Influence of Window Views on the Subjective Evaluation of Discomfort Glare

Transcription

1 Original Paper Indoor and Built Indoor Built Environ 2011;20;1:65 74 Accepted: October 9, 2010 Environment Influence of Window Views on the Subjective Evaluation of Discomfort Glare Geun Young Yun Ju Young Shin Jeong Tai Kim Department of Architectural Engineering, Kyung Hee University, Yongin , Korea Key Words Glare E Luminance conditions E Simulated window E Subjective evaluation E View E Visual discomfort psychological factors. Thus, the finding from this study would contribute to a more realistic evaluation of discomfort glare for future design of glare control systems. Abstract A window is an indispensable element in a building and acts as a view-giving component that keeps occupants in touch with the outside. This study investigated the potential effect of different window views on the subjective assessment of discomfort glare from a simulated window that rendered blank, natural and man-made views from far to close distances. Fortyeight subjects (24 men, 24 women) participated in the experiments. The experimental results confirmed that the subjective evaluation of discomfort glare can vary with the type of window views presented. The results also indicated that there were noticeable variations in the subjective assessments of discomfort glare over the same visual stimuli. A preliminary criterion for classification of subjects sensitivity to glare was illustrated and the difference for the glare-sensitive and glareinsensitive people was statistically significant. This study found that the psychological factor such as window views could be an important factor in the subjective evaluation of discomfort glare. Widely used glare evaluation formulas such as unified glare rating and daylight glare index would not consider Introduction A well-designed window is one of the most essential elements in buildings for comfortable (thermal and visual), carbon-efficient and healthy indoor environments. One of the important environmental functions of windows is to provide natural light to the interior spaces of buildings. In general, architects and occupants prefer daylighting to artificial lighting [1]. In addition, daylighting would potentially save considerable amount of lighting energy consumption [2,3], typically about 30 50% of the total energy used in a non-domestic building. Daylight from windows can cause discomfort glare in buildings and discomfort glare from a window is a common problem in buildings. Iwata et al. [4] used a simulated window in order to evaluate the subjective evaluation of discomfort glare and proposed the glare sensation vote as a new glare index. Iwata et al. [4] illustrated that the subjective evaluation of discomfort glares from actual windows were similar to the results using the simulated window. Fiekis et al. [5] tested the ß SAGE Publications 2010 Los Angeles, London, New Delhi, Singapore and Washington DC DOI: / X Accessible online at Figures 3, 5 and 9 appear in colour online Prof Jeong Tai Kim, Department of Architectural Engineering, Kyung Hee University, Yongin , Korea. Tel. þ , Fax þ , jtkim@khu.ac.kr

2 daylight glare index (DGI) and the unified glare rating (UGR) under real conditions and proposed the modifications to improve the prediction performance of the DGI and UGR. Piccolo and Simone [6] evaluated the effect of electrochromic (EC)-glazing on discomfort glares from windows and found that EC glazing would reduce the discomfort glare. Tuaycharoen and Tregenza [7] found that the tolerance of discomfort glare would be greater when interesting images were presented to the observers than for other images. Heerwagen and Heerwagen [8] also revealed that window views were a crucial factor affecting discomfort glare sensations. Based on these previous researches, this study was undertaken to reveal the potential effects of window views on the subjective evaluation of discomfort glare and to investigate whether there could be variations in glare assessments among subjects. Experimental Setting Experimental Apparatus A simulated window was developed to render the various luminance conditions. The size of the window was cm 3 (Figure 1). An aluminium frame of a 30 cm width was located on the edge of the window and a cm 2 view image was pasted in front of the window glass. A total of 196 incandescent lamps (100 W, OSRAM) were placed, 8 cm apart, in a 14-per-row arrangement inside the window glass. The lamps were simultaneously controlled by a 30 kw dimmer. The surface temperature of the window was monitored by a temperature control sensor on the left side of the window. When the window temperature reached 558C, the window light would automatically turn down. The luminance of the window could be set from 0 to 15,000 cdm 2. View and Luminance Conditions of the Window The window views investigated included natural and man-made scenes. Both near and distant views and mixed views of the natural and man-made scenes were examined. The three-layer image with the sky was designated as the distant view and the two-layer image without the sky was the near view. The forest image was selected to be the natural land view; lake in the mountain was the natural river view; the apartment buildings in an urban area was the man-made view; the apartment buildings in a countryside was the mixed land view and the apartment buildings near a lake was the mixed river view. Additionally, a blank view (white paper) was chosen as the reference view. A total of 11 views were tested in the experiment. Table 1 shows the views of the tested image. The mean window luminance conditions ranged from 1000 to 10,000 cdm 2 with an equal interval of 0.5 log luminance units. Luminance was measured from the eye position of the subject with Minolta CS-100. The mean window luminance was calculated from measurements taken at nine, equally spaced separate points on the 200 Ventilation fan Temp. control sensor 1200 Reflection panel Incandescent lamp Diameter Ventilation fan 750 Fig. 1. The appearance of the simulated window. 66 Indoor Built Environ 2011;20:65 74 Yun et al.

