerical software, that simulates the building thermal
overestimates the occupant responses at high temperatures and behaviour with complex topology, in transient conditions, was
underestimates them at low temperatures, which will lead to the developed and used in the study of kindergarten thermal response,
use of more air conditioning than is necessary. The adaptive models occupants’ thermal comfort and air quality, in mediterranean
also reveal that the thermal comfort temperature is a function of conditions. This study is a continuation of previous works made in
the outdoor temperature. several kinds of schools, namely, in high schools and universities. In
Ref. [24] conducted a field study in North-East India which is this work the application of a developed adaptive model in the
broadly classified into three climatic zones (warm and humid, cool evaluation of thermal comfort in ventilated kindergarten occupied
and humid and cold and cloudy) and for all four seasons. This field spaces is made. In the experimental tests environmental parame-
study on thermal comfort allowed developed theoretical models ters were measured and subjective responses survey was evalu-
based on adaptive coefficients varying for different seasons and ated, in order to obtain the PMV and the AMV values. The airflow
also for different climatic zones. This study field does validate rate was experimentally evaluated using the tracer gas concentra-
adaptive thermal comfort in conditioned building but the question tion. In the numerical simulations the inside environmental
remains as how long this adaptive model will continue to suit in thermal variables, carbon dioxide concentration, PMV and aPMV
the new buildings where occupants clothing patterns, activity were obtained.
192 E.Z.E. Conceição et al. / Building and Environment 50 (2012) 190e201

Fig. 1. Scheme of the kindergarten and surrounding buildings.

2. Numerical models these equations are evaluated by a sub-model that calculates the
incident solar radiation, internal view factors, radiative heat
2.1. Building thermal behaviour numerical model exchanges, glass radiative properties, occupation cycles, ventilation
strategy, human thermo-physiology and heat and mass transfer
In this numerical model an energy and mass balance integral coefficients by convection using empirical expressions.
equations philosophy (generated by the building geometry), an The mass balance integral equations are developed for the water
equations system resolution done by the Runge-Kutta-Felberg with vapour inside the spaces, the contaminants inside the spaces, the
error control and a human thermal comfort level evaluated through water vapour in the solid matrix for the building main body and
human thermo-physiology, are developed. interior bodies, the contaminants in the solid matrix for the
The energy balance integral equations are developed for the building main body and interior bodies, the water vapour at the
indoor spaces, opaque main bodies, transparent bodies and interior surface of the building main bodies, interior bodies and windows
bodies. These equations are based in conduction, convection, glasses and the contaminants at the surface of the building main
radiation and evaporation phenomena. All coefficients used in bodies and interior bodies.

W (4)
W(3)

W (2)

W(4 ) W(1)
W (3)
W (2)

S(6 )
S (5 )
S(4 )
S(3 )
S(2 )
W(1 )
S(1)
S W(1 )

Fig. 2. Kindergarten spaces identification. The windows numbers presented in this figures are associated with the caption of the Figs. 9 and 10.
 E.Z.E. Conceição et al. / Building and Environment 50 (2012) 190e201 193

Table 1 studies of comparison between numerical and experimental data in

Evolution of occupants’ number in each space. summer conditions (see Ref. [10] and in winter conditions (see
Time 0e8 8e9 9e12 12e13 13e15 15e16:30 16:30e18:30 18:30 to 24 Ref. [10].
(hours) The validation of the indoor air temperatures, for four occupied
Spaces strategic rooms, was measured in a typical summer day Ref. [10]. In
3 27 27 this study two kinds of numerical simulations were made: without
7 2 4 and with surrounding buildings. In accordance with the obtained
10 30 4 68 20 20 results was verified that the introduction of the surrounding
13 22 22
buildings decreases the indoor air temperature level, mainly in the
16 22 22
18 2 2 2 1 1 1
beginning of the morning and in the end of the afternoon. The
indoor air temperature values experimentally obtained are, in
general, higher than the numerical values during the night and
In the thermal comfort evaluation the PMV and PPD indexes are lower than the numerical values during the day. The numerical
considered. These indexes consider four environmental variables model, when surrounding buildings are considered, reproduces
[air temperature, air relative humidity, air velocity (calculated better the experimental values, being in general the difference
based in the air renovation and recirculation) and mean radiant between numerical and experimental indoor air temperature
temperature (calculated based in weighted mean value of the values lower than 2  C.
surrounding surfaces)] and two personal parameters (activity and The indoor air temperatures experimental data, used in the
clothing level, calculated in accord the [21] for winter and summer numerical model validation, for three occupied strategic rooms
conditions). In the numerical simulation the four environmental were also measured in a typical winter day Ref. [11]. In accordance
variables are numerically calculated in each iteration, while the two with the obtained results, it was possible to conclude that the
personal parameters are defined for each simulation day. numerical model reproduces the experimental indoor air temper-
Before the numerical model is applied, it was subjected to atures behaviour. In general, the difference between numerical and
different validation tests, where the numerical values were experimental indoor air temperature values is lower than 2.5  C.
compared with experimental values. The results of these valida- Some discrepancies verified between the numerical values and the
tions tests were made for winter and summer conditions and two experimental results, can be seen in Ref. [11].
situations were considered: school building with and without
surrounding buildings.
The present numerical model was validated in a school building 2.2. Adaptive thermal comfort model
in winter conditions (see [12] and in summer conditions (see [9]. In
these validation tests, without surrounding buildings, experimental The Predicted Mean Vote (PMV) and the Predicted Percentage of
values measured in a real building were compared with numerical Dissatisfied people (PPD) method developed by Fanger Ref. [19] has
values calculated by this developed software. been used worldwide to predict and assess indoor thermal comfort
In Ref. [12] was verified that the numerical model reproduces in buildings. The method takes into account the combined impact
well the measured experimental data. In general, the difference of the activity level (M), the clothing level ðIcl Þ, the air temperature
between indoor air temperature numerical values and experi- ðTair Þ, the air velocity ðVair Þ, the air relative humidity ðRHair Þ and the
mental data is lower than 2  C. A maximum difference of 4  C in mean radiant temperature ðMRTÞ on thermal comfort level (PMV):
compartments with windows turned towards South, only during
PMV ¼ f ðM; Icl ; Tair ; Vair ; RHair ; MRTÞ (1)
some hours in the afternoon, was verified. The difference between
indoor air relative humidity numerical values and experimental Field study surveys on thermal comfort have brought the PMV
data, in general, changes between 10 and 20%. The highest differ- model into question because it does not take into account adap-
ence is verified in the first hours of the day. tations of the human body, which play a key role in determining
In Ref. [9] the validation tests showed a good agreement subjective thermal sensation and perception Refs. [5,6,14]. The field
between experimental data and numerical values in all analyzed studies developed by Refs [13,24,25,27] shown that exist
spaces. In general, a difference around 1  C was verified. Some a discrepancy between actual mean votes (AMV), given by the
verified discrepancies are presented in Ref. [9]. thermal sensation votes recorded on the ASHRAE seven-point scale
In a similar geometry with the analysed in this work, with during comfort survey, and PMV values. A correction factor, called
surrounding buildings, the present numerical model was used in “adaptive coefficient”, is derived mathematically to match the AMV

