9
Comparison of thermal comfort algorithms in naturally ventilated office buildings Bassam Moujalled *, Richard Cantin, Ge ´ rard Guarracino Ecole Nationale des Travaux Publics de l’Etat, CNRS, URA 1652, De ´partement Ge ´nie Civil et Baˆtiment, Universite ´ de Lyon, 3 rue Maurice Audin, Vaulx-en-Velin F-69120, France 1. Introduction The energy consumption in office buildings is mainly used to create and maintain comfort conditions in the indoor environment, which also affect health and productivity of the occupants. In France, the HVAC systems account for more than 60% of energy consumption in these buildings. Also the recent years have shown a rise in the number of air-conditioning systems which creates supplementary loads during the warm season. With the urgent need to reduce the economic and environ- mental costs of energy consumption, the European and national French institutes gives top priority to the energy efficiency in the building sector. However, energy saving measures should be realized without detriment to the occupant’s comfort. Hence, the latest revision of the French thermal regulation aims to reduce the energy consumption due to air conditioning during the warm season and encourages passive cooling techniques such as natural ventilation which increases indoor air speed and improves the comfort by cooling down the human body and the building structure [1]. Naturally ventilated buildings typically use less than half as much as energy than those with air conditioning [2]. Several research projects [3,4] have shown that these buildings can be comfortable all over the year. Occupants of naturally ventilated buildings were found to accept and prefer a significantly wider range of temperatures that fall out of the standard comfort zone defined by ISO 7730. The standard’s comfort zones are based on laboratory experiments, and are suitable for air-conditioned buildings where the thermal conditions are static. In naturally ventilated buildings, indoor temperatures are variable and cycle or drift in response to the natural swings of the outdoor and indoor climate, especially in the warm season, as in the cold season these buildings are generally heated [5]. Thus, an adaptive approach has been developed using the results of the field studies. This approach considers that occupants of naturally ventilated buildings have different expectations and are able to adapt themselves to their thermal environments in buildings that afford them greater degrees of control over thermal conditions [6]. Adaptive comfort algorithms have been set for naturally ventilated buildings depending on the climatic conditions represented by the outdoor temperature. These algorithms have been introduced in the ASHRAE standard 55-2005 and the European standard EN15251 for the case of naturally ventilated buildings during warm season [7]. The purpose of this paper is to evaluate different algorithms from both static and adaptive approach in naturally ventilated Energy and Buildings 40 (2008) 2215–2223 ARTICLE INFO Article history: Received 16 November 2007 Received in revised form 30 May 2008 Accepted 18 June 2008 Keywords: Thermal comfort Adaptive algorithm Natural ventilation Office buildings ABSTRACT With the actual environmental issues of energy savings in buildings, there are more efforts to prevent any increase in energy use associated with installing air-conditioning systems. The actual standard of thermal comfort in buildings ISO 7730 is based on static model that is acceptable in air-conditioned buildings, but unreliable for the case of naturally ventilated buildings. The different field studies have shown that occupants of naturally ventilated buildings accept and prefer a significantly wider range of temperatures compared to occupants of air-conditioned buildings. The results of these field studies have contributed to develop the adaptive approach. Adaptive comfort algorithms have been integrated in EN15251 and ASHRAE standards to take into account the adaptive approach in naturally ventilated buildings. These adaptive algorithms seem to be more efficient for naturally ventilated buildings, but need to be assessed in field studies. This paper evaluates different algorithms from both static and adaptive approach in naturally ventilated buildings across a field survey that has been conducted in France in five naturally ventilated office buildings. The paper presents the methodology guidelines, and the thermal comfort algorithms considered. The results of application of different algorithms are provided with a comparative analysis to assess the applied algorithms. ß 2008 Elsevier B.V. All rights reserved. * Corresponding author. Tel.: +33 4 72 04 72 61; fax: +33 4 72 04 70 41. E-mail addresses: [email protected], [email protected] (B. Moujalled), [email protected] (R. Cantin), [email protected] (G. Guarracino). Contents lists available at ScienceDirect Energy and Buildings journal homepage: www.elsevier.com/locate/enbuild 0378-7788/$ – see front matter ß 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.enbuild.2008.06.014

Thermal comfort algorithms in naturally ventilated office buildings

Embed Size (px)

DESCRIPTION

Comparison of thermal comfort algorithms in naturally ventilated office buildings

Citation preview

Page 1: Thermal comfort algorithms in naturally ventilated office buildings

Comparison of thermal comfort algorithms in naturally ventilated office buildings

Bassam Moujalled *, Richard Cantin, Gerard Guarracino

Ecole Nationale des Travaux Publics de l’Etat, CNRS, URA 1652, Departement Genie Civil et Batiment, Universite de Lyon, 3 rue Maurice Audin, Vaulx-en-Velin F-69120, France

Energy and Buildings 40 (2008) 2215–2223

A R T I C L E I N F O

Article history:

Received 16 November 2007

Received in revised form 30 May 2008

Accepted 18 June 2008

Keywords:

