Rehva 2022 - 05

The REHVA
European HVAC Journal

Vo lum e: 59 Issue: 5 Oc tober 2022 w w w. re hva .eu
CLIMA 2022 papers on

Health & Comfort
with focus on aerosols, IAQ and comfort
light + building Save
the
Date!
Get in touch with
our sustainable
future at
ISH 2023
World’s leading trade fair for

HVAC + Water
13. – 17. 3. 2023

Frankfurt am Main
The REHVA www.rehva.eu
Contents
European HVAC Journal Download the articles from www.rehva.eu -> REHVA Journal
Volume: 59 Issue: 5 October 2022
Editor-in-Chief: EDITORIAL 69 Correlation of subjective and

objective air quality data in
Jaap Hogeling 4 Health & Comfort shopping centres as a function of air
[email protected] Jaap Hogeling temperature and relative humidity
Editorial Assistant:
Mahmoud El-Mokadem, Kai Rewitz &
CLIMA 2022 TOP PAPERS Dirk Müller
Marie Joannes
[email protected] 5 CLIMA 2022 conference papers on
the theme HEALTH & COMFORT ARTICLES
General Executive: 75 Cross-infection risk between two
7 Evaluation of preventive measures
Ismail Ceyhan, Turkey in mitigating the risk of airborne people standing close to each other
infection of COVID-19 at different room temperatures
REHVA BOARD Peter V. Nielsen
Yunus Emre Cetin & Martin Kriegel
President:
Cătălin Lungu 12 Aerosol transmission in rotary wheel 76 Smart air quality control in
heat exchangers residences for optimised energy use
Inland-vorab
Vice Presidents: Heinrich Huber, Thomas Richter, and improved health of occupants
Lada Hensen Centnerová Florian Brzezinski & Michael Riediker Elizabeth Cooper & Yan Wang
Livio Mazzarella 18 Development of a non-contact 81 Why a unified IAQ approach is critical
Pedro Vicente-Quiles modular screening clinic (NCMSC) for to securing public health
Johann Zirngibl COVID-19 Morten Schmelzer
Ivo Martinac
Jinkyun Cho, Jinho Kim, Jongwoon Song & 84 Next-Generation Energy
Seungmin Jang Performance Certificates. What
Kemal Gani Bayraktar
24 Thermal inactivation of the corona novel implementation do we need?
EDITORIAL BOARD virus (SARS-CoV-2) in air volumes Lina Seduikyte, Phoebe-Zoe Morsink-
Murat Cakan, Turkey André Schlott, Thomas Hutsch, Eileen Sauer, Georgali, Christiana Panteli, Panagiota
Guangyu Cao, Norway Jens Wetschky, Jana Hessel, Susanne Bailer, Chatzipanagiotidou, Koltsios Stavros,
John Laubert & Stefan Lösch Dimosthenis Ioannidis, Laura Stasiulienė,
DU: 29.09.2022
Tiberiu Catalina, Romania Paulius Spūdys, Darius Pupeikis, Andrius

Francesca R. d’Ambrosio, Italy 31 Health and energy assessment of Jurelionis & Paris Fokaides
Ioan Silviu Dobosi, Romania a demand controlled mechanical
extraction ventilation system 89 Healthy Homes Design Competition:
Lada Hensen, The Netherlands “reTHINK living”
Janneke Ghijsels, Klaas De Jonge &
Karel Kabele, Czech Republic
Jelle Laverge Caroline Reich & Amelie Reiser
Risto Kosonen, Finland
40 Low-temperature radiant cooling 94 Tiny Homes – A Tiny solution to a big
Jarek Kurnitski, Estonia
panel for hot and humid climate problem
Livio Mazzarella, Italy
Gongsheng Huang, Nan Zhang & Laura Denoyelle
Dejan Mumovic, United Kingdom Yuying Liang
Ilinca Nastase, Romania
46 Impact of future climate on the REHVA WORLD
Natasa Nord, Norway
performances of ground-source 98 Meet the BIM-SPEED Competition
Dusan Petras, Slovakia cooling system Winners & Finalists
Olli Seppänen, Finland Abantika Sengupta, Pieter Proot, Tom Trioen, Jasper Vermaut
Hilde Breesch & Marijke Steeman
73943_ISH_AZ_vorab_REHVA_210x297 • FOGRA 39 • CMYK • js |
Branislav Todorovic, Serbia

Peter Wouters, Belgium 56 A Multi-Domain Approach to INDUSTRY EXPERTISE
Explanatory and Predictive Thermal 102 Smart Industrial Thermal Imaging
CREATIVE DESIGN AND LAYOUT Comfort Modelling in Offices
Jarkko Narvanne, [email protected] Eugene Mamulova, Henk W. Brink, Device PX1 by iRay
Marcel G.L.C. Loomans, Roel C.G.M. Loonen
ADVERTISEMENTS & Helianthe S.M. Kort EVENTS & FAIRS
Nicoll Marucciová, [email protected]
63 Exploring futures of summer comfort 105 Exhibitions, Conferences and
SUBSCRIPTIONS AND in Dutch households Seminars in 2022 & 2023
CHANGES OF ADDRESSES Lenneke Kuijer & Lada Hensen Centnerová
REHVA OFFICE:
Washington Street 40 Advertisers
1050 Brussels, Belgium
Tel: +32-2-5141171 99 LIGHT + BUILDING.......................................... 2 99 WORLD SUSTAINABLE ENERGY DAYS 2023.. 62
[email protected], www.rehva.eu
99 SWEGON........................................................... 6 99 INFIRAY........................................................103
PUBLISHER 99 UPONOR.........................................................17 99 PURMO.........................................................104
TEKNİK SEKTÖR YAYINCILIĞI A.Ş.
Fikirtepe Mah., Rüzgar Sk. No: 44C 99 BELIMO...........................................................23 99 REHVA MEMBERS........................................106
A3 Blok, Kat:11 D:124 Kadıköy/Istanbul, Turkey 99 REHVA BRUSSELS SUMMIT...........................30 99 REHVA SUPPORTERS....................................107
REHVA Journal is distributed in over 50 countries through the 99 ACREX INDIA 2023.........................................55 99 REHVA BRUSSELS SUMMIT.........................108
Member Associations and other institutions. The views expressed
in the Journal are not necessarily those of REHVA or its members. Next issue of REHVA Journal
REHVA will not be under any liability whatsoever in respect of
contributed articles. Instructions for authors are available at www.rehva.eu (> Publications & Resources > Journal
Cover photo: / Information). Send the manuscripts of articles for the journal to Jaap Hogeling [email protected].
EDITORIAL
Health & Comfort
W
ith focus on aerosols, IAQ and comfort. At afford the high fuel prices. This will affect productivity
the same time, we are aware of EU policy at workplaces and may cause health issues in residential
actions like the Green Deal, Renovation buildings. Compensation with higher activity levels
Wave, Fit for 55 by 2030 and Repower EU plan which (think of the elderly) and more clothing is not always
are the drivers for the ongoing EPBD (2018) revision possible. Working behind the computer keyboard as
expected to be published beginning 2023. The updated many persons do, will be hindered by lower tempera-
draft EPBD provides more attention to health and tures of the hands and fingers.
comfort. In the renovation wave program, there is a
focus on tackling energy poverty and worst-energy Saving on fuel cost and at the same time ask for better
performing buildings towards healthy housing. ventilation to avoid high COVID-19 infection risk
levels is also a challenge. In many existing situations
By referring in EPBD Annex 1 to EN16798-1 there like in schools heat recovery systems are not present
is an incentive to include an IEQ performance indi- or designed for higher ventilation rates. At the same
cator in the EP Certificate and by doing so, include time school building management is confronted
the energy use of absent or underperforming building with high energy costs and the requirement to install
systems in the EP. In the expected EPBD this is still CO₂ sensors in classrooms to stimulate better (often
softly addressed in art.4 (related to MEPS), and in natural) ventilation. These are challenges which are
art.6 (New Buildings) IEQ issues shall be addressed. easier to cope with if during the last 10–20 years school
Annex 1 art.2 says: … indoor, conditions, and in order building operators had been more responsive on the
to optimise health, indoor air quality and comfort widely reported poor IEQ in school buildings.
levels defined by MS’s. We hope that this language
will stay, or even better, become stronger. We all know But we have to be realistic, these reduced comfort levels
that the health costs due to poor IEQ in buildings is and higher costs are small offerings compared to the
much higher than the energy use possibly related to suffering of the people of Ukraine.
a better IEQ!
The current situation in Europe, the high fuel prices

and shortages due to Europe’s dependency on Russian
fuel supply and the war in Ukraine requires many
people to accept lower thermal comfort levels during
this winter season. In many countries the government
advises or requires lower temperatures in buildings JAAP HOGELING
during the winter. This will lead to thermal discom- Editor-in-Chief
fort at workplaces and in houses where people cannot REHVA Journal
4 The REHVA European HVAC Journal — October 2022

TOP PAPERS
CLIMA 2022 conference papers on the

theme HEALTH & COMFORT
Twenty CLIMA 2022 papers have been acknowledged as high ranking on the theme HEALTH & COMFORT.
Ten are included in this issue:
Evaluation of preventive measures in mitigating the risk Low-temperature radiant cooling panel for hot and
of airborne infection of COVID-19 humid climate
Yunus Emre Cetin, Martin Kriegel Gongsheng Huang, Nan Zhang, Yuying Liang
Aerosol transmission in rotary wheel heat exchangers Impact of future climate on the performances of ground-
Heinrich Huber, Thomas Richter, Florian Brzezinski, Michael Riediker source cooling system
Abantika Sengupta, Pieter Proot, Tom Trioen, Hilde Breesch,
Development of a non-contact modular screening clinic Marijke Steeman.
(NCMSC) for COVID-19
Jinkyun Cho, Jinho Kim, Jongwoon Song, Seungmin Jang A Multi-Domain Approach to Explanatory and Predictive
Thermal Comfort Modelling in Offices
Thermal inactivation of the corona virus (SARS-CoV-2) in Eugene Mamulova, Henk W. Brink, Marcel G.L.C. Loomans,
air volumes Roel C.G.M. Loonen, Helianthe S.M. Kort
André Schlott, Thomas Hutsch, Eileen Sauer, Jens Wetschky,
Jana Hessel, Susanne Bailer, John Laubert, Stefan Lösch Exploring futures of summer comfort in Dutch households
Lenneke Kuijer, Lada Hensen Centnerová
Health and energy assessment of a demand controlled
mechanical extraction ventilation system Correlation of subjective and objective air quality data in
Janneke Ghijsels, Klaas De Jonge, Jelle Laverge shopping centers as a function of air temperature and
relative humidity
Mahmoud El-Mokadem, Kai Rewitz, Dirk Müller
The other ten papers can be accessed on-line::
Droplet Concentration Produced during Expiratory Investigation of PECS on the basis of a virtual building
Activities and Evaluation of Relative Infection Risk controller
Arisu Furusawa, Takashi Kurabuchi, Jeongil Kim, Masaki Shimizu, Katharina Boudier, Sabine Hoffmann
Haruki Taguchi
(Best paper award for the theme Health and Comfort) Investigation of the group differences in indoor
environmental quality
Control of the bed thermal environment by a ventilated Zheng Li, Ongun Berk Kazanci, Qingwen Zhang Bjarne W. Olesen
mattress
Ge Song, Mariya Petrova Bivolarova, Guoqiang Zhang, Arsen Self-reported health and comfort of outpatient workers
Krikor Melikov in six hospitals
AnneMarie Eijkelenboom, Philomena M. Bluyssen
Cooling effects on fatigue recovery during summer sleep
Noriko Umemiya, Mitsunori Suzuki Temperature calibration and Annual performance of
cooling for ceiling panels
Dynamics of metabolic rate in male individuals due to Seyed Shahabaldin Seyed Salehib, Karl-Villem Võsab,
the meal and regular office activities Jarek Kurnitskia, Martin Thalfeldta
Dolaana Khovalyg, Jiyoung Kwak
Thermal comfort perception and indoor climate: Results
Estimating Long-Term Indoor PM2.5 of Outdoor and from the OPSCHALER project
Indoor Origin using Low-Cost Sensors Arjen Meijer, Anastasia Petropoulou, Andrea Joseph Thaddeus,
Tongling Xia, Yue Qi, Xilei Dai, Ruoyu You, Junjie Liu, Chun Chen Laure Itard
The REHVA European HVAC Journal — October 2022 5

Swegon
Our need for knowledge
is greater than ever
To meet the challenges of our business, we at Swegon believe in

constantly upgrading our knowledge and going beyond our traditional
engineering skills. Whether it’s about decreasing the carbon footprint
or improving the indoor environments of our buildings, we are happy to
share what we have learnt during 90 years in the business. But we also
provide the award-winning platform Swegon Air Academy, connecting
experts from across the field of indoor climate to share knowledge with a
wider audience.
We believe this is more important than ever, and so we have just
launched a new Swegon Air Academy website – have a look and get
inspired.
swegonairacademy.com.
TOP PAPERS
Evaluation of preventive measures

in mitigating the risk of airborne
infection of COVID-19
YUNUS EMRE CETIN MARTIN KRIEGEL

Hermann-Riestchel-Institut, Hermann-Riestchel-Institut,
Chair of Energy, Comfort Chair of Energy, Comfort
and Health in Buildings, and Health in Buildings,
Technische Universitaet Technische Universitaet
Berlin, Berlin, Germany, Berlin, Berlin, Germany
[email protected]
Abstract: In this study, a systematic approach for estimating the infection probability
under different infection control strategies is presented for several indoor cases. Increased
airflow rates, ventilation schemes, air cleaning equipment, disinfection systems, and face
masks are considered according to existing guidelines and standards. These strategies
are implemented to care facilities, schools, and offices with varying scenarios. Infection
probability calculations are conducted based on the widely used Wells-Riley model. The
possible variation of the input parameters is evaluated by employing the Monte Carlo
approach to increase the representativeness of the findings. Results show that the infection
risk reduction of 15 to 99% is possible depending on the measure preferred.
Keywords: Indoor air quality, COVID-19, preventive measures, Wells-Riley model.
1. Introduction
risk modelling for ventilated indoor enclosures. In
The novel coronavirus COVID-19 pandemic impinged general, the infection risk is characterized by a prob-
millions of people [1]. Overwhelming numbers of ability between 0 to 1. Models preferred can provide a
reported cases brought into attention the importance quantitative risk assessment to deal with the ongoing
of preventive strategies to alleviate the propagation of epidemic and help to comprehend possible results
extremely infectious diseases. Multiple guidelines have of varying circumstances. The Wells-Riley model is
raised the concern on the indoor airborne transmission a simple and quick approach based on the quantum
of COVID-19 and many recommendations have been concept, which also considers infectivity and source
released by organizations [2–4]. These recommenda- strength. The infection risk prediction with the Wells
tions and strategies in terms of building, room, and Riley assumes that the pathogens are homogeneously
personal scale have been reviewed and discussed from distributed in a room. The dose-response model on
different perspectives [5–7]. In this context, infection the other hand can provide more precise and realistic
risk assessment is considered a useful tool that may outputs than the Wells-Riley model. Nevertheless, this
help to quantify and compare the effectiveness of cor- model is less handy since it requires infectious dose data
responding infection control measures. to construct the dose-response relationship [10,11].
Wells-Riley [8,9] and dose-response models are The Wells–Riley model and its modifications have been
known as two fundamental approaches in infection extensively used for the investigation and evaluation of

TOP PAPERS
the infection risk of numerous ventilated environments In this equation, ηs and ηI represents the mask filtra-
from different perspectives [12–17]. In corresponding tion efficiency of the susceptible and infected persons
studies, key parameters of the Wells-Riley model like respectively. V is the volume of the room (m³) and
quantum generation and breathing rate are evaluated α expresses the equivalent air change rate (given in
mostly as a constant. In fact, these fundamental param- Eq.3) which depends equivalent ventilation air change
eters have a varying character and considering them rate (λvent), inactivation rate of ultraviolet germicidal
as a constant may result in misleading conclusions. irradiation (kUV), and natural inactivation (kinact).
Also, the effect of preventive measures on infection
risk mitigation was rarely inspected and compared in 𝛼𝛼 = 𝜆𝜆𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣 + 𝑘𝑘𝑈𝑈𝑈𝑈 + 𝑘𝑘𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 (3)
a quantitative way. Hence, the motivation of this study
is to evaluate the effect of different infection preven- The equivalent ventilation rate (λvent) includes the air
tive strategies by employing the Wells-Riley model supply rate of the HVAC system (λHVAC) and portable
in which the probability distributions of unknown air cleaners (λPAC). Here, in order to reflect the imper-
parameters are considered. For this purpose, the sto- fect mixing case in different ventilation concepts like
chastic Monte Carlo approach is used to broaden the displacement ventilation an additional ventilation
representativeness of the results. The effect of displace- parameter (εHVAC) is also included to this equation as
ment ventilation, standalone air cleaners, installing seen below. This additional ventilation parameter is
partition, upper room UVGI systems, and wearing equal to one for a perfect mixing situation which is one
N95 masks are evaluated. The findings can be used in of the main assumptions in Wells-Riley consideration.
the ongoing struggle against COVID-19 by helping A similar factor is also employed in another modifica-
to understand effective countermeasures in infection. tion of Wells-Riley by Sun and Zhai [14].
𝜆𝜆𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣 = 𝜆𝜆𝐻𝐻𝐻𝐻𝐻𝐻𝐻𝐻 𝜀𝜀𝐻𝐻𝐻𝐻𝐻𝐻𝐻𝐻 + 𝜆𝜆𝑃𝑃𝑃𝑃𝑃𝑃 (4)

2. Methods
2.1 Infection risk model
The Wells-Riley equation is mainly described as Air supply rate of the HVAC system (λ HVAC) is
follows: composed of supplied air flow rate of outdoor air
𝐷𝐷 (λoutdoor) and recirculated air (λrecirculated). It is given
𝑃𝑃0 = = 1 − 𝑒𝑒 −𝐼𝐼𝐼𝐼𝑄𝑄𝑏𝑏 𝑡𝑡/𝑄𝑄 (1)
𝑆𝑆 in the following form:
where P 0 is the probability of infection, D is the 𝜆𝜆𝐻𝐻𝐻𝐻𝐻𝐻𝐻𝐻 = 𝜆𝜆𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜 + 𝜆𝜆𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟

number of cases, S is the number of susceptible, I is (5)
the number of infectors, q is the quanta emission rate
by one infector (quanta/h), Qb is the breathing rate where ηfilter is the filtration efficiency of the filters.
of the susceptible person (m³/h), and Q is the volume
flow rate of pathogen free air (m³/h). In this study, the 2.2 Cases considered
above version of the Wells-Riley equation is modified Three different base cases namely an elderly nursing
to include the use of N95 masks, air cleaners, displace- home, a waiting area at the doctor’s office, and a
ment ventilation, partition, and UVGI system. The classroom are evaluated. Layouts, occupancy levels,
modified equation is given as follows: duration of stay, ventilation configurations are assigned
based on literature and the most typical real practices.
𝐼𝐼𝐼𝐼𝑄𝑄𝑏𝑏 𝑡𝑡 Definitions and details regarding corresponding cases
𝑃𝑃 = 1 − 𝑒𝑒 −(1−𝜂𝜂𝑠𝑠 )(1−𝜂𝜂𝐼𝐼 ) 𝑉𝑉𝑉𝑉 (2)
are given in Table 1.
Table 1. Settings of the studied cases.

Space Elderly nursing Waiting area at Music lesson in Corridor in a Gym in
home doctor’s office a classroom school a school
Duration of stay (min) 60 60 60 15 90
Number of total people 2 10 25 40 25
Volume of space (m³) 3x4x2.7 4x5x2.7 5x8x3.2 30x1.25x3.2 15x27x5.5
Outdoor ventilation rate (l/s) 8.6 44 137 452.5
Quanta generation (h−1) 58±31 58 ± 31 970±390 251±134 492±270
Breathing rate (m³/h) 0.3±0.2 0.3± 0.2 1.3±0.85 1.3±0.85 2.5±1.75

TOP PAPERS
2.3 Model parameters

• Increased outdoor ventilation rate is analysed by
Two critical parameters in the Wells-Riley equation employing 100% outdoor air.
are quanta generation rate (q) and breathing rate(p).
The quantum generation rate depends on disease • Air distribution patterns affect the ventilation factor
type, infector activity, etc., and varies significantly (εHVAC) considerably. Displacement ventilation has
[18]. In this study, the quantum generation rate is the potential to reduce the exposure in the breathing
adapted from the studies of Shen et al. [19], Millet region so it is considered with a factor of 1.2 to 2
et al. [20], and Hartmann et al. [21]. Breathing rates [24].
are assigned up to the activity levels. In each scenario,
only one infectious pathogen emitter exists. It is also • Installing partition is considered by a factor of
assumed that the infectious aerosols become evenly 1.1 – 3 [24].
distributed throughout the space promptly. Quantum
generation and breathing rates are assumed to follow • Portable air cleaners are becoming popular recently.
the normal distribution. Variations of these inputs are The use of such air cleaners is assumed to supply
applied by using the stochastic Monte Carlo approach clean air with a rate of 12 m³/h per square meter,
on the calculations. In every setup, 50,000 trials are which is suitable with the current EPA guide [25].
simulated.
• Proper use of an upper room UVGI system is
Minimum outdoor ventilation rates are calculated in assumed to provide an air change rate of 2 to 6 h−1
accordance with Ashrae Standard 62.1 [22]. In these [26].
calculations space area and occupant number are taken
into account as seen in the following form: • N95 masks can filter the droplets significantly. Fil-
tration efficiency for both susceptible and infected
persons is assumed as between 70% to 95% [27,28].
𝜆𝜆𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜 = 𝑅𝑅𝑝𝑝 𝑃𝑃𝑧𝑧 + 𝑅𝑅𝑎𝑎 𝐴𝐴𝑧𝑧 (6)
3. Results and discussion
where Rp is the outdoor airflow rate required per The calculated infection rates for the base cases are
person (L/s), Pz is the occupant number, Ra is the shown in Table 2. The infection probabilities over
outdoor airflow rate required per unit area (L/s), and 10% are considered as high risky spaces and these
Az is the net floor area (m²). Required airflow rates values are bolded. As it is observed, infection rates
are determined by the minimum ventilation rates indicate a large variation in considered cases. The least
presented in Ashrae Standard 62.1 [22]. risky space is found as the waiting area at the doctor’s
For the base cases mixing ventilation (εHVAC = 1) is

applied. The fraction of outdoor ventilation on the
air supply rate of the HVAC system is specified as Table 2. Infection probabilities for base cases.
25% [23]. The filtration efficiency of the filters (ηfilter)
in recirculation is considered as 70% [22]. Natural
inactivation is assumed to have a uniform distribution Space Infection probability (%)
between 0 and 1 h−1.
Mean SD
For the base cases mixing ventilation (εHVAC = 1) is
applied. The fraction of outdoor ventilation on the Elderly nursing home 14 11
air supply rate of the HVAC system is specified as
25% [23]. The filtration efficiency of the filters (ηfilter)
Waiting area at doctor’s office 3.4 2.8
in recirculation is considered as 70% [22]. Natural
inactivation is assumed to have a uniform distribution
between 0 and 1 h−1. Music lesson in a classroom 40 23
At first, infection probability is calculated for the base Corridor in a school 32 22.6
cases. Then, six different mitigation strategy is applied,
and the effect of these strategies is evaluated individu-
Gym in a school 24 18.4
ally. Proposed strategies are as follows:

TOP PAPERS
office. As a result of the excessive quanta gen-

eration rate the music class configuration shows
the highest infection risk potential (40%) among
evaluated scenarios. As a result, it can be said that
without any mitigation strategy all the cases apart
from the waiting area show a considerable risk in
terms of infection probability.
Infection risk probabilities under different miti-

gation measures for the nursing home, waiting
area, classroom, corridor, and gym are depicted
in Figure 1, Figure 2, Figure 3, Figure 4, and
Figure 5 respectively. Mean value of the situ-
ational reproduction number (RS = P·S) [28] in
each case is also given in these figures. RS points
out the infection spreading in community. If
Figure 1. Infection risk predictions for the nursing home. RS > 1 it is considered that an epidemic occurred.
In general, it is seen that all the measures help to

decrease the infection probability to some extent.
For the nursing home shown in Figure 1, the
average infection probability decreases about
17% when the supply air is 100% outdoor
originated. Nevertheless, in this case, the average
infection risk is still higher than the threshold
level with an 11.6% infection probability. All the
other measures help to reduce the infection risk
between 30% to 98%. In all these cases average
infection risk is reduced to below 10% and the
limit is met. Also, since the RS > 1 in all cases the
spread of the disease is unlikely for the nursing
home.
As seen in Figure 2, in the case of the waiting

area considered measures alleviated the infection
Figure 2. Infection risk predictions for the waiting area.
probability in the range of 23-99%. In this case,
the lowest risk reduction is obtained with the use
of 100% outdoor air, and the highest reduction
is with the N95 face masking as expected. Both
the values of infection probability and RS point
out that the lowest risk is obtained for the waiting
area at doctor’s office.
Infection risk predictions for the classroom,

corridor and the gym are given in Figure 3,
Figure 4 and Figure 5 respectively. It is deduced
from the findings that the evaluated measures are
not adequate in terms of attaining the threshold
limit mostly. In all three cases high reproduction
number (RS > 1) points a risk of serious outbreak.
In every evaluated configuration, the use of N95
masks meets the threshold value and stands as
the only solution for the lowest infection prob-
Figure 3. Infection risk predictions for the classroom. ability. Still, it should be noted that it might not

TOP PAPERS
be a feasible solution in a music class. Also,

for a deeper analysis of the mask efficiency
leakage during inhalation and exhalation can
be considered in terms of personal protection
related factor [28]. For an efficient reduction
of the infection risk, the combined effect of the
multiple measures can also be evaluated with
including feasibility and cost considerations.
4. Conclusions
In the present paper, possible infection preven-
tive measures were analysed for five different
settings by employing the well-known Wells-
Riley model. Increased outdoor ventilation
rate, displacement ventilation, installing parti-
tion, portable air cleaners, UVGI systems, and
N95 face masks were evaluated. Related model
parameters were determined based on the litera-
Figure 4. Infection risk predictions for the corridor. ture and practices. The stochastic Monte Carlo
approach was used in calculations in order to
include the variations of the input parameters.
Future studies can evaluate the combined effect
of different risk-mitigating factors from a feasi-
bility and cost standpoint.
Important outcomes are summarized below:
• Predicted infection risk values show a devi-

ating figure depending on the boundary
conditions of the cases.
• Based on the evaluated measures risk reduc-
tion is possible between 15.5 to 99%.
• Infection risk-mitigating measures lower the
probability although this may not be suf-
ficient to achieve the predetermined limit
for some cases.
• The use of N95 masks may reduce the infec-
tion risk remarkably. This potential can be
Figure 5. Infection risk predictions for the gym. considered as an easy option for complicated
cases at first instance.
5. Acknowledgement
This study is supported by The Federal Institute for Research on Building, Urban Affairs and Spatial Development with funding
code of SWD-10.08.18.7-20.02.
6. References
Please find the full list of references in the original article at https://proceedings.open.tudelft.nl/clima2022/article/view/299

TOP PAPERS
Aerosol transmission in rotary wheel

heat exchangers
HEINRICH HUBER THOMAS RICHTER FLORIAN BRZEZINSKI MICHAEL RIEDIKER

Institute of Building Hoval Aktiengesellschaft, Institute of Building SCOEH: Swiss Centre
Technology and Energy, Vaduz, Lichtenstein Technology and Energy, for Occupational and
Lucerne University of Lucerne University of Environmental Health,
Applied Sciences and Applied Sciences and Winterthur, Switzerland
Arts, Horw, Switzerland Arts, Horw, Switzerland
[email protected]
Abstract: Transmission by aerosols is considered the main route of COVID-19 infections

indoors. Therefore, limiting air transfer between supply and extract air in ventilation systems
is critical. Rotary wheels are very efficient heat recovery devices, but have a higher exhaust
air transfer ratio (EATR) compared to other types. This raises the question whether a
relevant transfer of aerosols can take place and whether this is different from the EATR.
Experimental investigations were carried out with an aerosol whose properties are well
comparable to human lung aerosols. In parallel, the EATR was determined with tracer
gas. For Rotary wheels without purge sector the aerosol transfer ratio was typically 1 to
2 percentage points below the EATR. The results allow the conclusion that rotary wheels
designed and operated according to current standards transfer only a non-relevant small
amount of aerosols and thus do not pose an infection risk for COVID-19 in applications such
as offices where the frequency of highly infectious individuals is low to moderate. However,
aerosol transmission in hygienically relevant quantities is conceivable in poorly designed
systems, but this does not only affect rotary wheels.
Keywords: Ventilation, aerosol transmission, heat recovery, rotary wheel
1. Introduction
a higher exhaust air transfer ratio (EATR) can occur
Aerosols are considered a main cause of COVID-19 infec- than with other common heat recovery categories such
tions indoors. Therefore, limiting air transfer between as plate heat exchangers or run around coil systems.
supply and extract air in ventilation systems is critical. Measures to minimise and evaluate the EATR of
RHEs are well known and recommended e.g. in the
In bidirectional ventilations systems, heat recovery is REHVA COVID-19 Guide [1]. Special attention must
state of the art and is even required in the European be paid to the correct pressure ratios and purge sector.
regulation for ecodesign requirements for ventilation However, RHEs are not generally equipped with purge
units. Rotary heat exchangers (RHE) are an efficient sectors. Another aspect is the fact that the surface of
and economically interesting solution and are therefore an RHE is touched by both supply and extract air.
widely used. A disadvantage of RHE is that due to the Although this enables humidity recovery, it also carries
physical principle and the mechanical implementation the risk of transferring undesirable substances.

TOP PAPERS
The characteristics of RHEs raise the question of whether 1 – 3 m/s. On both sides air inlet temperatures were
relevant aerosol transfer can take place and whether this 20°C ± 1K and air inlet humidity 40% RH ± 10% RH.
differs from EATR. For hygienically demanding applica-
tions of RHEs, it is crucial to have more knowledge In order to make a conclusion of the aerosol transmis-
about the phenomenon of aerosol transfer. On the ini- sion, the EATR and the aerosol transfer ratio (ASTR)
tiative of a company, the University of Applied Sciences were determined and compared.
and Arts Lucerne (HSLU) together with the Swiss
Centre for Occupational and Environmental Health 2.2. Exhaust Air Transfer Ratio (EATR)
(SCOEH) carried out experimental investigations [2]. The EATR was measured based on EN 308:2022 [3]
by injecting Sulphur Hexafluoride (SF6) tracer gas in
the duct of the exhaust inlet. The EATR is calculated
2. Research methods
with the following formula:
2.1. Principle and test rig
The Building Technology Laboratory of the HSLU 𝑎𝑎22 − 𝑎𝑎21
𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸 = × 100 [%] (1)
operates a test rig for heat recovery devices. In Figure 1 𝑎𝑎11 − 𝑎𝑎21
the scheme of the test rig is shown. In all four air
where
streams temperature, humidity, air flow rate and tracer
gas concentration can be measured. For the aerosol a11 is SF6 concentration exhaust inlet [ppm],
measurements, a particle generator was installed in a21 is SF6 concentration supply inlet [ppm],
the exhaust air inlet duct and additional measuring
a22 is SF6 concentration supply outlet [ppm].
devices in all four air streams.
2.3. Aerosol Transfer Ratio (ASTR)
Test objects were two RHEs without purge sector and
with a free diameter of 1 000 mm: a condensation The test setup for the ASTR measurements was done
rotor (aluminium) and a sorption rotor (aluminium in the same way as for the EATR. Aerosol measuring
with molecular sieve coating). The nominal airflow devices (Sensirion SPS30) were installed in pairs in each
is between 1 400 to 4 199 m³/h for a face velocity of of the four air streams. The sensors count the particulate
TR PR
Tracer gas 901 902
injection
QR QR
Fog machine 104 204
Aerosol
SF6
MR
particles 103 MR
PDR
generator 504 203
TR PDR
FR 102 501 TR FR
101 202 201
Air conditioning
exhaust inlet static mixers
Fan Fan
1.2
1.1
PDR Rotor Air conditioning
503 supply inlet
Fan Fan
2.2 2.1
TR FR TR FR
402 401 PDR 302 301
502
MR MR
403 303
QR QR
404 304
Figure 1. Scheme of the test rig for heat recovery devices. Symbols: first letter F air flow rate, P pressure,
T temperature, M humidity, Q concentration; subsequent letter D difference; auxiliary letter R registration;
air flow type 1.1 exhaust inlet, 1.2 exhaust outlet; 2.1 supply inlet, 2.2 supply outlet.

