- Email: [email protected]
Switching frequency and energy analysis for photoelectric controls
Building and Environment 85 (2015) 205e210
Contents lists available at ScienceDirect
Building and Environment journal homepage: www.elsevier.com/locate/buildenv
Switching frequency and energy analysis for photoelectric controls Danny H.W. Li*, Angela C.K. Cheung, Stanley K.H. Chow, Joseph C. Lam Building Energy Research Group, Department of Architecture and Civil Engineering, City University of Hong Kong, Tat Chee Avenue, Kowloon, Hong Kong Special Administrative Region
a r t i c l e i n f o
a b s t r a c t
Article history: Received 7 October 2014 Received in revised form 18 November 2014 Accepted 22 November 2014 Available online 29 November 2014
In a well day-lit space when the daylight intensity is far more than the required value, photoelectric switching can result an excellent energy saving. However, a problem with the switching control type is the frequent and rapid switching of lights on and off, particularly during the unstable sky conditions when daylight levels are fluctuating around the switching lighting level. This can annoy occupants and lessen lamp life. This paper studies the simple and a few variants of photoelectric switching controls namely standard, differential, switching time delay, daylight time delay and solar reset types which can affect the number of switching operations and the corresponding lighting energy use. For a given time delay, daylight linked time delay performed better than switching linked time delay in terms of the number of switching operations. Several mathematical expressions were established to predict the lighting energy savings at various atrium floors. The findings provide important information and insight into the effects of different switching controls in terms of energy savings and real-time operations. © 2014 Elsevier Ltd. All rights reserved.
Keywords: Lighting controls Switching Daylight availability Atrium Corridor
1. Introduction Daylighting is an essential sustainable development to alleviating the problems in energy and pollution, and enriching the environment and visual comfort [1e4]. Daylight is the best source of light for good colour rendering and most closely matches human visual responses. The amount of natural daylight entering an interior space is mainly via window openings which give visual connection between internal and external environments, and brightens the internal spaces. For a good indoor space, as much glare free light as possible is essential [5]. The energy savings derived through the use of daylighting not only facilitate the sparing use of electric lighting and reduced peak electrical demand, but also reduce cooling loads and offer the potential for smaller airconditioning plants to be built [6,7]. It is argued that photoelectric controls should be installed for lamp fittings installed in day-lit spaces. Energy reduction can be achieved when the brightness from daylight is more than the design value. It can be attained with proper daylight-linked lighting controls to shut off or dim down the electric lightings such that the indoor illuminance levels can still meet the required values [8]. Generally, photoelectric dimming controls are more effective than
* Corresponding author. Tel.: þ852 34427063; fax: þ852 34420427. E-mail address: [email protected] (D.H.W. Li). http://dx.doi.org/10.1016/j.buildenv.2014.11.022 0360-1323/© 2014 Elsevier Ltd. All rights reserved.
photoelectric switching control particularly at high design illuminance levels [9]. In a well daylit room or low setting illuminance levels when daylight intensities are far more than the required values, the energy savings from photoelectric switching controls can be larger than those from the photoelectric dimming controls [10]. However, a problem with the oneoff control type is the frequent switching, annoying occupants and reducing the lamp life [11,12]. It is crucial to visual comfort, building energy and cost aspects. In circulation areas such as corridors, people expect the way ahead to be lit sufficiently. It has been reported that in daylit corridors photoelectric lighting controls can give excellent energy savings [13,14]. Daylighting design techniques are often best illustrated via field measurements to provide reliable operational data and establish design strategies [15e20]. It is essential to indicate the energy savings and the number of switching operations. This paper analyses the electric lighting energy savings and switching frequency for a number of daylight-linked lighting switching controls. Well day-lit corridors located in an atrium building were selected for the study. The findings are reported and design implications discussed. 2. Photoelectric lighting controls The methods of controlling lighting energy consumption can be classified into two basic categories. The first type of control allows the level to be set between maximum and minimum levels by
206
D.H.W. Li et al. / Building and Environment 85 (2015) 205e210
dimming (top-up), and the second type provides for either an on or off state. Automatic oneoff and dimming lighting controls can be employed in daylighting schemes. Dimming controls vary the light output of lamps in accordance with the prevailing daylight level. When daylight is insufficient to achieve the required design illuminance, the indoor lighting level is topped up by artificial lighting. Recently, high frequency dimming controls have been increasingly used and the electronic circuitry employed is more energy efficient than conventional ballast [21]. A high frequency dimming control does not have the ideal characteristic of light output being perfectly proportional to power consumed. The light output can be roughly assumed proportional to power consumed but the lamps cannot be dimmed to total extinction [22]. In normal operation, their residual light output and power consumption occur throughout working hours even if the illuminance level far exceeds the design value [23,24]. However, such operations may be less noticeable and less disturbing to occupants. An oneoff daylight-linked lighting control is designed to switch electric lighting on and off automatically as the daylight level falls and rises through a predetermined level. However, the main drawback with this lighting control type is the rapid and frequent switching of lights on and off, particularly during unstable weather conditions when daylight levels are fluctuating around the switching illuminance. This can disturb occupants and shorten the lamp life. Good switching controls would be the lighting system always off or the number of switching is very low. There are a few variants of the simple oneoff control to lower the number of switching [11,25]. The basic one is the ‘differential switching or dead-band’ photoelectric control that has two switching illuminance levels: one at which the lights are switch on (Eon), and another of higher illuminance, at which the lights are switched off (Eoff). The main strength is that it can reduce rapid switching on and off when the illuminance swings around the desired level. Also, it makes switching off less obtrusive, as it is performed when daylight represents a higher proportion of the illuminance to which the eye is accommodated [11]. It is very important to set appropriately the two switching illuminance values which affect both the number of switching and the electric lighting energy consumption. There is no simple analytical expression for lighting use under a differential switching control. To a good approximation, the fraction of the working year that electric lighting would be off under an oneoff control is simply given by the fraction of the working year that the daylight threshold illuminance level (Emean) which is the average of the two switching illuminances Eon and Eoff [11]. Another technique to decrease the number of switching is to introduce a time delay into the process. Two different types of time delay namely switching-linked and daylight-linked time delays are categorized. The switching-linked time delay only allows switching off until the pre-set delay had to elapse after the last switch on. For daylightlinked time delay, the lighting could only be switched off when the daylight illuminance had exceeded the target value for the preset time period. Delay in switching on is not considered as it could lead to illuminances falling well below desired levels. The third control mode is solar reset. In solar reset switching, the daylight illuminance is sampled at certain set times of day. When the daylight illuminance is more than the switching illuminance at any of the solar reset times, then the lighting is switched off. To ensure that the illuminance levels are not lower than the desired values, the electric light fittings will be automatically switched on whenever the daylight illuminance falls below the design values. There are several advantages for solar reset switching. Firstly, the switching frequency can be limited as the solar reset time can be set at certain periods. Moreover, the occupants can face with switching off at expected times, and the reset times can coincide with their schedule such as during lunchtime and afternoon breaks which