Table 2
Airflow (m3/s) used in winter conditions, for the different spaces. In all spaces not considered and not represented in this table, infiltrations with 0.11 renovation per hour, were
considered.

Time (hours) 0e8 8e9 9e12 12e13 13e15 15e16:30 16:30e18:30 18:30e24

Airflow direction
Outdoor / Greenhouse 2 0,2499 0,0333 0,5664 0,0100 0,1666 0,1666
Outdoor / Classroom 3 0,0053 0,0974 0,0053 0,0974 0,0053 0,0053
Outdoor / Playground 10 0,0110
Outdoor / Classroom 13 0,0059 0,1875 0,0059 0,1875 0,0059 0,0059
Outdoor / Classroom 16 0,0064 0,1519 0,0064 0,1519 0,0064 0,0064
Greenhouse 2/ Corridor 25 0,2499 0,0333 0,5664 0,1666 0,1666
Corridor 25/ Playground 10 0,2499 0,0333 0,5664 0,1666 0,1666
Classroom 3/ Outdoor 0,0053 0,0974 0,0053 0,0974 0,0053 0,0053
Playground 10/ Outdoor 0,0110
Classroom 13/ Outdoor 0,0059 0,1875 0,0059 0,1875 0,0059 0,0059
Classroom 16 / Outdoor 0,0064 0,1519 0,0064 0,1519 0,0064 0,0064
194 E.Z.E. Conceição et al. / Building and Environment 50 (2012) 190e201

Table 3
Airflow (m3/s) used in summer conditions, for the different spaces. In all spaces not considered and not represented in this table, infiltrations with 0.11 renovation per hour,
were considered.

Time (hours) 0e8 8e9 9e12 12e13 13e15 15e16:30 16:30e18:30 18:30e24

Airflow direction
Outdoor / Greenhouse 2 0,0100
Outdoor / Classroom 3 0,0053 0,2322 0,2322 0,2322 0,0053 0,0053
Outdoor / Playground 10 0,2499 0,0333 0,5664 0,0110 0,1666 0,1666
Outdoor / Classroom 13 0,0059 0,2764 0,2764 0,2764 0,0059 0,0059
Outdoor / Classroom 16 0,0064 0,3328 0,3328 0,3328 0,0064 0,0064
Playground 10 / Corridor 25 0,2499 0,0333 0,5664 0,1666 0,1666
Corridor 25 / Greenhouse 2 0,2499 0,0333 0,5664 0,1666 0,1666
Greenhouse 2 / Outdoor 0,2499 0,0333 0,5664 0,0100 0,1666 0,1666
Classroom 3/ Outdoor 0,0053 0,2322 0,2322 0,2322 0,0053 0,0053
Playground 10/ Outdoor 0,2499 0,0333 0,5664 0,0110 0,1666 0,1666
Classroom 13/ Outdoor 0,0059 0,2764 0,2764 0,2764 0,0059 0,0059
Classroom 16 / Outdoor 0,0064 0,3328 0,3328 0,3328 0,0064 0,0064

and PMV values. This correction factor is added to the PMV and is PMV ¼ G  d (2)
calculated for each outdoor air temperature. According to the above
analysis the index aPMV (adaptive PMV) is proposed Where:
Refs. [24,25,27], as a correction model to the Fanger PMV index. G e Transfer function e human physiology (thermoregulatory
All the adaptive actions are based on feedback mechanism and system) and thermal sensation;
present adaptation is the expectation and preference of past d e Physical stimuli.
experience. This is the starting point for further adaptive response According to the ‘Black Box’ theory, physiological adaptation is
Ref. [7]. It is also evident that occupants of the buildings are always contained within a black box, but psychological and behavioural
active towards the changes in thermal environment. For this stimuli will give an “adaptive (negative) feedback”. Then, aPMV can
reason, the adaptive actions are a feedback to the link between be obtained using the following equation:
outdoor and indoor climate and indoor climate and thermal stimuli.
aPMV ¼ G  d
aPMV  Kd  G (3)
According to the heat balance equation, PMV can be calculated
using equation: Where Kd is a psychological and behavioural impact coefficient.