TOP PAPERS
3. Results
matter based on laser scattering in a size of 0.3 to 10 μm
3.1. Condensation rotary heat exchanger
and provide a good assessment of concentrations over
a wide range [4]. To keep the aerosol concentration in Table 1 shows the results of the condensation RHE. The
the supply inlet air as low as possible, the installed filters measurements were conducted with three different air
of the class ISO ePM1 50% (F7) were replaced with face velocities (v) of 1, 2 and 3 m/s. Rotor speed (n) was
HEPA filters H14. During the measurement, aerosol 20 rpm and pressure difference between supply outlet
was applied in pulses in the duct of the exhaust inlet air. and exhaust inlet (Δp22-11) 10 Pa. One additional meas-
The aerosol used has an average diameter of just over uring point (MP) with a difference of 250 Pa between
one micrometer and is therefore comparable in size to supply outlet and exhaust inlet was conducted.
exhaled aerosol [5]. Like human exhaled aerosol, the
used aerosol is liquid at normal ambient temperatures, Figure 2 shows the EATR and ASTR as function of
hygroscopic and moderately viscous. It is produced by the face velocity. Even if the ASTR plus the upper
evaporation and condensation of a water-glycol mixture measurement uncertainty range partly exceeds the
by a fog machine and is stable in air for a longer time [6]. EATR, there is a clear tendency for the ASTR to be
smaller than the EATR.
The maximum peak heights at exhaust air inlet, supply
air inlet and outlet were used for evaluation the ASTR 3.2. Sorption rotary heat exchanger
according following formula: As with the condensation RHE, measurements with
the sorption RHE were carried out at three different
𝑏𝑏22 − 𝑏𝑏21 face velocities and a pressure difference of 10 Pa.
𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 = × 100 [%] (2)
𝑏𝑏11 − 𝑏𝑏21 Additional to the measurements with 20 rpm, meas-
urements with 10 rpm rotor speed were conducted.
where
Table 2 shows the results of the sorption RHE.
b11 is peak PM10 exhaust inlet [P/cm³],
b21 is peak PM10 supply inlet [P/cm³], Analogous to the results of the condensation RHE
Figure 3 shows that also with the sorption RHE the
b22 is peak PM10 supply outlet [P/cm³].
ASTR is below the EATR.
2.4. Measurement uncertainty
4. Discussion
For both EATR and ASTR, the measurement uncertainty
was determined by repeating each test point usually five 4.1. Significance of the results
times. Since linearity is given for both measurement AHUs that comply with the European Ecodesign
equipment, the measurement uncertainty can be deter- Regulation are typically designed for a face velocity
mined by the standard deviation of EATR and ASTR (related to the inner cross-sectional area of the casing)
and a coverage factor to reach a confidence level of 95%. of about 1.6 to 1.8 m/s. Since the face area of the of
12 12
EATR, dp = 10 Pa EATR, dp = 10 Pa
10 ASTR, dp = 10 Pa 10 ASTR, dp = 10 Pa
8 EATR, dp = 250 Pa 8 EATR, dp = 250 Pa

EATR, ASTR [%]
EATR, ASTR [%]
ASTR, dp = 250 Pa ASTR, dp = 250 Pa

6 6
4 4
2 2
0 0
1 2 3 1 2 3
Face Velocity [m/s] Face Velocity [m/s]
Figure 2. EATR and ASTR of the condensation rotor as Figure 3. EATR and ASTR of the sorption rotor as
function of the face velocity at isothermal conditions function of the face velocity at isothermal conditions
20°C, 40% RH. 20°C, 40% RH.

TOP PAPERS
4.2. State of the art in filters

the RHE is about 10 to 20% smaller than the cross-
sectional area of the AHU, the nominal face velocity In German speaking countries the VDI 6022-1 [8] is
of a RHE is typical approx. 2 m/s. In applications considered reflecting the state of the art. The minimum
such as office buildings and schools, face velocities in filter class for supply air is ISO ePM1 50%. The extract
the range of about 1 m/s often occur in partial load air before entering a RHE shall pass a filter of class
operation mode. On the other hand, the REHVA ISO ePM10 50%. As an estimate for the separation
COVID-19 guideline recommends not using partial efficiency of lung aerosols the gravimetric arrestance of
load operation in pandemic situations. Therefore, the the filters up to a particle size of PM10 can be assumed
measured EATR and ASTR at 2 m/s are considered to be 85% for class ISO ePM1 50% [9] and 50% for
for the infection risk assessment. class ISO ePM10 50% (according to the definition
of the class).
For condensation RHEs the rotor speed is in a range
of 10 to 20 rpm, depending on design characteristics, A typical solution is that in the outdoor air (before
e.g. foil thickness. For sorption RHEs 20 rpm can be entering the RHE) and in the extract air (before
seen as typical speed. Therefore, for the contamina- entering the RHE) an ISO ePM1 50% filter is placed.
tion risk assessment for both RHE types the measured Thus, the potentially contaminated air passes through
values at 20 rpm are chosen. For the calculation of the only one filter.
contamination risk an ASTR of 4% is used, which is
to understand as conservative value. 4.3. Estimation of the ASTR including filter
The aerosol transfer ratio of an AHU ASTRAHU can be
It goes without saying that these data are valid only for estimated as follow:
the tested RHE when using this specific fog aerosol.
Nevertheless, the RHEs examined are real products 𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 = 𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝑅𝑅𝑅𝑅𝑅𝑅 ∙ (1 − 𝑓𝑓𝐹𝐹,𝑒𝑒𝑒𝑒𝑒𝑒 ) ∙ (1 − 𝑓𝑓𝐹𝐹,𝑠𝑠𝑠𝑠𝑠𝑠 ) (3)
that are judged to be typical in a market comparison.
where
The draft revision of the European Ecodesign Regulation
from [7] serves as a comparison. In this draft a maximum ASTRRHE is the ASTR of the RHE, acc. to Equation (2),
EATR at nominal flow and nominal pressure of 5% is fF,eta is the separation efficiency of lung aerosols of
required. With reference to this source, a maximum an extract air filter (positioned before RHE),
EATR of 5% can be considered state of the art. fF,sup is the separation efficiency of lung aerosols of a
supply air filter (positioned after RHE).
However, for hygienically sensitive applications, such
as public buildings, the authors recommend taking
measures to achieve a lower EATR. For RHE with a With the values shown in chapters 4.1 and 4.2, an ISO
purge sector, an EATR below 0.5% can be achieved. ePM1 50% filter in extract air and no additional filter
Table 1. EATR and ASTR, condensation wheel.
MP, - v, m/s n, rpm Δp22-11, Pa EATR, % ASTR, %

1.1 1 20 10 10.4 ± 0.3 7.5 ± 3.7
1.2 2 20 10 4.7 ± 0.2 3.8 ± 2.3
1.3 3 20 10 3.3 ± 0.2 2.4 ± 0.6
1.4 3 20 250 2.6 ± 0.2 1.6 ± 0.2
Table 2. EATR and ASTR, sorption wheel.

MP, - v, m/s n, rpm Δp22-11, Pa EATR,% ASTR,%
2.1 1 20 10 8.2 ± 0.3 6.3 ± 2.9
2.2 2 20 10 4.7 ± 0.2 2.7 ± 0.5
2.3 3 20 10 3.2 ± 0.2 1.7 ± 0.3
2.4 1 10 10 4.5 ± 0.1 2.8 ± 0.8
2.5 2 10 10 2.3 ± 0.2 0.9 ± 0.2
2.6 3 10 10 1.5 ± 0.2 0.6 ± 0.2

TOP PAPERS
in supply air, the result is an APTRAHU of approx. 0.006 (breathing air volume 0.6 m³/h), the amount of inhaled
or 0.6% respectively. This applies to RHEs without viruses is two orders of magnitude below the critical value
purge sectors. With purge sector, values 10 times lower for an infection risk. However, in a hospital situation with
can be achieved. many highly emitting patients, the situation could rapidly
become critical. Thus, the recommendation to use heat
4.4. Estimation of the infection risk recovery systems without any risk of exhaust air transfer
The amount of virus released into the air can be done to supply air (e.g. tight plate heat exchanges or run around
by combining the concentration of viruses in the coil systems) in such settings is well warranted.
lung lining liquid with the size-distribution of micro-
droplets released during breathing and speaking and Eurovent 6/15 - 2021 [13] deals comprehensively and in
by taking into account the proportion that sediments detail with the prevention of air leakage in air handling
rapidly [10]. To estimate the viral concentration at units. According to this, an elementary measure is to ensure
steady state in different situations, we used an indoor correct pressure conditions in air handling units. This
scenario simulator that is based on this emission concept depends primarily on the fan positions. For hygienically
[11]. In all the calculated scenarios, we assumed that sensitive applications, a purge sector is also recommended.
a very contagious person (a so-called super-emitter)
is in a room of 100 m³ volume and 3 air changes per It should be mentioned that the risk of exhaust air
hour. We then simulated a quiet office, a loud office transfer also exists with other leakages in AHUs and
(e.g. call centre) and a hospital situation with coughing ventilation systems that are not the subject of this paper.
COVID-19 patients. We further assumed that 1000
people are in this building with a ventilation flow rate
5. Conclusion and outlook
of 30 000 m³/h. For the general population, a very
high infection rate of 1% is assumed, for the hospital The project investigated how aerosols, which behave
wing with COVID-19 patients a rate of 50% (50% similarly to human lung aerosols, are transferred in rotary
patients, 50% staff ). Table 3 summaries the simulated heat exchangers without a purge sector under isothermal
virus concentrations in these scenarios in the individual conditions. In the measurements on a condensation
rooms and in the extract air of the building. rotor, the exhaust air transfer ratio of the aerosols was
around 1 percentage point lower than the exhaust air
The concentrations indicate viral copies as assessed transfer ratio (EATR) of tracer gas. For a sorption rotor
by RNA assays. For the Delta variant about 1 in 300 it was 2 percentage points lower. The results suggest that
and for the Omicron variant about 1 in 100 of these there is no risk of the used aerosols being transferred
copies was found to be able to infect cells [12]. Thus, through the matrix with these rotary heat exchangers.
doses above 300 or even only 100 virus-copies seems
to be critical for viral infections. This is supported by In the follow-up project further measurements with dif-
simulation of super-spreading events where the virus ferent products were carried out. The test report [14] is
dose (the amount taken up) was estimated in the range freely available, but the results have not yet been prepared
of a few thousand viruses. for publication in a journal. However, it can be said that
for rotors without a purge sector, the ASTR was always
For the example of the call centre, it can be said that with below the EATR, even with different air conditions. With
an ASTRAHU of 0.6% and the virus concentrations in the a purge sector, EATR and ASTR were measured at a low
extract air according to Table 3, the virus concentration level of 0.0 to 0.5%. Taking into account the measure-
in the supply air is 0.2 copies/m³. Even if the personnel ment uncertainty, the EATR and the ASTR with purge
directly inhale the supply air during an 8-hour shift sector can be described as approximately the same.
Table 3. Simulated virus concentrations (copies/m³) in

individual rooms and in the extract air in different scenarios. 6. Acknowledgment
The authors would like to thank the company Hoval for
Scenario Infection Steady state Extract air, initiating the project, the technical exchange, and the
rate, in room, Copies /m³ financial support.
% Copies /m³
Quiet office 1 1200 1.2 7. References
Call centre 1 40 000 40 Please find the full list of references in the original article
at: https://proceedings.open.tudelft.nl/clima2022/article/
Hospital* 50 500 000 250 000
view/391
*COVID section

Uponor PEX pipes
50 years
and still in love
RESISTANT, VERSATILE In 1972, the first Uponor PEX pipes were launched worldwide and revolu-
AND EASY TO INSTALL tionized the entire industry. 50 years later, they remain resistant and
versatile. Uponor PEX Pipes, however, continue to evolve and create the
basis for further sustainable innovations, such as the bio-based
Uponor PEX Pipes Blue in 2022. Find out more about Uponor PEX pipes:
www.uponor.com/50-years-of-uponor-pex-pipes
Uponor
TOP PAPERS
Development of a non-contact modular

screening clinic (NCMSC) for COVID-19
JINKYUN CHO JINHO KIM JONGWOON SONG SEUNGMIN JANG

Dept. of Building and Dept. of Fire Protection, Energy Solution Dept. of Building and
Plant Engineering, Safety and Facilities, Technology, E-SOLTEC Plant Engineering,
Hanbat National Suwon Science College, Co., Ltd., Korea Hanbat National
University, Korea Korea University, Korea
[email protected]
Under the global landscape of the prolonged COVID-19 pandemic, the number of individuals
who need to be tested for COVID-19 through screening clinics is increasing. However, the risk
of viral infection during the screening process remains significant. To limit cross-infection in
screening clinics, a non-contact mobile screening clinic is developed. This study investigates
aerosol transmission and ventilation control for eliminating cross-infection and for rapid virus
removal from the indoor space using numerical analysis and experimental measurements.
Keywords: COVID-19, screening clinic, cross-infection, ventilation strategy, CFD, particle

image velocimetry (PIV)
Introduction
installation, and operation of these screening clinics
In the past, most types of viruses that reached the worldwide. A novel non-contact modular screening
pandemic level were respiratory infections such as clinic (NCMSC) was developed that addresses the
influenza and coronavirus [1]. Particularly, the coro- problems of existing screening clinics and the risk of
navirus causes a pandemic every 5 to 10 years because cross-infection during the COVID-19 testing process.
the cycle is getting shorter [2]. The WHO (word health
organization) declared COVID-19 a global pandemic
Non-contact modular screening clinic
in 2020, which persists at the present. In the beginning,
medical institutions isolated symptomatic patients A NCMSC that uses biosafety cabinets and negative
from general patients through screening clinics. The pressure booths enables safe, fast, and convenient
screening clinic plays a primary role in screening COVID-19 testing. The NCMSC is a mobile modular
suspected patients of infectious diseases. At the unit that can be quickly moved, installed, and operated
COVID-19 pandemic peak, more than 600 screening in the required area depending on the COVID-19
clinics (temporary, drive-through, walk-through and testing demand. This type of medical modular facility
etc.) were installed and operated in Korea. However, can reduce the risk of cross-infection between rooms by
there are no clear criteria and guidelines for the design, achieving the airtightness performance of the structure.

TOP PAPERS
Numerical analysis
In particular, a non-contact automated system was
applied to the entire testing process, from and body A quantitative analysis of the effect of the cross-infec-
temperature measurements to specimen transport, to tion prevention and a ventilation strategy to prevent
prevent infection from the source. the transmission are needed and should be established
in the developed NCMSC. The effects of the airflow
It increases the accessibility of patients to the screening clinic velocity and room pressure control based on the opera-
and provides adequate protection for healthcare workers tion of the ventilation system on the viral transmission
(HCWs). The NCMSC is a safe medical facility equipped were investigated. The dimension of the CFD domain
with negative pressure zones, an anteroom (AR) and a was 4100 × 3000 × 2400 mm (L×W×H). Figure 2
specimen collection booth (SCB), and positive pressure shows the division of SCB, AR, and ER. Both SA
zones, such as an examination room (ER), as shown in and EA systems were applied in the SCB for effective.
Figure 1. Moreover, it implemented two-stage negative On the other hand, only a SA system of ER was set
pressure control to prevent virus leakage. The air change to 6 ACH. Only an EA system was installed in AR
rate was set to more than 12 ACH [3], [4], which is the and the airflow rate of EA was 12 ACH for Baseline
standard for an airborne infectious isolation room, and the and 30 ACH for Cases 2 and 3. Meanwhile, an EA
pressure differential was set to maintain maximum 25 Pa outlet was installed in the SCB with an EA flowrate of
or above. Subsequently, ER maintained positive pressure
and HEPA filter (PM2.5 99.97%) were applied to prevent
infection among HCWs. The total air change rate of the
SCB was set to be maximum 30 ACH for an effective
discharge of viruses. The ventilation system provides a
safe air quality and space for HCWs and individuals to be
tested against infection. Therefore, the appropriate arrange-
ment of the supply air (SA) and exhaust air (EA) outlets
of the ventilation system is an important consideration
for adequate indoor airflow. The ventilation system and
pressure differential performance should be reviewed and
the airtightness and the area of opening of the structure
should be optimized to maintain pressure differential
through numerical analysis results and prevent aerosol viral
diffusion and infection between SCB and ER. Figure 2. NCMSC mesh of the CFD model.
Figure 1. Layout of the COVID-19 NCMSC showing the optimized of SA/EA inlet/outlet locations.

TOP PAPERS
12 ACH for Baseline. In addition, two EA outlets were Table 1. CFD boundary conditions with airflow rates.
installed for Cases 2 and 3 with total EA flowrate of
30 ACH. Furthermore, Case 3 applied the SA system Ventilation system Baseline Case 2 Case 3
was installed with a flowrate of 12 ACH. The negative
pressure was controlled in SCB, and analysed with Supply (ER) 160 m³/h 160 m³/h 160 m³/h
all doors closed. The shortage of SA for EA was sup- Transfer (ER to SCB) 25 m³/h 40 m³/h 30 m³/h
plemented through door gaps of adjacent rooms, and
Supply (SCB) N/A N/A 75 m³/h
the direction of airflow was from the ER to the SCB.
Exhaust (SCB) 70 m³/h 175 m³/h 175 m³/h
The boundary conditions of the simulations are listed
Exhaust (AR) 30 m³/h 75 m³/h 75 m³/h
in Table 1. The velocity of air supplied through the SA
inlet and gap of the door, the velocity of air exhausted Transfer (AR to SCB) 45 m³/h 135 m³/h 70 m³/h
through the EA outlet, and the pressure differential
Lying manikins Uniform heat flux: 62 W, no slip boundary
between rooms were evaluated. Then, the ventilation
performance through which the virus is assumed to be Walls 2 and 1 W/m² at ceiling/floor, no slip boundary
an aerosol of SCB is predicted.
Bedside Adiabatic wall boundary condition
Airflow velocity Grid cells 8,176,419
Figure 3(a) shows the horizontal airflow velocity
Turbulence model Standard k–ε model
profile at a height of 1.5 for each Case.
Figure 3. Results of CFD numerical analysis in the NCMSC for different ventilation conditions.

TOP PAPERS
The air change rates for AR and SCB in Baseline, which -3.02 Pa, respectively. However, the values for Baseline
only applied the EA system, was set to 12 ACH. The and Case 2 are not within the appropriate range of the
velocity values were in the range of 0.0374 to 0.0506 m/s recommended pressure differential of at least -2.5 Pa
by examining the average airflow velocity distribution based on the airborne infectious isolation room [5].
for each height of the SCB, indicating that the airflow
progressed slowly and the air was gradually exhausted. Airflow considerations
In the case of AR, the air was exhausted with a similar Figure 3(c) shows the airflow streamlines across the
velocity of approximately 0.0365~0.0414 m/s with some entire MCMSC space. It is apparent that for SCB,
of the air moved to the SCB. On the other hand, the which applied both SA and EA systems in Case 3, the
air change rates for AR and SCB in Case 2, which only ventilation is active across the entire room compared
applied the EA system were set to 30 ACH. The average to Baseline and Case 2, which only applied the EA
air velocity profile for each height of the SCB was in system. The airflow velocity results of 0.0587 m/s for
the range of 0.0852 to 0.0945 m/s, indicating that the Baleline and 0.112 m/s for Case 2 were obtained by
airflow velocity was increased twice than in Baseline. The examining the overall average airflow velocity of the
air in the AR is exhausted with a velocity range of 0.0931 room, indicating that the velocity of Case 2 increased
to 0.1003 m/s with some of the air moves to the SCB. by 1.9 times than Baseline. In addition, the airflow
velocity was 0.1786 m/s for Case 3, indicating a
Furthermore, the air change rates for AR and SCB in velocity increase of 1.6 times than Case 2 and 3.0 times
Case 3, which applied both EA and SA systems in the than Baseline. It is expected that Case 3 will enhance
SCB were set to 30 ACH. The average airflow velocity the ventilation performance and facilitate an effective
range was from 0.1236 to 0.1781 m/s. The average discharge of the aerosol COVID-19 viruses.
airflow velocity profile was increased by approximately
1.7 times than in Case 2. The air in AR was exhausted
Experimental analysis
with a velocity ranging from 0.1086 to 0.1166 m/s,
and some of the air moves to the SCB. Full-scale field measurements were performed under
similar conditions used in the numerical analysis.
Pressure differential PIV (particle image velocimetry) was used to conduct
A negative pressure should be maintained in the con- experiments for airflow behavior characterization and
taminated zone (SCB) and a positive pressure should examination of the leakage area through visualization
be maintained in the clean zone (ER) to ensure that of particles simulating viruses in SCB and to verify the
the aerosol viruses in SCB do not flow to the ER. safety of the developed NCMSC against cross-infection.
Figure 3(b) shows the pressure differential between
SCB and ER for each case. It is less likely that viruses Figure 4 shows the experimental setup and perspective
migrate from SCB to the ER if the pressure is great view of the PIV set-up. Two-dimensional flow fields
between these two rooms. were measured at different positions of the camera and
laser. At Position A, the camera was installed in the ER,
The average pressure differential for Baseline, Case 2 and and the laser and oil droplet generator were installed
3 were -14.62 Pa, -18.17 Pa and -25.25 Pa, respectively. in the SCB. Moreover, at Position B, the laser was
The analysis showed that the SCB was properly con- installed at the ER and the camera and the oil droplet
trolled for all cases to maintain the negative pressure. In generator was installed in the SCB to ensure that the
addition, the effect of the cross-infection prevention of droplet came out from the mouth of the manikin,
COVID-19 entering the ER is considerably enhanced an individual to be tested. Subsequently, the exhaust
because the pressure differential increases from Baseline airflow was observed. Four different PIV measure-
to Case 3. The average pressure differential between SCB ments were performed for four different combinations,
and AR for Baseline and Case 2, 3 were -1.39, -1.87 and as shown in Table 2. First, the PIV measurement was
performed at Position A for Case 2, where only the EA
Table 2. Measuring cases with the PIV equipment. system was operated, and Case 3, where both EA and
Measure Cases Position Door SA systems were simultaneously operated in the venti-
ment between lation system of the SCB. Subsequently, the ventilation
SCB and AR performance at Position B was examined with the door
PIV A1 Case 2 A Closed between the AR and SCB closed and open for Case 3.
PIV A2 Case 3 A Closed
PIV B1 Case 3 B Closed The pressure differential can determine the effect of
PIV B2 Case 3 B Open cross-infection prevention. The pressure differentials

TOP PAPERS
of SCB and ER with the ventilation system turned on

are ΔP = -21.8 and -29.3 Pa, respectively. The measured
values and numerical analysis results were very similar.
The negative pressure in the SCB was properly main-
tained for both cases (Cases 2 and 3). The experimental
results were divided into two parts based on the location
of the PIV measurements. Moreover, Figure 5 shows
the experimental results for the vertical airflow velocity.
Figure 5(a) shows the average velocity of Case 2 for PIV

A1. The particle movement velocity in the SCB was
found to be very slow with almost no airflow for an
average airflow velocity of 0.0098 m/s. On the other
hand, the average airflow velocity of PIV A2 in Case 3
(Figure 5(b)), where SA and EA systems were operated,
was 0.0541 m/s, indicating a four-time increase than PIV
Figure 4. Experimental set-up for the PIV. A1. The make-up air was smoothly supplied to improve
the exhaust efficiency. In addition, the cross-infection
by viruses is not expected to occur since there was no
airflow from the contaminated zone (SCB) to the clean
zone (ER) in PIV A1 and PIV A2. Figure 5(c) shows
the average velocity of Case 3 for PIV B1. The average
airflow velocity was 0.0536 m/s, and is the same as that
in PIV A1 under the same ventilation conditions. Finally,
Figure 5(d) shows the average velocity of Case 3 for PIV
B2. The same condition was applied to PIV B2 as PIV
B1, but the door to the AR was opened. In this case,
the results showed that the average airflow velocity was
0.1042 m/s, and the velocity of the generated particles
increased more than twice than that of PIV B1. However,
it is a principle to close the door during specimen col-
lection. Therefore, it is recommended to operate the
ventilation system with the door open before the next
individual to be tested enters to increase the cleaning and
disinfection effect after collecting the specimen.
Conclusion
Based on the results of this study, the standards for the
installation and operation of the screening clinics are
proposed. It is necessary to implement space configura-
tion and secure airtight performance to ensure that all
tests can be performed using non-contact methods.
Acknowledgment
This work is supported by the Korea Agency for
Infrastructure Technology Advancement (KAIA) grant
funded by the Ministry of Land, Infrastructure and
Transport (22TBIP-C161800-02).
References
Please find the full list of references in the original article
view/283
Figure 5. Time-averaged air distribution under PIV cases.

For a healthy
and comfortable
room climate.
Belimo
Webinar “The 7 Essentials for Healthy Indoor Air”

We assume that the air in buildings is “clean” and not harmful to our health. This makes it all the more surprising how
little building users and operators actually know about the air quality in their spaces. Key variables such as humidity, CO2
content or VOC concentration are almost never measured, much less displayed.
To find out where the about priorities lie in creating a healthy indoor air climate, Belimo conducted a global survey among
experts of the ventilation industry and emerged seven fundamental factors for healthy indoor air in functional buildings.
Register for the Webinar “The 7 Essentials for Healthy Indoor Air” on 29.11.2022 and learn from our Expert Mikko Gisin
how Belimo sensors and room operating units contribute to a healthy indoor air.
Register now
belimo.com
BELIMO Automation AG
[email protected], www.belimo.com
TOP PAPERS
Thermal inactivation of the corona

virus (SARS-CoV-2) in air volumes
ANDRÉ SCHLOTT THOMAS HUTSCH EILEEN SAUER JENS WETSCHKY

Fraunhofer Institute for Fraunhofer Institute for Fraunhofer Institute for Fraunhofer Institute for
Manufacturing Technology Manufacturing Technology Interfacial Engineering Interfacial Engineering
and Advanced Materials and Advanced Materials and Biotechnology IGB, and Biotechnology IGB,
IFAM, Branch Lab Dresden, IFAM, Branch Lab Dresden, Stuttgart, Germany Stuttgart, Germany
Dresden, Germany Dresden, Germany
[email protected]
JANA HESSEL SUSANNE BAILER JOHN LAUBERT STEFAN LÖSCH

Fraunhofer Institute for Fraunhofer Institute for Fraunhofer Institute for Fraunhofer Institute for
Interfacial Engineering Interfacial Engineering Manufacturing Technology Manufacturing Technology
and Biotechnology IGB, and Biotechnology IGB, and Advanced Materials and Advanced Materials
Stuttgart, Germany Stuttgart, Germany IFAM, Bremen, Germany IFAM, Bremen, Germany
Abstract: To control the spread of viruses (e.g., SARS-CoV-2) and other pathogens in
a pandemic situation, slowing down the rate of spread is an essential goal that can be
achieved by interrupting transmission chains. According to the current state of knowledge,
SARS-CoV-2 is mainly transmitted by droplet infection through virus-containing aerosol
clouds in the air. These aerosol clouds are mainly produced by exhalation and can be
reduced by wearing medical masks. In closed rooms, there is an increased probability of
infection by aerosols. Countermeasures include various recently developed air cleaning
technologies. Most of these technologies available on the market are based on filters with
a limited lifetime to remove the virus load from the air or different sterilisation methods like
UV irradiation. The air cleaning technology presented focuses on the thermal inactivation of
viruses beyond their temperature sensitivity by heating the air. In the developed apparatus,
the potentially germ-carrying ambient air is sucked in and conditioned in such a way that it
is exposed to a certain temperature for a defined period of time. Before the inactivated /
hygienised air is returned to the environment, it is cooled down to almost room temperature.
The recovered heat remains in the system and is used to heat the intake air. The use and
combination of different technologies enable the most efficient air disinfection possible. Four
different experiments were conducted. After determining a base line, the air was solely

TOP PAPERS
passed through the pump, through the whole ‘Virus-Grill’ pressure free and with an elevated
pressure of 1.5 bar. In all experiments (except the baseline) the number of active viruses
were reduced below the limit of detection.
Keywords: Health, Thermal Sterilisation, Thermal Inactivation, Air Treatment, Air Cleaning
Technology, Virus Inactivation, SARS-CoV-2, COVID-19, Heat Recovery, Energy Efficiency
1. Introduction
If this targeted process is carried out continuously, it
Aerosols containing viruses can accumulate indoors can be used as a kind of “virus sink”. In order to return
and lead to superspreading events [1]. Therefore, the inactivated air to the room, the air must be cooled
slowing down the rate of spread is an essential goal down to nearly room temperature. The use of a heat
in order to control the pandemic spread of viruses pump enables efficient temperature control within
(e.g. SARS-CoV-2) and other pathogens. That can be the apparatus. The interaction of the components is
achieved by interrupting transmission chains. ensured by a regulation and control software.
Researchers all over the world work on air purifica- The aim was to construct and build up a test rig for
tion technologies. Possible air purifying technologies the proof of the inactivation of viruses using a safety
include filtration with HEPA filters, where the viruses workbench. The developed principle itself can be inte-
are removed from the air volume. This is a common grated into climatization systems for closed volumes
approach to reduce contamination in the surrounding like rooms (bureau; waiting rooms; class rooms; pas-
air. With light in UV-C spectrum (200 to 280 nm), senger areas in aircrafts, hotels and ships).
viruses are deactivated through critical damage to the
genomic system of the microorganism by absorbing
2. Principle of Thermal Inactivation
photons [2, 3]. With UV light, surfaces, e.g. door
knobs, are sterilized [4]. To clean air volumes, an The concept of the ‘Virus-Grill’ is that a partial
upper-room mounting of UV lights is possible quantity is extracted from a closed volume of air by
[5]. Additionally, UV lights are installed in HVAC the compressor and is specifically exposed to a higher
channels. A third method is the plasma air cleaning, temperature (e.g. more than 70°C) in the apparatus
which uses high-voltage electrodes to create ions in for at defined dwell time (e.g. 10 minutes in a dwell
the air [6]. chamber). Before the air reaches this chamber, it is
heated in three steps as shown in Figure 1. The heating
The developed apparatus ‘Virus-Grill’ reduces the steps one and two utilize the already treated and still
probability of droplet infection by inactivating viruses. hot air. While the first step is a passive heat recovery
This inactivation is achieved by heating the viruses system, the second step uses an active heat pump to
beyond their temperature sensitivity. (e.g. 70°C), the heat the air. The final heating step is the compressor
virus proteins, especially the surface proteins needed itself, where the air reaches the temperature needed
to dock to receptor molecules on the host cell mem- during compression. After that, the compressed air is
branes, are denaturized. stored in a heated chamber to inactivate the viruses.
Figure 1. Flow chart of the sterilization process.

TOP PAPERS
The viruses will stay in the treated air but because principle of a piston compressor. This causes pressure
they are not infectious anymore, the air is considered (change) and temperature to act for a short time.
inactivated / purified. In addition, the pressure-resistant dwell chambers
(here 3 bar) are tempered to a certain temperature
After releasing the sterilized air from the chamber, it by an external heater. In order to use energy more
goes through the heat recovery and the evaporator efficiently, the heat pump supports the heat recovery
of the heat pump to reduce its temperature. Before unit used. The developed system of solenoid valves
leaving the apparatus, a throttle reduces the pressure enables almost continuous operation through appro-
of the air. Therefore, the air leaves the ‘Virus-Grill’ priate control. The question of effectiveness was
apparatus at ambient pressure and almost at ambient addressed as a complete unit as well as in partial
temperature. component tests.
The ‘Virus-Grill’ control includes the collection and

3. Experimental Test Rig – The ‘Virus-Grill’
processing of numerous temperature and pressure
Apparatus
sensor data and adjustable process parameters. Valves
The set-up was done considering the planned tests and compressor are controlled by control electronics
with surrogate viruses in a safety workbench to char- containing market available and self-developed compo-
acterise the effectiveness of the ‘Virus-Grill’. Because nents. The ‘Virus-Grill’ is monitored by 16 temperature
of the limited space in the workbench, the test rig had sensors (PT1000) and 16 pressure transmitters, whose
to adhere to certain dimensions. Also, the required air data are processed by the controller. Switching opera-
volume of the aerosol generator and available electrical tions on solenoid valves and air compressors as well
connection were considered at the design process of as power control of the heat pump components are
the test rig. made possible by means of a Mosfet driver circuit and
an external power supply. The control of the chamber
Figure 2 shows the set-up of the ‘Virus-Grill’ during heating is realised via standardised interfaces.
the trials for the proof of efficacy.
A microcontroller board is used as the central element
The individual components play a special role in of the control system, which has sufficient computing
this process, the sum of which can achieve effective power and supports all necessary connections. The
inactivation. The compressor works according to the necessary firmware was programmed in C++ and
5
2
4b 4c 3
4a
1
Figure 2. Prototype of the ‘Virus-Grill’ in the sterile bench environment to characterise the efficacy in terms of inactivation
of aerosol-borne viruses. (1) aerosol generator, (2) virus chamber, (3) compressor, (4a-4c) holding chambers, (5) clean
chamber, (6) heat pump, (7) measuring computer with control software, (8) safety workbench (sterile bench).

TOP PAPERS
processes the measurement data, interprets process

parameters and gives corresponding control signals.
A wireless LAN access point including a webserver
enables the user to interact with the ‘Virus-Grill’.
A control software developed in C# establishes the nec-

essary connection and displays the measurement data
using a circuit diagram of the ‘Virus-Grill’ (Figure 3).
The measurement data are stored in parallel for later
evaluation. It is also possible to adjust the process
parameters of the ‘Virus-Grill’.
In order to take suitable samples, aerosol samples must

be collected from a closed system, such as the virus
and clean chamber, without manipulating the atmos-
phere in the system or allowing air contaminated with
viruses to escape. Therefore, a small airlock for swabs
was developed and four polycarbonate 3D printed
components were created for each airlock, which were
assembled with UV-resistant adhesive.
The construction is shown in Figure 4. The three-part

airlock consists of a handpiece, a piston insert and
the actual airlock chamber as well as the piston. The
handpiece serves as a carrier/holder for the swab and
at the same time as the “outer sluice gate”. The piston
insert combines inner and outer sluice gate including
the necessary piston seals. The piston itself determines
the volume of air exchanged when opening and closing Figure 4. Three-part airlock for sampling (above:
the airlock. Handpiece/swab holder; middle: piston insert with
sealing ring; bottom: printed components)
Figure 3. GUI of the control software.

