- Email: [email protected]
Thermal comfort and behavioural strategies in office buildings located in a hot-arid climate
Journal of Thermal Biology 26 (2001) 409–414
Thermal comfort and behavioural strategies in office buildings located in a hot-arid climate Krzysztof Cenaa,*, Richard de Dearb a
Department of Environmental Science, Murdoch University, Perth 6150, Australia Environmental and Life Sciences, Macquarie University, Sydney 2109, Australia
b
Abstract The effects of indoor climates on thermal perceptions and adaptive behaviour of office workers during a large field study in Kalgoorlie-Boulder, located in a hot-arid region of Western Australia, are discussed. Clothing insulation levels were 0.5 clo in summer and 0.7 in winter. Thermal neutrality, according to responses on the American Society of Heating, Refrigerating and Air-Conditioning Engineers seven-point sensation scale, occurred at 20.31C in winter and at 23.31C in summer. The effect of hot-dry/cool-dry seasonality on thermal comfort responses of office workers was significant. Future research into how the overcooling of office buildings in hot-dry climates can be reduced is called for. r 2000 Elsevier Science Ltd. All rights reserved. Keywords: Thermal comfort; Hot-arid climate; Clothing insulation; Office occupants; Behavioural strategies
1. Introduction It is now widely accepted that the previously used climate chambers fail to provide the participating humans with so called ‘‘experiential realism’’ in determining thermal comfort. Cleaning-up and simplification of comfort problems to make them suitable for chamber research unfortunately removes many of the contextual factors that we now recognise as part of the thermal comfort matrix, like competing demands for attention, expectations and attitudes. Field studies of thermal comfort do not suffer from this flaw because they are conducted in actual buildings under normal conditions of occupancy, and involve much larger and diverse samples of ‘‘real’’ occupants as opposed to ‘‘paid college-age subjects’’ (Cena, 1994). This positive change has been facilitated by a gradual development of laboratory-grade field instruments and an enhanced set of protocols that, together, make comfort data obtained in the field every bit as rigorous as the climate chamber counterpart. Field studies also allow for analyses of *Corresponding author. Tel.: +61-8-9360-2883; fax: +61-89310-4997. E-mail address: [email protected] (K. Cena).
other factors than those that can be simulated in chambers, as the subjects provide responses in their everyday habitats, wearing their everyday clothing and behaving without any additional restrictions. American Society of Heating, Refrigerating and AirConditioning Engineers (ASHRAE) has commissioned a series of state-of-the-art, fully compatible field experiments across a variety of climate zones including: temperate (Schiller et al., 1988), hot-humid (de Dear and Fountain, 1994) and cold (Donnini et al., 1996). The current paper is based upon the fourth (final) instalment in this series conducted by Cena and de Dear (1999) in a hot-arid climate in Kalgoorlie-Boulder in Western Australia. There have been relatively few comparable studies conducted in hot-arid locations. These include those by Baker and Standeven (1994) in residential buildings during the summer in Athens, Greece. Their results indicated the importance of adaptive opportunity in order for building occupants to accept and achieve comfort in temperatures warmer than 241C. Abdelrahman (1990) measured thermal comfort in a few classrooms in a Saudi university and concluded that his instruments’ comfort algorithms, based on Fanger’s method, were valid (Fanger, 1970). Saeed (1993)
0306-4565/01/$ - see front matter r 2000 Elsevier Science Ltd. All rights reserved. PII: S 0 3 0 6 - 4 5 6 5 ( 0 1 ) 0 0 0 5 2 - 3
410
K. Cena, R. de Dear / Journal of Thermal Biology 26 (2001) 409–414
conducted field experiments in Riyadh, Saudi Arabia. Again, statistical analyses of their results indicated that Fanger’s equation did well at predicting subjective responses of this sample. A subsequent study by Saeed (1996) of indoor climate on 525 subjects exposed to hot conditions in a mosque confirmed his earlier conclusion. Unfortunately for the middle-eastern studies above, we have no indication of the subjects’ actual level of acclimatization to the hot-dry setting. Given the pervasive nature of refrigerated air-conditioning in such cheap-energy countries one suspects that acclimatization would be negligible. In such situations the adaptive model of thermal comfort (de Dear and Brager, 1998) predicts that the subjects become adapted to the indoor climates they spend most of their lives in, not their outdoor climatic setting. So the close agreement with Fanger’s PMV model is not surprising since it (or a close relative of it at least) has been used to actually design and operate the air conditioning systems. The present paper will discuss the main results of a large field study (Cena and de Dear, 1999) conducted in a hot-arid climate and will focus on the effects of indoor climates on thermal perceptions and adaptive behaviour of office workers.
2. Methods Kalgoorlie-Boulder is situated in the desert area of Western Australia. Rainfall is minimal (only 258 mm per annum, over 300 days in a year are without any rainfall) with a pronounced seasonal variation. Mean minimum temperatures during winter and summer sample periods were 9.61C and 16.71C, respectively. Mean maximum temperatures of the winter and summer sample periods were 18.51C and 30.71C, respectively. Over the summer study period, the mean 9 a.m. relative humidity (RH) and the average daily maximum temperature were 39% and 30.71C, respectively. Twenty-two of the largest office buildings in Kalgoorlie-Boulder were chosen for the study. Sample sizes of 640 and 589 subjects were achieved in winter and summer surveys, respectively. This total of 1229 sets of data was provided by 935 respondents, of whom 294 were interviewed in both seasons. Forty-eight percent of the sample was female, and the average age of all subjects was 35 years. A mobile measurement system (Fig. 1) was used onsite in Kalgoorlie to collect measurements of the indoor atmospheric environment. The system collected concurrent physical data (air temperature, dew-point temperature, RH, globe temperature, radiant asymmetry and air velocity) from three arrays of transducers placed 0.1, 0.6, and 1.1 m above floor level, representing the immediate environment of the seated subject’s ankles, body and neck, respectively.