a 8,0
7,8
a 8,0
Ln (CO2) = -3,521 t + 7,394 Ln (CO2) = -2,843 t + 7,836
7,8
7,6 R2 = 0,859 2
R = 0,950
7,6
7,4
7,4
7,2
Ln (CO2)

7,2
7,0
Ln (CO2)

7,0
6,8
6,8
6,6
6,6
6,4
6,4
6,2
6,2
6,0
6,0
0,00 0,05 0,10 0,15 0,20 0,25 0,30 0,35
0,00 0,05 0,10 0,15 0,20 0,25 0,30 0,35
t (hours)
t (hours)
b
b
8,5
8,5
8,3 Ln (CO2)= -5,1923 t + 8,151
2
R = 0,9772 8,3 Ln (CO2) = -6,224 t + 8,060
8,1 2
R = 0,963
8,1
7,9
7,9
7,7
Ln (CO2)

7,7
Ln (CO2)

7,5
7,5
7,3
7,3
7,1
7,1
6,9
6,9
6,7
6,7
6,5
6,5
0,00 0,05 0,10 0,15 0,20 0,25
0,00 0,05 0,10 0,15 0,20 0,25
t (hours) t (hours)

Fig. 3. Evolution of carbon dioxide concentration for 50% (a) And 100% (b) open Fig. 4. Evolution of carbon dioxide concentration for 50% (a) And 100% (b) open
window area, obtained in the classroom 13. window area, obtained in the classroom 16.
 E.Z.E. Conceição et al. / Building and Environment 50 (2012) 190e201 195

In order to find out the value of the adaptive coefficient a field

a 8,0
study has been carried out by using a subjective questionnaire survey
7,8 Ln (CO2) = -1,999 t + 7,808
2
R = 0,951 simultaneously with thermal environmental parameters monitoring.
7,6
The authors calculated the values for l in warm conditions and in cold
7,4 conditions. In the present model, used in this work, for other outdoor
7,2 air temperature, a linear variation was used in the evolution of l.
Ln (CO2)

7,0
6,8
3. Experimental setup
6,6
6,4 In the experimental tests the air exchange rate and the adaptive
6,2 thermal comfort model are developed. The airflow rate obtained in
6,0 the classroom is used as input value and the adaptive thermal
0,00 0,05 0,10 0,15 0,20 0,25 0,30 0,35 comfort model is used in the comfort evaluation in the kinder-
t (hours) garten thermal response numerical simulation.
The experimental tests are made in three real classrooms with
b 8,5
a volume of 153.4 m3 (“classroom 13”, in Fig. 2), 150.5 m3 (“class-
Ln (CO2) = -4,410 t + 8,296
8,3 2
R = 0,996 room 16”, in Fig. 2) and 148.2 m3 (“classroom 3”, in Fig. 2). Each
8,1 classroom, in a school building located in the South of Portugal, had
7,9 one or two main windows and one door.
7,7 The environment variables, during the tests, are measured by
Ln (CO2)

7,5 Babuc-A, of LSI, with a rate of 2 samples per minute. In general, the
tests are made in the room central area, 1.2 m above the floor level.
7,3
In general, the airflow rate is made by infiltration, when the
7,1
windows are closed, or by crossed natural ventilation, when the
6,9 windows are open. In the first two classrooms (numbers 13 and 16)
6,7 two windows located in different walls are identified, while in the
6,5 last classroom only one window is identified. The first ones, with
0,00 0,05 0,10 0,15 0,20 0,25 a maximum open area with 3.6 m2, are located in the North and
t (hours) East walls, the second ones, with a maximum open area with
2.3 m2, are located in the North and West walls, while the third one,
Fig. 5. Evolution of carbon dioxide concentration for 50% (a) And 100% (b) open with a maximum open area with 1.6 m2, is located in the South wall.
window area, obtained in the classroom 3.
The airflow rate (or the air exchange rate) is evaluated using the
tracer gases decay method. In these experimental tests the carbon
dioxide and the air were well mixed before the beginning of the test
Substituting Equation (2) into Equation (3), then
using two mixing ventilators. The measurements are made during

