Monitoring Daylight Provision and Glare Perception in Office Environments

Transcription

1 Monitoring Daylight Provision and Glare Perception in Office Environments Painter, B. Institute of Energy and Sustainable Development, De Montfort University, UK ( Mardaljevic, J Institute of Energy and Sustainable Development, De Montfort University, UK ( jm@dmu.ac.uk) Fan, D. Institute of Energy and Sustainable Development, De Montfort University, UK, ( dfan@dmu.ac.uk) Abstract It is generally accepted that there is an urgent need to understand the link between daylight and building users visual comfort. However, current behavioural models do not adequately describe user experience under daylight conditions. This is largely due to the lack of real life monitoring data that includes both, physical measurements and qualitative values, i.e. from user feedback. A survey methodology is presented that links the measurement of physical conditions, using luminance maps derived from high dynamic range (HDR) images, with the collection of user perception data at normal office workstations. After an initial setup phase, the method requires minimal intervention by the researchers and is thus particularly suitable for long-term field studies. The main challenges in the development of this method arose due to the need to keep the interference with the participants usual work activities to a minimum while still providing data suitable for meaningful analysis. The paper describes how these issues were addressed. The practical application of the method in post-occupancy investigations is explored using a case study example. Keywords: high dynamic range imaging, HDR, luminance measurement, participant fatigue 148

2 1. Introduction It is generally acknowledged that a good provision of daylight is considered to be highly desirable in terms of building occupants well-being and productivity as well as its contribution to energy savings (e.g. Boyce et al., 2003 and Galiasu and Veitch, 2006). Daylight however can cause visual discomfort by inducing glare and veiling reflections. It can also result in high solar gains which may lead to uncomfortable thermal conditions and a need for mechanical cooling. It is therefore important to assess the effect of daylight ingress on visual and thermal comfort during building design. Although several formulations have been proposed to quantify the glare experienced by occupants in daylit environments (e.g. Hopkinson, 1972; CIE, 1992; Nazzal, 2001), they have been proved to be inadequate for accurately determining discomfort glare from daylight, mainly because these indices were derived from experiments that used artificial light sources (Osterhaus, 2005). It is generally accepted that the glare sensation experienced in real-world daylit environments is likely to differ significantly from those laboratory conditions (Velds, 2002; Tuaycharoen and Treganza, 2005). In order, therefore, to develop new metrics that account for people s perception of the luminous conditions in daylit environments, comprehensive new data sets are required that include physical measurements and qualitative values, i.e. from user feedback (Osterhaus, 2005). A first step towards this aim was made by Wienold and Christoffersen (2006) who monitored physical conditions and participant glare perception in a daylit laboratory environment and developed a promising new glare index, called Daylight Glare Probability. The authors acknowledge, however, that data from other environments are required to validate the new index. Additional visual comfort field studies are also needed to investigate whether people s perception of visual comfort is different outside laboratory conditions, and longer-term studies are required in order to study the effect of seasonal variability of daylight conditions on visual comfort perception. Based on these considerations, a method was developed for long-term glare studies in real working environments. The method uses standard components and automated data collection procedures together with sophisticated measurement techniques in order to gather quantitative and qualitative data of daylight conditions and glare perception at a relatively high measurement frequency. An overview of the technical details on the measurement techniques has been given in Fan et al. (2009) and preliminary results were presented in Painter et al. (2009). This paper focuses on a discussion of the challenges encountered during the development of the method. The solutions for individual aspects are described and an overview of a case study is presented in order to illustrate the practical application of the method and the data set derived with it. 2. Overview of the method The method is designed to simultaneously collect quantitative data and qualitative data. It automatically records field-of-view luminance maps and subjective glare ratings from workstation users (glare survey) at regular intervals. The following sections give a brief description of the main 149