Fig. 1. Mobile and portable system used for measurements of indoor climates. It incorporates three arrays of sensors positioned at 0.1, 0.6 and 1.1 m and allows for automated recording of air temperature, dew-point temperature, RH, globe temperature, radiant asymmetry, and air velocity.
The questionnaires (approved by Murdoch University’s Human Ethics Committee) covered several areas including demographics, work area satisfaction, personal environmental control, job satisfaction and health. The questionnaires also included the traditional scales of thermal sensation and thermal preference, current clothing garment and metabolic activity checklists, and two scales focusing on air movement perceptions. The thermal sensation scale was the ASHRAE seven-point scale of warmth ranging from cold (3) to hot (+3) with neutral (0) in the middle. The thermal preference scale asked on a three-point scale whether the respondent would like a change in the thermal environment. Possible responses were ‘‘want warmer’’, ‘‘want no change’’, or ‘‘want cooler’’. Clothing garment checklists were adapted to the regional customs prevailing in Kalgoorlie-Boulder and compiled from the extensive lists published in ASHRAE Standard 55-92. Metabolic rates were assessed by a checklist of office activities and were based on the detailed databases published in ASHRAE Standard 55-92 and ISO 7730. Various indices were calculated using software by Fountain
411
K. Cena, R. de Dear / Journal of Thermal Biology 26 (2001) 409–414
and Huizenga (1996) and included operative temperature ðto Þ; mean radiant temperature, effective temperature ðETÞ; predicted mean vote (PMV), predicted percentage dissatisfied (PPD), TSENS, along with the PD index (draft risk) incorporated into ANSI/ASHRAE Standard 55-1992 (ASHRAE, 1992). All index calculations requiring a clo value input included an estimate of the thermal insulation increment resulting from subjects being seated.
3. Results Intrinsic clothing ensemble insulation was estimated on the basis of garment values published in ANSI/ ASHRAE Standard 55-1992 and was very close to the Standard’s assumed summer value of 0.5 clo, but with females being about 0.1 clo lower than the males (Table 1). The winter season’s ensemble average was about 0.2 clo higher than the summer’s. Concomitant with this seasonal difference was approximately a 50% decrease in the inter-individual variability (standard deviation) from 0.23 clo in the winter season to 0.14 clo in the summer season. This suggests that the degree of freedom for thermoregulation and accommodation of interindividual differences in thermal preferences by clothing adjustments diminished as the amount of clothing approached the socially acceptable minimum. Estimated average chair insulation of 0.15 clo lifted the subjects’ thermal insulation values to about 0.8 and 0.6 clo in the winter and summer surveys, respectively. Metabolic rates of the subjects were estimated to be, on average, 77 W/m2 or 1.3 met in both seasons and for both sexes, representative of sedentary activity. Mean height of the subjects was about 171 cm and their mean weight was about 75 kg. Table 2 provides statistical summaries of the indoors climatic measurements for the Kalgoorlie winter and summer season samples, respectively. These data represent the main thermal variables recorded at subjects’ workstations by the mobile system. Mean air and radiant temperatures (averaged across the three heights 0.1, 0.6 and 1.1 m) generally fell within the 22–261C interval for both seasons. Average RH fell within the 30–45% range for the summer seasons and increased on average by about 5% in winter. Air velocities (averaged across three heights) were low in winter, with a mean of 0.13 m/s, but this increased to 0.20 m/s in the summer season. Table 2 also presents a statistical summary of the thermal environmental and comfort indices broken down by season. On average, to and ET values fell within the 22–251C range while the PMV calculations indicate neutral conditions (0 to 0.1). The questionnaire recorded subjects’ assessments of their workstation thermal environments on a variety of scales at the same time as the measurements of indoor
Table 1 Summary of two samples of building occupants and their personal thermal variables Season Sample size (male/female)
Winter 640 (326/314)
Summer 589 (315/271)
Gender (% of sample) Male Female
50.9 49.1
53.8 46.2
Age (years) Mean Standard deviation Maximum Minimum
35.4 10.2 67.0 16.0
35.6 10.1 66.0 17.0
Years in Kalgoorlie region Mean Standard deviation Maximum Minimum
10.1 12.7 57.0 0
10.0 12.9 58.0 0
Highest education level (% of sample) High school Diploma/degree Postgraduate/university
33.8 52.5 13.7
28.8 55.6 15.6
0.69 (0.72/0.66) 0.23 (0.22/0.24) 1.57 (1.46/1.57) 0.43 (0.44/0.43)
0.49 (0.54/0.43) 0.14 (0.13/0.14) 1.21 (1.20/1.21) 0.33 (0.34/0.33)
Intrinsic clothing insulation (clo) Whole sample (male/female) Mean Standard deviation Maximum Minimum
Clothing and chair insulation (clo) 0.15 clo added to the above=mean
0.84 0.64 (0.87/0.81) (0.69/0.58)
Metabolism (W/m2), 58 W/m2=1 met Whole sample (male/female) Mean 76.5 (77.2/75.9) Standard deviation 11.4 (11.8/10.9) Maximum 100 (100/100) Minimum 60 (60/60)
77.2 (77.1/77.3) 11.1 (10.8/11.6) 100 (100/100) 60 (60/60)
climatic conditions were recorded. The scales included thermal sensation (ASHRAE seven-point scale), thermal acceptability (binary), thermal preference (warmer, cooler, no change), and general comfort
412
K. Cena, R. de Dear / Journal of Thermal Biology 26 (2001) 409–414
(six-point scale ranging from very comfortable to very uncomfortable). Table 2 indicates that mean thermal sensations on the ASHRAE seven-point scale were marginally warmer than neutral (+0.4) during the winter season but in summer the average vote came very close to neutral Table 2 Summary of indoor climatic data and calculated indoor climatic and thermal comfort indices, and the mean thermal sensation as rated by subjects on scale 3 (cold) to +3 (hot). ET; TSENS and PMV have added chair insulation included Season Sample size