TOP PAPERS
The greatest difficulty in the CAD-supported devel- of the surrogate virus, P. syringae was cultivated over-
opment was the sealing ability of the piston rings. night in tryptic soy broth (TSB) at 25°C and 170
The tubes used, made of special solvent-resistant and rpm. Bacterial culture was subcultivated in fresh TSB
UV-permeable acrylic glass, showed strong geometric for starter culture. At an optical density of 0.3, Phi6
tolerances, so that commercially available sealing solu- was added 1:10 followed by incubation to an optical
tions with rubber rings and lubricants and sealants did density of 0.08. To obtain pure phage lysate, the culture
not achieve sufficient sealing. Therefore, piston sealing was centrifuged to pellet the bacteria, and supernatant
rings made of flexible filament were developed and was filtered through a 0.45 µm filter (Fisher Scientific).
manufactured by means of 3D printing, with which
tolerances could be compensated and a sufficient seal For all experiments performed, Phi6 lysate was rebuff-
achieved at both sluice gates. These are shown in red ered in water. The titer of the obtained Phi6 stock
in Figure 4. Smooth sliding of the piston insert was lysate was determined using a bacterial plaque Assay
also ensured. [7] and stored at 4°C until further usage. For the
experimental procedures, the stock lysate was adjusted
To close the sluice gates securely, neodymium magnets to 1 × 1010 PFU/mL and used for aerosolization.
are used to magnetically attract a metal ring on the
handpiece itself or on the outer end of the piston, 4.2 Virological Data Acquisition
depending on the position of the handpiece. These The individual steps of sampling with this airlock are
metal rings required precise manufacturing in order to shown in Figure 5. After inserting the swab into the
minimise the gap between the magnets and the metal handpiece, it was mounted in the piston insert. At the
ring so that the neodymium magnets can generate a same time, the outer airlock is closed. For sampling,
sufficiently large magnetic attraction force. the entire piston insert is guided into the interior of the
system, exposing the swab to the aerosols. When the
piston insert is pulled out, the inner lock closes and the
4. Virological Test Setup
handpiece with the swab can be removed again. The
The current and still ongoing pandemic situation escape of aerosols can thus be reduced to a minimum.
regarding SARS-CoV-2 requires standardised methods
for the detection and analysis of airborne viruses in the
form of aerosol. Since SARS-CoV-2 is classified as a
biosafety level 3 virus, the number of laboratories that
can culture and analyse infectious viruses is severely
limited and the experiments require a high level of
security measures for containment. For this reason,
the bacteriophage Phi6 was established as a surrogate
virus. Phi6 shows a high degree of similarity to SARS-
CoV-2 in particle size, external structure, and the type
of genome. In contrast to other surrogate viruses, Phi6
is a bacteriophage that only infects bacteria and poses
no danger to humans, animals or plants. In order
to assess the effectiveness of disinfection measures,
two parameters were applied. i) The number of viral
genomes that accounts for all viruses whether dead
or alive, was determined by the molecular biological
method of qPCR. ii) The viral activity or infectivity
that accounts the number of infectious viruses was
determined by the plaque assay.
4.1 Sample Preparation of surrogate viruses

The bacteriophage Phi6 was used as surrogate virus
for the evaluation of the ‘Virus-Grill’. Phi6 was propa-
gated using its host bacterium Pseudomonas syringae.
Stock solutions of Phi6 and P. syringae were kindly
provided by Prof. Dr. Martin Hessling and Dr. Petra
Vatter (Technische Hochschule Ulm). For propagation Figure 5. Sampling concept with airlock and swab

TOP PAPERS
5. Measurements and Results

The detected virus genomes in Figure 7 show an
For the experimental set-up, a droplet aerosol with an increasing reduction of virus genomes when using the
average droplet size of 0.15 µm was generated from a pump in experiment 2, the flow through the ‘Virus-
phosphate-buffered saline (PBS) containing Phi6. The Grill’ in experiment 3 up to the greatest reduction
aerosol was collected and analysed at three positions of virus genomes when operating the ‘Virus-Grill’ at
before the experimental setup (virus chamber) and room temperature and a system pressure of 1.5 bar in
after the experimental setup (clean chamber) using experiment 4.
swabs moistened in PBS. All samples were analysed
for virus activity and total virus count. Taken together, it was possible to aerosolise surrogate
viruses in the experimental set-up, to transport them
Four different experiments were conducted to examine from the virus chamber into the clean chamber and
partial components of the ‘Virus-Grill’ and finally the to detect them. The use of the pump already leads to
entire test rig. In the first experiment (Exp. 1), the complete inactivation of the viruses used. This leads to
virus-containing aerosol was sucked from the virus the hypothesis that continuous inactivation is possible
chamber into the clean chamber. The pump was behind by using compressors based on the principle of the
the clean chamber. This was done to find a baseline piston compressor for SARS-CoV-2. Furthermore, the
of surrogate viruses in the virus chamber and clean use of the ‘Virus-Grill’ was able to reduce the total
chamber with no treatment whatsoever. In the second number of virus genomes.
experiment (Exp. 2), the virus-containing aerosol was
pumped from the virus chamber through the pump
6. Conclusions and outlook
into the clean chamber. In the third experiment (Exp.
3), the virus-containing aerosol was pumped pressure- The efficacy of the ‘Virus-Grill’ was validated with
free from the virus chamber through the ‘Virus-Grill’ four experiments. Experiment 1 defined a baseline
into the clean chamber. Again, the pump was located at of viruses reaching the clean chamber with no treat-
the end of / after the clean chamber. And in the fourth ment whatsoever. Experiments 2 to 4 tested different
experiment (Exp. 4), the virus-containing aerosol was ‘Virus-Grill’ configurations. Experiment 2 pulls the
pumped from the virus chamber through the ‘Virus- virus-containing aerosol only through the piston pump
Grill’, through the pump into the clean chamber at into the clean chamber. In experiments 3 and 4 the
room temperature and 1.5 bar overpressure. contaminated air flows through the complete ‘Virus-
Grill’ at ambient pressure and at an elevated pressure of
The results of the detection of the virus activity and the 1.5 bar, respectively. The number of active viruses were
determined virus genome count are shown in Figure 6 reduced below the limit of detection in all three cases.
and Figure 7. It was shown that only in experiment The number of viral genomes in the clean chamber
1, in which the viruses were sucked directly from one shows that there are inactivated viruses carried through
chamber to the other, were active viruses detectable in the ‘Virus-Grill’.
the clean chamber.
Figure 7. Detected viral genomes in virus chamber

Figure 6. Comparison of virus activity in virus chamber and clean chamber The viral genomes detect both
and clean chamber. inactivated and infectious viruses.

TOP PAPERS
This process has very high potential in applications

that dispense mechanical filtration. Here, maritime Data Statement
applications and integration into existing ventilation The datasets generated and analysed during the current
systems are mentioned in particular. The concept can study are not available because of the ongoing validation
also be adapted in the medical field for the ventilation and up-scaling process but the authors will make every
reasonable effort to publish them in the near future.
of patients. Thus, a wide range of flow rates from a
few l/min to several m³/min can be addressed through
suitable design and scaling of the components.
7. Acknowledgment 8. References
The presented work was part of the project AVATOR and Please find the full list of references in the original article
funded by the Fraunhofer-Society within the Fraunhofer at: https://proceedings.open.tudelft.nl/clima2022/article/
vs. Corona campaign which is gratefully acknowledged. view/214
REHVA is happy to welcome ZoonEx Systems

as a new REHVA Supporter
Even with the highest level of manual cleaning, surfaces The coating bonds to the surface and forms a micro-
are at continual risk of contamination and the transfer- scopic layer of pins that rupture bacteria and enveloped
ence of microbes to multiply and grow. viruses, giving building occupants a protected/treated
environment to work in.
Mechanical Ventilation Systems can provide ideal condi-
tions for microbial growth bringing unclean air into office More information: https://zoonexsystems.com/
spaces causing ‘Sick Building Syndrome’ (SBS), aiding the
spread of viruses like the common cold, influenza and
Coronavirus family among building users.
The ZoonEx™ System allows an antimicrobial treatment
to bond to a clean surface, giving effective protec-
tion against bacteria and enveloped viruses (including
Coronavirus family) for up to 30 days between treatments.
join us in Brussels!
REHVA
SESSION 1 SESSION 2
Achieving healthy
Brussels REPowER EU:
Zero Emission
Summit phasing-out fossil
Buildings with the
fuels in buildings
EPBD
NOVEMBER 15, FROM 10 TILL 12 NOVEMBER 15, FROM 13 TILL 15
POLICY CONFRENCE
MORE INFORMATION AND REGISTRATION
AT THE THON HOTEL BRISTOL STEPHANIE

TOP PAPERS
Health and energy assessment of

a demand controlled mechanical
extraction ventilation system
JANNEKE GHIJSELS KLAAS DE JONGE JELLE LAVERGE

Ghent University Ghent University & Ghent University
FWO, Flanders Research
Foundation (1SA7619N)
Abstract: Today, the assessment of residential demand-controlled ventilation systems only

considers the perceived indoor air quality in terms of comfort, with CO₂ and humidity as the
main parameters to investigate. However, the ventilation system and its controls also have an
impact on the health aspect of Indoor Air Quality (IAQ) due to the higher exposure to unhealthy
pollutants (Volatile Organic Compounds (VOCs), fine dust particles, e.g. PM2.5). In this paper,
two demand controlled mechanical extraction ventilation systems (DCV) and a continuous
mechanical extraction ventilation system (MEV) of a typical Belgian apartment are modelled using
Modelica. This allows to simulate the combined effect and interaction of temperature, airflow
and IAQ. The model includes sources of CO₂, humidity, VOCs and PM2.5 to the indoor air. The
combined approach using Modelica allows to do an in-depth analysis of the indoor air quality. A
two-stage assessment method is performed, resulting in an overall performance (in terms of IAQ
and energy use) of a DCV system in relation to the performance of the MEV reference system.
Keywords: Smart ventilation system, Indoor Air Quality, VOC, health and energy assessment.
1. Introduction
health effects. Therefore, there is a need for a designed
In the recent decades there has been an increasing ventilation system that brings fresh air in and evacuates
awareness that the energy demand for buildings must polluted air out of the home, preferably in a comfort-
be greatly reduced. Today, our buildings are better able way.
insulated and high-temperature heating is replaced
by low-temperature surface-heating. The stricter A continuous, constant, airflow ventilation system will
insulation standards ensure an energy reduction and guarantee a good IAQ but will also provide more cold
a better thermal comfort. The disadvantage is that, airflow that needs to be heated then strictly necessary to
if we only focus on the thermal comfort, an adverse ensure comfort. This results in an increase of the energy
effect will be induced on the indoor air quality (IAQ) use. The two main methods used in western-European
of our homes. In old houses there is natural ventilation residential ventilation systems to tackle this increase in
through cracks and crevices [1] but in more modern energy use are the use of an air-to-air heat exchanger
houses, that are build more airtight, the pollutants will (heat recovery) and the use of pollution sensors to
accumulate in the indoor air. This creates a greater risk measure the indoor air quality and lower the ventila-
of concentration problems, fatigue and other serious tion flowrate when and where it is possible without

TOP PAPERS
comprising the indoor air quality: demand-controlled are also extra CO₂-sensors in the dry spaces that will
ventilation systems (DCV). A DCV system can reduce increase the extraction flow rate in the wet spaces if
the heating energy related to ventilation and electricity the CO₂-concentration in the dry spaces becomes too
use of the ventilation system by 20 to 50% [2]. high. For the increase of the extraction flow rate only
the dry space with the maximum CO₂-concentration
Nowadays, the assessment of a DCV system only con- will be considered. The increase of extraction creates
siders the perceived IAQ in terms of comfort criteria negative pressure in the building which force more
(such as CO₂, humidity and odour) [3]. However, the fresh air through the trickle vents resulting in a larger
big disadvantage of a DCV system is the accumulation supply of fresh air in the dry spaces.
of indoor pollutants in times of low occupancy. When
the airflow rates are reduced, the VOC emissions of The second DCV system (DCV2) is also a theoretical
building materials and furniture will accumulate in the control system based on controls that can currently
indoor air, resulting in harmful VOC concentrations be found on the Belgian market. The ventilation flow
and a poor IAQ. Therefore, the assessment of a DCV rates (Q) are, just like DCV1, adapted by a local detec-
system must be extended from only comfort criteria tion and local control in the wet spaces. Supplementary
to both comfort and health criteria. to these extraction in the wet spaces, there is an
additional extraction in the dry spaces based on local
CO₂-sensors. Due to the direct extraction in the dry
2. Research methods
spaces, the amount of supply through the trickle vents
2.1 Simulation model can be guaranteed. In addition, the extraction works in
The simulation model is made in Dymola, an integrated two zones, namely the bedrooms and the living space.
environment for developing models in the Modelica The zone with the highest CO₂-concentration will be
language. This allows to simulate the combined effect controlled based on this concentration and the flow
of heat, moisture, airflow and indoor concentrations. rate of the other zone is lowered to the minimal flow
In this study, the IDEAS library [4] is used in combina- rate. In that way, the zone with the highest occupation,
tion with proprietary models for modelling the airflows receives the highest ventilation flow rate.
and pollutant sources. Figure 1 shows the floor plan of
the modelled three-bedroom apartment. This typical The working principles of both DCV systems is
Belgian apartment has already been used several times graphical represented in Figure 2. The nominal ven-
and has been described in Heijmans, Van Den Bossche, tilation flow rates according to the NBN-D50-001 are
Janssens (2007); Laverge, Janssens (2013) and De represented for each zone in Table 1. Qnom are the
Jonge, Janssens, Laverge (2018). During modelling, a nominal flow rates for both systems and Qnom_addition
lot of attention is paid to the multi-zone representation are the nominal flow rates for the additional extraction
of the apartment, the building envelope, the elements in the dry spaces for DCV2. The zones are numbered
of the various ventilation systems, the occupant sched- like the floor plan in Figure 1. Lastly, the different
ules, the ventilation controls, the emissions from the sensors and their controls on the ventilation flow rates
occupant activities and the emissions from the building are represented in Table 2. The first four controls are
materials and furniture. for both DCV systems. The fifth and sixth control are
respectively for DCV1 and DCV2.
2.2 Investigated DCV systems
The performance of two demand controlled
mechanical extraction ventilation systems (DCV) are
being compared to the performance of a continuous
mechanical extraction ventilation system (MEV). The
two DCV systems follow the same principles: fresh
air is naturally brought into the dry spaces through
trickle vents and will be mechanically extracted in
the wet spaces. The first DCV system (DCV1) is a
theoretical control system based on controls that can
currently be found on the Belgian market. The ventila-
tion flow rates (Q) are adapted on the one hand by a
local detection and a local control in the wet areas.
The bathroom is controlled on humidity, the kitchen Figure 1. Floor plan of the reference apartment. Zone
on CO₂ and the toilet on VOC. Additionally, there 1-4 are the dry spaces, zone 5-9 are the wet spaces.

TOP PAPERS
2.3 Pollutants of concern and their

emissions
More than 100 indoor pollutants are currently identi-
fied as (potentially) hazardous to our health. To obtain
a priority list of target pollutants, 7 large studies are
reviewed, each a conclusion of many other studies. The
most important study is the AIVC-CR17 [5] study
where, for Belgium, the concentrations of harmful
pollutants were measured in more than 400 homes.
As a result, 6 indoor pollutants and 3 outdoor pol-
lutants are prioritized for the Belgian residential
application, namely benzene, formaldehyde, naph-
thalene, limonene, toluene and particulate matter
(PM2.5) as indoor pollutants and (PM2.5), nitrogen
dioxide (NO₂) and ozone (O₃) as outdoor pollutants.
The concentration of the outdoor pollutants will be
modelled as constants. In the future, this can be further
investigated.
To allow a clear representation of which emissions are

implemented for each pollutant, the emissions will be
divided into three categories. namely emissions from
building materials and furniture, emissions from occu-
pants and emissions from occupant activities.
Figure 2. Graphical representation of the working

Table 2. Controls.
principle of DCV1 and DCV.2.
Sensor Control Q (m³/h)
Table 1. Ventilation flow rates. RV < 30% 10 %

1. RV – Bathroom and 30% < RV < 65% 30 %
service room 65% < RV < 95% 60%
Zone Qnom Qnom_addition, RV > 95% 100%
(m³/h) DCV2 (m³/h)
2. RV – Bathroom ΔRV > 2% in 5min 100%
1: Bedroom 43.92 25
CO₂ > 850 10%
3. CO₂ (ppm) Kitchen 850 < CO₂ < 950 Linear
2: Bedroom 35.26 25 CO₂ > 950 100%
3: Bedroom 38.88 25 No presence 10%

4. VOC – Toilet
Presence 100%
4: Living room 108.32 60 CO₂ < 1,000 10%

5. CO₂ (ppm) DCV1
1000 < CO₂ < 1,200 Linear
Max. of dry spaces
5: Hall 16 - CO₂ > 1,200 100%
Zone 1:
6: Kitchen 60 - CO₂ > 850 10%
zone with
850 < CO₂ < 950 Linear
maximum CO₂-
6. CO₂ (ppm) CO₂ > 950 100%
7: Bathroom 60 - concentration.
DCV2
Dry spaces
8: Toilet 30 - Zone 2:
extraction
zone with
- 10%
smaller CO₂-
9: service room 60 - concentration

TOP PAPERS
A. Emissions from building materials and B. Emissions from occupant activities

furniture
The impact of the occupant activities on the VOC con-
To determine the emissions of the building materials centrations is significant. To determine which activities
and furniture, it is assumed that the apartment is must be implemented, the original occupant schedules,
refurbished or newly built so that can be concluded used in Belgian simulation studies for the determi-
that the floor and furniture are new. Therefore, all nation of ventilation legislation [8], were reviewed.
the emissions will be determined after a lifetime of The original activities were cooking, showering and
28 days. The furniture is calculated for an occupancy washing clothes and only the emission of moisture was
of two adults, two children and two babies. The recorded. Eventually, the activities were expanded with
furniture is considered wood, synthetic or gypsum. cleaning, washing dishes and using deodorant spray.
This means that only the pollutants benzene, formal- The emissions of the activities, including the original
dehyde, naphthalene and toluene are considered for activities, were expanded with limonene, naphthalene
these emissions. All the emissions of building mate- and particulate matter emissions. Existing research
rials and furniture are determined using the Pandora on emission values is very limited, which means that
Database [6]. The summary of these emissions is given assumptions often must be made. When newer or more
in Table 3. All these emissions are assumed to have a accurate research is published, the emission values can
constant emission rate. One exception is made for the easily be adjusted in the model.
formaldehyde emission by the floor. The emission rate
of the flooring is a dynamic source model based on the For cooking, PM2.5-emissions were added based on the
air temperature and relative humidity in the zone [7]. relationship to the moisture emission in the study of
Poirier et al (2021) [9]. Extensive cooking results in a
Table 3. Summary of the emissions from building greater moisture and PM2.5-production. Important is
materials and furniture for each pollutant. that the cooking emissions are considerably reduced by
the implementation of a cooker hood with a flow rate
Emission formaldehyde benzene naphthalene toluene of 200 m³/h and a capture efficiency of 0.7. This means
[ug/h/m²]
that 70% of the emissions are captured by the cooker
Floor (wood) 9.91 negligible negligible negligible hood. For the activity of showering, the use of shampoo
Furniture and shower gel was added, resulting in limonene [10]
3.06 1.40 5.68 - and naphthalene [11] emissions. For the activity of
(wood)
washing clothes, the use of washing liquid (wash pods
Door (wood) 4.50 - - -
of 27 gram) was added, resulting in limonene emis-
Other furniture
3.00 2.00 - 11.00 sions [12]. For the use of deodorant, a PM2.5 [13] and
(synthetic) limonene [14] emission was added to the occupant
Carpet 4.27 0.21 0.47 0.20 itself. The occupant uses the deodorant 3 times a day
(0.5 gram) and carries these emissions around the house.
Walls (gypsum) negligible negligible negligible 0.50
Also, the emissions for the activity cleaning, were added
to the occupant itself. When the occupant is cleaning,
Table 4. Summary of the emissions from building moisture [15] and limonene [6] emissions are released
materials and furniture for each pollutant. into the air where the occupant is situated.
Emission moisture limonene naphthalene PM2.5
[ug/h/m²] All these emissions are summarized in Table 4. The
Cleaning 5.00 g/m² 1912 ug/h/m² references are given in the text and in the table. For a
- -
(3) (floor) [15] (floor) [6] more in-depth explanation on how the emissions are
0.60 L/s * 1260 ug/min [9] obtained, reference is made to “Health-based assess-
Cooking 1.00 L/s * - - 1910 ug/min
ment method for residential DCV systems” by Janneke
1.50 L/s * 2550 ug/min
Ghijsels (2022).
Washing
4.20 e-04 L/s 24.8 ug/h - -
dishes
C. Emissions from occupants
1200 ug/h
Showering 0.50 L/s *
[10]
3.76 ug/h [11] - Occupants produce both CO₂, H₂O and human odour.
The CO₂- and H₂O-emissions are shown in Table 5 for
Deodorant 1438 ug/use
use
-
[14]
- 12 ug/use [13] a metabolism (the degree of activity) equal to 1.6 met
(very active). In the simulation these productions are
Washing 6.50 x 7833 ug/h
clothes e-2 L/s * [12]
- - scaled according to the metabolism of each occupant
at each timestep.
* CEN 14788: Ventilation of buildings - Design and dimensioning of
residential ventilation systems

TOP PAPERS
2.4. Assessment method

After the control of the exposure concentrations, the
To make a complete analysis of the impact of a lifetime average daily dose (LADD) is calculated for
DCV system on IAQ, the assessment method will be each pollutant [20]. The formula of the LADD is
divided into two stages. The first stage is a health given in equation (1). Because the LADD considers
performance checklist that will rule out the possi- both body weight and inhalation rate it is possible to
bility that the exposure concentrations cause harmful obtain an estimation of the health effects for sensitive
health effects for the occupants. If this criterion is not occupants, for example babies.
met, the controls can be adapted (e.g. increasing the
nominal flow rates, increasing the minimal flow rates Babies will have a higher lifetime average daily dose
or adjusting the boundaries.) than an adult, even though the exposure concentra-
tions of the pollutants are the same.
When the quality of the indoor air is sufficient for
the health of the occupants, the DCV system can 𝐸𝐸𝑖𝑖 𝑥𝑥 𝐼𝐼𝐼𝐼 𝑥𝑥 𝐸𝐸𝑓𝑓 𝑥𝑥 𝐸𝐸𝑑𝑑
be analysed by the second stage of this assessment 𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿 = 𝑥𝑥 𝜖𝜖 (1)
𝐵𝐵𝐵𝐵 𝑥𝑥 𝐴𝐴𝑡𝑡
method, namely the overall performance rating in
terms of health and energy. A comparison will be made
with the performance of the MEV reference system. In
this way, a pareto optimum can be explored for each Where Ei is the timeweighted exposure (ug/m³), IR is
DCV system in which both the energy and health the inhalation rate (m³/day) [21], Ef is the exposure
performance are better than the performance of the frequency (day/year), Ed is the exposure duration (day),
MEV reference system. BW is the bodyweight (kg) where in this study 70 kg
is used for adults, 23 kg for children and 11 kg for
A. Health performance checklist babies. At is the simulation time, in this study 365 days
For the assessment of the health performance, a and ϵ is the absorption factor of each pollutant (for
checklist will be followed in which first the exposure example 0.9 for formaldehyde).
concentrations of each pollutant will be compared
with the limit concentrations of the chosen exposure The use of LADD makes it possible to exclude non-
metrics. Both acute and chronic exposure concentra- carcinogenic health effects by calculating the hazard
tions will be checked to exclude both acute and chronic quotient (HQ), given in equation (2). The LADD
health effects. is compared by the reference doses (RfD) (obtained
by U.S. EPA) [22]. When HQ is less than 1, the
The peak concentrations will be checked by the risk of non-carcinogenic health effects is considered
10-minute AEGL-1 value (obtained by U.S. EPA) [18] negligible.
and the average exposure concentrations over a time
interval of 1 hour and 8 hours will be checked by the
acute REL values (obtained by OEHHA) [19]. The 𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿
chronic exposure concentrations will be checked by 𝐻𝐻𝐻𝐻 = < 1 Health effects are negligible (2)
𝑅𝑅𝑓𝑓 𝐷𝐷
the chronic REL value. The summary of these limit
concentrations is given in Table 6.
Table 5. Emissions by occupants themselves. Table 6. Emissions by occupants themselves.
Production 1.6 met (light activity) Reference 10 min Acute Acute Chronic
Cexposure (ug/m³) AEGL-1 REL-1h REL-8h REL
CO₂-production
Benzene 415,000 27 3 3
- Adult 19.0 l/h *
- Child 12.6 l/h [16] Formaldehyde 1105 55 9 9
- Baby 6.7 l/h * Naphthalene - - 9 9
Limonene - - - 9000
H₂O-production
- Adult 55.0 g/h * Toluene 252,000 - 800 400
- Child 41.3 g/h [17] PM2.5 - - 25 (24h) * 10*
- Baby 18.3 g/h *
* WHO guidelines: air quality guidelines for particulate matter,
ozone, nitrogen dioxide and sulphur dioxide: summary of risk
* Norm CEN 14788 assessment

TOP PAPERS
B. Overall performance rating

of the WHO and is only exceeded by DCV2. The
If the DCV system passes the health performance check- chronic exposure concentration is exceeded by all
list, it is evaluated by the second part of this assessment the ventilation systems. This is a result of a constant
method. In this assessment method, the DALY-index outside PM2.5-concentration of 14 ug/m³ (according
(Disabled Adjusted Life Years) is used as health indicator. to MIRA 2019) [26]. In the future, it can be important
It quantifies the total years lost due to death or disability to change the approach of the outside pollutants to
due to poor IAQ. It scales the harmfulness of the different more variable concentrations according to the environ-
VOC and PM2.5-concentrations to allow a general health ment (e.g. temperature) and the location (e.g. nearby
rating. The total DALYs are calculated based on the study industry, heavy traffic).
of Logue et al. (2012) [23]. The DALYs of the outdoor
pollutants (i.e PM2.5, NO₂ and O₃) are calculated using
the IND-method. The DALYs of the indoor pollutants
(i.e. benzene, formaldehyde, naphthalene, limonene
and toluene) are calculated using the ID-method where
the study of Huijbrechts et al. (2005) [24] provides the
information on the (∂D/∂I)-factors.
When the total DALYs are calculated for each DCV

system, this health indicator can be compared with the
energy use of each DCV system. In that way it is possible
to rate the overall performance of the DCV system.
The performance of a DCV system is considered suf-
ficient when there is a pareto optimum compared to
the continuous MEV reference system. This means that
both the energy use and the health impact of the DCV
system must be lower than those of the reference system.
3. Results – Assessment method

Each DCV system and their controls are modelled Figure 2. Comparison of the exposure concentration of
in the Modelica model together with the different formaldehyde between DCV1, DCV2 and MEV_ref.
emissions and the activity schedules of the occupants.
For each ventilation system 10 different families are
simulated. The results are average exposure concentra-
tion of these 10 scenarios.
3.1 Health performance checklist

In this paper, the focus will be on one constant emission
source (i.e. formaldehyde) and on one emission source
that depends on the activities of the occupants (i.e.
PM2.5), because these two pollutants show to have
the highest impact on our health. In Figure 2 the
exposure concentration of formaldehyde is compared
for the two DCV systems and the MEV reference
system. The acute concentration limit of 55 ug/m³ is
not exceeded by any system. The chronic concentra-
tion limit of 9 ug/m³ is exceeded by the two DCV
systems. DCV1 has a chronic exposure concentration
of 11 ug/m³, while DCV2 has a much larger chronic
exposure concentration of 18.8 ug/m³.
Figure 3 shows the exposure concentration of PM2.5

for the three systems. The acute exposure concentra- Figure 3. Comparison of the exposure concentration of
tion is compared with the 24 hours limit concentration PM2.5 between DCV1, DCV2 and MEV_ref.

TOP PAPERS
In Table 7 all the concentrations of the pollutants flow rates. This adaptation changes the LADD of
of concern are summarized. DCV2 scores too high benzene for occupant 5 (Baby 1) from 0.01062 to
for formaldehyde and PM2.5. DCV2 has also higher 0.00784 mg/ kg/day and for occupant 6 (Baby 2) from
exposure concentrations for all the other pollutants 0.00915 to 0.00676 mg/ kg/day. All the average daily
than DCV1. doses are now below the reference doses. Hence, it can
be said that no important negative health effects will
To ensure that the IAQ, caused by system DCV2 does occur due to a poor IAQ. In the next paragraph, it is
not cause any health effects on the sensitive occupants, examined whether there is a pareto optimum between
the LADD is calculated. Subsequently the HQ of each the two DCV systems, incl. the new DCV2 system and
pollutant is calculated by comparing the LADD to the the MEV-reference system. The overall performance of
reference dose (obtained by U.S. EPA). An example is the DCV systems, both in terms of energy and health,
worked out for benzene. In Table 8 the average LADD should perform better than the overall performance of
of benzene of the 10 scenarios simulations is calcu- the MEV reference system.
lated for each occupant in the simulation. For DCV2,
the LADD of benzene is higher than the reference
dose (8.57 × 10−3 mg/ kg/day) for both the smaller Table 8. Summary of all LADD of benzene for each occupant
children. Therefore, DCV2 will not ensure a good IAQ in the simulation for MEV_ref and the two DCV systems.
for the sensitive occupants.
It is necessary to adjust DCV2 to meet the minimum

requirements of the health performance checklist. A
new simulation is carried out where the minimum flow
rates are increased from 10% to 30% of the nominal
Table 7. Summary of all the acute and chronic exposure concentrations of the pollutants of concern for DCV1,
DCV2 and MEV_ref.

TOP PAPERS
3.2 Overall performance rating

The total DALYs per 100,000 persons per year are
calculated for each pollutant using the IND and ID
method. The results are given in Figure 4. It becomes
clear that the total DALYs are for more than 80%
caused by PM2.5. The second major pollutant is formal-
dehyde. All the other VOCs seem to have a very small
impact and are therefore less harmful for our health.
DCV1 has the smallest total number of DALYs. Even
smaller than the reference system. This is caused by the
smaller influence of the outdoor pollutants in periods
of less ventilation. DCV2 has the largest total number
of DALYs. This is caused by the higher formaldehyde
concentrations and the higher PM2.5-concentration in
the kitchen during cooking periods.
The new DCV2 system, where the minimum flow are

adapted to 30% of the nominal flow rates (instead of
10%) reduces the total number of DALYs from 41.1 to
32.8 DALYs per 100,000 persons per year. This is a
reduction of more than 20% (8.3 DALY). Figure 4. The share of each pollutant in the total
number of DALYs (the health indicator in this research)
Now that the total DALYs are known, it is possible to per 100,000 persons per year compared for the two
generate an overall performance rating of the DCV original DCV systems and the MEV reference system.
systems. In Figure 5 the total number of DALYs are
compared with the energy use (electricity use of the
fans and ventilation heat losses) of the ventilation
system. To create a pareto optimum it is necessary that
the DCV systems perform better in terms of health
and energy. This means that the DCV systems must
be located in the green frame shown in Figure 5. The
system that is situated on the bottom, left, is the system
that generate the best pareto optimum. At first, system
DCV2 had a very low energy use. After the adapta-
tion, the IAQ is improved with 20% (8.3 DALY) and
the energy use increased with 35% (1,040 kWh/year).
This increase in energy use seems very high, but the
total energy use is still 40% (1,845kWh/year) lower
than the energy use of the reference system with con-
tinuous flow rates. The overall performance of the new
DCV2 is even better than the overall performance
of DCV1. It is stated that with correct adaptions, it
is possible to find a pareto optimum for each DCV
system.
4. Discussion and conclusion

In this research, a very extensive emission model was
combined with a dynamic temperature and occu- Figure 5. The overall performance of a DCV system
pancy model in the Dymola software. This allows an in terms of energy and health compared to the MEV
assessment of IAQ at every timestep for the different reference system. The overall performance of a DCV
occupants with different ages, habits and metabolisms. system is sufficient if there is a pareto optimum
Based on the determination of the exposure concentra- compared to the MEV reference system. DCV2_new
tion at each time step and for each occupant, both scores best on the overall performance.

TOP PAPERS
the average daily dose (LADD) and the intake can be for each system are compared with the energy use of
calculated. These two parameters are both very impor- each system, resulting in an overall performance of
tant in the assessment method that was developed to the DCV system in comparison to the MEV reference
determine whether a system ensures a good indoor air system.
quality. The necessity for such an assessment method
is high since the DCV systems are nowadays only The conclusion of the two DCV systems is that, if
assessed on comfort criteria such as CO₂, humidity there is minimum air flow rate of 10%, DCV1 ensures
and odour. Because the danger lies in the increased a better indoor air quality. This means that increasing
VOC concentrations in times of less ventilation, it the extraction flow rate by using CO₂-sensors in the
is necessary that a DCV system, designed to ensure dry spaces, works sufficient. The disadvantage is that
an energy saving, also ensures a sufficiently IAQ in DCV1 only ensures an energy reduction of around
terms of health. 25% (1,200 kWh/year) in comparison to the con-
tinuous MEV system, which is rather low. When the
The health and energy assessment method was designed minimal flow rates of DCV2 are increased to 30%, it
as a two-stage assessment method in which first the guarantees a better IAQ and a higher energy reduc-
exposure concentrations are checked on health risks for tion then DCV1. That’s why the adaption of a DCV
the occupants. The peak concentrations, acute concen- system that initially guarantees a high energy reduc-
trations and chronic concentrations are compared with tion but a low IAQ, ensures mostly an improvement
the corresponding limit concentration from relevant to the health/energy contradiction. In comparison to
exposure metrics. In this paper, the limit concentra- the continuous MEV system, the energy reduction
tions of the reference exposure levels [19] were used. It of DCV2_new is around 40% (1,845 kWh/year) and
is possible to use other limit concentrations of exposure the improvement of IAQ is around 15% (5.5 DALY).
metrics that are for example drawn up in function of a
legislation in the country where the research is being The investigated cases show that DCV systems can
conducted. be an effective measure to save energy and provide a
healthier indoor air. Both systems, DCV1 and DCV2_
The second stage of the assessment method is an new, guarantee a pareto optimum in comparison to the
overall performance rating where one health indicator, continuous reference system. It is stated that for every
the total DALYs, is used as general health indicator of DCV system, an optimalisation can be found where
the system. The total DALYs scales the harmfulness of the health/energy contradiction disappears.
exposure to the different pollutants. In that way, the
health performance of a system that ventilates more in References
the kitchen and induces lower PM2.5-concentrations, Please find the full list of references in the original article
can be compared to the health performance of a system at: https://proceedings.open.tudelft.nl/clima2022/article/
that ventilates more in the living rooms and induces view/155
lower formaldehyde concentrations. The total DALYs
Energy Efficient Renovation of Existing

Buildings for HVAC professionals
Energy Efficient Renovation of Existing Buildings for HVAC professionals.