Thermal comfort

Adaptive algorithm

Natural ventilation

Office buildings

A B S T R A C T

With the actual environmental issues of energy savings in buildings, there are more efforts to prevent any

increase in energy use associated with installing air-conditioning systems. The actual standard of thermal

comfort in buildings ISO 7730 is based on static model that is acceptable in air-conditioned buildings, but

unreliable for the case of naturally ventilated buildings. The different field studies have shown that

occupants of naturally ventilated buildings accept and prefer a significantly wider range of temperatures

compared to occupants of air-conditioned buildings. The results of these field studies have contributed to

develop the adaptive approach. Adaptive comfort algorithms have been integrated in EN15251 and

ASHRAE standards to take into account the adaptive approach in naturally ventilated buildings. These

adaptive algorithms seem to be more efficient for naturally ventilated buildings, but need to be assessed

in field studies. This paper evaluates different algorithms from both static and adaptive approach in

naturally ventilated buildings across a field survey that has been conducted in France in five naturally

ventilated office buildings. The paper presents the methodology guidelines, and the thermal comfort

algorithms considered. The results of application of different algorithms are provided with a comparative

analysis to assess the applied algorithms.

� 2008 Elsevier B.V. All rights reserved.

Contents lists available at ScienceDirect

Energy and Buildings

journal homepage: www.e lsev ier .com/ locate /enbui ld

1. Introduction

The energy consumption in office buildings is mainly used tocreate and maintain comfort conditions in the indoor environment,which also affect health and productivity of the occupants. InFrance, the HVAC systems account for more than 60% of energyconsumption in these buildings. Also the recent years have showna rise in the number of air-conditioning systems which createssupplementary loads during the warm season.

With the urgent need to reduce the economic and environ-mental costs of energy consumption, the European and nationalFrench institutes gives top priority to the energy efficiency in thebuilding sector. However, energy saving measures should berealized without detriment to the occupant’s comfort. Hence, thelatest revision of the French thermal regulation aims to reduce theenergy consumption due to air conditioning during the warmseason and encourages passive cooling techniques such as naturalventilation which increases indoor air speed and improves thecomfort by cooling down the human body and the buildingstructure [1].

* Corresponding author. Tel.: +33 4 72 04 72 61; fax: +33 4 72 04 70 41.

E-mail addresses: [email protected], [email protected]

(B. Moujalled), [email protected] (R. Cantin), [email protected] (G. Guarracino).

0378-7788/$ – see front matter � 2008 Elsevier B.V. All rights reserved.

doi:10.1016/j.enbuild.2008.06.014

Naturally ventilated buildings typically use less than half asmuch as energy than those with air conditioning [2]. Severalresearch projects [3,4] have shown that these buildings can becomfortable all over the year. Occupants of naturally ventilatedbuildings were found to accept and prefer a significantly widerrange of temperatures that fall out of the standard comfort zonedefined by ISO 7730. The standard’s comfort zones are based onlaboratory experiments, and are suitable for air-conditionedbuildings where the thermal conditions are static. In naturallyventilated buildings, indoor temperatures are variable and cycle ordrift in response to the natural swings of the outdoor and indoorclimate, especially in the warm season, as in the cold season thesebuildings are generally heated [5]. Thus, an adaptive approach hasbeen developed using the results of the field studies. This approachconsiders that occupants of naturally ventilated buildings havedifferent expectations and are able to adapt themselves to theirthermal environments in buildings that afford them greaterdegrees of control over thermal conditions [6]. Adaptive comfortalgorithms have been set for naturally ventilated buildingsdepending on the climatic conditions represented by the outdoortemperature. These algorithms have been introduced in theASHRAE standard 55-2005 and the European standard EN15251for the case of naturally ventilated buildings during warm season[7].

The purpose of this paper is to evaluate different algorithmsfrom both static and adaptive approach in naturally ventilated

Page 2: Thermal comfort algorithms in naturally ventilated office buildings

Fig. 1. Buildings E1 and E2.

Fig. 2. Frequency distribution of participants by age and gender.

B. Moujalled et al. / Energy and Buildings 40 (2008) 2215–22232216

office buildings for the French climatic context. Therefore a fieldsurvey has been conducted in five naturally ventilated officebuildings. First the paper presents the methodology guidelines,and the thermal comfort algorithms considered in this study. Then,results of application of different algorithms are provided with ananalysis of the indoor climates in the surveyed buildings. Finally acomparative analysis is developed which allows us to evaluate theapplied algorithms.

2. Methodology guidelines

2.1. Characteristics of surveyed buildings

In order to cover a wide interval of indoor conditions, thetransversal survey type [6] was adopted with multiple visits ineach of the surveyed buildings. Five buildings were surveyed. Theyare located in the southeast region of France near Lyon (Fig. 1).

The buildings have a low-rise concrete structure (3 up to 4levels) and the buildings’ facades are largely glazed. They werechosen for this study upon the following selection criteria: the useof natural ventilation for cooling during hot period with animportant thermal mass, solar protection for exposed glazing andoptional local fan in the local, besides the availability and will ofpeople to take part in the investigation. The surveyed buildings arepresented in Table 1.

