air-conditioner use effects on thermal environment in...

Proceedingsof9thWindsorConference:Making Comfort Relevant CumberlandLodge,Windsor,UK,7-10April2016.NetworkforComfortandEnergyUseinBuildings,http://nceub.org.uk

Air-ConditionerUseEffectsonThermalEnvironmentinBedroomsandSleepQualityduringSummer–AnalysisofUniversityStudentsinOsaka

N.Umemiya*,Y.Tachibana*,Y.Nakayama*,J.Arai**andT.Kobayashi*

*OsakaCityUniversity,**DaikinIndustries,Ltd.

AbstractAnalysesofthermalenvironment,thermalcontrolbehaviors,thermalsensations,andsleepqualityyieldedthefollowing results indicating air conditioner (AC) use effects on sleep quality in 11 bedrooms during threeperiodsofsummer.1)OutsidetemperatureswerehigheronACusedaysbutroomtemperaturesduringsleepwerekeptalmostidentical(25.7°C).2)SleepqualityasevaluatedbytheOSAscorewashigheronACusedays.However,itwashotterandlessacceptableonACusedays,althoughthermalcomfortwasnotdifferent.3)OnAC non-use days, the relation between thermal comfort and thermal sensation was stronger. Thermallyacceptablerangeswerewideronnon-usedays.4)RelationbetweentheOSAscoreandthedegreeofsoundsleepwasweakeronACusedays.TheOSAscorewasmorestronglyrelatedtothermalacceptabilityonACusedaysthannon-usedays.5)Differences insleepqualityweresmallerbetweenACusedaysandnon-usedaysthanbetweenthesurveyedperiods.Lowersleepqualityinmid-summerwascausednotbyAC,butbyhigherroomtemperatures.

Keywords:SleepQuality,ThermalSensation,AirConditioneruse

1 IntroductionSultry nights with temperatures that remain higher than 25°C have become increasinglycommon in theurbanareaofOsaka, Japan,perhapsbecauseof globalwarmingandheatislandphenomena.Therewere23,27,10,50,57,and55sultrynights,respectively,in1960,1970,1980,1990,2000,and2010.

Common households possess 2–3 air-conditioners (AC), using them during sleep times.Decreasedsleepqualitycausedbyheatmightaffectworkperformanceandhealth.Recently,increased heat stroke risk has because become amatter of greater concern. AC use hasbecomeaheatstroke countermeasure.However, electrical power shortages related to thesevere accident at the Fukushima Daiichi nuclear power plant caused by the Tohokuearthquake on March 11, 2011 possibly affected AC use. A government campaign of‘electricityconservation’,COOLBIZ,callsforanACtemperaturesettingof28°C.

Sakaneetal.(2012)conductedaquestionnairesurveyofACuseby362residentsinOsakaapartmenthousesinautumn2012.Resultsshowthat24.6%answered‘veryfrequentlyuse’AC and29.7%answered ‘frequently use’ACon a four-point scale in daytime. In addition,36.5% answered ‘very frequently use’ AC; 30.7% answered that they ‘frequently use’ ACduring nighttime. However, 54.4% and 29.0% answered ‘keep open’ in daytime andnighttime on a five-point scale; 22.5% and 38.4% answered ‘keep closed’ windows indaytimeandnighttime.ResultsshowthatpeoplereportedchoosingACuseaccordingtoadailypreference.

Windsor Conference 2016 - Making Comfort Relevant - Proceedings 292 of 1332

In such circumstances, this study objects to clarify the followings. 1) How are indoor andoutdoorthermalenvironmentdifferentbetweenAC-useandnon-usedays?2)Howaresleepquality, thermal sensation, comfort, and acceptability improved by AC use? 3) How therelations between sleep quality and thermal senations different in between AC use andnon-usedays?4)Aretherelationsdifferentamongperiodsinsummer?

2 MethodsElevenuniversitystudentswholivedinornearOsakacityparticipatedinthesurveyduringJune 18–25 in early summer, August 27 through September 5 in mid-summer, andSeptember 17–24 in late summer. Each survey was continued for more than sevensuccessive days during the university summer vacation season. For that reason, some ofthemwentoutsidetheirordinarybedroomsbecauseoftravelorreturnedhomeortowork;therebyeachbecameunabletoattendtheentiresurvey.Table1presentsattributesofthesubjectsandthebedrooms.

Thesubjectsmeasuredthebedroomtemperatureatintervalsof10minusingdataloggers.Relative humidity was recorded before and after sleep. Every morning, they filledquestionnaire sheets related to the degree of sound sleep in four categories, thermalsensation in seven-point-scale, thermal comfort in four-point-scale, and acceptability inthree-point-scaleinbedroomsofthepriornight.SleepqualitywasalsoevaluatedusingOSAscalesinmid-summerandlatesummer.Atthesametime,theykeptadiaryrecordingtheirabsenceorpresenceinbedrooms,personalACuse,windowopeningofthebedroomsandsleepatintervalsof30min.

