lable at ScienceDirect

Building and Environment 46 (2011) 2342e2350

Contents lists avai

Building and Environment

journal homepage: www.elsevier .com/locate/bui ldenv

Thermal sensations of the whole body and head under local cooling and heatingconditions during step-changes between workstation and ambient environment

Quan Jin a,*, Lin Duanmu a, Hui Zhang b, Xiangli Li a, Hongbo Xu a

a Institute of Building Environment and Facility Engineering, Dalian University of Technology, Dalian, Liaoning 116024, PR ChinabCenter for the Built Environment, University of California, Berkeley 94720, USA

a r t i c l e i n f o

Article history:Received 13 February 2011Received in revised form29 April 2011Accepted 18 May 2011

Keywords:Thermal sensationStep-changeNon-uniformLocal ventilationPersonal environment control

* Corresponding author.E-mail address: jq51yyy@gmail.com (Q. Jin).

0360-1323/$ e see front matter � 2011 Elsevier Ltd.doi:10.1016/j.buildenv.2011.05.017

a b s t r a c t

This paper examines people’s thermal sensations during step-changes between ambient and workstationenvironments with a local ventilation device installed to supply-air motion around heads. We conductedhuman subject tests in a controlled environment chamber for summer and winter conditions. We per-formed 29 tests. The ambient air temperatures were 28 and 30 �C for summer conditions and 19 �C forwinter conditions. The local supply-air temperatures were at 24, 28 and 30 �C for summer and 50 �C forwinter. The supply-air velocities of the local ventilation device were at 3, 3.5, and 5 m/s for summer and3.5 m/s for winter. The air temperatures near heads were 26e30 �C for summer and 32 �C for winter. Thevelocities along the jet-flow line at a distance of 10 cm from heads were 1.4e2.6 m/s for summer and1.8 m/s for winter. In total, 23 subjects participated in the tests, and each subject participated in 1w2 testconditions. Both the dynamic and stable thermal sensations of head and whole body were analyzed.When head is cooled by local ventilation, head thermal sensation has an effect on overall thermalsensation. When subjects moved from the workstation, where local devices were installed, to theambient environment that was warmer in summer and colder in winter than the workstation, bothovershooting and hysteresis were found. These thermal sensation changing trends in non-uniform step-change environments are helpful in personalizing environment control designs and exploring thepossibilities of saving energy in buildings.

� 2011 Elsevier Ltd. All rights reserved.

1. Introduction

Step-changes between two environments are common in dailylife, such as walking from the indoors to the outdoors, walking fromthe office into the hallway and getting into and out of a car. Wheninstalling local ventilation device at a workstation (also calledpersonal environmental control systems, PEC systems), there isanother type of step-change that occurs between the workstationand the ambient environment. This paper examines changingpatterns of overall and head thermal sensation during step-changesbetween a workstation at where a local ventilation device isinstalled to blow air toward occupants’ heads, and the ambientenvironment where there is no local ventilation device installed.

There have been a number of studies to examine thermalsensation responses in the step-change process. Gagge [1] in 1967carried out experiments that had subjects move between twouniform environments with a temperature difference of 6w20 �C

All rights reserved.

and proposed the important phenomena of thermal sensation’santicipatory and hysteresis. When the step-change is from cold toneutral or to warm, or from hot to neutral or to cold, anticipatoryoccurs which is probably caused by the sense of comfort that occursbefore the body temperature changes. When step-change is fromneutral to cold or to warm, hysteresis is especially obvious. Later,Wyon [2], Glickman [3] and Nagano [4] also found similarresponses. de Dear [5] studied overall thermal sensation responsesduring up-step and down-step-changes when people movedbetween two twin chambers. Immediate sensations resulting fromthe temperature up-steps (from neutral to slightly warm, fromslightly cool to neutral, and from neutral to neutral) showed a sharpincrease, which approximately equaled the final steady-state value,while initial impressions of temperature down-steps (slightlywarm to slightly cool, slightly warm to neutral, and neutral toslightly cool) overshot the final steady-state responses consider-ably. The author explained that the dynamic component of thethermal sensory system is capable of anticipating the steady-stateresponse to a suddenly warmer environment. While the over-shoot in sudden step-down tests appeared to result from coldthermo receptors being closer to the skin surface than heat thermo

Fig. 1. Experimental facilities a) Ambient environment b) Head local ventilation andfabric devices.

Q. Jin et al. / Building and Environment 46 (2011) 2342e2350 2343

receptors and that cold receptors are more sensitive to the skintemperature changing rate than heat receptors. In Zhang’s disser-tation [6], the quick overshooting appeared in the following step-changes. When face warming is applied, facial thermal sensationdisplayed gradual pattern and overall thermal sensation hada quick overshooting pattern. When the warm stimulation wasremoved, both of the two thermal sensations displayed transientovershooting. When face cooling was applied and removedsuddenly, both thermal sensations showed transient overshooting.When the head (excluding the face) experienced the step-changes,the overshoot did not occur.