In each building, offices have been selected in order to havevarious orientations and conditions encountered within thebuilding. Participants were selected in order to have a samplerepresentative of the occupants in the offices. Fig. 2 presents thefrequency distribution of the sample by age and gender.

2.2. Physical measurements of indoor climate

The field investigation was realized according to the level II [8]respecting the specifications set out in the ISO standard 7726 [9].The level II consists of physical measurements of variables that arenecessary for the calculation of thermal comfort indices. Thesemeasurements are accompanied by comfort questionnairescollected at the same time and place.

Table 1Presentation of the investigated NV buildings

Code Building City

E1 ENTPE existant Vaulx-en-Velin

E2 ENTPE extension Vaulx-en-Velin

C CETE L’Isle d’Abeau

L LASH Vaulx-en-Velin

P Palais de justice Lyon

The physical measurements aim at quantifying indoor climateand in particular thermal environment. In this investigation, themain thermal comfort parameters have been measured duringeach visit near each participant in order to calculate differentthermal comfort indices, especially the PMV index.

To achieve these measurements, Vivo equipment has beenchosen for several reasons. It offers portable and robust devicesthat measure air temperature, operative temperature, relativehumidity and air velocity in compliance to the specifications ofthe ISO 7726 standard. These devices have an importantcapacity of data storage up to 20,000 measurements. A batteryensures the energy autonomy for one day of full measurements.The devices can be programmed and controlled with a handhelddevice via an infrared port. Fig. 3 shows the measurementequipments.

The meteorological data were obtained from the weatherstation that is located on the roof of the building E1. Data includethe records of air temperature, relative humidity, wind speed anddirection and solar radiations, and can be downloaded on theinternet via the following link ‘‘http://idmp.entpe.fr/vaulx/mesfr.htm’’.

Survey period No. of observations

August 2004 and March 2005 95

August 2004 and March 2005 51

September 2004 and March 2005 110

August 2004 and March 2005 37

June 2005 37

Page 3: Thermal comfort algorithms in naturally ventilated office buildings

Fig. 3. Measurements instruments. V1 (vivo unit): air speed; V2 (vivo unit): operative temperature; V3 (vivo unit): relative humidity; V4 (vivo unit): battery; L: Minolta CL-

200 chromameter, C: CO2 analyser—Anagas CD98; S: Solo sonometer; R: Raynger MX2 – Infrared temperature and T: Tinytag recorder.

B. Moujalled et al. / Energy and Buildings 40 (2008) 2215–2223 2217

2.3. Description of the questionnaire

The physical measurements were accompanied with ques-tionnaire that is intended to evaluate thermal sensation andpreference of participants. The influence of the thermal environ-ment is assessed using subjective judgement scales according tothe ISO 10551 standard [10]. The questionnaire is also used tocollect data about the clothing and the activities of participants inorder to calculate thermal comfort indices.

The questionnaire is divided in four sections. The first sectionasks the subjects to evaluate their thermal environment at themoment of measurements using the perceptual, evaluation andpreferential judgement scales recommended by the ISO 10551standard. Table 2 shows the wordings of the three scales. Twoquestions about the acceptability of the thermal sensation and theair movement in the local are also included. The second section isthe clothing and activity checklists. Separated clothing checklistshave been developed for male and female on the basis of the ISO9920 [11]. The activity checklist inquires about physical activity,eating, drinking (hot or cold) and smoking during the hourprevious to taking the survey. In the third section, the subject hasto evaluate the interior air quality, the lighting and the soundquality, and the overall quality of the indoor environment at themoment of measurements on a seven-point-scale. The last sectionis a checklist on the use of different thermal environment controlmeans: windows, local fan, shading device.

2.4. Data collection procedure

Each building was surveyed five times during a weekalternating the visits between the morning and the afternoon.Only one person is needed to realize the complete data collection

Table 2The wording of the subjective scales used to assess the influence of the thermal

environment

Thermal perception scale Thermal evaluative scale Thermal preference scale

Cold Acceptable Much warmer

Cool Slightly unacceptable Warmer

Slightly cool Unacceptable Slightly warmer

Neutral Very unacceptable Neither warmer nor cooler

Slightly warm Slightly cooler

Warm Cooler

Hot Much cooler

procedure. The participants were surveyed at their workstationsaccording to the following steps. In the first step, the participanthas to fill the questionnaire. At the same time, the Vivo devices areplaced on the participant’s desk. The devices are set to make thethermal measurements during a period of 10 min. At the end of the10 min, the filled questionnaire and the devices are recovered.These different steps took about 12 min per participant. Duringeach visit, a total of 15–20 subjects were surveyed.

The field study was conducted during August and September2004 and March 2005, covering warm and cold conditions.

3. Calculated thermal comfort indices and algorithms

The collected physical measurements were entered in an excelfile. The questionnaires data were also transferred to an excel file.Clothing thermal insulation values were determined, for eachsubject, from the individual clothing articles indicated in thesurvey responses. Metabolic rates values were also determinedfrom the activity checklist inquired by participants. Then thephysical measurements and questionnaires data were merged on asingle excel work sheet to facilitate statistical analysis. From thisdatabase, different thermal comfort indices and algorithms can becalculated.

