Adaptive Fuzzy Control of Thermal Comfort in Smart

Houses

Bogdan Walek Michal Janošek

Department of Informatics and Computers Institute for Research and Applications of Fuzzy Modeling
University of Ostrava University of Ostrava
Ostrava, Czech Republic Ostrava, Czech Republic
[email protected] [email protected]

Jaroslav Žáþek Radim Farana

Institute for Research and Applications of Fuzzy Modeling Department of Informatics and Computers
University of Ostrava University of Ostrava
Ostrava, Czech Republic Ostrava, Czech Republic
[email protected] [email protected]

Abstract—This paper deals with an adaptive fuzzy mechanism house by CCTV cameras, or to play the latest films or music?
for heating control in smart houses. A fuzzy approach is Mostly it is the comfort which is the most significant criterion
connected with an expert system and considers various input when deciding whether to get a smart house or not.
data which can enter the system. The proposed adaptive
mechanism adjusts the thermal comfort rate based on the input One of the required functions is to ensure requested thermal
data and IF-THEN rules. The knowledge base of the expert comfort. This can only be accomplished by cooperation of
system is created based on input data and it is possible to extend several systems and evaluation of a number of data. Moreover,
it to other possible input values. The proposed adaptive not every necessary measurement is possible to carry out
mechanism is demonstrated on a particular example. accurately enough and for a reasonable price.

Keywords— fuzzy; fuzzy control; smart house; thermal comfort; Owing to the fact that there is a mutual communication
adaptive; expert system working among the particular systems within smart housing
(often via a central control unit, such as PLC), these are not
separate systems. The system of the smart house control can
I. INTRODUCTION take into consideration a number of factors, such as air quality,
With the expansion of the smart housing concept, the light comfort, renewable sources, or energy consumptions.
requirement for a smart house able to handle a large number of
advanced functions appears more and more. Suppliers of smart The paper focuses on the smart housing functions heading
home solutions mostly divide the whole platform into several to ensure thermal comfort on the basis of selected measurable
logical areas regarding which area the particular platform physical quantities in association with the control of
focuses on. It is particularly the case of security, economy, technology for heating or air conditioning.
comfort, entertainment, and ecology areas.
II. THERMAL COMFORT
The whole concept of smart houses came into existence
because of these areas, which are important for an ordinary Thermal comfort influences the quality of using interiors
user. Security of the house is vital for its inhabitants. After all, and even though it depends on several physical measurable
security systems belong among the first remote monitored and quantities, it can be very subjective. For someone, the degree of
controlled systems by the user. In the course of time, functions thermal comfort can be winter clothing and 15°C, for someone
as turning on the heating, turning off the lights and others were else 27°C and swimwear. Since thermal comfort ensuring
added, all that via a mobile phone and electronic security influences energy performance of buildings, it can be applied
system. Economy is another significant criterion, both for the in low-energy or passive houses as well [3].
house construction and for its operation. If this attribute of Factors affecting thermal comfort include the rate of
economy also has a positive impact on ecology, it is an metabolism and the clothing insulation of the particular person,
important decision-making criterion for a current user. All of air temperature, relative humidity of air, mean radiant
that by means of elaborated statistics of energy usage, which a temperature and air speed in the room. Thermal comfort can
house can compile due to monitoring of the consumptions. Of also be influenced by more subjective factors, such as the
course, the main part consists of entertainment and comfort. mental well-being of the particular person [2]. The season of
Who would not want, comfortably by their phone, to control the year or current weather situation can also make a difference
the gate on the opposite side of the garden, to observe the