A number of studies have been performed on PEC to improvehuman thermal comfort (Melikov et al. [7], Bauman et al. [8], Brookand Parson [9], and Knudsen and Melikov [10]). The air movementenlarges the range of ambient temperatures in which people arecomfortable and improves occupants’ perceived air quality (Bau-man et al. [8] and Arens E et al. [11]). Some studies showed that thehead is an influential body part for whole body thermal comfortand local ventilation on head is effective for cooling (Taniguchi et al.[12], Zhang [13], and Arens et al. [14]). One of the physiologicalreason is that head is a very important body part for heat dissipa-tion of thewhole body. Zhang’s dissertation [6] studied in detail thethermal comfort responses for different body parts and the wholebody. She found that the local comfort of the face, feet, and handspredominate in determining a person’s overall thermal comfort inwarm and cool conditions. Later, she performed subjective exper-iments with a PEC system, which included local cooling on the headand hands by air motion and local warming on the hands and feetby conduction and radiation. The results showed that comfort iswell maintained with the PEC system in the tested environmentaltemperatures of 18w30 �C, and the energy savings of HVAC systemscan be as high as 30%. Zhang Y [15] examined the effects of localexposures of the face, chest, and back on overall thermal sensation,comfort, and acceptability. When the face is cooled by local venti-lation, overall thermal sensation changed significantly. The authorexplained the physiological reason is that the head, which has thearea perfused by a rich superficial vascular supply, acts as a radiator.When face cooling is provided, it improved thermal acceptabilitymore than other body parts and the boundary of the acceptablerange of room temperature can shift from 26 to 30.5 �C with lessthan 20% dissatisfaction. It provides evidence that there isa reasonable increase for the ambient temperature when usinglocal ventilation to cool the face.

From the results summarized above, three things are clear. First,local ventilation on the face could improve thermal comfort andenlarge the ambient temperature range for comfort, so energyconsumptionwill be reduced. Second, the face is an influential bodypart when applying local ventilation, which can produce consid-erable change on overall thermal sensation. Third, thermal sensa-tion appears as overshooting and hysteresis during step-changes.However, therewas no step-change test done between the ambientenvironment and theworkstation, to explain how people feel whenthey step outside their workstations with PEC systems into theambient that does not have a PEC system. In Zhang’s study [6], therewas a sudden stimulus on a local body part, such as head, hand andfeet, which appears to be similar to the step-change from theambient to the workstation. However, when the stimulus isremoved, it is not similar to the step-change from a workstationinto an ambient environment, because the environment change isgradual in Zhang’s study, which is different from the sudden changefrom a workstation to the ambient environment, so the signalsreceived by the thermal receptors are different. Also, the supply-airtemperature in Zhang’s tests was low, which is not realistic in officeenvironments and does not use energy efficiently. The purpose ofthis study is to test human responses during a step-change between

ambient environment and a workstation with a small temperaturedifference between the air from the local supply nozzle and theambient environment.

We should note that different terms are used in literature toexpress the two types of thermal sensation responses after step-changes, anticipatory or overshooting and hysteresis or gradualpattern. In this paper, we use overshooting and hysteresis todescribe the phenomena, which refer to what occurs after step-changes, in which the initial thermal sensation reaches the peakvalue immediately and gradually, respectively.

2. Methods

2.1. Experimental equipment and test conditions

The experiments were carried out in a controlled environmentchamber at Dalian University of Technology. The chamber’s airtemperature and humidity were controlled automatically, and theaccuracy controls were �0.5 �C and �5%. There were two areas,which were the workstation environment and the ambient envi-ronment (see Fig. 1a and b). In the workstation, pictures and bookswere added to create an office atmosphere. The ventilation methodin the chamber had a ceiling supply and floor return. The airsupplied to the chamber ambient is mixed of fresh air and re-circulated air. The air velocity in the ambient area is less than

Table 1Environmental parameters in the ambient environment and the workstation.

Ambient/Supply-airtemperature (�C)

n Ambient radianttemperature (�C)

Air temperaturenear head (�C)

Air temperature nearchest (�C)

Air temperaturenear leg (�C)

Air motion velocity on theline of jet-flow nearhead/outlet (m/s)

30/30 5 30.1 30.3 30.4 30.3 2.4/4.028/28 7 28.3 28.3 28.4 28.3 1.2/2.530/24 4 30.2 27.0 30.3 30.0 2.6/5.028/24 7 28.4 26.5 28.6 28.3 1.8/3.519.5/50 6 19.3 32.0 23.0 20.2 1.8/3.5

Q. Jin et al. / Building and Environment 46 (2011) 2342e23502344

0.1 m/s. The local ventilation device was composed of two nozzleson each side of the workstation to aim air at the head from thesides, and there is a fabric device to block the air-flow from thenozzles to other parts of the body (Fig. 1b). The purpose of the aircoming from the sides was to avoid dry-eye sensation in theoccupants caused by front ventilation. The nozzle directions andthe design of the fabric were determined by preliminary tests.