Winter 640
Summer 589
Air temperature (1C) (average of three heights) Mean Standard deviation Maximum Minimum
22.0 1.4 24.5 16.0
23.4 1.4 30.5 19.1
Mean radiant temperature (1C) (average of three heights) Mean Standard deviation Maximum Minimum
22.3 1.3 24.7 16.8
24.0 1.4 32.8 20.2
Relative humidity (%) Mean Standard deviation Maximum Minimum
46.1 6.6 69.8 31.6
41.5 8.8 66.1 24.5
Table 2 (continued) Season Sample size
Winter 640
Summer 589
Predicted mean vote, PMV, on scale 3 to +3 Mean Standard deviation Maximum Minimum
0.1 0.5 1.0 2.2
0.02 0.9 1.9 3.0
ASHRAE thermal sensation rated by subjects, scale 3 to +3 Mean Standard deviation Maximum Minimum
0.4 1.0 3.0 3.0
0.1 1.1 3.0 3.0
(+0.1). For further analysis, the data were binned into 0.51C intervals and percentages of subjects voting ‘‘warmer-than-neutral’’ and ‘‘cooler-than-neutral’’ for each bin were calculated (with subjects voting ‘‘neutral’’ being split 50 : 50). The operative temperature ðto Þ neutralities were determined to be 20.31C in the winter and 23.31C in the summer. The winter survey’s fitted equation was ½mean ASHRAE vote ¼ 0:21to 4:28: The equation that best fitted the summer survey’s data was ½mean ASHRAE vote ¼ 0:27to 6:29:
Air velocity (m/s) (average of three heights) Mean Standard deviation Maximum Minimum
0.13 0.06 0.68 0.04
0.20 0.11 1.57 0.05
o
Operative temperature , to ( C) Mean Standard deviation Maximum Minimum
22.1 1.3 24.5 16.6
23.7 1.4 31.7 19.8
Effective temperature, ET (oC) Mean Standard deviation Maximum Minimum
22.1 1.3 24.5 16.7
23.5 1.3 29.7 19.6
Thermal sensation, TSENS, on scale 3 (cold) to +3 (hot) Mean Standard deviation Maximum Minimum
0.02 0.19 0.82 0.43
0.01 0.18 1.23 0.63
4. Discussion Both in winter and in summer the standard PMV calculations indicated slightly cooler than neutral conditions (the difference was up to a full PMV unit), but when the thermal insulation effect of office chairs was taken into account this difference was reduced to within 0.5 of a PMV unit. PMV predictions overestimated observed neutrality by less than three degrees (C) in winter. On the basis of the adaptive model of thermal comfort (de Dear and Brager, 1998) one might have predicted that acclimatization to Kalgoorlie’s hot and dry climate, especially during the summer season, would push the observed neutrality warmer than that predicted on the basis of PMV. One possible explanation for this counterintuitive outcome is that the occupants of these air-conditioned buildings actually adapt to their indoor climates. The psychological dimension of thermal adaptation comes into play and building occupants come to expect the cool temperatures they experience in
K. Cena, R. de Dear / Journal of Thermal Biology 26 (2001) 409–414
office buildings, even if they are being overcooled. The slope of the PMV-on-operative temperature relationship was, however, quite small which indicates that the subjects used other methods like clothing insulation and/ or adjustments to their activity levels to compensate for any departures in the ambient conditions from neutrality but they were apparently ignoring these adaptive responses when casting their votes on the ASHRAE sensation scale. It was as if they were saying: ‘‘Because I had to put on a pullover to stay comfortable I must be cooler than neutral’’! This latter pattern was particularly evident in Kalgoorlie’s winter. Overall about half of the sample wished for ‘‘more air movement’’, with requests for ‘‘more’’ in winter being greater than in summer by about 7%. Over 40% of subjects wanted ‘‘no change’’. The remaining 6% wanted ‘‘less’’ air movement. This was most evident during the summer season. Actual air speeds in that season were generally measured between 0.15 and 0.25 m/s. Furthermore, a majority (about 80%) of subjects who requested cooler temperatures also requested more air movement. At the same time, over 40% of the subjects who indicated preference for higher temperatures also requested more air movement, despite the fact that this would make them feel even cooler. Forty-eight percent of the total sample was female. Earlier field studies of thermal comfort have noted differences in the thermal comfort responses of males and females (e.g. Fishman and Pimbert, 1971). Usually the differences have been explained in terms of clothing differences, with females having greater inter- and intraseasonal variability in clo levels than males. These gender differences mean that air conditioning systems are trying to meet the thermal needs of two disparate sub-populations with a single set of environmental conditions, so it is not surprising that the levels of thermal acceptability obtained in field studies such as these in Kalgoorlie oftentimes fall well short of the ideal levels assumed in the comfort standards and models. The between-sex difference in clo values was about 0.1 clo units in Kalgoorlie and this translates into about 0.81C difference in preferred temperatures on the basis of the PMV model. In the present study the females’ seasonal range in mean clo values was 28% greater than the male’s. The mean thermal sensation cast by male subjects during the winter season’s survey was +0.26 on the ASHRAE seven-point scale, which was marginally cooler than the females’ mean of +0.58. In summer the difference narrowed to just 0.06 sensation units, with the males again being cooler than the females. In the winter survey, 12% of males registered votes of ‘‘thermally unacceptable’’ whereas 19% of females voted that way under the same conditions. In summer, the females were again more prone to express thermal dissatisfaction with 14% voting ‘‘unacceptable’’ compared to just 8% of males.
413
Only 10% of the subjects claimed that they did not have air-conditioning at home or in their car. During summer, over 80% of all subjects actually used airconditioners in their cars. Two-thirds to three-quarters of Kalgoorlie-Boulder subjects also used evaporative air conditioning in their homes in summer. These findings indicate that the majority of subjects in the present study are not exposed in summer to the natural ambient conditions as they live, travel and work in airconditioned spaces, calling into question their degree of acclimatization. In effect they are being cocooned from the heat and aridity of the climate zone in which they live, but only at a cost of significant energy input to their built environments. As the ethos of low-energy and passive architecture gains acceptance elsewhere in the world (e.g. the hybrid ventilation Commerzbank headquarters in Frankfurt and other landmark ‘‘green’’ buildings) we are confronted with the paradox of significantly colder buildings in hotter parts of the world and vice versa. Clearly further research into the processes used by occupants of buildings in a hot-arid and other warmer climates zones to adapt to their thermal indoor environments is suggested as an important goal for thermal comfort specialists. The interaction of all (home and work) indoor environments in shaping behavioural strategies of occupants would be of special interest. Furthermore, behavioural strategies for ‘‘weaning’’ any building occupants off possible excessive cooling would be desirable research goals as long as they can be implemented without being overly disruptive to occupant comfort or productivity. Is it simply a case of lifting set-points to more appropriate levels in one day, or is an incremental approach more likely to meet with occupant acceptance? Additionally we need to know more about what the ‘‘costs’’ of adaptation to warmer indoor temperatures are, if any.
5. Conclusions 1. A total of 935 subjects provided 1229 sets of data for winter and summer in a hot-arid location of Kalgoorlie-Boulder in the desert region of Western Australia. Clothing insulation levels recorded in Kalgoorlie-Boulder were 0.5 clo in summer and 0.7 in winter. Office chairs were estimated to add 0.15 clo to the clothing insulation. Metabolic rates were estimated to be on average 77 W/m2 or 1.3 met for both seasons and for both sexes. 2. Thermal neutrality, according to responses on the ASHRAE seven-point sensation scale, occurred at 20.31C in winter and at 23.31C in summer. Preferred temperature, defined as a minimum of subjects requesting temperature change, was 22.21C for both seasons. Thermal acceptability showed little or no
414
3.
4.
5.
6.
K. Cena, R. de Dear / Journal of Thermal Biology 26 (2001) 409–414
systematic relationship with the thermal environmental conditions. After the effect of chair insulation was accounted for, the PMV index adequately predicted optimum summer-time temperatures for the KalgoorlieBoulder subjects, whether defined in terms of thermal neutrality, thermal acceptability or thermal preference. The PMV predicted neutrality and preference for winter were less useful. There was little difference (particularly in summer) between the sexes in terms of thermal sensations, although there were significantly more expressions of thermal dissatisfaction from the females. The effects of Kalgoorlie-Boulder hot-dry/cool-dry seasonality on thermal comfort responses of office workers was significant, amounting to a 31C shift in neutrality and was within the range expected on the basis of the clothing insulation differences of approximately 0.2 clo between seasons. Future research into how the overcooling of office buildings in hot-dry climates can be reduced without disrupting the comfort and productivity of their occupants is called for.
Acknowledgements Financial support from the American Society of Heating, Refrigerating and Air-Conditioning Engineers (ASHRAE) under research grant RP-921 is acknowledged. The work reported was performed by Murdoch University (Perth, Australia) and Macquarie University (Sydney, Australia). R. de D. is on the faculty of the Division of Environmental and Life Sciences, Macquarie University.
References Abdelrahman, M.A., 1990. Field evaluation of a comfort meter. ASHRAE Trans. 96, 212–215.
ASHRAE, 1992. ANSI/ASHRAE Standard 55-1992. Thermal environmental conditions for human occupancy. American Society of Heating, Refrigerating and Air-Conditioning Engineers, Atlanta. Baker, N., Standeven, M., 1994. Comfort criteria for passively cooled buildings. A PASCOOL task, Renewable Energy 5, 977–984. Cena, K., 1994. Thermal and non-thermal aspects of comfort surveys in homes and offices. In: Oseland, N.A., Humphreys, M.A. (Eds.), Thermal Comfort: Past, Present and Future. Building Research Establishment, Watford, UK, pp. 73–87. Cena, K., de Dear, R.J., 1999. Field study of occupant comfort and office thermal environments in a hot, arid climate. ASHRAE Trans. 105, 204–217. de Dear, R.J., Fountain, M.E., 1994. Field experiments on occupant comfort and office thermal environments in a hothumid climate. ASHRAE Trans. 100, 457–475. de Dear, R.J., Brager, S.G., 1998. Developing an adaptive model of thermal comfort and preference. ASHRAE Trans. 104, 145–167. Donnini, G., Molina, J., Martello, C., Lai, D.H.C., Lai, H.K., Chang, C.Y., Laflamme, M., Nguyen, V.H., Haghighat, F., 1996. Field study of occupant comfort and office thermal environments in a cold climate. ASHRAE Trans. 103. Fanger, P.O., 1970. Thermal comfort. Danish Technical Press, Lyngby. Fishman, D.S., Pimbert, S.L., 1971. Survey of the subjective responses to the thermal environment in offices. In: Fanger, P.O., Valbjorn, O. (Eds.), Indoor Climate. Danish Building Research Institute, Copenhagen, pp. 677–689. Fountain, M.E., Huizenga, C., 1996. A thermal comfort prediction tool. ASHRAE J. 38, 39–42. Saeed, S.A.R., 1993. Thermal comfort requirements in hot dry regions with special reference to Riyadh, Part 1: for university students. Int. J. Ambient Energy 14, 147–154. Saeed, S.A.R., 1996. Thermal comfort requirements in hot dry regions with special reference to Riyadh, Part 2: Friday prayer. Int. J. Ambient Energy 17, 17–21. Schiller, G.E., Arens, E., Bauman, F., Benton, C., Fountain, M., Doherty, T., Craik, K., 1988. A field study of thermal environments and comfort in office buildings. ASHRAE Trans. 94, 1–10.
Thermal comfort and behavioural strategies in office buildings located in a hot-arid climate Krzysztof Cenaa,*, Richard de Dearb a
Department of Environmental Science, Murdoch University, Perth 6150, Australia Environmental and Life Sciences, Macquarie University, Sydney 2109, Australia
b
Abstract The effects of indoor climates on thermal perceptions and adaptive behaviour of office workers during a large field study in Kalgoorlie-Boulder, located in a hot-arid region of Western Australia, are discussed. Clothing insulation levels were 0.5 clo in summer and 0.7 in winter. Thermal neutrality, according to responses on the American Society of Heating, Refrigerating and Air-Conditioning Engineers seven-point sensation scale, occurred at 20.31C in winter and at 23.31C in summer. The effect of hot-dry/cool-dry seasonality on thermal comfort responses of office workers was significant. Future research into how the overcooling of office buildings in hot-dry climates can be reduced is called for. r 2000 Elsevier Science Ltd. All rights reserved. Keywords: Thermal comfort; Hot-arid climate; Clothing insulation; Office occupants; Behavioural strategies
1. Introduction It is now widely accepted that the previously used climate chambers fail to provide the participating humans with so called ‘‘experiential realism’’ in determining thermal comfort. Cleaning-up and simplification of comfort problems to make them suitable for chamber research unfortunately removes many of the contextual factors that we now recognise as part of the thermal comfort matrix, like competing demands for attention, expectations and attitudes. Field studies of thermal comfort do not suffer from this flaw because they are conducted in actual buildings under normal conditions of occupancy, and involve much larger and diverse samples of ‘‘real’’ occupants as opposed to ‘‘paid college-age subjects’’ (Cena, 1994). This positive change has been facilitated by a gradual development of laboratory-grade field instruments and an enhanced set of protocols that, together, make comfort data obtained in the field every bit as rigorous as the climate chamber counterpart. Field studies also allow for analyses of *Corresponding author. Tel.: +61-8-9360-2883; fax: +61-89310-4997. E-mail address: [email protected] (K. Cena).
other factors than those that can be simulated in chambers, as the subjects provide responses in their everyday habitats, wearing their everyday clothing and behaving without any additional restrictions. American Society of Heating, Refrigerating and AirConditioning Engineers (ASHRAE) has commissioned a series of state-of-the-art, fully compatible field experiments across a variety of climate zones including: temperate (Schiller et al., 1988), hot-humid (de Dear and Fountain, 1994) and cold (Donnini et al., 1996). The current paper is based upon the fourth (final) instalment in this series conducted by Cena and de Dear (1999) in a hot-arid climate in Kalgoorlie-Boulder in Western Australia. There have been relatively few comparable studies conducted in hot-arid locations. These include those by Baker and Standeven (1994) in residential buildings during the summer in Athens, Greece. Their results indicated the importance of adaptive opportunity in order for building occupants to accept and achieve comfort in temperatures warmer than 241C. Abdelrahman (1990) measured thermal comfort in a few classrooms in a Saudi university and concluded that his instruments’ comfort algorithms, based on Fanger’s method, were valid (Fanger, 1970). Saeed (1993)
0306-4565/01/$ - see front matter r 2000 Elsevier Science Ltd. All rights reserved. PII: S 0 3 0 6 - 4 5 6 5 ( 0 1 ) 0 0 0 5 2 - 3
410
K. Cena, R. de Dear / Journal of Thermal Biology 26 (2001) 409–414
conducted field experiments in Riyadh, Saudi Arabia. Again, statistical analyses of their results indicated that Fanger’s equation did well at predicting subjective responses of this sample. A subsequent study by Saeed (1996) of indoor climate on 525 subjects exposed to hot conditions in a mosque confirmed his earlier conclusion. Unfortunately for the middle-eastern studies above, we have no indication of the subjects’ actual level of acclimatization to the hot-dry setting. Given the pervasive nature of refrigerated air-conditioning in such cheap-energy countries one suspects that acclimatization would be negligible. In such situations the adaptive model of thermal comfort (de Dear and Brager, 1998) predicts that the subjects become adapted to the indoor climates they spend most of their lives in, not their outdoor climatic setting. So the close agreement with Fanger’s PMV model is not surprising since it (or a close relative of it at least) has been used to actually design and operate the air conditioning systems. The present paper will discuss the main results of a large field study (Cena and de Dear, 1999) conducted in a hot-arid climate and will focus on the effects of indoor climates on thermal perceptions and adaptive behaviour of office workers.
2. Methods Kalgoorlie-Boulder is situated in the desert area of Western Australia. Rainfall is minimal (only 258 mm per annum, over 300 days in a year are without any rainfall) with a pronounced seasonal variation. Mean minimum temperatures during winter and summer sample periods were 9.61C and 16.71C, respectively. Mean maximum temperatures of the winter and summer sample periods were 18.51C and 30.71C, respectively. Over the summer study period, the mean 9 a.m. relative humidity (RH) and the average daily maximum temperature were 39% and 30.71C, respectively. Twenty-two of the largest office buildings in Kalgoorlie-Boulder were chosen for the study. Sample sizes of 640 and 589 subjects were achieved in winter and summer surveys, respectively. This total of 1229 sets of data was provided by 935 respondents, of whom 294 were interviewed in both seasons. Forty-eight percent of the sample was female, and the average age of all subjects was 35 years. A mobile measurement system (Fig. 1) was used onsite in Kalgoorlie to collect measurements of the indoor atmospheric environment. The system collected concurrent physical data (air temperature, dew-point temperature, RH, globe temperature, radiant asymmetry and air velocity) from three arrays of transducers placed 0.1, 0.6, and 1.1 m above floor level, representing the immediate environment of the seated subject’s ankles, body and neck, respectively.
Fig. 1. Mobile and portable system used for measurements of indoor climates. It incorporates three arrays of sensors positioned at 0.1, 0.6 and 1.1 m and allows for automated recording of air temperature, dew-point temperature, RH, globe temperature, radiant asymmetry, and air velocity.
The questionnaires (approved by Murdoch University’s Human Ethics Committee) covered several areas including demographics, work area satisfaction, personal environmental control, job satisfaction and health. The questionnaires also included the traditional scales of thermal sensation and thermal preference, current clothing garment and metabolic activity checklists, and two scales focusing on air movement perceptions. The thermal sensation scale was the ASHRAE seven-point scale of warmth ranging from cold (3) to hot (+3) with neutral (0) in the middle. The thermal preference scale asked on a three-point scale whether the respondent would like a change in the thermal environment. Possible responses were ‘‘want warmer’’, ‘‘want no change’’, or ‘‘want cooler’’. Clothing garment checklists were adapted to the regional customs prevailing in Kalgoorlie-Boulder and compiled from the extensive lists published in ASHRAE Standard 55-92. Metabolic rates were assessed by a checklist of office activities and were based on the detailed databases published in ASHRAE Standard 55-92 and ISO 7730. Various indices were calculated using software by Fountain
411
K. Cena, R. de Dear / Journal of Thermal Biology 26 (2001) 409–414
and Huizenga (1996) and included operative temperature ðto Þ; mean radiant temperature, effective temperature ðETÞ; predicted mean vote (PMV), predicted percentage dissatisfied (PPD), TSENS, along with the PD index (draft risk) incorporated into ANSI/ASHRAE Standard 55-1992 (ASHRAE, 1992). All index calculations requiring a clo value input included an estimate of the thermal insulation increment resulting from subjects being seated.
3. Results Intrinsic clothing ensemble insulation was estimated on the basis of garment values published in ANSI/ ASHRAE Standard 55-1992 and was very close to the Standard’s assumed summer value of 0.5 clo, but with females being about 0.1 clo lower than the males (Table 1). The winter season’s ensemble average was about 0.2 clo higher than the summer’s. Concomitant with this seasonal difference was approximately a 50% decrease in the inter-individual variability (standard deviation) from 0.23 clo in the winter season to 0.14 clo in the summer season. This suggests that the degree of freedom for thermoregulation and accommodation of interindividual differences in thermal preferences by clothing adjustments diminished as the amount of clothing approached the socially acceptable minimum. Estimated average chair insulation of 0.15 clo lifted the subjects’ thermal insulation values to about 0.8 and 0.6 clo in the winter and summer surveys, respectively. Metabolic rates of the subjects were estimated to be, on average, 77 W/m2 or 1.3 met in both seasons and for both sexes, representative of sedentary activity. Mean height of the subjects was about 171 cm and their mean weight was about 75 kg. Table 2 provides statistical summaries of the indoors climatic measurements for the Kalgoorlie winter and summer season samples, respectively. These data represent the main thermal variables recorded at subjects’ workstations by the mobile system. Mean air and radiant temperatures (averaged across the three heights 0.1, 0.6 and 1.1 m) generally fell within the 22–261C interval for both seasons. Average RH fell within the 30–45% range for the summer seasons and increased on average by about 5% in winter. Air velocities (averaged across three heights) were low in winter, with a mean of 0.13 m/s, but this increased to 0.20 m/s in the summer season. Table 2 also presents a statistical summary of the thermal environmental and comfort indices broken down by season. On average, to and ET values fell within the 22–251C range while the PMV calculations indicate neutral conditions (0 to 0.1). The questionnaire recorded subjects’ assessments of their workstation thermal environments on a variety of scales at the same time as the measurements of indoor
Table 1 Summary of two samples of building occupants and their personal thermal variables Season Sample size (male/female)
Winter 640 (326/314)
Summer 589 (315/271)
Gender (% of sample) Male Female
50.9 49.1
53.8 46.2
Age (years) Mean Standard deviation Maximum Minimum
35.4 10.2 67.0 16.0
35.6 10.1 66.0 17.0
Years in Kalgoorlie region Mean Standard deviation Maximum Minimum
10.1 12.7 57.0 0
10.0 12.9 58.0 0
Highest education level (% of sample) High school Diploma/degree Postgraduate/university
33.8 52.5 13.7
28.8 55.6 15.6
0.69 (0.72/0.66) 0.23 (0.22/0.24) 1.57 (1.46/1.57) 0.43 (0.44/0.43)
0.49 (0.54/0.43) 0.14 (0.13/0.14) 1.21 (1.20/1.21) 0.33 (0.34/0.33)
Intrinsic clothing insulation (clo) Whole sample (male/female) Mean Standard deviation Maximum Minimum
Clothing and chair insulation (clo) 0.15 clo added to the above=mean
0.84 0.64 (0.87/0.81) (0.69/0.58)
Metabolism (W/m2), 58 W/m2=1 met Whole sample (male/female) Mean 76.5 (77.2/75.9) Standard deviation 11.4 (11.8/10.9) Maximum 100 (100/100) Minimum 60 (60/60)
77.2 (77.1/77.3) 11.1 (10.8/11.6) 100 (100/100) 60 (60/60)
climatic conditions were recorded. The scales included thermal sensation (ASHRAE seven-point scale), thermal acceptability (binary), thermal preference (warmer, cooler, no change), and general comfort
412
K. Cena, R. de Dear / Journal of Thermal Biology 26 (2001) 409–414
(six-point scale ranging from very comfortable to very uncomfortable). Table 2 indicates that mean thermal sensations on the ASHRAE seven-point scale were marginally warmer than neutral (+0.4) during the winter season but in summer the average vote came very close to neutral Table 2 Summary of indoor climatic data and calculated indoor climatic and thermal comfort indices, and the mean thermal sensation as rated by subjects on scale 3 (cold) to +3 (hot). ET; TSENS and PMV have added chair insulation included Season Sample size