3 components of the data collection setup. The considerations and design decisions made during the development of the method are discussed in section Quantit ative measurements Luminance measurements are taken with high dynamic range image capture. The camera is installed at the participant s workstation and positioned to capture field-of-view images of the workstation area as seen by the participant (Figure 1). Luminance maps containing high resolution measurement data can be generated from these images using High Dynamic Range (HDR) imaging techniques. The HDR approach works as follows: A sequence of images of the field-of-view is captured at different exposures. The images are then combined into one HDR image (luminance map) that provides a luminance value for each pixel in the scene (Reinhard et al., 2005). An HDR image contains the wide dynamic range of luminance conditions as seen by the human eye. It thus provides a record of the magnitude and spatial variation of the luminance in the captured scene. This approach provides a significant improvement over the traditionally used method of collecting spot measurements with a hand held photometer. As the whole scene can be captured, a more sophisticated data analysis is possible. For example, rather than selecting specific points for spot measurements, based on assumptions of where the main glare sources may be, complex glare sources (or other areas of interest) can be investigated by extracting the relevant pixels from the luminance maps and analysing them in terms of their luminance distribution and statistics (mean, max etc). HDR imaging techniques have been demonstrated to be accurate with a 10% error margin for a wide range of conditions (Inanici, 2005) and have already been used in daylight glare studies (e.g. Wienold and Christoffersen, 2006). The approach can be automated by using a computer to control the camera and create the HDR images. A description of the image capture and calibration procedures is given by Fan et al. (2009). Figure 1: Typical camera position and example of a monitored workstation. 150

4 2.2 Qual itative measurements A short, Java-based, survey form, which is displayed on the participant s computer screen, is used to record their feedback regarding the glare experienced when the quantitative measurements are taken. The survey form allows participants to mark the level of discomfort experienced by moving a slider control along a continuous scale that ranges from imperceptible to intolerable, similar to the one described by Osterhaus (2008). The participants then mark the source of the discomfort on a field-ofview image of their workstation environment. A comment box is provided for further feedback. 2.3 Networking and timing The data collection procedure is controlled by two programs, a client and a server application as shown in Figure 2. The survey software is installed on the participants desktop computer (client application), and run in the background when the machine is switched on. All tasks can be scheduled automatically at regular intervals. At the scheduled time, a dialog box appears on the participant s computer screen, asking them to either continue to the survey (Yes) or postpone it for 5 minutes (Later). If the subject chooses Yes the survey loads onto the screen. After submitting the survey, the subject response data are sent to the computer (MacMini) that controls the digital camera (the server application). This activates the capture software to carry out the HDR luminance measurements. Both, HDR image and participant s rating are saved to the MacMini. At the end of each day the survey data and HDR images are transferred via the wireless network to the main survey workstation for backup and archiving. Figure 2: Data collection applications 151

5 3. Practicabil ity considerations meeting researchers and participants requirements A number of issues had to be resolved during the development of the method in order to ensure that all requirements for the collection of the data set were met. Problems mainly arose due to the conflict between the desire for high frequency data collection and the need to keep the disturbance of the participants to a minimum. The next section outlines the technical requirements, followed by a discussion of the participant related considerations. 3.1 Researcher requirements for data collection Ideally, the method should meet the following requirements: Easy to install, low cost, easy to replicate. Suitable for long-term studies. Require minimal effort for researchers after initial setting up. Generate a data set suitable for automated analysis. In order to address the first point, the method was designed to use standard, off-the-shelf components, which included a digital SLR camera, a small computer to control the camera, standard fixing clamps and cables as well as a storage device for backing up the data. Data transfer is managed through scripts that utilize standard protocols of the wireless and wired networks. Due to the use of custom written programs for automated data capture and archiving, data can be collected over a long period of time with minimal input from the researchers. The on-screen survey and HDR image capture were designed to provide luminance measurements and user assessment data in a formats suitable for programmable data processing and analysis tools, in order to facilitate efficient handling of the large amounts of collected data. 3.2 Participant related design considerations The main aim was to develop a method that would allow measurement at high frequency in order to obtain a data set that was sufficiently large to be the basis for meaningful analysis. However, since it is more likely that participants provide accurate, useful feedback if they feel positive towards the study, it is crucial to avoid participant fatigue (Krosnick, 1999). This is particularly important for studies that require participation of study subjects in their usual working environments, as opposed to participants which have been specifically recruited to take part in (laboratory) experiments. Therefore, reducing participant disturbance and keeping the required interaction to a minimum was one of the main considerations during the design of the measurement method. Each aspect of the method design was thus affected by the conflict between having to keep the interference with the participants usual work activities to a minimum while still obtaining a useful data set. Table 1 gives an overview of the 152