In this study, there are two methods for the local ventilation airsupply. An air supply with a temperature difference, as 30/24 �Cand 28/24 �C which represent ambient temperature/supply-airtemperature, is provided by a heat pump that sends fresh air intothe local air supply nozzles after cooling or heating. Air supplywithout temperature difference, as 30/30 �C and 28/28 �C, using re-circulated air from the ambient in the chamber, is conducted bya fan installed in the chamber into the air supply nozzles. There aretwo wind volume control valves to adjust supply-air volumes, onein the pipe in the floor plenum and one at the end of the pipe justbefore the connections to the two air supply nozzles. The supply-airvelocity in the outlet is from 2.5 m/s to 5.0 m/s. The supply-airvelocity in the outlet is from 2.5 m/s to 5.0 m/s. Heater control-lers were installed near the two volume control valves to accuratelycontrol the supply-air temperature. The heating wire on the pipeclose to the nozzle and a transformer are used together for a finecontrol of the air temperature. Local ventilation outlets are circleejector nozzles with a diameter of 50 mm that direct the supply-airat the head, not other body parts. The air delivery direction can beadjusted flexibly. The temperatures in outlets of the nozzles andaround subjects are tested by PT1000 thermal resistance and a dataacquisition system, which could be maintained with a precision of0.2 �C. Both a hot ball anemometer and a hot wire anemometertested the air velocities at the outlet of the nozzles and aroundsubjects. The accuracy is within �0.2 m/s.

2.2. Experimental conditions

Twenty-nine tests were carried out in this experiment. Of the 23subjects who participated, 11 were female and 12 were male. Bothhead cooling in summer and head warming in winter were per-formed by supplying cool, isothermal, or warm air around the head.Four sets of air temperature conditions for summer and one set ofair temperature condition for winter were tested. For summer, theambient temperatures were set at 28 �C and 30 �C, and localsupply-air temperatures were set at 28 �C or 30 �C, and 24 �C. Thepurpose of supplying re-circulated air in the chamber was to testthe cooling effect of just using fans without a temperature

Table 2Clothing insulation information [17].

Summer Winter

Bra Panties(tricot)

Sportshirt

Straight longtrousers

Slipper Long underwetop

0.01 0.03 0.17 0.15 0.03 0.20

Total 0.4 Total

difference because it has the potential to save more energy inbuildings. For winter, the room air temperature was set at 19.5 �C,and the supply-air temperature was set at 50 �C. The reason to setthe supplying air temperature high is to reduce the air velocitywhich causes draft sense but provide certain amount of heating tothe head region. Though the air supply temperature is high at 50 �C,after mixed with the ambient air during its travel toward head, theair temperature near head is decreased at 32 �C with a velocity of1.8 m/s on the line of flow jet. During the pilot tests, we found thistemperature acceptable. The humidity was kept at 40% in winterand 60% in summer. The summer tests were conducted from lateAugust to middle October, and the winter tests were conducted inJanuary. During the test period, the average outdoor air tempera-tures in summer and in winter were 24 �C and 4 �C respectively.

In Table 1, the details of the environmental air temperature andair velocity near the head, chest and legs are shown for each testcondition. The air temperature and velocity were tested at a pointat a distance of 10e15 cm away from the body parts, while for thehead it was tested along the line of the jet-flow. The supply-airvolume is small, the velocity attenuated when it arrived on thehead. According to the subject reports, in all tests conditions, therewas no draft sensation. The surrounding air temperatures near thechest and legs were similar to the ambient temperature, whichindicates that the fabric device was effective in directing the air-flow to the head only.

For summer conditions, test subjects were required to wear a T-shirt, thin cotton trousers, and slippers. The clothing resistance isapproximately 0.43 clo. For winter conditions, the participantswere required to wear long underpants, under shirts, thin coats,trousers, cotton socks and slippers. The clothing resistance isapproximately 0.93 clo (see Table 2). This clothing level is typical innewer buildings and is slightly less than the level worn by people inolder buildings. During the tests, subjects were required to do somecomputer work, and the metabolic level was approximately 1.0met [16].

The chamber was a closed space in the basement withoutsunlight coming through the windows and other radiant sources.The radiant temperature was assumed to be the same as the airtemperature.

2.3. Experimental procedure

The participants in this test were chosen from college studentswith a mean age of 24 � 2.5 yr, a mean weight of 59.1 � 8.6 kg, anda mean height of 168.8 � 7.3 cm. They were in good health,

ar Long underwearbottom

Straight Longtrousers

Single-breastedsuit jacket

Cottonsocks

0.15 0.15 0.36 0.03

0.93

Table 3Experiment code for local cooling ventilation in summer.

Experiment code A B C D

Step-change Procedure From workstation to ambient From ambient to workstation From ambient to workstation From workstation to ambientand to workstation

Thermal sensation From neutral to slightlywarm/warm; slightly warm toslightly warm; slightly cool toslightly warm

From warm/slightly warm toneutral; neutral to neutral

From warmer to cooler From cooler to warmer andto cooler

Fig. 2. Overall thermal sensation changing curve (ambient air temperature/localSupply-air temperature in outlet 28/24 �C).

Q. Jin et al. / Building and Environment 46 (2011) 2342e2350 2345

followed a disciplined sleep schedule, and ate normal meals. Theywere not allowed to drink alcoholic or caffeine. No intense exerciseswere allowed before coming to the test. The subjects were paidafter they finished all the test conditions, at 15w20 RMB per hour.