3.1. Rational comfort indices

The typical rational comfort indices are the PMV and PPDindices used by the ISO 7730 standard. To calculate them, aprogram has been developed in MATLAB software based on thealgorithm proposed in the ISO 7730 standard [12].

The effective temperatures ET* and SET used by the ASHRAEstandard were also calculated with another program developed inMATLAB. This program is base on the Gagge two-node model [13].It calculates in dynamic conditions the physiological parameters ofthe human body centre and skin according to the model suggestedby Gagge.

3.2. Adaptive comfort algorithms

The ASHRAE Standard 55 has included recently an alternativecompliance for naturally ventilated buildings based on theadaptive comfort standard developed by de Dear [3]. The adaptivestandard defines the ‘‘optimum’’ temperature as a function of themean monthly outdoor temperature of a location. It includes also

Page 4: Thermal comfort algorithms in naturally ventilated office buildings

Table 3Statistical summary of indoor climate measurements

Warm season Cold season

E1a E2a Ca La Pa alla E1a E2a Ca La Pa alla

48b 29b 70b 37b 37b 221b 47b 22b 40b 24b 20b 153b

Air temperature (8C)

Mean 28.0 29.8 24.9 29.8 25.5 27.1 23.0 23.4 24.0 25.5 23.6 23.8

S.D. 1.9 1.1 1.6 2.9 2.1 2.8 0.9 0.9 0.8 1.5 0.5 1.3

Min 23.7 27.2 21.9 23.2 22.7 21.9 21.1 21.7 21.6 21.7 22.7 21.1

Max 32.4 32.3 28.7 32.9 28.8 32.9 25.2 25.4 25.8 27.8 24.6 27.8

Operative temperature (8C)

Mean 28.3 30.0 25.1 29.8 25.3 27.3 23.1 23.4 23.8 25.4 23.4 23.7

S.D. 1.7 1.1 1.6 2.7 2.2 2.8 1.1 1.1 1.0 1.5 0.6 1.3

Min 24.6 27.4 21.7 23.5 22.4 21.7 20.9 21.6 21.7 21.6 22.4 20.9

Max 32.5 32.6 28.7 32.7 28.8 32.7 25.3 25.6 26.0 27.8 24.6 27.8

Air velocity (m/s)

Mean 0.20 0.25 0.06 0.25 0.09 0.15 0.05 0.05 0.06 0.05 0.07 0.06

S.D. 0.16 0.12 0.06 0.12 0.05 0.13 0.02 0.02 0.02 0.02 0.03 0.02

Min 0.03 0.04 0.02 0.06 0.01 0.01 0.00 0.02 0.02 0.02 0.01 0.00

Max 0.53 0.43 0.34 0.55 0.21 0.55 0.10 0.09 0.09 0.09 0.12 0.12

Relative humidity (%)

Mean 46.6 42.8 46.0 44.8 34.0 43.5 25.7 25.3 28.7 23.9 27.4 26.3

S.D. 9.1 6.5 6.9 3.3 8.9 8.5 2.4 2.8 4.0 1.9 6.9 4.0

Min 32.0 33.0 33.0 41.0 0.1 0.1 19.8 20.1 23.0 20.0 0.1 0.1

Max 65.0 58.0 58.0 57.0 44.0 65.0 31.0 31.0 35.0 27.0 33.9 35.0

a Building.b Sample size.

B. Moujalled et al. / Energy and Buildings 40 (2008) 2215–22232218

an acceptable range of temperatures based on criteria that either80% or 90% of the occupants will be comfortable within thoserespective ranges. The mean outdoor dry bulb temperature wascalculated from the climatic data file for each point of the data set.The comfort temperature can be found using the equationindicated in the standard. A mean comfort zone band of 5 and7 8C for 80% has been considered for 90 and 80% acceptability,respectively.

The European standard EN12521:2007 [14] gives also separatecomfort criteria for the naturally ventilated office buildings withoperable windows under the occupants’ control. The method isbased on the adaptive algorithm of the European study SCATs [4].The comfort temperature limits are given as a function of theexponentially weighted running mean of the outdoor temperaturewhich gives a higher weighting on today’s and yesterday’s outdoortemperatures instead of plain monthly average [7]. Threecategories of comfort limits are considered upon the level of

Fig. 4. Clo value frequency by g

satisfaction: 90, 80 and 65%. The first category which correspondsto the highest level of expectation is recommended for spacesoccupied by very sensitive and fragile persons, the second categorywith normal level of expectation to be used for new buildings andrenovations, and the third category with moderate level ofexpectation for existing buildings. The running mean outdoortemperature was also calculated from the climatic data file for eachpoint of the data set. The comfort temperatures can then be foundfor the three categories using the equations given in the standard.

4. Results of questionnaires and measurements

4.1. Physical measurements of indoor climate

Statistical summaries of measured physical parameters areprovided in Table 3 for the total data set broken down by buildingand by season. The total number of participants was 120 subjects

ender during cold season.