6 participant related challenges that were encountered, together with the practical solutions that were found and applied in the case study. Details for each aspect are given in the following sections Ca mera position In order to ensure that the luminance data from the HDR images can be usefully linked to the glare rating reported by the participant via the survey, the camera for the luminance measurements needs to be installed as closely as possible to the occupant s seating position at head height. Although it has been shown that the deviations in angle and distance between the camera and the actual field of view position can result in significant measurement errors, experiments in a test room showed that these errors are limited if angle deviations are kept <30 (Fan et al., 2009). Experience from the pilot study shows that it is relatively easy to find a camera position that fulfilled this requirement and was acceptable for the participants Sur vey form Contents and layout The use of a long and detailed survey form to capture all aspects of the visual environment, as would typically be used for post-occupancy investigations, was not suitable for our study due to the time required to answer a relatively large number of questions. Since the aim was to collect survey responses very frequently (30 minute interval) over a long study period, there was a risk of the data quality being affected by participant fatigue. Participant fatigue, also called respondent burden, occurs if study participants tire of being asked for feedback (Krosnick, 1999). Particularly in long term studies this can affect the quality of the data. Therefore, the effort required by the participants to complete the survey at each measurement instance needed to be kept to a minimum (Sharp and Frankel, 1983). It was thus decided to focus solely on asking participants about their perception of glare. Table 1: Overview of main participant related issues encountered during method design Issue Ideal situation Constraints Best solution found Camera position Head position of participant to capture exact field-of-view Physically impossible, intruding into participant s personal space. Install camera as closely as possible/accessible to participant. Survey form Wide range of questions regarding luminous environment and comfort at workstation Participants time is limited. Participant fatigue due to repeatedly answering numerous questions. Short survey form, limited to inquiry about glare level and position of glare source. Training and feedback option provided. Measurement timing Every 10 minutes to increase chance of capturing glare conditions, e.g. sunny spells on overcast days Participant fatigue due to frequent survey prompts. Once every 30 minutes. Survey can be postponed. Survey trigger script can be modified to (1) adapt to personal preference and (2) take weather data as input. 153

7 Privacy No privacy issues Camera based method means that participants may feel uncomfortable about having their computer screen photographed or being photographed themselves. Study duration Covering diurnal variability (morningevening) Covering seasonal variability (summer/winter) Participant fatigue due to long period with repeated survey prompts. Participant s working hours may not cover required period. Participants can hide sensitive screen content before image capture is triggered. Training provided: participants are informed that image analysis is carried out programmatically, i.e. nobody looks at the photos. Tested in pilot study. Having optimized solutions for issues 1-4, participants continued to take part throughout 12-month study. It is acknowledged that this limits the scope of analysis, i.e. other sources of visual discomfort are neglected, but it allows more frequent, automated collection of data specific to the aim of the study, which was the recording of user ratings for daylight glare at workstations under naturally varying conditions. Additional information was collected with a one-off pre-study questionnaire/interview, which included questions regarding the participants demographics, their working patterns, use of task lighting and shading devices as well as their general perception of visual conditions at their workstation. During the development of the on-screen survey form, a number of different layouts were trialled and refined based on participant feedback. The layout finally used in the pilot study is very simple (Figure 3). It consists of a set of Yes/No radio buttons, a glare scale and an area to record the glare source position. Initially, only the main question Are you experiencing glare is active, with the rest of the form greyed out. This allows the participant to quickly focus on the question, without being presented with the additional detail which is irrelevant if no glare is being reported. If the answer is No, the participant only has to click on the Submit button on the bottom of the form to complete and submit the survey. If Yes is selected, the remaining components of the survey form become active, and the participant is prompted for details regarding the glare conditions he is experiencing. 154