 35 min. However, when the airflow rate (or the air exchange rate) is
K  PMV
aPMV ¼ PMV= 1 þ d (4) high this time is reduced.
d
During the experimental test in the classroom, without occu-
Assuming l ¼ Kd =d, the index aPMV is done by: pation (in order not to influence the experimental measurements),
the steady-state conditions are maintained, the sky is clean and the
PMV
aPMV ¼ (5) outdoor wind velocity is very low.
1 þ l  PMV In these tests two ventilations strategies are analyzed:
where l is the “adaptive coefficient”.
When the adaptive coefficient is positive, that means the indoor  Test I, the half (50%) of the total window open area is made;
temperature is greater than comfort temperature. This situation is  Test II, the total (100%) window open area is made.
generally common on summer for naturally ventilated buildings.
From Equation (5), for l > 0 situations, the value of aPMV is smaller
than PMV, which means aPMV predicts a cooler vote than PMV does at Table 5
Statistical data of the students participated in the survey.
the same thermal condition. At l < 0, that means the indoor
temperature is lower than comfort temperature. This situation usually Number of children “Classroom 3” 20
occurs in winter for naturally ventilated buildings. From Equation (5), “Classroom 13” 15
it is observed that aPMV predicts a warmer vote than PMV does at the
“Classroom 16” 17
same thermal condition. At l ¼ 0, that means the indoor temperature
Total 52
is equal to comfort temperature or aPMV ¼ PMV, condition theo-
retically possible, but unlikely to happen. Respondent gender Female 22
Male 30
Average age (years) 5.4
Average weight (kg) 21.41
Table 4 Average height (m) 1.16
Air exchange rate, obtained in the three classroom spaces, for 50% and 100% open Estimate metabolic rate (met) 1.2
window area. Clothing level (clo) Winter 1.1
Summer 0.5
50% of open window area 100% of open window area Number of surveys “Classroom 3” 120
Classroom 3 1999 1/h 4410 1/h “Classroom 13” 90
Classroom 13 3521 1/h 5192 1/h “Classroom 16” 102
Classroom 16 2843 1/h 6224 1/h Total 312
196 E.Z.E. Conceição et al. / Building and Environment 50 (2012) 190e201

1,60 environmental parameters measured during the survey with the

AMV metabolism value and the clothing insulation value were used in
1,40
aPMV
the computational technique developed by Ref. [21] standards to
calculate the PMV values. The clothing level was obtained based in
aPMV/AMV

1,20
the children subjective responses, while the metabolism level was
1,00 obtained by observation of the children activity (see ISO 7730,
2005) [21].
0,80

0,60 4. Building description

0,40
In this work the kindergarten, for 3, 4 and 5 years old children,
1,40 1,45 1,50 1,55 1,60 1,65 1,70 1,75 1,80 1,85 1,90
as well as two schools (placed in West area in relation to the
PMV
kindergarten) and three residential (placed in East, South-East and
Fig. 6. The relationships between the PMV, aPMV and AMV in warm conditions. South area in relation to the kindergarten) surrounding buildings
(see Fig. 1), are considered.
In the numerical simulation the 25 compartments (Classrooms,
In order to obtain the adaptive thermal comfort model, the indoor playground, office, kitchen, teachers’ room, non-teacher’s
indoor environmental variables (air temperature, air velocity, air room, corridors, hall, storage rooms, WC and garret), the 25 wood
relative humidity and mean radiant temperature) and the outdoor doors and the 42 glass windows are considered.
air temperature are measured by sensors. Simultaneously, the In passive strategies the kindergarten is equipped with:
subjective survey is made.
Respondents always try to adapt themselves to the existing Multiple inclined aluminium shading devices placed above one
thermal environment. These adaptations are largely governed by transparent window level facing South and in front one trans-
their own socio-cultural background and traditions. In naturally parent door facing South. These transparent window and door,
ventilated buildings, the air temperatures corresponding to installed in the classroom number 3, guarantee the maximum
comfortable thermal environment keep changing due to changes in passage of solar radiation in winter solstice (21st December)
indoor and outdoor environmental conditions. Simultaneous conditions and none passage of solar radiation in summer
indoor and outdoor environmental parameters monitoring and solstice (22nd June) conditions.
a questionnaire survey were conducted during the period of Horizontal fabrics shading devices placed above the transparent
experimental analysis. panels’ levels facing South, South-West and West. These opaque
The questionnaire was enquire to reflect the respondents’ panels are installed in the greenhouse number 2 and do not
subjective assessment of the indoor thermal environment. The allow the solar radiation passage in the South transparent panel
main contents of the survey include: during the children lesson times (until 15 pm), in the South-
West and West transparent panels, in summer solstice condi-
1 e Respondents’ background, such as gender, age, weight, tions. However, in winter solstice conditions, these opaque
height, etc.; panels allow the maximum solar radiation passage. The trans-
2 e Respondents’ clothes and activity rate; parent panels facing South-West and West are not subjected to
3 e Respondents’ subjective thermal sensation: the ASHRAE shading devices due to the surrounding buildings placed in the
seven-point thermal sensation scale was used in the survey to East area.
help respondents express their own thermal sensation. They
are, respectively,
3 cold,
2 cool,
1 slightly cool, 0 neu- This kindergarten, used in this study, similar that one located in
tral, þ1 slightly warm, þ2 warm and þ3 hot. Olhão (South of Portugal), has 10 spaces with higher occupation
4 e Respondents’ sensation about the classroom space pleas- and 15 spaces with lower occupation. As main spaces with higher
antness: yes or no. occupation are considered the classrooms (numbers 3, 13 and 16),
indoor playground (number 10), kitchen (number 18), teachers’
The thermal sensation respondents’ votes recorded during room (number 7) and corridor (number 25), while in the main
comfort survey is considered as actual mean votes (AMV). The spaces with lower occupation the hall (number 2) is considered.