may be appropriate to offices and schools. For a very short reset time interval, the performance would be close to the standard photoelectric control. 3. Background information The analysis was based on the long-term measured daylight illuminance data in an institutional building. It is a 13-storey block located in Hong Kong within the subtropical region at latitude of 22.3 N and longitude of 114.2 E. The building was designed with a skylight and an enclosed stepped atrium to harvest daylight. The stepped atrium increases daylight and improves sky views by splaying the well walls away from the vertical, but the sky component and hence the daylight factor of the lower floors are significantly reduced [26]. The lower part of the atrium between 2/F and 9/F contains typical classrooms, with four open corridors surrounding the atrium. The upper part, from 10/F to 13/F, consists of largely mechanical rooms and offices. The atrium corridors are used for circulation among rooms at different rooms. There are 67 numbers of ceiling-mounted energy-efficient T5 fluorescent tubes, with rated power ranging from 14 W to 35 W for the corridors at every floor. Daylight-linked dimming controls were installed in the corridors at 9th floor. The system detected both the reflected electric light and the daylight values to give a ‘closed loop’ control. The recorded lighting levels were sent to the dimmable electronic ballasts which adjust the light outputs of the lamp fittings accordingly. Four photoelectric sensors were mounted on the ceiling of the four corridors to record the light intensity. The daylight performance and energy use due to the daylight-linked dimming controls in the 9/F corridors were examined and reported [23,24]. There are three additional ceiling mounted photoelectric light level sensors installed respectively along the same plummet line from 6/F to 8/F to record the daylight illuminance. The data were transmitted to a logging system for storage. The recorded illuminance readings only provided the daylight availability for the corridors and the sensor itself did not form any part of the lighting controls. The present study analyzes the various types of control algorithms, namely differential switching, time delay and solar reset based on the daylight illuminance data measured in the atrium corridors. 4. Data analysis and simulation The daylight illuminance data were recorded at 2.5 min intervals from 9:00 am to 6:00 pm. each day between February 2012 and July 2013. Totally, there are 216 daylight illuminance readings collected for each photo-sensor every day without any missing record. The analysis was on daily basis. The whole day illuminance data would be rejected if some data were omitted. It is inevitable that there would be some short periods of missing data due to different reasons including instrumentation malfunction and power failure. Considerable efforts were made to obtain a continuous record of data and in all about 257,000 data were recorded for use. The illuminance levels at a particular time were assumed to have applied for the whole of the time interval of 2.5 min since the previous scan. As pointed out by Littlefair [11] that such an assumption in practice might cause to a slight underestimate of the switching numbers during periods of very substantial daylight illuminance variation (i.e. on a less than 2.5 min timescale). Under a standard switching control, the daylight illuminance recorded by every photocell at each time was checked and if this was less the target illuminance (Et), the control was assumed to have kept the lights on since the last scan. If it was larger than the Et, the lights were assumed to have been off. At the end of the working day, the percentage of time the lights off was computed. The number of
D.H.W. Li et al. / Building and Environment 85 (2015) 205e210
switch offs was counted by keeping a record of when the control changed state and the switch off at the end of the day was excluded in the totals. If at 09:00 the daylight illuminance was above Et then the lighting was assumed to be switched off at first. It should be pointed out that no electric lighting was actually operated under such switching controls. The results were simulated according to the measured daylight illuminance under such a ‘perfect’ control. It means switching exactly when daylight reached Et. The cumulative frequency distribution of daylight availability can show the percentage of the working year in which a given illuminance is exceeded. Fig. 1 presents the cumulative frequency distributions for the daylight illuminance data at the 8th, 7th and 6th Floors. For all setting illuminance values, 8th Floor scored the largest frequency, 7th Floor second and 6th Floor the third. The cumulative frequencies for the 3 floors at individual illuminance levels are valuable for determining the necessary use of artificial lighting, and the probable energy savings from a standard oneoff control. Table 1 summaries the range and the average number of switch offs per day and the percentage of lighting energy saving at these three floors under the standard switching control for the switching illuminances of 100, 125 and 150 lux. The recorded minimum number was zero for all cases but the daily maximum number of switch offs can be up to 29 which imply at least 57 switching operations in a day if switch ons are also included. Such a large number of switching operations could result in remarkable annoyance to occupants. The daily average number of switch offs ranged between 4.7 and 6.8. At the recommended illuminance level for a corridor of 100 lux, the percentage of lighting energy savings were 81.7%, 83.5% and 90.2% at 6th, 7th and 8th Floor, respectively. The lighting energy saving was quite good but the switching number was not acceptable. To reduce the number of switching operating, differential switching controls at target illuminance (Eon ¼ 100 lux) were considered. Five different illuminance combinations with Eoff ¼ 110, 120, 150, 180 and 200 lux were examined and Table 2 gives the results. The number of switch offs was considerably reduced using differential switching controls. The peak switch offs ranged from 24 at 8th Floor when Eoff ¼ 110 to 6 at 6th Floor when Eoff ¼ 200 lux. When the Eoff was 10 lux more than Eon (i.e. Eon ¼ 100 lux and Eoff ¼ 110 lux) the average number of switch offs dropped gently ranging between 4.3 and 4.9. When the Eoff ¼ 200 lux the average
207
Table 1 Number of switch offs per day and percentage of working day (09:00e18:00) that the lights are switched off under standard switching controls. Control switching illuminances (lux)
6/F
7/F
8/F
Range Average number % of working day Range Average number % of working day Range Average number % of working day
100
125
150
0e23 5.4 81.7% 0e27 5.6 83.5% 0e27 4.7 90.2%
0e22 6.2 68.1% 0e28 6.8 76.3% 0e28 5.3 86.4%
0e27 6.4 51.3% 0e25 6.7 68.9% 0e29 5.5 82.5%
numbers of switch offs were 2.7, 2 and 0.9 at 8th, 7th and 6th Floors, respectively. Lower the switch offs means less the lighting energy savings. The lighting energy saving was 54.7% when the average number of switch offs was 0.9. As mentioned, the electric lighting energy under the differential switching control can be considered as the fraction of the working year of the mean values of Eon and Eoff (Emean) [11]. When the Emean ¼ 105 lux (i.e. Eon ¼ 100 lux and Eoff ¼ 110 lux) the simulated lighting energy saving was 2.3% less than when Et ¼ 100 lux under the standard oneoff control. For the other two Emean of 125 lux (i.e. Eon ¼ 100 and Eoff ¼ 150 lux) and 150 lux (i.e. Eon ¼ 100 lux and Eoff ¼ 200 lux) the difference was within 3.5% from the corresponding switching illuminances (i.e. Et ¼ 125 lux and Et ¼ 150 lux) under the standard switching control. The findings were in good agreement with other independent studies [11]. Two different types of time delay viz. switching and daylight linked time delay photoelectric controls were studied. For individual control type, six different time delays from 5 up to 60 min at switching illuminance of 100 lux were considered. Tables 3 and 4 show the switch offs per day and percentage of energy saving for respectively switching and daylight linked time delay photoelectric controls. A short time delay (e.g. 10 and 15 min) gives a substantial reduction in switching particularly for the daylight linked time delay control. For the same average switch offs per day, the maximum switch offs using time delay controls were far less than those using the differential control. It shows that time delay
Fig. 1. Cumulative frequencies of the indoor daylight availability for 6th, 7th and 8th Floors.