3 Table 1. The images of the simulated view Distant Natural land Natural river Man made Mixed land Mixed river Near Just perceptible Noticeable Just acceptable Fig. 2. Discomfort glare evaluation scale. Acceptable Just uncomfortable Uncomfortable Just intolerable Intolerable window. During the experiment, the luminance was reproduced within 5% deviation of the reference luminance and the difference between the rendered and reference luminance levels was shown to be not statistically significant. Background luminance, spectrum distribution, correlation colour temperature and colour-rendering index were also measured in all luminance conditions. Evaluation Scale of Discomfort Glare The multiple criterion discomfort glare scale used by Hopkinson [9] was adopted for this study (Figure 2). s with short descriptions both in English and Korean were printed on the questionnaire. The subjects were requested to score their sensation accordingly, using a particular description given in the scale. Figure 2 shows the discomfort glare evaluation scale. Laboratory Condition Experiment was conducted in a m 3 laboratory space. The subject was requested to take their seat on a chair at 1.5 m position away from the simulated window and the researcher was stationed at 2.5 m away from the simulated window to measure the luminance conditions. The interior surfaces were painted white. The laboratory was fully masked from the outside light. Figure 3 shows the experimental layout in the laboratory Simulated window Bookshelf Diameter 1500 Fig. 3. Experimental layout Subject Experiment Experimental Procedure On arrival at the test room, the subject was placed on a chair 1.5 m away from the window. The subject was fully briefed about the procedures; and made to understand the purpose and definition of the glare scale used for the experiment. The subject was allowed 2 min to adapt their eyes to the lighting condition in the laboratory before the experiment began. The researcher, then, set up the luminance conditions and turned on the light. The subject was asked to focus on the centre of the window for about Influence of Window Views on Discomfort Glare Indoor Built Environ 2011;20:

4 Start Adaptation Window light on 1min 1min 5s 30s Look at Complete the window the sheet Repeat (5 luminance conditions) Take a break / change the views 3min End Fig. 4. The procedure of the experiment. Repeat (11 view conditions) Fig. 5. A scene taken from the experiment. 5 s and give the score, based on the discomfort glare scale, on the questionnaire. The above experiment was repeated for five luminance conditions of each view. There was a 3-min relaxation period after finishing each view experiment and the total experiment time for 11 views for testing of one subject was about 50 min. Each luminance and view conditions were randomly allocated for each subject being tested. Figure 4 illustrates the experimental flow and Figure 5 shows a scene from the experiment. The Subjects A total of 48 subjects (24 male, 24 female) participated in the experiment. All of them were university students. The age ranged from 20 to 32. Among the participants, 33 subjects wore glasses or contact lens. The data were analysed without discrimination of the gender, age, eyesight and eye conditions. A financial incentive of US$ 21 was paid to a subject for participating in the experiment. Results The Effects of Luminance Figure 6 shows the changes in the discomfort glare evaluation of the subjects as a function of the luminance of the blank view. The glare response votes of the subjects increased with an increase in the level of luminance. The glare response vote was 1.24 with a mean luminance of 1000 cdm 2, 2.30 with a mean luminance of 3200 cdm 2 and 3.97 with a mean luminance of 10,000 cdm 2 (F (4, 235) ¼ , p-value50.001). This confirms that the luminance from the window was an influential factor affecting the subjective assessment of discomfort glare. In this study, the discomfort zone (i.e. the multiple criterion discomfort glare scale greater than 3) assessed with the blank view began with the luminance value of 5600 cdm 2, that is the subjects started feeling uncomfortable with the blank view with a luminance of 5600 cdm 2. The mean glare response vote was 3.06 with the standard deviation (SD) of 0.58, although there were variations in glare response votes, as shown in Figure 6. The lowest vote was 1.5 while the highest vote was 4.5. The Effect of Window Views Table 2 compares the mean glare response vote for the blank view with the subjective glare evaluation for the simulated views at the luminance of 5600 cdm 2, from which the subjects started experiencing discomfort glare (Figure 6). The subject s tolerance to discomfort glare tended to increase with the simulated views; hence we see lower glare response vote, compared with the blank view. Seven out of 10 simulated views were given lower glare response votes than the response for the blank view. The differences were statistically significant at the level of at least The glare response votes for three simulated views were higher than that for the blank view. However, this was not statistically significant as the p-values for the t-tests were higher than 0.05 (Table 2). This indicates that the simulated views, whether they were natural, man-made or mixed scenes, had a positive effect on alleviating the discomfort glare sensation of the subjects. The positive effects of the simulated views were more distinctive when the luminance of 10,000 cdm 2 was used for views (Table 3). Lower glare response votes were given for all the simulated views than for the blank view. The glare response vote for the blank view was Indoor Built Environ 2011;20:65 74 Yun et al.