Page 5: Thermal comfort algorithms in naturally ventilated office buildings

Fig. 5. Clo value frequency by gender during warm season.

Fig. 6. Frequency distribution of thermal sensation vote by season.

Fig. 7. Frequency distribution of thermal preference vote by season.

B. Moujalled et al. / Energy and Buildings 40 (2008) 2215–2223 2219

making a total of 374 observations: 221 during the warm seasonand 153 during the cold season.

Air temperatures ranged from 21.9 8C to as high as 32.9 8C in thewhole buildings sample during the warm season, making anaverage of 27.1 8C. The highest temperatures have occurred inhighly glazed locals facing east or west and presenting a lack in thesolar shading devices. During the cold season, the air temperaturevariations were narrower, averaging around 23.8 8C with aminimum of 21.1 8C and a maximum of 27.8 8C. Because the useof local fans during the warm season, air velocities were higher andaveraged around 0.15 m/s (with a maximum of 0.55 m/s) against0.06 m/s during the cold season. Relative humidity showed lowvalues during the cold season with a mean of 26.3% against 43.5%during the warm season.

4.2. Assessment of clothing insulation and activity level

Figs. 4 and 5 show the calculated thermal insulation ofparticipants clothing for men and women during cold seasonand hot season, respectively.

The calculated clo values averaged around 0.6 clo during thewarm season and 0.9 clo during the cold season. In the warmseason as the indoor temperatures are generally high, thevariations of the clothing levels were limited between 0.3 cloand 0.7 clo. During the cold season, the clothing levels have widelyvaried from a minimum of 0.3 clo up to a maximum of 1.3 clo. Asthe indoor climates are heated during the cold season, theparticipants were more able to adjust their clothing to suit theirthermal preferences with the indoor conditions. In the warmseason, the clothing insulation levels are kept near to the lowestvalues possible for the office activity. In the questionnaire, theparticipants were asked if they consider the weather conditionswhen they get dressed in the morning. The results showed thatmore than 80% of the participants have adapted their clothing tothe weather conditions, especially the outdoor temperature.

The activity levels were approximately the same (1.2 met) asthe participants exercise the same type of activity in the five officebuildings. It seems also that the activity level was not affected bythe season.

4.3. Thermal sensation and preference of participants

The surveyed participants cast their votes on the thermalperception, evaluative and preference scales in response to the

immediate conditions at their desks. The distributions of votes onperception and preference scales are shown in Figs. 6 and 7 for thecold and warm seasons, respectively. The thermal sensationsdistribution is not the same in the warm and cold seasons. In thecold season, Fig. 6 shows that almost all the votes (92%) are withinthe central neutral category (between ‘‘slightly cool’’ and ‘‘slightlywarm’’ or�1 to +1 on the perception scale) and the two third of theparticipants want no change. The average thermal sensation is 0.4.In the warm season, Fig. 7 shows that only 66% of the thermal

Page 6: Thermal comfort algorithms in naturally ventilated office buildings

Fig. 8. Average thermal preference vote, coincident with thermal sensation vote for

the warm season. (Thermal preference scale: �3—colder, �2—cooler, �1: slightly

cooler; 0—no change; 1—slightly warmer; 2—warmer; 3—hotter).

Fig. 10. Linear regressions of thermal sensation votes vs. operative temperature for

the warm season (regression model was weighted by the number of votes falling

into each of the temperature bins on the x-axis).

Fig. 11. Linear regressions of thermal sensation votes vs. operative temperature for

the cold season (regression model was weighted by the number of votes falling into

each of the temperature bins on the x-axis).

B. Moujalled et al. / Energy and Buildings 40 (2008) 2215–22232220

sensation votes are within the central category (�1 to +1) and 38%on the warm side (+2 to +3). The average thermal sensation is 1.1.Almost 60% of the participants judged their thermal sensation to beunacceptable and want to be cooler with more air movement.

Figs. 8 and 9 show the mean thermal preference votes andcoincident distribution for each category of the thermal sensationscale for the warm and cold season, respectively. These figuresshow that the acceptability of warm and cool sensation is notsymmetric in both seasons. In the warm season, subjects’preference to return to neutral is stronger on the warm side thanon the cool side. In the cold season, the situation is reversed. Theneutral sensation is well accepted in both seasons and subjectsneeded no change.

4.4. Thermal neutralities

Simple linear regression was performed between thermalsensation and operative temperature to determine the strength ofthe relationship between them. The regression model is weightedby the number of observations making up the mean response ineach bin of the operative temperature.

Figs. 10 and 11 show the regressions obtained in the warm andcold seasons, respectively. For both season, the thermal sensation

Fig. 9. Average thermal preference vote, coincident with thermal sensation vote for

the cold season. (Thermal preference scale: �3—colder; �2—cooler; �1—slightly

cooler; 0—no change; 1—slightly warmer; 2—warmer; 3—hotter).

votes correlated strongly with the operative temperature that canexplain more than 80% of the variability of the thermal sensationvotes (r2 = 0.82 in the warm season and 0.85 in the cold season).The gradients of the regression models were slightly different:0.21/8C in the warm season against 0.29/8C in the cold season. It islower in the case of the hot season. As the gradient of the regressionmodels measures the thermal sensitivity, this means thatoccupants of naturally ventilated buildings were less sensitiveto the variations of the operative temperature in the warm seasonthan in the cold season. On average, mean thermal sensationschanged one unit every 3.5 8C of operative temperature in the coldseason, whereas in the hot season, 5 8C were needed to shift meanthermal sensation jump by one unit. In the hot season, as thevariations in the indoor temperature are more important in thesebuildings, occupants became less sensitive to the temperature rise.