208
D.H.W. Li et al. / Building and Environment 85 (2015) 205e210
Table 2 Number of switch offs per day and percentage of working day (09:00e18:00) that the lights are switched off under differential switching controls. Control switching illuminances (lux)
6/F
7/F
8/F
Range Average number % of working day Range Average number % of working day Range Average number % of working day
100/110
100/120
100/150
100/180
100/200
0e19 4.3 79.4% 0e23 4.9 82.2% 0e24 4.3 89.6%
0e15 3.2 76.9% 0e21 4.4 81.3% 0e24 4.2 89.3%
0e11 1.8 68% 0e18 2.9 77.5% 0e19 3.3 87.3%
0e6 1.2 58.8% 0e15 2.4 74.7% 0e19 3 85.5%
0e6 0.9 54.7% 0e13 2 72.4% 0e17 2.7 84.1%
controls give a smaller variation on number of switching. The number of switch offs drops with longer time delay but for very long time delays such as an hour the return diminishes. The lighting energy saving is related to the number of switch offs. More switch operations means less lighting energy consumption. Based on these two control strategies, daylight linked time delay performed better than switching linked time delay in terms of the number of switch offs. For the same number of switch offs, daylight linked time delay always has a shorter time delay. For the time delay of 60 min, the average number of switch offs was around 2 for the switching linked time delay control and 1.3 for the daylight linked time delay control. The lighting energy savings ranged from 57.8% to 88.9% depending on the number of switch offs. The last photoelectric control for analysis is the solar reset switching. Four different control patterns were evaluated: a single reset at 10:30, two resets per day at 10:00 and 12:00; a 2 hourly reset and an hourly reset. Table 5 presents the number of switch offs and the percentage savings for above four solar reset types. As expected, the number of switch offs per day was dropped to very low value. The maximum numbers of switch offs were less than 7 for all the four solar resets. The average number of switch offs for hourly and the single resets were 2 and around 1, respectively. As the number of resets raises so does the number of switch offs. For the hourly reset, the number of switch offs is comparable to those for the time delay and differential switching controls. The percentage savings are affected by the number of switching. The savings ranged from 59% to 79% when the average number of switch offs was between 1 and 2. The electric lighting energy savings under various daylight linked switching controls strongly relate to the corresponding number of switch offs. As observed the data in Tables 1e5, lowered number of switch offs resulting less lighting energy savings and vice versa. Figs. 2e4 depict the correlations between the
Table 3 Number of switch offs per day and percentage of working day (09:00e18:00) that the lights are switched off under switching linked time delay controls at switching illuminance of 100 lux.
Table 4 Number of switch offs per day and percentage of working day (09:00e18:00) that the lights are switched off under daylight linked time delay controls at switching illuminance of 100 lux. Time delay (minutes)
6/F
7/F
8/F
Range Average number % of working day Range Average number % of working day Range Average number % of working day
7/F
8/F
Range Average number % of working day Range Average number % of working day Range Average number % of working day
10
15
30
60
0e10 2.9 74.6% 1e11 2.9 76.2% 1e12 2.6 83.6%
0e8 2.5 72.2% 1e10 2.5 73.7% 1e8 2.3 81.4%
0e6 1.9 66.2% 0e9 1.9 67.7% 0e6 1.8 75.7%
0e4 1.3 57.8% 0e4 1.3 59.2% 0e4 1.3 67.3%
percentage savings and average number of switch offs per day for the 6th, 7th and 8th Floors, respectively. The scatter of data points in the figures can be explained by the variety of energy savings for a particular switching off numbers. Through regression analysis, three simple mathematical equations to relate the average switch offs per day (N) and the lighting energy saving (LES) for the three floors were obtained.
LES ¼ 0:15 Ln N þ 0:576 LES ¼ 0:141 Ln N þ 0:6
R2 ¼ 0:96 R2 ¼ 0:93
LES ¼ 0:156 Ln N þ 0:672
6th Floor
(1)
7th Floor
R2 ¼ 0:93
(2)
8th Floor
(3)
All the regression equations have high coefficients of determination (R2), ranging from 0.93 to 0.96. It indicates that 93e96 % of the variations in percentage of lighting energy savings can be explained by the variations in the average number of switch offs. The strength of correlation is thus considered very good. The mathematical models can provide an alternative to estimate electric lighting energy saving from a given number of switch offs under various daylight linked switching control such that the appropriate type can be selected. 5. Conclusions Photoelectric switching can save energy by shutting down lamp fittings when daylight illuminance is sufficient. However, a problem with such control type is the frequent switching, annoying building users and reducing the lamp life. The best strategy should be a small number of switching off with a high saving. Field
Table 5 Number of switch offs per day and percentage of working day (09:00e18:00) that the lights are switched off under solar reset controls at switching illuminance of 100 lux.
Time delay (minutes)
6/F
5 1e14 3.5 77.5% 1e14 3.5 79.1% 1e16 3.1 86.4%
Reset type
5
10
15
30
60
1e19 4.9 80.6% 1e22 5.1 82.3% 1e25 4.2 88.9%
1e17 4.3 78.7% 1e18 4.4 80.3% 1e19 3.7 86.9%
1e14 3.8 77.1% 1e15 3.9 78.6% 1e16 3.3 85.2%
1e11 3.1 72.8% 1e11 3.1 74% 1e11 2.7 80.5%
1e6 2.2 67.2% 1e7 2.3 68.1% 1e7 2 73.2%
6/F
7/F
8/F
Range Average number % of working day Range Average number % of working day Range Average number % of working day
Single 10:30
10:00 and 12:00
2 hourly
Hourly
0e2 1.1 59% 0e2 1.1 60% 0e2 1.2 69%
0e3 1.3 64% 0e3 1.3 65% 0e3 1.3 73%
0e5 1.6 66% 0e5 1.6 68% 0e5 1.6 76%
0e7 2 70% 0e7 2 72% 0e7 1.9 79%
D.H.W. Li et al. / Building and Environment 85 (2015) 205e210
209
Fig. 2. Percentage of energy savings under various average switch offs per day using different forms of switching controls at 6th Floor.