5 Largest value within 1.5 box lengths 75 th percentile Median 25 th percentile Smallest value within 1.5 box lengths Luminance (cd/m 2 ) Fig. 6. The subjective evaluation of glare for a blank view as a function of luminance. Table 2. Comparison of subjective glare votes between the blank view and simulated views at the luminance of 5600 cdm 2 View Glare response vote (SD) Difference with a blank view t-test p-value Blank 3.06 (0.58) Distant mixed river 3.01 (0.73) Near mixed river 2.64 (0.57) Distant mixed land 2.67 (0.65) Near mixed land 3.17 (0.62) Distant natural land 2.84 (0.45) Near natural land 3.21 (0.58) Distant natural river 2.81 (0.53) Near natural river 3.13 (0.56) Distant man-made 2.50 (0.61) Near man-made 2.75 (0.68) Table 3. Comparison of subjective glare votes between a blank-window view and simulated window views at the luminance of 10,000 cdm 2 View Glare response vote (SD) Difference with a blank view t-test p-value Blank 3.97 (0.42) Distant mixed river 3.80 (0.63) Near mixed river 3.57 (0.57) Distant mixed land 3.43 (0.64) Near mixed land 3.84 (0.55) Distant natural land 3.82 (0.58) Near natural land 3.89 (0.52) Distant natural river 3.57 (0.56) Near natural river 3.73 (0.47) Distant man-made 3.41 (0.67) Near man-made 3.56 (0.53) (SD ¼ 0.42), while the distant man-made view had the lowest glare response vote (i.e. 3.41, SD ¼ 0.67, t-test ¼ 4.959, p-value50.001). The glare response votes for the distant mixed river, near mixed land, distant natural land and near natural land views were slightly smaller than the glare evaluation for the blank view. However, it was not statistically significant ( p-values ). Figure 7 compares the subjective discomfort glare votes for the blank view with those for the simulated views. The Y axis in Figure 7(b) represents the bias of the votes for the simulated views from the subjective evaluations for the blank view (i.e. blank view votes minus corresponding simulated view votes). Thus, when the bias is positive, this would indicate less glare response given by the subjects with the simulated views than with the blank view. Influence of Window Views on Discomfort Glare Indoor Built Environ 2011;20:

6 Mixed river distant Mixed river near Mixed land distant Mixed land near Natural river distant Natural river near Natural land distant Natural land near Man made distant Man made near Blank Luminance (cd/m 2 ) 0.8 Mixed river distant Mixed river near Mixed land distant Mixed land near Natural river distant Natural river near Natural land distant Natural land near Man made distant Man made near Blank -0.4 Luminance (cd/m 2 ) Fig. 7. Comparison of the glare response votes for the blank and simulated views. (a) The subjects glare response votes as a function of the luminance. (b) The bias of the subjective discomfort glare votes from the glare evaluation for the blank view. Analysis of covariance (ANCOVA) analysis results show that the regression lines for the 11 views were not the same (F (20, 2618) ¼ 5.279, p-value50.001). This implies that the subjective evaluation of discomfort glare from windows could vary with the type of views. Figure 7 illustrates that the glare response votes for the simulated views were generally lower than the votes for blank views. The distant man-made view had consistently the lowest glare response votes at illuminance level ranging from 1000 to 10,000 cdm 2. The difference between the regression slopes of the distant man-made and blank views were significant (t-test ¼ 2.219, p-value ¼ 0.027). The difference in glare response votes between the near manmade and blank views became larger as the level of luminance increased with the views. The difference was 0.05 at the luminance of 1000 cdm 2 and this increased to 0.41 when the luminance level was 10,000 cdm 2. The subjective evaluation of the glare sensations for the natural river view from a near distance showed different patterns from those for the other simulated views. The glare votes for the near natural river view were greater than those for the blank view at luminance values of 1000, 1800, 3200 and 5600 cdm 2 (t-test ¼ 2.098, p-value ¼ 0.036). This was in contrast to the patterns found in the other simulated views. However, the near natural river views had a lower vote than the blank view at a luminance of 10,000 cdm 2. Variations in Glare Evaluations Figure 8 illustrates that there could be wide variations in the discomfort glare votes given by the subjects. Although the majority of the subjects assessments were Just uncomfortable with the luminance of 5600 cdm 2, there were 6 subjects voting as Uncomfortable or 70 Indoor Built Environ 2011;20:65 74 Yun et al.