0,80 1,5
Classroom_3
0,60
AMV Classroom_16
0,40
1
Classroom_13 λ(3) = 0,1539Tout - 3,3798
aPMV Linear (Classroom_3) R = 0,8258
0,20
0,5 Linear (Classroom_16)
aPMV/AMV

0,00 Linear (Classroom_13) λ(13) = 0,0916Tout - 1,8988

R = 0,7479
0
λ

-0,20
λ(16) = 0,141Tout - 3,1325
-0,40 R = 0,9477
-0,5
-0,60
-0,80 -1
-1,00
-1,20 -1,5
-1,40 -1,20 -1,00 -0,80 -0,60 -0,40 -0,20 0,00 12 16 20 24 28

PMV Tout (ºC)

Fig. 7. The relationships between the PMV, aPMV and AMV in cool conditions. Fig. 8. Variation of l as a function of the outdoor temperature for each classroom.
 E.Z.E. Conceição et al. / Building and Environment 50 (2012) 190e201 197

5000
South (1) South (2) South (4)
4500 South (5) South (6) West (1)
South-West (1) South (3) East (1)
4000 East (2) East (3) East (4)
West (2) West (3) West (4)
3500
3000

I (W) 2500
2000
1500
1000
500
0
7 8 9 10 11 12 13 14 15 16 17
Time (hours)

Fig. 9. Evolution of transmitted solar radiation calculated in mains windows in winter conditions.

In Fig. 2 the kindergarten spaces identification is presented. The In this study two numerical simulations were made. In both
caption of the spaces in the figure is the following: studies the natural ventilation in the classroom are measured and
the forced ventilation in the indoor playground, corridor and hall
2 Greenhouse; are introduced:
3 Classroom;
7 Meeting room;  In the first simulation, made in winter conditions, the natural
10 Indoor playground; ventilation in the classrooms, with 50% of window open area,
13 Classroom; was introduced. The forced ventilation is introduced from the
16 Classroom; hall warm space (worked as greenhouse) to the indoor
18 Kitchen; playground;
25 Corridor.  In the second simulation, made in summer conditions, the
natural ventilation in the classrooms, with 100% of window
open area, was introduced. The forced ventilation is introduced
5. Input data from the indoor playground (cold space) to the hall;

As input external environmental information (air temperature, In Table 1 the evolution of occupants during one typical day, in
air relative humidity, wind velocity and wind direction) the each space, are presented, while the airflow used in the winter and
experimental data measured in 21st December, in winter condi- summer conditions, for the different intervals, in the occupied
tions, and 22nd June, in summer conditions, were considered. The spaces, are presented, respectively, in Tables 2 and 3.
external environmental information experimentally used in this In rooms where no measurements were made an air exchange
work is considered typical as winter and summer conditions in the rate of 0.11 1=h was considered. This value was obtained from the
Algarve (South Portugal). In the school building numerical simu- average of the experimentally obtained values using the tracer
lations, in order to evaluate the real building thermal inertia, the gases decay method in the rooms where measurements were made
previous 5 days were also simulated. when windows and door were closed (by infiltrations).

6000
South (1) South (2) South (4)
South (5) South (6) West (1)
5000 South-West (1) South (3) East (1)
East (2) East (3) East (4)
West (2) West (3) West (4)
4000
I (W)

3000

2000

1000

0
7 9 11 13 15 17 19
Time (hours)

Fig. 10. Evolution of transmitted solar radiation calculated in mains windows in summer conditions.
198 E.Z.E. Conceição et al. / Building and Environment 50 (2012) 190e201

6.2. Adaptive model

In Table 5 the statistical data of the trained occupants that

participated in the survey is presented.
In order to find out the value of the adaptive coefficient l in the
aPMV model, for a kindergarten, a subjective questionnaire survey,
simultaneously with thermal environmental parameter moni-
toring, is made. The field survey is conducted simultaneously in 3
classrooms of a kindergarten, in a typical day of January and of May
of 2007. The kindergarten children were prepared and briefed on
the survey questions. Statistical data of the students participated in
the survey are shown in Table 5. For each outdoor temperature, the
least square method is used to fit the sets of data obtained from the
Fig. 11. Evolution of internal air temperature calculated in winter conditions. field study. Accordingly to Ref. [27]; with the results of the surveys
in the three rooms and the determination of Fanger’s PMV corre-
sponding to each survey, for each outdoor temperature, is possible
to determine the adaptive coefficient from:
6. Results and discussion
Pn
i ¼ 1 ðYi
Xi Þ
In this study, the software that simulates the building thermal l ¼ (6)
n
response with complex topology, in transient conditions, is used in
a ventilated kindergarten. In the simulation the indoor radiative Where: Yi e AMV obtained from the survey’s answers; Xi e PMV
heat exchanges and the indoor curtains are not considered. obtained from Fanger’s equation; N e number of data sets.
In the first part, the experimental natural ventilation is deter- The relationships between the PMV, aPMV and AMV in warm
mined, in the second part an adaptive comfort model is developed, conditions are shown in Fig. 6 and the relationships between the
using trained occupants, in the third part the external incident solar PMV, aPMV and AMV in cool conditions are shown in Fig. 7. The
radiation, that the surrounding kindergarten surfaces are subjected Figs. 6 and 7 show that the fit between actual votes and the pre-
in winter and summer conditions, is analyzed and in the fourth and dicted votes using aPMV model is good within the range of
1; 4 
fifth parts the internal thermal conditions, using the internal air PMV  2.
temperature field, the occupants’ thermal comfort level (using the For each outdoor temperature, from Equation (6) it gets the
PMV and the aPMV index), and the internal air quality (using the points present in Fig. 8. By linear regression it is obtained the
dioxide carbon concentration), for winter and summer conditions, variation of l as a function of the outdoor temperature for each
respectively, are analyzed. considered classroom. Other set of data obtained in other studies