Table1Attributeofthesubjectsandthebedrooms

earlysummer

middlesummer

latesummer

no.1 ○ ○ ○ boarder apartment n.a. tatamimat 7.9no.2 ○ △ ○ commuter apartment 15 bed 6.4no.3 ○ ○ ○ commuter detached 12 bed 6.4no.4 ○ ○ ○ boarder apartment 13 bed 8.6no.5 ○ △ ○ commuter detached 12 bed 8.5no.6 ○ × ○ commuter detached 9 tatamimat 8.8no.7 ○ × × commuter detached 9 tatamimat 10.2no.8 ○ ○ ○ commuter detached 12 bed 8.1no.9 ○ × ○ boarder apartment 12 tatamimat 8.5no.10 × × ○ commuter detached n.a. bed 7.0no.11 ○ × × commuter apartment n.a. n.a. 4.7

SubjectBuildingtype

Beddingtype

Sleepingtime(hrs.)

Attendanceinthesurvey Boarderorcommuter

Bedroomarea(m2)

Ogurietal. (1985)developedtheOSAsleep inventorytoevaluatesubjectivesleepquality.Yamamotoetal.(1999)revisedit.Generally,OSAisusedforclinicalpurposesinJapan.Itisastandarddeviationcalculatedfromsubjectiveratingscalesofsleepquality.Itevaluatessleepat the timeofawakening,basedon responses to16questions. Eachquestion response isgivenonafour-responsescale:verygood,somewhatgood,somewhatbad,verybad.Thesequestionsyieldstandardscoresoffivefactors.AveragedstandardscoresoffivefactorsaredefinedastheOSAscore.Thehigherthescoreis,thehigherthesleepqualityisindicated.The five factors are Factor I (Drowsiness when waking such as ‘I have the power ofconcentration’,‘Ifeelasenseofliberation’and‘Ifeelclear-headed’),FactorII(Fallingasleepandmaintaining sleep such as ’Iwas able to sleep soundly’, ‘I dozed off until I finally fell


asleep’, ‘I got to sleep easily’, ‘I oftenwoke up from sleep’ and ‘The sleepwas shallow’),Factor III (Dreaming, suchas ‘Ihadmanynightmares’and ‘Ihadmanydreams’),Factor IV(Fatigue recovery such as ‘Fatigue persists afterwaking up’, ‘I feel languorous’ and ‘I feelunwell’),andFactorV(Sleepingdurationsuchas ‘Ihavegenerallygoodappetite’and’Thesleepdurationwaslong’).

3 Results3.1 Comparisonamongperiodsinsummer3.1.1 ThermalenvironmentAnexampleofbedroomtemperatureandhumidity,outdoortemperatureandhumidity,andbehaviors inmid-summer for subject no.1 inmid-summer is presented in Fig. 1.We useoutdoor temperature measured by local weather observatories. The respondent wentoutsidethebedroomonthesixthandseventhnight,sothesurveywascontinueduntiltheninth day. He lived with his family in an RC apartment. He sometimes used AC, andsometimes opened windows. Indoor temperatures were sometimes lower than outdoortemperaturesindaytimebutusuallyhigherthanoutdoortemperaturesinnighttime.

Fig.2presentsthemeandaily indoorandoutdoortemperaturesforthesurveyedperiods.Themean indoor temperatures (during sleep)were 26.5°C (25.7°C ), 27.7 °C (27.2°C) and25.7°C(25.4°C)inearly,middleandlatesummer.Meanoutdoortemperatures(duringsleep)were24.1°C(21.8°C),26.0°C(23.3°C)and22.6°C(20.2°C)inearly,middleandlatesummer.

303540455055606570

05

101520253035

08/2700

:00

08/2800

:00

08/2900

:00

08/3000

:00

08/3100

:00

09/0100

:00

09/0200

:00

09/0300

:00

09/0400

:00

09/0500

:00

humidity[％

]

tempe

rature[℃

]

Time

roomtemp.outdoortemp.outdoorhumidityACuse

stayinbedroomsleep

windowopening

Fig.1Anexampleoftherecordsofthermalenvironmentandthermalcontrolbehaviors

Fig.2Dailymeantemperaturesforthreeperiods Fig.3Timeratioofbehaviorsfortheperiods

202224262830

2014/06/19

2014/06/20

2014/06/21

2014/06/22

2014/06/23

2014/06/24

2014/06/25

2014/08/27

2014/08/28

2014/08/29

2014/08/30

2014/08/31

2014/09/01

2014/09/02

2014/09/03

2014/09/17

2014/09/18

2014/09/19

2014/09/20

2014/09/21

2014/09/22

2014/09/23

2014/09/24

2014/09/25

temp.[℃

]

date

roomtemp.outdoortemp.

early summer midsummer latesummer

15.5

8.2

30.6

28.9

38.8

15.9

49.8

32.5

22.8

2.6

55.5

30.2

0 10 20 30 40 50 60

ACuseofthesubject

ACuseinbedroom

windowopening

sleeping

timeratio (%]

latesummermidsummerearlysummer


Fig.4Frequencydistributionofthermalsensation,comfortandacceptabilityforperiods

3.1.2 BehaviorsFig.3showstheaveragetimeratioofsleep,windowopeningandACuse in thebedroomand AC use of the respondents for 24 hours for different periods. These time ratios arebased on records from diaries of respondents kept at 30min intervals. The time ratio ofpresence inbedroomwas51.2%.Respondents rarely stayed inbedroomsduringdaytime.Thetimeratioofsleepwasnotdifferentamongperiods.Thetimeratioofwindowopeninginbedroomswas30.6%inearlysummer,49.8%inmid-summerand55.5%inlatesummer.Windowsofbedroomsweresometimeskeptopenduringabsence.ThetimeratioofACuseinbedroomswas8.2%inearlysummer,15.9%formid-summerand2.6%for latesummer.RespondentsdidnotusuallyuseACanddidnotopenwindowsinearlysummer.TheACusedaysduringsleepwere11daysinearlysummer,15daysinmid-summer,and4daysinlatesummer. The AC non-use days during sleep were 54 days in early summer, 27 days inmid-summer,and48daysinlatesummer.