Table 4The neutral temperatures in each building

Building Neutral temperature

Warm season (8C) Cold season (8C)

E1 24.2 22.9

E2 23.8 22.8

L 25.2 22.6

C 23.6 22.0

P 22.7 23.0

All 23.4 22.5

Page 7: Thermal comfort algorithms in naturally ventilated office buildings

Fig. 12. Linear regressions of thermal sensation votes vs. calculated PMV indices

(regression model was weighted by the number of votes falling into each of the

temperature bins on the x-axis).

B. Moujalled et al. / Energy and Buildings 40 (2008) 2215–2223 2221

This can be explained by the diversification of thermal experiencesof occupants and the interactions between occupants and theirenvironments as suggested by Nicol [15].

The neutral temperature (the temperature at which the meanthermal sensation for the group is ‘‘neutral’’) can be estimated fromthe regression model by solving the linear equation for a meanthermal sensation value of zero. The neutral operative temperatureis used to define ideal comfort temperatures. The calculatedneutral temperatures are presented in Table 4 for each building inboth seasons. The overall neutral temperature is 23 8C in the warmseason and 22.5 8C in the cold season. In buildings E1, E2 and Lwhere subjects were experiencing higher indoor temperatures, theneutral temperatures are found to be higher by 1 or 2 8C than in theother buildings. They are also higher in the warm season than inthe cold season. This difference can be explained by the seasonaladjustment of clothing. The thermal experience of occupants andthe amount of control that offers the naturally ventilated buildingsto the occupants over their environments help also in stretchingthe neutral temperature and the comfort zone.

5. Comparative analysis of comfort algorithms

5.1. Compliances of comfort votes with standard and adaptive

comfort zone

Physical measurements and comfort votes have been comparedwith the requirements of ISO 7730 standard and the adaptivecomfort ranges of ASHRAE and EN15251. The results are shown inTable 5.

The first line presents the percentage of people finding thethermal environment acceptable. In the next three lines arepresented the percentage of votes within the comfort zones of theISO7730, ASHRAE and EN15251 standards, respectively. For theISO 7730, the votes having the PPD (calculated from themeasurements) lower than 10% are considered within the comfortrange. For the ASHRAE and EN15251, the adaptive comfort zonesare applicable only to the naturally ventilated buildings during thewarm season. The considered comfort zones are those of 90%acceptability.

During the cold season, more than 80% of the participants findthe thermal environment acceptable, whereas in the warm season,almost the half find it unacceptable. When compared with therequirements of the ISO 7730, only 67% of the votes are within thecomfort range of the standard during the cold season, and 31%during the warm season. These values are lower than thoseobtained from the participants’ votes in both season. With theASHRAE and EN15251, almost half of the votes fall within thecomfort range during the warm season, and the results are veryclose to those obtained from the votes. The standard PPD of the ISO7730 underestimates the level of comfort in naturally ventilatedbuildings during the cold and the warm seasons, while theadaptive comfort algorithms of both ASHRAE and EN15251 givemore accurate results during the warm season in this case.

Table 5Comparison of comfort votes with thermal comfort standard ISO7730 and the

adaptive comfort algorithms of EN15251 and ASHRAE

Cold

season (%)

Warm

season (%)

People finding the thermal environment acceptable 83 48

Votes within the ISO7730 comfort zone (PPD < 10%) 67 31

Votes within the ASHRAE comfort zone for

NV buildings (90% satisfaction)

– 41

Votes within the EN15251 comfort zone for

NV buildings (category I)

– 52

The following paragraphs examine the relationship betweenthe perceived comfort and the PMV than with the adaptivealgorithms.

5.2. Comparison between perceived votes and PMV predicted votes

Simple linear regression has been performed between theperceived thermal sensation votes and the predicted thermalsensation by the PMV index to determine the strength of therelationship between them. The regression model is weighted bythe number of observations making up the mean response in eachbin of the PMV index.

Fig. 12 shows the regressions obtained from data of bothseasons. The thermal sensation votes correlated well with the PMVindex, but the gradient of the regression model was lower than one(0.62). It means that the PMV predicts warmer sensations on thewarm side of the scale, and cooler sensations on the cool side. Thus,the PMV overestimates the warm sensation in the range of warm/hot sensations, and underestimates it in the range of neutral andcool/cold sensations.

Fig. 13 presents a comparison between the percentages ofcomfort votes calculated from the questionnaires and from PPDagainst the operative temperature. The PPD predicts lowerproportions of comfort votes at different values of operativetemperature, and the difference becomes more important as thetemperature becomes higher and rises above 28 8C.