6.1. Air exchange rate

In Figures 3, 4 and 5 the evolution of carbon dioxide concen-

tration, obtained respectively in the classroom 13, 16 and 3, are
presented. The figures (a) are associated to 50% open window area
and the figures (b) are associated to 100% of open window area.
In Table 4 the air exchange rate, obtained in the three classroom
spaces, for 50% and 100% open window area, are presented.
According to the obtained results, as is expected, the air
exchange rate increase with the open window area. The classrooms
with two windows located in different walls, classrooms 13 and 16,
presented higher air exchange rate than the classroom with only
one window, classrooms 3.

Fig. 13. Evolution of thermal comfort level calculated in winter conditions with the
Fanger model (a) and with the adaptive model (b). The horizontal line with the PMV
equal to þ0,7 represents the warmer thermal comfort acceptable limit, while the
Fig. 12. Evolution of internal carbon dioxide concentration calculated in winter horizontal line with the PMV equal to
0,7 represents the cooler thermal comfort
conditions. The horizontal line represents the acceptable carbon dioxide limit. acceptable limit.
 E.Z.E. Conceição et al. / Building and Environment 50 (2012) 190e201 199

comfort level (a) and adaptive comfort level (b) are presented. In all
figures the space number, space name and window directions are
included in the caption.
According to the presented results is possible to conclude that:

- The internal air quality level, evaluated using the carbon

dioxide concentration, is not acceptable in classrooms (sub-
jected with natural ventilation) occupied spaces, however, in
the indoor playground (subjected with forced ventilation) is
acceptable, in accord to the Ref. [3,15];
- According with the Fanger model the spaces with windows
turned to South are thermally comfortable, by warmer PMV
values, while the spaces with windows turned to East, North and
Fig. 14. Evolution of internal air temperature calculated in summer conditions. West are thermally uncomfortable during the morning and ther-
mally comfortable during the afternoon, by cooler PMV values;
- Using the adaptive comfort model, all spaces are thermally
suggests that there is a linear relation between the adaptive coef- comfortable. The space with windows facing South with
ficients and the outdoor temperature. warmer aPMV values and the other spaces by cooler aPMV
For the other rooms, it was considered the weighted average values;
results obtained for the three classrooms, and thus the respective - The application of the adaptive model in spaces with cooler
adaptive coefficient is given by: aPMV values suggested that the comfort conditions are better
than the predicted by the Fanger model. However, the appli-
lother ¼ 0:1247,Tout
2:7041 (7) cation of the adaptive model in spaces with warmer aPMV
C values, in general, suggest that the comfort conditions are
The results shown for an outdoor temperature of 21.7 (spring
slightly worse than the predicted by the Fanger model, even so
and autumn conditions in the region) the adaptive coefficient is
inside the comfort caption.
zero, and thus aPMV is equal to PMV. For an outdoor temperature of
5.5  C above (summer conditions), the aPMV is around 50% lower
than the PMV. For an outdoor temperature 5.5  C below (winter
conditions), the aPMV is around 30% lower than the PMV. 6.5. Comfort in summer conditions

In Figs 14, 15 and 16 are presented, for summer conditions, the

6.3. External solar radiation evolution of internal air temperature, internal carbon dioxide
concentration and internal thermal comfort level, that the occu-
In Fig. 9 And 10 the evolution of solar radiation, respectively for pants are subjected, respectively. In Fig. 16 the Fanger thermal
winter and summer conditions, are presented. The caption of the
South, West, East and South-West window are represented in the
Fig. 2.
According to the numerical results is possible to verify that in
summer conditions the shading devices guarantee, in the main
occupation time, low radiation level in the windows surfaces.
However, in winter conditions the opposite is observed.

6.4. Comfort in winter conditions

In Figs. 11, 12 and 13 are presented, for winter conditions, the

evolution of internal air temperature, internal carbon dioxide
concentration and internal thermal comfort level, that the occu-
pants are subjected, respectively. In Fig. 13 the Fanger thermal

Fig. 16. Evolution of thermal comfort level calculated in summer conditions with
Fig. 15. Evolution of internal carbon dioxide concentration calculated in summer Fanger model (a) and with the adaptive model (b). The horizontal line with the PMV
conditions. The horizontal line represents the acceptable carbon dioxide limit. equal to þ0,7 represents the warmer thermal comfort acceptable limit.
200 E.Z.E. Conceição et al. / Building and Environment 50 (2012) 190e201