7 Fig. 8. Histogram of glare response votes for the blank view. Intolerable and 10 subjects voting as Noticeable or Just perceptible. This implied that the subjective assessments of discomfort glare can differ from each other. Based on the above analysis, the subjects were classified into glare-sensitive and glare-insensitive groups. Glare-sensitive subjects were those who evaluated the blank view as Just intolerable or Intolerable when subjected to a luminance of 5600 cdm 2. Individuals who gave a vote less than 2 (i.e. Acceptable ) for the other simulated views with the same luminance were excluded from the glare-sensitive group. Consequently, 20 out of 48 subjects were classified as glare-sensitive. The glare-insensitive subjects met the following two conditions. First, their glare response votes for the blank view was less than or equal to 2 (i.e. Acceptable ) when subjected to a luminance of 5600 cdm 2 and their votes were not greater than 3.5 for the other views with the same luminance. Thirteen out of 48 subjects were categorized as glare-insensitive. Table 4 compares the subjective evaluation of discomfort glare between glare-sensitive and glare-insensitive groups. There were differences in the glare evaluations that were noted between the two groups. There was a general tendency for the glare-sensitive people to give a significantly higher glare response votes than the glareinsensitive people when the luminance level was 1800 cdm 2 or higher ( p-values50.05). For example, the glare-sensitive vote for the near man-made view with a luminance of 3200 cdm 2 was 2.81 (SD ¼ 0.25) and 1.73 (SD ¼ 0.47) was given by the glare-insensitive (t-test ¼ 8.538, p-value50.001). There was consistency in the discomfort evaluations given by the two groups, although the assessment criterion was different for each group (Table 4). The criterion on discomfort glare evaluation was maintained for each group and this criterion did not change with the type of views and luminance levels. The glare response votes given by the glare-sensitive group were consistently greater than the votes given by the glare-insensitive group for all the views considered in this study. This demonstrated that there were actual differences in glare sensitivities between the groups. However, the difference between the two groups when viewed at a luminance of 1000 cdm 2 was not always statistically significant. For example, the glare response vote of the distant man-made river view for the glaresensitive group was greater than that for the glareinsensitive group by 0.2 and this difference was not significant (t-test ¼ 1.391, p-value ¼ 0.18). On the other hand, the vote of the near man-made view was 1.65 (SD ¼ 0.43) for the glare-sensitive group and was 1.10 (SD ¼ 0.26) for the glare-insensitive group (t-test ¼ 4.190, p-value50.001). Figure 9 shows the distributions of the glare response votes for the glare-sensitive and glare-insensitive groups as a function of the luminance of the view. The variations in sensitivies to discomfort glare of the two groups were considerable. The glare-insensitive group experienced less discomfort glare than the glare-sensitive group when Influence of Window Views on Discomfort Glare Indoor Built Environ 2011;20:

8 Table 4. Glare response votes of the glare-sensitive and glare-insensitive groups for the blank and simulate views at luminance levels of 1000, 1800, 3200, 5600 and 10,000 cdm 2 View Luminance Sensitive Insensitive t-test p-value Glare scale SD Glare scale SD Mixed river distant , Mixed river near , Mixed land distant , Mixed land near , Natural land distant , Natural land near , Natural river distant , Natural river near , Man made distant , Man made near , Blank , Indoor Built Environ 2011;20:65 74 Yun et al.