In this case study, the PMV and PPD failed to evaluate the levelof the thermal comfort in the surveyed naturally ventilatedbuildings. In such buildings, the adaptive mechanisms are

Fig. 13. The proportion of subjects who voted comfortable at different indoor

temperatures.

Page 8: Thermal comfort algorithms in naturally ventilated office buildings

Fig. 14. Indoor operative temperatures compared to the adaptive comfort zones of

the ASHRAE in each NV buildings in the warm season.

Fig. 16. Measured neutral temperatures compared to the adaptive comfort zone of

the ASHRAE in each NV buildings in the warm season.

B. Moujalled et al. / Energy and Buildings 40 (2008) 2215–22232222

primordial in achieving thermal comfort because of the variabilityof the indoor conditions and the amount of control given to theoccupants over their environments. The PMV and PPD can beconsidered as partially adaptive as they take into account theadjustments of clothing and air speed, but they exclude thepsychological mechanisms of the adaptation such as the expecta-tion and the habituation [6].

5.3. Comparison of the perceived comfort with the adaptive comfort

zone

The thermal comfort votes are compared with the adaptivebased comfort zones of ASHRAE and EN15251 standards during thewarm season.

Figs. 14 and 15 show the distribution of the operativetemperature against the outdoor temperature with the comfortlimits of the EN15251 and AHSRAE, respectively. For the building Cand P, almost all points fall within the comfort limits of bothstandards. For buildings E1, E2 and L, there are more points outsidethe comfort limits as the temperatures are higher especially with

Fig. 15. Indoor operative temperatures compared to the adaptive comfort zones of

the EN15251 in each NV buildings in the warm season.

the ASHRAE. In these buildings, the neutral temperatures are foundto be higher than in the two others.

Figs. 16 and 17 compare the neutral temperatures in the fivebuildings with the comfort limits of EN15251 and ASHRAEstandards, respectively. The neutral temperatures are closer tothe centre range of the ASHRAE comfort limits, while they arecloser to the lowest range of the EN15251 comfort limits. However,they are all within the comfort limits of both standards for the fivebuildings. The comfort limits of both methods correspond well tothe neutral temperatures found in this study.

6. Implication of comfort algorithms on the energy use

The application of the adaptive algorithms can help to reducethe energy consumption in buildings; in fact it allows the variationof comfort temperature in buildings that are not mechanicallycooled. Therefore, it allows the possibility of using passive meanswhich cannot guaranty a constant indoor temperature instead ofthe mechanical cooling that is very energy intensive. The EN 15251and ASHRAE standards give the categories of summer tempera-tures mainly used for the design of buildings without mechanicalcooling to prevent the overheating of the buildings by usingappropriate orientation, solar shading, utilisation of the buildings’thermal mass, natural ventilation, etc. [7]

In order to evaluate the impact of the use of the adaptivealgorithm in assessing thermal comfort, a naturally ventilatedoffice has been simulated using the TRNSYS software undersummer conditions.

The simulated local represents a typical office and was selectedfrom the surveyed offices of the field study. The office is located inthe building E2, and has registered the highest indoor tempera-tures during the warm period. It is an individual office of 30 m2

Fig. 17. Measured neutral temperatures compared to the adaptive comfort zone of

the EN15251 in each NV buildings in the warm season.

Page 9: Thermal comfort algorithms in naturally ventilated office buildings

Fig. 18. Quality of thermal environment of the simulated NV office in % of time of

occupation in the four categories of the EN15251 standard.

B. Moujalled et al. / Energy and Buildings 40 (2008) 2215–2223 2223

floor area and facing southwest with an operable window underthe control of the occupant. It has 6 m2 of glazing area equippedwith exterior blind. The simulation was run between June andAugust with the climatic data of Lyon (France). Fig. 18 shows theresult of the simulation in terms of the percentage of time ofoccupation where the indoor operative temperature is within eachof the four categories of the EN15251. During more than 70% of theoccupied time, the operative temperature falls outside of the threecomfort categories of the standard. As it was found during the fieldstudy, the thermal conditions in the local are unacceptable duringthe warm period. In fact, the local is disadvantaged by the largeglazing area and the southwest orientation. In order to enhance thethermal conditions without using mechanical cooling, the nightventilation technique was tested. The night ventilation has beenshown to be an appropriate strategy for office buildings that couldcover the majority of required cooling during the summer period. Itallows the exposed thermal mass of the building to use the coolnight air to discard the heat absorbed during the day [16]. Thesimulation was rerun with night ventilation of 3 ach when outsidetemperature is lower than the indoor temperature. This time, theindoor temperature is within the first comfort category during 63%of the occupied time and almost all the time within the thirdcategory that is recommended for existing buildings. Thus, thenight ventilation prevents the overheating in this case andprovides the comfort conditions. The application of the adaptivealgorithm of EN15251 can ensure comfort conditions in naturallyventilated buildings using passive means that need less energycompared to the mechanical cooling.

7. Summary and general conclusions

A field study has been conducted in five naturally ventilatedoffice buildings in south east of France during hot and cold seasons.This study has allowed to assess the different existent thermalcomfort algorithms. Findings from analysing the data gathered areas follows:

� T

he thermal indoor climate was in general warm during thewarm season, and more than the half of the participants weredissatisfied from the indoor thermal conditions and want to havemore air movement. However, during the cold season more than90% were comfortable with the thermal conditions. � T he thermal sensations are well correlated with the operative

temperature. The occupants were less sensitive to the rise oftemperature during the warm season. The variability of indoorconditions in the naturally ventilated buildings during the warmseason with the availability of thermal control has contributed toa relaxation of expectation and greater tolerance of temperaturerise.

� T

he application of the actual thermal comfort standard ISO7730has shown that the standard did not match with the results of thecomfort votes from the field study. The PMV has overestimatedthe warm sensation during the warm season, and the coolsensation during the cold season. This study confirms the lack ofreliability of the PMV index to predict thermal comfort innaturally ventilated buildings in both warm and cold season. � T he adaptive comfort algorithms of EN15251 and ASHRAE are in

close agreement with the measured comfort votes. It predictswell the thermal comfort of subjects in naturally ventilated officebuildings in our case.

� T he simulation of an office using TRNSYS shows the interest of

the adaptive approach for the comfort and the energy use innaturally ventilated buildings. The adaptive approach shows thatthe thermal comfort can be guaranteed during the large part ofthe occupation time with less expensive energy means such asnight ventilation.

Thus, the adaptive algorithms are more reliable to evaluate thethermal comfort in naturally ventilated buildings than thestandard PMV index. The use of the PMV to determine thermalcomfort conditions, as indicated by the standard ISO7730, wouldresult by overheating buildings during the cold season and air-conditioning during the warm season. The use of PMV is penalizingfor the naturally ventilated buildings while the adaptive algo-rithms are more advantageous for both thermal comfort andenergy use in buildings as they took into account for the variabilityof indoor comfort conditions. Other field studies in differentbuilding types and outdoor climate conditions are needed toconfirm these results.

References

[1] M. Santamouris, K. Pavlou, A. Synnefa, K. Niachou, D. Kolokotsa, Recent progresson passive cooling techniques—advanced technological developments to improvesurvivability levels in low-income households, Energy and Buildings 39 (7) (2007)859–866.

[2] J.F. Nicol, M.A. Humphreys, A stochastic approach to thermal comfort—occupantbehaviour and energy use in buildings, ASHARE Transactions 110 (2) (2004) 554–568.

[3] R.J. deDear, G.S. Brager, Thermal comfort in naturally ventilated buildings: revi-sions to ASHRAE Standard 55, Energy and Buildings 34 (6) (2002) 549–561.

[4] K.J. McCartney, J.F. Nicol, Developing an adaptive control algorithm for Europe,Energy and Buildings 34 (6) (2002) 623–636.

[5] G.S. Brager, G. Paliaga, R. deDear, Operable windows, personal control, andoccupant comfort, ASHRAE Transactions 110 (2) (2004) 17–35.

[6] M.A. Humphreys, J.F. Nicol, Understanding the adaptive approach to thermalcomfort, ASHRAE Transactions 104 (1B) (1998) 991–1004.

[7] B.W. Olesen, The philosophy behind EN15251: indoor environmental criteria fordesign and calculation of energy performance of buildings, Energy and Buildings39 (7) (2007) 740–749.

[8] R.J. deDear, A global database of thermal comfort field experiments, ASHRAETransactions 104 (1b) (1998) 1141–1152.

[9] NF EN ISO Standard 7726, Ergonomics of the Thermal Environment—Instrumentsfor Measuring Physical Quantities, AFNOR, Paris, France, 2003.

[10] NF EN ISO Standard 10551, Ergonomics of the Thermal Environment—Assessmentof the Influence of the Thermal Environment using Subjective Judgement Scale,AFNOR, Paris, France, 2003.

[11] NF EN ISO Standard 9920, Ergonomics of the Thermal Environment—Estimationof the Thermal Insulation and Evaporative Resistance of a Clothing Ensemble,AFNOR, Paris, France, 2003.

[12] NF EN ISO Standard 7730, Moderate Thermal Environments—Determination ofthe PMV and PPD Indices and Specification of the Conditions for Thermal Comfort,AFNOR, Paris, France, 2003.

[13] A.P. Gagge, A.P. Fobelets, L.G. Berglund, A standard predictive index of humanresponse to the thermal environment, ASHRAE Transactions 92 (2B) (1986) 709–731.

[14] NF EN15251, Indoor Environmental Input Parameters for Design and Assessmentof Energy Performance of Buildings—Addressing Indoor aur Quality, ThermalEnvironment, Lighting and Acoustics, AFNOR, Paris, France, 2007.

[15] J.F. Nicol, M.A. Humphreys, Adaptive thermal comfort and sustainable thermalstandards for buildings, Energy and Buildings 34 (6) (2002) 563–572.

[16] M. Kolokotroni, B.C. Webb, S.D. Hayes, Summer cooling with night ventilation foroffice buildings in moderate climates, Energy and Buildings 27 (3) (1998) 231–237.