Only one person participates in a test at a time to avoid distur-bance from other people. The test ran from 8 a.m. to 10 a.m. Thesubjects were required to finish breakfast 1 h before participating.Before the tests began, subjects were trained to be familiar with thevoting procedure and questionnaire. However, the details of theexperimental conditions were kept confidential. The subjects wereassigned randomly to the test conditions and one subject partici-pated in 1w2 test conditions.

A continuous seven-point sensation scale was used in our testand covered a range from ‘cold’ to ‘hot’. During each experiment,the survey questionnaires (regarding local and overall thermalsensation) appeared automatically on subjects’ computer screen ina previously defined schedule.

The whole test procedure was separated into three phases:workstation before step-change, ambient environment after step-change, and workstation after step-change. Before the test star-ted, subjects were asked to stay in the chamber for 15w30 min.Then, they experienced three consecutive test conditions: localventilation in the workstation, uniform temperature in the ambientenvironment, and local ventilation in the workstation. Each phasetook approximately 60 min. In order to capture rapid thermalsensation changes during the step-change process, different timeinterval for votingwere designed. In any of the three test phases, forthe first 10 min, votes were recorded every 2 min. During next10 min, the vote interval was 3 min, and for the following 10 min, itwas 4 min. During the last 30 min, subjects voted for thermalsensation only every 5 min. The advantage in increasing the voteinterval was to decrease vote frequency and to avoid mistakes orannoyance with too many questionnaires. After being in theworkstation for 60 min, subjects stepped into the ambient envi-ronment quickly and began to evaluate thermal responses imme-diately. After another 60 min, they were moved back to theworkstation again and began to perceive local ventilation. Duringthe entire experiment, subjects were allowed to read and talk, butthey were forbidden to discuss their thermal responses.

3. Results

In this paper, the thermal sensation votes (TSV) are plotted indetail. To explore the changing trend of thermal sensation betweenworkstation and ambient environments, individual result andaverage result were both analyzed.

3.1. Cooling in summer

As we know, different individuals have different thermalperceptions to thermal environments, even in the same conditions.From the primary results, we found that in the same test conditions(shown in Table 1), individual thermal sensation votes are obviousdifferent. Thermal sensations’ mean value can not reflect the

thermal sensation for the whole participants. The conclusion fromone step-change test condition is only applicable in this case and islimited by the environmental condition. In regard of this, to abstractthe changing trend among various test conditions, we summarizedthermal sensationvotes before andafter step-change and found thatsame thermal sensation changing phenomenahappened referred tothe thermal sensations. It was not limited by the environmentalconditions. Thus, experiment conditions are further classified bythermal sensation in the workstation and ambient. See Table 3. Theanalysis of the results presented below follows these classifications(presented in experiment code A, B, C, and D).

3.1.1. Overall thermal sensation’s changing trendFig. 2 shows individualoverall thermal sensationcurves following

the time sequences (from the workstation to the ambient environ-ment and to the workstation again) in the environmental conditionof 28/24 �C listed inTable 1. The labels for the x-axis, 55.5/56min and111/112/114 min represent the initial voting time after the subjectsstepped into the ambient environment or the workstation.

From the graphs shown in Fig. 2, some trends can be observedfrom individual results. All samples except for ID30 showed over-shooting after stepping into the workstation. ID3, ID11, ID30 andID27 showed hysteresis after stepping into the ambient environ-ment, and ID9, ID31 and ID11 showed overshooting in the ambientenvironment.

Table 4 shows the initial and steady mean thermal sensationvotes during step-changes, and the differences just before and afterstep-changes for four experiment codes (A, B, C, D). To verify thesignificant difference of the mean value of thermal sensation in aninitial state and a steady-state, corresponding paired sample t-testswere performed. These results could help explore changing trendsin thermal sensation after step-changes and the instant coolingeffect on thermal sensation. In Table 4, we found the following:

1) Hysteresis. In experiment code A, (from neutral workstation toslightly warm/warm ambient environment, from slightly warm

Table 4Overall thermal sensation changing value in different experiment codes.

Experimentcode

Samplequantity

Workstationenvironmentbefore step-change

Ambient environment Workstationenvironment afterstep-change

Changing value DTSV Statistic T-testfor DTSV

n Steady value (T1) Initialvalue (T2)

Steadyvalue (T3)

Initialvalue (T4)

Steadyvalue (T5)

T5eT4 T4eT3 T4eT1 T3eT2 P value

A 9 e 0.51 1.30 e e e e e 0.79 0.003B 15 e e e �0.46 �0.03 0.43 e e e 0.013C 21 e e 0.64 �0.38 e e 1.02 e e 0.001D 21 0.33 e e �0.38 e e e 0.71 e 0.025

Q. Jin et al. / Building and Environment 46 (2011) 2342e23502346

to slightly warm or from slightly cool to slightly warm), overallthermal sensation (average of nine people) became warmergradually. From the average initial thermal sensation of 0.51 tothe average steady thermal sensation of 1.3, there is a 0.79 scaleunit increase. This pattern is statistically significant(p value < 0.003). It indicates that the mean thermal sensationof the 9 subjects’ votes displayed hysteresis phenomena afterthe step-change.

2) Overshoot. In code B, which is from a warm/slightly warmambient environment to a neutralworkstation or fromneutral toneutral, overshooting phenomena was found. The initial sensa-tion overshot at �0.46, and gradually went back to �0.03(average of 15 subjects). The statistic test shows that p value<0.05.