8 Glare scale categories Values below the noticeable range of the scale apply if, having been asked, the participant can see that there is some glare in his/her field-ofview, but it does not affect him/her at all. The borderline just noticeable refers to lighting conditions which are uncomfortable but could be tolerated for the duration of a working day. If those conditions persisted longer, the participant would attempt to remedy the situation. The borderline just disturbing marks lighting conditions which the participant could tolerate while completing the present task (for approximately 15 to 30 minutes). If those glare conditions persisted longer, however, the participant would attempt to alter his/her working environment. The borderline just intolerable refers to a luminous environment with extreme glare which the participant cannot tolerate and in which he/she would require an immediate change of the lighting conditions in order to continue working. Figure 3: Example of survey form, also showing glare definitions based on Osterhaus (2008) Choice of glare scale The participant s perception of the level of glare experienced is reported on a continuous graphic scale, which ranges from imperceptible through noticeable and disturbing to intolerable. The category labels and the borderlines on the scale were derived based on those used by Osterhaus (2008), and are defined as shown in Figure 3. Although other definitions and scales have been used to record glare ratings (e.g. Iwata et al., 1992 and Velds, 2002), there is some debate as to which words to use to label the categories and borderlines (Fisekis, 2003). The Osterhaus scale seems to be the most consistent and sensibly worded. Moreover, a continuous scale has the advantage that it is easy to understand and potentially allows a more detailed comparison of participants glare perception. Since the survey is presented as an on-screen form, rather than a paper based one, it was possible to take advantage of the slider control available for Java forms. This is very intuitive and easy to operate, and provides a numerical output value for the slider position, thus facilitating automated data analysis of the large data set accumulated during the study. Glare source location 155

9 The participant is further asked to mark the location of the glare source on the field-of-view image provided on the survey form. The image is taken at the typical head position of the participant and thus provides a good representation of the visual area surrounding the workstation. It is inserted into the Java form before the study commences. To mark the source of glare on the image, the participant simply clicks on the image and drags the mouse, thus drawing a rectangular region around the glare source. Using this approach means that the participant can feed back where the glare occurred. It thus provides a far more practicable solution than alternatives, such as pre-selecting possible glare sources and adding them on the form as a list for the participant to choose from, or a text input option where the participant would enter a description of the glare source. The click-draw approach is intuitive and easy to use, even for casual computer users, and provides numerical data that can be automatically processed for analysis purposes. It is less error prone than transcriptions from text input fields and less restrictive than a list of possible sources. The coordinates of the rectangular region are saved out for use of the analysis, where they can be linked to the measured values on the HDR luminance maps. Comments Free text feedback can provide additional detail that cannot be captured by the glare scale and clickdraw image area. A comment box is provided on the form, which allows the participant to provide extra information, such as the occurrence of unusual conditions, details of additional sources, information regarding the arrangement of shading devices and the use of artificial lighting. During the pilot study participants were also encouraged to use the comment box to provide feedback about their satisfaction with the survey form and its timing interval. The survey form was trialled and found to be very easy to use by the participants. They felt that the effort to complete it did not overly interfere with their work since it can be completed very easily: two mouse clicks to submit the form if no glare is experienced and four clicks to report occurrence, level and location of glare if any is observed. The output of the form includes (1) binary feedback as to whether glare was experienced or not, (2) a numerical value for glare slider position on the continuous scale, (3) coordinates for a rectangular region around the main glare source in the participant s field of view and (4) additional comments added by the participant Measurement timing and privacy After trialling data collection with different measurement intervals, the 30 minute interval was chosen for use in the case study. It was found to be generally acceptable by the participants and still allowed frequent measurements, i.e. sufficient to capture the diurnal variability in conditions and to maximise the chance of measuring during glare conditions. One participant asked for a modification of the procedure since he felt that the survey interrupted his work too much and he was likely to not report accurate feedback if asked every day. As a compromise, a one week on, one week off approach was tried, which, according to his feedback, reduced the burden enough for him to be cooperative and positive towards the study during the weeks when the measurements did take place. The option of postponing the survey for a short period of time, e.g. five minutes, was added based on feedback during the initial trials of the survey form. Instead of displaying the survey form straight 156