3.1.3 ThermalsensationFig. 4 shows a frequency distribution of thermal sensation, comfort, and acceptability indifferent periods. Results show that, ‘slightly hot’ was most frequent (48.6%) in earlysummer;‘cool’wasmostfrequent(30.6%)inmid-summer;and‘neutral’wasmostfrequent(42.3%)inlatesummer(inuniformitytestamongperiods,p<.0001).Itwashottestinearlysummerandmostcoolinmid-summer.’Slightlyuncomfortable’wasmostfrequent(59.5%)in early summer; ‘comfortable’ was most frequent (59.2%) in mid-summer, but 18.9%responded ‘uncomfortable’; ‘comfortable’ was most frequent (63.5%) in late summer(p=.007).‘Comfortable’wastheleastinearlysummerand‘uncomfortable’wasthemostinlatesummer.‘Acceptable’wasmostfrequentlyreportedinallperiodsand‘Acceptable’was91.6%inlatesummerand‘sometimesnotacceptable’was39.8%Inmid-summer.Buttherewerelittlestatisticallysignificantdifferencebyperiodsinthermalacceptability(p=.092).

Fig. 5 portrays the ratio of ‘comfortable’ and ‘acceptable’ in each category of thermalsensation indifferentperiods.Relationsbetween thermal sensationsand thermal comfortchangedwiththeperiod.Theratiobecamehigherasthethermalsensationbecamewarmerinearlysummer.Itwasunrelatedtothermalsensationinmid-summer.Itbecamelowerasthethermalsensationbecamecoolerinlatesummer.

1.4

5.335.6

48.6

9.1

30.028.6

22.2

19.1

2.2

22.923.0

42.3

6.92.7

0 20 40 60

cold

cool

sl.cool

neutral

sl.hot

hot

veryhot

frequency[%]

Thermalsensationlatesummermidsummerearlysummer

32.5

59.5

8.0

59.2

21.9

18.9

63.5

28.2

8.2

0 20 40 60 80

comfortable

sluncomf.

uncomf.

frequency[%]

Comfort latesummermidsummerearlysummer

65.8

29.6

4.6

60.2

39.8

0.0

91.6

8.4

0.0

0 20 40 60 80 100

acceptable

sometimesnotaccept.

notaccept.

frequency[%]

Acceptability latesummermidsummerearlysummer


a)‘comfortable’ b)‘acceptable’

Fig.5Comparisonsofratiosof‘comfortable’and‘acceptable’ineachcategoryofthermalsensationamongperiods

Relations between thermal sensation and thermal acceptability changedwith the period.Theratiobecamehigheras itbecamehotter inearlysummer. Itwasunrelatedtothermalsensations in mid-summer. It became lower as the thermal sensation became cooler orwarmerinlatesummer.

3.1.4 SleepFig. 6 portrays frequency distributions of the degree of sound sleep for periods. Thefrequencydistributionwassimilarforearlyandlatesummer.Actually,‘somewhatgood’wasthemost frequent response forbothperiods: 47.8% for early summerand45.7% for latesummer.Formid-summer,‘somewhatgood’wasthemostfrequent.Itwas44.2%,but‘sleepsomewhatbad’was30.0%.Thedegreeofsoundsleepwassimilarinearlyandlatesummer(inuniformitytestamongperiods,p=.22).

Fig.7presentsOSAscorescomparedbetweenmid-summerandlatesummerineachsubject.Eachscorerepresentsthemeanofsevennights.Theaveragescorewas45.5inmid-summerand 48.1 in late summer. TheOSA scorewas lower inmid-summer than in late summer,exceptforsubjectno.3.

54.7

91.3

60.6

90.5

60.8

13.6

53.1

29.6

0 20 40 60 80 100

cold

cool

slcool

neutral

slhot

hot

veryhot

ratioof'comfortable'[%]


13.3

55.9

87.5

77.8

85.8

54.4

76.8

44.7

86.1

81.3

0.0

0 20 40 60 80 100

cold

cool

slcool

neutral

slhot

hot

veryhot

ratioof'acceptable'[%]


40.8

47.8

8.3

3.1

23.6

44.2

30.0

2.2

38.8

45.7

14.1

1.4

0 20 40 60

verygood

somewhatgood

somewhatbad

verybad

frequency[%]


40

45

50

55

no.1

no.2

no.3

no.4

no.5

no.6

no.8

no.9

no.10

OSAscore

Subjectmidsummer45.5latesummer48.1

midsummer latesummer

p=.22

Fig.6Frequencydistributionofdegreeofsleepforperiods

Fig.7ComparisonofmeanOSAscoresineachsubjectbetweenperiods


Fig. 8 presents a comparison of profiles of the mean vote of each OSA scale betweenmid-summerandlatesummer.ScalesareorderedinthefigurefromfactorItoV.Thelowervalueshowsbettersleep.Completesleepwasbetter in latesummerthan inmid-summer.MeanvotesofFactorI(Drowsinesswhenwaking),FactorIV(Fatiguerecovery)andFactorV(Sleeping duration) were higher in late summer than in mid-summer, although fewdifferenceswerefoundinFactorII(Fallingasleepandmaintainingsleep),andnodifferencesinFactorIII(Dreaming)betweenmiddleandlatesummer.DifferencesinliberationinFactorI,languorousinFactorIVandgoodappetiteinFactorVweresignificantint-test.

Fig.9showsthemeanOSAscoresforrespectivecategoriesofthermalsensation,comfort,acceptability, anddegreeof sound sleep formiddle and late summer. TheOSA scorewasalmost identical foreachcategoryof thermalcomfortboth inmiddleand latesummer (inKruskal-Wallistest,p=.57and.61,respectively).TheOSAscorewashigherfor‘comfortable’inmid-summer(p=.022),butnodifferencewasapparentinlatesummer(p=.37).Differencesof mean OSA scores were not 3.1 in mid-summer, but 1.6 in late summer between‘acceptable’and‘sometimesnotacceptable’(p=.14and.54,respectively).

Fig.10showsthemeanbedroomtemperatureduringsleepforeachcategoryofdegreeofsoundsleepforperiods.TheOSAscorebecamelowerfrom‘verygood’to‘somewhatbad’inthedegreeofsoundsleepinthemiddleandlatesummer(p=.002and.0006).

Fig.11presentsthemeanbedroomtemperatureduringsleepforeachcategoryofdegreeof

1

2

3

4

I.con

centratio

n

I.liberation

I.clear-heade

d

II.soun

dly

II.do

zed

II.gottosleep

easily

II.oftenwokeup

II.shallow

III.nightmares

III.manydreams

IV.fatigue

IV.languo

rous

IV.fe

elunw

ell

V.good

ape

tite

V.longduration

verybad

somewhatbad

somewhatgood

verygood

midsummer latesummer

Fig.8ComparisonofprofileofOSAscalesbetweenperiods

Fig.9ComparisonofmeanOSAscoresforeachcategoryofthermalsensation,comfortandacceptabilitybetweenperiods

45.9

45.9

45.3

43.4

46.5

49.2

46.3

48.2

48.351.2

30 35 40 45 50 55 60

cold

cool

slcool

neutral

slhot

hot

OSAscorelatesummermidsummer

47.6

42.7

42.6

48.1

47.2

51.3

30 35 40 45 50 55 60

comfortable

sluncomf.

uncomf.

OSAscorelatesummermidsummer

46.6

43.5

48.4

46.8

30 35 40 45 50 55 60

acceptable

sometimesnot

acceptable

OSAScorelatesummermidsummer

** **

*

** p<1%* p<5%

Factorsandscales


soundsleepforearly,middle,andlatesummer.Thedegreeofsoundsleepwasunrelatedtothebedroom temperature in early summer (p=.50). Themeanbedroom temperaturewas26.1°C for ‘very good’, 27.6°C for ‘somewhat good’, and ‘somewhat bad’ in mid-summer(p=.26).Itwas24.2°Cfor‘verygood’,25.2°Cfor‘somewhatgood’,and25.4°Cfor‘somewhatbad’ in late summer (p=.071). Differences of the temperature between ‘very good’ and‘somewhat good’ were larger than that between ‘somewhat good’ and ‘somewhat bad’.Roomtemperaturewasrelatedtothedegreeofsoundsleepin latesummer,althoughtherelationwasweakinmid-summer.

Fig.12depictsrelationsbetweenthemeanbedroomtemperatureduringsleepandtheOSAscore for middle and late summer. The OSA score became slightly higher for the lowerbedroom temperature in mid-summer (r=-.22), although little correlation was foundbetweenOSAscoreandroomtemperatureinlatesummer(r=-.09).

3.2 ComparisonbetweenACusedaysandnon-usedaysTheACnon-usedayisdefinedasthedayonwhichrespondentsdidnotrecordACuseinthediaryatallduringsleep.However,theACusedaysincludefromdaysofACusethroughthenighttodaysofACuseforonly30min.

3.2.1 ThermalenvironmentFig.13showsthedailymeanoutdoortemperatureandmeanoutdoortemperatureduringsleepforACuseandnon-usedays.Thedailymeanoutdoortemperatureswere24.6°Cand23.4 °C for use and non-use days (in t-test, p=.005).Mean outdoor temperatures during

y=-1.1006x+76.627R²=0.0503

30

35

40

45

50

55

60

20 22 24 26 28 30

OSAScore

roomtemp.duringsleep[deg.C]

a)midsummer

y=-0.3134x+55.917R²=0.007830

35

40

45

50

55

60

20 22 24 26 28 30

OSAScore


b)latesummer

Fig.12RelationsbetweenroomtemperatureduringsleepandOSAscoreformid-summerandlatesummer

25.9

25.4

26.1

25.7

26.1

27.6

27.6

28.5

24.2

25.2

25.4

26.2

24 25 26 27 28 29

verygood

somewhatgood

somewhatbad

verybad


latesummermidsummerearlysummer49.4

46.7

41.2

30.6

51.0

48.3

41.0

43.9

30 35 40 45 50 55

verygood

somewhatgood

somewhatbad

verybad

OSAscore

latesummermidsummer

Fig.10ComparisonofmeanOSAscoresforeachcategoryofsoundsleepbetweenperiods

Fig.11Comparisonofmeanroomtemp.foreachcategoryofsoundsleepbetweenperiods


sleepwere22.6°Cand21.4°Cforuseandnon-usedays(p=.002).Differencesbetweenuseand non-use days were 1.2 K for both daily temperature and temperature during sleep.However, p-values showed that subjects chose to use AC according to the outdoortemperatureduringsleep.

Fig.14presentsacomparisonofmeanbedroomtemperatureduringsleepbetweenuseandnon-use days for three periods.Differences betweenuse and non-use dayswere 1.2 K inearly summer, 1.1 K in mid-summer and 0.9 K in late summer. Slight differences wereapparent inbedroomtemperatureduringsleepbetweenuseandnon-usedays.Themeanbedroom temperature during sleep was 26.4°C (25.7°C) and 26.3°C (25.7°C) for use andnon-usedaysfortotalthreeperiods.

3.2.2 BehaviorsFig.15showsaveragetimeratiosofsleep,windowopening,andACuseinbedroomsandACuse of the respondents for AC use and non-use days. The time ratio of sleep was notdifferentbetweenuseandnon-usedays.Timeratiosofwindowopeninginbedroomswere24.5%forusedaysand47.8%fornon-usedays.Windowswerekeptclosedforabouthalfofthedayevenfornon-usedays.ThetimeratioofACuseoftherespondentswas18.4%.ThetimeratioofACuseinbedroomswas4.7%fornon-usedays.However,theACuseratioandAC use ratio in bedrooms were similar for use days. Respondents used AC outside thebedroomsfornon-usedays.

3.2.3 ThermalsensationFig.16showsafrequencydistributionofthermalsensation,comfortandacceptabilityforACuseandnon-usedays.Resultsshowthat‘slightlyhot’wasthemostfrequent(23.4%)forusedays,although‘neutral’wasthemostfrequent(38.3%)fornon-usedays(inuniformitytestbetween AC use and non-use, p=.19). Themean bedroom temperature during sleepwasalmostidenticalforuseandnon-usedaysasnotedin3.2.1,butrespondentsfelthotterforusedays,although‘cool’and‘neutral’werethesecondmostfrequent.

Actually, ’slightly uncomfortable’ was most frequent (43.3%) for use days, although‘comfortable’ was the most frequent (51.9%) for non-use days. Little difference wasapparentinthermalcomfortbetweenuseandnon-usedays(p=.80).

For use days, ‘acceptable’ was 53.3%; ‘sometimes not acceptable’ was 43.3%. Actually,‘acceptable’was75.6%and‘sometimesnot’was22.7%fornon-usedays.Forbothuseand

Fig.13Comparisonofmeanoutdoor Fig.14Comparisonofmeanroom Fig.15Comparisonoftimeratio temp.betweenACuseand temp.betweenACuseand ofbehaviorsbetweenAC useandnon-usedays non-usedays non-usedays

25.9

27.7

24.8

24.7

26.8

25.7

25.7

27.3

24.9

24 25 26 27 28

earlysummer

midsummer

latesummer


summerｔotalACusedaysnon-usedays

23.4

21.4

24.6

22.6

20 21 22 23 24 25

outdoortemp.

outdoortemp.during

sleep

temp.[deg.C]

ACusedaysnon-usedays

18.4

4.7

47.8

32.2

44.8

55.0

24.5

33.3

0 20 40 60

ACuse

ACuseinbedroom

windowopening

sleeping

timeratio[%]ACusedaysnon-usedays


non-usedays, ‘acceptable’was themost frequent,but less frequent forusedays than fornon-usedays(p=.031).

Fig.16Frequencydistributionofthermalsensation,comfortandacceptabilityforACuseandnon-usedays

Fig. 17 presents ratios of ‘comfortable’ and ‘acceptable’ for each category of thermalsensationforuseandnon-usedays.Theratioof‘comfortable’washigherin‘cool’, ‘slightlycool’, and ‘neutral’ forusedays than fornon-usedays. The ratioof ‘comfortable’was thehighestin‘slightlycool’fornon-usedays,althoughtheratioin‘neutral’wassimilarforuseandnon-usedays.Theratioof‘acceptable’washigherin‘slightlycool’,‘neutral’and‘slightlyhot’fornon-usedaysthanusedays.Figuresshowthatusedaysweremorecomfortablebutlessacceptablethannon-usedays,irrespectiveofthermalsensation.

3.2.4 SleepFig.18displaysthefrequencydistributionofthedegreeofsoundsleepforACusedaysandnon-usedays.Actually,‘verygood’was50%;‘somewhatgood’was40%forusedays.Resultsshow that ‘very good’was 32.6% and ‘somewhat good’was 46.5% for non-use days. Thedegreeofsoundsleeptendedtobebetterforusedaysthanfornon-usedays(inuniformitytestbetweenACuseandnon-use,p=.238).

Fig.19presentsOSAscorescomparedbetweenusedaysandnon-usedaysineachsubject.TheOSAscorewasbetterforusedaysthanfornon-usedaysinnos.1,3,6and9.However,nos.2and6weretheopposite.

a) ‘comfortable’ b) ‘acceptable’

Fig.17Comparisonsofratiosof‘comfortable’and‘acceptable’ineachcategoryofthermalsensation

betweenACuseandnon-usedays

ComparisonofRatios of ‘comfortable’ and ‘acceptable’ inthermalsensationcategoriesforACuseandnon-usedays

0.814.1

20.338.3

23.43.1

0.0

20.7

13.8

20.734.5

10.3

0.0

0 10 20 30 40

cold

cool

slcool

neutral

slhot

hot

veryhot

frequency[%]

Thermalsensation ACusedaysnon-usedays

51.9

41.1

7.0

46.7

43.3

10.0

0 20 40 60

comfortable

sluncomf.

uncomf.

frequency[%]

Comfort ACusedaysnon-usedays

76.6

22.7

0.8

53.3

43.3

3.3

0 20 40 60 80

acceptable

sometimesnotaccept.

notaccept.

frequency[%]

Acceptability ACusedaysnon-usedays

66.7 90.0

0.0

55.6 80.8

63.3 13.3

0.0

0 20 40 60 80 100

cold

cool

slcool

neutral

slhot

hot

veryhot

ratioof'comfortable'[%]

non-usedaysACusedays

66.7

75.0 83.3

30.0 0.0

0.0

61.1

91.7 53.3

0.0

0 20 40 60 80 100

cold

cool

slcool

neutral

slhot

hot

veryhot

ratioof'acceptable'[%]

non-usedaysACusedays


Fig.20showsmeanOSAscoresforuseandnon-usedaysformid-summerandlatesummer.ThemeanOSAscorewas47.9 inmid-summerand50.2 in latesummerforusedays.Theywere 44.1 inmid-summer and 48.0 in late summer for non-use days.Difference ofmeanscores and p-values of t-tests between use and non-use days were 3.9 and 0.023 inmid-summer,and2.3and0.39inlatesummer.TheOSAscorewashigherforusedaysthanfornon-usedaysinmid-summer.However, littledifferencewasapparentbetweenuseandnon-usedaysinlatesummer.

Fig.21presentsacomparisonofprofilesofmeanvoteofeachOSAscalebetweenuseandnon-usedays. Sleep forusedayswasbetter thannon-usedays.Concentration inFactor I,soundly in Factor II, many dreams in Factor III, fatigue in Factor IV and good appetite inFactorVwerehigherforusedays.Littletendencywasapparentamongfactors.

3.2.5 RelationbetweenthermalenvironmentandsleepFig.22showsacomparisonofthemeanOSAscoreofeachcategoryofthermalsensation,comfortandacceptabilitybetweenACuseandnon-usedays.OSAscoreslightlycorrelatedtothermalsensationonlyforusedays(p=.0659).OSAscorebecamehigherforcoolersensationfor use days, but no relation was found between OSA score and thermal sensation. OSAscoredidnot so relate to thermalcomfort forbothuseandnon-usedays.ThemeanOSAscore for ‘acceptable’was 52.0 and ‘slightly acceptable’was 45.0 for usedays.OSA scorewashighwhenthermalacceptabilitywashighforusedays(p=.0062).However,littlerelationexistedbetweenOSAscoreandthermalacceptabilityfornon-usedays(p=.18).OSArelatedtothermalsensationandacceptability,butnot tothermalcomfort forusedays.OSA littlerelatedtothermalsensation,comfortandacceptabilityfornon-usedays.

1

2

3

4

I.con

centratio

n

I.liberation

I.clear-heade

d

II.soun

dly

II.do

zed

II.gottosleep

easily

II.oftenwokeup

II.shallow

III.nightmares

III.manydreams

IV.fatigue

IV.languo

rous

IV.fe

elunw

ell

V.good

ape

tite

V.longduration

verybad

somewhatbad

somewhatgood

verygood

ACusedays ACnon-usedays

Fig.20ComparisonofmeanOSAscoresfor Fig.21ComparisonofprofilesofOSAscalesperiodsbetweenACuseandnon-usedays betweenACuseandnon-usedays

Fig.18Frequencydistributionofdegreeof Fig.19ComparisonofmeanOSAscoresineachsoundsleepforACuseandnon-usedays subjectbetweenACuseandnon-usedays

40

45

50

55

no.1

no.2

no.3

no.4

no.5

no.6

no.8

no.9

no.10

OSAscore

SubjectACusedays47.1non-usedays47.2

ACusedays non-usedays

44.1

47.9

46.6

48.0

50.2

48.5

40 45 50 55

midsummer

latesummer

summertotal

OSAScoreACusedaysnon-usedays

32.6

46.5

17.1

3.9

50.0

40.0

10.0

0.0

0 20 40 60

verygood

somewhatgood

somewhatbad

verybad

frequency[%]

ACusedaysnon-usedays

* *

Factorsandscales

* p<5%


Fig.23 shows themeanOSAscoresof respectivecategoriesof thedegreeof soundsleepbetweenuseandnon-usedays.TheOSAscorewasunrelatedtothedegreeofsoundsleepforusedays(p=.058).However,themeanOSAscorefor‘sleepsoundly’was50.4,for‘sleepsomewhatsoundly’was47.4,for‘notsleepsomewhatsoundly’was41.7andfor‘notsleepsoundly’was37.3fornon-usedays.TheOSAscorewasrelatedtothedegreeofsoundsleepfornon-usedays(p=.0001).ThedegreeofsoundsleepisonescaleofOSAFactorII.FactorIIwas higher inOSA for non-usedays andOSA scorewas lowerwhen thedegreeof soundsleep was low for non-use days, although factors other than Factor II became dominantwhenthedegreeofsoundsleepwashighforusedays.

Fig.22ComparisonsofmeanOSAscoresforeachcategoryofthermalsensations,comfortandacceptabilitybetweenACuseandnon-usedays

Fig.23ComparisonofmeanOSAscoresforeach Fig.24ComparisonofmeanroomtemperaturesforcategoryofsoundsleepbetweenACuse eachcategoryofsoundsleepbetweenACuse

andnon-usedays andnon-usedays

Fig.25RelationsofroomtemperatureduringsleepandOSAscoreforACuseandnon-usedays

y=-0.3614x+59.734R²=0.0151

30

40

50

60

20 22 24 26 28 30

OSAscore


a)ACusedays

y=-0.7845x+67.058R²=0.065330

40

50

60

20 22 24 26 28 30

OSAscore


b)non-usedays

46.546.7

45.8

47.5

44.6

51.1

50.1

44.044.9

51.2

40 45 50 55

cold

cool

slcool

neutral

slhot

hot


47.3

44.7

49.9

51.1

45.3

45.4

40 45 50 55

comfortable

sluncomf.

uncomf.


47.1

44.6

52.0

45.0

40 45 50 55

acceptable

sometimesnotacceptable


50.4

47.4

41.7

37.3

50.8

49.1

36.3

35 40 45 50 55

verygood

somewhatgood

somewhatbad

verybad

OSAScoreACusenon-use

25.2

25.7

26.4

26.3

25.5

25.8

26.9

24 25 26 27 28 29

verygood

somewhatgood

somewhatbad

verybad


ACusenon-use


Fig. 24 shows themeanbedroom temperatureduring sleep for each categoryofdegreeofsound sleep for use and non-use days. Degree of sound sleep was unrelated to bedroomtemperatureforACusedays(p=.68).However,themeanbedroomtemperaturewas25.2for‘sleepsoundly’,25.7for‘sleepsomewhatsoundly’,26.4for‘notsleepsomewhatsoundly’,and26.3for‘notsleepsoundly’fornon-usedays(p=.064).Therelationwasnotsignificant,butthedegreeofsoundsleeptendedtobelowwhenthebedroomtemperaturewashighinnon-usedays.AsFig.25shows,thecoefficientofcorrelationbetweenthemeanbedroomtemperatureduring sleep andOSA scorewas -0.26 for non-use days,where little correlationwas foundbetween roomtemperatureand theOSAscore forusedays. These figures show that roomtemperaturesdidnotrelatetothedegreeofsoundsleeporsleepqualityforusedays.

4 Discussion4.1 ComparisonamongperiodsMean room temperatures were 27.2°C for mid-summer and 25.4°C for late summer.However,‘cool’wasmostfrequentformid-summer;itfeltcooler.Therelativefrequenciesof‘comfortable’ responses were around 60% for mid-summer and late summer. However,mid-summer was more uncomfortable because about 20% felt ‘uncomfortable’ formid-summer. ‘acceptable’responseswereabout60%formid-summerandmorethan90%forlatesummer.Itcanbesaidthatitwascoolerformid-summerthanforlatesummer,butmid-summerwaslesscomfortableandlessacceptable.

The ratios of ‘comfortable’ and ‘acceptable’ were not different by thermal sensation formid-summer. Thermal comfort and acceptability did not dependon thermal sensations inmid-summer.

Regarding the degree of sound sleep ‘somewhat good’ was reported by 40–50% ofrespondentsforbothperiods,butthefrequencyof‘slightlybad’was30%formid-summer.ThemeanOSAscorewas45.5formid-summerand48.1forlatesummer.MeanOSAscoresforrespondentswerelowerformid-summerthanforlatesummer.Itcanbesaidthatsleepqualitywaslowerformid-summerthanforlatesummer.ScoresofFactorIII(dreaming)andFactor II (maintenance)weresimilar,butscoresofFactor I (wakingup),Factor IV(fatigue),and Factor V (duration) were lower for mid-summer. The difference of the OSA scorespresumablyresultedfromdifferencesofFactorI,IV,andVscores.

TheOSAscoreswerehigherfor‘comfortable’formid-summer.ThermalcomfortwasrelatedtoOSAscoreformid-summer,althoughOSAscoreswerenotdifferentbythermalcomfortfor late summer.Thermalacceptabilitywas related toOSAscore formid-summerbecausethe difference in OSA score for ‘acceptable’ and ‘slightly unacceptable’ was larger formid-summerthanforlatesummer.Thermalcomfortandacceptabilitywererelatedtosleepqualityformid-summer.

RoomtemperaturewasrelatedtotheOSAscoreordegreeofsoundsleepformid-summer,butnotrelatedforlatesummer.TheOSAscorewasrelatedtothedegreeofsoundsleepforbothmid-summerand late summer.RelationsbetweenOSAscoresand thermal sensationwereweakforbothmid-summerandlatesummer.

4.2 ComparisonbetweenAir-conditionerusedaysandnon-usedaysThe room temperature during sleep was 25.7°C for both AC use days and non-use days,although outdoor temperatures during sleepwere 23.4°C for AC use days and 21.4°C fornon-use days. ‘slightly hot’ was 23.4% andmost frequent on use days and ‘neutral’ was


38.3% and most frequent. Use days were significantly hotter than non-use days. Nodifferencewas found in thermalcomfortbetweenuseandnon-usedays. ‘acceptable’was53.3%onusedaysand75.6%onnon-usedays.Usedaysweremorethermallyacceptablethannon-usedayswere.Itcanbesaidthatthermalcomfortwasnotdifferentbetweenuseandnon-usedays,butnon-usedayswerecoolerandmoreacceptable.

Theratioof‘comfortable’wasnotdifferentbythermalsensationonusedays.Ontheotherhand,theratioof ‘slightlycool’wasthehighestandrelatedmoretothermalsensationonnon-usedays thanonusedays. The ratioof ‘acceptable’ onnon-usedayswas about twotimes higher than on use days. It was hotter but more comfortable. The range ofacceptabilitywaswideronnon-usedays.

Regarding the degree of sound sleep, ‘very good’ was 50% on use days and 32.6% onnon-usedays.MeanOSAscoreswere48.3onusedaysand46.0onnon-usedays.Thescoreonusedayswas slightlyhigher than thatonnon-usedays.Thedifferenceofmeanscoresamong respondents was not significant. Little difference was found between use andnon-use days for late summer, although themean score on use dayswas higher than onnon-usedaysformid-summer.LittledifferencewasfoundinOSAfactorscoresbetweenuseand non-use days. It can be said that the difference of sleep quality was not so greatbetweenuseandnon-usedays.

The OSA score differed by thermal sensation and comfort on use days, although littledifference was found in the OSA score by different thermal sensation, comfort andacceptability. The relation between OSA score and thermal sensation, comfort andacceptabilitywasweakeronusedaysthanonnon-usedays.

OSAscorewasrelatedtodegreeofsoundsleeptightlyonnon-usedays,butslightlyrelatedonusedays.RoomtemperaturewasnotrelatedtoOSAscoresonuse-days,althoughOSAwaslowerforhigherroomtemperatureonnon-usedays.

4.3 Air-ConditionerUseEffectsonSleepQualityACusedayswere15inmid-summer;non-usedayswere27.Inlatesummer,theACusedayswere4;non-usedayswere48.

Itwascooler inmid-summerthan in latesummer inspiteofhigherroomtemperatures inmid-summer.Air-conditionerusemightreducetemperatures,butroomswereslightlycooleron AC non-use days than on use days. AC use did not cause the coolness. It was ratherslightlyhotteronusedaysthanonnon-usedays,althoughroomtemperatureswerealmostidentical on use days and non-use days. Thewide acceptable range on non-use dayswasrathernotable.Itwasslightlyhot,butcomfortableonnon-usedays.

Sleepqualitywasworseformid-summerthanforlatesummer,andworseonnon-usedaysthanonusedays.However,thedifferencebetweenusedaysandnon-usedayswassmallerthan that between mid-summer and late summer. Not AC use, but higher roomtemperaturesinmid-summercausedlowersleepquality.

The OSA score was related to the degree of sound sleep both for mid-summer and latesummer.However,therelationwasweakonusedays.TherelationbetweentheOSAscoreandthermalacceptabilitywasweakbothformid-summerandforlatesummer,buttheOSAscorewasrelatedtoacceptabilityonusedays.Presumably,ACusecausedtheweakrelationbetween the OSA score and the degree of sleep quality. Furthermore, AC use caused astrongrelationbetweentheOSAscoreandthermalacceptability.


5 ConclusionsA survey of the thermal environment of bedrooms, sleep quality, and thermal sensationswereconductedinsummerduringsevensuccessivedaysforthreeperiodsfor11bedrooms.TheACusedaysandnon-usedayswerecomparedrelative to thermalenvironment,sleepquality, and thermal sensations consideringdifferencesofperiods.NumberofACuseandnon-usedaysduringsleepinearly,middle,andlatesummerwere,respectively,11,15,and4,and54,27,and48.

1) OutsidetemperatureswerehigheronACusedaysbutroomtemperaturesduringsleepwerekeptalmostidentical(25.7°C)forACusedaysandnon-usedays.

2) SleepqualityasevaluatedbytheOSAscorewashigheronACusedays.However,itwashotter and less acceptable on AC use days than on non-use days, although thermalcomfortwasnotreportedasdifferentbetweenACuseandnon-usedays.

3) On AC use days, the relation between thermal comfort and thermal sensation wasweaker.ThermallyacceptablerangeswerenarroweronACusedays.

4) Differences in sleepqualitywere smallerbetweenACusedaysandnon-usedays thanbetweenmid-summerandlatesummer.Lowersleepqualityinmid-summerwascausednotbyAC,butbyhigherroomtemperatures.

5) TheOSAscorewasrelatedtothedegreeofsoundsleepbothformid-summerandlatesummer.However,therelationwasweakonACusedays.RelationsbetweenOSAscoreand thermal acceptabilitywereweak both formid-summer and late summer, but theOSAscorewasrelatedtotheacceptabilityonACusedays.

ReferencesJapanMeteorologicalAgency,2016.Latestweather,http://www.data.jma.go.jpMinistryoftheEnvironment,GovernmentofJapan,2016.Launchofanewclimatechangecampaign,http://www.env.go.jpOguri.Metal.,1985.DevelopmentofOSAsleepqualityinventory–Statisticalconstructionand standardization of rating scale for evaluation of subjective sleep sensation. ClinicalPsychiatry,27(7),pp791-798.Sakane. T et al., 2013. Relation among sleep, thermal comfort and thermal control useduring summer in 2012. Proceedings of annualmeeting of Kinki branch of the Society ofHeating,Air-ConditioningSanitaryEngineersofJapan,vol.42,pp173-176.Yamamoto. Y et al., 1999. Standardization of revised version of OSA sleep inventory formiddleageandaged.BrainScienceandMentalDisorders,10(4),pp401-409.


air-conditioner use effects on thermal environment in...

Documents