This REHVA Guidebook shows the baseline for specific energy efficiency and other
renovation measures in existing buildings for which the HVAC systems play an
important role. It presents the best available techniques and solutions that can be
used as part of the energy modernization of the HVAC systems.
REHVA 40 RUE WASHINGTON 1050 BRUSSELS, BELGIUM

+32-2-5141171 [email protected] WWW.REHVA.EU

TOP PAPERS
Low-temperature radiant cooling panel

for hot and humid climate
GONGSHENG HUANG NAN ZHANG YUYING LIANG
Department of Department of Department of
Architecture and Civil Architecture and Architecture and
Engineering, City Civil Engineering, Civil Engineering,
University of Hong Kong, City University City University
Kowloon, Hong Kong of Hong Kong, of Hong Kong,
[email protected] Kowloon, Kowloon,
Hong Kong Hong Kong
Abstract: In this paper, the heat transfer and thermal environment of air-layer integrated
radiant cooling panel (AiCRCP) was studied experimentally. AiCRCP was proposed in 1963,
which was characterized by the use of an infrared-radiation transparent (IRT) membrane to
separate the panel’s radiant cooling surface from its external air-contact surface. Therefore,
the panel’s radiant cooling surface temperature can be reduced to increase the cooling
capacity, while its external air-contact surface, due to the thermal resistance provided by the
air layer and the IRT membrane, can be easily maintained at a high temperature to reduce
condensation risks. The thermal performance of AiCRCP was investigated using a prototype.
Several scenarios were tested to analyse the thermal performance of prototype, and the
cooling capacity of the AiCRCP was also investigated according to the thermal performance
of the prototype. The results demonstrated that this new type of radiant cooling systems
could be more preferable to be implemented in hot and humid climates.
Keywords: Radiant cooling, air-layer integrated ceiling radiant cooling panel, IRT
membrane, condensation, cooling capacity
1. Introduction
other air-cooling systems in practical applications to
Radiant cooling has many benefits, such as its high provide enough cooling for thermal comfort.
thermal comfort, lower energy use, quiet operation,
and smaller equipment footprint, compared with the Many attempts have therefore been made to solve the
methods of cooling air of indoor spaces through air condensation problem while increasing the cooling
circulation [1]. However, radiant cooling applications capacity [4]. A typical method is to maintain a low
face great challenges in hot and humid climates from relative humidity so low radiant temperature can be
condensation and their limited cooling capacity [2]. used [5], but this dehumidification is costly and risks
In its cooling mechanisms, whether through panel degrading the thermal comfort. By realizing that the
cooling or slab cooling, the radiant surface is also the radiant surface must be isolated from the air-contact
air-contact surface, as shown in Figure 1(a), where a surface to essentially solving this dilemma, Morse
conventional ceiling radiant-cooling panel (CRCP) is proposed a radiant-cooling panel covered with a sheet
used. In this CRCP, the cooling capacity increases with spaced several centimetres from it, filled with dry air,
the decrease of the radiant-cooling surface tempera- and sealed to prevent room air from contacting the cold
ture, but this will increase the risk of condensation. radiant-cooling surface [6], referred to as an air-layer-
This dilemma has inhibited the commercialization of integrated radiant-cooing panel (AiCRCP) and illustrated
radiant cooling in hot and humid climates [3], and in Figure 1(b). Thus, the air-contact surface and radiant
explains why CRCPs should be used together with cooling surface in the AiCRCP are physically separated.

TOP PAPERS
It should be noted that the cover sheet must be maintain give them potential applications in hot and
transparent to the energy radiated from a body at a humid climates. However, most of the work reviewed
temperature range of 29.4°C to 35°C, and thus titled above was based on simulation or numerical studies.
as infrared transparent (IRT) membrane. It enables In this paper, the cooling performance of an AiCRCP
the panel to act as a radiation heat sink, and heat prototype and the thermal environment created by the
from occupants in the vicinity can radiate through the AiCRCP were investigated using experiments, which
transparent cover to the cold plate behind. Teitelbaum showed that the AiCRCP could provide higher cooling
et al. [7] revisited this design, investigating several capacity and better condensation prevention.
manufactured materials, such as low/high density poly-
ethylene and polypropylene, using Fourier transform
2. Theoretical background
infrared (FTIR) spectroscopy to analyse its thermal
performance. They also investigated the panel depth The heat transfer process of the AiCRCP is shown
(spacing between the radiant-cooling panel and the in Figure 1(b), where the heat exchange between
membrane) to balance radiation, conduction, and the AiCRCP and its thermal environment is mainly
convection when AiRCPs was applied to an outdoor through two mechanisms. One is radiation that occurs
environment. directly between the AiCRCP and its thermal environ-
ment; and other is the combination of convection and
In our research group, Zhang et al. established a conduction.
two-flux heat transfer model for the AiCRCP, to analyse
the optical, physical and thermal properties of the In the first mechanism, since the IRT membrane is
IR-transparent membrane [8]. Liang et al. investigated assumed to have a poor ability to absorb infrared radia-
the thermal environment and thermal comfort created tive heat flux and the air-layer is transparent to infrared
by an AiCRCP using CFD simulations, and they dem- radiation, the radiative cooling power of the radiant
onstrated that general thermal comfort indices could cooling surface will not be much affected by the IRT
be satisfied even when the AiCRCP operated at a very membrane. In the second mechanism, heat is firstly
low radiant temperature (e.g. −2.3°C) [9]. transferred to the IRT membrane from the air sur-
rounding the AiCRCP through convection, and to the
Previous literature has demonstrated that the enhanced radiant cooling surface through the dry air convection
cooling capacity and reduced condensation risk of (major) and conduction (minor). It should be noted
AiCRCPs in addition to the thermal comfort they can that the IRT membrane is thin, its internal surface
and external surface temperature could be consider as
the same in the analysis of the thermal performance
of AiCRCP.
Because the air layer has a large thermal resistance

to both convection and conduction as shown in the
previous work [8], the IRT membrane can be main-
tained at a high temperature even if the radiant cooling
temperature is low, for example 5°C (much lower than
17°C used in conventional CRCPs). Therefore, a low
radiant cooling temperature can be used to enhance
the cooling capacity of the AiCRCP without increasing
the condensation risk.
3. Experiment setup
3.1 Prototype of AiCRCP
Currently there is one prototype of AiCRCP in our
laboratory with the dimension of 1 m × 1 m. This
prototype uses a piece of aluminium plate: on one
side high emissivity paint (ε ≈ 0.95) was coated and
used as cooling radiant surface, and on another side
Figure 1. Diagrams of (a) the conventional radiant heat transfer fluid (HTF) pipes were attached and fixed
cooling unit, (b) the air-integrated CRCP. using screws.

TOP PAPERS
Thermal paste was used to increase the thermal con-

ductivity between the HTF pipe and the aluminium
plate, shown in the photo of Figure 2(a). This side
was covered with a insulation layer (Nitrile rubber) to
prevent cooling loss. A photo of the prototype is given
in Figure 2(b). The prototype was connected with an
air-cooled chiller that can provide heat transfer fluid
with the temperature from −10°C to 20°C with the
rated cooling capacity of 5.67 kW and coefficient of
performance (COP) of 3.86. A 20 µm thick low-den-
sity Polyethylene (LDPE) was used as the IR-transparent
membrane to seal a dry-air layer to separate the air-
contact and radiant-cooling surfaces. The membrane
has a good IR transparency property as shown in
Figure 2(c), where the spectral transmittance was
measured using FTIR (Spectrum Two, PerkinElmer).
3.2 Thermal chamber and measurements

The thermal chamber is placed on a movable structure
as shown in the photo of Figure 3(a). Figure 3(b)
shows the internal view of the chamber.
A number of sensors/flow meters were installed to

measure the temperature, flow and humidity. The
measured qualities, number of sensors and sensor
uncertainty were summarized in Table 1.
Figure 2. (a) The radiant cooling panel covered by an insulation layer; (b) LDPE membrane covering the panel and
sealing a dry air layer; (c) IR spectral transmittance through LDPE membrane using FTIR.
Table 1. Measured quantities, number and uncertainty of sensors.

Measured Quantity Sensor Number Uncertainty
HTF Fluid temperature T-type Thermocouple 2 ±0.5°C
Surface temperature T-type Thermocouple 18 ±0.5°C
Dry bulb air temperature Swema 03+ 1 ±0.1°C
Indoor air velocity Swema 03+ 1 ±0.04m/s
Relative humidity Swema HC2A-S 1 ±0.8%RH
Globe temperature Swema 05 1 ±0.1°C
HTF flow rate Turbine pulse flowmeter 2 ±0.5%
Infrared temperature FLIR infrared camera 1 ±4%

TOP PAPERS
4. Experimental results and analysis

the thermal chamber, the cooling panel was installed hori-
4.1 Testing scenarios
zontally on the ceiling, facing downward and working as
During the experiment, the thermal chamber was moved ceiling cooling. In the test, the controlled variables were
into a conditioned large hall, where the space temperature the indoor air temperature inside the thermal chamber and
and humidity was maintained to be relatively stable. Inside the radiant cooling temperature of the AiCRCP.
Supporting
structure for
AiCRCPs
Sensors (wires)
Computer
HTF pipes
Data logger
AiCRCP (not
yet installed on
the ceiling)
Figure 3. (a) an external view of the thermal chamber and (b) an internal view of the chamber.

TOP PAPERS
5. Cooling capacity of the AiCRCP

There were three testing scenarios, defined according
to the total power of the heaters. In scenario 1, the total The cooling capacity of an AiCRCP depends on
power of the heater was the highest; while it was the the thermal environment that is conditioned by the
lowest in scenario 2. In each scenario, the HTF supply AiCRCP. Here we considered a room with the dimen-
temperature was varied in the range of −5~15°C, and sions of 4 m × 4m × 3m. We assumed that the whole
the cooling panel radiant surface temperature had a ceiling was used as AiCRCP. The walls and floors had
gradient change between 1°C and 19°C. The RH of the emissivity of 0.9. Referring to the Figure 2(d),
the indoor air temperature was controlled at 3 levels, the IRT transmittance of the membrane was assumed
i.e. 70%, 60% and 50%. When any variable was to be 80%. The temperature of the room air and the
changed, the temperatures were measured 20 min after wall/floor surfaces were assumed to be 26°C.
stability. No mechanical ventilation was used during
the experiment. The cooling capacity was the sum of radiative and
convective heat fluxes. The cooling capacity of the
4.2 Temperature of the IRT membrane AiCRCP was compared with that of the conventional
The temperature of the IRT membrane surface (toward CRCP in Figure 5. When the RH was 70%, 60%,
indoor air side) is an important variable that indi- and 50%, the dew point temperatures of the room
cates the capability of the AiCRCP for condensation air were 20.12°C, 17.66°C and 14.81°C, respectively.
prevention. Please note that it is required at least Therefore, the cooling panel temperature of the con-
1°C higher than the dew point of indoor air to avoid ventional CRCP were set as 21°C, 19°C and 16°C
condensation risk [1]. The average temperature of the (approximated 1°C above the dew point). To maintain
IRT membrane under different radiant cooling panel the IRT membrane temperature at 21°C, 19°C and
temperature and different indoor air temperature and 16°C, the minimum allowable cooling panel tem-
humidity were measured. peratures of the AiCRCP were 13°C, 8°C and 2°C
respectively.
When RH was maintained at 70%, the cooling
panel temperature should be higher than 13.5°C, Figure 5(a) showed the comparison when RH was
8.5°C and 3°C when the indoor air temperature was 70%. The cooling capacity of the conventional CRCP
28.9°C, 25°C and 19.4°C, respectively. However, when was 40 W/m² using a radiant temperature of 21°C.
RH=60%, the cooling panel temperature should be 7.5% cooling capacity enhancement was achieved by
higher than 6°C when the environmental temperature the AiCRCP when its radiant cooling temperature was
was 29°C. When the indoor air was lower than 25°C, 17°C. Since the minimum allowable cooling panel
the condensation would not occur on the membrane temperature of the AiCRCP could be down to 13°C,
even if the cooling surface temperature was reduced the maximum cooling capacity of AiCRCP reached
to below 2°C. When RH=50%, the results showed 63.63 W/m², improved by 59.1%.
that no condensation occurred on the membrane in
these three different indoor air temperatures even the Figure 5(b) showed the comparison when RH was
cooling panel temperature was reduced to 1°C. 60%. At this condition, the cooling capacity of the
To further analyse the capability of the condensation

prevention of the AiCRCP, the minimum allowable
cooling panel temperature allowed was defined, which
was the minimal temperature of the panel that can
guarantee the IRT membrane temperature being 1°C
higher than the dew point of the indoor air. The differ-
ences between the minimum allowable cooling panel
temperature of the AiCRCP and the conventional
ceiling radiant cooling panel were shown in Figure 4
when RH=70% and RH=60%. When RH=70% the
minimum allowable temperature of the AiCRCP was
around 10.8°C lower than the conventional CRCP;
and when RH=60%, the value was about 15°C.
Therefore, the AiCRCP has a good performance in
preventing condensation in hot and humid climates Figure 4. The minimum allowable cooling panel
with lower radiant temperature. temperature of the AiCRCP and the conventional CRCP.

TOP PAPERS
conventional CRCP was 59 W/m² using a radiant of 16°C. When the radiant cooling temperature of the
temperature of 19°C. Similar cooling capacity was AiCRCP was 8°C, 2.2% cooling capacity improve-
achieved by the AiCRCP when its radiant cooling ment was achieved. Similarly, when we considered the
temperature was 14°C. Since the minimum allow- minimum allowable cooling panel temperature of the
able cooling panel temperature of the AiCRCP could AiCRCP that could be down to 2°C, the maximum
be down to 8°C, the maximum cooling capacity of cooling capacity of AiCRCP reached 121 W/m²,
AiCRCP was improved by 52.2%, reaching 90 W/m². improved by 35.9%.
The comparison when RH was 50% was shown in Figure 5 also showed that due to a higher air contact
Figure 5(c), where the cooling capacity of the conven- surface temperature and a lower radiant surface tem-
tional CRCP was 89 W/m² using a radiant temperature perature, the radiative heat flux of the AiCRCP was
increased, while the convective flux was decreased
when compared to the conventional CRCP. The
enhanced radiative heat flux will benefit the heat
exchange directly between heating sources (such as
occupants) and the cooling panel.
6. Concluding remarks
This paper investigated the thermal performance
of a prototype of AiCRCP experimentally and then
analysed its cooling capacity of AiCRCP in a simple
room environment based on the thermal performance
of the AiCRCP prototype. The results have shown that
• The AiCRCP has much enhanced capacity to prevent

condensation even in hot and humid climates. Due
to a large thermal resistance from the sealed air layer,
the IRT membrane can be maintained at a high
temperature even when the radiant temperature is
controlled to a very low temperature.
• Due the possibility of using a low temperature, the
cooling capacity of the AiCRCP can be enhanced
significantly. At a higher humid environment, for
example RH = 70%, the cooling capacity can be
improved around 60%.
Thus, the thermal performance of the AiCRCP could

make it more preferable when the technique of ceiling
radiant cooling is adopted in hot and humid climates.
7. Acknowledgement
The research work presented in this paper was
supported by a grant from the Research Grants Council
of the Hong Kong Special Administrative Region, China
(Project No. 11212919).
8. References
Figure 5. The cooling capacity of the AiCRCP and view/395
conventional CRCP.

TOP PAPERS
Impact of future climate on the

performances of ground-source
cooling system
ABANTIKA SENGUPTA PIETER PROOT TOM TRIOEN HILDE BREESCH MARIJKE STEEMAN
PhD student, Faculty of Master Thesis Master Thesis Associate Professor, Associate Professor,
Engineering Technology, student, Faculty student, Faculty Faculty of Engineering Department of
KU Leuven, Belgium of Engineering of Engineering Technology, KU Architecture and
[email protected] Technology, KU Technology, KU Leuven, Belgium Urban Planning, Ghent
Leuven, Belgium Leuven, Belgium University, Belgium
Abstract: Newly constructed and renovated dwellings in Belgium are designed for the
current climate context. However, due to the effects of global warming, extreme weather
conditions like warmer summers and frequent heatwaves are expected in future climate
scenarios. Future climate scenarios are nowadays mostly not taken into consideration
during the building design process. This paper studies a case study dwelling equipped
with a ground-water heat pump coupled with a heat exchanger to provide passive floor
cooling, derived from two vertical boreholes with a depth of 100 meters. The aim of this
study the impact of future climate scenario on the thermal comfort and the performance
of radiant floor cooling system in a Belgian dwelling. Monitoring of the case study building
(April-October 2020) and Building Energy Simulations (BES) in Open studio and EnergyPlus
were conducted. Future weather files (future mid-term-2050s and future long-term-
2090s according to the RCP8.5-scenario) were developed in the framework of IEA EBC
Annex 80 Resilient Cooling of Buildings. The performance of the floor cooling system was
analysed for four different climate scenarios for Melle, Belgium: typical historical-2010s,
2020 including a heatwave (observational data obtained from RMI), typical future mid-term-
2050s and long-term 2090s. The evaluation was based on two parameters, [1] thermal
comfort and [2] cooling capacity of the ground heat exchanger. Results demonstrate that
in the future (long-term) the current design of the building including the floor cooling system
is inadequate to provide a good thermal comfort. Due to the rising indoor and ground
temperatures, the maximum cooling capacity will decrease 22,5% in future long-term
scenario compared to the typical historical weather scenario. Results also confirm that the
occupancy has a big impact on the thermal comfort, especially in the sleeping rooms. This
study also indicates the importance of implementing shading as a good solution to obtain a
better thermal comfort in future climate scenarios.
Keywords: Climate Change, Future climate scenarios, Passive cooling strategies, Cooling
capacity, Thermal Comfort.

TOP PAPERS
1. Introduction
cooling energy need in future climate scenario, even
IPCC’s Special Report on Global Warming of 1.5°C in moderate climates. The effect of climate change
concludes there is a growing risk of overheating in (increase in outdoor air temperature, solar radiation
buildings and an increase in severity and frequency of and ground temperature) is assessed to evaluate the
heatwaves in future climate scenarios [1]. As seen in performance of the floor cooling to ensure robust
Figure 1, if the current trend of global greenhouse gas thermal summer comfort in a dwelling in Flanders
emissions continues, that is Business as Usual (BAU), (Belgium). In the following section, the case study
RCP8.5, by 2100 an increase of the outdoor tempera- building, model validation and methodology are
ture of approximately 5°C in Belgium is expected [2]. described in detail, followed by a discussion of the
An increase of the outdoor temperature will have an main results and conclusions.
enormous impact on the environment, soil and indoor
climate. Given the uncertainty in future climate, miti-
2. Materials and Methods
gating the adverse effects of climate change is a high
priority for the EU. To reduce the sensitivity of highly 2.1. Case Study Building
insulated dwellings to overheating, cooling systems The case study building examined is a terraced dwelling,
(active or passive) are needed and will have an impor- located in Geel, Belgium and constructed in 2014.
tant role in the future [3]. This building is also a part of the SCOOLS-project
In recent years, policies in Europe and worldwide

focused on the energy efficiency of both new and reno-
vated buildings. To avoid excessive increase of energy
use for cooling, there is a need to examine alternative
cooling concepts and passive cooling strategies in order
to achieve the goals of the EU’s Climate Change miti-
gation policies [4]. In new dwellings in Flanders, built
between 2006 and 2018 [5], between 15-20% of new
dwellings are equipped with a heat pump, out of which
30% are ground-water heat pump. Even though cur-
rently they are mainly used for heating, these numbers
suggest a high potential for ground-source cooling.
Ground-source cooling systems are gaining significant
market share amongst low energy cooling technologies
[6] [7].
Figure 1. Evolution of the average annual temperature
The working principle of ground-source cooling is in Belgium. [2]
based on the fact that the ground temperature below
approximately 10 m remains fairly constant all year
round at about mean annual ambient air tempera-
ture [9]. It rejects heat to the ground by circulating
a working fluid through ground heat exchangers.
Ground-source cooling can be classified as direct
ground cooling (passive) or ground source heat pump
(active). As seen in Figure 1 and Figure 2, in future
climate scenarios, the outside air temperature and
ground temperature will be higher, this will affect the
cooling capacity of ground-source cooling [8]. Rising
ground temperatures lead to the decrease in perfor-
mance of a ground-water heat pump, as shown by
a decrease in COP value during the summer. In the
winter the COP will increase [10,11].
The objective of this study is to assess the perfor-

mance of radiant floor cooling in future climate Figure 2. Evolution of ground temperature in future
scenarios. This study aims to evaluate the increase in climate scenario. [8]

TOP PAPERS
(2018-2021) which aims to evaluate the performance

of low-energy cooling systems [12]. The building is
South-West oriented and designed for a family con-
sisting of 4 people. Figure 3 shows the (southwest)
facade of the building.
The building, apart from the parking in the ground

floor, consists of two floors (3 m high each), with a
total volume of 825 m³, an external surface of 440 m²
and a compactness ratio (Surface area/gross heated
volume) of 0.53. The dwelling is well-insulated and
has a heavy thermal mass (calculated based on NBN
EN ISO 52016-1) [13]. The average U-value of the
construction is 0.42 W/m²K. The U-values of the con-
struction elements are given in Table 1. The windows Figure 3. South-West facade of the case study building.
and the doors have a g-value of 0.55. The U-value of
the fixed window and the skylights are 1.11 W/m²K
and 1 W/m²K respectively. The glazing to floor ratio
is 14%.
As seen in Figure 4, the case building is equipped with

a geothermal heat pump of 8kW capacity coupled with
a heat exchanger to provide floor heating in winter
and during summer period, passive floor cooling from
two vertical boreholes with a depth of 100 meters.
One drilling of 100 m provides a cooling capacity of
2.5 kW and a heating capacity of 5 kW. Furthermore,
a balanced mechanical ventilation system with total Figure 4. The scheme of the geothermal heat pump
Airflow of 275 m³/h, with heat recovery is installed, and the measurement locations.
which allows free cooling during the night.
2.2. Simulation Model

OpenStudio and EnergyPlus was used to perform the Table 1. U-values of construction elements.
dynamic simulations [14] [15]. First, the building Construction U-value (W/m²K)
envelope is drawn in SketchUp using the OpenStudio Ground Floor slab 0.21
SketchUp Plugin [16]. Then loads, schedules, HVAC- External Wall 0.19
systems etc. are modelled in accordance with the real Common Wall 0.31
building, to set up a detailed simulation model. Internal Wall 2.39
Pitched Roof 0.13
The building was divided into 3 main floors, where: (a) Flat Roof 0.15
ground floor was for entrance and storage, (b) first floor
(day use)-living room, kitchen and (c) second floor
(night use)- 3 bedrooms, bathroom, attic. However, Table 2. Thermal Zones and Ventilation flow rates.
for the simulation, the building has been divided into
Thermal Zone Ventilation Flow
13 thermal zones (see Figure 4 and Table 2). rates (m³/h)
Zone 1 (Entrance hall) -
The internal loads are assigned for each zone and corre- Zone 2 (Kitchen + dining) +
112.1
spond to the heat gains due to occupancy, lighting and Zone 3 (Living Room)
equipment (see Table 3). The occupancy is scheduled Zone 4 (Home Office/ Desk) 39.1
separately for weekday, weekend and summer vacations Zone 5 (Toilet-WC) 39.1
(See Figure 6 for a typical weekday Schedule). Four Zone 6 (Bedroom 1&2+ Dressing) 75
people are assumed to be at work/school from 9h to 17h Zone 7 (Bedroom 3) 32.8
from Monday to Friday except Wednesday afternoon Zone 8 (Bathroom) 75
from 13 h. Hours of occupancy are 132 h/week. The Zone 9 (Laundry) 62.6

TOP PAPERS
‘radiant fraction’ is set to 0.5 for the persons, electrical is only applied to the windows located on the south-west
equipment and lighting. This value is recommended by side of the building, both for the bedrooms (thermal
EN ISO 52016-1when performing simulations [17]. zone 6) and the living space (thermal zone 2 and 3). The
solar shading is modelled to be ON, when the global
Hours of occupancy in a space is a crucial factor solar radiation on the window reaches exceeds a value
impacting thermal comfort. In zone 2 and 3 (living of 250 W/m². It remains ON for 15 min, after which
room), it is assumed that the spaces are occupied during the control checks the radiation on the window. If the
the daytime. Concerning zone 6 (bedrooms), two dif- solar radiation on the window exceeds 250 W/m², the
ferent scenarios are created for simulation. In the first shading remains ON and if the value is lower than the
scenario, the bedrooms are only occupied during the threshold value of 250W/m², the shading is turned OFF.
night (11 p.m. – 7 a.m.). For the second scenario, an
occupancy from 4 p.m. till 7 a.m. is assumed. Since 2.3. Model Validation
the case study building was not equipped with solar Before evaluating the impact of future climate sce-
shading, a simulation scenario to evaluate the impact of narios on the performance of the floor cooling system,
solar shading on thermal comfort was implemented. An the simulation model must be validated. To do this,
automatic solar shading with g-value of 0.55 has been the indoor temperatures obtained from the model are
implemented in the simulation model. The solar shading compared with the indoor temperatures measured by
sensors on site. For this reason, the operative indoor
Table 3. Internal loads for each space for heat gains. [17] zone temperatures (from the model) are compared
with the data from measuring devices placed on-site. A
Room Persons Lighting Equipment weather station TMK was placed with 2 minutes time
(number) (W/m²) (W) step for monitoring the outdoor dry bulb temperature,
Desk (Zone 4) 1 2 250 relative humidity and solar radiation. Sensor HUBO
MX1102 with time step 10 minutes was placed in
level 1
Living area (Zone 3) 2 2 330 the bedroom (See Figure 5 for the position of the
sensor in the bedroom) to monitor the temperature,
Dining + kitchen (Zone 2) 2 1.7 108
Bedroom 1 + dressing (Zone 6) 2 2 0

Bedroom 2 (Zone 6) 1 2 30
level 2
Bedroom 3 1 2 30
Bathroom 1 4.6 0
Laundry room 0 0 3 200
Table 4. Adaptations in the simulation model for model

validation.
Adaptation Description
––Setpoint cooling: 23°C ±1°C
––Supply temperature: 18°C
Floor ––Measure operative temperatures
1
Cooling ––Max Flow rate cooling: 1200l/h
––Check parameters of circulation pump and heat
exchangers
––Measure: Add zone mixing Object
2 Air Flow
––Adjustments in the air flows between zones
3 Infiltration ––N50 = 2 h−1
––Ground Thermal conductivity: 2.1 W/Mk
––Ground thermal heat capacity: 3400000 PA:K
––Ground Temperature: 13°C
Ground Heat
4 ––Pipe thermal conductivity: 0.42 W/mK
Exchanger
––U_tube distance: 0.06 m
––Pipe Thickness: 0.003 m
––Flow rate loop: 0.00032 m³/s
Figure 5. Thermal Zones.

TOP PAPERS
relative humidity, CO₂ and dew point temperature. Table 5. Model validation (Summer period:
This sensor has a range between 0°C to 50°C with 1st July–14th September 2020).
an accuracy of ±0.21°C in the given range. The com-
Difference between simulated and Validation Status
parison is performed for two periods of 10 days in July measured temperature condition
(6.07-16.07) and August (17.08-27.08-post heat wave
living room 0.80°C <1°C
period) of the year 2020 (see the weather data from
MAE
bedroom 0.83°C <1°C
July and August 2020 in Figure 7). Validated
living room 1.03°C <1.5°C
RMSE
Figure 8 shows the measured data between 6th to bedroom 1.03°C <1.5°C
16th July 2020. The outdoor temperature reaches
a maximum of 24.1°C on 13/07/2020 at 2 pm.
The maximum indoor temperature for bedroom
(23.5°C) and living room (25.2°C) was observed on
12/07/2020 at 8.11 pm.
Improvements on the simulation model were made on

parameters of the floor cooling system, airflows, infil-
tration rates and ground heat exchanger to obtain the
best result (see Table 4). The entire summer period of
measured temperatures and simulated temperatures is
compared (1st July – 14th September 2020). There
are two conditions that the results of the simulation
model must meet to be considered that the indoor Figure 6. Weekday occupancy schedule.
temperatures as validated- (a) The MAE (Mean
absolute value of error) should be less than 1°C and
(b)RMSE (Root mean squared error) should be less
than 1.5°C [18].
2.4. Climate Scenarios

The impact on the performances of a floor cooling
system is assessed for four possible climate sce-
narios, representing historical, extreme weather data
(heatwave period) and future weather data (midterm-
2050s and longterm-2100s). The historical weather
data of 2010 represent a moderate climate and are
used as the reference scenario. The rising ground
temperature is also taken into consideration for the Figure 7. Weather data from 1st July to 31st August
dynamic-simulations. The weather data of Melle, 2020 indicating the heat wave period in the beginning
for Scenario 1, 3 and 4 are based on RCP8.5 and of August.
developed adopting the methodology IEA EBC
Annex 80-Resilient Cooling of Buildings [19].
Weather data of Melle for scenario 2 is based on the
observations by RMI for year 2020 [20].
In Figure 10, we can observe, the maximum tempera-

tures recorded in the summer months (April-August),
shows rising temperature trend from the historical to
the future long -term scenario. However, 2020, was an
exceptionally warm year with heat waves in July and
August and the observed weather data of 2020 had
higher maximum temperature for the summer months
than mid-term climate scenario. However, for radia-
tion, the variation between the 2 scenarios (historical Figure 8. Measured outdoor and indoor temperature
and midterm) in summer months are not significant. between 6th to 16th July 2020.

TOP PAPERS
But for 2020, heat wave scenario, the solar radiation

is significantly higher compared to the long-term
scenario.
The OpenStudio/EnergyPlus manual shows that

weather data (excluding specific ground temperatures)
only has an impact on the ground temperature at the
surface (up to 0.5 m). The ground heat exchanger is
placed to a depth of 100 meters. It is therefore crucial
that the ground temperature for the ground heat
exchanger is manually adjusted for the future climate
scenarios. For climate scenario 1 and 2, the ground
Figure 9. Comparison of the measured and Simulated temperature 13°C, 13.7°C for scenario 3 and 14.7°C
temperature of the living room. for scenario 4 has been implemented. With boreholes
up to 100 m deep, the choice was made to follow the
fastest increasing scenario at a depth of 75 m [8].
Table 6. Climate scenarios used for performance
assessment. 2.5. Thermal Comfort and Cooling Capacity
Assessment
Scenarios Description Method A as described in Annex F of the EN 15251 (15)
is selected for the evaluation of summer comfort. For this
1 Historical weather data 2010s (2000-2020)
study, the percentage of occupied hours when the opera-
2 Weather data -Melle 2020 (heatwave) tive indoor temperature is above 25°C, 26°C and 28°C
3 Future mid-term 2050s (2040-2060) is evaluated for a period from the 1st of April until 31st
October. The temperature thresholds also verify with
4 Future long-term 2090s (2080-2100)
the heat stress impact on human body, studied by [21].
The analysed period is an extension of the meteorological
summer because, assuming future climate scenarios,
summers can be longer. If the percentage above 25°C
is lower than 5%, the thermal comfort is considered
acceptable while lower than 3% is considered good.
The cooling capacity is calculated using the energy

equation (1). The cooling capacity of the floor cooling
system is increases/decreases depending on the inlet and
outlet temperature of the ground heat exchanger (∆T).
𝑄𝑄 = 𝑚𝑚̇ ∙ 𝑐𝑐 ∙ ∆𝑇𝑇 (1)
For the calculations on the supply side of the heat

exchanger, the following parameters have been
extracted from the simulation model (per time step
of 1 hour):
• heat exchanger: inlet temperature (°C)

• heat exchanger: Outlet temperature (°C)
• Flow rate flowing through the heat exchanger (kg/s)
To evaluate the thermal comfort, only the most critical

zones (living space (2&3) and bedrooms [6]) are con-
Figure 10. Temperature and solar radiation trends for sidered. To improve the thermal comfort, the impact
the summer months for all 4 climate scenarios. of solar shading is also investigated for all the zones.