3) Impact of head cooling on overall thermal sensation. In code C,after the step-change from a warm ambient to a cool work-station, the mean thermal sensation of all 21 subjects changedone scale unit. Some subjects changed nearly two scale units.The statistical analysis shows that this change is significant (pvalue<0.001). It indicates that local ventilation on the head hasa strong instant effect on overall thermal sensation.

4) Instant cooling lower thermal sensation. In code D, afterreturning to the workstation, the overall thermal sensation islower than that before the step-change from theworkstation tothe ambient environment. Comparing steady thermal sensa-tion values in the workstation before the step-change with theinitial thermal sensation values after stepping back to theworkstation for all 21 subjects’ votes, there was a significantdecrease in overall thermal sensation (p value <0.05). It showsthat in the same workstation environment, the sensation iscooler after people experienced warm ambient environments.

5) For the two cooling methods of 30/30 �C and 28/28 �C with re-circulated air and 30/24 �C, 28/24 �C, with a temperaturedifference, overall thermal sensation was neutral (see Table 5,column2). Themethodof re-circulatedairmotionby fancanalsosatisfy the cooling requirement in tested conditions. In column4,from the thermal sensation changing value, it was found thatafter stepping into the ambient environment, the condition at30/24 �C had a bigger value than the condition at 30/30 �C. Thecondition at 28 �C had similar results. Also, the condition at 28 �Chad bigger changing values than 30 �C. It indicates that the re-

Table 5Overall thermal sensation changing values in different experiment conditions.

Condition Workstationbefore step-changeSteady TSV

Ambient afterstep-changeInitial TSV

DTSV

30/30 �C (re-circulating room air) 0.30 0.47 0.1730/24 �C(temperature difference) 0.26 0.59 0.3328/28 �C (re-circulating room air) 0.23 0.53 0.2028/24 �C(temperature difference) �0.27 0.44 0.71

circulated air method has a smaller increase on thermal sensa-tion after the step-change into an ambient environment withrespect to the temperature difference method.

3.1.2. Overall thermal sensation and head thermal sensationHead and overall thermal sensation votes in the workstation

when local ventilation is supplied and in the ambient environmentare displayed in the scatter graphs in Fig. 3 with correlation coef-ficients and residual standard variations.

From the results, we found that:

1) Using all the votes, the coefficient shows that there is a goodcorrelation (R ¼ 0.7) between head and whole body in theworkstation and ambient when local ventilation is applied tothe head (Fig. 3a). When the environment was set at 28 �C(Fig. 3d, e) rather than 30 �C (Fig. 3b and c), or the air supplytemperature is lowat 24 �C in 28 �C condition, the correlation issignificant. In a higher ambient temperature, head is not able todissipate most heat enough for the whole body, and even thehead is cooled, the thermal condition for the whole body maynot change much. Thus, in 30 �C ambient environment, thecooler the head, the more difference between the head and thewhole body. From Table 6 column 7, after stepping into theworkstation, the initial overall thermal sensation showsa significant instant decrease.

2) Stepping from the workstation to ambient environment at thesame ambient air temperatures (28 �C and 30 �C), people feelthe head is warmer when supply-air temperatures are lower inthe workstation (28/24 �C and 30/24 �C) than in the conditionswhere supply-air temperatures are higher (30/30 �C, 28/28 �C)(see steady votes in workstation and initial votes for theambient environment in Table 6).

3) Head thermal sensations show overshooting when steppingfrom the ambient environment to a workstation where a localventilation is applied (initial votes were cooler than the steadyvotes, see Table 6).

3.2. Warming in winter

The following graphs (Fig. 4) show overall thermal sensationchanging trends when subjects moved between a workstation andthe ambient environment. In the graph, ID17 (55.5/78) means (theminute ID17 stepped into the ambient/the time ID 17 stepped intothe workstation). In the workstation with warming air from thenozzles aimed toward the head at 50 �C, people felt neutral orslightly warm. After a step-change to the slightly cool or coolambient environment (19.5 �C), overall thermal sensation dis-played hysteresis. The initial neutral or slightly cool thermalsensation gradually became cool (see Table 7). This was differentfrom the trend when people do the step-change between twoenvironments with a large temperature difference (6w20 �C) [11].

Fig. 3. Overall thermal sensation versus head thermal sensation a) Condition 30/30 �C, 30/24 �C, 28/28 �C, 28/24 �C b) Condition 30/30 �C c) Condition 30/24 �C d) Condition 28/28 �C e) Condition 28/24 �C.

Q. Jin et al. / Building and Environment 46 (2011) 2342e2350 2347

Although the thermal sensation was also from neutral or warmenvironment to cool environment in that study, the votes showedan obvious overshooting pattern. The statistical analysis for thisstudy shows that the difference between initial value and steadyvalue was significant in the ambient environment after a step-change (p value <0.05, Fig. 5).