675
978-1-4799-3528-4/14/$31.00 ©2014 IEEE

Authorized licensed use limited to: Pontificia Universidad Javeriana. Downloaded on August 19,2021 at 03:01:59 UTC from IEEE Xplore. Restrictions apply.
for some people. Heating control units often use the following are two main reasons for the application of the weather-
quantities for calculating thermal comfort: air temperature, compensated control:
relative humidity of air, mean radiant temperature and air speed
in the room (Table I.). A number of models are used for this a) higher thermal comfort on the whole, which arises
thermal comfort calculation, of which the most frequently used due to the suppression of the temperature dynamics
and acclaimed one is PMV [1]. This model belongs to static (fluctuation) in the room
models. The second type is a so-called adaptive model, based b) energy economy – it is not necessary to heat up the
on the idea that the outside climate influences the inside heat source to the maximum, just to such power
thermal comfort as people are able to adapt to different which is sufficient to warm the room to the required
temperatures in the course of the year. ASHRAE standard deals temperature with regards to the outside one
not only with this model [4].
The disadvantage of the stated process is mainly the
TABLE I. INSIDE CLIMATE QUANTITIES [3] difficult way of obtaining the mean radiant temperature value.
The problem is primarily a shortage of currently produced
Parameter Symbol Quantity Notes
means for such measurement. From this reason, PMV value is
air temperature ta °C measured mostly determined according to so-called operating temperature,
within +5 to +35 which is measured by ball thermometer. Its disadvantages are
°C measurement inaccuracies and significantly different time
relative humidity Ma % various types of
of air hygrometers are
constants negatively influencing the thermal comfort
used evaluation. Moreover, they are rather expensive. The mean
mean radiant tr °C measured mostly radiant temperature is often substituted with the reference
temperature within radiant temperature trt, which is calculated from the external
-25 to +75 °C (outdoor) and internal (indoor) temperatures - tex a tin (2) [3].
air speed va, Sa m/s air speed and Similarly, the value of the heat transfer coefficient through the
draught are
measured by wall without considering the heat transfer coefficient on the
anemometer inner side U st and the value of the heat transfer coefficient
through the wall with considering the heat transfer coefficient
The result, which the control unit is able to calculate, are *
on the inner side U st are necessary for the calculation (1).
PMV values (Predicted Mean Vote), PPD values (Predicted
Percentage of Dissatisfied) and DR (Draught Raging) [1].
q st = U st (t in − t ex ) = U st* (t rt − t ex ) (1)

Control PMV specific heat flux through the wall (1)

unit
PPD
U st (2)
DR
Weather- t rt = (tin − tex ) + tex
compensat U st*
uPMV(τ) ed control

tex(τ) reference radiant temperature (2)

tut(τ)
III. FUZZY APPROACH
Heating/ Due to these particular reasons, this paper proposes an
cooling adaptive fuzzy approach, which will count the weather-
compensated control values on the basis of the input values and
expert systems rules. The expert system will contain the
Fig. 1. Thermal comfort control circuit [3]. knowledge base consisting of IF-THEN rules. The proposed
fuzzy approach is graphically demonstrated in the following
In the scheme shown in Fig. 1 above, first the control unit picture:
calculates PMV, PPD and PRV values on the basis of input
values shown in Table I. Then uPMV(t) is calculated by means
of thermometers in the room. This value, together with the
external and the internal temperature values - tex(t) a tin(t), is
afterwards the input into a so-called weather-compensated
control. The weather-compensated control sets the temperature
of the heating water with regards to the outside temperature.
This control thus ensures lower or higher temperature of the
supplied heating water on the basis of the input values. There

2014 15th International Carpathian Control Conference (ICCC) 676

Authorized licensed use limited to: Pontificia Universidad Javeriana. Downloaded on August 19,2021 at 03:01:59 UTC from IEEE Xplore. Restrictions apply.
 a) air temperature – values (ex sm – extra small, ve sm-
very small, sm – small, me – medium, bi – big, ve bi-
very big, ex bi – extra big)
b) relative humidity of air – values (ex sm – extra small,
ve sm – very small, sm – small, me – medium, bi –
big, ve bi – very big, ex bi – extra big)
c) air speed – values (ex sm – extra small, ve sm – very
small, sm – small, me – medium, bi – big, ve bi –
very big, ex bi – extra big)
d) outside temperature – values (ex sm – extra small, ve
sm – very small, sm – small, me – medium, bi – big,
ve bi – very big, ex bi – extra big)

The output value is

a) Heating water temperature – values (ex sm – extra
small, ve sm – very small, sm – small, me – medium,
bi – big, ve bi – very big, ex bi – extra big)