TOP PAPERS
3. Results and Discussion

However, with the intervention of shading, all 4 sce-
3.1. Thermal comfort assessment
narios are within the acceptable limit. Thus, Solar
3.1.1. Bedroom (Zone 6)
shading will reduce 50% of occupied hours above
Zone 6 (Bedroom) is only occupied during the night. 26°C% for the long-term scenario and will guarantee
Two types of occupancy profile (11pm to 7 am) and
(4 pm to 7 am) has been simulated. Both the occu-
pancy scenarios are simulated with and without solar
shading interventions.
As seen in Figure 11, a good thermal comfort can be

guaranteed in all 4 climate scenarios with a maximum
of 1.4% of occupied hours above 25°C in the long-
term scenario. Occupancy profile 1 (11 pm to 7 am)
do not have any occupied hours above 26°C and 28°C
threshold with or without shading. However, when
the bedrooms are occupied for a longer period (4 am
to 7 am), the percentage of occupied hours above
25°C in long-term scenario increases to 7.1% (see
Figure 12). With higher occupancy, the percentage Figure 11. % occupied hours >25°C occupancy
of exceeding hours in Future long-term scenario can between 11 pm to 7 am, with and without shading
be decreased from 7.1% to 3.5% with the interven-
tion of solar shading (see Figure 12). For the climate
scenarios-2020, historical and also midterm, solar
shading decreases the exceeding hour percentage by
50%. There are no exceeding hours above 28°C with
solar shading. However, without solar shading, there
are 3.1% occupied hours exceeding 26°C, and 1.8%
occupied hours exceeding 28°C.
3.1.2. Living Room (Zone 2&3)

Zone 2&3 (Kitchen+ dining and Living room) is
occupied during the day. Zone 2 &3 is simulated with
and without solar shading interventions. As seen in
Figure 13, without solar shading, the percentage of
occupied hours above 25°C is higher than the accept-
able limit in all 4 climate scenarios. The percentage
of occupied hours above 25°C threshold reaches up Figure 12. % occupied hours >25°C (Zone 6- Bedroom
to 14.5 % in the long-term scenario. with Occupancy between 4 pm to 7 am, with and
without shading).
With the intervention of solar shading, the percentage
of occupied hours in historical and mid-term scenario
can be improved and brought back below the accept-
able 5% limit. However, even with solar shading,
percentage of occupied hours above 25°C is above the
acceptable level for the year 2020 (heat wave period)
and for long-term climate scenarios (see Figure 13).
Figure 14 illustrates the percentage of occupied hours

above 26°C threshold in Zone 2 &3, without and with
shading. Without shading, only in long-term scenario,
the percentage of occupied hours are above acceptable
limit (9.8%). Historical and mid-term scenario are in
good limits (below 3%), whereas year 2020 is 3.5%. Figure 13. % occupied hours >25°C (zone 2&3) without
and with shading.

TOP PAPERS
a good thermal comfort in the other scenarios. In the negative impact on the cooling capacity, even for
long-term, the indoor temperature will rise above cooling devices used in this case study.
28°C (0.5%) without solar shading. With shading,
no indoor temperatures above 28°C are measured. However, the technical data sheet of the system shows
a peak cooling capacity of 5Kw. This value was never
3.2. Cooling capacity assessment exceeded during all 4 climate scenarios, proving the
The impact of future climate scenarios on the cooling current system should be able to deliver the peak
capacity is illustrated in Figure 15. The results from cooling capacity in future climate scenarios. However,
the simulation confirm, climate change will have a peak cooling power decreases 1 kW over 80 years. This
is due to the rising ground temperature as a result
of global warming. Figure 16 illustrates the inlet and
outlet temperatures on the supply side of floor cooling.
The supply temperature of the floor cooling system was
set at 18°C to avoid condensation and this setpoint was
used in the actual system on site as well. For climate
scenarios 2, 3 and 4, this supply temperature cannot
longer be guaranteed. There is a rise in the inlet tem-
perature (0.15°C) for 2020, 0.16°C during mid-term
and almost by 1°C for the long-term scenario.
Analysis of the cooling capacity shows that the

maximum power output will decrease in the future
climate scenarios. For example, the maximum
power output will decrease by approximately 17%
Figure 14. % of occupied hours in Zone 2 & 3 above
by 2050 and 22.5% by 2090. The average cooling
26 without and with solar shading.
capacity will remain approximately the same. The
decrease in maximum cooling capacity is due to the
faster increase of the outlet temperature compared
to the inlet temperature of the heat exchanger. For
example, in climate scenario 4 (2090) the supply tem-
perature of 18 degrees cannot be guaranteed over the
entire period. On average, the supply temperature rises
to almost 19°C, which is the result of a rising source
temperature in the boreholes.
4. Conclusions
The results of the thermal comfort show that occu-
Figure 15. Cooling Capacity heat exchanger. pancy has a major impact on whether or not the
predetermined upper limits are exceeded. If the analysis
is performed for the living space with an upper limit
of 25°C, good thermal comfort cannot be guaranteed
for any climate scenario. If the upper limit is raised
to 26°C, the floor cooling system will only fail in
2090 and in rest 3 climate scenarios, it can guarantee
acceptable hours of thermal comfort. In the long-term,
only 0.5% of occupied hours was measured, with the
indoor temperature exceeding the upper limit of 28°C.
For the bedroom, assuming occupancy only at night
in the bedrooms, good thermal comfort is achieved
for all climate scenarios. With the occupancy profile
from 4 pm to 7 am, good thermal comfort cannot be
Figure 16. Average inlet and outlet temperature (Floor guaranteed 7.1% occupied hours above 25°C in 2090.
cooling). In the bedroom, assuming an upper limit of 25°C

TOP PAPERS
for both occupancy patterns, good thermal comfort the floor cooling system will not be able to reduce the
can still be guaranteed. However, implementation temperature below 26°C in each thermal zone. The
of solar shading shows much improvement in the inlet temperature of 18°C can’t be guaranteed in future
thermal comfort in the bedroom and especially in climate scenarios, which results in a lower maximum
the living-dining room (Zone 2 &3). In the living cooling capacity.
space, good thermal comfort is only obtained if the
comfort temperature limit is raised to 26°C, even in Thus, it can be concluded that in future climate
historical climate scenario. However, with intervention scenarios, shading is indispensable. Also, the cooling
of solar shading this can be improved for the 4 climate systems needs to be dimensioned keeping in mind the
scenarios. This verifies the necessity of solar shading rise in air and ground temperature to guarantee good
especially in highly glazed and south or west facing thermal comfort to the users.
zones in the buildings.
5. Acknowledgement
The decrease in cooling capacity of the floor cooling
This study was conducted as part of the curriculum for
in mid-term climate scenario is 17%. In the long- Master thesis of students Pieter Proot and Tom Trioen
term, the impact of climate change is greater. Even under the supervision of Prof. Hilde Breesch and PhD
assuming an upper limit of 26°C, good thermal student Abantika Sengupta. The data of the case study
building and the monitoring data was provided by
comfort is not obtained. With regard to the cooling Thomas More University of Applied Sciences as a part of
capacity of the floor cooling, the decrease is also their sCOOLS project.
greater-22.5% compared to the existing condition.
In the future, the system will therefore have to be
6. References
dimensioned differently, for example by increasing
the flow rates or additional (active or passive) cooling at: https://proceedings.open.tudelft.nl/clima2022/article/
systems will have to be implemented. In the long- view/326.
term, the current design of this dwelling including
SESSION 1
Achieving healthy Zero Emission
Buildings with the EPBD
u s !
FROM 10 TILL 12
j o i n
SESSION 2
REPowER EU: phasing-out fossil
fuels in buildings JOIN US!
FROM 13 TILL 15
POLICY CONFRENCE ON 15 NOVEMBER

AT THE THON HOTEL BRISTOL STEPHANIE
Acrex India
2023
TOP PAPERS
A Multi-Domain Approach to
Explanatory and Predictive Thermal
Comfort Modelling in Offices
EUGENE HENK W. BRINK MARCEL G.L.C. ROEL C.G.M. HELIANTHE S.M.

MAMULOVA Eindhoven University LOOMANS LOONEN KORT
Eindhoven of Technology, Eindhoven Eindhoven Eindhoven University
University of Eindhoven, the University of University of of Technology,
Technology, Netherlands & Hanze Technology, Technology, Eindhoven, the
Eindhoven, the University of Applied Eindhoven, the Eindhoven, the Netherlands
Netherlands Sciences, Groningen, Netherlands Netherlands
[email protected] the Netherlands
Abstract: It is well known that physical variables, such as temperature, exert a significant
influence on occupants’ thermal comfort in office buildings. Despite this knowledge, models
that are currently used to predict thermal comfort fail to do so accurately, resulting in a
mismatch between design conditions and actual thermal comfort conditions. The assumption
is that exclusive attention to physical variables is insufficient for understanding or predicting
thermal comfort. The question arises as to how a multi-domain approach can aid in
explaining and predicting thermal comfort in offices. In this study, a unique dataset containing
indoor environment, demographic, occupancy and personality related variables is used to
construct two types of thermal comfort models. The dataset contains 524 observations,
collected during summertime in two office buildings in the Netherlands. Firstly, structural
equation modelling (SEM) is used to construct an explanatory model, with the aim to identify
significant variables affecting thermal comfort, as well as the interactions between them.
Secondly, machine learning is used to train four binary classification models to predict
thermal discomfort. For the investigated cases, SEM suggests that thermal discomfort is
significantly affected by (i) temperature, (ii) sound pressure level, (iii) the interaction between
temperature, sound pressure level and illuminance, and (iv) the interaction between
gregariousness and occupancy count. The four predictive models are subsequently trained
using only the significant variables. Nevertheless, the weighted F₁-score for all four models
ranges between 0.55 and 0.59, indicating weak predictive performance. The results show
that significant influencers are not necessarily good predictors of thermal discomfort. Future
researchers are encouraged to combine explanatory and predictive modelling techniques,
in order to test whether variables that are relevant to the domain are useful for prediction.
Keywords: Thermal comfort, multi-domain, personal domain, interaction effects, structural
equation modelling, machine learning.

TOP PAPERS
1. Introduction 1.2 thermal comfort variables

Thermal comfort is that condition of mind that Existing multi-domain studies identify several vari-
expresses satisfaction with the thermal environment ables that are of interest to thermal comfort modelling.
[1]. Building engineers refer to building standards to A list of main effects and interaction effects that are
predict the thermal comfort conditions for a given supported or rejected by existing research on multi-
design. However, current standards do not always domain thermal comfort in offices is composed [4],
produce adequate thermal comfort predictions [2]. leading to the following hypotheses:
Researchers in the field of thermal comfort seek
to understand and predict thermal comfort, using M1: Air temperature exerts a positive, exponential,
explanatory and predictive models. Explanatory effect on thermal discomfort.
models typically employ statistical techniques that M2: Sound pressure exerts a positive effect on
provide insight into what influences thermal comfort thermal discomfort.
in offices. Predictive models are built to forecast the M3: Occupant gregariousness exerts a negative effect
thermal comfort conditions for a given office space. on thermal discomfort.
I1: Air temperature exerts a negative effect on the
Recent research efforts have focused on multi-domain interaction effect between sound pressure level
approaches that treat thermal comfort as a combina- and illuminance on thermal discomfort.
tion of variables belonging to four domains, outlined I2: Occupant assertiveness exerts a positive effect
in Figure 1 [3]. Their relevance is apparent but their on the effect of air temperature on thermal
presence in existing thermal comfort models is limited discomfort.
[4]. The combined presence of all four domains is I3: Occupancy count exerts a positive effect on the
almost non-existent [4]. Moreover, the majority of effect of occupant gregariousness on thermal
existing studies focus on explanatory modelling [4]. The discomfort.
rift between design conditions and real-world condi-
tions is in part attributable to the absence of a suitable The aforementioned hypotheses are tested via an
thermal comfort model. In consequence, it is important explanatory model, using field measurement data. The
to pursue better prediction of thermal comfort in office results are used to train a model that aims to predict
buildings and it is worthwhile doing so using the multi- whether office employees are experiencing thermal
domain approach. This study looks at existing thermal comfort or discomfort. The articulation of the mod-
comfort models to identify potential variables that may elling outcome is unprecedented in current literature,
aid in better explaining and predicting thermal comfort. covering three physical variables (air temperature,
THERMAL COMFORT
PHYSICAL CONTEXTUAL SOCIAL PERSONAL

DOMAIN DOMAIN DOMAIN DOMAIN
Indoor Geography Demographic Psychology

Thermal / V isual / Climate / Culture Gender / Age / Personality traits /
Acoustical / Body Mass Index State of af fect
Air Quality Design
Envelope / Activity Physiology
Outdoor Façade / Space / Task / Activity Physiological
Thermal Interior signals
Work
System Tenure / Job type /
Type / Operation / Working hours
Control / Access
Status
Occupancy Economy /
Education LEGEND
Time
Time of day / Clothing level / DOMAIN
Weekday / Season Exposure Factor
Variable
Figure 1. Physical, social, contextual and personal variables present in literature, adapted from [3].

TOP PAPERS
illuminance and sound pressure level), one contex- Nocc are continuous. Variables a1 − a2 and g1 − g4 are
tual variable (occupancy count), two personal variables ordinal. All variables are normalized using min-max
(occupant assertiveness and gregariousness) and one feature scaling. To account for multivariate non-
social variable (gender), in the interest of testing normality, robust diagonally weighted least squares
whether such a multi-domain approach can aid in a (DWLS) estimation, known as weighted least square
better understanding or prediction of thermal (dis) mean and variance adjusted estimation (WLSMV) in
comfort in offices. ‘lavaan’ package, is used to compute the parameter
estimates, robust standard errors and fit indices.
2. Research methods 2.2 predictive modelling
The data was collected prior to this study, in two office The predictive model takes the form of a binary classi-
buildings in the Netherlands. The cross-sectional fier that predicts whether a participant is experiencing
campaign was conducted during the years 2015-2018. thermal comfort or discomfort. The variables included
The applied measurement protocol is described in a in the model are listed in Table 2. Two linear and
publication by Brink and Mobach [5]. The data points two non-linear classification algorithms are selected
used in this study are limited to the warmer months of and trained using the scikitlearn Python library [8].
June and July 2016. 623 office employees participated Po is used for linear algorithms, while P1 is used for
in the measurements. The final sample size is equal non-linear algorithms, as the latter are expected to
to 522. The data consists of objective and subjective capture non-linear relationships. The linear algorithms
measurements.
2.1 explanatory modelling Table 2. Variables used for prediction.

Explanatory modelling is performed via structural
equation modelling (SEM); a covariance-based tech- Variable Symbol
nique that enables the inclusion of observable and Indoor temperature exponent Po
unobservable variables. Visualization is done using Indoor temperature P1
standard LISREL matrix notation [6]. The computa- Sound pressure level P2
tion is performed via the ‘lavaan’ package [7]. Sound × illuminance × temperature P3
Gregariousness × occupancy count P4
Table 1 provides an overview of the variables used,
Gender P5
along with their notation. Variables Tin, SPL, E and
Table 1. Direct and indirect effects included in the SEM model.

Effect Domain Symbol SEM Variable Range [unit]
expTin x8 Air temperature 20 – 26 [°C]
Physical SPL x9 Sound pressure level 40 – 70 [dB(A)]
E Illuminancea 0 – 2,000 [lx]
g1 x1 Gregariousness
Direct g2 x2 Gregariousness
g3 x3 Gregariousness
Personal
h1 y1 General body discomfort
h2 y2 Lower body discomfort
h3 y3 Upper body discomfort
Physical SPL · E · Tin x10 Sound, illuminance and temperature
Tin · a1 x4 Temperature and assertiveness

Indirect Physical and personal
Tin · a2 x5 Temperature and assertiveness
Nocc · g1 x6 Occupancy count and gregariousness Nocc < 20
Contextual and personal
Nocc · g4 x7 Occupancy count and gregariousness Nocc < 20
a The direct effect of illuminance is excluded but illuminance is used to compute SPL · E · Tin.

TOP PAPERS
are logistic regression (LR) and linear support-vector

machine (L-SVM), while the non-linear algorithms are Table 3. Parameter estimates for the thermal comfort
random forest ensemble (RF) and non-linear support- variables included in the SEM model.
vector machine that uses the radial basis function
kernel (RBF-SVM). During the testing phase, the Estimate SE Z P(<|z|)
models are retrained on 308 observations, comprising x8 0.643 0.203 3.177 0.001a
the training and validation sets, and are tested on the x9 0.357 0.151 2.368 0.018b
remaining 77 observations. Common metrics such as x10 −0.383 0.196 −1.951 0.051c
the F1-score, accuracy (ACC) and the area under the
ξ1 −0.128 0.174 −0.736 0.462
ROC curve (AUC) are used.
ξ2 0.104 0.382 0.272 0.785
ξ3 0.394 0.198 1.988 0.047c
4. Results a CI – 99.9%. b CI – 98%. c CI – 95%.
3.1 structural equation modelling
The outcome of the explanatory modelling phase is a on η1. The main effect of sound pressure level x9 is
SEM model. Figure 2 shows the parameter estimates, found to be positive and significant at approximately
variance/covariance estimates and factor loadings for 98% confidence (z > 2.33, p < 0.02).
the explanatory model. The parameter estimates are
also shown in Table 3. The exponent of air tempera- The interaction between indoor temperature, sound
ture x8 is expected to have a positive effect on thermal pressure level and illuminance x10 is expected to exert
discomfort η1. According to the results, the effect of a negative effect on η1, such that an increase in indoor
x8 on η1 is positive (see Figure 2) and significant temperature will result in a decreased audio-visual
at 99.9% confidence (z > 3.09, p < 0.001). Sound influence. The parameter estimate for the three-way
pressure level x9 is expected to exert a positive effect interaction x10 is found to be negative and significant at
Figure 2. Graphical representation of model estimation.

TOP PAPERS
95% confidence (z > 1.96, p < 0.05). Gregariousness above random guessing (= 0.50) and is not sufficient
x1 is expected to exert a negative effect on η1. The effect for predicting thermal (dis)comfort.
of x1 on η1 is found to be negative but it is not found
to be significant. The interaction between assertiveness
5. Explaining thermal comfort
and indoor temperature x2 is expected to be positive,
to the extent that an increase in temperature will result The interpretation of the SEM model addresses the
in an increased influence of assertiveness on η1. The hypotheses M1− M3 and I1− I3. The model estimates
two-way interaction x2 is found to be positive but it is do not reject M1, M2, I2 and I3, leading to several
not found to be significant. The interaction between implications that may be of interest to the under-
gregariousness and occupancy count x3 is expected to standing of thermal comfort in offices:
be positive, such that an increase in occupancy count
will result in an increased influence of gregariousness • During the cooling season, an increase in indoor
on η1. The two-way interaction x3 is found to be sig- temperature results in an exponential increase in
nificant at approximately 95% confidence (z > 1.96, thermal discomfort.
p < 0.05). As a result, hypotheses M1, M2, I1 and I3 • An increase in sound pressure level results in an
are not rejected. increase in thermal discomfort.
• An increase in air temperature decreases the effect
3.2 binary classification that the interaction between sound pressure level and
The outcome of the predictive modelling phase are illuminance has on thermal discomfort, resulting in
four models; LR, L-SVM, RF and RBF-SVM. LR is a negative three-way interaction.
fitted as shown in equation (1). The polarity of the • An increase in occupancy count increases the effect
parameter estimates is consistent with hypotheses M1, of occupant gregariousness on thermal discomfort,
M2, I1 and I3, suggesting the model learned a similar resulting in a positive two-way interaction.
pattern to the one captured using SEM.
The results support the notion that the model may be
used to explain thermal comfort. However, the exist-
𝑃𝑃(𝑌𝑌= =
𝑃𝑃(𝑌𝑌 1|𝑋𝑋)
1|𝑋𝑋)
(1)
ence of a near-equivalent model is likely. The reliability
of the subjective data, particularly assertiveness and
𝑒𝑒𝑒𝑒𝑒𝑒(−0.27 + 0.29𝑋𝑋
𝑒𝑒𝑒𝑒𝑒𝑒(−0.27 + 0.30𝑋𝑋2 +
+0 0.29𝑋𝑋 − 0.13𝑋𝑋3 + 0.17𝑋𝑋
0.30𝑋𝑋 4 + 0.23𝑋𝑋5 ) + 0.23𝑋𝑋5 )
= 0 2 − 0.13𝑋𝑋3 + 0.17𝑋𝑋4gregariousness, is questionable. A better fit may be
=1 + 𝑒𝑒𝑒𝑒𝑒𝑒(−0.27 + 0.29𝑋𝑋0 + 0.30𝑋𝑋2 − 0.13𝑋𝑋3 + 0.17𝑋𝑋4 + 0.23𝑋𝑋5 )
1 + 𝑒𝑒𝑒𝑒𝑒𝑒(−0.27 + 0.29𝑋𝑋0 + 0.30𝑋𝑋2 − 0.13𝑋𝑋3 + 0.17𝑋𝑋 via5 )the use of a more extensive and well-
4 + 0.23𝑋𝑋
achieved
known scale, such as the IPIP-NEO-120 [9].
The performance metrics for the validation and testing
phases are reported in Table 4. The difference in per-
6. Predicting thermal comfort
formance across the models is very slight and all four
models yield similar scores across all three metrics. The SEM model suggests that Po−P5 significantly
While L-SVM and RF show better ACC and weighted affect thermal comfort in offices. Yet, the four predic-
F1 on the validation set, they no longer outperform tive models are not capable of adequately predicting
the other models on the test set. The increase in ACC thermal (dis)comfort. Looking at all four outcomes,
during the testing phase for all four predictive models the quality of the data may have introduced noise,
could be attributed to random variation between data masking the patterns necessary for making reliable
splits. The predictive performance of the models is just predictions. However, real-world data is noisy and
constitutes a pitfall for even the most prevalent models.
Table 4. Performance metrics (validation and testing). A predictive model can be expected to perform even
worse in practice than it does on the mother data set.
Model Set AUC ACC F1
The results show that thermal comfort is a complex,
Valid 0.58 0.56 0.53 multi-domain construct that is difficult to predict.
LR
Test 0.68 0.56 0.56 However, the performance of the four predictive
Valid 0.58 0.61 0.61 models does not cast a definitive shadow over the
L-SVM
Test 0.67 0.55 0.55 prospect of accurate prediction. Predictive models that
Valid 0.62 0.60 0.60 include a larger number of thermal comfort variables
RF and higher quality subjective measurements may yield
Test 0.64 0.58 0.59
better predictions. Moreover, other, more advanced
Valid 0.57 0.52 0.48
RBF-SVM modelling techniques, such as stochastic modelling,
Test 0.66 0.57 0.58
may be better suited for thermal comfort prediction.

TOP PAPERS
7. Conclusion
This study is part of a broader research effort to achieve
This study applies the multi-domain approach to better prediction of thermal comfort in offices, which
thermal-comfort modelling. An explanatory model is is an essential step in the building design process. The
constructed using SEM. The specified model examines results formulate a basis for further research on the influ-
the influence of indoor temperature, illuminance, ence of indoor climate, occupancy and personality traits
sound pressure level, occupancy count, gregariousness on thermal comfort in offices, as well as the interaction
and assertiveness on thermal discomfort. The SEM between the different influences. Moreover, the findings
model is unique, as it is the first explanatory model, have direct implications for the engineering sector, as
derived from field measurements, to include multiple they suggest that influences such as sound pressure level,
physical and personal variables, while also including occupancy and personality traits, should be considered
contextual variables. The following conclusions are when designing for optimal thermal conditions.
derived from the explanatory model:
6.1 limitations
• Thermal discomfort increases at higher indoor This research is subject to several limitations, the
temperatures and higher sound pressure levels, mitigation of which is encouraged in the future.
suggesting that both should be optimized and Firstly, prominent variables such as correlated colour
maintained. temperature and air velocity are not included in the
study. Similarly, variables such as age, relative humidity,
• Uncomfortably high indoor temperatures decrease clothing insulation and metabolic rate are excluded due
the effect that sound pressure level and illuminance to insufficient variability in the measured data. Secondly,
otherwise have in a comfortable thermal environ- extreme indoor conditions are not observed during field
ment. This highlights the importance of designing measurements. In addition, the measurements are limited
for optimal temperature conditions and constitutes to summer conditions in the context of the Netherlands
a basis for the use of personalized strategies. and are not representative of cooler conditions or other
climate regions. Due to this limitation, the relation-
• Gregarious individuals may be more thermally ship between temperature and thermal discomfort is
comfortable than non-gregarious individuals when assumed to be exponential. Future studies are encour-
there are many occupants in the room. Designers aged to include cold sensation data and thereby model a
are encouraged to account for inter-individual dif- parabolic relationship between temperature and thermal
ferences by providing flexible working conditions. discomfort, where thermal discomfort increases at lower
and higher temperatures both. Thirdly, the internal con-
Four predictive models LR, L-SVM, RF and RBF-SVM sistency of the personal variables is poor and they are
are trained using significant variables Po−P5. The not sufficiently representative of the Big Five personality
models examine the predictive potential of the explana- traits. Lastly, the quality of the predictive models may
tory model. All models struggle to predict thermal (dis) be improved via the use of advanced hyper-parameter
comfort, despite the inclusion of significant thermal tuning, a larger variety of machine learning algorithms
comfort variables. The results bring to light several and more advanced modelling methods.
conclusions:
• Significant thermal comfort influences are not

7. Acknowledgement
always adequate predictors thereof.
The data is provided by the research group of Mark
Mobach at the Hanze University of Applied Sciences. The
• Researchers are advised to precede future thermal authors acknowledge the contribution of Yasin Toparlar,
comfort studies with explanatory modelling, to Deerns Groep B.V.
facilitate the creation of predictive models that
contain a large variety of variables.
• Combined use of explanatory and predictive model-

8. References
ling is necessary, to test whether variables considered
in thermal comfort research hold theoretical rel- at: https://proceedings.open.tudelft.nl/clima2022/article/
evance, predictive potential, both or, perhaps, view/181
neither.

International Conference
World Sustainable
Energy Days 2023
28 February - 3 March 2023
Wels/Austria
World
Sustainable Energy
transition
Energy Days Energy

security!
Conferences:
2023
European Energy Efficiency Conference
European Pellet Conference
Industrial Energy Efficiency Conference
Smart E-Mobility Conference
Young Energy Researchers Conference
Innovation Workshops
www.wsed.at
OÖ Energiesparverband, Linz/Austria, Tel. +43 732-7720-14386, [email protected], www.wsed.at
TOP PAPERS
Exploring futures of summer comfort in

Dutch households
LENNEKE KUIJER LADA HENSEN CENTNEROVÁ
Future Everyday Group, Department of Building Performance Group,
Industrial Design and Eindhoven Institute Department of the Built
Photo by Angeline Swinkels
for Renewable Energy Systems, Eindhoven Environment, Eindhoven

University of Technology, Eindhoven, University of Technology,
the Netherlands Eindhoven, the Netherlands
[email protected]
Abstract: This article describes the results of a qualitative study into cultural aspects of
summer comfort in Dutch households. Results show that two main ways of preventing
overheating – shading and summer night ventilation – compete with active cooling. Moreover,
it identifies potential for technologies, policies and procedures to support adaptation to
higher temperatures.
Keywords: Summer comfort, Netherlands, cultural change, qualitative study, dwellings, energy.
Introduction
Actions in the present, such as building policies,
The Dutch Meteorological Institute (KNMI) defines a proposing standards (e.g. EN ISO 52000-1,
heatwave as a period of at least five consecutive days with EN ISO 52016-1), designing infrastructure, building
daily maximum temperatures exceeding 25°C, with at technologies, and passing on instructions have an
least three of the five days reaching maximum tempera- effect on shaping futures of summer comfort in Dutch
tures above 30°C. In the coming decades, heatwaves are households. These actions, in turn, are informed by
expected to become longer, warmer, and more frequent visions, assumptions and expectations of what futures
[1]. On top of this, dwellings in the Netherlands have are possible, desirable and likely to come about.
long been built with a focus on keeping warm during However, insight into these futures from a Dutch
winter, resulting in high levels of insulation and airtight- household perspective is so far limited.
ness that increase the risk of overheating. Other factors
contributing to risks of overheating are large windows, This raises questions like: To what extent are main-
urbanisation, and an ageing society. stream Dutch households equipped and able to equip
themselves to deal with longer, warmer, and more
Dutch households are beginning to adjust their lives frequent heatwaves? Which strategies do households
and homes to these new circumstances. For those who apply and aspire to achieve comfort in times of hot
can afford it, mitigating discomforts and health risks weather, and which currently not? What are possible
of hot weather are within reach, but tend to require consequences of these strategies for levels and patterns
high amounts of energy. Essent, one of the main energy of energy demand, and general well-being? What are
providers in the Netherlands reported a 30% increase in developments outside of these households that may
energy demand during the August 2020 heatwave [2], affect these strategies?
and a 2021 study by research institute TNO [3] showed
that 20% of Dutch households already have a form of This article summarizes a selection of findings from
active cooling, while 26% is considering to get it. a qualitative exploration of the future of domestic
summer comfort in the Netherlands. A full version of
At this point in time, Dutch responses to global warming the study results is available in an open access stake-
can still go in many directions, some of which are holder report [4]. While the focus of the study is on the
undesirable from health, inclusivity, and environmental Netherlands, its outcomes may be relevant beyond this
points of view. Therefore, it is important to try and context, particularly in countries where active cooling
anticipate, and where possible redirect these pathways. is currently on the rise due to global warming.

TOP PAPERS
Method
contribute to summer comfort in a low-energy manner.
The study consisted of a set of interrelated research However, their establishment is hampered by existing
activities, primarily involving 21 household interviews infrastructures, historically shaped cultural values and
and 10 domain expert interviews. habits, and competition between different activities.
The sections below elaborate on these points.
The household interviews were designed to capture
the main daily domestic activities that are relevant to Actual and experienced overheating
summer comfort. These were: (1) cooking and eating, Overheating is already an issue in Dutch households.
(2) personal care and clothing, (3) laundering and From the perspective of the NZEB (Net Zero Energy
cleaning, (4) home working, (5) free time, (6) sleeping, Buildings) standards, several dwellings in the study (all
and (7) ventilating, shading and cooling. apartments) exceed the threshold of 450 WHOs (the
Weighted Overheating Hours), which in the Dutch TOjuli
The interviews were conducted around a heat wave -indicator start counting above 27°C. Overall, reported
in August 2020 (Figure 1). No actual measurements indoor daytime temperatures during the heat wave ranged
of indoor temperatures were made, so conclusions from 24°C to 45°C for dwellings without cooling. The
regarding temperatures are based on self-reported dwellings that reported temperatures over 30°C in their
values, which tend to be less reliable. The use of work- main living areas were all rented city apartments.
books with daily exercises, which primed participants
to notice their indoor temperature values in the week Stories of residents confirm that these dwellings become
before the interview, partly compensated for this. practically unliveable during a heat wave. Moreover,
most households in the study considered their dwelling
The domain experts included HVAC, sleep, physiology, to be overheated well below the formal overheating
fashion, architecture, building standards, social housing threshold. Indoor temperatures above 25°C were con-
and domestic shading. These interviews and additional sidered too high by all but four participants, particularly
sources, such as trade fairs and observations were used to for working and sleeping. These higher temperatures
identify and extrapolate current cultural, demographic inhibited their freedom of movement and capability to
and technological trends in domestic summer comfort. go about their daily business, such as focusing on work,
sleeping well, and performing housework.
Findings
These issues were absent for households with active
Overall, the study confirms the expectation that the use cooling, but they were inhibited in other ways,
of active cooling in Dutch dwellings is likely to increase particularly, in being outdoors. The data indicated
in the future. This is reflected in growing sales figures of that households with active cooling have a stronger
cooling systems, but also in actual and experienced over- tendency to take the car instead of walk or cycle,
heating in dwellings, as well as frictions with emerging because stepping outside from a cooled space felt
practices of shading and ventilation. The latter, along like ‘hitting a wall’. In other words, spending time in
with embodied acclimatisation, have potential to cooled spaces reduced their willingness and ability to
tolerate the higher outdoor temperatures.
With climate change, these overheating issues are

expected to grow. The next sections go deeper into the
strategies that households currently apply and aspire
to deal with these issues.
Acclimatizing
Participants that were able to enjoy or accept the heat
and modify their daily schedules around it were most
capable of getting through the heatwave without too
much discomfort. For example, families that had their
summer holidays during the heat wave or a young
person practicing mindfulness. However, the freedom
to adjust one’s daily schedule is not accessible for
Figure 1. Timing of household interviews in relation to the everyone, especially if heatwaves are to occur more
August 2020 heatwave. Graph shows average temperatures often outside of summer holidays. Moreover, not all
measured at De Bilt weather station obtained from KNMI.nl. [4]

TOP PAPERS
bodies are equally capable of dealing with heat and Active cooling is more explicitly approached with reser-
these capabilities decrease with age [5,6]. vation. Participants without active cooling are familiar
with air conditioning, but find it too energy-consuming,
However, the study shows that common knowledge noisy, expensive, and uncomfortable. However, even
among the general public on bodily responses to heat highly committed, knowledgeable residents in modern
show a gap with state-of-the-art research, particularly homes, equipped with the latest shading and ventilation
regarding the role of sweat in dealing with heat (it is technologies, had trouble maintaining a comfortable
mostly seen as something negative) and the capability of indoor climate without the use of active cooling. Many
bodies to adjust to higher temperatures over time. While anticipated getting some form of active cooling in the
research shows that people can adapt to heat by as much future. Those who already had cooling were mostly
as 1°C per day as confirmed by the physiology expert, content with their systems (except for mobile air con-
none of the participants referred to concepts of bodily ditioners). Although there are cultural and practical
adjustment to heat over time. This finding indicates that frictions to integrate active cooling into Dutch house-
adjustments in knowledge, available products, skills, and holds, they seem easier to overcome than those related
attitudes to acclimatize could reduce people’s experi- to shading and ventilating. Added to this lower barrier
ences of being locked into their homes and bodies, and to uptake is the risk that active cooling creates a further
contribute to well-being in a low-energy manner. Put threat for shading and ventilation practices to reach their
more strongly, the adverse mental and physical effects potential because they compete.
of reduced physical activity in hot weather might be
partly mitigated if people (are facilitated to) acclimatize. Shading, ventilating and cooling compete
Shading and active cooling can complement each
Cultural frictions with shading, ventilation other in dwellings, but the study illustrates how
and cooling they compete in the market. Both active cooling and
A seemingly embedded friction that arose from the outdoor shading require considerable investment.
interviews is the relationship that ‘the Dutch’ have
with warm weather. Warm and sunny weather is associ-
ated with being outdoors and enjoying the light and
warmth of the sun. In the spring, when days get longer
and warmer, people open doors and windows to let
fresh air in, extending their living spaces onto balconies
and into gardens. Fluctuating temperatures mean that
Dutch summers can have relatively cool spells that
precede heatwaves. When temperatures go up, the
sun is initially welcomed into the home. But when
temperatures rise, this behaviour leads to overheating,
which is then difficult to correct.
Proper, disciplined outdoor shading and summer night

ventilation routines could reduce the extent to which
indoor spaces heat up [7], but adopting these routines
requires more than new equipment and behaviours.
Viewing the sun as an ‘enemy’ instead of a ‘friend’ for
part of the year requires a cultural shift. The Dutch
friendship with the sun is deeply embedded in customs
(opening doors, curtains and windows to enjoy light,
views and fresh air), the built environment (ample, sun
facing windows) and related professional practices such
as architecture (disliking and sometimes prohibiting
outdoor shading). For most of the year, the sun is and
will remain a friend, helping to light and warm dwell-
ings, and keep people healthy and cheerful. Learning
to occasionally ‘cool’ this friendship requires a cultural Figure 2. Examples of inconsistent Dutch practices
shift that is necessary for the potential of shading and of shading and ventilating during hot weather (28°C).
ventilation practices to develop (Figure 2). [Lenneke Kuijer, August 2022]

TOP PAPERS
If households have an opportunity to only invest in equipment that could hamper the development of
one, then cooling has the better position in terms of other strategies to deal with a warming climate.
low-effort comfort. This competition is also visible in
the current NZEB requirements, where adding a form Automation can play a role here. For example, when
of active cooling eliminates incentives for other, low- shading responds automatically to levels of solar gain,
energy measures against overheating such as shading. rain and wind, and ventilation to temperature and
humidity differences inside and outside the dwelling.
Cooling and summer night ventilation compete However, the role of residents cannot be ignored. Not
directly in the dwelling. While the cooling system is only their autonomy in deciding whether to have these
on, windows and doors need to be closed to retain the systems at all, but also in the ways they are used.
microclimate. This effect is even stronger for mobile air The study revealed a wide array of circumstances in
conditioners, used in three participating households, which people might disable automated shading, such
because securing the hose in the window can further as feeling locked-in, wanting more light, annoyance
hamper the opening of windows when the device is not with repeated movement, wanting to open windows,
in use. During the day, the hose in the window neces- etc. For ventilation systems, it became clear that their
sitates a (partly) open window enabling hot outside automated responses to CO₂ or humidity levels can
air to enter the dwelling (Figure 3). conflict with summer comfort by drawing in hot
outside air during the day, while summer night venti-
In general, active cooling, when properly designed and lation, in most homes, requires residents to open and
installed, can secure comfortable temperatures in the close windows while they are sleeping. Further research
dwelling regardless of other measures such as shading or is needed into this direction.
ventilation. Shading and ventilation require the active
involvement of residents. With active cooling in place, Pathways for active cooling
the incentive to invest money, time, and effort in them is So far, active cooling is discussed as one practice, but
reduced. Mobile air conditioners have a particularly prob- in fact, different forms of active cooling are currently
lematic position in this respect because of their relatively developing in parallel. Main pathways are radiant
low threshold, and energy-efficiency. While they can be cooling and air-conditioning. Radiant cooling, mainly
life-savers on the scale of individual users, in the broader in underfloor settings powered by heat pumps or
picture these appliances form undesirable symptoms of district cooling are relatively slow systems that cool
overheating in Dutch dwellings that contribute to the the building mass. Such systems are likely to run
problem of climate change and heat islands [8]. continuously during hot weather. An advantage of
these systems is that for ground source heat pumps,
Smart automation cooling can be provided on low-energy demand, or
These insights can be used to design measures that even energy-positive manners when heat is stored for
might slow down or prevent Dutch households from use in winter (Figure 4).
becoming dependent on energy intensive cooling
WINTER SUMMER
COLD WARM WARM COLD

WATER WATER WATER WATER
Figure 3. (Mobile) air-conditioners interfere with Figure 4. Storing summer heat in a ground source for
ventilating practices. [Lenneke Kuijer, August 2022] use in winter. [4]

TOP PAPERS
Air-conditioners work more quickly by directly cooling materials, media and installer instructions for cooling
the indoor air and are more likely to be used based systems recommend setting the system to 18°C or
on occupancy and direct demand. Spaces also heat 27°C, introduce some other metric like a combined
up again relatively quickly when they are off. Mobile humidity/temperature value, or are designed to offer
air-conditioners allow for even more directed, person- a variable temperature that moves with the outdoor
oriented, albeit fleeting forms of cooling. Apart from temperature and slowly increases over time to support
their efficiency, these different patterns of use are likely acclimatisation.
to affect their overall energy demand.
Additional effects of warming on everyday life
Moreover, levels of energy demand for active cooling
do not only depend on the type of systems and when The use of active cooling in the home seems to lead
it is used, but also on the set-temperature. At present, to a dependence on cooled spaces that extends beyond
the temperatures to which households will set their the dwelling. The examples in the study indicate a
cooling systems has not been settled or stabilised, but trend towards spending more time indoors, and the
it is likely that norms around acceptable and normal car becoming preferred over other means of transporta-
temperature ranges will develop in the coming years tion. Beside increases in CO₂ emissions and energy
and decades. With heating, for example, Dutch house- costs that accompany the increased use of most forms
holds presently tend to set their thermostat somewhere of active cooling, these trends indicate undesirable
between 18 and 22°C. This normalised temperature health effects resulting from lower activity levels and
range has formed and changed over a long time period lower natural vitamin D intake.
[9], and varies per cultural context [10].
Other areas in which increases in energy demand are
As illustrated in the introduction, building norms, likely to arise according to our finding are in increased
standards and system design can play an important capacity for cold food storage, showering and laun-
role in shaping these norms. Considering that tech- dering. Several households reported that fruits and
nologies co-shape practices, it makes a huge difference vegetables that are normally kept in dry storage are
for the way in which summer comfort practices moved to refrigerators during hot weather, where they
will develop whether default settings, promotion compete for space with more cooled drinks (Figure 5).
Figure 5. Demand for cold food storage increases in hot weather, but fridges and freezers heat up indoor spaces.

TOP PAPERS
This leads to an increased demand for (larger) fridges explicit demand for shading and cooling seems low
and freezers—appliances which, in turn, directly con- among social housing tenants. This could have all
tribute to overheating in dwellings due to the heat kinds of causes such as other more pressing issues on
they produce. the tenants’ minds, a fear of unmanageable rises in
rent, unfamiliarity with the effects of shading and
The study also indicated that shower frequencies ventilation on overheating, or better skills of acclima-
increase during hot spells. The main reason for more tizing. Despite various efforts to involve low-incomes
showers within the sample was not to cool down, but households, they were only present indirectly in the
to rinse off sweat. This requires water, as well as energy study through stories of higher income tenants, experts
to heat it. With a trend towards better insulated homes, and observations during fieldwork. More research is
these secondary effects could become significant. needed into the specific issues, wishes and strategies
of this group.
Conclusions
This study set out to gain more in-depth insight into
the ways in which Dutch households are likely to deal
with hot weather in their dwellings. Several opportuni-
ties were identified that might direct Dutch domestic
practices of summer comfort onto more inclusive,
healthy, and less energy-intensive pathways.
A range of opportunities present themselves around

acclimatisation, i.e., modifying bodily relations with
hot weather. There seems to be a gap between state-
of-the-art physiological research on how bodies deal
with heat and everyday knowledge among the house-
holds. The benefits of sweating (when combined with
drinking enough water) as an effective way to deal
with heat is not fully acknowledged. Moreover, none
Different consequences for different types of of the participants talked about bodily adjustment to
households and dwellings heat over time, while experts confirm that this effect
While it is difficult to draw a strict line, some of the can be as strong as 1°C per day.
dwellings in the study were clearly overheated. These
examples represent a larger group of households for Outdoor shading during the day and ventilation
which homes become unliveable for part of the year. during the night can reduce or prevent overheating in
Smaller, well-insulated dwellings, with higher window- low-energy ways. The study shows that active cooling,
to-content ratios, high sun exposure (e.g. in high-rise), while spreading quickly, reduces residents’ acceptance
little shading and ventilation opportunities, located in of the experienced downsides of shading (including
cities (heat islands) heat up quickly. Such dwellings are the costs) and lowers incentives and opportunities to
more likely to be occupied by lower-income house- utilise cooler night air.
holds and are more often rented than owned. This
might also mean that the potentially higher amount of
time spent at home by the residents due to lower levels Acknowledgement
of employment could add to the overheating issues. We want to thank our experts and participants. This
research was funded by the Dutch Research Council
Also judging from recently introduced building stand- (NWO) under grant number VENI17343.
ards in the Netherlands and elsewhere [11], overheating
is slowly starting to be acknowledged as an issue, and
social housing providers and landlords are beginning to References
contemplate on how to intervene. The study indicates Please find the full list of references in the original article
that the costs of installing and maintaining outdoor at: https://proceedings.open.tudelft.nl/clima2022/article/
shading on non-ground floor windows plays a role view/388
in hampering tenants and owners to act. Moreover,

TOP PAPERS
Correlation of subjective and objective

air quality data in shopping centres
as a function of air temperature and
relative humidity
MAHMOUD EL-MOKADEM KAI REWITZ DIRK MÜLLER

RWTH Aachen University, E.ON RWTH Aachen University, E.ON RWTH Aachen University, E.ON
Energy Research Center, Institute for Energy Research Center, Institute for Energy Research Center, Institute for
Energy Efficient Buildings and Indoor Energy Efficient Buildings and Indoor Energy Efficient Buildings and Indoor
Climate (EBC), Aachen, Germany Climate (EBC), Aachen, Germany Climate (EBC), Aachen, Germany
Abstract: Germany has 493 shopping centres mostly located in urban cities. According
to STASTICA, the number of shopping centres was doubled in the last two decades [1].
For consumers, good indoor air quality (IAQ) is a basic requirement for their shopping
experience. This leads to very high air exchange rates for current operation of HVAC
systems in shopping centres. Accordingly, achieving good IAQ in combination with increasing
energy efficiency is a main issue for operation of the shopping centres. Thus, in previous
studies the intensity of shopping product emissions was evaluated by trained subject panels.
In this paper, we analyse the IAQ parameters not only by trained human panel, but also by
analysing the volatile organic compounds (VOC) through objective tests. In a first step, we
cluster five different product groups: books, clothing, shoes, coffee and perfume. Regarding
these groups, we measure the emissions with objective and subjective tests for a variation
of temperature and relative humidity. We use the results to investigate the influence of these
environmental parameters on the subjective and objective intensity of the VOCs. Finally,
we analyse the subjective data along with the objective data, to correlate the subjective
evaluations with the measured sensor signal of the multi VOC sensor system. The evaluation
is done with statistical data analysis methods such as Friedman test. The results show the
potential for the metal oxide semiconductor sensors technology for detection of VOCs and
for prediction of perceived intensity based on objective data. Furthermore, the results show
an influence of air temperature and humidity on subjective perception.
Keywords: Indoor air quality, shopping centre, total volatile organic compounds, E-Nose, VOC
1. Introduction
volume flow control. However, running the ventilation
Currently, air quality control in shopping centres is based system with fixed volume flows does not lead to energy
on fixed high air exchange rates or on the measured CO₂ efficient operation and controlling the ventilation
concentration, which is a good indicator for the emis- system based on only one parameter (CO₂) may not be
sions of persons and allows for demand-based supply sufficient from an air quality perspective. For example,

TOP PAPERS
2.3 Subjective and objective tests

odours and contaminants emitted from shopping centre
products are not detectable by CO₂ sensors. The results of previous studies with subjective tests
confirm the significant decrement of the acceptability
Odours perceived by humans can often be traced back of perceived air quality with increasing temperature
to so-called volatile organic compounds (VOCs). and humidity [5–8].
These substances occur in the air in the form of gases
and vapours. Even low VOC emissions are often The aim of the current study is to collect both subjec-
associated with significant odour perceptions and can tive and objective data for the evaluation of odours or
lead to health problems. In contrast to CO₂ sensors, air quality for selected product groups and to derive
VOC sensors can detect mixtures of substances in the possible limit values. In particular, the influences
indoor air with a characteristic signal [2]. However, of air temperature and relative humidity are to be
the control of ventilation using VOC sensors is rarely analysed. The objective evaluation is carried out with
implemented, since no specific limit value is defined the developed VOC sensor system. The subjective
that could be used to as a setpoint. evaluation is carried out with a group of test persons
trained according to DIN ISO 16000-28 with regard
Due to the large number of VOC emissions in to acceptance and perceived intensity by using acetone
shopping centres, the goal is to define acceptable comparative scale [9].
objective limits for different product groups. Hence,
extending current CO₂-based control by VOC-based 2.3.1 Emission chambers
control can improve air quality. The study is conducted in the air quality laboratory
of the Institute for Energy Efficient Buildings and
Indoor Climate (EBC), RWTH Aachen University.
2. Methodology
The product groups are divided into 5 categories:
2.1 E-Nose: Sensor Methodology clothing, books, shoes, perfume and coffee.
In this study, metal oxide semiconductor sensors are
selected for the detection of volatile organic com- The products are filled into so-called emission
pounds. Their operating principle is based on the chambers. The chambers are made of stainless steel
dependence of the electrical conductivity of metal to minimize the influence of oxidation reactions on
oxides on the gas concentrations and gas types present. the air. In addition, these chambers have a connec-
Metal oxide semiconductor sensors can be divided into tion for introducing conditioned supply air via a
thin-film and thick-film sensors and exhibit, among central ventilation unit and an outlet fitted with a
other things, high sensitivity to low gas concentrations, glass cylinder where test subjects can evaluate the air
a long service life and a low price, which potentially quality. Control of the supply air volume flow rate
qualifies them for use in the demand-based control of by measuring the actual condition via an orifice plate
ventilation systems. On the other hand, their non-linear allows a precise adjustment of the air exchange in the
sensor characteristics depending on the gas concentra- emission chambers. To achieve good mixing even at
tion make calibration difficult. Moreover, drift and low air changes, an additional mixing fan, which cir-
aging behaviour cannot be neglected. Furthermore, culates the air in the chamber, is installed in each case.
there is a cross-sensitivity of additional parameters such
as relative humidity and air temperature [3].
2.2 E-Nose: Sensor system development

For the setup of the VOC sensor system, different
metal oxide sensors of the MQ sensor series are used
to enable the detection of several gases, such as CO,
CO₂, H₂, NH₃, NOx, ETOH, alcohol, and formalde-
hyde. The MQ sensors are connected to a single board
Arduino Mega microcontroller for power supply and
data transfer. In addition, a temperature sensor and a
humidity sensor are integrated to compensate for cross-
sensitivities. With the help of a TVOC sensor, the sum
of volatile organic compounds can be recorded. Thus,
total air pollution can additionally be estimated [4].
Figure 1 shows the final VOC sensor system. Figure 1. Developed E-Nose.

TOP PAPERS
2.3.3 VOC classes for shopping centre

Figure 2 shows an example of the positioning of the
products and method of detection
clothing and the sensor system in one of the emission
chambers. The sensor system is positioned directly in A classification of product groups from shopping centres
front of the mixing fan. or their emissions with respect to different VOC classes
are investigated. Shoes, clothing, books, perfume and
2.3.2 Acetone comparative scale and subject coffee, representing the food courts, are identified as
panel relevant and suitable product groups for the main study.
To evaluate the intensity of the shopping products Sensors with correspondingly high sensitivity are also
emissions, an acetone comparative scale is used assigned to the VOC classes. For example, formaldehyde
according to DIN ISO 16000-28. The test rig allows which can be emitted from new clothing, new shoes, or
reproducible acetone concentrations. According new books can be detected with HCHO sensor [12].
to DIN ISO 16000-28, the acetone concentration Sanaeifar et al. use the MQ-135 sensor for the detec-
in mg/m³ can be converted into PI. Here, 0 PI cor- tion of aromatic compounds emitted from food [13].
responds to an acetone concentration of 20 mg/m³. Perfumes and detergents contain alcohols, which can be
The increase by 1 PI corresponds to a linear increase detected with the MQ-3 sensor. Cosmetics containing
of the acetone concentration of 20 mg/m³. alkanes can be detected with the MQ-2 sensor [14].
In previous investigations Hegemann et. al used a Since only three emission chambers are available, the
maximum intensity of 14 PI for a shopping centre study is divided into two sub-studies. In the first sub-
product (shoes) [10]. Since the current study also study, clothing, shoes and perfume (variant 1) and
tests other products such as perfume, for which in the second sub-study, books, coffee and perfume
significantly higher ratings are expected, the upper (variant 2) are evaluated. Perfume is tested for two
end of the comparative acetone scale is extended source strengths, as the highest ratings were recorded
to 28 PI (580 mg/m³). In choosing this value, the here. The second variant corresponds to a 60% reduc-
stated limits of 590 mg/m³ by the National Institute tion in source strength compared to the first variant. In
for Occupational Safety and Health (NIOSH) for each sub-study, the group of test subjects is also divided
exposure over 24 hours are met [11]. into two sub-groups due to hygiene protection with
regard to the COVID-19 pandemic, so that the first
According to DIN ISO 16000-28, the minimum size sub-group participated in the study in the morning and
of a subjective group, for the evaluation of acceptance, the second sub-group in the afternoon.
is 15 persons (untrained) and for the evaluation of
perceived intensity is eight persons (trained). For this 2.3.4 Test procedure
study, 17 successfully trained test persons participated We selected one setting per day for the air temperature
in the tests. and the relative humidity of the supply air, which is kept
constant during the day. Two hours before the arrival
of the test subjects, the test stand is switched on to
create stationary boundary conditions. After a welcome
and brief acclimatization of the subjects, the evaluation
begins. Once an evaluation is completed for all three
emission chambers, the volume flow rate or air exchange
rate is varied. During the test run, the subject group is
almost exclusively in the air quality laboratory. After all
assessments are completed, the subjects are dismissed.
The parameter variations carried out were chosen to be
inside the acceptable thermal comfort range boundaries
[15]. Temperatures are chosen to 20°C, 23°C and 27°C,
relative humidity is selected to 30% and 50% and air
change rate is varied in 6 steps between 5.1 to 1 h−1.
This leads to 36 evaluations per subject and product.
In addition to the evaluation of the perceived intensity,

the percentage of dissatisfied subjects is analysed using
Figure 2. Positioning of the E-Nose and the product the question: “Imagine you were exposed to the air
inside the emission chamber. from the emission test chamber for several hours a day.

TOP PAPERS
Would you rate the odour as acceptable?” “Yes” and

“No” are available as response options. The proportion
of no ratings is defined as the proportion of dissatisfied.
3. Results
3.1 Influence of temperature and humidity
on the perceived intensity
In Figure 3, the ratings per air exchange setting
averaged over all test subjects are plotted against tem-
perature for a relative humidity of 50%. These mean
values are shown as coloured circles per product group.
Based on the results, a direct proportional relationship
between air temperature and perceived intensity can be
seen for each product groups. Furthermore, the differ- Figure 3. Influence of temperature on the perceived
ence in perceived intensity due to the product groups is intensity at 50% relative humidity.
significantly higher than the influence of temperature.
Figure 4 shows the effect of relative humidity on

the perceived intensity for perfume (variant 1). The
results for the relative humidity of 30% are shown as
red circles, respectively as black crosses for 50%. In
addition, the regression lines are drawn. Both regres-
sion lines run almost parallel with the same slope.
However, the perceived intensity at 50% humidity is
rated higher by about 2 PI compared to 30% humidity.
The results for the other product groups show similar
behaviour. Overall, it can be concluded that an increase
in temperature and humidity leads to an increase in
perceived intensity.
To evaluate the statistical significance of the tem-

perature and humidity variation with respect to the
subjects’ evaluation, the Friedman test is used. The Figure 4. Influence of relative humidity (rel. hum.) on
Friedman test is a statistical test suitable for evaluating the perceived intensity for perfume (variant 1).
non-parametric data as it does not assume a normal
distribution of the data [15]. As a result of the test
statistic, the “P-value” is compared. In our case, if
P > 0.05, the null hypothesis cannot be rejected, and
a difference between the samples distributions cannot
be detected. If P < 0.05, the null hypothesis is rejected,
and the alternative hypothesis is accepted.
The Friedman test leads to P-values smaller than the

defined significance level (P < 0.05) with range from
3.30e−13 to 6.95e−55, so that statistically significant
differences due to the variation in temperature and
humidity can be assumed.

on acceptance
Figure 5 shows the percentage of dissatisfied people
for the perfume product (variant 1) across the different Figure 5. Influence of temperature and humidity on
scenarios. The values in each scenario represent an percentage of dissatisfied for perfume (variant 1).

TOP PAPERS
average value over all test subjects and analysed air

change rates. The percentage of dissatisfied people
increases with rising air temperature and rising relative
humidity. The influence of the temperature is higher
at a relative humidity of 50% than at 30%.
In Figure 6, the percentage of dissatisfied people is

plotted against the perceived intensity. The coloured
symbols are mean values of the 17 evaluations by the
test persons for each combination of product, tempera-
ture, relative humidity and air change rate. The range
of ratings for the products studied varies. The highest
ratings for perceived intensity and the proportion
of dissatisfied persons exist for the product perfume
Figure 6. Correlation between perceived intensity and (variant 1). The lowest ratings occur for clothing and
percentage of dissatisfied. books, with an intensity of approximately 5 PI resulting
in a proportion of dissatisfied of approximately 20%.
The data shows that the proportion of dissatisfied
people increases with the perceived intensity.

on the sensor signal
In the following, the results for the MQ-3 sensor are
presented since it has the highest sensitivity to the
products investigated. Figure 7 shows the sensor signal
as a function of temperature for various products at
a relative humidity of 50%. For all products except
clothing, the sensor signal increases with increasing
air temperature.
Comparing this with the subjective data from Figure 3,

a similar behaviour can be seen for the sensor signal as
a function of temperature. When looking at the order
of the perceived intensities with the measured sensor
Figure 7. Influence of temperature on the sensor signal signals, it can also be seen changing in the ranking
at 50% relative humidity for MQ-3 sensor. of the products between the subjective and objective
output. For example, the subjects rate coffee as more
intense than shoes, books and clothing, whereas the
sensor signal is only weaker for clothing than for coffee.
Figure 8 shows the sensor signal as a function of tem-

perature for perfume (variant 1) at a relative humidity
of 30% and 50% for the MQ-3 sensor. The sensor
signal increases with increasing relative humidity. This
behaviour is similar to the observation for the evalua-
tion of perceived intensity in Figure 4.
3.4 Correlations between objective and

subjective evaluation
To use VOC sensors for demand-controlled ventilation
in shopping centres, limit values are needed that can
be used as set points for the controlled variables. In
Figure 8. Influence of relative humidity on the sensor Figure 9, the perceived intensity (subjective values) is
signal for perfume (variant 1) for MQ-3 sensor. plotted against the sensor signal for the MQ-3 sensor.

TOP PAPERS
For the different product groups the point clouds in on the basis of the five product groups investigated,
the diagram have different positions. In consequence, whose functions, the coefficients of determination R²
the correlation of the perceived intensity and the and the root mean square error (RMSE) are contained
sensor signal should be done depending on the product in Table 1.
groups. In addition, linear regression lines are shown
In each case, a linear regression approach is used to
derive a correlation between the sensor signal in mV
Table 1. Derived correlations between sensor signal in
and the perceived intensity in PI. The coefficient of
mV (x) and perceived intensity in PI (y).
determination for the separate analysis of the product
Data Correlation R² RMSE groups is smaller than the overall regression data. It is
noticeable that for books, coffee, and shoes, the slope
All data y = 0.05 · x + 4.75 0.65 3.2 PI
of the regression function is the same, and only the
Books y = 0.026 · x + 3.72 0.40 1.4 PI Y-axis intercept differs. In Figure 10, the percentage of
Coffee y = 0.026 · x + 11.48 0.30 1.5 PI
dissatisfied persons is plotted against the sensor signal
of the MQ-3 sensor for the product groups investi-
Clothes y = 0.053 · x + 3.39 0.40 0.9 PI gated. It can be observed, that with increasing sensor
Perfume y = 0.05 · x + 5.20 0.65 2.5 PI signal the percentage of dissatisfied is rising. However,
a strong scattering of the data along the regression line
Shoes y = 0.026 · x + 7.09 0.20 1.5 PI
drawn can be seen. This makes it difficult to derive
an exact objective limit value to reach a specific per-
centage of satisfied people or not to be too restrictive
to guarantee a minimum percentage of dissatisfied.
4. Conclusion
In order to detect emissions from building materials
and products in shopping centres, VOC sensors are
suitable in principle. Based on the results of the study
carried out, it is possible to correlate the objective
sensor signals with subjective evaluations for individual
product groups. However, the determination of exact
limit values in order to achieve a certain acceptance
of the air quality must be set with caution due to
the scattering of the data. With a larger amount of
data, which would require further studies, it would
Figure 9. Correlation between the sensor signal of MQ- also be possible to use machine-learning methods to
3 and the perceived intensity. analyse the data. With these, it might be possible to
derive correlations between multiple sensor signals and
subjective evaluations much more easily and robustly.
Important findings were also obtained in connection
with the influence of air temperature and humidity
on subjective perception. Thus, for all products inves-
tigated, the perceived intensity and the proportion
of dissatisfied persons increase with rising tempera-
ture and relative humidity. These findings can help
implement an energy-efficient heating, cooling, and
ventilation system operation in shopping centres.
5. References
Figure 10. Correlation between the sensor signal of view/418
MQ-3 and the percentage of dissatisfied.

ARTICLES
Cross-infection risk between two people standing
close to each other at different room temperatures
A COBEE 2022 conference paper shows new experi-
ments with different room temperatures, Nielsen et al.
PETER V. NIELSEN (2022). The paper shows that the different lengths and
Aalborg University
[email protected]
directions of exhalation, obtained at different room
temperatures, will influence the cross-infection risk
between two people standing close to each other.
I
f we look at the exhalation flow from a person standing The paper shows that hot surroundings do increase
in a room with Mixing Ventilation and an air temper- the cross-infection risk to a high level at high room
ature Ta of 21°C, we will see that the exhalation will temperatures. It also shows that the cross-infection risk
take an upward direction. The reason is that the pulsating between two people standing close to each other will
exhalation flow has an decrease in cold surroundings.
initial temperature Texh of
34°C. The exhalation will Figure 3 shows video stills of the exhalation flow from the
be entrained with room index manikin. The exhalation flow is primarily governed
air in a forward movement by momentum in the first part, independent of the sur-
and will decrease at tem- rounding room temperature. It is obvious that the buoyancy
perature level, but it will effect changes the movement in an upward direction for
keep a temperature higher the room temperatures 17°C and 25°C in the later flow.
than the surroundings and
therefore have an upward Figure 1. Pulsating exhala-
movement, in principle to tion flow in surroundings with
the ceiling level, Figure 1. fully mixed flow. Liu et al. (2009)
The horizontal length of the exhalation flow is a function

of the temperature difference Texh − Ta. If the tempera-
ture difference is large, as in cold surroundings, we will
obtain a short length, but we will on the other hand have
a substantial length in hot surroundings. The horizontal
lengths will also depend on the person’s activity level and
on activities such as speaking, singing, coughing etc.
Things are different in

a room with stratified
flow as in the case of
Displacement Ventilation. Figure 3. Exhalation flow from an index manikin standing
Figure 2 shows that the 1.0 m in front of a susceptible (target) manikin Each row
exhalation will be strati- shows the flow at one of the three room temperatures:
fied at a certain height, 17°C, 25°C and 33°C.
which could be just above
the height of a person’s Figure 2. Pulsating exhala- When the room temperature is 33°C, there will be almost
mouth. The increased tion flow in surroundings isothermal surroundings due to an exhalation tempera-
temperature versus height with stratified flow. The room ture of the same level and the buoyancy effect will not
will lock the exhalation at temperature at the height of be present. The horizontal direction seems to be blocked
a certain height. the mouth is 21°C. Liu et al. (2009) by the thermal plume from the target manikin. Figure 3
shows that the growth rate of the vertical height of the
When we conduct research, we often investigate the exhalation flow is reduced in the Ta = 33°C case, indicating
usual situations. For example, we consider the room that it is a stratified flow with a reduced turbulence level.
temperature to be at comfort level, around 21°C.
This is not the case in many practice situations where The cross-infection risk is increased from 2 to 6,
it could be more than 34°C in the summer in some expressed as normalized exposure in case of mixing
countries and it could be 10°C in shops, for example, ventilation, when the room temperature increases from
handling foods. 23 to 33°C. (Distance between persons are 0.35 m).
Literature: please see the online version at rehva.eu

ARTICLES
Smart air quality control in residences

for optimised energy use and improved
health of occupants
ELIZABETH COOPER YAN WANG

Dr, Lecturer Dr, Research Associate
UCL Institute for Environmental UCL Institute for Environmental
Design & Engineering Design & Engineering
Introduction
and carbon emissions are generally the intentions
The premise of the work described here was to deter- behind the design of a building control system. These
mine if a tailored application of building information systems, in general, share a common structure; first,
and communication technologies might improve the sensor measures an environmental parameter (e.g.,
occupant comfort and health. This work explored the temperature, CO₂ concentration); second, the sensor
novel concept of connecting the predicted effects of a sends collected data to the controller which then uses
building control system with a health impact assess- pre-programmed control logic to determine the direc-
ment, an important and innovative step in the creation tion of change (if any). Window control systems have
of holistic and responsive building controls. become a research topic gaining increasing interest in
recent years, and research into HAPs has, indepen-
People in the UK, as with most of the Global North, dently, become a growing area of study. Although both
spend nearly 65% of their time at home, where con- automatic window controls and HAPs have, in recent
centrations of particulate matter less than 2.5 µm in years, but of increasing interest, an integrated system,
diameter, (PM2.5) can be much higher than in outdoor which controls both window operations and HAPs,
air due to occupant behaviours such as cooking and has not yet been investigated.
smoking. PM2.5 has been linked to many serious health
effects, including lung cancer, stroke, heart disease,
Methods
and asthma. The good news is that previous research
has reported that portable home air purifiers (HAPs) The proposed control framework included two modes;
equipped with High Efficiency Particulate Air (HEPA) for the non-heating period, the control framework has
filters can effectively reduce PM2.5 levels in the rooms both HAP and window controls running in parallel;
in which they are used. An important factor which has during the heating period, windows are set to be closed,
the potential to conflict with air filtration strategies and the mechanical ventilation with heat recovery
in buildings, is occupants’ operation of windows. The (MVHR) system operates continuously to provide
operation of windows also exerts a substantial impact background ventilation, and the HAP is enabled.
on thermal comfort and building energy consumption. The two control modes are shown schematically in
In this context, the research presented here aimed to Figures 1 and 2. In both modes, when the indoor
develop a novel building control framework in which PM2.5 concentration reaches the defined ‘HAP-on’
the operation of windows and the use of portable home threshold (15 µg/m³) the HAP switches on and con-
air purifiers were optimised for energy efficiency and tinues to operate until the concentration falls below
occupant health. the defined ‘HAP-off ’ threshold (5 µg/m³).
Building control systems play a central role in building When the residence is occupied, the default state of
operations and performance. Satisfying occupants’ the window is fully open to optimise natural ven-
comfort and minimising building energy consumption tilation. However, the window is set to fully close

ARTICLES
when certain conditions are met; either Yes

Flat vacant at step i
the indoor temperature falls outside
the limits of EN 16798-1 Category II No
adaptive comfort temperature, or when
both the outdoor PM2.5 concentration is
higher than that indoors, and the indoor Yes
temperature is within the comfort zone, PM2.5 in (i) > PM2.5 upper ?
to reduce the working load of the HAP.

No
Quantitative health impact assessments

are used to estimate future rates of
No
mortality and morbidity from different HAP off HAP on at step (i–1) ? HAP on
interventions compared to what is pre-
Yes
dicted without such changes. In the
work presented here, life-table models
were used to quantify the impacts on
mortality from reductions in indoor Yes No
PM2.5 in (i) < PM2.5 lower ?
PM2.5-concentrations due to the imple-
mentation of the control framework.
Figure 1. Algorithm for HAP control. (PM2.5 in: indoor PM2.5 concentration,
The life-table method is based on age- PM2.5 upper: HAP-on threshold, PM2.5 lower: HAP-off threshold, i: time step).
and sex-specific mortality rates, which are
used to calculate probabilities of survival
by year-of-age and calendar year. An Flat vacant at step i
Yes
impact assessment is performed using
the underlying mortality rates which No
are adjusted to reflect changes in mor- Yes

tality risk from changes in exposure by Is it night �me at step i ?
applying relative risks calculated using No

available epidemiological evidence.
Individual single-year survival probabili- Yes Tcomfort.lower ≤ Tin (i) ≤ Tcomfort.upper
ties are multiplied together to calculate
cumulative probabilities of survival over
No
multiple years. These cumulative survival
probabilities are applied to a population
Yes No Tin (i) > Tcomfort.upper No Window
allowing the calculation of life years lived PM2.5 in (i) > PM2.5 upper?
closed
by the population (where one life year is Yes
a full year of life lived by one person),
which in turn can be used to estimate PM2.5 out (i) > No Yes No
PM2.5 in (i) + �PM2.5 PM2.5 out (i) < PM2.5 lower? Tin (i) > Tout (i)
the average remaining life expectancy per
person by age. Yes
Yes No
No
Results Tin (i) > Tin (i - 1)
The proposed hybrid control framework, Window Yes

unchanged
along with either HAP or auto-window
Window
control alone for comparison, was tested open
in building simulations for a one-bed-
room apartment. The simulation results
for a summer week are presented below. Figure 2. Algorithm for window control. (Tin/Tout: indoor / outdoor
temperature, Tcomfort.upper / Tcomfort.lower: upper / lower limit of comfort
In the baseline scenario, without any temperature, PM2.5 in / PM2.5 out: indoor / outdoor PM2.5‑concentration,
control measures, there were morning ΔPM2.5: the maximum accepted difference between indoor and outdoor
and evening peaks of indoor PM 2.5 PM2.5‑concentration, i: time step).