4. Discussion

4.1. Comparison to other studies

This paper studied the step-change between an ambient envi-ronment and a workstation with a PEC system, comparing it withother studies on step-changes between two uniform environmentswith a large temperature difference [1] [5] or on step-changes

Table 6Head and overall thermal sensations changing values during step-changes between the

Condition Workstation before step-change Ambient after step-change

Head steady value Head initial value Overall ste

30/30 �C �0.23 0.17 0.5330/24 �C �1.02 0.87 0.5828/28 �C �0.55 0.12 0.6928/24 �C �0.53 0.67 0.75

before and after local air motion was supplied at the workstation[6] [15]. See Table 8.

First, in this paper, compared with uniform test results byothers, overshooting occurs when subjects stepped from a warmerambient environment into a neutral workstation, a conditionincluded in Gagge and De Dear’s studies in uniform environmentalconditions. In this paper, hysteresis occured when subjects steppedfrom a neutral workstation to a non-neutral ambient environment,which is the same as Gagge’s results, but different from de Dear’s inthe condition from neutral to slightly cool. In de Dear’s result, it isovershooting, while, in this paper, it is hysteresis. One reason mightbe that with local ventilation on the head, when stepping intocooler ambient, the warm thermal sensation maintained on thehead has a big effect on overall thermal sensation, which sloweroverall thermal sensation to change to be cool immediately. Heat is

workstation and the ambient environment.

Workstation after step-change

ady value Head initial value Head steady value Overall initial value

�0.91 �0.58 �0.71�0.83 �0.83 0.28�0.59 �0.51 �0.20�1.19 �0.70 �0.74

Fig. 4. Overall thermal sensation changing curve in the condition of 19.5/50 �C with local warming ventilation.

Q. Jin et al. / Building and Environment 46 (2011) 2342e23502348

transported by the blood flow between each tissue inside humanbody [18]. About 15%e20% of blood flow coming from the heartgoes to the brain [19]. There are rich facial blood vessels. These allhelp the head to dissipate body heat efficiently. After a step-changeto a cooler ambient, blood vessels are vasoconstrictor and the heattransported to the head from other body parts is decreased grad-ually. Thus, the whole body will not feel cool immediately. Someliterature [6,12,13] also showed that head is a very importantexposed body part to the whole body and has a significant effect onoverall thermal sensation. Secondly, compared with the local airsupplying and removing study [6] [15], the results are different,which is due to the different change signals to thermo receptorfrom the environment, one is sudden environment temperaturechange when step into ambient in this paper and the other isgradual environment temperature change when supply or removeair motion. Warm and cold thermo receptors in the skin are used tosense the thermal environment [20]. When a change of thermoreceptor temperature occurs, the receptor is strongly stimulatedand the frequency of its impulses increases which is sent to thebrain. The brain interprets the signal from thermo receptors tocreate thermal sensation [21]. The thermal environment can bedrifts and ramps-change, step-change or cycling change. In step-change study, Gagge [1] found that thermal sensation changeimmediately with the change of air temperature while there wasa delay before skin temperature. In Zhang’s [6] local cooling andheating study, thermal sensation can be quantified well from skintemperatures. Thermo receptors detect different environmentchange signals and may give different patterns of impulse. Whenthe research is on the step-change between a workstation and theambient environment, the test design is very different from local airmotion’s supplying and removing as in Zhang’s study [6]. Thirdly, inthis study, the temperature difference of step-change environ-ments is the smallest comparing to other studies, 0w3 �C insummer, but the overshooting and hysteresis changing values arebig, 0.5 and 0.8 scale units, compared with the 0w1.0 scale unit in

Table 7Overall thermal sensation changing value in ambient environment in wintercondition.

1 2 3 4 5 6

Ambient initial value �2.07 0 �0.59 0 0.22 �0.65Ambient steady value �2.39 �1.87 �1.38 �0.97 �2.16 �1.38

the 20 �C temperature difference condition. Also, in winter, therewas a 1.1 value change in the scale unit, and a 0w0.5 scale unitchange even in the 20 �C temperature difference condition.Therefore, obvious overshooting and hysteresis can happen in thehead cooling, with a slight temperature difference or even withouttemperature difference, because head can mostly maintain a warmsensation with a step-change into a cool ambient environment inwinter, and quickly perceive a cool sensation when stepping intoa cool workstation in summer. Finally, this obvious hysteresis in anambient environment and overshooting in aworkstation for certainthermal conditions (experiment codes A and B in Table 3), areadvantageous for energy savings in ambient and workstationenvironments, and should be considered during air conditioningdesigns to adjust the air temperature set points for different func-tional zones.

4.2. Re-circulated local ventilation method

In this paper, two local ventilation methods were designed,respectively, to supply-air motion with temperature difference anduse re-circulated air motion in the chamber. The design of the re-circulated air considered energy savings from cooling the airsupply and convenient utilization. The cooling effect on overallthermal sensation is as good as the temperature difference method,in which overall thermal sensation can be neutral (see Table 5).

Fig. 5. Statistical t-test for mean TSV difference of steady and initial state in ambientenvironment after step-change in winter condition.

Table 8Results comparison on step-change studies.