comfort level (a) and adaptive comfort level (b) are presented. In all the indoor air temperature, indoor air velocity, indoor air relative
figures the space number, space name and window directions are humidity and indoor mean radiant temperature (obtained through
included in the legend. the weighted mean value of the compartment surrounding
According to the results is possible to conclude that: surfaces temperatures), outdoor air temperature, for a pre-defined
clothing and activity levels. The considered clothing and the
- The internal air quality level, evaluated using the carbon activity level in winter conditions are, respectively, 1 Clo. and 1.2
dioxide concentration, is acceptable, according to the [3,15]; in Met, while in summer conditions are, respectively, 0.5 Clo. and 1.2
all occupied spaces; Met.
- The thermal comfort level, evaluated by the Fanger model, is
thermally uncomfortable, for warmer PMV values, for all
Acknowledgements
spaces. The indoor playground presents highest uncomfortable
levels because two main windows are subjected to direct solar
This research activity is being developed inside a project
radiation during all morning;
approved and sponsored by the Portuguese Foundation for Science
- The application of the adaptive comfort model, in general,
and Technology, POCI 2010, sponsored also by the European
present acceptable thermal comfort conditions, by slightly
Comunitary Fund FEDER.
warmer aPMV values.
The authors are grateful for the collaboration of Isilda Moreno,
Vitalina Pereira, Helena Uva Sancho and Paula Charrano in the
7. Conclusions
experimental subjective test responses.
This research activity is being developed inside a project
In this work was developed an adaptive model and applied for
financed by the City Council of Olhão.
the evaluation of thermal comfort in a kindergarten equipped with
natural and forced ventilation. The study, developed in real
conditions, is made in winter and summer thermal conditions, in References
mediterranean environment.
[1] Aitken D. Transitioning to a Renewable energy Future. International Solar
The airflow rate (or the air exchange rate) was evaluated using Energy Society; 2003.
tracer gases techniques, while the adaptive model was developed [2] ANSI/ASHRAE Standard 55. Thermal Environmental Conditions for Human
using subjective responses and experimental environmental vari- Occupancy. American Society of Heating, Refrigerating and Air-Conditioning
Engineers, Inc; 2004.
ables. The results allowed defining strategies for building ventila-
[3] ANSI/ASHRAE Standard 62.1. ASHRAE Standard – Ventilation for acceptable
tion in order to obtain the best compromise between the indoor air indoor air quality, Atlanta, American Society of Heating, Refrigerating and Air-
quality and thermal comfort. In winter conditions, the strategies Conditioning Engineers, Inc; 2004
considered natural ventilation in the classrooms, with 50% of [4] Auliciems A. Towards a psychophysiological model of thermal perception. Int J
Biometeorol 1981;27:83e6.
window open area, and forced ventilation from the hall (working as [5] Becker R, Paciuk M. Thermal comfort in residential buildings: failure to predict
greenhouse) to the indoor playground, while in summer condi- by standard model. Build Environ 2009;44(5):948e60.
tions, considered natural ventilation in the classrooms, with 100% [6] Bouden C, Ghrab N. An adaptive thermal comfort model for the Tunisian
context: a field study results. Energy Build 2005;37(9):952e63.
of window open area, and forced ventilation from the indoor [7] Brager G, de Dear R. Thermal adaptation in the built environment: literature
playground (cold space) to the hall. review. Energy Build 1998;27(1):83e96.
In this study, the used adaptive coefficient is based in a linear [8] Brager G, de Dear R. A standard for natural ventilation,. ASHRAE J 2000;
42(10):21e8.
regression and is function of the outdoor temperature. When l is [9] Conceição EZE, Lúcio MMJR. Numerical study of thermal response of school
equal to zero, the aPMV is equal to PMV, which is the lab-based buildings in summer conditions, proceedings of healthy buildings; 2006
condition with no adaptation. The simulation program and the [Lisbon. Portugal].
[10] Conceição EZE, Lopes MC, Lúcio, Ma MJR. Indoor environment comfort in
experimental results show that in warm conditions the aPMV is
a kindergarten school building in the South of Portugal in Summer conditions,
lower than the PMV, this means that the occupants’ thermal In Proceedings of the Indoor Air 2008; 2008. Copenhagen, Denmark, 17e22
comfort sensation in a certain environment could feel less warm August of 2008.
[11] Conceição EZE, Lopes MC, Lúcio Ma MJR. 2008b, Energy and thermal
than the predicted PMV index. In cool conditions the aPMV is
comfort management in a kindergarten school building in the South of
greater than PMV, this means that the occupants’ thermal comfort portugal in Winter conditions, In Proceedings of the 4th WSEAS/IASME
sensation in a certain environment could feel less cold than the International Conference on Energy, Environment, Ecosystems and
predicted PMV index. In winter conditions, according to the Sustainable Development (EEESD ’08) and 1st WSEAS International
Conference on Landscape Architecture (LA ’08), Faro, Portugal, 11 to 13
numerical results, the PMV, by warmer values, is slightly lower than June 2008.
aPMV, even so aPMV is inside the comfort limits. The experimental [12] Conceição EZE, Silva A, Lúcio Ma MJR. 2004. Numerical study of thermal
results show something slightly different: a PMV slightly cooler and response of school buildings in Winter conditions, In Proceedings of the 9th
Conference on Air Distribution in Rooms e RoomVent 2004. Coimbra, 5 to 8
an aPMV slightly warmer because in warm rooms the occupants September of 2004.
feel more heat than actually exists because they experience [13] Conceição E, Nunes A, Gomes J, Lúcio M. Application of a school building
a greater temperature difference between the room and the thermal response numerical model in the evolution of the adaptive thermal
comfort level in Mediterranean environment. Int J Vent 2010;9(3):287e304.
outside. [14] Corgnati S, Ansaldi R, Filippi M. Building and Environment 2009;44(4):
According to the measured air exchange rate, the adaptive 785e92.
model ensures, in winter conditions, good levels of thermal [15] Decreto-Lei n 79. Regulamento dos sistemas energéticos de climatização em
edifícios (RSECE). Diário da República, I Série- A, Nr 67, April 4th.; 2006.
comfort within the category C, and, in summer conditions, [16] de Dear R, Auliciems A. Validation of the Predicted Mean Vote model of
ensures acceptable levels of thermal comfort close to the category thermal comfort in six Australian field studies. ASHRAE Trans 1985;91(2b):
C. The indoor air quality, in summer conditions, is acceptable, 452e68.
[17] de Dear R, Brager G, Cooper D. Developing an adaptive model of thermal
with dioxide concentration below 1500 mg/m3. In winter condi-
comfort and preference. Final Report ASHRAE Rp-884; 1997.
tions, on the other hand, the indoor air quality is not acceptable; [18] de Dear R, Brager G. Developing an adaptive model of thermal comfort and
however the carbon dioxide concentration is near the acceptable preference. ASHRAE Trans 1998;4(1):145e67.
limit. [19] Fanger PO. Thermal comfort: analysis and applications in environmental
engineering. USA: McGraw-Hill Book Company; 1970.
The PMV index is calculated, by the software, for each space, [20] Humphreys M. Comfortable indoor temperatures related to the outdoor air
using the numerical values, calculated by the numerical model, of temperature. Build Serv Eng 1976;44:5e27.
 E.Z.E. Conceição et al. / Building and Environment 50 (2012) 190e201 201