ARTICLES
concentration, and the daily mean of indoor PM2.5 In the auto-window control mode (Figure 4), there was
concentration exceeded the WHO 24-hour limit a significant reduction in the peaks of indoor PM2.5
(15 µg/m³) for most of the week. However, the indoor concentration. The number of days exceeding the WHO
temperature remained within the comfort range the limit of PM2.5 concentration fell by two from the baseline
entire week (Figure 3). and occupant thermal comfort was still satisfied.
Figure 3. Summer week: Baseline. (a) Indoor and outdoor PM2.5-concentrations with the daily mean of indoor
PM2.5 concentration compared with the WHO guideline; (b) indoor, outdoor and adaptive comfort temperatures;
(c) window state schedule; (d) occupancy schedule
Figure 4. Summer week: Auto-window mode. (a) Indoor and outdoor PM2.5-concentrations with the daily mean
of indoor PM2.5 concentration compared with the WHO guideline; (b) indoor, outdoor and adaptive comfort
temperatures; (c) window state schedule.

ARTICLES
Figure 5 shows the HAP control mode where the The major advantage of the joint control of HAPs and
peaks of indoor PM2.5-concentrations were further windows was that the window could be shut when
reduced from both baseline and auto-window control. outdoor pollution was high before an accumulation
However, because of high outdoor PM2.5 levels, even of PM2.5 could occur indoors. In this way, the indoor
with the use of HAPs there were still two days when PM2.5 concentration was lowered while optimising the
the daily mean concentration of indoor PM 2.5 was HAP operation and efficiency. Meanwhile, the control
above the WHO limit. algorithm was directed, when the outdoor air quality
was good, to open the window to allow for natural
The hybrid control mode, shown in Figure 6, repre- ventilation.
sents the proposed control method. With both HAPs
and windows controlled, the indoor PM2.5 concentra- The simulated reductions in indoor PM 2.5 -
tion never decreased exceeded the WHO daily limit, concentrations, from baseline, for each of the control
whilst indoor temperature was maintained within the modes were used to estimate the mean years of life
comfort range. gained (YLGs) for females and males in the UK over
Figure 5. Summer week: HAP mode. (a) Indoor and outdoor PM2.5-concentrations with the daily mean of indoor
PM2.5 concentration compared with the WHO guideline; (b) HAP operation schedule.
Figure 6. Summer week: Hybrid mode. (a) Indoor and outdoor PM2.5-concentrations with the daily mean of indoor
PM2.5 concentration compared with the WHO guideline; (b) indoor, outdoor and adaptive comfort temperatures;
(c) window state schedule; (d) HAP operation schedule.

ARTICLES
a lifetime (97 years). The mean YLGs for males were Table 1. Summary of life-table model estimates of
approximately 6.5 million, 15 million, and 18 million changes in mortality from different environmental
for the automatic window/MVHR, HAP, and hybrid control strategies based on modelled PM2.5-
modes respectively. For females YLGs were approxi- concentrations in case study flat.
mately 6 million, 14 million, and 16 million for the Males Females
three intervention scenarios, respectively (Table 1).
The reduction in exposure to PM2.5 from the imple- Mean Mean
Control YLG YLG
days days
mentation of the hybrid mode added a mean of nearly mode Mean
gained
Mean
gained
6 months of life for both males and females.
Auto-
window / 6,557,926 73 5,948,462 64
Key Messages MVHR
For building designers and engineers, this work high- HAP 15,209,453 168 13,739,074 148
lights the importance of recognising health impacts, as
well as energy efficiency and environmental impacts, Hybrid 17,940,660 199 16,188,821 175
related to occupant-centric building design and
operation.
The implementation of smart building control systems,
For policymakers, this research adds technical evidence such as those modelled here, has the potential to reduce
of the impact of including health metrics in the exposure to indoor pollutants such as PM 2.5 which
building sector. could have substantial population health benefits.
Air Filtration in
HVAC Systems
No.11
This Guidebook will help the designer and user
to understand the background and criteria for air
filtration, how to select air filters and avoid problems
associated with hygienic and other conditions at
operation of air filters. The selection of air filters
is based on external conditions such as levels of
existing pollutants, indoor air quality and energy
efficiency requirements.
REHVA 40 RUE WASHINGTON 1050 BRUSSELS, BELGIUM

+32-2-5141171 [email protected] WWW.REHVA.EU

ARTICLES
Why a unified IAQ approach is critical to

securing public health
MORTEN SCHMELZER
Technical Marketing Director,
Head of Public Affairs,
Systemair Group
[email protected]
Morten Schmelzer, Technical Marketing Director, Head of Public Affairs, Systemair Group,
traces the rise and fall of IAQ in the public and political consciousness, the shortfalls of
standards and legislation promoting good indoor air, and how addressing the lack of common
language and related discussions is critical to achieving a healthier population.
How IAQ has evolved over the years

There have been positive movements globally, with
Over the years, the market’s acceptance of Indoor Air schemes such as WELL and RESET, initiated around
Quality (IAQ) has varied. In response to the 1973 2010, gaining further traction. These schemes primarily
and 1979 energy crises, the building industry focused aimed to qualify IAQ in major industrial cities con-
mainly on insulation with limited attention to IAQ. fronted with pollution. Meanwhile, IAQ has become a
As modern buildings became increasingly tighter to criterion in LEED and BREEAM certification schemes.
optimise energy usage, ventilation was (re-) discovered
to bring fresh air into these tight buildings. Yet, still, The COVID-19-pandemic has given the correct atten-
there was no genuine interest in the health aspects of tion to IAQ worldwide, as everyone had to face the
indoor air from ventilation. Architects and regulatory fact that good ventilation is an essential element of
committees neglected the topic, and there was little public health. It has been at the forefront of discus-
to no concern about IAQ improving the health of sions related to legislation ever since. Unfortunately,
occupants. the energy crisis in Europe and rising energy prices are
putting the momentum towards IAQ awareness at risk
In time there was a shift in consciousness among archi- as energy optimisation is becoming a higher priority
tects and building specialists. Since the 1990s, IAQ once again.
has gathered momentum, attracting scientific interest,
following the development of innovative IAQ equip-
Falling short in IAQ standards
ment within the HVAC industry.
Given these global trends, it is critical to take a closer
During the early days that IAQ was gaining traction, look at the deficits in the standards and regulations
it was mostly considered an element of building that further hamper wider acceptance and adoption
protection, with a strong focus on mould. While it of better IAQ practices and the reasons behind them.
became a consideration for special applications such as
operating theatres in hospitals, it remained primarily Today, there remains no clear framework for IAQ.
overlooked in schools, offices, and residential build- The approaches to and uptake of IAQ have been dif-
ings where the only ventilation practice commonly ferent among the EU Member States. In many Member
used was opening windows. States, IAQ has been added or is being added to the

ARTICLES
national transpositions of the Energy Performance of will occur when the new European Commission is in
Buildings Directive (EPBD). Depending on the relative place, and the EPBD is reviewed anew, as it is set for
strength of local manufacturers, some see a preference recast. The proposed EPBD recast aims to translate
for specific solutions in the EPBD transpositions. the actions proposed in the Renovation Wave, placing
more emphasis on the need for improved IAQ through
Airflow volumes in Member States differ. Between well-maintained mechanical ventilation systems in new
The Netherlands and Germany, for example, the and existing buildings. In the report by the European
rules are different and typical standards have very Parliament Committee on Industry, Research and
high airflow volumes oriented to Scandinavian require- Energy (ITRE) rapporteur, there is a proposal to
ments. The flexibility among the Member States also address Indoor Environmental Quality (IEQ), not
varies. Some can design lower or higher, and some are IAQ. The proposal for a new article 11a would require
more restrained or based on certain factors. In France, Member States to set minimum IEQ standards. These
for instance, minimum flow is based on mould protec- IEQ standards would have to be according to a meth-
tion for residential buildings. Even during Covid, there odological framework defined by the Commission.
were different interpretations among the Member Unfortunately, this framework is not yet available.
States on the national level.
Part of the lack of alignment stems from the fact that
Regional standards and guidelines related to IAQ exist. the industry has not been pushing for voluntary stand-
An example is ISO 16890, a series of product standards at the CEN level, either as a separate Technical
ards on air filters for general ventilation. However, Committee or integrated into existing Technical
its acceptance among Member States deserves to be Committees. This may be because many people in the
further improved. There are also standards concerning industry now regard standardisation too much as only
IEQ classifications available. Examples include: the development of mandated standards and therefore
are no longer developing standards voluntarily, which
• EN 16798-1 Energy performance of buildings could be fit for future legislation.
- Ventilation for buildings - Part 1: Indoor environ-
mental input parameters for design and assessment Some industry associations have given up their
of energy performance of buildings addressing role in defining state-of-the-art rules and codes of
indoor air quality, thermal environment, lighting good practice. The latter may be difficult for IAQ
and acoustics - Module M1-6. because it would also have to liaise with and rely on
• ISO 17772-1 Energy performance of buildings — expertise from the medical and health sector. A risk
Indoor environmental quality — Part 1: Indoor is that IAQ would be developed separately in dif-
environmental input parameters for the design and ferent Technical Committees, leading to confusion
assessment of energy performance of buildings.
These two equivalent standards provide a classification

of the IEQ and are the reference for IEQ classifica-
tion in buildings. The current EPBD revision proposal
references EN 16798-1 in Annex 1. This means that
the EU Member States must include this information
in their national Energy Performance labelling scheme.
Gaps in legislation
Despite such efforts in standardisation and legislation,
the market still falls short as there remain no minimum
IAQ standards across Europe. As the IAQ industry is
still very dispersed, with few major manufacturers, it
has not yet managed to draw sufficient attention to
IAQ at the political level to promote and push for
inclusion in legislation.
Overall, EU legislation lacks an IAQ focus. The best

guess is that real attention towards this issue in Europe

ARTICLES
and market fragmentation. Due to this complexity, standards, there are different approaches. Within the
ventilation rates and IAQ may likely have to be devel- HVAC sector, there is no common understanding of
oped for specific situations, at home, in the office, what constitutes a good IAQ.
and in schools.
Typically, benchmarks of IAQ still relate to outdoor
air quality. Initially, criteria for IAQ were developed
An issue of market protection
for the elements in the outdoor Ambient Air Quality
The lack of alignment in European standards and Directive sulphur dioxide, nitrogen dioxide and oxides
legislation can also be an offshoot of differing indi- of nitrogen, and particulate matter. IAQ was focused
vidual company strategies. The inconsistency of IAQ on Volatile Organic Compounds (VOCs) emitted from
consideration between residential and non-residential paints, furniture, and other equipment in buildings.
provides a clue as to why. CO₂ and VOC are widely accepted indicators because
they are measurable and used as common reference
In non-residential applications, the disparity in will- points.
ingness to invest in IAQ is insignificant. There is a
large European market, and products are not linked. Yet, it can happen that if ventilation is weak, even if
Generally, companies and engineers across European CO₂ is low, air quality in that building can be bad.
countries have aligned principles and understand what This is logical as CO₂ is used as a tracer for human
constitutes a good solution, with range differences of occupancy. If no or minimal persons are in a room, the
between 10% more or lower air volumes. ventilation system should still run at a level that the
building emissions are be removed. This factor is also
The residential market is more fragmented, with typi- addressed in EN16798-1 and should be considered
cally smaller ventilation markets, small product ranges when designing the ventilation system.
and a very simplified market with a wide variety of
projects. Companies designed to address local needs The perceived air quality is a commonly accepted
may see specific country-wise barriers to protect their approach in Scandinavia and also addressed in
market from a larger European supplier. As a result, EN 16798-1. While it is a good approach, it is complex
some companies can be more in favour of having indi- and faces challenges related to measurement. Let’s
vidual specified markets because it gives an element of not forget that bacteria and viruses are currently not
market protection. included, and any definition related to IAQ would also
be beneficial if this can be considered. (For reference,
However, a harmonised IAQ approach should not we recommend reading the Systemair article on “Why
be considered a threat to commercial interests. Not we ventilate”).
only does this somewhat block innovation, but better
IAQ is essentially technology-agnostic. A harmonised Before any regulation, the industry must work to define
approach that will create better IAQ awareness would what it means to have a good IAQ so the industry has a
allow individual solutions locally marketed to grow common language, interpretation and understanding.
because they will not have to fight to split a smaller Addressing this is the key to unlocking better IAQ,
market share due to rising demands. allowing us to be better prepared for events such as the
COVID-19-pandemic. However, when doing so, we
should also not get stuck in details, which tends to be
The root of the problem: A lack of common
common in Europe. A pragmatic approach is necessary
understanding
to finally push ahead – rather tomorrow than in 10
Despite the progress of standards and guidelines thus years.
far, much remains to be done. The legal background of
EPBD is energy efficiency. Therefore, we still need an To progress, the industry must invest time and
IAQ regulation tackling awareness and information on resources in people’s health first. This would involve
buildings or, perhaps, a building certificate. Hence, the advocating IAQ as a prime objective, meaning it
consumer is better informed or would be empowered would have a priority in line with energy saving and
to choose a facility with better IAQ. decarbonisation of heating and cooling. While this
may be difficult in uncertain times where energy prices
Such a classification would be helpful, but we must are fluctuating, investing now in promoting a better
consider that today, there is not even a commonly understanding and application of IAQ would lead to
accepted definition of IAQ. If you read the different a healthier population for the future.

ARTICLES
Next-Generation Energy
Performance Certificates. What novel
implementation do we need?
LINA PHOEBE-ZOE CHRISTIANA PANAGIOTA KOLTSIOS

SEDUIKYTE MORSINK-GEORGALI PANTELI CHATZIPANAGIOTIDOU STAVROS
[email protected]
DIMOSTHENIS LAURA PAULIUS DARIUS ANDRIUS PARIS

IOANNIDIS STASIULIENĖ SPŪDYS PUPEIKIS JURELIONIS FOKAIDES
a Faculty of Civil Engineering and Architecture, Kaunas University of Technology, Kaunas, Lithuania
b Frederick Research Center, Frederick University, Nicosia, Cyprus
c Cleopa Gmbh, Hennigsdorf, Germany
d Centre for Research and Technology Hellas, Information Technologies Institute, Thessaloniki, Greece
Abstract: This study performed under the H2020 project “Next-generation Dynamic Digital
EPCs for Enhanced Quality and User Awareness (D^2EPC)”, aims to analyse the quality and
weaknesses of the current EPC schemes and aspires to identify the technical challenges
that currently exist, setting the grounds for the next generation dynamic EPCs.
Keywords: EPC, SRI, LCA, BIM, DT, GIS, human comfort, D^2EPC.
The novelty of the dynamic EPC

the introduction of novel aspects in the certification
The aforementioned shortcomings of national EPC process and the simplification thereof, the strength-
schemes urge the development of a holistic framework ening of its user-friendliness, as well as the conformity
that will strengthen and improve the quality and appli- with national and European legislations, can be accom-
cation of EPCs. The former can be achieved with the plished using a standard collection of indicators based
introduction of novel and cost-effective approaches for on a specific methodology. All upgrade needs of EPCs
assessing the energy performance of building envelopes can be met by choosing acceptable output indicators
and systems. According to the collected information, supported their automated estimation.

ARTICLES
New indicators
Human comfort. People in developed countries spend
The introduction of novel aspects into the energy per- more than 90% of their time in closed environments
formance certification process in our project includes - buildings and transport [2]. Air quality indoors
three indicators – the smart-readiness level of the is 2-5 times lower than outdoors [3]. These values
buildings (SRI), human comfort-related indicators, can be even lower if we consider the future effects
and environmental aspects (LCA). of climate change (extreme temperatures, heat waves,
heavy rainfalls, air pollution). Therefore, significant
Smart readiness indicator. The scheme for rating attention should be paid from researchers, businesses,
the smartness of buildings was presented in 2018 in and standardization organizations to the field of indoor
a revision of the Energy Performance of Buildings environment quality (IEQ) [4-6].
Directive (EPBD). It was established that the smart
readiness of buildings should be optionally evaluated The main indicators that assess the IEQ of a building
by the smart readiness indicator (SRI) [1]. According to and human comfort/wellbeing can be described by an
the EPBD, this indicator reflects the building’s ability to integrated multi-comfort concept that includes indoor
adapt to the needs of its occupants and outdoor energy air comfort/quality, thermal, visual, and acoustic
infrastructure, improving its overall energy performance. comfort. The indoor air quality (IAQ) examines how
fresh the air is in a building and the concentration in
To define the smartness of buildings’ services, three the air of certain pollutants (e.g. CO₂, VOC). Thermal
main functionalities of smart readiness are introduced comfort provides a state of satisfaction with the existing
in the methodology: thermal environment. Visual comfort ensures that the
luminance levels are within acceptable levels. Acoustic
• The ability of the building to adapt its energy comfort creates a comfortable acoustic environment
consumption based on the needs in an energy- without uncomfortable noise or vibrations.
efficient way;
• The ability of the building to adapt its operation to Within the H2020 D^2EPC project, human comfort /
occupant’s needs; wellbeing parameters will be measured and used in
• The building’s flexibility to its overall electricity the calculation engine allowing dynamic input for
demand, as well as its ability to participate in dynamic EPCs.
demand-response, in relation to the grid.
Life cycle assessment. LCA indicators such as “energy
Within the H2020 D^2EPC project, both SRI and savings”, expressed in “embodied energy/m²” and “carbon
EPC methodologies will be included in the same calcu- reductions”, expressed in “carbon dioxide equivalent/m²”,
lation engine allowing, where possible, to merge these will be included in the dynamic EPCs calculation engine.
two methodologies to progress the SRI to a higher level. This will provide to the building design team the option
HORIZON 2020 PROJECT
DYNAMIC
DIGITAL
ENERGY
PERFORMANCE
CERTIFICATES
Next-generation Dynamic Digital EPCs for Enhanced Quality and User Awareness

ARTICLES
to improve and optimize the environmental performance Figure 1 represents the definition of BIM and DT
of the building, based on changes to be integrated at the concept regarding the energy efficiency of the building
initial design stages of the building. throughout its life cycle stages: plan and design >
produce and construct > use and maintain.
In the D^2EPC project, the LCA Indicators for EPCs
will allow maximizing energy saving and carbon reduc- It is evident that BIM and DT overlap in the construc-
tion of the buildings, introducing this way the aspect tion phase since BIM can deliver object-based data
of building’s sustainability as part of the EPC issuance utilized for the DT.
process. This could speed up the transaction into NZEBs
as well as control the building’s energy demand, reduce Introduction of GIS in EPC
carbon emissions, and enhance public awareness. In the D^2EPC project and GIS context build-
ings are described and considered in the concept of
The D^2EPC project will propose additional indi- BuildingsExtended3D, i.e. with correct geometric
cators, which will demonstrate the environmental dimensions, proportions, scale, but not considering
performance of buildings, for their introduction in geolocation of a particular building. The use of geo-
the next-generation EPCs. spatial technologies and accurate data location could
improve the processes related to the data needed to assess
The Introduction of BIM and Digital Twin the energy performance and needs of buildings and
Concepts for the Next-Generation EPCs urban areas. In addition, the use of geolocation practices
The use of BIM technology helps to improve the col- can increase decision-making effectiveness by different
laboration of stakeholders from the design to asset stakeholders (policymakers, technicians, citizens).
maintenance phases. While BIM delivers static data,
Digital Twin (DT) focuses on linking physical objects Introduction of financial schemes
to their respective digital replicas using periodically Introducing financial schemes in EPC is suggested in
updated (dynamic) data flow. The key features of DT this study. Based on the well-established principle of
are sensing and monitoring, data linkage, Internet of lifecycle costing, a set of financial indicators could be
Things (IoT) implementation, simulation, predictions, developed to allow the individual elements of buildings’
and controls. energy efficiency to be interpreted into standardized
De�inition of BIM and Digital Twin concept regarding Energy Performance

Explore / Plan / Design / Engineer Produce / Construct / Assemble Use / Maintain / Exploit
Lighting PIR
Virtual
Semi-physical
Asset Layer
Physical
Representation
representation representation
BIM data Sustainability Climate Smart
requirements database Survey Sensors As built BMS IoT LiDAR metering
data data data data data data data data data data data data
Analysis/change layer
Remake Remake Remake
Analyse > Predict > Decide Analyse > Predict > Decide Analyse > Predict > Decide
Energy Performance Digital Dynamic Energy

Energy Performance
Simulation Performace Simulation and
Certi�ication
Certi�ication
BIM Overlap zone

Digital Twin
Digital data layer
PIM –
Project Information Model AIM –
Asset Information Model
Figure 1. Definition of BIM and DT concept regarding energy performance.

ARTICLES
D^2EPC System Architecture

numerical values. The delivery of such indicators could
allow the use of EPCs for the financial evaluation of A novel methodology for dynamic EPC is being
energy upgrading measures for buildings. For example, developed within the H2020 D^2EPC project, which
financial awards (e.g. tax reliefs) should be included introduces the aspects of SRI, occupant comfort,
if the building owner exceeds new EPC requirements LCA, integration with DT, and GIS systems. Key
and class. In the opposite case – penalties should be functionalities of D^2EPC architecture are presented
imposed based on the “polluter pays” principle. in Figure 2.
Infrastructure / Sensors / Devices / WEATHER API

BMS / SCADA / ...
Physical layer Actuators
Interoperability Informa�on Management Layer

layer Applica�on
Common
IoT Interfaces Informa�on
Layer
Interfaces
Service / BIM-Based Digital Twin

Processing Layer Building
BIM Parser Behaviour
Profiling
Calcula�on Engine
Building
Opera�onal
Asset Ra�ng Performance
Ra�ng
Module
Roadmapping Tool Energy Performance

Added Value Service Suite for dEPCs
Verifica�on & Credibility
Applica�ons Toolkit
RE Asset
Extended dEPCs
DSS Data
Management Network
Quality
Monitoring
Tool
AI-Driven Performance
Forecasts Building Energy
Adapt.
Perf. Performance Benchmarking
Lightweight
Forecast
Training
Performance Alerts &

No�fica�ons
D^2EPC Repository
BIM, Data, KPIs
Representa�on
Layer
D^2EPC Web GIS
D^2EPC Web Credibility and
Pla�orm Verifica�on UI
Mobile App
Third Party Pla�orms

& Tools
D^2EPC End Users
Figure 2. D^2EPC System Architecture. [13]

ARTICLES
D^2EPC framework consists of four layers:
• Infrastructure/Physical Layer,
• Interoperability Layer,
• Service/Processing Layer, and
• Representation Layer [7].
Conclusions
New technologies that didn’t existed at the time when
the current EPCs schemes were developed, enable
new approaches towards building energy certifica-
tion. D^2EPC platform aims to integrate IoT, AI,
and other novel technologies to enhance end-user
awareness and facilitate a more sustainable life cycle of
buildings. Nevertheless, integrating these technologies
into a coherent unified tool is still a challenging task.
D^2EPC aims to provide a demonstrator platform
that will help increase the understanding of European
building stock’s EPCs.
References
[1] “BPIE of the European Parliament and of the Council of
30 May 2018 on the energy performance of buildings
(recast)”.
[2] U.S. Environmental Protection Agency. 1989. Report to

Congress on indoor air quality: Volume 2. EPA/400/1-
9/001C. Washington, DC.
[3] U.S. Environmental Protection Agency. 1987. The total

exposure assessment methodology (TEAM) study:
Summary and analysis. EPA/600/6-87/002a. Washington,
DC.
[4] ASHRAE, “Indoor Air Quality Guide Best Practices for

Design, Construction, and Commissioning, Guide”, 2012.
[5] ASRHAE, “Position paper on IEQ”, 2015.
[6] American Society on Heating, Refrigeration and Air-

Conditioning Engineers, “Standard 55. (2004) Thermal
Environmental Conditions for Human Occupancy”, 2004.
[7] Koltsios S., Tsolakis A.C., Fokaides P.A., Katsifaraki

A., Cebrat G., Jurelionis A., Contopoulos C.,
Chatzipanagiotidou P., Malavazos C., Ioannidis D.,
Tzovaras D. D²EPC: Next Generation Digital and Dynamic
Energy Performance Certificates. 6th International
Conference on Smart and Sustainable Technologies, Bol Acknowledgement
and Split, Croatia, 2021.
This work is part of the D^2EPC project that has received
funding from the European Union’s Horizon 2020
Full article: https://proceedings.open.tudelft.nl/clima2022/ research and innovation program under grant agreement
article/view/348 No. 892984.

reTHINK living
How will we live tomorrow? ARTICLES
Healthy Homes Design Competition:

“reTHINK living”
CAROLINE REICH AMELIE REISER
master student, energy master student, energy
efficiency design (e2d), efficiency design (e2d),
University of Applied University of Applied
Sciences, Augsburg, Germany Sciences, Augsburg, Germany
[email protected]
REHVA announced the Healthy Homes Design Competition 2022 with the aim to encourage
students in different building disciplines to design a building with increased comfort quality
while also tackling the problems of climate and demographic change. The team “re²THINK
tank” convinced the international jury with both their space-efficient and user-focused
proposal as well as their clean technical approach. The two master students from Augsburg,
Germany were able to achieve the third prize.
Keywords: healthy homes, demographics, wellbeing, housing shortage, user-centric design,
space-efficiency, communal space, synergies, comfort, daylight simulation, PV generation
T
he submission of the project “reTHINK living” alarmingly high housing shortage in the Netherlands.
depicts a radical answer to the question of how Therefore, the huge gap between the number of needed
we will live together. accommodation and number of affordable and avail-
able ones is calling for new solutions regarding the
Before the implementation of any design idea, a complex question of how we will live tomorrow.
healthy home was defined as a way of life that con-
siders all kinds of wellbeing: social, physical, mental
The concept of “reTHINK living”
and eudaemonic health, hedonic and subjective well-
being, productivity, as well as environmental quality The concept of reTHINK living questions the given
and comfort. This groundwork formed the basis of a apartment sizes and suggests a solution that is remarkably
user-centric design of this building complex in Pernis. more space-efficient and results in a smaller per capita
living space than in any form of conventional housing.
An analysis of the demographics of the Netherlands,
of Rotterdam and of Pernis in general gave an idea of The left side of Figure 1 depicts a conventional
the kind of future residents that could be expected. building organisation. The coloured area stands for
The fact that a household has to wait 5.5 years for an the communal area inside every flat, usually the living
affordable flat on average is just one out of many issues room. The concept of “reTHINK living”, as seen on
that occur because of the increasing problem of the the right side of Figure 1, removes this space and
LESS = MORE
private communal space, communal space to connect,
separated in each apartment, designated for different functions
disconnected from each other and for everyone to use
Figure 1. Conventional living becomes reTHINK living.

ARTICLES
shares it with all the other residents of the building. The organisation of the different kinds of common
The result of less private space and a reduced flat areas is depicted in Figure 2. It depends on the impor-
size is compensated by a maximised common area tance of the room for the household. Ultimately, this
that can be used by every resident in a specific and concept allows to build synergies among the residents
meaningful way. throughout the building complex and strengthens the
neighbourhood.
Synergies between the future residents
Urban design principles
Coming from the demographics and the social infra-
structure of Pernis, four different household types The location of the competition at the Pernis water-
have been created: families, friends + couples, seniors, front presented an exciting but challenging site. The
and singles. They all differ in their needs and desires, weaknesses include constant emissions from the nearby
particularly when it comes to the kinds of communal motorway in the west, a continuously present noise
areas that need to be close to them, be it a communal level coming from mainly the west as well, the smell
room for “sports + health” or “silence + rest”, a room of Diesel that is only partly shielded from the existing
for childcare or some shared office spaces. trees and lastly a poor social and functional mix.
youth’s space
let’s jam
sports + health
FAMILIES FRIENDS + COUPLES
silence + rest
child care
home cinema
shared office spaces
common room chat + rest
meet + connect
atelier
SENIORS
sports + health
SINGLES
Figure 2. The four household types and the synergies among them.
TOP VIEW NORTH-EASTERN VIEW SOUTHERN VIEW
Figure 3. Perspective views.

ARTICLES
An urban analysis led to the following three urban design heat island effect. Small parts of the ground floor area
principles that define the cubature of the building: are used for the technical equipment of the building.
1. Blocking the west side of the site to prevent the

Cubature
existing smell and noise.
2. Opening the complex to the east and south and In order to improve the incidence of daylight, the southern
alongside the solar path. facade is tilted by 70° and the eastern and western facades
3. Strengthening the sightlines to Pernis as well as facing the inner courtyard are tilted by 80°.
those to Rotterdam downtown.
The roof is tilted as well. There are five floors on the
Building organisation and flood concept western side and four floors on the eastern side of the
building to strengthen the urban principles.
The building complex is vertically organised, which
means the privacy increases with the floors of the
Access and building organisation
building. The ground floor consists of a communal area
for everyone to use, including the non-residents. The The access to the flats and to the common areas is via
previously mentioned communal rooms are located on a circumferential and completely glazed arcade and two
the first floor. The apartments can be found on the staircases, each of them contains a lift. This makes the
second to fourth floor. The reason why there are no building completely barrier-free which marks the simplest
rooms on the ground floor is the flooding concept of way of inclusion. The plans in Figure 5 also illustrate that
the building. As a maximum flooding of the site of one each flat gets light from at least two directions, so there is
meter had to be considered (a number that is estimated no apartment with an insufficient exposure. In addition
to rise as a result of the climate crisis), the building was to that every flat has at least one loggia which allows each
elevated by one floor level. This way, the ventilation of resident to enjoy their own private outdoor space. This
the courtyard can be increased, which counteracts the creates a connection to the outside space and nature.
block the west side of the site to open the complex to the east and strengthen the sightlines to Pernis
prevent smell and noise south and alongside the solar path as well as to Rotterdam downtown
Figure 4. Three urban design principles.
SENIORS
daylight factor of 2 %
in 71 % of the floor surface
35 m² inside
7 m² loggia
FAMILIES
daylight factor of 2 %
in 66 % of the floor surface
110 m² inside
12 m² loggiasv
Figure 5. Detailed floor plans of two out of the four household types (seniors and families).

ARTICLES
Detailed floor plans

brings with it a high storage capacity as well as a high
The differences of the four household types result in level of comfort due to a uniform heat radiation.
different floor plans. Two out of the four kinds of floor
plans are depicted below. The choice of ventilation concept is a hybrid ventila-
tion. This allows the building to be naturally ventilated
The seniors’ floor plans are primarily located close to in summer, in our case by cross-ventilation. In winter,
staircases and elevators to shorten the access routes the mechanical ventilation system with heat recovery is
within the building. Each of their floor plans are used to ventilate the building sufficiently while only a
designed to be extra-spacious to allow an independent negligible amount of thermal energy is released into the
life regardless of age or inability. Since the relative need outside air. The choice of a hybrid ventilation system
for daylight increases with age, the south-facing window that draws in air through an air tower in the south
areas are relatively larger than in other floor plans. with a fine filter also gives the chance to optimise the
indoor air quality. Thus, the residents can decide for
The families enjoy a spacious entrance situation with themselves whether to ventilate via natural ventilation
the living room as the heart of the apartment refer- by means of cross-ventilation, which is made possible
ring to both the location and communication inside by the arcade in which the upper part of the glazing
the apartment. Several loggias designated to different can be opened at any time, or whether to use the
rooms maximise private outdoor areas. mechanical ventilation system to ventilate the room.
The common features of the floor plans are that all cir-
Winter and summer comfort policy
culation areas have been reduced to a minimum to keep
the usable space as generous as possible. In addition, To improve the thermal comfort in winter on one
the bathroom is always located next to the entrance hand, the building is heated by a water-water heat
area facing the north in order to maximise the southern pump. An underfloor heating system using concrete
exposition of living and sleeping rooms in general. core activation serves as a radiator which allows the
heat to be kept in the building as efficiently as possible.
Hybrid ventilation, which ventilates the building
Heating / cooling / ventilation concept
mainly in winter, releases as little thermal energy as
Passive cooling and heating of the building is provided possible to the outside air, improving thermal comfort
by a water-water heat pump, which uses the heating in the building in winter.
and cooling energy from the river water. The heat
pump delivers the correspondingly cooled or heated To avoid overheating in summer on the other hand,
medium to a hot water tank and a heating water buffer the external blinds, which are attached to each exterior
tank. The hot water is then transferred from the hot window facing south and the inner courtyard, were
water storage tank to all water points in the house. controlled with a sun sensor. In addition, it is possible
The cooled or heated medium is passed on from the to override the control of the external blinds by means
heating water buffer tank to the underfloor heating of the sun sensor, so that the user can also control the
with concrete core activation. Concrete core activation external blinds individually. The building is equipped with
cross-ven�la�on
individually controllable
external vene�an blinds
underfloor hea�ng with
concrete core ac�va�on
hwbt = hea�ng water buffer tank

hwbt hwbt
hwt hwt = hot water tank
hp rat hwt hp
hp = heat pump
outside air intake
via air tower in the
sip well sip well south with fine filter
suc�on well suc�on well
Figure 6. Heating / cooling / ventilation: summer (left) and winter (right) situation.

ARTICLES
a water-water heat pump, which extracts the heat or cold the building, it is stored in an electricity storage unit
from the adjacent river, so that overheating of the building for use at a later time. If there is still surplus electricity,
can be avoided by cooling through the underfloor heating it is fed into the power grid.
with concrete core activation. The river has an average
temperature of 18°C in the warm months, which makes
Overall lighting concept
it ideal for cooling the building with a water-water heat
pump and thus optimising comfort in the summer. Floor-to-ceiling windows maximise the visual axes to
the south and the inner courtyard. The view to the
north and thus towards the water is also improved by a
Integrated water systems
completely glazed arcade which means that depending
In terms of water use, all the rainwater from the roof of on preferences the occupants can enjoy the view towards
the building is collected in a rainwater storage tank. The the water or towards nature and neighbourhood.
water is then filtered and transported to the toilets or
washing machines via domestic waterworks or a pump,
Strategy to minimize overall energy use
so that the rainwater can be used for these purposes.
In order to minimize overall energy use of the concept,
passive cooling by using a water-water heat pump was
Energy concept and on-site energy use
applied on the one hand, and the 3E-strategy was
Regarding to the on-site energy use concept, a 370 m² purposed on the other hand:
photovoltaic system will be installed on the roof of the
building. This photovoltaic system generates a total 1. Sufficiency = energy saving
of roughly 78 430 kWh/a, so that theoretically the → less floor space by removing and reducing the
entire electricity demand of 61 914 kWh/a (including common area from the flats
the electricity demand to supply the heat pump) can 2. Efficiency = energy efficiency
be covered by the photovoltaic system through the → use of a heat pump, a hybrid ventilation system
integration of an electricity storage unit – this creates and clever energy management throughout the
a plus-energy building. building
3. Consistency = renewable energy
The electricity generated by the photovoltaic system → availability of the required energy as efficiently as
is first used to cover the current electricity demand in possible. In this case, all the energy required on
the building (washing machine, PCs, TVs, kitchen, the site is produced on the roof of the building
e-mobile charging stations). If the photovoltaic and partly stored in an electricity storage system
system generates more electricity than is needed in to be available for use at any time.
power grid
hot water house waterworks

tank —> pump
heat pump heat pump
rainwater storage
sip well tank with filter electricity storage
suc�on well
Figure 7. Integrated water systems and energy concept.

ARTICLES
Tiny Homes – A Tiny solution to a big problem

LAURA DENOYELLE
Ir. Arch.
[email protected]
onetinystep.eu
Master thesis: A Tiny solution to a big problem. How Tiny Homes
could benefit you, society and how they could play their part in
climate change communication.
The Healthy Homes design competition organized by CLIMA 2022 asked for a design of an
innovative apartment building. They asked big, but I decided to hand in a proposal with Tiny
Homes of 20-40 m² instead.
Why tiny homes?

Humanity has a big problem on their hands: climate
change. It is affecting everyone, and it is everyone’s
responsibility to do their part. Whether that is eating
less meat, revolutionizing solar power, or looking at
sustainable industry from the perspective of ... a Tiny
Home. Exactly what I looked at in my master thesis.
But why choose to look at Tiny Homes? Put simply:

I was living in one myself. Considering the struggle,
I encountered to live there legally, my original motivation
was to legalize my own house using my thesis research.
Somewhere along the road the scope changed drastically,
and the research focused more and more on how Tiny
Homes can help the transitioning to a sustainable society The Tiny Home Design created in master thesis
and what role they could help play in climate change ‘A tiny solution to a big problem’.
communication, preparedness, and mitigation.
The thesis explored many fields in relation to Tiny design point; it is a reality we face today, and will face
Homes, through the methodology of research by more frequently in the future: Rotterdam has depleted
design. The practical aspect of the design helped keep its land use within protected dikes. In America certain
all elements connected, and focus on a design that is zones are exempt from insurance as the risks for fire
feasible to build today. As we need these homes … hazard due to a changing climate are deemed too high.
well, now!
Luckily most of the time we are warned for these disas-
While I did not win the CLIMA 2022 competition, ters – in this case a Tiny Home proves to be an elegant
I did impress the jury with my proposal, following is solution: In case of disaster, you can drive it away and
how. wait out the storm at a safe location. A Tiny Home
does make you a bit slower and we have seen with the
floodings in Belgium and Germany that a warning is
Climate preparedness
not always present. But even if you need to leave your
The brief of the competition asked for a very specific home behind in a rush, the financial repercussions
building: An apartment building that will be unaffected and options are less severe than they would be for a
by potential flooding. This is not just an interesting conventional home.

ARTICLES
Carbon footprint (= climate mitigation)

location is near an industrial site. Having been raised
Preparedness is not the only way Tiny Homes relate in Hoboken in Belgium I have experienced first-hand
to climate change. The Tiny Home designed has a what that means. An industry scandal 20 years ago
carbon footprint of 15 000 kg CO₂eq. As the thesis showed extreme lead values damaging children’s growth
focused on Tiny Homes in a suburb/rural environ- and health. Two years back a documentary proved the
ment the comparison was made to the typical house protection measures taken were actually insufficient.
a two-person household will acquire in this environ- This is not an insulated event. It has been shown on
ment. The LCA for this home showed that the average multiple occasions that effects of nearby industrial
house of a two-person household in Danish suburbs is activity are known too late, and mostly affect children.
150 000 kg CO₂eq. The Tiny Home footprint is no Thus, an argument is made that the site itself is not
less than ten times smaller! suitable for families. Adults however are safe, and the
perfect demographic for smaller homes, or Tiny ones.
A demographic mismatch
Zooming in to the level of a building, or a Tiny Home,
Thus, Tiny Homes are sustainable housing when it there are two elements related to Health. The most
comes to climate change, but who will live in them? commonly addressed is the building’s influence on
Looking at the demographics of Denmark – and they the physical health. On the other hand, following the
are quite similar in most of Europe - shows that 49% COVID-19 pandemic, a lot of research has been done
of people live alone or as a couple, with the average on buildings’ social and mental health.
household being 2.1 people. If we look at the residen-
tial buildings being made. However, we see that about
Buildings’ impact on physical health
80% of these are aimed at families. Those numbers do
not add up on paper. And in reality, they affect the Tiny Homes provide unique challenges when it comes
lives of a whole generation, one that cannot afford to to Indoor environmental quality, the extremely low
buy a home anymore. Tiny Homes are an affordable volume creates higher pollution levels. Thus, for this
option for this target market existing of 50% of the article we will focus on air quality. Two factors were
population. considered: materials and ventilation. Volatile organic
compounds (VOC’s) are so present that, even when
we bring clean air into a room, by the time it reaches
Healthy homes
the occupant it might be polluted again. This is why
There was a more important reason however, that the first step for the thesis was to design a Tiny Home
I chose to submit Tiny Homes into the competition. without VOC’s. Again, here I was motivated by own
One that has everything to do with the key word of experience of a period with extreme asthma at univer-
its title: Healthy Homes. The proposed competition sity. In my Tiny Home that is build VOC free, I had
A carbon footprint comparison, buying the traditional Danish home versus a Tiny Home.

ARTICLES
no health problems. But I often had to leave classes also regulates humidity differently than traditional
early at university, however, as I had no air, despite systems. And a third obstacle was the size of the Tiny
the ventilation. Home. For example, infiltration is calculated using
exposed surface/volume ratio, and the size of a Tiny
To address the ventilation, bringing in a range of Home completely skews those results. I did manage to
pipes in a Tiny Home is not really feasible, space is in speak to some of the PhD-researchers on this topic and
short demand in tiny houses after all. After thorough got some help to correctly set up a simulation model.
investigation, I ended up working with decentral-
ized ventilation for the design. This has been widely An argument against decentralized heating I often
used in Tiny Homes, specifically the lunos e2. It was hear is the noise. It is also a main argument I’ve heard
even analysed in a research paper studying two Tiny against centralized ventilation. But in all the research
Homes. One of these homes had perfect air quality I did I have never seen this argument pop up in relation
and humidity control. The second didn’t have the same to Tiny Homes. I think an important difference is the
good results, there several cats were also living in the owners’ involvement in building the tiny home, and
space though and the owners manually choose to use their awareness of how the home functions. This makes
the lowest mode of the equipment. people more tolerant towards discomfort. Secondly
as the Tiny Homes are in such a small volume that
Simulating the ventilation in my own design proved to serves all the functions of a home, there might be more
be a challenge. I used IDA ICE and decentralized ven- obstacles between the actual sleeping resident and the
tilation is just not included yet. Decentralized heating ventilation to limit the sound discomfort.
IDA ICE simulation summary.

ARTICLES
Buildings’ impact on mental health

of designed outdoor space and biophilic design was
A subject usually not addressed in the building also never mentioned in Tiny Homes. Most of these
industry: what influence does a building have on our end up in the countryside however, where nature and
mental health? A number has been floating around outdoor space are plenty.
for a while now: we spend 90% of our time indoors.
The lockdown probably made this number rise to a
Back to the original question
solid 99% for many people. Different papers were
released following lockdown, investigating tangible Right now, Tiny Homes are largely illegal, or at least
parameters that can improve our home designs. Tiny not entirely legal. European law however allows a
Homes faced similar challenges to lockdown apart- country to implement a different building code for
ments. How do you use a small space as a bedroom, buildings under 50 m². Thus, what if we consider
office, bathroom, kitchen. The thesis summarized the Tiny Homes as a housing typology? They serve 50%
research done on these subjects and came to a sur- of our society and have a carbon footprint that is
prising conclusion: the design principles suggested for 10 times lower. They are also fast to build, and cheap
Tiny Homes and mental health are largely the same. compared to large projects. The estimate for the
entire building site development of the CLIMA 2022
Guidelines emphasize a flexible use of space. One competition with Tiny Homes was 2.7 million euros
room should be able to have multiple functions such compared to the 7.35 million euro’s development
as kitchen, workspace etc. Flexible furniture can play that the asked for apartment building would cost.
a big role in this. What was different from Tiny Home Last but not least they have been pioneers in off grid
living that usually features open lifts was the emphasis living, and they are the perfect field for optimizing
on a closed bedrooms for biorhythm. The importance new technologies.
Design guidelines for social/mental health of a home.

REHVA WORLD
Meet the BIM-SPEED

Competition Winners & Finalists
JASPER VERMAUT
Policy & Project Officer
at REHVA
A s part of the BIM-SPEED project¹, REHVA co-organised

the EU BIM for Building Renovation Competition
together with other Brussels-based EU associations
The Competition took place from June 2021 until April 2022
and aimed to engage professionals and students active in
the design and construction industry to present a residential
Architect Council of Europe (ACE), European Builders building renovation project that applies the BIM tools and
Confederation (EBC) and the European Construction methods developed by the BIM-SPEED consortium². The
Industry Federation (FIEC). BIM-SPEED with technical challenge was to demonstrate a renovation project,
support from other partners in the consortium such as using BIM-SPEED platform for collaboration, in a way that
the Technical University of Berlin and Erasmus University allows energy saving for the occupants, improves their
Rotterdam. comfort while reducing the time and the cost of the
overall process. The participants had to do this by making
use of one or more of the tools developed within the BIM-
SPEED project and made available through the project’s
platform³. After submission the results were assessed by
a jury of professionals who have a long experience with
BIM from different perspectives:
•• András Rónai: Mechanical Engineer M.Sc.; HVAC+R

and BIM – Óbuda Group – MMK.
•• Chiara Dipasquale: Expert in Innovation and
Sustainability – Volksbank.
•• Olga Venetsianou: Architect PhD, MA in Digital Arts
ASFA - Representative from the Technical Chamber
of Greece to the Architects’ Council of Europe BIM
Working Group.
On the following pages we’re delighted to present you

the winners and finalists of the competition. REHVA
interviewed both teams to give the opportunity to present
themselves, their experiences with the competition and
the BIM-SPEED platform, and what future they see for
the digitalization of the construction sector.
¹ https://www.bim-speed.eu/en ² https://www.bim-speed.eu/en/competition ³ https://www.bim-speed.eu/en/training-materials

REHVA WORLD
Interview with the Competition Winners: Team ENSTP from the

National Advanced School of Public Works (Yaounde, Cameroon)
Idriss Tchaheu Tchaheu

•• Works at Consultation et Project D’Afrique (CPA) in charge of
digitalization of the construction of French-speaking Saharan Africa;
•• Founding member of the the non-profit association BIMUP AFRICA

which aims to create awareness on BIM to students and professionals
in French-speaking Sub-Saharan Africa.
•• Master’s Degree in Civil Engineering at National Advanced School of

Public Works (NASPW) in Yaounde (Cameroon) in partnership with the
University of Padova (Italy) with the thesis topic “BIM applied to the
structural assessment of a multi-story reinforced concrete building;
•• Team leader of ENSTP BIM team who the second prize of the student
BIM competition by BIM Harambee Africa in 2021;
Charlène Delavictoire Sobgoum Jiogo

•• Student in Fifth year of Architecture at National Advanced School of
Public Works (NASPW), Yaounde (Cameroon) in partnership with the
University of Padova (Italy).
•• Founding member of the non-profit association BIMUP AFRICA which

aims to create awareness on BIM to students and professionals in
French-speaking Sub-Saharan Africa;
•• Member of the ENSTP BIM team who won the second prize of the
student BIM competition launched by BIM Harambee Africa in 2021;
•• Team leader of BIM ENSTP team who won the EU BIM FOR BUILDING
RENOVATION COMPETITION launched by BIM SPEED 2022.
Congratulations on winning the EU BIM for In 2021, we were members of the team that won the
Building Renovation Competition! Tell us more second prize in the student BIM competition launched by
about your team, specialization and career
BIMHarambee Africa. We are also members of the team that
The ENSTP BIM Team that participated in the EU BIM created in 2022 the non-profit association BIMUP AFRICA
SPEED 2022 competition is a team composed of Idriss in order to raise awareness about BIM among students
Tchaheu Tchaheu and Charlène Delavictoire Sobgoum and professionals in French-speaking sub-Saharan Africa.
Jiogo from the National Advanced School of Public Works
(NASPW) of Yaoundé (Cameroon) in partnership with Can you briefly describe the project that you
the University of Padova in Italy. We are passionate worked for the Competition and how you used
the BIM-SPEED tools available?
about digital technologies that are revolutionizing the
construction industry and helping to improve productivity. The project submitted to our study is located in France in the
We are determined to contribute to the technological municipality Massy precisely at 15 Avenue de la Republique,
transition process of the construction sector in Africa. the building is a basement, R + 10 storey for residential
purpose. The objective of the study was to renovate the
Idriss is a civil engineer graduated from (NASPW); he building while using the BIM Speeds tools integrated to the
works at the consulting firm Consultation et Projects collaborative platform kroqi. The BIM Speed tools used in our
D’Afrique (CPA) which operates for the digitization of project are the Mereen Weather Service and the file naming
the construction sector in Saharan Francophone Africa. convention service which was respectively used to collect
Charlène Delavictoire SOBGOUM JIOGO is a student in historical climate data and to define naming convention
fifth year of architecture at (NASPW). standard across project files to ensure standardization.

REHVA WORLD
How does the BIM-SPEED platform contribute The Mereen Weather Service that is available through the
to this? Can you tell us more about your platform allowed us to collect free of charge historical climate
experiences with the platform?
data from 1999 to 2021 in EPW format. This collected climate
The platform allowed us to collaborate smoothly data was used in Graitec Archiwizard software to simulate
and efficiently throughout our project by deploying a the daylighting analysis as well as the energy analysis through
common data environment and workspace, allowing us which we managed to get the cost savings.
to work in a hierarchical manner. In addition, the naming
convention defined for the files allowed us to discern Another available tool in the platform allowed us
with precision the discipline, phase and version of the to increase time efficiency, which is the File Naming
software used for the model. All the data was stored in a Convention Service that uses an automate workflow which
hierarchical and easily accessible way inside BIM-SPEED consisted on creating; managing and applying a naming
platform that is available through Kroqi. standard across project files to ensure standardization.
Interview with finalists of the BIM-SPEED Competition: Team

from the Federal University of São Carlos (São Paulo, Brazil)
Clélia Mendonça de Moraes

•• Postdoctoral Researcher at the Federal University of São Carlos
(2018 – ongoing), Postdoctoral Researcher at Institute for
Technological Research (IPT) (2013 – 2018);
•• Ph.D. Degree in Mechanical Engineering (UNICAMP), thesis on

“Thermal comfort in the classroom in Brazil: experimental analysis
and numeric” in collaboration with the Indoor Environment &
Energy Department in DTU;
•• Architect and Urban Planner for the Department of Transportation

and Urban Mobility in the Araraquara Prefecture;
•• Publications: https://bv.fapesp.br/pt/pesquisador/54287/
clelia-mendonca-de-moraes & https://lattes.cnpq.
br/4970119616148506
Everson de Castro Rodrigues

•• Graduate Civil Engineering at Faculdade Estácio de Belém (2020);
•• Graduate student BIM Manager at Unyleya (2022);
•• Designer in architecture, structures, hydraulics and electrical;
•• Publications at scientific national events: https://www.linkedin.

com/in/engcivileverson/
Anderson André Lima de Souza

•• Graduate Civil Engineering at Faculdade Estácio de Belém (2020);
•• Graduate student BIM Manager at Unyleya (2022);
•• Designer in architecture, structures, hydraulics and electrical;
•• Publications at scientific national events: https://www.linkedin.

com/in/engcivileverson/

REHVA WORLD
Congratulations on winning the EU BIM for different parameters in our model for the sustainable
Building Renovation Competition! Tell us more use of materials for different element in the buildings,
about your team, specialization and career
for example of the roof.
Clélia: I’m a researcher at the Federal University of São
Carlos in Brazil at the department of Urban Engineering and Clélia: As last we’ve also made use of the GIS Data
work as an architect on smart buildings and urban mobility Provider to do a study of the terrain while using active
for the municipality of Araraquara, in São Paulo state. My and passive technologies to make the building an active
specialisation is in sustainable design, environmental and generator which contributes to the local grid.
thermal comfort, BIM management and the integration of
renewable energy in buildings, airplane and cities. I was From your perspective, what do you see as the
also part of the scientific committee of the CLIMA: REHVA main opportunities & challenges for the uptake
of BIM within the building Industry in Brazil?
HVAC World Congress in 2013 and 2019.
Anderson: It can be challenging in Brazil to work with as
Everson: I’ve graduated in civil engineering, during it takes time to get to know new software and platforms
college I started working with BIM and specialised in it that are related to BIM. Many platforms in Brazil that
further since my graduation in 2020. As a BIM manager work with BIM can only be used in quite isolated cases,
I’ve contributed to research articles related to civil while the BIM-SPEED platform is more universal, which
engineering and how BIM can be used for different types is a key element for the uptake of BIM in our country.
of buildings.
Everson: The platform and the available tools don’t just
Anderson: Similar to Everson, I’ve graduated from civil focus on one element but consider the full life-cycle of
engineering and working on BIM management but with a building, from design to the post-construction phase,
a focus on industrial design and concrete structures. which makes it very useful to analyse the full impact of
a building.
Can you tell us more about the project that
you worked on that was submitted in the Clélia: The BIM-Speed platform is universal and allows
competition and your experiences with the BIM-
specialists from different fields to work together. For me
SPEED platform?
it’s not just important to look at individual buildings but
Clélia: Our project was a residential building for a family also look at city-level, which is what I focus on for my work
in a sustainable urban neighbourhood in Araraquara, São for the local government in Araraquara (São Paulo). By
Paulo state. The aim of the project was to lower energy looking at the wider level we can see the wider impact
consumption and costs while improving the IAQ and that buildings have on the environment around them,
comfort levels of the building. The BIM-SPEED platform which is made a lot easier through digital technologies.
made it a lot easier for us to work with each other on this We look which neighbourhoods are less energy efficient
project in the same digital workspace, even if we live in and how we can better tackle this at a district level. This
different regions (Clélia lives in São Paulo while Everson is why I’ve been advocating for a long time already to
and Anderson live on the other side of Brazil in Belém). increase the uptake of digital technologies, such as BIM,
at the local government.
Everson: We used different tools that were made
available through the platform. First there’s the File Name
Convention tool which allowed us to more efficiently have Acknowledgement
control over the different documents that we used within
This article is prepared within the scope of the BIM-
the project. We also used the Mereen Weather Service SPEED project, which has received funding from the
on the platform, which allowed us to assess the climate European Union’s Horizon 2020 research and innovation
and meteorological data in the region by inputting the programme under the grant agreement number 820553.
latitude and longitude. From there we used BIM to better The European Union is not liable for any use that may
analyse the surroundings of the building and the amount be made of the information contained in this document,
which is merely representing the authors’ views.
of solar exposure.
Anderson: To a certain extend we’ve also used the

BIMSpeed Library, or at least the concept of it, as the
data on the platform was adapted to different European
regions. We’ve used the software and adapted it with data
from our own region. Through this we could examine

INDUSTRY EXPERTISE – ADVERTORIAL
Smart Industrial Thermal Imaging Device PX1

– A Building HVAC Protection Tool – Escorts the Heating Season
Application Background leakage and wind pouring in of the house caused by poor
sealing, and blocking the cold wind outside promptly to
As an efficient, scientific, and intuitive detection tool for
ensure the warmth inside.
detecting building HVAC faults, the thermal camera has
been widely received by the market for its “nondestructive Hollow External Wall Detection
testing” performance. Thermal Imaging is widely used
in the building HVAC field for its competitiveness such Due to different thermal conductivity coefficient, the
as visualization of temperature data, which can detect temperature transfer will be hindered when there are
problems before hidden dangers occur, and also quickly hollows or cavities on the walls of a building. Therefore,
determine the fault location in case of failure. a non-uniform temperature field will be formed on the
surface during temperature rise or fall. For example,
When the heating season comes, problems such as heating when the indoor temperature is low and the outdoor
failure, water seepage at welds or joints, floor heating temperature is high by sunlight, heat will accumulate at the
leakage, and surface deformation, and aging are the most hollow spot. Compared with the traditional hand-knocking
common. Thermal Imaging can detect these problems method, Thermal Imaging does not need to erect scaffolds
to help engineers efficiently complete daily maintenance for detection and can shoot in a large area. And the area
work, ensuring warmth for citizens throughout the winter. and extent of damage can be clearly distinguished through
infrared images.
HVAC Fault Detection
As a nondestructive testing tool that visualizes temperature Water Leakage Detection for Buildings
data, Thermal Imaging can visualize the heat distribution Leakage in building roof, window wall corners, top and
on the ground, making it clear to see the temperature of bottom slab is a common quality common problem. With
the pipeline from the floor heating to the ground, and can excellent thermal sensitivity, Thermal Imaging can quickly
scan the area of the leaks, which is conducive to accurate and clearly display subtle temperature differences, search
and rapid positioning of the underground leaks, facilitating and locate the leakage point, and help routine inspectors
maintenance, reducing energy loss and ensuring normal efficiently troubleshoot the wall faults so as to formulate
heating in winter. effective maintenance plans to improve building quality.
Air Tightness Test for Houses Electrical Device Security Monitoring

The air tightness of the house is crucial in the cold winter. As temperature drops, there will be an increase in electricity
If there are gaps in the windows, doors, or walls, the chill consumption in both enterprises and communities.
wind will blow into the house along the gaps, plummeting Whether power supply devices and electrical devices
indoor temperatures. Thermal Imaging can quickly locate in the distribution stations work properly affects the
the gaps in the walls, doors, or windows to figure out where production operations of the park and the livelihood of the
the chill wind comes from, thus solving the problem of wind community residents. In the event of a malfunction, power

INDUSTRY EXPERTISE – ADVERTORIAL
failure, voltage instability, and other electrical accidents IP68/IP69K protection rank is reached, complying with
may ensue. Thermal Imaging can timely discover a thermal MIL-STD-810H standard
defect and hidden thermal danger of electrical devices, 1.5 m-drop protection, as well as water and dust
thus preventing line thermal accidents, and ensuring the resistance, makes it solid and durable
security of electricity in winter.
5500 mA ultra-large capacity battery supports long
outdoor use
Introduction to Smart Industrial Thermal
Imaging Device PX1
Application Value of Smart Industrial Thermal
PX1 is built with an InfiRay self-developed 12 μm high Imaging Device PX1
performance 256×192 infrared detector, a dual night Non-contact measurement can be realized, which allows
vision system and a 48-megapixel main camera, which can for remote measuring without changing the structure of
support routine inspection Thermal Imaging in any light. the measured target, thus ensuring inspectors’ safety
High IP grade and large-capacity battery can meet the
Professional measurement tools are provided to freely
requirements of engineers for long outdoor inspection in
monitor the areas selected and automatically obtain
harsh environment.
the highest temperature point to implement intelligent
Three-spectrum sensing of Thermal Imaging, visible photographing and fault diagnosis of devices
light, and low light night vision is available to support
Photographing and video recording are supported to
engineers for routine inspections at any time without
transmit inspection data to the background, which is
any blind spots
helpful for secondary analysis and remote diagnosis of
The minimum temperature difference of 0.04°C can be on-site faults
distinguished, and the measurement accuracy can reach
App development is supported to help customers
±2°C, with the frame rate up to 25Hz
enhance the level of intelligent inspection for efficiency
Professional SDK and technical support can be provided improvement and cost savings
to help customers develop their own apps
Lightweight structure makes it easy to carry, and
Professional APP for analysis on infrared temperature enables an all-in-one design, which lightens the burden
measurement is built in, analyzing while shooting to of inspectors carrying many devices
quickly get accurate data
InfiRay

Innovative systems for
better performance
Purmo
Let’s create the systems that take indoor
climate comfort further
By delivering complete solutions rather than just components, we can offer indoor
climate comfort that takes multiple aspects of performance optimisation into account.
This includes supporting our customers with advice and guidance to make sure our
solutions match their needs exactly. At the same time, our solution approach enables us
to better deliver higher energy efficiency. For example, by combining low temperature
heat emitters with underfloor heating, we can deliver an optimal balance between swift
changes in room temperatures and energy efficiency. Our configuration software is
very useful at product selection and system design stages. Product information can be
easily forwarded to BIM design systems and from there, support the entire process
from planning to installation.
Discover more at www.purmogroup.com

Please send information of your event to Ms Nicoll Marucciova [email protected] EVENTS
Exhibitions, Conferences and Seminars

in 2022 & 2023
Conferences and fairs 2022

November-December
3 Nov 2022 Climatization Days Lisbon, Portugal
14-15 Nov 2022 Brussels Summit 2022 (rehva.eu) Brussels, Belgium
30 Nov - 2 Dec 2022 53rd International HVAC&R Congress and Exhibition Belgrade, Serbia
(kgh-kongres.rs)
8-10 Dec 2022 REFCOLD India 2022 (refcoldindia.com) Mahatma Mandir, Gandhinagar, Ahmedaba, India
Conferences and fairs 2023

February
4-8 Feb 2023 2023 ASHRAE Winter Conference (ashrae.org) Atlanta, Georgia, USA
28 Feb - 3 Mar 2023 World Sustainable Energy days (wsed.at) Wels, Austria
March
6-8 Mar 2023 HVAC Cold Climate Conference 2023 (ashrae.org) Anchorage, Alaska
13-17 Mar 2023 ISH 2023 (ish.messefrankfurt.com) Frankfurt am Main, Germany
14-16 Mar 2023 ACREX 2023 (acrex.in) Mumbai, India
20-23 May 2023 IAQVEC 2023 (iaqvec2023.org) Tokyo, Japan
Due to the COVID-19 circumstances, the dates of events might change. Please follow the event’s official website

MEMBERS
ATIC vzw–asbl – Belgium BAOVK – Bulgaria STP – Czech Republic DANVAK – Denmark EKVU – Estonia
www.atic.be www.baovk.bg www.stpcr.cz www.danvak.dk www.ekvy.ee
FINVAC – Finland
www.finvac.org
REHVAAICVF – France
www.aicvf.org
VDI–e.V. – Germany
www.vdi.de
ÉTÉ – Hungary
www.eptud.org
MMK – Hungary
www.mmk.hu
AiCARR – Italy
www.aicarr.org
Members
AHGWTEL/LATVAC – Latvia
www.lsgutis.lv
LITES – Lithuania
www.listia.lt
AIIRM – Republic of Moldova
www.aiirm.md
TVVL – The Netherlands
www.tvvl.nl
NEMITEK – Norway PZITS – Poland ORDEM DOS ENGENHEIROS – Portugal AFCR – Romania AGFR – Romania
www.nemitek.no www.pzits.pl www.ordemengenheiros.pt www.criofrig.ro www.agfro.ro
AIIR – Romania KGH c/o SMEITS – Serbia SSTP – Slovakia SITHOK – Slovenia ATECYR – Spain
www.aiiro.ro www.smeits.rs www.sstp.sk https://web.fs.uni-lj.si/sithok/ www.atecyr.org
SWEDVAC – Sweden DIE PLANER – Switzerland TTMD – Turkey CIBSE – United Kingdom
www.energi-miljo.se www.die-planer.ch www.ttmd.org.tr www.cibse.org
Network of 26 European HVAC Associations joining 120 000 professionals

REHVA Office: Rue Washington 40, 1050 Brussels - Belgium • Tel: + 32 2 514 11 71 - [email protected] - www.rehva.eu
SUPPORTERS
Leaders in Building Services
Daikin Europe – Belgium EPEE – Belgium EVIA – Belgium Velux – Denmark Granlund – Finland
www.daikin.eu www.epeeglobal.org www.evia.eu www.velux.com www.granlund.fi
Halton – Finland Uponor – Finland Eurovent Certita Certification – LG Electronics – France Viega – Germany
www.halton.com www.uponor.com France www.lgeaircon.com www.viega.com
www.eurovent-certification.com
Aermec – Italy
www.aermec.com
REHVA
Evapco Europe – Italy
www.evapco.eu
Rhoss – Italy
www.rhoss.com
Purmo Group – The Netherlands
www.purmogroup.com
Royal Haskoning DHV –
The Netherlands
Supporters
www.royalhaskoningdhv.com
East European Business Center
SMAY – Poland E.E.B.C. – Romania Dosetimpex – Romania Testo – Romania Camfil – Sweden
www.smay.eu www.eebc.ro www.dosetimpex.ro www.testo.com www.camfil.com
Fläkt Group – Sweden Lindab – Sweden Swegon – Sweden Systemair – Sweden Belimo Automation – Switzerland
www.flaktgroup.com www.lindab.com www.swegon.com www.systemair.com www.belimo.com
Arçelik – Turkey Friterm Termik Cihazlar Zoonex – United Kingdom

www.arcelikglobal.com Sanayi ve Ticaret – Turkey www.zoonexsystems.com
www.friterm.com
REHVA Associate Organisations:
Enerbrain srl – Italia Enviromech – United Kingdom ISIB – Turkey ECI – Belgium OAER – Romania
www.enerbrain.com www.enviromech.co.uk www.isib.org.tr copperalliance.org www.oaer.ro
Network of 26 European HVAC Associations joining 120 000 professionals

REHVA Office: Rue Washington 40, 1050 Brussels - Belgium • Tel: + 32 2 514 11 71 - [email protected] - www.rehva.eu
welcome to the
REHVA
Brussels Summit

Rehva 2022 - 05

Uploaded by

Copyright:

Available Formats

Rehva 2022 - 05

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Rehva 2022 - 05

Uploaded by

Copyright:

Available Formats

The REHVA

European HVAC Journal

CLIMA 2022 papers on

World’s leading trade fair for

13. – 17. 3. 2023

Editor-in-Chief: EDITORIAL 69 Correlation of subjective and

Tiberiu Catalina, Romania Paulius Spūdys, Darius Pupeikis, Andrius

Branislav Todorovic, Serbia

Health & Comfort

The current situation in Europe, the high fuel prices

4 The REHVA European HVAC Journal — October 2022

CLIMA 2022 conference papers on the

Ten are included in this issue:

The REHVA European HVAC Journal — October 2022 5

To meet the challenges of our business, we at Swegon believe in

Evaluation of preventive measures

YUNUS EMRE CETIN MARTIN KRIEGEL

The REHVA European HVAC Journal — October 2022 7

𝜆𝜆𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣 = 𝜆𝜆𝐻𝐻𝐻𝐻𝐻𝐻𝐻𝐻 𝜀𝜀𝐻𝐻𝐻𝐻𝐻𝐻𝐻𝐻 + 𝜆𝜆𝑃𝑃𝑃𝑃𝑃𝑃 (4)

where P 0 is the probability of infection, D is the 𝜆𝜆𝐻𝐻𝐻𝐻𝐻𝐻𝐻𝐻 = 𝜆𝜆𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜 + 𝜆𝜆𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟

Table 1. Settings of the studied cases.

8 The REHVA European HVAC Journal — October 2022

2.3 Model parameters

For the base cases mixing ventilation (εHVAC = 1) is

The REHVA European HVAC Journal — October 2022 9

office. As a result of the excessive quanta gen-

Infection risk probabilities under different miti-

In general, it is seen that all the measures help to

As seen in Figure 2, in the case of the waiting

Infection risk predictions for the classroom,

10 The REHVA European HVAC Journal — October 2022

be a feasible solution in a music class. Also,

Important outcomes are summarized below:

• Predicted infection risk values show a devi-

The REHVA European HVAC Journal — October 2022 11

Aerosol transmission in rotary wheel

HEINRICH HUBER THOMAS RICHTER FLORIAN BRZEZINSKI MICHAEL RIEDIKER

Abstract: Transmission by aerosols is considered the main route of COVID-19 infections

12 The REHVA European HVAC Journal — October 2022

The REHVA European HVAC Journal — October 2022 13

8 EATR, dp = 250 Pa 8 EATR, dp = 250 Pa

EATR, ASTR [%]

ASTR, dp = 250 Pa ASTR, dp = 250 Pa

14 The REHVA European HVAC Journal — October 2022

4.2. State of the art in filters

Table 1. EATR and ASTR, condensation wheel.

MP, - v, m/s n, rpm Δp22-11, Pa EATR, % ASTR, %

Table 2. EATR and ASTR, sorption wheel.

The REHVA European HVAC Journal — October 2022 15

Table 3. Simulated virus concentrations (copies/m³) in

16 The REHVA European HVAC Journal — October 2022

Development of a non-contact modular

JINKYUN CHO JINHO KIM JONGWOON SONG SEUNGMIN JANG

Keywords: COVID-19, screening clinic, cross-infection, ventilation strategy, CFD, particle

18 The REHVA European HVAC Journal — October 2022

The REHVA European HVAC Journal — October 2022 19

20 The REHVA European HVAC Journal — October 2022

The REHVA European HVAC Journal — October 2022 21

of SCB and ER with the ventilation system turned on