measurements of the daylight illuminance in atrium corridors were undertaken. Based on the measured daylight availability, a number of daylight linked switching controls namely the standard, differential, switching and daylight time delays and solar reset oneoff controls were analysed. Their performances in terms of lighting energy reduction and number of switching were simulated and evaluated. Under the standard oneoff control, the energy saving was very good but the frequent switching was unaccepted. At the switching illuminance of 100 lux, the lighting energy savings were found ranging from 81.7% to 90.2% and the average number of switch offs between 4.7 and 5.6. Based on the simulated results, the variants of the switching control can reduce the number of switching but increase the energy consumption. Using the differential switching control, the number of switch offs can be dropped
to 0.9 with an energy saving of 54.7% at 6th Floor when Eon ¼ 100 lux and Eoff ¼ 200 lux. The switching and daylight linked time delay controls were examined. Provided an appropriate delay time is chosen, a daylight linked time delay can result less number of switch offs with lower energy savings. For solar reset switching, four different control patterns were studied. The average number of switch offs under the single reset was just about 1.1 for the three floors. The correlation between the percentage of lighting energy savings and the number of switch offs was evaluated. Three simple mathematical models for the three floors were developed. The findings can give an alternative to estimate the electric energy savings from the selected number of switch offs based on the various photoelectric switching controls. The important issue would be whether the number of switch offs can be achieved by a
Fig. 3. Percentage of energy savings under various average switch offs per day using different forms of switching controls at 7th Floor.
210
D.H.W. Li et al. / Building and Environment 85 (2015) 205e210
Fig. 4. Percentage of energy savings under various average switch offs per day using different forms of switching controls at 8th Floor.
particular switching control. Photoelectric electric lighting controls are appropriate building energy conservation schemes and should therefore be widely applied in day-lit areas. Further research work is required to fit the approach for various daylighting design criteria such as daylight factor so that the analysis would be more generalized. The theory proposed has been supported by on-site measurements in a building which is fully operational. This has not been done before. Given the growing concern about energy use in buildings and its likely impact on the environment, this study is timely and worthwhile. It is believed the work and its findings presented in this paper would be of use to the researchers and building professionals, particularly those who are concerned about energy use in buildings and its impact on the environment. Although the work was conducted in subtropical Hong Kong, the procedures and techniques developed could be applied in other locations. Acknowledgements The work described in this paper was fully supported by a General Research Fund from the Research Grant Council of the Hong Kong Special Administrative Region, China [Project no. 9041896 (CityU 117713)]. References [1] Chiogna M, Albatici R, Frattari A. Electric lighting at the workplace in offices: efficiency improvement margins of automation systems. Light Res Technol 2013;45(5):550e67. [2] Hwang T, Jeong TK. Effects of indoor lighting on occupants' visual comfort and eye health in a green building. Indoor Built Environ 2011;20(1):75e90. [3] Iversen A, Nielsen TR, Svendsen SH. Illuminance level in the urban fabric and in the room. Indoor Built Environ 2011;20(4):456e63. [4] Roetzel A, Tsangrassoulis A, Dietrich U. Impact of building design and occupancy on office comfort and energy performance in different climates. Build Environ 2014;71:165e75. [5] Li DHW, Tsang EKW. An analysis of daylighting performance for office buildings in Hong Kong. Build Environ 2008;43(9):1446e58. [6] Li DHW, Lam JC, Wong SL. Daylighting and its implications to overall thermal transfer value (OTTV) determinations. Energy 2002;27(11):991e1008. [7] Li DHW, Lam JC, Wong SL. Daylighting and its effects on peak load determination. Energy 2005;30(10):1817e31.
[8] Li DHW, Mak AHL, Chan WWH, Cheng CCK. Predicting energy saving and life cycle cost analysis for lighting and daylighting schemes. Int J Green Energy 2009;6(4):359e70. [9] Lynes JA, Littlefair PJ. Lighting energy savings from daylight: estimation at the sketch design stage. Light Res Technol 1990;22(3):129e37. [10] Li DHW, Cheung KL, Wong SL, Lam TNT. An analysis of energy-efficient light fittings and lighting controls. Appl Energy 2010;87(2):558e67. [11] Littlefair PJ. Photoelectric control: the effectiveness of techniques to reduce switching frequency. Light Res Technol 2001;33(1):43e58. [12] Li DHW, Lam JC. An analysis of lighting energy savings and switching frequency for a daylit corridor under various indoor design illuminance levels. Appl Energy 2003;76(4):363e78. [13] Li DHW, Lam JC. An investigation of daylighting performance and energy saving in a daylit corridor. Energy Build 2003;35(4):365e73. [14] Li DHW, Tsang EKW. An analysis of measured and simulated daylight illuminance and lighting savings in a daylit corridor. Build Environ 2005;40(7): 973e82. [15] Galasiu A, Atif MR. Applicability of daylighting computer modeling in real case studies: comparison between measured and simulated daylight availability and lighting consumption. Build Environ 2002;37(4):363e77. [16] Littlefair PJ, Aizlewood ME. Measuring daylight in real buildings. In: Proceedings of the CIBSE national lighting conference, Bath, UK; 1996. [17] Li DHW, Lam JC. Evaluation of lighting performance in office buildings with daylighting controls. Energy Build 2001;33(8):793e803. [18] Li DHW, Lam JC, Lau CCS, Huan TW. Lighting and energy performance of solar film coating in air-conditioned cellular offices. Renew Energy 2004;29(6): 921e37. [19] Li DHW, Lam TNT, Wong SL. Lighting and energy performance for an office using high frequency dimming controls. Energy Convers Manag 2006;47(9e10):1133e45. [20] Li DHW, Lam TNT, Wong SL, Tsang EKW. Lighting and cooling energy consumption in an open plan office using solar film coating. Energy 2008;33(8): 1288e97. [21] Slater AI. Energy efficiency in lighting. Energy Manag 1991;12(2):24e6. March/April. [22] Ihm P, Nemri A, Krarti M. Estimation of lighting energy savings from daylighting. Build Environ 2009;44(3):509e14. [23] Chow SKH, Li DHW, Lee EWM, Lam JC. Analysis and prediction of daylighting and energy performance in atrium spaces using daylight-linked lighting controls. Appl Energy 2013;112:1016e24. [24] Li DHW, Cheung ACK, Chow SKH, Lee EWM. Study of daylight data and lighting energy savings for atrium corridors with lighting dimming controls. Energy Build 2014;72:457e64. [25] Littlefair PJ. Predicting lighting energy use under daylight linked lighting controls. Build Res Inf 1998;26(4):208e22. [26] Samant S. A critical review of article published on atrium geometry and surface reflectances on daylighting in an atrium and its adjoining spaces. Archit Sci Rev 2010;53(2):145e56.
Contents lists available at ScienceDirect
Building and Environment journal homepage: www.elsevier.com/locate/buildenv
Switching frequency and energy analysis for photoelectric controls Danny H.W. Li*, Angela C.K. Cheung, Stanley K.H. Chow, Joseph C. Lam Building Energy Research Group, Department of Architecture and Civil Engineering, City University of Hong Kong, Tat Chee Avenue, Kowloon, Hong Kong Special Administrative Region
a r t i c l e i n f o
a b s t r a c t
Article history: Received 7 October 2014 Received in revised form 18 November 2014 Accepted 22 November 2014 Available online 29 November 2014
In a well day-lit space when the daylight intensity is far more than the required value, photoelectric switching can result an excellent energy saving. However, a problem with the switching control type is the frequent and rapid switching of lights on and off, particularly during the unstable sky conditions when daylight levels are fluctuating around the switching lighting level. This can annoy occupants and lessen lamp life. This paper studies the simple and a few variants of photoelectric switching controls namely standard, differential, switching time delay, daylight time delay and solar reset types which can affect the number of switching operations and the corresponding lighting energy use. For a given time delay, daylight linked time delay performed better than switching linked time delay in terms of the number of switching operations. Several mathematical expressions were established to predict the lighting energy savings at various atrium floors. The findings provide important information and insight into the effects of different switching controls in terms of energy savings and real-time operations. © 2014 Elsevier Ltd. All rights reserved.
Keywords: Lighting controls Switching Daylight availability Atrium Corridor
1. Introduction Daylighting is an essential sustainable development to alleviating the problems in energy and pollution, and enriching the environment and visual comfort [1e4]. Daylight is the best source of light for good colour rendering and most closely matches human visual responses. The amount of natural daylight entering an interior space is mainly via window openings which give visual connection between internal and external environments, and brightens the internal spaces. For a good indoor space, as much glare free light as possible is essential [5]. The energy savings derived through the use of daylighting not only facilitate the sparing use of electric lighting and reduced peak electrical demand, but also reduce cooling loads and offer the potential for smaller airconditioning plants to be built [6,7]. It is argued that photoelectric controls should be installed for lamp fittings installed in day-lit spaces. Energy reduction can be achieved when the brightness from daylight is more than the design value. It can be attained with proper daylight-linked lighting controls to shut off or dim down the electric lightings such that the indoor illuminance levels can still meet the required values [8]. Generally, photoelectric dimming controls are more effective than
* Corresponding author. Tel.: þ852 34427063; fax: þ852 34420427. E-mail address: [email protected] (D.H.W. Li). http://dx.doi.org/10.1016/j.buildenv.2014.11.022 0360-1323/© 2014 Elsevier Ltd. All rights reserved.
photoelectric switching control particularly at high design illuminance levels [9]. In a well daylit room or low setting illuminance levels when daylight intensities are far more than the required values, the energy savings from photoelectric switching controls can be larger than those from the photoelectric dimming controls [10]. However, a problem with the oneoff control type is the frequent switching, annoying occupants and reducing the lamp life [11,12]. It is crucial to visual comfort, building energy and cost aspects. In circulation areas such as corridors, people expect the way ahead to be lit sufficiently. It has been reported that in daylit corridors photoelectric lighting controls can give excellent energy savings [13,14]. Daylighting design techniques are often best illustrated via field measurements to provide reliable operational data and establish design strategies [15e20]. It is essential to indicate the energy savings and the number of switching operations. This paper analyses the electric lighting energy savings and switching frequency for a number of daylight-linked lighting switching controls. Well day-lit corridors located in an atrium building were selected for the study. The findings are reported and design implications discussed. 2. Photoelectric lighting controls The methods of controlling lighting energy consumption can be classified into two basic categories. The first type of control allows the level to be set between maximum and minimum levels by
206
D.H.W. Li et al. / Building and Environment 85 (2015) 205e210
dimming (top-up), and the second type provides for either an on or off state. Automatic oneoff and dimming lighting controls can be employed in daylighting schemes. Dimming controls vary the light output of lamps in accordance with the prevailing daylight level. When daylight is insufficient to achieve the required design illuminance, the indoor lighting level is topped up by artificial lighting. Recently, high frequency dimming controls have been increasingly used and the electronic circuitry employed is more energy efficient than conventional ballast [21]. A high frequency dimming control does not have the ideal characteristic of light output being perfectly proportional to power consumed. The light output can be roughly assumed proportional to power consumed but the lamps cannot be dimmed to total extinction [22]. In normal operation, their residual light output and power consumption occur throughout working hours even if the illuminance level far exceeds the design value [23,24]. However, such operations may be less noticeable and less disturbing to occupants. An oneoff daylight-linked lighting control is designed to switch electric lighting on and off automatically as the daylight level falls and rises through a predetermined level. However, the main drawback with this lighting control type is the rapid and frequent switching of lights on and off, particularly during unstable weather conditions when daylight levels are fluctuating around the switching illuminance. This can disturb occupants and shorten the lamp life. Good switching controls would be the lighting system always off or the number of switching is very low. There are a few variants of the simple oneoff control to lower the number of switching [11,25]. The basic one is the ‘differential switching or dead-band’ photoelectric control that has two switching illuminance levels: one at which the lights are switch on (Eon), and another of higher illuminance, at which the lights are switched off (Eoff). The main strength is that it can reduce rapid switching on and off when the illuminance swings around the desired level. Also, it makes switching off less obtrusive, as it is performed when daylight represents a higher proportion of the illuminance to which the eye is accommodated [11]. It is very important to set appropriately the two switching illuminance values which affect both the number of switching and the electric lighting energy consumption. There is no simple analytical expression for lighting use under a differential switching control. To a good approximation, the fraction of the working year that electric lighting would be off under an oneoff control is simply given by the fraction of the working year that the daylight threshold illuminance level (Emean) which is the average of the two switching illuminances Eon and Eoff [11]. Another technique to decrease the number of switching is to introduce a time delay into the process. Two different types of time delay namely switching-linked and daylight-linked time delays are categorized. The switching-linked time delay only allows switching off until the pre-set delay had to elapse after the last switch on. For daylightlinked time delay, the lighting could only be switched off when the daylight illuminance had exceeded the target value for the preset time period. Delay in switching on is not considered as it could lead to illuminances falling well below desired levels. The third control mode is solar reset. In solar reset switching, the daylight illuminance is sampled at certain set times of day. When the daylight illuminance is more than the switching illuminance at any of the solar reset times, then the lighting is switched off. To ensure that the illuminance levels are not lower than the desired values, the electric light fittings will be automatically switched on whenever the daylight illuminance falls below the design values. There are several advantages for solar reset switching. Firstly, the switching frequency can be limited as the solar reset time can be set at certain periods. Moreover, the occupants can face with switching off at expected times, and the reset times can coincide with their schedule such as during lunchtime and afternoon breaks which
may be appropriate to offices and schools. For a very short reset time interval, the performance would be close to the standard photoelectric control. 3. Background information The analysis was based on the long-term measured daylight illuminance data in an institutional building. It is a 13-storey block located in Hong Kong within the subtropical region at latitude of 22.3 N and longitude of 114.2 E. The building was designed with a skylight and an enclosed stepped atrium to harvest daylight. The stepped atrium increases daylight and improves sky views by splaying the well walls away from the vertical, but the sky component and hence the daylight factor of the lower floors are significantly reduced [26]. The lower part of the atrium between 2/F and 9/F contains typical classrooms, with four open corridors surrounding the atrium. The upper part, from 10/F to 13/F, consists of largely mechanical rooms and offices. The atrium corridors are used for circulation among rooms at different rooms. There are 67 numbers of ceiling-mounted energy-efficient T5 fluorescent tubes, with rated power ranging from 14 W to 35 W for the corridors at every floor. Daylight-linked dimming controls were installed in the corridors at 9th floor. The system detected both the reflected electric light and the daylight values to give a ‘closed loop’ control. The recorded lighting levels were sent to the dimmable electronic ballasts which adjust the light outputs of the lamp fittings accordingly. Four photoelectric sensors were mounted on the ceiling of the four corridors to record the light intensity. The daylight performance and energy use due to the daylight-linked dimming controls in the 9/F corridors were examined and reported [23,24]. There are three additional ceiling mounted photoelectric light level sensors installed respectively along the same plummet line from 6/F to 8/F to record the daylight illuminance. The data were transmitted to a logging system for storage. The recorded illuminance readings only provided the daylight availability for the corridors and the sensor itself did not form any part of the lighting controls. The present study analyzes the various types of control algorithms, namely differential switching, time delay and solar reset based on the daylight illuminance data measured in the atrium corridors. 4. Data analysis and simulation The daylight illuminance data were recorded at 2.5 min intervals from 9:00 am to 6:00 pm. each day between February 2012 and July 2013. Totally, there are 216 daylight illuminance readings collected for each photo-sensor every day without any missing record. The analysis was on daily basis. The whole day illuminance data would be rejected if some data were omitted. It is inevitable that there would be some short periods of missing data due to different reasons including instrumentation malfunction and power failure. Considerable efforts were made to obtain a continuous record of data and in all about 257,000 data were recorded for use. The illuminance levels at a particular time were assumed to have applied for the whole of the time interval of 2.5 min since the previous scan. As pointed out by Littlefair [11] that such an assumption in practice might cause to a slight underestimate of the switching numbers during periods of very substantial daylight illuminance variation (i.e. on a less than 2.5 min timescale). Under a standard switching control, the daylight illuminance recorded by every photocell at each time was checked and if this was less the target illuminance (Et), the control was assumed to have kept the lights on since the last scan. If it was larger than the Et, the lights were assumed to have been off. At the end of the working day, the percentage of time the lights off was computed. The number of
D.H.W. Li et al. / Building and Environment 85 (2015) 205e210
switch offs was counted by keeping a record of when the control changed state and the switch off at the end of the day was excluded in the totals. If at 09:00 the daylight illuminance was above Et then the lighting was assumed to be switched off at first. It should be pointed out that no electric lighting was actually operated under such switching controls. The results were simulated according to the measured daylight illuminance under such a ‘perfect’ control. It means switching exactly when daylight reached Et. The cumulative frequency distribution of daylight availability can show the percentage of the working year in which a given illuminance is exceeded. Fig. 1 presents the cumulative frequency distributions for the daylight illuminance data at the 8th, 7th and 6th Floors. For all setting illuminance values, 8th Floor scored the largest frequency, 7th Floor second and 6th Floor the third. The cumulative frequencies for the 3 floors at individual illuminance levels are valuable for determining the necessary use of artificial lighting, and the probable energy savings from a standard oneoff control. Table 1 summaries the range and the average number of switch offs per day and the percentage of lighting energy saving at these three floors under the standard switching control for the switching illuminances of 100, 125 and 150 lux. The recorded minimum number was zero for all cases but the daily maximum number of switch offs can be up to 29 which imply at least 57 switching operations in a day if switch ons are also included. Such a large number of switching operations could result in remarkable annoyance to occupants. The daily average number of switch offs ranged between 4.7 and 6.8. At the recommended illuminance level for a corridor of 100 lux, the percentage of lighting energy savings were 81.7%, 83.5% and 90.2% at 6th, 7th and 8th Floor, respectively. The lighting energy saving was quite good but the switching number was not acceptable. To reduce the number of switching operating, differential switching controls at target illuminance (Eon ¼ 100 lux) were considered. Five different illuminance combinations with Eoff ¼ 110, 120, 150, 180 and 200 lux were examined and Table 2 gives the results. The number of switch offs was considerably reduced using differential switching controls. The peak switch offs ranged from 24 at 8th Floor when Eoff ¼ 110 to 6 at 6th Floor when Eoff ¼ 200 lux. When the Eoff was 10 lux more than Eon (i.e. Eon ¼ 100 lux and Eoff ¼ 110 lux) the average number of switch offs dropped gently ranging between 4.3 and 4.9. When the Eoff ¼ 200 lux the average
207
Table 1 Number of switch offs per day and percentage of working day (09:00e18:00) that the lights are switched off under standard switching controls. Control switching illuminances (lux)
6/F
7/F
8/F
Range Average number % of working day Range Average number % of working day Range Average number % of working day
100
125
150
0e23 5.4 81.7% 0e27 5.6 83.5% 0e27 4.7 90.2%
0e22 6.2 68.1% 0e28 6.8 76.3% 0e28 5.3 86.4%
0e27 6.4 51.3% 0e25 6.7 68.9% 0e29 5.5 82.5%
numbers of switch offs were 2.7, 2 and 0.9 at 8th, 7th and 6th Floors, respectively. Lower the switch offs means less the lighting energy savings. The lighting energy saving was 54.7% when the average number of switch offs was 0.9. As mentioned, the electric lighting energy under the differential switching control can be considered as the fraction of the working year of the mean values of Eon and Eoff (Emean) [11]. When the Emean ¼ 105 lux (i.e. Eon ¼ 100 lux and Eoff ¼ 110 lux) the simulated lighting energy saving was 2.3% less than when Et ¼ 100 lux under the standard oneoff control. For the other two Emean of 125 lux (i.e. Eon ¼ 100 and Eoff ¼ 150 lux) and 150 lux (i.e. Eon ¼ 100 lux and Eoff ¼ 200 lux) the difference was within 3.5% from the corresponding switching illuminances (i.e. Et ¼ 125 lux and Et ¼ 150 lux) under the standard switching control. The findings were in good agreement with other independent studies [11]. Two different types of time delay viz. switching and daylight linked time delay photoelectric controls were studied. For individual control type, six different time delays from 5 up to 60 min at switching illuminance of 100 lux were considered. Tables 3 and 4 show the switch offs per day and percentage of energy saving for respectively switching and daylight linked time delay photoelectric controls. A short time delay (e.g. 10 and 15 min) gives a substantial reduction in switching particularly for the daylight linked time delay control. For the same average switch offs per day, the maximum switch offs using time delay controls were far less than those using the differential control. It shows that time delay
Fig. 1. Cumulative frequencies of the indoor daylight availability for 6th, 7th and 8th Floors.
208
D.H.W. Li et al. / Building and Environment 85 (2015) 205e210
Table 2 Number of switch offs per day and percentage of working day (09:00e18:00) that the lights are switched off under differential switching controls. Control switching illuminances (lux)
6/F
7/F
8/F
Range Average number % of working day Range Average number % of working day Range Average number % of working day
100/110
100/120
100/150
100/180
100/200
0e19 4.3 79.4% 0e23 4.9 82.2% 0e24 4.3 89.6%
0e15 3.2 76.9% 0e21 4.4 81.3% 0e24 4.2 89.3%
0e11 1.8 68% 0e18 2.9 77.5% 0e19 3.3 87.3%
0e6 1.2 58.8% 0e15 2.4 74.7% 0e19 3 85.5%
0e6 0.9 54.7% 0e13 2 72.4% 0e17 2.7 84.1%
controls give a smaller variation on number of switching. The number of switch offs drops with longer time delay but for very long time delays such as an hour the return diminishes. The lighting energy saving is related to the number of switch offs. More switch operations means less lighting energy consumption. Based on these two control strategies, daylight linked time delay performed better than switching linked time delay in terms of the number of switch offs. For the same number of switch offs, daylight linked time delay always has a shorter time delay. For the time delay of 60 min, the average number of switch offs was around 2 for the switching linked time delay control and 1.3 for the daylight linked time delay control. The lighting energy savings ranged from 57.8% to 88.9% depending on the number of switch offs. The last photoelectric control for analysis is the solar reset switching. Four different control patterns were evaluated: a single reset at 10:30, two resets per day at 10:00 and 12:00; a 2 hourly reset and an hourly reset. Table 5 presents the number of switch offs and the percentage savings for above four solar reset types. As expected, the number of switch offs per day was dropped to very low value. The maximum numbers of switch offs were less than 7 for all the four solar resets. The average number of switch offs for hourly and the single resets were 2 and around 1, respectively. As the number of resets raises so does the number of switch offs. For the hourly reset, the number of switch offs is comparable to those for the time delay and differential switching controls. The percentage savings are affected by the number of switching. The savings ranged from 59% to 79% when the average number of switch offs was between 1 and 2. The electric lighting energy savings under various daylight linked switching controls strongly relate to the corresponding number of switch offs. As observed the data in Tables 1e5, lowered number of switch offs resulting less lighting energy savings and vice versa. Figs. 2e4 depict the correlations between the
Table 3 Number of switch offs per day and percentage of working day (09:00e18:00) that the lights are switched off under switching linked time delay controls at switching illuminance of 100 lux.
Table 4 Number of switch offs per day and percentage of working day (09:00e18:00) that the lights are switched off under daylight linked time delay controls at switching illuminance of 100 lux. Time delay (minutes)
6/F
7/F
8/F
Range Average number % of working day Range Average number % of working day Range Average number % of working day
7/F
8/F
Range Average number % of working day Range Average number % of working day Range Average number % of working day
10
15
30
60
0e10 2.9 74.6% 1e11 2.9 76.2% 1e12 2.6 83.6%
0e8 2.5 72.2% 1e10 2.5 73.7% 1e8 2.3 81.4%
0e6 1.9 66.2% 0e9 1.9 67.7% 0e6 1.8 75.7%
0e4 1.3 57.8% 0e4 1.3 59.2% 0e4 1.3 67.3%
percentage savings and average number of switch offs per day for the 6th, 7th and 8th Floors, respectively. The scatter of data points in the figures can be explained by the variety of energy savings for a particular switching off numbers. Through regression analysis, three simple mathematical equations to relate the average switch offs per day (N) and the lighting energy saving (LES) for the three floors were obtained.
LES ¼ 0:15 Ln N þ 0:576 LES ¼ 0:141 Ln N þ 0:6
R2 ¼ 0:96 R2 ¼ 0:93
LES ¼ 0:156 Ln N þ 0:672
6th Floor
(1)
7th Floor
R2 ¼ 0:93
(2)
8th Floor
(3)
All the regression equations have high coefficients of determination (R2), ranging from 0.93 to 0.96. It indicates that 93e96 % of the variations in percentage of lighting energy savings can be explained by the variations in the average number of switch offs. The strength of correlation is thus considered very good. The mathematical models can provide an alternative to estimate electric lighting energy saving from a given number of switch offs under various daylight linked switching control such that the appropriate type can be selected. 5. Conclusions Photoelectric switching can save energy by shutting down lamp fittings when daylight illuminance is sufficient. However, a problem with such control type is the frequent switching, annoying building users and reducing the lamp life. The best strategy should be a small number of switching off with a high saving. Field
Table 5 Number of switch offs per day and percentage of working day (09:00e18:00) that the lights are switched off under solar reset controls at switching illuminance of 100 lux.
Time delay (minutes)
6/F
5 1e14 3.5 77.5% 1e14 3.5 79.1% 1e16 3.1 86.4%
Reset type
5
10
15
30
60
1e19 4.9 80.6% 1e22 5.1 82.3% 1e25 4.2 88.9%
1e17 4.3 78.7% 1e18 4.4 80.3% 1e19 3.7 86.9%
1e14 3.8 77.1% 1e15 3.9 78.6% 1e16 3.3 85.2%
1e11 3.1 72.8% 1e11 3.1 74% 1e11 2.7 80.5%
1e6 2.2 67.2% 1e7 2.3 68.1% 1e7 2 73.2%
6/F
7/F
8/F
Range Average number % of working day Range Average number % of working day Range Average number % of working day
Single 10:30
10:00 and 12:00
2 hourly
Hourly
0e2 1.1 59% 0e2 1.1 60% 0e2 1.2 69%
0e3 1.3 64% 0e3 1.3 65% 0e3 1.3 73%
0e5 1.6 66% 0e5 1.6 68% 0e5 1.6 76%
0e7 2 70% 0e7 2 72% 0e7 1.9 79%
D.H.W. Li et al. / Building and Environment 85 (2015) 205e210
209
Fig. 2. Percentage of energy savings under various average switch offs per day using different forms of switching controls at 6th Floor.
measurements of the daylight illuminance in atrium corridors were undertaken. Based on the measured daylight availability, a number of daylight linked switching controls namely the standard, differential, switching and daylight time delays and solar reset oneoff controls were analysed. Their performances in terms of lighting energy reduction and number of switching were simulated and evaluated. Under the standard oneoff control, the energy saving was very good but the frequent switching was unaccepted. At the switching illuminance of 100 lux, the lighting energy savings were found ranging from 81.7% to 90.2% and the average number of switch offs between 4.7 and 5.6. Based on the simulated results, the variants of the switching control can reduce the number of switching but increase the energy consumption. Using the differential switching control, the number of switch offs can be dropped
to 0.9 with an energy saving of 54.7% at 6th Floor when Eon ¼ 100 lux and Eoff ¼ 200 lux. The switching and daylight linked time delay controls were examined. Provided an appropriate delay time is chosen, a daylight linked time delay can result less number of switch offs with lower energy savings. For solar reset switching, four different control patterns were studied. The average number of switch offs under the single reset was just about 1.1 for the three floors. The correlation between the percentage of lighting energy savings and the number of switch offs was evaluated. Three simple mathematical models for the three floors were developed. The findings can give an alternative to estimate the electric energy savings from the selected number of switch offs based on the various photoelectric switching controls. The important issue would be whether the number of switch offs can be achieved by a
Fig. 3. Percentage of energy savings under various average switch offs per day using different forms of switching controls at 7th Floor.
210
D.H.W. Li et al. / Building and Environment 85 (2015) 205e210
Fig. 4. Percentage of energy savings under various average switch offs per day using different forms of switching controls at 8th Floor.
particular switching control. Photoelectric electric lighting controls are appropriate building energy conservation schemes and should therefore be widely applied in day-lit areas. Further research work is required to fit the approach for various daylighting design criteria such as daylight factor so that the analysis would be more generalized. The theory proposed has been supported by on-site measurements in a building which is fully operational. This has not been done before. Given the growing concern about energy use in buildings and its likely impact on the environment, this study is timely and worthwhile. It is believed the work and its findings presented in this paper would be of use to the researchers and building professionals, particularly those who are concerned about energy use in buildings and its impact on the environment. Although the work was conducted in subtropical Hong Kong, the procedures and techniques developed could be applied in other locations. Acknowledgements The work described in this paper was fully supported by a General Research Fund from the Research Grant Council of the Hong Kong Special Administrative Region, China [Project no. 9041896 (CityU 117713)]. References [1] Chiogna M, Albatici R, Frattari A. Electric lighting at the workplace in offices: efficiency improvement margins of automation systems. Light Res Technol 2013;45(5):550e67. [2] Hwang T, Jeong TK. Effects of indoor lighting on occupants' visual comfort and eye health in a green building. Indoor Built Environ 2011;20(1):75e90. [3] Iversen A, Nielsen TR, Svendsen SH. Illuminance level in the urban fabric and in the room. Indoor Built Environ 2011;20(4):456e63. [4] Roetzel A, Tsangrassoulis A, Dietrich U. Impact of building design and occupancy on office comfort and energy performance in different climates. Build Environ 2014;71:165e75. [5] Li DHW, Tsang EKW. An analysis of daylighting performance for office buildings in Hong Kong. Build Environ 2008;43(9):1446e58. [6] Li DHW, Lam JC, Wong SL. Daylighting and its implications to overall thermal transfer value (OTTV) determinations. Energy 2002;27(11):991e1008. [7] Li DHW, Lam JC, Wong SL. Daylighting and its effects on peak load determination. Energy 2005;30(10):1817e31.
[8] Li DHW, Mak AHL, Chan WWH, Cheng CCK. Predicting energy saving and life cycle cost analysis for lighting and daylighting schemes. Int J Green Energy 2009;6(4):359e70. [9] Lynes JA, Littlefair PJ. Lighting energy savings from daylight: estimation at the sketch design stage. Light Res Technol 1990;22(3):129e37. [10] Li DHW, Cheung KL, Wong SL, Lam TNT. An analysis of energy-efficient light fittings and lighting controls. Appl Energy 2010;87(2):558e67. [11] Littlefair PJ. Photoelectric control: the effectiveness of techniques to reduce switching frequency. Light Res Technol 2001;33(1):43e58. [12] Li DHW, Lam JC. An analysis of lighting energy savings and switching frequency for a daylit corridor under various indoor design illuminance levels. Appl Energy 2003;76(4):363e78. [13] Li DHW, Lam JC. An investigation of daylighting performance and energy saving in a daylit corridor. Energy Build 2003;35(4):365e73. [14] Li DHW, Tsang EKW. An analysis of measured and simulated daylight illuminance and lighting savings in a daylit corridor. Build Environ 2005;40(7): 973e82. [15] Galasiu A, Atif MR. Applicability of daylighting computer modeling in real case studies: comparison between measured and simulated daylight availability and lighting consumption. Build Environ 2002;37(4):363e77. [16] Littlefair PJ, Aizlewood ME. Measuring daylight in real buildings. In: Proceedings of the CIBSE national lighting conference, Bath, UK; 1996. [17] Li DHW, Lam JC. Evaluation of lighting performance in office buildings with daylighting controls. Energy Build 2001;33(8):793e803. [18] Li DHW, Lam JC, Lau CCS, Huan TW. Lighting and energy performance of solar film coating in air-conditioned cellular offices. Renew Energy 2004;29(6): 921e37. [19] Li DHW, Lam TNT, Wong SL. Lighting and energy performance for an office using high frequency dimming controls. Energy Convers Manag 2006;47(9e10):1133e45. [20] Li DHW, Lam TNT, Wong SL, Tsang EKW. Lighting and cooling energy consumption in an open plan office using solar film coating. Energy 2008;33(8): 1288e97. [21] Slater AI. Energy efficiency in lighting. Energy Manag 1991;12(2):24e6. March/April. [22] Ihm P, Nemri A, Krarti M. Estimation of lighting energy savings from daylighting. Build Environ 2009;44(3):509e14. [23] Chow SKH, Li DHW, Lee EWM, Lam JC. Analysis and prediction of daylighting and energy performance in atrium spaces using daylight-linked lighting controls. Appl Energy 2013;112:1016e24. [24] Li DHW, Cheung ACK, Chow SKH, Lee EWM. Study of daylight data and lighting energy savings for atrium corridors with lighting dimming controls. Energy Build 2014;72:457e64. [25] Littlefair PJ. Predicting lighting energy use under daylight linked lighting controls. Build Res Inf 1998;26(4):208e22. [26] Samant S. A critical review of article published on atrium geometry and surface reflectances on daylighting in an atrium and its adjoining spaces. Archit Sci Rev 2010;53(2):145e56.