Gagge A.P [1]. de Dear R [5]. Zhang H [6]. Zhang Y.F [15]. This paper

Step-change pattern Two uniformenvironments

Two uniformenvironments

Before and after localair supply inworkstation

Before and afterlocal air supplyin workstation

Workstation andambientenvironments

1967 1993 2003 2007 2010Temperaturedifference6e20 �C

Temperature difference2.3e7.8 �C

Local warming andcooling on face

Local cooling on face,temperature difference13 �C between ambientand center of face

Local cooling andwarming on head,temperaturedifference 0e3 �Ccooling and 12 �Cwarming betweenambient and headsurrounding

Overshooting conditions(overall thermal sensation)

From cold to neutralor to warm,hot to neutralor to cold

From slightly warm toslightly cool, slightlywarm to neutral,neutral to slightly cool

Face warming orcooling applied,face warming orcooling removed

e From warm orslightly warmambient to neutralworkstation, fromneutral to neutral

Overshooting value(overall thermal sensation)

1.0 0 or 0.5 e e 0.5

Hysteresis conditions(overall thermal sensation)

From neutral to coldor to warm

From neutral to slightlywarm, from slightlycool to neutral,from neutral to neutral

e From slightlywarm to neutral

From neutralworkstation to slightlywarm or warm ambient,from slightly warmto slightly warm, fromslightly cool to slightlywarm, from neutral orslightly warm workstationto slightly cool or coolambient

Hysteresis value(overall thermal sensation)

<0.5 0 or 0.5 e 0.3 0.8 or 1.1

Q. Jin et al. / Building and Environment 46 (2011) 2342e2350 2349

Additionally, with the high air supply velocity, there is no draftsensation for the subjects, which was the result of the draft surveyduring the tests. Further, by comparing overall thermal sensationchange after a step-change into an ambient environment, there isan interesting phenomenon that occurs in the re-circulated airsupply method. Fig. 6 shows that the re-circulated method hasa smaller increase on thermal sensation after a step into theambient environment with respect to the temperature differencemethod (see Table 5). One reason may be the cooling intensity. Inhigh cooling intensity conditions, like 28/24 �C or 30/24 �C, aftersubjects have adapted to, their expectations are also heightened forthe ambient environment. In other words, a strong intense airconditioning environment will weaken occupants’ adaptivecapacity to the thermal environment. Thus, to save energy andthermal adaptability, we should use a re-circulated air supply asmuch as possible.

Fig. 6. Overall thermal sensation change between steady-state in workstation beforestep-change and initial state in ambient after step-change for re-circulated methodand temperature difference air supply method (ambient temperature/air supplytemperature in outlet).

4.3. Correlations of overall and head thermal sensations

This study suggests that overall thermal sensation and headthermal sensation are highly correlatedwhen there is cooling on thehead in the workstation and a step-change into an ambient envi-ronment (see Fig. 3a). In Table 6, the initial thermal sensation voteafter stepping into the workstation shows a significant instantdecrease, which indicates that the method for cooling the head iseffective in cooling the whole body. This result correlates withearlier studies [6,12e15] that conclude that the head is an influentialbody part for applying local ventilation. Simultaneously, we foundthat the correlation coefficient, the R-value, is different in differentconditions. The correlation was much higher for lower ambientconditions. A higher correlation appears with a method of usinga temperature difference in the air supply. It indicates that ina particular ambient environment, themore cooling intensity on thehead, the higher the correlation between the whole body and thehead. Whereas in a higher ambient condition, the temperaturedifference method didn’t show high correlation between the headand overall thermal sensation, whichmay be caused by the fact thatcooling intensityon theheaddidnot greatlyaffect thewholebodyorotherbodyparts surroundedbyhighambient temperature, affectingthe whole body more than the head in this strong non-uniformenvironment. It seems that in a very high ambient environment,cooling on multi body parts should be considered in later studies.

5. Conclusions

This paper examined thermal sensation changes on human testsubjects during step-changes between an ambient environmentand a workstation with local ventilation installed to supply-airmotion around people’s heads. The study focused on a smalltemperature difference between the ambient and the supply-airfrom the local ventilation device. Cooling in summer and

Q. Jin et al. / Building and Environment 46 (2011) 2342e23502350

warming in winter by the local ventilation device were analyzedseparately. From the analysis on the thermal sensation votes, thefollowing conclusions were drawn and based on them futurestudies are proposed.

1) With local ventilation supply-air around 5 �C lower than theambient air temperature, during step-changes, from neutralworkstation to slightly warm/warm ambient environment,from slightly warm to slightly warm or from slightly cool toslightly warm, overall thermal sensation displayed hysteresis.From a warm/slightly warm ambient environment to a neutralworkstation or from neutral to neutral, overall thermal sensa-tion showed overshooting. After step-changes from theambient environment to the workstation, overall thermalsensation could change by nearly two scales units.

2) With local ventilation supplying warming air in the worksta-tion during a step-change, from neutral or slightly warmworkstation to a slightly cool or cool ambient environment,overall thermal sensation displayed hysteresis, which isdifferent from the overshooting that happened in the step-change between two uniform environments with a bigtemperature difference.

3) Overall thermal sensation correlated very well with headthermal sensation during the entire step-change processbetween the workstation and the ambient environment. Fromthe analysis of the initial thermal sensation vote after steppinginto the workstation, overall thermal sensation showeda significant instant decrease, which indicates that the head hasstrong effect on overall thermal sensation.