Winter 640
Summer 589
Air temperature (1C) (average of three heights) Mean Standard deviation Maximum Minimum
22.0 1.4 24.5 16.0
23.4 1.4 30.5 19.1
Mean radiant temperature (1C) (average of three heights) Mean Standard deviation Maximum Minimum
22.3 1.3 24.7 16.8
24.0 1.4 32.8 20.2
Relative humidity (%) Mean Standard deviation Maximum Minimum
46.1 6.6 69.8 31.6
41.5 8.8 66.1 24.5
Table 2 (continued) Season Sample size
Winter 640
Summer 589
Predicted mean vote, PMV, on scale 3 to +3 Mean Standard deviation Maximum Minimum
0.1 0.5 1.0 2.2
0.02 0.9 1.9 3.0
ASHRAE thermal sensation rated by subjects, scale 3 to +3 Mean Standard deviation Maximum Minimum
0.4 1.0 3.0 3.0
0.1 1.1 3.0 3.0
(+0.1). For further analysis, the data were binned into 0.51C intervals and percentages of subjects voting ‘‘warmer-than-neutral’’ and ‘‘cooler-than-neutral’’ for each bin were calculated (with subjects voting ‘‘neutral’’ being split 50 : 50). The operative temperature ðto Þ neutralities were determined to be 20.31C in the winter and 23.31C in the summer. The winter survey’s fitted equation was ½mean ASHRAE vote ¼ 0:21to 4:28: The equation that best fitted the summer survey’s data was ½mean ASHRAE vote ¼ 0:27to 6:29:
Air velocity (m/s) (average of three heights) Mean Standard deviation Maximum Minimum
0.13 0.06 0.68 0.04
0.20 0.11 1.57 0.05
o
Operative temperature , to ( C) Mean Standard deviation Maximum Minimum
22.1 1.3 24.5 16.6
23.7 1.4 31.7 19.8
Effective temperature, ET (oC) Mean Standard deviation Maximum Minimum
22.1 1.3 24.5 16.7
23.5 1.3 29.7 19.6
Thermal sensation, TSENS, on scale 3 (cold) to +3 (hot) Mean Standard deviation Maximum Minimum
0.02 0.19 0.82 0.43
0.01 0.18 1.23 0.63
4. Discussion Both in winter and in summer the standard PMV calculations indicated slightly cooler than neutral conditions (the difference was up to a full PMV unit), but when the thermal insulation effect of office chairs was taken into account this difference was reduced to within 0.5 of a PMV unit. PMV predictions overestimated observed neutrality by less than three degrees (C) in winter. On the basis of the adaptive model of thermal comfort (de Dear and Brager, 1998) one might have predicted that acclimatization to Kalgoorlie’s hot and dry climate, especially during the summer season, would push the observed neutrality warmer than that predicted on the basis of PMV. One possible explanation for this counterintuitive outcome is that the occupants of these air-conditioned buildings actually adapt to their indoor climates. The psychological dimension of thermal adaptation comes into play and building occupants come to expect the cool temperatures they experience in
K. Cena, R. de Dear / Journal of Thermal Biology 26 (2001) 409–414
office buildings, even if they are being overcooled. The slope of the PMV-on-operative temperature relationship was, however, quite small which indicates that the subjects used other methods like clothing insulation and/ or adjustments to their activity levels to compensate for any departures in the ambient conditions from neutrality but they were apparently ignoring these adaptive responses when casting their votes on the ASHRAE sensation scale. It was as if they were saying: ‘‘Because I had to put on a pullover to stay comfortable I must be cooler than neutral’’! This latter pattern was particularly evident in Kalgoorlie’s winter. Overall about half of the sample wished for ‘‘more air movement’’, with requests for ‘‘more’’ in winter being greater than in summer by about 7%. Over 40% of subjects wanted ‘‘no change’’. The remaining 6% wanted ‘‘less’’ air movement. This was most evident during the summer season. Actual air speeds in that season were generally measured between 0.15 and 0.25 m/s. Furthermore, a majority (about 80%) of subjects who requested cooler temperatures also requested more air movement. At the same time, over 40% of the subjects who indicated preference for higher temperatures also requested more air movement, despite the fact that this would make them feel even cooler. Forty-eight percent of the total sample was female. Earlier field studies of thermal comfort have noted differences in the thermal comfort responses of males and females (e.g. Fishman and Pimbert, 1971). Usually the differences have been explained in terms of clothing differences, with females having greater inter- and intraseasonal variability in clo levels than males. These gender differences mean that air conditioning systems are trying to meet the thermal needs of two disparate sub-populations with a single set of environmental conditions, so it is not surprising that the levels of thermal acceptability obtained in field studies such as these in Kalgoorlie oftentimes fall well short of the ideal levels assumed in the comfort standards and models. The between-sex difference in clo values was about 0.1 clo units in Kalgoorlie and this translates into about 0.81C difference in preferred temperatures on the basis of the PMV model. In the present study the females’ seasonal range in mean clo values was 28% greater than the male’s. The mean thermal sensation cast by male subjects during the winter season’s survey was +0.26 on the ASHRAE seven-point scale, which was marginally cooler than the females’ mean of +0.58. In summer the difference narrowed to just 0.06 sensation units, with the males again being cooler than the females. In the winter survey, 12% of males registered votes of ‘‘thermally unacceptable’’ whereas 19% of females voted that way under the same conditions. In summer, the females were again more prone to express thermal dissatisfaction with 14% voting ‘‘unacceptable’’ compared to just 8% of males.
413
Only 10% of the subjects claimed that they did not have air-conditioning at home or in their car. During summer, over 80% of all subjects actually used airconditioners in their cars. Two-thirds to three-quarters of Kalgoorlie-Boulder subjects also used evaporative air conditioning in their homes in summer. These findings indicate that the majority of subjects in the present study are not exposed in summer to the natural ambient conditions as they live, travel and work in airconditioned spaces, calling into question their degree of acclimatization. In effect they are being cocooned from the heat and aridity of the climate zone in which they live, but only at a cost of significant energy input to their built environments. As the ethos of low-energy and passive architecture gains acceptance elsewhere in the world (e.g. the hybrid ventilation Commerzbank headquarters in Frankfurt and other landmark ‘‘green’’ buildings) we are confronted with the paradox of significantly colder buildings in hotter parts of the world and vice versa. Clearly further research into the processes used by occupants of buildings in a hot-arid and other warmer climates zones to adapt to their thermal indoor environments is suggested as an important goal for thermal comfort specialists. The interaction of all (home and work) indoor environments in shaping behavioural strategies of occupants would be of special interest. Furthermore, behavioural strategies for ‘‘weaning’’ any building occupants off possible excessive cooling would be desirable research goals as long as they can be implemented without being overly disruptive to occupant comfort or productivity. Is it simply a case of lifting set-points to more appropriate levels in one day, or is an incremental approach more likely to meet with occupant acceptance? Additionally we need to know more about what the ‘‘costs’’ of adaptation to warmer indoor temperatures are, if any.
5. Conclusions 1. A total of 935 subjects provided 1229 sets of data for winter and summer in a hot-arid location of Kalgoorlie-Boulder in the desert region of Western Australia. Clothing insulation levels recorded in Kalgoorlie-Boulder were 0.5 clo in summer and 0.7 in winter. Office chairs were estimated to add 0.15 clo to the clothing insulation. Metabolic rates were estimated to be on average 77 W/m2 or 1.3 met for both seasons and for both sexes. 2. Thermal neutrality, according to responses on the ASHRAE seven-point sensation scale, occurred at 20.31C in winter and at 23.31C in summer. Preferred temperature, defined as a minimum of subjects requesting temperature change, was 22.21C for both seasons. Thermal acceptability showed little or no
414
3.
4.
5.
6.
K. Cena, R. de Dear / Journal of Thermal Biology 26 (2001) 409–414
systematic relationship with the thermal environmental conditions. After the effect of chair insulation was accounted for, the PMV index adequately predicted optimum summer-time temperatures for the KalgoorlieBoulder subjects, whether defined in terms of thermal neutrality, thermal acceptability or thermal preference. The PMV predicted neutrality and preference for winter were less useful. There was little difference (particularly in summer) between the sexes in terms of thermal sensations, although there were significantly more expressions of thermal dissatisfaction from the females. The effects of Kalgoorlie-Boulder hot-dry/cool-dry seasonality on thermal comfort responses of office workers was significant, amounting to a 31C shift in neutrality and was within the range expected on the basis of the clothing insulation differences of approximately 0.2 clo between seasons. Future research into how the overcooling of office buildings in hot-dry climates can be reduced without disrupting the comfort and productivity of their occupants is called for.
Acknowledgements Financial support from the American Society of Heating, Refrigerating and Air-Conditioning Engineers (ASHRAE) under research grant RP-921 is acknowledged. The work reported was performed by Murdoch University (Perth, Australia) and Macquarie University (Sydney, Australia). R. de D. is on the faculty of the Division of Environmental and Life Sciences, Macquarie University.
References Abdelrahman, M.A., 1990. Field evaluation of a comfort meter. ASHRAE Trans. 96, 212–215.
ASHRAE, 1992. ANSI/ASHRAE Standard 55-1992. Thermal environmental conditions for human occupancy. American Society of Heating, Refrigerating and Air-Conditioning Engineers, Atlanta. Baker, N., Standeven, M., 1994. Comfort criteria for passively cooled buildings. A PASCOOL task, Renewable Energy 5, 977–984. Cena, K., 1994. Thermal and non-thermal aspects of comfort surveys in homes and offices. In: Oseland, N.A., Humphreys, M.A. (Eds.), Thermal Comfort: Past, Present and Future. Building Research Establishment, Watford, UK, pp. 73–87. Cena, K., de Dear, R.J., 1999. Field study of occupant comfort and office thermal environments in a hot, arid climate. ASHRAE Trans. 105, 204–217. de Dear, R.J., Fountain, M.E., 1994. Field experiments on occupant comfort and office thermal environments in a hothumid climate. ASHRAE Trans. 100, 457–475. de Dear, R.J., Brager, S.G., 1998. Developing an adaptive model of thermal comfort and preference. ASHRAE Trans. 104, 145–167. Donnini, G., Molina, J., Martello, C., Lai, D.H.C., Lai, H.K., Chang, C.Y., Laflamme, M., Nguyen, V.H., Haghighat, F., 1996. Field study of occupant comfort and office thermal environments in a cold climate. ASHRAE Trans. 103. Fanger, P.O., 1970. Thermal comfort. Danish Technical Press, Lyngby. Fishman, D.S., Pimbert, S.L., 1971. Survey of the subjective responses to the thermal environment in offices. In: Fanger, P.O., Valbjorn, O. (Eds.), Indoor Climate. Danish Building Research Institute, Copenhagen, pp. 677–689. Fountain, M.E., Huizenga, C., 1996. A thermal comfort prediction tool. ASHRAE J. 38, 39–42. Saeed, S.A.R., 1993. Thermal comfort requirements in hot dry regions with special reference to Riyadh, Part 1: for university students. Int. J. Ambient Energy 14, 147–154. Saeed, S.A.R., 1996. Thermal comfort requirements in hot dry regions with special reference to Riyadh, Part 2: Friday prayer. Int. J. Ambient Energy 17, 17–21. Schiller, G.E., Arens, E., Bauman, F., Benton, C., Fountain, M., Doherty, T., Craik, K., 1988. A field study of thermal environments and comfort in office buildings. ASHRAE Trans. 94, 1–10.