9 (a) <1 Cumulative frequency (%) (b) Cumulative frequency (%) (c) Cumulative frequency (%) (d) Cumulative frequency (%) < < <1.5 <1.5 <1.5 Insensitive Sensitive <2 <2.5 <3 <3.5 Insensitive Sensitive <2 <2.5 <3 <3.5 <2 <2.5 <3 <3.5 Insensitive Sensitive <4.5 subjected to the same luminous conditions. For example, more than 26% of the votes from the glare-sensitive group were greater than 2.5 ( Acceptable ) at a luminance of 1800 cdm 2, while less than 2% of the glareinsensitive group gave 2.5. This tendency for the <4 <4 Insensitive Sensitive <4 <4.5 <4.5 <1 <1.5 <2 <2.5 <3 <3.5 <4 <4.5 Fig. 9. The cumulative frequency values of discomfort glare scale for the glare-sensitive and glare-insensitive groups: (a) 1800 cdm 2 luminance condition; (b) 3200 cdm 2 luminance condition; (c) 5600 cdm 2 luminance condition; and (d) 10,000 cdm 2 luminance condition. glare-sensitive group to experience more glare was also found in other luminance conditions. At a luminance of 10,000 cdm 2, the proportion of cumulative votes greater than 4 ( Just intolerable ) was 25% for the glareinsensitive group and 89% for the glare-sensitive group. Discussion and Conclusions This study has confirmed the effect of window views on the assessment of discomfort glare based on experiments using a simulated window that rendered blank, natural and man-made views from far to near distances. The tolerance of subjects to discomfort glare for simulated views were greater than the tolerance for a blank view. This confirms previous research on relationship between views and discomfort glare [7]. Although previous research on discomfort glare analysed physical factors such as luminance distributions in the field of view, the size of, and the distance from, glare source [9 14], discomfort glare from windows can depend on various factors. This study found that psychological factors such as window views could be an important factor in the subjective evaluation of discomfort glare. Widely used glare evaluation formulas such as UGR [15] and DGI [9] would not consider psychological factors. Thus, the finding from this study would contribute to the more realistic evaluation of discomfort glare. The results also indicated that there were noticeable variations in the subjective assessment of discomfort glare over the same visual stimuli. This was consistent with a study reported by Hopkinson [16]. Our study, in particular, revealed that there could be a group of subjects with a strong response to glare while another group of subjects were less sensitive to the same level of luminance. We also provided a preliminary criterion for classification of subjects sensitivity to glare and illustrated that the difference was statistically significant for the glaresensitive and glare-insensitive people. This implies that individual differences in assessing discomfort glare should be considered in the design process of daylighting systems. Acknowledgements This research was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology (No ). Influence of Window Views on Discomfort Glare Indoor Built Environ 2011;20:

10 References 1 Robbins CL: Daylighting: Design and Analysis. New York, Van Nostrand Reinhold, Yun GY, Shin HY, Kim JT: Monitoring and evaluation of a light-pipe system used in Korea: Indoor Built Environ 2010;19(1): Yun GY, Hwang T, Kim JT: Performance prediction by modelling of a light-pipe system used under the climate conditions of Korea: Indoor Built Environ 2010;19(1): Iwata T, Shukuya M, Somekawa N, Kimura K: Experimental study on discomfort glare caused by windows: J Archit Plan Environ Eng Trans Archit Inst Jpn 1992;432: Fisekis K, Davies M, Kolokotroni M, Langford P: Prediction of discomfort glare from windows: Lighting Res Technol 2003;35(4): Piccolo A, Simone F: Effect of switchable glazing on discomfort glare from windows: Build Environ 2009;44(6): Tuaycharoen N, Tregenza PR: Discomfort glare from interesting images: Lighting Res Technol 2005;37(4): Heerwagen J, Heerwagen D: Lighting and psychological comfort: Lighting Des Appl 1986;16(4): Hopkinson RG: Glare from daylighting in buildings: Appl Ergon 1972;3(4): Kim W, Kim JT: Effects of background luminance on discomfort glare in relation to glare source size: Indoor Built Environ 2010;19(1): Kim G, Kim JT: Visual environment within the already-built underground development in South Korea: Indoor Built Environ 2010;19(1): Dahlan ND, Jones PJ, Alexander DK, Salleh E, Alias J: Daylight ratio, luminance, and visual comfort assessments in typical Malaysian hostels: Indoor Built Environ 2009;18(4): Chow TT, Pei G, Chan LS, Fong KF: A comparative study of PV glazing performance in warm climate: Indoor Built Environ 2009;18(1): Snyder ME, Megri AC, Haghighat F: Analysis and design of a new roof monitor for day and night illumination of a large volume room: Indoor Built Environ 2008;17(5): CIE : Discomfort Glare in Interior Lighting. Vienna, Austria, CIE, Hopkinson RG: Evaluation of glare: Illum Eng 1957; 52: Indoor Built Environ 2011;20:65 74 Yun et al.