With regard to these values, the knowledge base consisting

of IF-THEN rules is then created. The knowledge base is filled
up by an expert in a residential premises heating domain. This
expert is (on the basis of his experience) able to determine
particular output values according to various combinations of
the input values.
The expert system is created in LFLC (Linguistic Fuzzy
Fig. 2. Proposed fuzzy approach
Logic Controller) tool. LFLC tool is more described in [7].
The proposed approach relies on four measurable values This system enables to create a knowledge base of an expert
(air temperature ta, relative humidity of air Ma, air speed va, and system and also ensures the process of decision making
outdoor temperature tex). The radiant temperature value will be together with a specific output, which is possible to use
replaced with air temperature and outdoor temperature. Since straightforward as the inputs into the weather-compensated
fuzzy logic can operate with fuzzy concepts, it can also control.
determine the resultant temperature of heating water on the The input values can contain linguistic variables which are
basis of not very accurate input data. Moreover, the user demonstrated by means of fuzzy sets. In the following picture,
himself, while contemplating about the temperature, does not the linguistic expression “small” is highlighted by red color [5]:
think in accurate numbers. The user rather prefers expressions
such as pleasant temperature, higher than pleasant temperature,
unpleasant temperature etc. The values of the four measurable
quantities are at the same time the input values of the expert
system. Consequently, the expert system proposes the
requirement of the optimal temperature of the heating water on
the basis of the input values and knowledge base created by an Fig. 3. Fuzzy sets of linguistic variables
expert in the area of residential premises heating. Afterwards,
according to this temperature, the weather-compensated control The knowledge base consists of IF-THEN rules which are
sets continuous optimal power of the heat source. Thanks to composed of input and output linguistic variables described
that, thermal comfort will be ensured for the residents. earlier. The knowledge base was created in LFLC tool. A part
of the knowledge base is shown in the picture below:
IV. EXPERT SYSTEM
The main part of the proposed fuzzy approach is the expert
system which executes the decision making with regard to the
input values and knowledge base. The input values will be
represented by means of linguistic variables, which will be the
basic structural unit of the expert system and will enable easier
creation of the knowledge base and its filling-up [5, 6].
The input values of the expert system are (ex sm – extra
small, ve sm – very small, sm – small, me – medium, bi – big,
ve bi – very big, ex bi – extra big):
Fig. 4. A part of the knowledge base of the expert system

2014 15th International Carpathian Control Conference (ICCC) 677

Authorized licensed use limited to: Pontificia Universidad Javeriana. Downloaded on August 19,2021 at 03:01:59 UTC from IEEE Xplore. Restrictions apply.
 The picture clearly shows that if, at the same time, the air very often are) different, which may lead to various problems
temperature is high, the relative humidity of air is high, the air nowadays.
speed is low, and the outside temperature is high, then the
heating water temperature is extremely low. This means that ACKNOWLEDGMENT
basically there is no need to heat, rather to use air conditioning.
The opposite case is the situation when the air temperature is This work was supported by the European Regional
low, the relative humidity of air is low, the air speed is high, Development Fund in the IT4Innovations Centre of Excellence
and the outside temperature is very low; then the heating water project (CZ.1.05/1.1.00/02.0070) and by the European project
temperature must be extremely high. In other words, it is OU OP VK (CZ.1.07/2.2.00/28.0245).
necessary to heat up significantly so that the optimal thermal
comfort is ensured. REFERENCES

V. CONCLUSION [1] Fanger, P.O., „Thermal Comfort,“ New York, Mc-Graw-Hill Book
Company, 1970, 224 p.
This paper proposed a fuzzy adaptive approach which
[2] Andreoni G., Piccini L., Maggi L., „Active Customized Control of
suggests the optimal heating water temperature, and Thermal Comfort,“ International Encyclopedia of Ergonomics and
consequently the optimal degree of thermal comfort on the Human Factors, Second Edition - 3 Volume Set. Edited by Waldemar
basis of measurable values in the house and the expert system. Karwowski. CRC Press 2006. ISBN: 978-0-415-30430-6.
The proposed approach includes an expert system whose main [3] Hruška, F. „Tepelná pohoda a inteligentní dĤm/Thermal Comfort and
parts were described in the paper in detail. The expert system intelligent house.“ Automa 2013, 19 (4), 14–16. ISSN 1210-9592.
was verified on a simple example. The advantage of the fuzzy [4] de Dear, R., Brager, G.: (1998). "Developing an adaptive model of
approach compared to PMV is the ability to work with fuzzy thermal comfort and preference. ASHRAE Transactions, Vol 104 (1):
145–67.
variables, which is also user’s requirement for thermal comfort.
[5] Zadeh, L. A., “Fuzzy sets, Information & Control,” vol. 8, 1965, pp.
In the future, the focus of the authors of the paper will be 338-353.
on the calculation of thermal comfort based on user’s behavior [6] Novák, V. and Perfilieva, I., “Evaluating Linguistic Expressions and
and his mental well-being. These subjective quantities can have Functional Fuzzy Theories in Fuzzy Logic,” Computing with Words in
Information/Intelligent Systems 1, L. A. Zadeh a J. Kacpryk (eds.),
a relatively high impact on the requirement for the optimal Springer-Verlag, Heidelberg, 1999, pp. 383-406.
thermal comfort. In addition, attention will be paid to [7] H. Habiballa, V. Novák, A. DvoĜák, V. Pavliska, “Using software
calculation of the thermal comfort for more users located in the package LFLC 2000”, 2nd International Conference Aplimat 2003,
same room. Their thermal comfort requirements can be (and Bratislava, 2003, pp. 355-358.

2014 15th International Carpathian Control Conference (ICCC) 678

Authorized licensed use limited to: Pontificia Universidad Javeriana. Downloaded on August 19,2021 at 03:01:59 UTC from IEEE Xplore. Restrictions apply.