10 away, a small dialog box appears first, asking the participant whether they are ready to complete the survey at that time. This allows the participant to postpone the survey in order to carry on with their current task. Also, they are able to control what information is present or on their screen during image capture. This gives the participant more control and helps to ensure that any concerns about privacy/security issues regarding the items visible on the screen are avoided. The scripts controlling the data capture can be easily modified and adapted to varying requirements, such as change in data capture timing mentioned above. Moreover, provision is made in the scripts to use input of cloud cover or sunshine data from a weather station to suppress measurements during periods where glare is unlikely to occur, e.g. consistent cloud cover/overcast skies. This helps to further reduce participant fatigue Setting up and training In order to ensure that participants remained cooperative during the long-term study, a thorough setup and training procedure was carried out. This included working with each participant to determine suitable camera position, demonstrating the operation of the survey form and explaining the definitions of the categories and boundary lines on the glare scale. A hard copy of the definitions was provided for quick reference during the study. An initial visual environment survey was carried out for each participant and workstation and the physical dimensions of the desk layout and window positions were recorded. It was explained to the participants how the HDR capture worked and that images would be analysed automatically. This addressed the privacy concerns associated with image capture in non-laboratory conditions, i.e. the privacy of the participants would be protected since researchers would not actually look at individual images but specifically written routines would extract relevant data for further analysis. Participants were also given guidance on the types of comments to include, such as happiness with study or information regarding unusual glare conditions. During the pilot we also asked for suggestions regarding improvements of the procedures and survey form. Throughout the duration of the study the setup, equipment and participant satisfaction/status was reviewed repeatedly through informal feedback sessions. 4. Practical application of the method The newly developed data collection method was applied during a one year study, which started in May The study was carried out in three mainly daylit open plan offices in the Queens Building at De Montfort University (Leicester, UK). Five workstations with different layouts were selected, including two workstations characterised by small windows, two workstations directly next to a large glazed façade and one directly below a small window that faces west into a courtyard. A digital camera was installed as closely as possible to each participant s seating position and the data capture, including glare survey and HDR luminance measurements, was triggered every 30 minutes during the working hours. A detailed description of the measurement setup and calibration procedures is given in Fan et al. (2009). 157

11 Although the high frequency and of the measurements placed a significant burden on the participants, they continued to take part throughout the 12 month study period. According to repeated informal discussions, they found the disturbance caused by the survey and data capture procedures acceptable and remained positive towards the study. After the initial setup procedure and participant training, data collection proceeded automatically. Very limited involvement on part of the researchers was required during the study. This included occasional informal talks with the participants and resolving a temporary problem with the wireless network. Table 2: Summary of initial results from Queens Building case study (Painter et al., 2009) Glare reports and measured values Glare ratings vs. existing glare metrics Results showed that, generally, the luminance values experienced at all workstations are relatively low for mainly daylit offices. Even at the workstations next to a glazed facade, which have the highest and most frequent exposure to direct daylight, relatively low illuminance values were recorded. This can be taken as a credit to the building design, which uses deep window reveals and overhangs to reduce the ingress of direct light. However, the survey responses show that glare is still regularly experienced by all participants. As would be expected, reports of glare occurrence and level of discomfort varied according to workstation layout. It was also found that personal preference had a strong effect, i.e. similar luminance conditions were rated quite differently by different participants. Values for the most typically used glare metrics were calculated from the luminance maps and compared with the glare ratings recorded during the study. No clear correlation could be found for any of the existing glare metrics, which confirms they are not suitable for the assessment of glare from daylight sources. Since the luminance maps contain high resolution data, it is possible to extract other parameters that better characterise the luminous environment at the workstations during glare and non-glare conditions. It was found that parameters such as the average or maximum luminance of the glare source in the field-of-view the vertical eye illuminance, as can be measured with the HDR method, may provide a basis for developing or improving glare metrics for the application in daylit environments. Almost 4800 survey responses and associated glare measurements were collected during the study. The data archive including all luminance maps and the associated glare ratings is kept on a dedicated storage device which can be accessed through the program routines used by the researchers for analysis. Table 2 gives an overview of the results from the study. A more detailed description of the results, including luminance conditions and frequency of glare being recorded is given in Painter et al. (2009). There are of course other types of studies that require the simultaneous collection of qualitative and quantitative data from building occupants. For example, the method could be modified to include aspects of thermal comfort as well as visual comfort and thus become a valuable tool for integrated post-occupancy investigations. 158