[21] ISO 7730. Ergonomics of the Thermal Environments e Analytical determina- [32] Crawford RR, Neill PJO. Multi-Zone Airflow Measurements Using a Tracer Gas,
tion and interpretation of thermal comfort using calculation of the PMV and ASHRAE 1989, Building Systems, Room Air and Air Contaminant Distribution;
PPD indices and local thermal comfort criteria. Int Stand; 2005 [Switzerland]. 1989. pp. 103e105.
[22] Nicol J, Humphreys M. Adaptive thermal comfort and sustainable thermal [33] Niemelä R, Saamanen A, Koskela H. A comparison between the step-up,
standards for buildings. Energy Build 2002;34(6):563e72. step-down and pulse injection technique for the measurements of the mean
[23] Nicol F, Roaf S. Pioneering new indoor temperature standard: the Pakistan age of air. 10th AIVC Conference, Dipoli, Finland, 25e28 September;
project. Energy Build 1996;23:169e74. 1989:17e30.
[24] Singh M, Mahapatra S, Atreya S. Adaptive thermal comfort model for different [34] Niemelä R, Lefevre A, Muller JP, Aubertin G. Comparison of Three Tracer Gases
climatic zones of North-East India. Appl Energy 2011;88:2420e8. for Determining Ventilation Effectiveness and Capture Efficiency. Ann Occup
[25] Wei X, Xiangguang C, Jun Z. An adaptive Predicted Mean Vote (aPMV) model Hyg 1991;35(4):405e17.
in office. 201 Int Conf Mech Autom Control Eng; 2010:1887e991. [35] Roulet CA, Compagnon R, Jakob M. A Simple Method Using Tracer Gas to
[26] Weiss W, Bergmann I, Faninger G. Solar heating worldwide e markets and Identify the Main Airflow and Contaminant Paths Within a Room. Indoor Air
contribution to the energy supply. IEA SHC; 2005. 1991;1(3):311e22.
[27] Yao R, Li B, Liu J. A theoretical model of thermal comfort e Adaptive Predicted [36] Breum NO, Skotte J. Displacement Air Flow in a Printing Plant Measured With
Mean Vote (aPMV). Build Environ 2009;44:2089e96. a Rapid Reponse Tracer Gas System. Building Serv Eng Res Technol 1991;12:
[28] Yao R, Liu J, Li B. Occupants’ adaptive responses and perception of thermal 39e43.
environment in naturally conditioned university classrooms. Appl Energy [37] Roulet CA. Use of tracer gases to assess indoor air quality. ASHRAE Symposium
2010;87:1015e22. on Indoor Air Quality and Energy Consumption. Montreal: September; 1991.
[29] Sandberg M. What is ventilation efficiency? Build Environ 1981;16(2): [38] Breum NO. Flow Fields of Simulated Body Odor in an Office Ventilated From
123e35. the Displacement Design Principle, ROOMVENT 92, Air Distribution in Rooms,
[30] Sandberg M, Sjöberg M. The use of moments for assessing air quality in Third International Conference, Aalborg, Denmark, September, Vol. 3; 1992,
ventilated rooms. Build Environ 1983;18(4):181e97. pp. 181e194.
[31] Breum NO. The Air Exchange Efficiency of a Lecture Hall. Proceedings of [39] Roulet CA, Cretton P. Field comparison of age of air measurement techniques.
Second International Symposium on Ventilation for Contaminant Control. ROOMVENT'1992, Air Distribution in Rooms. Third International Conference,
London: England; 1988. p. 373e380. Aalborg, Denmark, Vol. 3; 1992, pp. 213e229.