4) The two methods for local cooling, re-circulated air supply andtemperature difference air supply can make subjects feelneutral in the workstation. The difference in overall sensationwas negligible and might be caused by the mixing of thesupply-air with the ambient air in workstation. Therefore,there is no obvious advantage to supply cooler air here. On theother hand, when stepping into the warm ambient environ-ment, both overall and head thermal sensation increased whenusing the air temperature difference method.

5) After returning to the cool workstation, overall thermalsensation is lower than before the step-change from worksta-tion to ambient environment. Even in the same workstationenvironment, the sensation was cooler after people experi-enced warm ambient environments.

6) Some preliminary suggestions for workstation local ventilationcan be provided based on the above analysis and discussions.Head cooling is a goodchoice to cool thewholebody. The supply-air direction, coming from two sides toward head providescooling but avoids dry-eye discomfort. When the air supplytemperature near head is 27 �C, 2.6 m/s air velocity will notinducedraft discomfort for theparticipants.When thesupply-airtemperature near head is 32 �C in the workstation, 1.8 m/s airvelocity gives a sense of warm air motion for the participants.

Acknowledgments

The authors are grateful to: the National Natural Science Foun-dation of China (Project No. 50678030 and Project No. 51078052),to Institute of Building Environment and Equipment Engineering,Dalian University of Technology and to the Center for the BuiltEnvironment, UC, Berkeley.

References

[1] Gagge AP, Stolwijk JAJ, Hardy JD. Comfort and thermal sensations and asso-ciated physiological responses at various ambient temperatures. Environ-mental Research 1967;1:1e20.

[2] Wyon DP. Methodology for indoor environmental research. ICIE symposium,Copenhagen, Denmark 2001.

[3] Glickman N, Inoye T, Keeton RW, Callen IR, Hick FK, Fahnestock MK. Physio-logical adjustment of normal subjects and cordial patients to sudden changein environment. ASHRAE Transactions 1949;55:27e38.

[4] Nagano K, Takaki A, Hirakawa M, Tochihara Y. Effects of ambient temperaturesteps on thermal comfort requirements. International Journal of Biometeo-rology 2005;50:33e9.

[5] de Dear R, Ring JW, Fanger PO. Thermal sensations resulting from suddenambient temperature changes. Indoor Air 1993;3:181e92.

[6] Zhang H. Human thermal sensation and comfort in transient and non-uniformthermal environments. Ph.D. thesis, UC Berkeley, 2003.

[7] Melikov AK, Arakelian RS, Halkjaer L, Fanger PO. Spot cooling part 1:human responses to cooling with air jets. ASHRAE Transaction 1994;100(2):476e99.

[8] Bauman FS, Carter TG, Baughman AV, Arens EA. Field study of the impact ofa desktop task/ambient conditioning system in office buildings. ASHRAETransaction 1998;104:1153e71.

[9] Brook JE, Parsons KC. An ergonomics investigation into human thermalcomfort using an automobile seat heated with encapsulated carbonized fabric.Ergonomics 1999;42(5):661e73.

[10] Knudsen GL, Melikov AK. Human response to individually controlled envi-ronment. In: Proceedings of indoor air 2005, 4e9 September 2005, Beijing,paper no. 1.4-02.

[11] Arens E, Zhang H, Kim D, Buchberger E, Bauman F, Huizenga C, et al. Impact ofa task-ambient ventilation system on perceived air quality. Copenhagen,Denmark: Indoor Air; August 15e20, 2008.

[12] Taniguchi Y, Aoki H, Fujikake K, Tanaka H, Kitada M. Study on car air condi-tioning systemcontrolled by car occupants’ skin temperatures - Part 1: researchon amethod of quantitative evaluation of car occupants’ thermal sensations byskin temperatures. SAE Technical Paper Series 1992, 920169: 13e19.

[13] Zhang H, Arens E, Kim D, Buchberger E, Bauman F, Huizenga C. Comfort,perceived air quality, and work performance in a low-power taskeambientconditioning system. Building and Environment 2010;45(1):29e39.

[14] Arens E, Zhang H, Huizenga C. Partial- and whole body thermal sensation andcomfort, part I: uniform environmental conditions. Journal of Thermal Biology2006;31:53e9.

[15] Zhang Y, Zhao R. Effect of local exposure on human responses. Building andEnvironment 2007;42:2737e45.

[16] ASHRAE. ASHRAE standard 55-2004 (Thermal environmental conditions forhuman occupancy). Atlanta: ASHRAE, Inc; 2004.

[17] McCullough EA, Jones BW, Huck J. A comprehensive data base for estimatingclothing insulation. ASHRAE Transactions 1985;91(2):29e47.

[18] Williams BA, Shitzer A. Modular liquid-cooled helmet liner for thermalcomfort. Aerospace Medicine 1974;45(9):1030e6.

[19] Kandel ER, Schwartz JH, Jessell TM. Principles of neural science. Appleton &Lange; 2000.

[20] Bazett C, McGlone B. Experiments on the mechanism of stimulation ofendorgans for cold. America Journal of Physiology 1930;93:632.

[21] Benzinger TH, Pratt A, Kitzinger C. The thermostatic control of human meta-bolic heat production. Proceedings of the National Academy of Sciences of theUnited States of America 1961;47:730.