12 5. Conclusions The data collection method described here has been designed to facilitate automated collection of qualitative and quantitative data for glare studies. The method uses HDR imaging techniques to measure luminance at office workstations and collects occupant perception data via an on-screen glare survey form. The method has been refined in order to reduce participant fatigue and is therefore suitable for medium and long-term studies. Due to the use of standard components and protocols the method is easy to implement and adaptable for specific requirements. The method has been used successfully in a 12 month study of glare perception. It was found that data collection was truly automated, requiring minimal intervention of the researchers after the initial setup. The method has provided a data set that allows linking physical condition with building occupant assessments. The data set now needs to be expanded, i.e. more measurements in other office layouts and for other demographics are necessary. The described method is the ideal tool for this purpose due to its easy installation and limited effort required after setup. The method can be modified for other purposes that require collection of occupant feedback simultaneously with measurement of physical conditions, such as post-occupancy studies. References Boyce P, Hunter C and Howlett O (2003) The Benefits of daylight through windows, Review from the Daylight Dividends Programme, Lighting Research Centre, Rensselaer Polytechnic Institute, NY, (accessed March 2010). CIE (1992) Discomfort Glare in the Interior Lighting, Technical Committee TC-3.13, Division 4, Interior Environment and Lighting Design. Fan D, Painter B and Mardaljevic J (2009) A data collection method for long-term field studies of visual comfort in real-world daylit office environments, Proc. of PLEA 2009, Quebec. Fisekis K, Davies M, Kolokotroni M and Langford P (2003) Prediction of discomfort glare from windows, Lighting Research and Technology 35(4): Galasiu A D and Veitch J A (2006) Occupant preferences and satisfaction with the luminous environment and control systems in daylit offices, Energy and Buildings 38(7): Hopkinson, R. G., 1972, "Glare from daylighting in buildings," Applied Ergonomics 4. Inanici M (2005) Per-pixel lighting data analysis, Technical Report, Ernest Orlando Lawrence Berkeley National Laboratory, Berkeley, U.S.A. Iwata T, Tokura M, Shukuya M and Kimura K-I (1992) Experimental study on discomfort glare caused by windows, International Journal of Architecture 439:

13 Krosnick J A (1999) Survey Research, Annual Review of Psychology 50: Nazzal A A (2001) A new daylight glare evaluation method: Introduction of the monitoring protocol and calculation method, Energy and Buildings 33: Newsham G R and Arsenault C D (2009) A Camera as a sensor for lighting and shading control, Lighting Research & Technology, 41(2): Osterhaus W K E (2005) Discomfort glare assessment and prevention for daylight applications in offices, Solar Energy 79: Osterhaus W K E (2008) Analysis of luminance histograms for the assessment of discomfort glare in daylit office, Proc. of Balkan Light 2008, Ljubljana, Slowenia, 8-10 October Painter B, Fan D and Mardaljevic J (2009) Evidence-based daylight research: Development of a new visual comfort monitoring method, Proc. of Lux Europa 2009, Istanbul, Sept Reinhard E, Ward G, Pattanaik S and Debevec P (2005) HDR Image Capture, in High Dynamic Range Imaging: Acquisition, Display, and Image-Based Lighting, Morgan Kaufmann. Sharp L M and Frankel, J (1983) Respondent burden: A test of some common assumptions. The Public Opinion Quarterly 47(1): Tuaycharoen N and Tregenza P R (2005), Discomfort glare from interesting images, Lighting Research & Technology 37(4): Velds M (2002) User acceptance studies to evaluate discomfort glare in daylit rooms, Solar Energy 73(2): Wienold J and Christoffersen J (2006), Evaluation methods and development of a new glare prediction model for daylight environments with the use of CCD cameras, Energy and Buildings 38: