University of Southern California Dissertations and Theses

Evaluation of daylighting circadian effects: Integrating non-visual effects of lighting in the evaluation of daylighting designs

(USC Thesis Other)

Evaluation of daylighting circadian effects: Integrating non-visual effects of lighting in the evaluation of daylighting designs

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content 1 Evaluation of daylighting circadian effects Integrating non-visual effects of lighting in the evaluation of daylighting designs by Geli Qiu A Thesis Presented to the SCHOOL OF ARCHITECTURE UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree MASTER OF BUILDING SCIENCE AUGUST 2018 2 COMMITTEE CHAIR: Kyle Konis, Ph.D., AIA Assistant Professor USC School of Architecture kkonis@usc.edu COMMITTEE MEMBER #2: Marc E. Schiler Professor USC School of Architecture marcs@usc.edu (213)740-4591 COMMITTEE MEMBER #3: Victor Albert Regnier FAIA Associate Dean of Research USC School of Architecture regnier@usc.edu (310)773-1260 3 ACKNOWLEDGEMENTS I would like to express my deepest appreciation to my thesis chair, Professor Kyle Konis, and my committee members, Professor Marc E. Schiler and Professor Victor Albert Regnier. This project would not have been possible without your great ideas, creative inputs, detailed instructions and valuable comments. You have been the excellent committee members throughout not only my thesis process, but also my entire graduate student life. Furthermore, I would like to express my deepest gratitude to my family. I would not be able to make it without your help along the way. You are always there, being with me through countless ups and downs. You pave the way for me and give me the chance to do what I love to do and never ask anything for return. Last but not the least, I would like to thank the faculty and friends in Master of Building Science program, as well as all the friends in School of Architecture, University of Southern California for your supports through my academic career. You make my dreams come true. 4 CONTENTS 1. CHAPTER 1 INTRODUCTION ....................................................................................................................12 1.1. Circadian rhythm..................................................................................................................................12 1.2. Daylight ...............................................................................................................................................13 1.3. Daylight and the circadian system .........................................................................................................15 1.4. Causes of circadian disruption ..............................................................................................................15 1.4.1. Delayed Sleep Phase Syndrome (DSPS) and Advanced Sleep Phase Syndrome (ASPS) ....................15 1.4.2. Free-Running Disorder (FRD) ..........................................................................................................16 1.4.3. Irregular Sleep-Wake Rhythm (ISWR) .............................................................................................16 1.4.4. Shift-work sleep disorder (SWSD)....................................................................................................16 1.4.5. Jet Lag disorder (JLD) ......................................................................................................................16 1.5. The effects of circadian disruption ........................................................................................................17 1.6. Weather data file for daylight simulation ..............................................................................................17 1.6.1. International Weather for Energy Calculations 2.0 (IWEC2) .............................................................17 1.6.2. Test Reference Year (TRY) ..............................................................................................................17 1.6.3. Typical Meteorological Year (TMY) ................................................................................................18 1.6.4. Comparison of three weather files ....................................................................................................18 1.7. Terminology ........................................................................................................................................18 1.7.1. Spectral Power Distribution ..............................................................................................................18 1.7.2. Correlated Color Temperature (CCT) ...............................................................................................19 1.7.3. Circadian Stimulus (CS) ...................................................................................................................19 1.7.4. Equivalent Melanopic Lux (EML) ....................................................................................................20 1.8. Scope of thesis .....................................................................................................................................20 1.9. Objectives of thesis ..............................................................................................................................20 2. CHAPTER 2 BACKGROUND AND LITERATURE REVIEW ....................................................................21 2.1. Introduction .........................................................................................................................................21 2.2. Daylight design objectives ....................................................................................................................21 2.2.1. Visual light quality ...........................................................................................................................21 2.2.2. Solar gain.........................................................................................................................................22 2.2.3. Energy consumption.........................................................................................................................22 2.2.4. Glare................................................................................................................................................23 2.3. Circadian lighting .................................................................................................................................24 2.3.1. Ganglion cells ..................................................................................................................................24 2.3.2. Calculation method ..........................................................................................................................25 2.3.3. Circadian lighting evaluation metric .................................................................................................25 5 2.4. Standard ...............................................................................................................................................26 2.4.1. LEED ..............................................................................................................................................27 2.4.2. WELL building Standard..................................................................................................................27 2.5. Evaluation Tool ....................................................................................................................................27 2.5.1. Rhinoceros & Grasshopper ...............................................................................................................27 2.5.2. Ladybug & Honeybee ......................................................................................................................28 2.5.3. Circadian stimulus calculator ............................................................................................................28 2.5.4. Lark Spectral Lighting......................................................................................................................28 2.6. Summary .............................................................................................................................................29 3. CHAPTER 3 METHODOLOGY ...................................................................................................................30 3.1. Geometry .............................................................................................................................................30 3.1.1. Building orientation .........................................................................................................................32 3.1.2. Window to wall ratio (WWR)...........................................................................................................32 3.1.3. Ceiling height ..................................................................................................................................33 3.1.4. Glazing material ...............................................................................................................................33 3.1.5. Exterior shading ...............................................................................................................................33 3.2. Energy performance .............................................................................................................................34 3.3. Daylighting quality ...............................................................................................................................34 3.3.1. Visual lighting quality ......................................................................................................................35 3.3.2. Circadian lighting quality .................................................................................................................36 3.4. Iterative analysis .................................................................................................................................37 3.5. Interior finish selection .........................................................................................................................39 3.6. Detailed circadian lighting evaluation ...................................................................................................39 3.6.1. Convert SPD component ..................................................................................................................40 3.6.2. Spectral Sky .....................................................................................................................................41 3.6.3. Multi-channel simulation ..................................................................................................................42 3.6.4. Results .............................................................................................................................................43 3.6.5. Circadian lighting changes ...............................................................................................................43 3.7. Summary .............................................................................................................................................43 4. CHAPTER 4 RESULTS ................................................................................................................................44 4.1. Iterative analysis results........................................................................................................................44 4.2. Interior finishes circadian lighting quality results ..................................................................................47 4.3. Detailed circadian lighting evaluation results ........................................................................................49 4.3.1. View direction selection result ..........................................................................................................49 4.3.2. Results of circadian lighting effects shift in both space and time .......................................................51 6 4.4. Summary .............................................................................................................................................52 5. CHAPTER 5 DISCUSSION ..........................................................................................................................53 5.1. The significance of geometry design and GA based iterative analysis ....................................................53 5.2. The significance of interior finish color.................................................................................................54 5.3. The significance of view direction selection ..........................................................................................54 5.4. An approach to artificial lighting design for circadian effects ................................................................54 5.5. Summary .............................................................................................................................................57 6. CHAPTER 6 CONCLUSION ........................................................................................................................58 7. CHAPTER 7 LIMITATIONS AND FUTURE WORK ..................................................................................60 7.1. Future work for this research ................................................................................................................60 7.2. Future work in circadian lighting design ...............................................................................................60 7.3. Summary .............................................................................................................................................61 8. Reference ......................................................................................................................................................62 7 LIST OF FIGURES Figure 1-1 Circadian Rhythm Phase Shifts (Drake et al. 2004) ...........................................................................13 Figure 1-2 Biological Clock (“Biological_clock.png (700×527)” 2018)..............................................................13 Figure 1-3 Natural light category .......................................................................................................................14 Figure 1-4 Shortest daytime in Helsinki (“Sunrise and Sunset Times in Helsinki” 2017) .....................................14 Figure 1-5 Longest daytime in Helsinki (“Sunrise and Sunset Times in Helsinki” 2017) ....................................15 Figure 1-6 Relative spectral power distribution for CIE D65 (Kevin Houser 2006) .............................................18 Figure 1-7 Comparison of spectral power distribution curves from three daylight conditions ..............................19 Figure 2-1 Daylight design based on solar analysis (Curtis and Mistrick, n.d.) ....................................................22 Figure 2-2 Commercial electricity use per square foot of commercial floorspace falls (“EIA - Annual Energy Outlook 2018” n.d.) ..........................................................................................................................................23 Figure 2-3 Non-visual light response generation ((Lucas et al. 2014)) ................................................................25 Figure 2-4 Melanopic and Visual Curves (“Appendix C: Tables | WELL Feature Library” 2017) .......................26 Figure 2-5 Circadian Stimulus Calculator main display (“Light and Health | Research Programs | LRC” 2017) ...28 Figure 3-1 New Daylight Design Workflow .......................................................................................................30 Figure 3-2 FlexLab (“FLEXLAB® | FLEXLAB” 2017).....................................................................................31 Figure 3-3 FlexLab inside view .........................................................................................................................31 Figure 3-4 Digital Test Space ............................................................................................................................32 Figure 3-5 Shading Creation ..............................................................................................................................34 Figure 3-6 Local Weather Analysis....................................................................................................................35 Figure 3-7 Annual daylighting simulation illuminance results ............................................................................36 Figure 3-8 Optimization tool interface ...............................................................................................................38 Figure 3-9 Detailed Circadian Lighting Effects Evaluation Workflow ................................................................40 Figure 3-10 Fourth_Madision Glazing SPD data structure ..................................................................................41 Figure 3-11 Fourth_Madision output data structure ............................................................................................41 Figure 3-12 Irradiance data structure .................................................................................................................42 Figure 3-13 Relative sun position changes .........................................................................................................43 Figure 4-1 Iterative analysis results from "Octopus" ...........................................................................................45 Figure 4-2 Comparison of configurations of reference and characteristic solutions .............................................46 Figure 4-3 Comparison of overall better solution and worst energy consumption solution ...................................47 Figure 4-4 Comparison of overall better solution and worst visual lighting quality solution ................................47 Figure 4-5 Comparison of overall better solution and worst circadian lighting quality solution ...........................47 Figure 4-6 Arrangement of view directions ........................................................................................................48 Figure 4-7 views from 4 directions ....................................................................................................................48 Figure 4-8 Comparison of EML fluctuation based on three interior finishes, Mar 21st, 9:00................................49 Figure 4-9 Comparison of all sun positions and typical sun positions .................................................................50 8 Figure 4-10 Comparison of circadian effective percentage based on view directions ...........................................51 Figure 4-11 Summary of EML results of typical days ........................................................................................52 Figure 5-1 Summary of circadian lighting stimulus level of each portion, March 21st, 9:00 ................................55 Figure 5-2 Lighting fixture layout and operation template for March and September at 9:00 ...............................55 Figure 5-3 Template of artificial lighting for circadian lighting design................................................................56 Figure 5-4 Comparison of lighting fixture output in the advanced circadian lighting design ................................57 9 LIST OF TABLES Table 2-1 Tailored Methods Lighting Power Allowances ...................................................................................22 Table 2-2 Melanopic ratio from WELL Building Standard .................................................................................26 Table 2-3 Summary of WELL building standard, section 54 ..............................................................................27 Table 3-1 Summary of glazing materials ............................................................................................................33 Table 3-2 Thermal setting for energy simulation ................................................................................................34 Table 3-3 Summary of parameters and ranges ....................................................................................................37 Table 3-4 Summary of Objective Function.........................................................................................................37 Table 3-5 Summary of optimization setting .......................................................................................................38 Table 3-6 Geometry setting for interior finish selection ......................................................................................39 Table 3-7 Geometry setting for detailed circadian lighting evaluation .................................................................40 Table 4-1 Comparison of 8 case solutions from iterative analysis .......................................................................47 Table 4-2 Summary of typical days and the days that they represent ...................................................................50 Table 5-1 Comparison of poor circadian lighting design and advanced circadian lighting design ........................56 Table 6-1 Category of portions based on circadian lighting stimulus ..................................................................59 10 ABSTRACT This research proposed a new daylighting design workflow for architects to develop parametric designs for commercial buildings based on multiple building performance indicators. Conventional daylighting design mainly focuses on building performance like visual lighting quality and energy consumption. However, with a deeper understanding of daylight and its effect on human beings’ wellness, more factors need to be considered, such as its effect on people’s “circadian rhythms”. Several circadian lighting evaluation methods have been introduced but not integrated into a daylighting design workflow. The proposed workflow adapted parametric design including window wall ratio, ceiling height, building orientation and glazing type. The parametric design aimed at seeking for a better design solution based on three building performance indicators which are visual lighting quality, circadian lighting quality, and energy consumption. In this research, visual lighting quality took spatial Daylight Autonomy (sDA) as the metric while energy consumption took Energy Use Intensity (EUI) as the metric. Similar to sDA, this research proposed a new metric named Circadian Effective Percentage (CEP) and adapted it as the metric for circadian lighting quality. Also, the parametric design has another parameter which is interior finish. Different than other parameters, interior finish color only affects circadian lighting quality. Besides, this workflow ended with a detailed circadian lighting evaluation of daylighting. The evaluation is aimed at presenting circadian lighting effect shifts in both time and space so that lighting designers can layout lighting fixtures and use pre-programed lighting control systems as supplements. The application of this workflow is qualified for Title 24 (Section 140.6: General Illumination Level) and WELL building standard (Section 54: Circadian Lighting Design). Architects can adapt the proposed workflow for a design solution that maximizes daylight as a resource for circadian stimulus and entrainment. A case study took FLEX LAB in Lawrence Berkeley National Laboratory as a prototype to demonstrate the viability of the proposed workflow. 11 HYPOTHESIS In the early-stage of daylighting design, it is possible for architects to seek for a balanced building geometries design based on multiple building performance indicators including visual lighting quality, circadian lighting quality, and energy consumption. Also, with studies on circadian lighting effects that shift in both time and space, the lighting designer can take the layout of lighting fixtures and a pre-programmed lighting control system to supply “biological darkness” space with circadian lighting effects from daylighting. 12 1. CHAPTER 1 INTRODUCTION Conventional daylight design workflow focused on solar radiation, visual light quality, energy consumption and so on, but not include circadian lighting quality. This research proposed a new simulation-based workflow for daylight design in the early design stage which brought circadian lighting effects into consideration. Apart from optimization in an early stage of design, a detailed evaluation of circadian effects of daylighting provided interior lighting designers a guide for layout and control of artificial lighting. Daylight design involves multidisciplinary teams including architects, owners, mechanical engineers, structual engineers, curtain wall consultants, lighting designers and so on. As architects, several strategies could be applied in the early design stage such as building form, building orientation, fenestration and shading devices. All teams worked hard to optimize daylight values based on human beings’ vision, energy performance for a long time while a new objective was brought into designers’ scope named “circadian effect”. Circadian photobiology was proposed as another indicator apart from vision (M.G. Figueiro et al. 2007). With more research done by different teams such as the Lighting Research Center (LRC), circadian effects drew public attention and were addressed in the WELL Building Standard. To start with, this chapter introduces circadian-related terminology, the importance of the circadian system, daylight features and the way to describe daylight. 1.1. Circadian rhythm Circadian rhythm is a biological process which has a period of around 24 hours. It occurs in plants, animals, fungi, and cyanobacteria. Circadian rhythm is dominated by a biological clock, which is controlled by nerve cells named suprachiasmatic nuclei (SCN). SCN are located at the anterior hypothalamus, a part at the base of the brain (Vitaterna, Takahashi, and Turek 2001). There is another clock known as an astronomical clock, which refers to a special mechanism indicating the position of sun. Another way to express the astronomical clock is “day and night cycle which occurs every 24 hours. If a human circadian system is in a healthy state of entrainment, the biological clock keeps resetting itself to follow the pattern with the astronomical clock every day. As intermediaries, zeitgebers (external cues) like light and timekeeping can either extend or shorten the approximately 24-hour circadian rhythm. Circadian rhythm is critical for human beings’ health especially when people are trying to challenge the biological clock. For example, people who suffer from a rotating work schedule have to work during the nighttime and sleep during the daytime. When they are working, it is nighttime and the biological clock suggests to them to go to sleep. Then the conflict occurs. Normally, this situation causes excessive sleepiness when they are working (nighttime) and causes insomnia when they should bee sleeping (daytime). If people have access to a certain amount of light for a long time at night, then their circadian system gets constant external stimuli which obey the instruction from the biological clock. In that case, disruption of circadian rhythm occurs, and they will suffer from that. For example, some people nowadays work through the night, so they have a longer biological process than normal people. That will disrupt their next period of the biological process. Figure 1-1 compares a normal circadian rhythm phase and other problematic circadian rhythm phases. Figure 1-2 explains typical time spots in a normal biological clock. For example, at 7:30, the stop of melatonin secretion is a biological signal that this is the time to wake up. Then at 21:00, melatonin secretion starts indicating that it is time to sleep. 13 Figure 1-1 Circadian Rhythm Phase Shifts (Drake et al. 2004) Figure 1-2 Biological Clock (“Biological_clock.png (700×527)” 2018) 1.2. Daylight As a part of natural light, daylight refers to the light that received during the daytime. Daylight varies from several factors including local time, latitude, longitude, and weather conditions. Figure 1-3 explains how natural light is categorized. Generally, daylight is passed by either a direct way which refers to sunlight or an indirect way which refers to sky light. Sky light usually gets diffused by the overcast sky or gets reflected by ground or walls. Ambient daylight refers to the volume of natural light that enters a building. So, when people talk about interior daylighting conditions, technically they are talking about ambient daylighting, which is an architectural designer’s topic as well. For daylight, extreme conditions happen in the northernmost and southernmost regions named polar night which refers to the nighttime that lasts for 24 hours and polar day which refers to the daytime for 24 hours. Taking Helsinki 14 as an example, Figure 1-4 shows that it has the shortest daytime which is 5h 49min and Figure 1-5 shows that it has the longest daytime which is 18h 56min in 2017. Figure 1-3 Natural light category Figure 1-4 Shortest daytime in Helsinki (“Sunrise and Sunset Times in Helsinki” 2017) 15 Figure 1-5 Longest daytime in Helsinki (“Sunrise and Sunset Times in Helsinki” 2017) 1.3. Daylight and the circadian system As an external cue, daylight is critical to maintaining circadian rhythm in a healthy state of entrainment. When daylight travels to the retina through the eye, it is converted into signal which is carried by neurons and then goes all the way down to the suprachiasmatic nucleus (SCN). Located in hypothalamus, the SCN serves as a functional clock which bridges the gap between external cue and circadian rhythm (“The Human Suprachiasmatic Nucleus | HHMI BioInteractive” n.d.). What happens with daylight savings time(DST) is an example for proving the function of SCN. When DST comes, it affects people’s health by causing disorder between the biological clock and the astronomical clock and that is how circadian disruption occurs. To solve this disruption, SCN receives a signal converted by daylight and uses it as a cue to reset the biological clock until it matches the astronomical clock. Previous research showed that the whole circadian rhythm transfer process takes from one day up to two weeks when daylight serves as external cues (Kantermann et al. 2007). 1.4. Causes of circadian disruption For the people who have their circadian rhythm in a good state, their biological clock is corresponding to the astronomical clock. However, there are certain groups of people who suffer from the disorder between the biological clock and the astronomical clock. Consequently, circadian disruption occurs which harms their wellness. There are 5- key factors causing circadian disruption and they have different effects based on age, gender and other individual differences. The following sections describe the 5 common causes and which typical group of people suffer from them. 1.4.1. Delayed Sleep Phase Syndrome (DSPS) and Advanced Sleep Phase Syndrome (ASPS) Both delayed sleep phase syndrome and advanced sleep phase syndrome are circadian rhythm disorders. Normal circadian rhythm has a range of time for people to wake up. However, for those people who wake up either earlier or later than the range will have circadian disruption. To maintain a consistent pattern, DSPS and ASPS patients have the same sleep and wake time length. If people wake up earlier than normal, consequently they go to sleep earlier than normal and this is advanced sleep phase syndrome. Otherwise, when the wake up and go to sleep cycle is behind schedule, it is a delayed sleep phase syndrome. Secretion of melatonin impacts the productivity of people because higher melatonin secretion makes people fall asleep. Once people stop secreting melatonin, they are unsleeping. DSPS is a delay of melatonin secretion. The symptom of that is a delayed sleep. By shifting the 16 biological clock, patients’ energetic peak time is also shifted which brings problems. For example, a normal working schedule starts at 8 a.m. and ends at 5 p.m. People with DSPS would fall asleep from 8 a.m. to 10 a.m. or even longer. So, DSPS patients’ effective working hours decrease by 25% for every day. In other cases, even though people with ASPS would be effective from 8a.m., they would fall asleep around 3p.m. or even earlier which is at least 25% less effective hours as well. DSPS is more common among young people because of late night work or activities. More specifically, about 7% of teenagers aged from 7 to 20 have trouble with DSPS. Waking up in the morning is the challenge for this group of people (“6 Circadian Rhythm Sleep Disorders That May Be Disrupting Your Sleep” n.d.). They suffered from late night insomnia, excessive daytime sleepiness and higher risk of depression (“6 Circadian Rhythm Sleep Disorders That May Be Disrupting Your Sleep” n.d.). Other than DSPS, aging people represent a large part of the patients who are suffering from ASPS. They undergo a higher risk of depression and excessive sleepiness but early morning insomnia. 1.4.2. Free-Running Disorder (FRD) Non-24-hour sleep-wake syndrome, also called Free-Running Disorder (FRD), refers to those patients who have their sleep-wake cycle longer than 24 hours. Different than DSPS and ASPS, a non-24-hour sleep-wake cycle suggests that it is not only mismatched with a normal schedule but also varies from day to day. Patients cannot fall asleep even though they are in bed. Consequently, their wake-up time does not correspond to a normal schedule. The non-24-hour sleep-wake syndrome happens among most blind people. This suggests that daylight is a critical factor for FRD. When it comes to why most blind people undergo FRD, one assumption would be that the signal converted by daylight failed to get to SCN so SCN can’t reset the biological clock (“6 Circadian Rhythm Sleep Disorders That May Be Disrupting Your Sleep” n.d.). One study showed that most blind people can entrain their circadian rhythm by administrating melatonin (Lewy; 2000). 1.4.3. Irregular Sleep-Wake Rhythm (ISWR) Irregular sleep-wake rhythm means the sleep period and wake-up period are not consistent with an approximately 24- hour cycle. Patients with ISWR divided their sleep time into several short parts and usually poor sleep quality comes along with that. An earlier experiment showed that animals without SCN have a similar sleep-wake pattern to aging people, especially those patients in dementia facilities (Mistlberger 2005). Both age and light exposure are risk factors. One group claimed that napping problem increases with age(Foley et al. 1995) while another group hold the opinion that it increases with medical syndrome (Maurice M. Ohayon et al. 2004). Also, lower lighting exposure resulting in higher secretion of melatonin is a risk factor of ISWR. There is a higher chance that aging people especially the ones with mental problems suffer from ISWR. 1.4.4. Shift-work sleep disorder (SWSD) Shift-work sleep disorder occurs when the work schedule is not aligned with the biological clock. Generally, shift work is divided into four categories which are early morning shift, late evening shift, the night shift and rotating work schedules (“6 Circadian Rhythm Sleep Disorders That May Be Disrupting Your Sleep” n.d.). Patients with SWSD suffered from excessive sleepiness during the circadian rhythm wake-up period and insomnia during the circadian rhythm sleep period. An experiment with 2570 samples in ages from 18 to 65 years old showed that approximately 10% of rotating schedule workers have SWSD (Drake et al. 2004). Lack of External stimulus like daylight stimuli is another reason for this disorder. Patients did not get enough lighting exposure in their “wake-up” period (nighttime) so their circadian rhythm has no access to external stimulus to reset their biological clock. 1.4.5. Jet Lag disorder (JLD) Jet lag disorder occurs when travelers fly across several time zones. During such a long-distance flight, light conditions in the cabin remain dark for most of the time. In that case, both light schedule and light level are not 17 enough to provide travelers effective circadian lighting stimulus. On the plane, circadian disruption already occurred. Even worse, when aircraft landed on another time zone, the light condition at destination has a different pattern than the place of departure. The new astronomical clock (at the destination) had a bigger gap comparing to travelers’ biological clock. For example, it took travelers 11.75 hours to fly from Shanghai (PVG) to Los Angeles (LAX). The flight departed at 23:20 (GMT+8) which is nighttime in Shanghai and arrived at 19:05 (GMT-8) which is early evening in Los Angeles. Based on GMT+8-time zone, it is about time for sleeping when travelers departed and time to wake up when they arrived Los Angeles. This sleep-wake cycle generally follows an internal clock. However, when they arrived Los Angeles, local time was 19:05 which means the astronomical clock suggests travelers to go to sleep in the next 2 hours. The result is that when travelers arrive Los Angeles, the biological clock indicates that they should wake up while the astronomical clock suggests that they should go to sleep. This conflict causes JLD. Aside from light exposure, body temperature is another factor which causes JLD. The time duration of resetting the biological clock depends on the direction of the flight, quantity of time zones that they crossed, travelers’ sleep quality in the cabin, external stimulus at the destination and individual differences (Sack et al. 2007). Based on job type, cabin crews have a higher risk to have JLD. A study with 33 samples showed that pilots over 50 years old have a lower influence with JLD than younger pilots (Tresguerres et al. 2001). 1.5. The effects of circadian disruption Circadian disruption may cause a lot of diseases including but not limited to cancer, low defense of the immune system, and sundowning syndrome. Protein fluctuates day by day when circadian disruption occurs because circadian genes control clock function. The consistent fluctuation results in regulation of cell cycles which eventually induces cancer (Savvidis and Koutsilieris 2012). An early study by analyzing the clock gene expression provided the evidence that there is a complex connection between the immune system and circadian rhythm (Lange, Dimitrov, and Born 2010). More commonly in aging people, a phenomenon called late-day confusion, also referred as sundowning, changes patients’ action dramatically in the late afternoon or early evening. Sundown syndrome causes confusion, disorientation, anxiety, agitation, pacing, wandering, resistance to redirection, screaming, yelling and so forth (Khachiyants et al. 2011). An early study with 25 Alzheimer’s disease samples demonstrated that sundowning is related to circadian rhythm disruption to some extent (Volicer et al. 2001). 1.6. Weather data file for daylight simulation Weather files vary in format, data source and data structure. This section analyzes weather data from those 3 aspects to apply one for this simulation-based research. Three commonly used weather data files are listed below. They are widely applied in building energy simulation and daylighting simulation. 1.6.1. International Weather for Energy Calculations 2.0 (IWEC2) Provided by the American Society of Heating, Refrigerating and Air-Conditioning Engineers (ASHRAE) and White Box Technologies (WBT), IWEC2 weather file contains 3012 Typical Year Weather (TYW) Files all over the world (Parakou et al., n.d.). Maintained by National Climatic Data Center (NCDC), Integrated Surface Hourly (ISH) database contains sky cover, wind speed and direction, dry-bulb temperature, atmospheric pressure and liquid precipitation. The weather data enables researchers and designers to run both energy simulation and daylight simulation. Instead of using measurement data, the ISH database calculated hourly total solar radiation, the direct normal solar radiation and global horizontal illuminance, direct normal illuminance, diffuse horizontal illuminance and zenith luminance as input for daylight simulation. The IWEC2 files are adaptable for different format with records of 4 times per day on average for at least 12-years period up to 25-years period. DVD version comes in CSV format while individual files comes with EPW, BINM and CSV formats(“Frequently Asked Questions about ASHRAE IWEC2 Weather Data and White Box Technologies” n.d.). 1.6.2. Test Reference Year (TRY) As one of the earliest set of weather data, Test Reference Year does not have original radiation data for software to run daylighting simulation. Instead of read data from TRY weather file, the software needs to calculate radiation data for energy simulation. Weather data for TRY was picked from 1948-1975. Based on its picking principle, the final data for reference year is mild, without any extreme high or low but true data(Crawley 1998). In some cases, TRY data that researcher used for their simulation were generated by researcher’s own principal (De Miguel and Bilbao 2005). 18 1.6.3. Typical Meteorological Year (TMY) Typical Meteorological Year was set by cooperation of the National Climate Data Center (NCDC) and Sandia National Laboratory (SLN). TMY broke the limitation of TRY by recording total horizontal and direct normal solar radiation data for 234 locations in the U.S. Among these 234 locations, 26 of their solar data came from measurement. Solar data for the rest locations were calculated based on cloud cover and type. TMY is a set of data selected from 1952- 1975. Different than TRY, TMY data for each month is not from the same year. The selection principle is based on a balance between solar radiation, dry-bulb temperature, dew point and wind velocity (Crawley 1998). For example, the TMY weather data (epw. format) for San Francisco takes January data from 1974, February data from 1960, March data from 1963, April data from 1974, May data from 1972, June data from 1974, July data from 1960, August data from 1963, September data from 1961, October data from 1966, November data from 1974 and December data from 1966. These monthly data from different year made up the TMY data. 1.6.4. Comparison of three weather files One lamination of IWECS2 is that there are only 4 times per day data representing the whole day which failed to address the hourly changes. Also, both IWECS 2 and TRY calculate solar radiation for daylight simulation. This approach fails to count cloud cover and light density which have drastically impacts on daylight simulation results. This research takes TMY weather data for the daylight simulation because the radiation data were directly measured. the simulation took cloud cover and light density into consideration for a more accurate result. However, TMY weather data has its own limitation which is its logic for monthly data selection. Based on TMY monthly data selection principle, no extreme weather condition was included but it happens in a real situation. 1.7. Terminology 1.7.1. Spectral Power Distribution A light source is a combination of different wavelengths. Each wavelength has its own radiant power. For example, a visible region of light consists of wavelengths from 360 nm to 770 nm. Spectral power distribution is the statistical data represented by radiant and wavelength (“Lighting Research Center | Education | Learning | Terminology | Spectral Power Distribution” n.d.). Figure 1-6 shows the spectral power distribution curves of standard CIE D65 (CCT=6504K) in visible wavelength. Relative power refers to any radiometric or photometric quantity which covers most biological representations of irradiance(McCluney 1994). The figure 1-6 indicates that relative power reaches its peak at the wavelength of 460 nanometers. Figure 1-6 Relative spectral power distribution for CIE D65 (Kevin Houser 2006) From SPD’s most completed form, more information can be derived from this range of spectroradiometers (Lucas et al. 2014). Instead of measuring SPD directly, an alternative way to get SPD data is to derive it from international glazing database (IGDB). For all five potential photoreceptors that can affect non-visual response, spectral power distribution is capable for outputting any other unit of measurement(Lucas et al. 2014). In txt. Format, SPD file contains at least two columns of data. The first column represents wavelength in nm while the second column is relative power in W·m -2 ·nm -1 . Figure 1-7 is a comparison of relative power distribution curves from three daylight 19 corelated color temperature which are 5000 K, 7000 K and 25000 K. This research took CCT=7000 K as an example. Figure 1-7 Comparison of spectral power distribution curves from three daylight conditions 1.7.2. Correlated Color Temperature (CCT) To describe a color, one common way is a 2-dimension approach with hue and saturation. It also known as chromaticity. To simplify the way to describe color, lighting fixture manufacturers use Correlated Color Temperature instead. Correlated Color Temperature (CCT) is a single number as a measurement of lighting source color appearance with K as its unit (“What Is Correlated Color Temperature? | Light Sources and Color | Lighting Answers | NLPIP” n.d.). It is an easy way to represent the equivalent value to the surface of luminance at certain temperature. However, it divided it into several points (certain temperature) while it is actually a smooth change. Also, the CCT assumes the receptor is a blackboy, which is an ideal model but never happens in the real world. However, as a proximity, CCT works well. For sunlight above the atmosphere, the CCT is around 5900 K. CCT of sunlight changes when it was scattered. CCT is an index for defining the sky when sunlight source need to be specified for daylight simulation. 1.7.3. Circadian Stimulus (CS) Circadian stimulus represents the effectiveness of the spectrally weighed irradiance at the cornea (Mariana G. Figueiro, Gonzales, and Pedler 2016). Ranging from 0.1 (threshold) to 0.7 (saturation), circadian stimulus is measured by acute melatonin suppression after an hour exposure (Mariana G. Figueiro, Gonzales, and Pedler 2016). Spectral power distribution serves as an input for the calculation of irradiance. The proposers addressed the following hints for researchers when they design with CS (Mariana G. Figueiro, Gonzales, and Pedler 2016). • CCT alone cannot describe SPD of a light source in a comprehensive way. It is possible that a light source with a lower CCT provided a greater CS than a light source with a higher CCT. • Unlike visual lighting measurement, CS should be considered at the eye level on vertical planes. • A study showed that the horizontal illuminance to vertical illuminance ratio is critical to efficacy of luminaire when designing with CS. • Both light level and spectrum affect CS design. 0 50 100 150 200 250 300 300 320 340 360 380 400 420 440 460 480 500 520 540 560 580 600 620 640 660 680 700 720 740 760 780 800 820 Relstive power (W·m -2 ·nm -1 ) Wavelength(nm) 5000k 7000k 25000k 20 • CS is related to the whole day light condition. • CS varies from building type and occupants. • Light layer is an alternative way when certain restrictions occurred. 1.7.4. Equivalent Melanopic Lux (EML) . 1.8. Scope of thesis This thesis proposes a workflow for architects to improve daylighting design quality at an early design stage. Bringing in visual lighting quality, circadian lighting quality and energy consumption, this parametric design workflow can be applied to help decision makers make choices on building orientation, window wall ratio and ceiling height, glazing type and interior finish. This research took a work space in San Francisco, California as a prototype. Well Building Standard was applied as a criterion for circadian-effects evaluation. The same workflow can be applied to other space type like resident space or space of dementia facilities. 1.9. Objectives of thesis There are two objectives within the entire workflow. An architect can apply this workflow for their research and provide an evaluation report of circadian lighting performance of daylighting for artificial lighting designers. Two objectives are listed below: • A workflow with an optimization tool to help architects make decisions in parametric design. This parametric design aiming at improving building performance based on energy consumption, effects of vision and effects of human beings’ circadian system. According to the designer’s preference, this workflow helps the decision maker to customize a building geometry for improving building performance at an early design stage. • A workflow for an architect to evaluate the circadian effects of daylighting. The evaluation report can serve as a guide for artificial lighting designers. For evaluating circadian lighting effects, a multi-channel approach was applied, aiming at more accurate evaluation results. The evaluation results indicating the biological dark area, circadian effects shift in both time and space. The evaluation report can be treated as a reference when a lighting designer is doing the lighting fixture layout and lighting control system as a supplement for a biological dark area. 21 2. CHAPTER 2 BACKGROUND AND LITERATURE REVIEW 2.1. Introduction Sustainability and occupants’ wellness of buildings draws building the attention of all those involved in the building process. Nowadays, building energy consumption is increasing due to comfort and equipment requirement. Based on a typical working schedule in the US which is 8:00 AM to 5:00 PM, energy consumption can be lower with a better architectural design integrating with daylighting. Daylight design focuses on solar radiation and visual lighting quality first. However, it is time to step further since circadian lighting concept have been raised. Architectural design is brought to another level which considers occupants’ wellness. An optional building standard named the WELL building standard proposes building performance indicators focusing on human beings’ health in architecture. In this standard, section 54 proposed circadian lighting requirement for 4 space types. Existing circadian lighting designs were rely on tunable lighting fixtures to increase productivity (Culver n.d.). Architects can make positive contributions to circadian lighting effects as well. With decisions on building geometry, architects can take advantage of daylight to enhance indoor circadian lighting quality for occupants’ wellness. This chapter focuses on the daylighting design objectives and circadian lighting measuring methods. 2.2. Daylight design objectives Knowing that daylight is an essential architectural component, architects work with building geometry to maximize its value. However, architects cannot make every aspect of daylighting effects perfect. So, it’s important for them to achieve a balance between objectives. Several daylight designs integrating with objectives like light level (Curtis and Mistrick, n.d.) (Konis, Gamas, and Kensek 2016), solar gain analysis (Curtis and Mistrick, n.d.), cost (Curtis and Mistrick, n.d.) (Wong 2017), exterior shading (Konis, Gamas, and Kensek 2016), energy consumption (Maltais and Gosselin 2017) (Konis, Gamas, and Kensek 2016) (Wong 2017), glare (Maltais and Gosselin 2017). Following sections described 5 objectives that commonly seen in a daylighting design workflow. 2.2.1. Visual light quality Visual lighting quality affects occupants’ comfort in the most straightforward way. It requires daylight to provide interior space a comfortable visual light level which is neither too high or too low. Usually, daylight level is evaluated by spatial daylight autonomy (sDA), but the threshold varies between designers. For buildings in California, Title 24 provides general illumination levels (lux) for different functional areas. Table 2-1, named “AS TAILORED METHOD LIGHTING POWER ALLOWANCES” showed value of general illumination level as well as required lighting power density (LPD). Primary Function Area General Illumination Level (Lux) Wall Display Lighting Power Density (W/ft) Allowed Combined Floor Display Power and Task Lighting Power Density (W/ft 2 ) Allowed Ornamental/Special Effect Lighting Power Density (W/ft 2 ) Auditorium Area 300 2.25 0.3 0.5 Civic Meeting Place 300 3.15 0.2 0.5 Convention, Conference, Multipurpose, and Meeting Center Areas 300 2.50 0.4 0.5 Dining Areas 200 1.50 0.6 0.5 Exhibit, Museum Areas 150 15.00 1.2 0.5 Financial Transaction Area 300 3.15 0.2 0.5 Grocery Store Area 500 8.00 0.9 0.5 22 Hotel Function Area 400 2.25 0.2 0.5 Hotel Lobby 200 3.15 0.2 0.5 Main Entry Lobby 200 0 0.2 0 Lounge Area 200 7.00 0 0.5 Malls and Atria 300 3.50 0.5 0.5 Religious Worship Area 300 1.50 0.5 0.5 Retail Merchandise Sale, and Showroom Areas 400 14.00 1.0 0.5 Motion Picture Theater Area 200 3.00 0 0.5 Performance Theater Area 200 6.00 0 0.5 Transportation Function Area 300 3.15 0.3 0.5 Waiting Area 300 3.15 0.2 0.5 Table 2-1 Tailored Methods Lighting Power Allowances 2.2.2. Solar gain Solar gain refers to a temperature increase caused by solar radiation. One design team divided solar gain into three categories which are cooling mode solar gain (good solar), cooling mode solar gain (bad solar) and net solar gain (net gain) (Curtis and Mistrick, n.d.). The optimization process of geometry was based on achieving higher good solar and lower bad solar like it shows in figure 2-1. Figure 2-1 Daylight design based on solar analysis (Curtis and Mistrick, n.d.) The advanced building takes exterior shading or interior shading as a solution for a better daylighting design. In the early design stage, it would be ideal to bring exterior shading. Exterior shading includes fixed shading and movable ones with a control system. Some research aiming at only shading strategies took moveable shading for an operation system to solve the problem (Osterhaus 2009) while most building geometry optimization workflow took exterior shading as a strategy. 2.2.3. Energy consumption Energy consumption related to daylight design consists of energy for HVAC systems and energy for artificial lighting. One research used annual energy for artificial lighting (AEL) (klm-h/m 2 ) evaluation (Maltais and Gosselin 2017). By involving the electricity cost into the performance indicator, it successfully provides a straightforward number indicating how much money can be saved by daylighting design but failed to take energy on the HVAC system. Figure 2-2 shows a comparison of electricity falls of commercial floorspace between 2017 and 2015. This chart presented by 23 U.S. Energy Information Administration showed a potential electricity saving on lighting. This can be achieved by the advanced daylight design strategies. An advanced daylight design strategy can also save electricity on heating or cooling commercial floorspace. Another performance indicator is Energy Use Intensity (EUI). It is an energy consumption summary on the operation of HVAC system. This one-dimension index is an ideal way to evaluate energy consumption with parametric design. Also, EUI is simply indicator which higher value means bad energy performance while lower value means better energy performance. Figure 2-2 Commercial electricity use per square foot of commercial floorspace falls (“EIA - Annual Energy Outlook 2018” n.d.) 2.2.4. Glare Glare refers to a phenomenon in vision made by redundant and unsuppressed brightness (“What Is Glare? | Light Pollution | Lighting Answers | NLPIP” n.d.). Generally, a glare issue is caused by one or more reasons listed below (Pierson, Wienold, and Bodart 2017): • High luminance intensity of lighting source • Adaption ability of human beings • Size of lighting source • Perspective position There are several different performance indicators and different thresholds for glare evaluation. The commonly used indicator in daylight design workflow is luminance value. Threshold varies from researchers and designers. Luminance is a measurement of luminous intensity per unit area of light coming from a specific direction. The higher the luminance value is, the greater amount of light is passed from the light source with in a specific solid angle (“Luminance” n.d.). One research indicated that a luminance level above 2000 cd/m 2 causes glare in a room (Dubois, n.d.). Another research showed a higher threshold which is 2500 cd/m 2 (Osterhaus 2009). Later researches stepped further to bring “discomfort level” into the threshold table. One research proposed a threshold which breaks down luminance levels. It suggested that a luminance value under 2000 cd/m 2 when referring to perceptible glare, 2000 cd/m 2 to 4000 cd/m 2 referring to reasonable glare, 4000cd/m 2 to 6000 cd/m 2 referring to uncomfortable glare and luminance value above 8000 cd/m 2 referring to the unacceptable glare. As a huge step further on glare evaluation, several glare indexes like Daylight Glare Probability (DGP) (Jens Christoffersen 2005)(Wienold and Christoffersen 2006), Daylight Glare Index (DGI) (Hopkinson 1972), Unified Glare Rating (UGR), Visual Comfort Probability (VCP) and CIE Glare Index (CGI) addressed the importance of contrast ratio and proved that absolute luminance is a problematic method. A luminance level of 2000 cd/m2 are possibly acceptable when people stay outside at a sunny day. However, it would be a potential glare if people stay in the dark space. For the glare evaluation in a building, DGP, DGI, UGR, VCP and VCP are relative reliable.. A research showed DGP has the best function of intolerable 24 glare while another research showed that GDP has the most robust results for daylighting simulation (Suk, Schiler, and Kensek 2017) (Jakubiec and Reinhart 2012). DGP severed as a glare performance indicator in detailed evaluation and more specific description are listed below (Jakubiec and Reinhart 2012): • Imperceptible Glare [0.30 > DGP] • Perceptible Glare [0.35 > DGP >= 0.30] • Disturbing Glare [0.45 > DGP >= 0.35] • Intolerable Glare [DGP >= 0.45] As another index to evaluate glare in the building, especially the glare without direct sunlight or specular reflections, DGI proposed by Hopkinson in 1972 is considered as a reliable source (Jakubiec and Reinhart 2012). Relations between degrees of perceived glare and DGI values are listed below (Jakubiec and Reinhart 2012). • Imperceptible Glare [18 > DGI] • Perceptible Glare [24 > DGI >= 18] • Disturbing Glare [31 > DGI >= 24] • Intolerable Glare [DGI >= 31] However, all glare evaluation methods and indices were proposed under the situation of a specific time spot and a given solid angle. Currently, there is no index, and it is hard to propose one representing the glare problem for the whole year. As an objective for optimization workflow, a glare problem should be taken only for a period of time instead of only single time spot. So, this research did not bring glare as an objective into optimization workflow. 2.3. Circadian lighting Based on melatonin synthesis suppression, circadian lighting refers to spectrally weighted retinal irradiance which contributes to human circadian system (Rea et al. 2010). To make positive contributions to the human circadian system with architectural design, designers should understand how light was converted into signal and how that affects human beings. This section introduces cells which be part of circadian lighting signals generation and how to quantify the circadian lighting effects. Since circadian lighting draw public attention, this section also shows two optional building standards that providing methods and requirements to integrating circadian lighting effects into architecture design. 2.3.1. Ganglion cells As a stimulus, light is converted into signal by ganglion cells in the brain. Figure 2-3 shows how ganglion cells work for the non-visual part of light. Rods and cones are the most well-known ganglion cells that work for human beings’ visual comfort. One research showed that they also works for non-visual comfort, but not critical (Berson 2002) (Hattar 2002). Another ganglion cell, called intrinsically photoreceptive retinal ganglion cell, outside the retina is critical to non-visual comfort. Of the photoreceptors, the rod is most sensitive to the wavelength at 500nm with photopigment called rod opsin (Lucas et al. 2014). Other photoreceptors, named cones, consist of three subcomponents which are short-wavelength cones (most sensitive to 420nm) with photopigment called S-cone photopsin, medium-wavelength cones (most sensitive to 535nm) with photopigment called M-cone photopsin and long-wavelength cones (most sensitive to 565nm) with photopigment called L-cone photopsin (Roenneberg and Merrow 2016). The last piece of photoreceptor named ipRGC is most sensitive to wavelength at 480nm with photopigment called melanopsin (Panda 2005). S-cone photopsin, M-cone photopsin, L-cone photopsin together with rod opsin are called extrinsic signals and melanopsin is intrinsic melanopsin photoreception. An ipRGC firing pattern consists of these two components while it is not the only pattern. A research showed there are at least five patterns while researchers are continuously digging more information (Ecker et al. 2010). Even though ipRGC only takes a small part in ganglion cells according to quantity, it is still critical because it projected to most non-visual response (Gooley et al. 2001). Without working in retina with such a big impact to non-visual comfort is answer to the question that why blind people still get circadian lighting effects even though they cannot see anything. 25 Figure 2-3 Non-visual light response generation ((Lucas et al. 2014)) 2.3.2. Calculation method Generally speaking, there are two ways to quantify lighting effects which are radiometry and photometry (DiLaura, David L., Kevin W. Houser, Richard G. Mistrick 2011). Based on how non-visual lighting effects were generated, a group of researchers proposed a circadian lighting calculation method according to spectral sensitivity (Lucas et al. 2014). This approach is suitable for the following reasons: • This is a one-dimensional unit which can be applied to all response under all circumstances • This method is based on a fundamental context of lighting which minimized the possibility of pre-processed errors The most commonly used approach for visual lighting calculation integrating with spectral sensitivity is a 3-channel approach which refers to dividing visible light into 3 wavelengths (R, G and B). It works well for the reason that photopigment of cones corresponds to these 3 wavelength ranges, so it can precisely mimic a light source. However, non-visual lighting effects came from 5 pieces and the part from cones is relatively little. A multi-channel approach was proposed by a research which tripling 3 wavelength ranges into totally 9 wavelength ranges for a more accurate calculation result (Inanici, Brennan, and Clark 2008). 2.3.3. Circadian lighting evaluation metric Measured on a vertical plane at humans’ eye level, Equivalent Melanopic Lux (EML) is a metric of circadian lighting (Lucas et al. 2014). Lux is the calculation unit of light that weighted in cones while EML is the calculation unit that weighted in intrinsically photosensitive retinal ganglion cells (ipRGCs). Figure 2-4 shows the relative power of both melanopin and vision through the wavelength. At the blue range of light, melanopic reaches its peak point indicating that circadian rhythm is most sensitive to blue light. Given melanopic curve and visual curve, the melanopic ratio can be determined and feeds calculation of EML. The formulation below shows the calculation process of the melanopic ratio. Start with the melanopic ratio, calculation processes, and both curves are shown below. WELL Building Standard also provides ratios for 9 common light sources which are listed in Table 2-2. 26 Figure 2-4 Melanopic and Visual Curves (“Appendix C: Tables | WELL Feature Library” 2017) 𝑀𝑒𝑙𝑎𝑛𝑜𝑝𝑖𝑐 𝑟𝑎𝑡𝑖𝑜=1.218× ∑ 𝑙𝑖𝑔ℎ𝑡 𝑜𝑢𝑡𝑝𝑢𝑡×𝑚𝑒𝑙𝑎𝑛𝑜𝑝𝑖𝑐 𝑐𝑢𝑟𝑣𝑒 9:; <:; ∑ 𝑙𝑖𝑔ℎ𝑡 𝑜𝑢𝑡𝑝𝑢𝑡×𝑣𝑖𝑠𝑢𝑎𝑙 𝑐𝑢𝑟𝑣𝑒 9:; <:; CCT (K) Light Source Ratio 2700 LED 0.45 3000 Fluorescent 0.45 2800 Incandescent 0.54 4000 Fluorescent 0.58 4000 LED 0.76 5450 CIE E (Equal Energy) 1.00 6500 Fluorescent 1.02 6500 Daylight 1.10 7500 Fluorescent 1.11 Table 2-2 Melanopic ratio from WELL Building Standard Once melanopic ratio was calculated, EML can be calculated by following the formulation “EML=L×R”. For example, assume a room with a single light source achieved 200 lux on the workplane. If the light source is 4000K LED, then it provided 152 (200×0.76) equivalent melanopic lux. If the light source is 6500K Daylight, then it provided 220 (200×1.1) equivalent melanopic lux (“Appendix C: Tables | WELL Feature Library” n.d.) However, EML failed to consider all variables including the time, intensity, duration wavelength and light exposure history (Chang, Scheer, and Czeisler 2011). A novel evaluation metric was proposed which integrates duration into EML (Konis 2017). As a supplement of the existing metric, this novel one considers duration and frequency which provides a well-considered evaluation metric. 2.4. Standard For the purpose of highly efficient, cost-saving and human health and wellness, Leadership in Energy and Environmental Design (LEED) and WELL Building Standard are most commonly used building standards nowadays. Both standards mention about circadian lighting in one of their section but have different focuses. 0 0.2 0.4 0.6 0.8 1 380 395 410 425 440 455 470 485 500 515 530 545 560 575 590 605 620 635 650 665 680 695 710 725 740 755 770 Relative Power Wavelength(nm) Melanopic and Visual Curve Melanopic Curve Visual Curve 27 2.4.1. LEED Provided by U.S. Green Building Council, the LEED ranking system mentions circadian lighting in its Environment Quality (EQ) credit, the Daylight section (USGBC 2016). Points in this section aim at saving energy by introducing more daylighting instead of using artificial lighting and reinforce circadian rhythms. LEED provides three approaches listed below for achieving points (USGBC 2016): • Spatial Daylight Autonomy and Annual Sunlight Exposure (simulation) • Illuminance Calculation (simulation) • Measurement All three options have an index for visual comfort and no direct relation to circadian rhythm or circadian lighting index. 2.4.2. WELL building Standard Developed by Delos Living LLC and managed by International WELL Building Institute (IWBI), the WELL Building Standard stepped further for its distinguishing visual lighting and circadian lighting. Under the Light category, in section 54, WELL uses “Melanopic Light Intensity” as an index for circadian lighting evaluation of mainly four categories which are working areas, living environments, breakrooms, and learning areas (International WELL building institute 2015). Melanopic light intensity is calculated by a unit of “Equivalent Melanopic Lux” which was discussed in section 2.3.3. WELL provides a relatively easy approach for architects to design based on humans’ circadian rhythm. Requirements for four space type were listed in Table 2-3. Space Type Requirement (or) Note Work Areas At least 200 EML are presented at more than 75% of workstations between 9:00 and 13:00 Measured on the vertical plane at the height of 1.2 m At least 150 EML provided by electrical lights on workstations Living Environments At least 200 EML Measured at the height of 1.2 m, facing the wall in the center of the room At least 50 EML Measured at the height of 0.76 m Breakrooms At least 250 EML on average Measured on the vertical plane at the height of 1.2 m Learning Areas At least 200 EML are presented at more than 75% of desks Measured on the vertical plane at the height of 1.2 m, for early education, elementary, middle and high schools, and adult education for student primary under 25 years of age At least 150 EML provided by electrical lights Ambient lights provide maintained illuminance on the vertical plane of equivalent melanopic lux greater than or equal to the lux recommendations in the Vertical (Ev) Targets in Table 3 of IES-ANSI RP-3-13, following the age group category most appropriate for the population serviced by the school Table 2-3 Summary of WELL building standard, section 54 2.5. Evaluation Tool Modeling software and simulation engines are used for the daylight design workflow. A series of software corresponding to each other are applied in this research for an accurate circadian lighting calculation. 2.5.1. Rhinoceros & Grasshopper Developed by Robert McNeel & Associates, Rhinoceros is a 3D modeling software for users to customize geometry design (“Rhinoceros” n.d.). A free plug-in software named ‘Grasshopper’ for rhinoceros is a parametric design tool 28 and also an interface for an analysis engineer like EnergyPlus and Radiance (“Grasshopper - Algorithmic Modeling for Rhino” n.d.). 2.5.2. Ladybug & Honeybee Ladybug is a radiance based daylighting simulation tool while Honeybee is an Energy Plus based energy simulation tool designed for Grasshopper (“Ladybug Tools - Grasshopper” n.d.). Ladybug and Honeybee provide a user-friendly interface for designers to combine parametric design comes with energy and daylight analysis. 2.5.3. Circadian stimulus calculator For an effective way of calculating circadian stimulus, the Lighting Research Center (LRC) at Rensselaer Polytechnic Institute released a free, open-access circadian stimulus calculator (“Light and Health | Research Programs | LRC” 2017). Figure 2-5 shows the interface of circadian stimulus calculator. An updated calculator can calculate the circadian stimulus with multiple light sources based on their SPDs. The chart below shows the main display of Circadian Stimulus calculator. As an Excel-based tool, the circadian stimulus calculator enables the user to manually select the light source from the database. Preloaded source including CIE D65, CIE A, and other artificial lighting source came with their SPD curves. With the selected light source, every combined illuminance value (lux) will output corresponding Circadian Light (CLA) and Circadian Stimulus based on SPD curve. An example below shows CLA=779 and CS=0.49 when the combined illuminance is 500 lux under standard CIE D65. One note here is that all circadian stimulus here is assuming that people receive this amount of light for an hour. Figure 2-5 Circadian Stimulus Calculator main display (“Light and Health | Research Programs | LRC” 2017) 2.5.4. Lark Spectral Lighting Developed by ZGF Architects LLP and University of Washington, a free, open-source plug-in in grasshopper named ‘Lark Spectral lighting’ provides four components which are “Convert SPD component”, “Spectral Sky Component”, “3-Channel Metrics” and “9-Channel Metrics” for researchers to make the multi-channel calculation. “Convert SPD component” converts spectral data into radiance materials which is a bonus compared to normal the EML calculation method. “Spectral Sky Component” works with DaySim to specify diffuse and direct irradiance under a specific location, sky condition, and time spot. “3-Channel Metrics” and “9-Channel Metrics” calculate both illuminance and luminance of circadian lighting. 29 2.6. Summary Architecture involvers including architects and global communities proposed circadian lighting concept in architectural design while this concept hasn’t been widely applied. The limitations of existing daylighting design workflow and focuses of this research are listed below: • Existing daylighting design workflows achieved the goal for a better building performance based on energy, solar heat gain, visual lighting and so on. However, these daylighting design workflows failed to consider about circadian lighting concept. This research proposed a simulation-based daylighting design workflow for early design stage which filled the gap between workflows and new concept. • The Conventional EML calculation method, which brings illuminance values and melanopic ratio is easy to approach, but it failed to address the features of materials and sky component. This research applied a multi- channel EML calculation approach which for more accurate results. • Lighting designers use artificial lighting as a supplement to meet the requirement of circadian lighting effects. However, lighting fixture layout failed to consider circadian effects from daylighting. In most space that has access to daylight, circadian effects have reached the required level (WELL Building Standard) during the daytime without any artificial lighting. This research proposes a circadian lighting evaluation report based on an informed parametric design which addresses daylighting circadian effects. This report provides a way for the lighting designer to make a smart lighting fixture layout and an advanced lighting control system. • This research takes advantage of visual programming software named “grasshopper” to run both daylighting simulation and iterative analysis. Also, a customized Python script in Grasshopper simplifies operation procedure which enhances user experience. 30 3. CHAPTER 3 METHODOLOGY Since the circadian lighting concept was brought into architectural design, it is possible for an architect to add this concept to daylight design workflow. This chapter proposes a new simulation-based daylighting design workflow for the early design stage. Also, a detailed circadian lighting evaluation report is provided to show changes in both time and space. Figure 3-1 shows the methodology diagram of this research. First, five sections describe how annual daylight simulation data are calculated and how iterative analysis is recorded. A parametric design including window wall ratio, building orientation, ceiling height and glazing materials was adapted in this workflow. Objectives including visual lighting quality, circadian lighting quality, and energy consumption are the criteria for seeking for better parametric design solutions. Exterior shading is added as a basic high-performance building component to improve daylighting performance. Interior finish color is also optimized only based on effects on circadian lighting. Except for building geometry, the site also contributes to building performance under different daylight conditions. Three sky conditions including a sunny day, an intermediate day and an overcast day, provide architects an approach to simulate different conditions among the year. An optimization tool designed for multi-objective optimization process is named “Octopus” runs iterative analysis for a high-performance geometry combination. The following sections describe how this multi-channel approach works in detail. The EML data from multi-channel approach are analyzed in chapter 5 seeking for circadian lighting effect shifts in space and time. Figure 3-1 New Daylight Design Workflow 3.1. Geometry A workspace is chosen for geometry optimization. Consisting of workspace, commercial buildings can be two-story buildings or skyscrapers. This research focuses on an office area in a commercial building which has an open space. Considering simulation time and similarity of space layout, this research uses a simplified building model. The same method can be applied to other building types like assisted-living apartments or dementia facilities. This research takes Flexlab (“FLEXLAB® | FLEXLAB” n.d.)from Lawrence Berkeley National Laboratory(LBNL) (“Berkeley Lab — Lawrence Berkeley National Laboratory” n.d.) with 6.4m by 9.4m layout as a prototype. Figure 3-2 and 3-3 show both an outside view and an inside view of Flexlab. Figure 3-4 shows the digital model in Rhino for this research. This section is divided into detailed parametric designs, including building orientation, window to wall ratio, ceiling height, glazing materials and exterior shading. This workflow uses US metrics for the reason that a daylight simulation engine “Radiance” has its default parameter set based on meters. 31 Figure 3-2 FlexLab (“FLEXLAB® | FLEXLAB” 2017) Figure 3-3 FlexLab inside view 32 Figure 3-4 Digital Test Space 3.1.1. Building orientation Building orientation is vital for passive design due to its effects on solar gain, views to outside and light levels (Morrissey, Moore, and Horne 2011). Non-ideal building orientation choice causes low energy efficiency, limited views from inside and poor lighting conditions from daylight. The parametric design enables different building orientations to be tested. Since this is a rectangle-layout commercial building, a full circle orientation range and WWR on all four walls is not necessary to mimic in the iterative analysis. 90-degree rotation ranges together with WWR change on south-facing walls are capable to mimic all possible combinations and similar strategies can be applied to reduce repeating and unnecessary calculations. 3.1.2. Window to wall ratio (WWR) As the main component to introduce daylighting, windows are placed with consideration of light level, solar gain, and glare. Normally, large windows are put on the equatorial wall aiming at getting less direct sunlight which causes glare and solar radiation. For this research, San Francisco was chosen as the location which is a city in the northern hemisphere. Aiming at minimal direct sunlight, windows were only put on south-facing wall in this research. For a definition of window size, window to wall ratio was adapted here. Higher WWR values refer to the larger window on a wall. For one thing, more lighting is introduced into space which benefits occupants’ vision and circadian rhythm. On the other hand, increasing solar heat gain requires more energy consumption for cooling indoor space. Windows are defined in grasshopper based on size and location of the south-facing wall. Parametric design controls variables on south facing the wall to seek for best WWR setting for the whole building. Size of the windows is determined by WWR with the value from 0-1. More specifically, based on Title 24, WWR value cannot exceed 40% for commercial building. Also, based on the formula for vertical fenestration minimum Visible Transmittance (VT≥0.11/WWR), WWR value should be larger than 0.26 because minimum VT is 0.42 (fixed window). WWR for each wall was set from 0.26 to 0.3 with the step of 0.1. Location of windows is set with sill height which is the distance from the lower 33 edge to floor and window height. the window can be either a whole piece or broken up and evenly distribute on walls. Figure 3-4 shows window (WWR=0.35 on south-facing wall broke up windows with 0.8m sill height). 3.1.3. Ceiling height With the same WWR, higher ceiling height means bigger glazing which contributes to daylight in a commercial building while energy consumption increases due to more solar gain in summer. For flexibility of future floor plan changes and space comfort in an office area, 2.7m is required as a minimum ceiling height (“3.2 Space Planning” n.d.). Ceiling height from 3 m to 5 m is used as optional ceiling height in test geometry. Architects can set different range according to different projects. 3.1.4. Glazing material In a building unit, glazing system provided the access to daylighting. With same glazing system, different glazing system has various performance on filtering amount of light and energy. Glazing system has performance indicators like the amount of layers, U-values, shading coefficient (SC) solar heat gain coefficient (SHGC), visible transmittance (VT or Tvis), relative heat gain (RHG). Developed by Lawrence Berkeley National Laboratory (LBNL), A tool named “WINDOW” can provide glazing performance information based on the database (“WINDOW | Windows and Daylighting” n.d.). A Glazing system can be taken apart into glazing layers. Each layer has their own performance indicator. Also developed by LBNL, a tool named “Optics” has access to International Glazing Database (IGDB) and can provide spectral properties (“IGDB | Windows and Daylighting” n.d., “Optics | Windows and Daylighting” n.d.). Instead of applying glazing system, this research uses four glazing layers from IGDB to test which one performs better based on objectives. Table 3-1 shows the summary of glazing materials. This research applies a multi-channel approach to defining glazing materials in a simulation which means only spectral data of glazing materials changes. In following optimization process, only visual lighting quality and circadian lighting quality are changing with different glazing materials while energy consumption remains the same. Label Type Thickness (mm) NFRC ID Tvis Emissivity Front Back A Monolithic 4 6955 0.897 0.837 0.837 B Laminate 5.66 3306 0.865 0.836 0.198 C Coated 6 13527 0.776 0.025 0.837 D Applied Film 3.67 8611 0.86 0.840 0.901 Table 3-1 Summary of glazing materials 3.1.5. Exterior shading Ideal daylighting design should not only increase daylight levels but also improve daylight quality. Glare and excessive solar heat gain often come along with increased daylight level. Except for glazing size, position, and materials, a shading system serves as an assistant trying to reduce the frequency of glare and a lower solar heat gain in some scenarios (Beck et al. 1999). Compared with the internal shading device, the external shading device performs better on eliminating the influence of taking in solar heat gain (Kim et al. 2012). Because glare is not in the scope of work, an external shading device was chosen as one design strategy. Different than other geometries, exterior shading is automatically generated by this workflow based on the worst-case scenario. A static vertical exterior louver on south- facing wall is provided based on maximum solar radiation during the year trying to solve the problem caused by direct sunlight. In this test, September 21 st is chosen for peak solar radiation day and louver system is provided based on solar radiation distribution from 8:00 to 17:00. Figure 3-5 shows how shading system generated in daylighting design workflow. 34 Figure 3-5 Shading Creation 3.2. Energy performance Energy simulation runs are performed in grasshopper with Energy Plus as an engine. Energy Use Intensity (EUI) is the metric in this workflow indicating energy performance. EUI is a simple metric expressing the amount of energy consumption one-unit area of a building in one year. Measured in kBtu/ft/yr or kWh/m2/yr (applied in this research), EUI can be an index comparing energy consumption with any scale buildings. As architects, they expect a low EUI value which basically means a better parametric design solution regarding energy performance. In order to compare energy only based on building geometry, all HVAC systems and schedules keep default template for the iterative analysis. Table 3-2 shows the thermal settings for this research. Building Type Office Cooling Setpoint 23.89℃ Heating Setpoint 21.11℃ Natural Ventilation Type Window Natural Ventilation Max Indoor Temp for Nat Vent 25℃ Min Outdoor Temp for Nat Vent 19℃ Max Outdoor Temp for Nat Vent 26℃ Table 3-2 Thermal setting for energy simulation 3.3. Daylighting quality With the building geometry and local weather data file, daylight simulation is processed by Radiance (“Radiance — Radsite” n.d.) and Daysim (“Daysim” n.d.). Ladybug and Honeybee in Grasshopper provide the interface for daylighting simulation and results analyzed. This part used solar radiation data in weather file to define sky condition like Figure 3-6 showed. Also, it provides geometric information like latitude, longitude, time zone and sun vectors (for shading creation). Both visual light quality and non-visual light quality are considered as objectives in the optimization process. There are several different metrics for both visual light and non-visual light quality with both real-time evaluation and annual evaluation. As objectives for optimizing building geometry, annual data need to be applied for overall daylighting performance. The subsection describes the way to calculate annual daylighting quality for both visual lighting quality and circadian lighting quality. 35 Figure 3-6 Local Weather Analysis 3.3.1. Visual lighting quality The annual daylighting simulation starts with local weather analysis. As is mentioned in the previous section, TMY weather contains direct measured solar ration which defines sky condition and it was applied in this simulation. Evaluation points are set at a horizontal work plane which is 0.8 m above the finish floor. Each point has 1.5 meters spacing with each other and represents the surrounding illuminance level. The intensity of points can be higher based on user’s preference for a more accurate calculation result. Illuminance level with unit “lux” describes indoor visual lighting quality. To describe annual indoor visual lighting conditions, this research adopts a climate-based metric named Spatial Daylight Autonomy. It shows the efficient level of daylight of the entire year in an indoor environment (Illuminating Engineering Society 2013). Spatial daylight autonomy is a metric that indicates the floor percentage with photopic illuminance over 300 lux more than 50% of occupied hours. However, researchers and designers can customize occupied hours and the photopic illuminance level range for the different type of projects. Considered about building type, this research sets occupied daily hour from 8:00 to 17:00. Normally, designers take Useful Daylight Illuminance ranging from 300lux up to 2000 lux as the criteria (Curtis and Mistrick, n.d.). For a building with exterior shading design, a study developed new criteria which combined shading and occupants’ visual comfort by lower the limits to 100 lux (Konis 2013). All evaluation points with photopic illuminance under 100 lux were treated as inadequate indoor lighting and all evaluation points with photopic lux above 2000 lux were treated as a potential glare issue. The percentage of useful daylight illuminance is the indicator of visual light performance. The higher the percentage is, the better visual light performance provided by building geometry. Based on Title 24 (Section 140.6: General Illuminance Level), this research sets 300 lux as a threshold. Figure 3-7 shows the annual hourly-based daylighting simulation results. Any daytime illuminance values above 300 lux are marked as positive, otherwise, they are negative. The result is the percentage of annual daytime (8:00 ~ 17:00) illuminance values above 300 lux. As an objective for optimization process, this research uses the average value of 6 evaluation points’ percentage values to describe visual lighting quality. 36 Figure 3-7 Annual daylighting simulation illuminance results 3.3.2. Circadian lighting quality The same as the visual lighting illuminance calculation method, EML calculation method in this section uses solar radiation from weather data to define sky condition. Evaluation points are set at a vertical plane at height of eye level which is 1.2 m above the finish floor. From the top view, evaluation points layout is the same as evaluation points for a visual lighting test. However, each point has four vectors representing different view directions. The intensity of points can be higher based on the user’s preference for more accurate calculation results. In this section, the EML calculation method takes advantage of the illuminance file from annual daylighting simulation by dividing by melanopic ratio (ratio=1.1, CCT=6500K. Daylight). For one-hour data, there are 24 EML results representing the whole space. Based on WELL Building Standard (Section 54: Circadian lighting design), this research sets 200 equivalent melanopic lux as threshold and 9:00 ~ 13:00 as evaluation time period. Any evaluation point has EML value above 220 are marked as positive, otherwise, are marked as negative. The results are the percentage of time during the evaluation time period that achieved threshold. As an objective for optimization process, this research 37 used the average value of 24 EML percentage values to describe circadian lighting quality. In this research, circadian effective percentage (CEP) is the proposed metric for circadian lighting quality. Also, this research proposed a definition for different level of circadian lighting effects. In an indoor space, any portion that has over 200 EML is defined as “biologically bright” portion. This portion gets sufficient circadian lighting stimulus in that time. Then any portion has less than 200 EML but more than 100 EML is defined as “biologically grey” portion. This portion gets circadian lighting stimulus while it is not strong enough to make this portion “circadian effective”. The rest portion has less than 100 EML results are defined as “biologically dark” portion. This portion barely have circadian lighting stimulus. 3.4. Iterative analysis With building geometries affecting objectives, the informed design solution with an optimal building geometry combination is provided by Octopus. Four design parameters including building orientation, window wall ratio, ceiling height and glazing materials and their variation ranges are listed in table 3-3. Together with parameters, three building performance indicators including energy consumption, visual lighting quality and circadian lighting quality and their metrics are listed in table 3-4. In this case, high energy performance requires low WWR, but visual lighting quality decreased at the same time. To solve this conflict, a multi-objective problem solver with genetic algorithms (GA) is applied here. A research showed that GA is the best solution for the multi-objective problem for its easy access (Konak, Coit, and Smith 2006). In GA, all variables are called genes. In this case, building orientation, window wall ratio and ceiling height and glazing materials are genes. One combination of genes is a chromosome. Objective refers to a value. In an optimization workflow, the objective can be maximized or minimized. This research adapted values of EUI, 1- sDA, and 1-CEP as objectives. The difference between single-objective optimization and multi-objective optimization lies in the generation of chromosomes. GA for single-objective optimization generates chromosomes which contain the correct genes and seeking for the better chromosome. In that case, correct gene dominates the next chromosome generation which saves time for “best solution” seeking. Other than that, another GA which designed for multi- objective optimization aims at finding the diversity of chromosomes. To achieve that diverse chromosomes generation, two operators name crossover and mutation are critical to generating chromosomes. Crossover refers to the chance that two chromosomes exchanging their genes. A high crossover rate means a high probability that genes in a chromosome come from two existing chromosomes. Compared to crossover, mutation plays opposite role which defines genetic diversity in chromosome generation. High mutation refers to a big change of genes in next chromosome generation. An optimization tool in grasshopper entitled “Octopus” runs an iterative analysis to test all possible geometry combinations and finalize optional choices with a balance of three objectives. Figure 3-8 shows the interface of Octopus. Parameter Window Wall Ratio Ceiling Height (m) Building orientation (°) Glazing material Range [0.3,0.35] [3,5] [-30,30] A, B, C, D Table 3-3 Summary of parameters and ranges Performance Indicator Energy Consumption Visual lighting quality Circadian lighting quality Metric Energy Use Intensity, EUI (kWh/m 2 /yr) Spatial Daylight Autonomy, sDA (%) Circadian Effective Percentage, CEP (%) Note 9:00-17:00, evaluation points on a horizontal plane at the height of 0.8 m, lower limit: 300 lux 9:00-13:00, evaluation points on a vertical plane at the height of 1.2 m, lower limit: 200 EML Table 3-4 Summary of Objective Function 38 Figure 3-8 Optimization tool interface Octopus provides a process control panel to customize the optimization process. Other than two critical operators, octopus users can also set “population size” before iterative analysis which controls time duration of the whole process. Depend on design stage or complexity of building geometry, users can input a low population size which is “time- saving” strategy. The process can stop or restart any time which offers a flexible schedule. Items on algorithm setting panel listed below. • Mut. Probability • Crossover Rate • Population Size • Max. Generation • Record Interval • Save Interval The display panel offers choices to visible or invisible results based on user preference. Also, it provides result marks, calculation history records and other features for users to analysis, trace all solutions. Table 3-5 showed a summary of algorithm setting for this research. For the convenience of picking up an informed design, this research uses 1-sDA for visual lighting quality and 1-CEP for circadian lighting quality. Both 1-sDA and 1-CEP are ranging from 0-1, and both values close to 0 means a high quality. Together with an indicator for energy performance, EUI, any point on a coordinate system close to origin means a geometry design solution which refers to building geometries with overall high-performance. Elitism Mut. Probability Mutation Rate Crossover Rate Population Size Max Generations Record Interval 0.5 0.100 0.5 0.8 100 0 1 Table 3-5 Summary of optimization setting 39 3.5. Interior finish selection In a workplace, daylighting is bouncing from different surfaces. The spectral property of lighting changes every time through the bounces based on surface materials. Interior finish color is a factor for circadian lighting even though they have the same reflectance. This section is seeking for a better interior finish color based on circadian lighting effects. Three scenarios are tested here which are red walls, blue walls, and green walls. This section applied control variate method which is average reflectance of the wall is set as 0.7 while the appearance is different. Table 3-6 described simulation geometry setting for interior finish selection. Ceiling Height (m) Window Wall Ratio (%) Building Orientation (°) Glazing Material 3 35 0 Fourth_Madision Table 3-6 Geometry setting for interior finish selection 3.6. Detailed circadian lighting evaluation A well-developed circadian daylighting design should consist of both daylighting and artificial lighting. Architects try to achieve the goal that maximizing daylight value on circadian lighting effects. Consequently, the whole project can save money and energy from installing and using artificial lighting as supplements. It is an essential conservation between architects and lighting designers regarding circadian lighting performance. Architects should provide a detailed circadian daylight evaluation as a guide for the lighting designer to do lighting fixtures layout and lighting control system as supplements. More accurate results can be addressed by the multi-channel approach. Conventional EML calculation approach was developed based on photopic lux which is problematic. For visual comfort illuminance calculation, the visual spectrum was divided into three wavelength ranges which are R, G and B. It works for photopic illuminance because that is how cone works. Other than cones, ipRGCs is not working in that way so 3-channel (R, G and B) approach cannot accurately define the whole wavelength of circadian light. A multi-channel approach was proposed for EML calculation by tripling each three-channel. There are two typical ways to calculate EML. The most common way is provided by WELL Building Standard, using the formulation showed below EML=Lux × Ratio (R𝑎𝑡𝑖𝑜=1.218× ∑ ?@ABC DECFEC×GH?IJDF@K KELMH NOP QOP ∑ ?@ABC DECFEC×M@REI? KELMH NOP QOP ) The advantage of this approach lies in that “lux” is easy to get. Software like AGI 32 are well developed for calculation of “lux” (“Lighting Design Software by Lighting Analysts” n.d.). Within this method, users can calculate EML easily, While it brings the limitation that the “lux” we get is calculated by 3-channel-approach. Even though the ratio is accurate, the EML is still problematic. Also. Index “lux” was designed for visual effects so evaluation points are set at height of workstation (0.8 m) on a horizontal plane. It should be noticed that if Lux is taken from other sources, designers need to confirm that all evaluation points was set at height of eye level on a vertical plane. Another way takes data from SPD curve which can be expressed in 3-channel-approach, a multi-channel approach like it shows below or n-channel approach (Inanici, Brennan, and Clark 2008). 3-channel approach: 𝐸𝑀𝐿=149×(0.0018×𝐼𝑟𝑟𝑎𝑑(𝑅)+0.1532×𝐼𝑟𝑟𝑎𝑑(𝐺)+0.4024×𝐼𝑟𝑟𝑎𝑑(𝐵)) multi-channel-approach: 𝐸𝑀𝐿=149×(0.0166×𝐼𝑟𝑟𝑎𝑑(𝐵1)+0.1819×𝐼𝑟𝑟𝑎𝑑(𝐵2)+0.3973×𝐼𝑟𝑟𝑎𝑑(𝐵3)+ 0.2468×𝐼𝑟𝑟𝑎𝑑(𝐺1)+0.1204×𝐼𝑟𝑟𝑎𝑑(𝐺2)+0.0351×𝐼𝑟𝑟𝑎𝑑(𝐺3)+0.0018×𝐼𝑟𝑟𝑎𝑑(𝑅1)+0×𝐼𝑟𝑟𝑎𝑑(𝑅2)+ 0×𝐼𝑟𝑟𝑎𝑑(𝑅3)) This approach starts from the beginning of calculation which brings accuracy but is relatively harder to apply. However, a free plug-in in grasshopper named “Lark” offers several components for designers to realize multi- channel-approach. Figure 3-9 showed how detailed circadian lighting effects were evaluated. Table 3-7 described the geometry setting for simulation in this part. A circadian lighting calculation tool named “Lark” and its four components were described below (“Lark Spectral Lighting” n.d.). 40 Figure 3-9 Detailed Circadian Lighting Effects Evaluation Workflow Ceiling Height (m) Window Wall Ratio (%) Building Orientation (°) Glazing materials Wall materials 4 35 26 Fourth_Madision Macbeth_3 Table 3-7 Geometry setting for detailed circadian lighting evaluation 3.6.1. Convert SPD component This component serves as a converter which converts SPD data into radiance materials. The input file has at least two columns of data. The first column data is wavelength and the second column of data is transmittance (glazing) or reflectance (surface). Take optics 6 output glazing file named as “Fourth_Madision” as an example, Figure 3-10 shows the data structure. If channel 3 analysis method is activated, SPD data is divided into three parts based on wavelength range: [586,780] (R), [498,586] (G) and [380,498] (B). Otherwise, SPD data is divided into nine parts based on wavelength range: [380,422] (B1), [422,460] (B2), [460,498] (B3), [498,524] (G1), [524,550] (G2), [550,586] (G3), [586,650] (R1), [650,714] (R2) and [714,780] (R3). Then average transmittance was calculated based on each range (without transmittance of start and end wavelength of each range). With three or nine average transmittance values (based on channel type), this component converts transmittance (tn) to transmissivity (Tn) with formulation below: 𝑡𝑛 = (0.8402528435+0.0072522239×𝑇𝑛 e ) e 0.003626119×𝑇𝑛 Figure 3-11 shows the Fourth_Madision output file Structure. Line 1 shows the name of this material. Line 2 refers to roughness while line 3 refers to specularity. Four values in line 4 refer to material type and transmissivity. 41 Figure 3-10 Fourth_Madision Glazing SPD data structure Figure 3-11 Fourth_Madision output data structure 3.6.2. Spectral Sky This component aims at generating sky condition (sunny, intermediate or cloudy sky) at a specific location and time. It starts with converting global illuminance to diffuse horizontal irradiance by divided it by 179 lm/W (internal luminous efficacy model in Radiance). Take 3 channel approach as an example. 𝐷𝑖𝑓𝑓𝑢𝑠𝑒 ℎ𝑜𝑟𝑖𝑧𝑜𝑛𝑡𝑎𝑙 𝑖𝑙𝑙𝑢𝑚𝑖𝑛𝑎𝑛𝑐𝑒 = 𝑑𝑖𝑓𝑓𝑢𝑠𝑒 ℎ𝑜𝑟𝑖𝑧𝑜𝑛𝑡𝑎𝑙 𝑖𝑟𝑟𝑎𝑑𝑖𝑎𝑛𝑐𝑒 𝑟×𝑣 L +𝑔×𝑣 A +𝑏×𝑣 j r-average relative power of wavelength [586,780] of sky SPD; g-average relative power of wavelength [498,586] of sky SPD; b-average relative power of wavelength [380,498] of sky SPD; vr=0.2537, the photopic coefficient of channel r; 42 vg=0.6635, the photopic coefficient of channel g; vb=0.0622, the photopic coefficient of channel b; 3.6.3. Multi-channel simulation Figure 3-12 shows an irradiance data file which was fed in 3 channel metrics for illuminance calculation. As a grid- based simulation, 12 lines of data represent irradiance data of 12 points. Each column represents irradiance data for each wavelength range. Column one is irradiance data of wavelength from 586-780nm (R) and column two and three are irradiance data of wavelength from 498-586 nm (G) and 380-498 nm (B). It calculates photopic illuminance, circadian illuminance (Rea) and circadian illuminance (Lucas) with following formulations: 𝑃ℎ𝑜𝑡𝑜𝑝𝑖𝑐 𝐼𝑙𝑙𝑢𝑚𝑖𝑛𝑎𝑛𝑐𝑒 =179×(0.2537×𝐼𝑟𝑟𝑎𝑑(𝑅)+0.6635×𝐼𝑟𝑟𝑎𝑑(𝐺)+0.0622×𝐼𝑟𝑟𝑎𝑑(𝐵)) 𝐶𝑖𝑟_𝑅𝑒𝑎 𝐼𝑙𝑙𝑢𝑚𝑖𝑛𝑎𝑛𝑐𝑒 =130×(−0.0405×𝐼𝑟𝑟𝑎𝑑(𝑅)+0.1532×𝐼𝑟𝑟𝑎𝑑(𝐺)+0.8873×𝐼𝑟𝑟𝑎𝑑(𝐵)) 𝐶𝑖𝑟_𝐿𝑢𝑐𝑎𝑠 𝐼𝑙𝑙𝑢𝑚𝑖𝑛𝑎𝑛𝑐𝑒 =149×(0.0018×𝐼𝑟𝑟𝑎𝑑(𝑅)+0.1532×𝐼𝑟𝑟𝑎𝑑(𝐺)+0.4024×𝐼𝑟𝑟𝑎𝑑(𝐵)) Named as “The RGB Approximation”, this approach has a highly accurate illuminance measurement (“Measuring and Calculating Lux Values, Part 2” n.d.). More accurate illuminance can be achieved with a more accurate representation of the actual spectral composition just like 9 channel metrics did below (“Measuring and Calculating Lux Values, Part 2” n.d.). Figure 3-12 Irradiance data structure Similar to 3 channel metrics, 9 channel approach calculates photopic illuminance, circadian illuminance (Rea) and circadian illuminance (Lucas) with following formulations: 𝑃ℎ𝑜𝑡𝑜𝑝𝑖𝑐 𝐼𝑙𝑙𝑢𝑚𝑖𝑛𝑎𝑛𝑐𝑒 =179×(0.0004×𝐼𝑟𝑟𝑎𝑑(𝐵1)+0.0095×𝐼𝑟𝑟𝑎𝑑(𝐵2)+0.0522×𝐼𝑟𝑟𝑎𝑑(𝐵3) +0.1288×𝐼𝑟𝑟𝑎𝑑(𝐺1)+0.2231×𝐼𝑟𝑟𝑎𝑑(𝐺2)+0.3174×𝐼𝑟𝑟𝑎𝑑(𝐺3)+0.2521×𝐼𝑟𝑟𝑎𝑑(𝑅1) +0.0162×𝐼𝑟𝑟𝑎𝑑(𝑅2)+0.0002×𝐼𝑟𝑟𝑎𝑑(𝑅3)) 𝐶𝑖𝑟_𝑅𝑒𝑎 𝐼𝑙𝑙𝑢𝑚𝑖𝑛𝑎𝑛𝑐𝑒 =130×(0.0669×𝐼𝑟𝑟𝑎𝑑(𝐵1)+0.394×𝐼𝑟𝑟𝑎𝑑(𝐵2)+0.4264×𝐼𝑟𝑟𝑎𝑑(𝐵3) +0.1464×𝐼𝑟𝑟𝑎𝑑(𝐺1)+0.0362×𝐼𝑟𝑟𝑎𝑑(𝐺2)+(−0.0294)×𝐼𝑟𝑟𝑎𝑑(𝐺3)) +(−0.038)×𝐼𝑟𝑟𝑎𝑑(𝑅1))+(−0.0026)×𝐼𝑟𝑟𝑎𝑑(𝑅2)+0×𝐼𝑟𝑟𝑎𝑑(𝑅3)) 𝐶𝑖𝑟_𝐿𝑢𝑐𝑎𝑠 𝐼𝑙𝑙𝑢𝑚𝑖𝑛𝑎𝑛𝑐𝑒 =149×(0.0166×𝐼𝑟𝑟𝑎𝑑(𝐵1)+0.1819×𝐼𝑟𝑟𝑎𝑑(𝐵2)+0.3973×𝐼𝑟𝑟𝑎𝑑(𝐵3) +0.2468×𝐼𝑟𝑟𝑎𝑑(𝐺1)+0.1204×𝐼𝑟𝑟𝑎𝑑(𝐺2)+0.0351×𝐼𝑟𝑟𝑎𝑑(𝐺3)+0.0018×𝐼𝑟𝑟𝑎𝑑(𝑅1) +0×𝐼𝑟𝑟𝑎𝑑(𝑅2)+0×𝐼𝑟𝑟𝑎𝑑(𝑅3)) 43 3.6.4. Results Different than illuminance file-based approach, the multi-channel approach defines spectral sky and building materials based on SPD curve. Compared to relying on solar radiation data in the weather file, a multi-channel approach successfully described factors in the sky as well as factors in materials. This research uses annual hourly-based EML results from a multi-channel approach to show circadian lighting shift in both space and time. Based on WELL building standard, annual hourly-based data refers to data 9:00 to 13:00 every day. 3.6.5. Circadian lighting changes Since the light source is the sun, the circadian lighting changes mainly due to sky condition and relative position of the building and sun. This research focuses on changes caused by relative position changes. Figure 3-13 shows different sun positions through different hours of a day, different days of a month and different months of a year. Since daily change through a month does not change the relative position between the sun and building a lot, this research mainly focuses hourly change through a day and monthly change through a year. Figure 3-13 Relative sun position changes 3.6.5.1. Hourly change Annual hourly-based data of all vectors are evaluated to determine the changes in both space and time. In this section, circadian lighting change is presented in two ways. For one way, the point that all four vectors that achieved the minimum threshold (200 EML) is marked as Circadian-Effective (CE). Otherwise, that evaluation point is marked as Circadian-Ineffective (CI) (Konis 2017). Hourly changes through one day with either CE or CI are marked in a diagram for lighting designers to refer to. Instead of categorizing original data into CE or CI, another way took the average value of four vectors to represent circadian lighting value of that evaluation point. A dynamic heat map with representative values shows the hourly change of a specific day from 9:00 to 13:00. 3.6.5.2. Monthly change WELL building standard requires all the time from 9:00 to 13:00 that lighting can present over 200 EML. In this section, the author takes the average value on four vertical surfaces to represent circadian lighting value of that evaluation point. For 5 time-spots of a day (9:00 ~ 13:00), any portion with EML that all meet a minimum threshold (200 EML) is marked as CE. Otherwise, the evaluation point is marked as CI. A diagram showed monthly changed through a year is served as a guide for the lighting designer to make a fixture layout and lighting control system. 3.7. Summary The methodology section explains the entire workflow in detail. The multi-objective optimization process takes advantage of the ratio-EML calculation method to get annual hourly-based data while detailed circadian lighting evaluation takes advantage of the accuracy of multi-channel EML calculation approach. Iterative analysis pre- processes objective data to make better geometry combination close to the origin. Detailed evaluation filters and visualizes data for both lighting designer and owners to understand circadian lighting changes through time and space. 44 4. CHAPTER 4 RESULTS This research proposes a new workflow which can be applied in the early stage of daylight design for geometric solutions. This chapter reports results based on different steps of the workflow. The first sub-section shows the results from “Octopus”. Octopus is an interface for a genetic algorithm (GA) based multi-objective optimization. In this case, the input genes are building orientation, window wall ratio, ceiling height and glazing materials and objectives are energy use intensity (EUI), spatial Daylight Autonomy (sDA) and Circadian Effective Percentage (CEP). More than 50 solutions were analyzed and presented with their performance indicators. “EnergyPlus” serves as the engine to calculate energy consumption and “Radiance” serves as the engine to calculate both visual lighting effects and circadian lighting effects. 8 case solutions were picked to make a comparison of a single-objective best- case scenario, single-objective worst-case scenario, and overall best-case scenario. The raw results for remaining sub-sections are EML results of each evaluation point. The second sub-section shows the lighting condition comparison of three interior finishes. Circadian lighting effects on four view directions are the dominant criteria of three interior finish color selections. Also, three HDR images from the same evaluation point with the same view direction are presented for potential glare analysis. The last sub-section describes the EML results in both space and time. The EML results contain 5 time-spots per day (9:00~ 13:00), one day per month (21 st ) and five months for a year (March, June, April, October, December). For each time-spot, there are 35 (5 by 7) evaluation points at the height of 1.2 m and 4 view directions of each evaluation point. In total, there are 3500 EML results for detailed circadian lighting effects evaluation. This sub-section sorts EML values separately based on view directions and time. Done to here 4.1. Iterative analysis results These GA based iterative analysis results are presented with all analyzed solutions and selected-cases results. This multi-objective optimization workflow contains four genes which are building orientation, window wall ratio, ceiling height and glazing materials and three objectives which are energy use intensity (EUI), spatial daylight autonomy (sDA) and circadian effective percentage (CEP). For the convenience of data visualization, this research took EUI, 1-sDA, 1-CEP as objectives so that each objective close to 0 is indicating a better performance. Figure 4-1 shows iterative analysis results from “Octopus”. Each point in this figure represents one design solution. The location of the point in this space indicates the performance of this design solution. Each point can be described in coordinate format and each coordinate on the axis indicating one performance indicator value. In this case, any point close to the origin can be defined as an overall better solution. Figure 4-1 also shows the extremum of three objectives among all generated analyzed solutions. The points marked with yellow are worse case scenarios based on a single-objective while the points marked with purple are the overall better solutions. 45 Figure 4-1 Iterative analysis results from "Octopus" Figure 4-2 and table 4-1 pick 8 case solutions which are reference model, maximum energy consumption solution, minimum energy consumption, maximum visual lighting quality solution, minimum visual lighting quality solution, maximum circadian lighting quality solution, minimum circadian lighting quality solution, and overall best solution. Figure 4-2 is the comparison of configurations of case models while table 4-1 compares the statistics of both design geometries and the performance of the case models. Take circadian lighting as the only objective, the best-case scenario has 70% of the time (circadian lighting time: 9:00 – 13:00) as circadian effective when daylighting is the only light source, which is 89.19% higher than the worst-case scenario. However, the best solution based on only circadian lighting effects sacrificed the energy consumption by 15.85% and sacrificed visual lighting quality by 32.20%. The overall better solution balanced three objectives and has improved three aspects compared to the poor design solution. Considering energy performance, the overall better solution used 49.91 kWh/m 2 /yr, which is 13.7% lower than the highest energy consumption solution. Also, the overall better solution has 67% of the time as circadian effective which is 81.08% higher than circadian lighting worst-case solution and 70% of the time (visual lighting time: 9:00-17:00) meeting visual lighting requirement illuminance level which is 27.27% higher than visual lighting worst-case solution. Among all three objectives, circadian lighting effects show the biggest potential in this optimization workflow. Figure 4-3 to figure 4-5 show the comparison between the overall best solution and separate worst solution. The results prove the importance of building geometry design depend on the three objectives. So, the new workflow is applicable to early daylight design stage and circadian lighting effect notably draws the designer’s attention. 46 Figure 4-2 Comparison of configurations of reference and characteristic solutions Window Wall Ratio Ceiling Height (m) Building Orientation (°) Glazing Type Energy Use Intensity, EUI (kWh/m 2 /yr) Spatial Daylight Autonomy, sDA (%) Circadian Effective Percentage, CEP (%) Reference model 0.33 4 0 A 51.55 71 56 Maximum energy consumption solution 0.34 5 -25 C 57.88 79 61 Minimum energy consumption solution 0.34 3 27 B 45.56 61 57 Maximum visual lighting quality solution 0.35 5 21 C 55.29 82 74 Minimum visual lighting quality solution 0.31 3 11 C 46.64 55 47 47 Maximum circadian lighting quality solution 0.34 5 27 C 54.82 78 70 Minimum circadian lighting quality solution 0.31 3 -27 B 47.32 59 37 Overall best solution 0.34 4 26 C 49.91 70 67 Table 4-1 Comparison of 8 case solutions from iterative analysis Figure 4-3 Comparison of overall better solution and worst energy consumption solution Figure 4-4 Comparison of overall better solution and worst visual lighting quality solution Figure 4-5 Comparison of overall better solution and worst circadian lighting quality solution 4.2. Interior finishes circadian lighting quality results This section presents the results of circadian effective percentage between three interior finish colors which are red, blue and green. All 35 evaluation points are arranged on the horizontal plane at the height of 1.2 m and all EML results are calculated on vertical planes of points with four view directions. Figure 4-6 shows how four view directions are arranged. View_4 is looking towards the glazing and view_3 is looking backward the glazing. View_1 and view_2 are looking parallel to the wall with glazing. Figure 4-7 shows views from 4 directions. 44 46 48 50 52 54 56 58 60 Overall best solution Worst solution Energy Consumption, EUI (kWh/m2/yr) 0 10 20 30 40 50 60 70 80 Overall best solution Worst solution Visual Lighitng quality, sDA (%) 0 10 20 30 40 50 60 70 80 Overall best solution Worst solution Circadain lighting quality, CEP (%) 48 Figure 4-6 Arrangement of view directions Figure 4-7 views from 4 directions Figure 4-8 describes the fluctuation of EML results under three interior finish selections at 9:00, March 21 st . In this box & whisker chart, for each view, every interior finish color EML results contains 35 values which are 35 evaluation points on the horizontal plane. The higher the box is located on the chart, the better the circadian lighting quality achieved. From perspectives of view_1 and view_2, the blue interior finish has the most positive impact on circadian lighting quality while the red interior finish has the worst circadian lighting quality among these three selections. The blue range of light being reflected by the surface and other range of light being absorbed by the surface make the surface blue. In this case, comparing to other interior finish colors, blue interior surface absorbed a minimum blue range of light, which is the most sensitive range of light for circadian rhythm. With three selections, Circadian effective percentage doesn’t vary a lot based on view_4 while there is a huge fluctuation apparent in view_3. Interior finish has an impact on circadian lighting quality mainly because the light that the eye received is reflected by the surface. All daylight comes through the glazing first so glazing materials affect all light for occupants’ circadian rhythm. However, the interior finish does not affect the inside circadian lighting quality if all the circadian lighting that the eyes receives is coming directly from the glazing. Otherwise, the interior finish does have an impact on circadian lighting quality. The light that the eye receives bounces more from the surface, the circadian lighting quality changes more based on surface color. EML results on view_4 indicates that most light that view_4 received is coming directly through the glazing because the different interior surface color did not affect the circadian lighting quality a lot. Different from the results of view_4, EML results of view 3 49 suggests that most light that this view received is reflected by interior surfaces. View_1 and View_2 also indicate that there is a certain amount of light that the eye receives that is reflected by the interior surfaces. Take average EML results of green interior finish as a reference, for view_1, the blue interior finish improved circadian lighting quality by 10.52% and red interior finish sacrificed circadian lighting quality by 16.91%. Take the same evaluation method for view_3, the blue interior finish improved circadian lighting quality by 60.58% and the red interior finish sacrificed circadian lighting quality by 97.48%. By comparing these two view directions, it shows that the light that view_3 received bounces more times than the light that view_1 received because interior surfaces have a larger impact on view 3. If view 1 is chosen for the workstation layout, the blue interior finish can improve circadian effective percentage by 15.39% on March 21 st , 25.01% on June 21 st and 17.14% on December 21 st compared to red interior finish. If view 2 is chosen for workstation layout, the blue interior finish can improve circadian effective percentage by 58.33% on March 21 st , 78.58% on June 21 st and 31.26% on December 21 st compared to red interior finish. Figure 4-8 Comparison of EML fluctuation based on three interior finishes, Mar 21 st , 9:00 4.3. Detailed circadian lighting evaluation results This section sorts the data from the result database based on design objectives. The first part is the results of view direction selection. Circadian effective percentage of four view directions in a whole year is presented as a guide for an interior designer. The second part sorts the EML results showing the circadian lighting effects shift in both space and time. 4.3.1. View direction selection result With the same sky conditions, circadian lighting effects change with sun position changes. For the purpose of saving time, this research chose five typical days which are March 21 st , June 21 st , April 21 st , October 21 st and December 21 st to represent all scenarios through a year. Figure 4-9 shows all sun positions and typical sun positions and table 50 4-2 shows the summary of typical days and the days that they represent. Each day contains 5 time-spots and each time -spot comes with a circadian effective percentage value. Figure 4-9 Comparison of all sun positions and typical sun positions Typical day All days in the month March 21 st March/ September June 21 st June/ May/ July April 21 st April/ August October 21 st October/ February December 21 st December/ January/ November Table 4-2 Summary of typical days and the days that they represent Figure 4-10 shows the comparison of circadian effective percentages based on view directions. In this radar chart, the month marked in brackets indicates the input circadian effective percentage value. The larger the area covered, the better circadian lighting quality that view direction achieved. It shows that view 3 fails to provide effective circadian lighting at any time during the year while view 4 can provide effective circadian lighting for the most area during the whole year. However, there is a potential problem for view 4, which is a glare issue. View 4 is the direction towards the glazing. With fixed shading system it is hard to make sure no glare happens during a year. So, if view 4 is not protected, then either option of these two is ideal for workstation layout. Both view 1 and view 2 are parallel to the wall with the glazing system. Figure 4-10 shows that view 1 and view 2 have a similar circadian effective percentage from in March, April, August, and September. View 1 shows a huge improvement on circadian effective percentage from October to February while lower circadian effective percentage from May to July compared to view 2. However, the covered area on figure 4-10 indicates that view 1 has a larger circadian effective percentage among the whole year compared to view 2. In this case, view 1 is an overall best view direction among four options. 51 Figure 4-10 Comparison of circadian effective percentage based on view directions 4.3.2. Results of circadian lighting effects shift in both space and time As was mentioned in previous sections, circadian lighting effects change with sun position change. With the astronomical clock, sun position changes with hour, day and month. By analyzing relative sun position, it shows that different days in a month do not have such a huge impact on relative sun positions compared with different hours in a day and different month in a year. So, this section focuses on hourly and monthly circadian lighting effect shift in space. For the purpose of saving time, this section takes three days in one year for monthly shift study. The three typical days are March 21 st , June 21 st and December 21 st . The value of EML indicates the level of circadian lighting effects. This research takes the threshold from WELL building standard which is circadian illuminance above 200 EML is the effective circadian lighting. Otherwise, it is non-effective circadian lighting. As discussed in the previous section, view 1 was chosen as circadian lighting design view direction. Figure 4-11 shows circadian illuminance level on typical days. During the circadian lighting function period which is 9:00 to 13:00 every day, both March 21 st and December 21 st have their EML increasing with time passing by through hours. However, EML in Jun 21 st decreases when the time passing by. This indicates that during March and December, early morning is the worst-case scenario through the day. However, in June, early afternoon is the worst-case scenario during each day. From south to north, the space closer to glazing has a better circadian lighting effect because of the amount of daylight that the eyes received. From east to west, circadian lighting strong spot concentrates on the space slightly east to glazing system. Overall, the south-east corner has a better circadian lighting performance than south-west corner even though glazing system was evenly distributed on south façade. 52 Figure 4-11 Summary of EML results of typical days 4.4. Summary This chapter presented the results from three parts of the workflow. The first part is from the iterative analysis. The iterative analysis is a time-saving way to test more solutions for the decision maker to refer to. Designers can also benefit from interior finish analysis to test different options. The detailed circadian lighting evaluation helps designers to understand the importance of view direction and how circadian lighting effects shift in both space and time. 53 5. CHAPTER 5 DISCUSSION The results presented in the previous chapter proved that the proposed workflow can be applied to the decision- making stage. If designers want to apply circadian lighting effects in architectural design, it is critical to understand how it performs and how it changes through space and time. This chapter aims at answering the research question which is the possibility to bring circadian lighting effects into the architectural daylighting design. To clarify the answer, this research took Flex Lab as a prototype to analysis circadian lighting effects in a commercial building with simulation-based data. A similar workflow can be applied to another building type, but the threshold of performance needs to be updated based on standards. Parametric design components, including window wall ratio, ceiling height, four glazing materials and building orientation are input into a multi-objective optimization analysis. With the genetic algorithm, the optimization tool named “Octopus” ran the iterative analysis and output a large range of solutions. This chapter described case solutions and analyzed building performance based on different scenarios. Both worst case scenarios and overall best-case scenarios are compared to show how much impact that building geometry and affect circadian lighting effects. Then, with informed architectural design, this chapter analyzed hourly-based EML results to present a detailed circadian lighting evaluation. Both interior designer and lighting designer can benefit from the evaluation to be a better decision maker. 4 objectives in this chapter are listed below: • To prove the proposed workflow can be applied to early daylighting design stage and overall better solutions have an improved building performance comparing to poor solutions. • To evaluate the circadian lighting effects based on view directions and provide interior designer a guide for workstation layout. • To demonstrate interior finish has an impact on circadian lighting effects and provide an approach for an interior designer to pick interior finish based on simulation results. • To find “biological darkness” spots from only daylighting and contribute to artificial lighting fixture layout. 5.1. The significance of geometry design and GA based iterative analysis Building geometry has both visual effects and functional effects. In this research, functional effects include energy consumption, circadian lighting effects, visual lighting effects All the effects are changing with geometry changes simultaneously. The results in chapter 4 showed the tradeoff between design solutions. Comparison between 8 picked solutions showed the improvement on three objectives, especially the circadian lighting effects increment up to 81.08% than the worst design solution. Also, it is nearly impossible for any designer to figure out the relationship between all effects and how they affect each other. It is a “black box” with very complex relationships. A practical way is using a GA-based iterative analysis of an overall best solution. The iterative analysis with GA can go through solutions within a short time and help a decision maker making choices in the early design stage. By comparing overall best-case scenario and single- objective worst-case scenarios in chapter 4, the iterative analysis shows its potential for decision making. More parameters and objectives can be added to this workflow and GA based iterative analysis is capable of showing how each solution performs and which options are overall better solutions. Without GA iterative analysis, for one thing, it takes endless time for designers to calculate all performance indicators. For another thing, it is hard for a designer to present the results with multiple performance indicators. One research measured data for circadian lighting quality in 13 spaces over 13 weeks to evaluate circadian stimulus in dementia facility (Konis 2018). This is a trustworthy evaluation, but it took too much time and labor. Also, with circadian lighting calculation tool development, EML results from multi-channel approach can provide a reliable result for circadian lighting evaluation and design in architecture. 54 5.2. The significance of interior finish color Most of the light that eyes receive is reflected by interior surfaces, so interior finish color is critical for circadian lighting quality. By measuring data with a moving sensor, a research study proved that interior finish with a lighter color has a notable impact on circadian lighting quality (Andersen, Gochenour, and Lockley 2013). By comparing three interior finishes, it proved that the more blue range of light the surface can reflect, the higher circadian lighting quality can be achieved. The more amount of light received by eyes are reflected by the interior surface, the more it can benefit from the interior finish. 5.3. The significance of view direction selection For visual lighting quality in a space, the seating direction does not matter a lot because occupants are looking toward workstation surface. However, circadian lighting quality varies a lot from view direction because the lighting effects occupants’ circadian system is received on the vertical plane. The results in section 4.3.1 showed that the whole space is almost “biologically dark” from view 3 but it is almost a circadian effective area during a year from view 4. This result draws the attention because the “wrong” view direction can put occupants’ circadian system in danger. No matter which direction the occupant is facing, they receive the same level of visual light, which they can directly feel it. At the same time, they might receive completely effective circadian lighting or completely non- effective circadian lighting that they do not even notice. 5.4. An approach to artificial lighting design for circadian effects This research took Flexlab as a prototype, which has a simple rectangle floorplan. The proposed method applied a grid representing the location of the possible lighting fixtures. In this case, a 1.2 m by 1.2 m grid was applied in this 6.4 m by 9.4 m rectangle layout. The whole space is divided into 35 portions. Each portion has a potential lighting fixture hung above providing extra circadian lighting. The multi-channel approach was used for output EML results for each portion. Section 4.3.1 claimed that view 1 is the overall best view direction based on circadian lighting. So, this section assumed all occupants are seating with view 1. The EML results of view 1 were divided into three levels which are “biologically bright”, “biologically grey” and “biologically dark”. A heat map is generated based on each portion’s category. Figure 5-1 showed categorized portions on March 21 st at 9:00. This portion map provides both fixture layout information and lighting control suggestions for the lighting designer. Any portion marked with white means they are “biologically bright”. Daylight already provided that area enough circadian lighting stimulus, so no extra lighting fixture is necessary. Both portions marked with black and grey mean extra circadian lighting stimulus is necessary. So, these portions indicate where should lighting fixture get installed. The difference between a biological dark portion and a biological grey portion lies in the required amount of light. This difference will be filled up by pre- programmed lighting control system. With a dimmer for each lighting fixture, lighting control system has the access 55 to individually control each fixture to lower the output. For lighting fixtures in black portions, control system set them to full output. For grey portions, control system sets the lighting fixtures to half output. Figure 5-1 Summary of circadian lighting stimulus level of each portion, March 21st, 9:00 With same weather and sky condition, the shift of circadian lighting comes along with relative sun position changes. Instead of defining all scenarios, five typical days with five hour-spots was chosen as circadian lighting design templates. Figure 5-2 shows artificial lighting layout and operation template for all days in March and September at 9:00. Figure 5-2 Lighting fixture layout and operation template for March and September at 9:00 Figure 5-3 presents all 25 templates of circadian lighting design for Flexlab. It shows that there are 4 portions out 35 don’t need extra circadian lighting as a supplement. For the remaining portion, part of the fixtures doesn’t need to be 100% on all the time. Without circadian evaluation, there is an alternative way, which is installing a lighting fixture in every portion and turning them on all the time. Table 5-1 compares two design strategies based on required fixture numbers and operating schedule. Instead of doing poor artificial circadian lighting design, which turns all lighting fixtures on all the time, advanced artificial lighting design breaks it down into three output modes. Figure 5-4 showed the percentage of different fixture output category within a year. Assume 50% output saved half energy comparing to 100% output, advanced circadian lighting design saved 88.36% regardless of 4 lighting installation costs. 56 Figure 5-3 Template of artificial lighting for circadian lighting design Design strategy Number of lighting fixtures needed Percentage of hours requires 100% output (assume with 35 fixtures) Percentage of hours requires 50% output (assume with 35 fixtures) Percentage of hours requires 0% output (assume with 35 fixtures) Poor circadian lighting design 35 100% 0% 0% Advanced circadian lighting design 31 45.53% 31.20% 23.27% Table 5-1 Comparison of poor circadian lighting design and advanced circadian lighting design 57 Figure 5-4 Comparison of lighting fixture output in the advanced circadian lighting design 5.5. Summary This chapter proved that the proposed workflow is applicable for finding design solutions which cover all three objectives. The EML results showed the difference in geometry design between different case scenarios and comparison based on three design objectives. artificial lighting design for circadian lighting effects varies from building parameters. By analyzing circadian effects that vary based on view directions, interior finishes, also circadian lighting effects shift in both time and space, this research proposed a new method for artificial lighting design based on circadian effects. This method took 25 typical scenarios based on sun position, representing five time-spots of each day in a year. The new method turned out to be promising based on the energy-saving comparing to poor circadian lighting design. 58 6. CHAPTER 6 CONCLUSION This research proposed a new workflow which brings circadian lighting effects into daylight design during the early stages of design. The topic, circadian lighting, was raised up and widely discussed especially when the 2017 Nobel Prize in Physiology or Medicine was awarded to Jeffrey C. Hall, Michael Rosbash and Michael W. Yong for their discovery of molecular mechanisms controlling the circadian rhythm. This is a chance, as well as a new challenge for an architect to think about how the circadian lighting effect can be an influencing factor in architecture and how to integrate it in architectural design. This research took a lab named flexlab at LBNL as a prototype to prove the feasibility of a new workflow. The workflow is qualified for WELL building standard based on section 54. The research had two main objectives: 1. To provide a parametric design method and an optimization process based on GA for architects finding a better daylighting design solution based on circadian lighting effects and another performance indicator. 2. To develop an interior lighting control system aiming at using artificial lighting as supplements for circadian lighting effects. The research question is the possibility of achieving a circadian lighting effects qualified building with less energy from electrical lighting fixtures. This question was broken down into two parts, which are two objectives listed above. The first objective saved energy through the passive design strategy which is a better building parameter solution. The second objective saved energy by minimizing fixture numbers and dimming lighting fixtures, but it can still meet the requirement of WELL building standard. Both parts came with notable outcomes. The first outcome is a list of design solutions with building performance indicators. The parameters including building orientation, ceiling height, window wall ratio and glazing materials were run in a GA based multi-objective iterative analysis. Each solution came with three objectives which are visual lighting effects, circadian lighting effects, and energy consumption. With the help of GA, this iterative analysis can run wide-spread solutions seeking for non-dominated objectives solutions which are a time-saving way to find an overall best solution. By comparing different case scenarios, an overall improvement on three objectives is a hint that this approach worked well in the early design stage. Also, the simulation-based approach relies on rhinoceros and grasshopper, which are two major software applied in an architecture firm. Without any experience with GA, the user can still run the iterative analysis with pre-defined evolutionary algorithms in “Octopus”. Provided with a user-friendly interface from Octopus, users can customize evolutionary algorithms, set time duration of iterative analysis, check iterative analysis history, mark solutions and so on which benefit them in the decision making stage. For example, in the very early design stage, the model can be a simple parametric design about building geometry. The maximum generation number can be low for the purpose of saving time. At this stage, solutions can keep the design on the right track. Building geometry, interior designs, occupants’ positions and so on are all factors for both visual lighting effects and circadian lighting effects. Another iterative analysis of complex parametric designs with evaluation points in occupants’ positions can help architects understand the circadian lighting effects more precisely. A large population size can be applied in this stage seeking for more choices which brings a better solution. With an informed geometry, designers can also take advantage of this workflow to make a decision about occupants’ view direction by arranging the layout of workstations and interior color selection. The second outcome is a strategy of circadian lighting design with artificial lighting as supplements. To provide effective circadian stimulus, poor artificial circadian lighting design turns the lighting fixtures on during 9:00 to 13:00 every day. This is a practical way, but not energy efficient because it ignored the help of daylight. To integrate with daylight, a design strategy based on studying circadian lighting effect shift in both time and space can save energy at least 88.36% comparing to poor circadian lighting design. This example divided circadian lighting stimulus into three levels. Level 1, the space described as dark portion has circadian stimulus from 0 EML to 100 EML. In the design stage, this portion is treated as 0 EML and provided artificial lighting with 200 EML output. The space with 100 EML to 200 EML from daylight is defined as a grey portion. Artificial lighting provided half output for this part. The remaining space, with over 200 EML from daylight, is defined as a white portion which doesn’t require extra artificial lighting. This method can be customized by designers. For example, designers can divide the 59 entire space into 2 portions like the category that table 6-1 shown below. In this case, the portion was defined more precisely, which means it saves energy more than 88.36%. Melanopic lux (EML) Portion code Lighting fixture output percentage (%) [0, 50) 1 100 [50, 100) 2 75 [100, 150) 3 50 [150, 200) 4 25 [200, +∞) 5 0 Table 6-1 Category of portions based on circadian lighting stimulus Instead of providing a guideline for each day within a year, this method selected 5 typical days with 5 hour-spots of each day based on relative sun positions. This representative look-up table saved designers time from programming lighting control system by decreasing number of scenarios from 1825 to 25. This chapter proved the feasibility and promising impact of the proposed daylight design workflow. With a GA based optimization tool, it is possible to compare a certain number of solutions and pick the one with balanced performances. Also, the results from multi-channel circadian lighting calculation approach showed the potential energy saving when integrating artificial lighting controls with daylight conditions. 60 7. CHAPTER 7 LIMITATIONS AND FUTURE WORK This research brought a new lighting concept into an existing daylighting design workflow. A new performance requirement means more opportunities, as well as more challenges. Some work, like updating building standards and developing digital design tools will be the topic for other parties. However, as an architect, circadian lighting concepts can be expressed clearer and more precisely. To achieve that, three suggestions are listed below. 7.1. Future work for this research This optimization took four parameters as variables for an overall better solution. However, a real project is likely to be much more complicated than this prototype. Other parameters like complex building form and adjacent buildings are factors which bring effects on circadian lighting and other building performance. Also, interior design should not be limited to interior finish color and view direction of the workstation. Some strategy, like workstation layout, can be applied in interior design as well. For example, the portion with high melanopic lux should have a high workstation intensity while the portion with low melanopic lux should be designed as a break space or a kitchen. Distinct from lighting from daylight and lighting fixtures, lighting from monitors is another factor which should be considered in optimization workflow. The way people work now is substantially different than the past. People worked on papers and spent most of their time looking down on a horizontal work plane in the past. Nowadays, people work with digital devices like desktop computers and projectors, some of them even spend the whole working time on monitors. In this case, the monitor not only provides circadian lighting but also blocks part of the circadian lighting from daylight. A circadian lighting optimization based on that specific group of people should draw designers’ attention. A melanopic lux calculation method like a multi-channel approach should be applied to optimization processes with annual data. This research took a “melanopic ratio + photopic lux” approach to obtain annual melanopic lux. This is an easy way but not accurate enough. There is no well-developed tool to use a multi-channel approach for annual melanopic lux calculation, right now. However, a multi-channel approach or a more accurate calculation method will take over “melanopic ratio + photopic lux” approach in this workflow. A pre-decision about view direction should be made before an iterative analysis. This research took the average of CEP of four view directions as the value representing circadian lighting level. This is a problematic setting since the workstation view is not changing with time. For example, if view 1 was chosen for layout of the workstation, then it means an average of four view directions has a high value. Assume both solution 1 and solution 2 have the same CEP for view 1. However, solution 2 has a high CEP value for other view directions. In this case, if view 1 was chosen for layout design, both solutions have the same circadian lighting level, while the solution lost the chance to evaluate that in the iterative analysis. For a better optimization process, view direction should be defined, and CEP of that view direction should serve as circadian lighting performance indicator. 7.2. Future work in circadian lighting design Except for building geometry design, other building elements can be applied to improving circadian lighting effects. A list below showed potential strategies which are also future works. • Dynamic shading systems Daylight condition changes with time. A fixed shading system cannot handle a dynamic daylight condition, even though a fixed shading system can lower building energy consumption to some extent. However, an advanced building should have a dynamic shading system to make sure daylight and heat gain effects can be controlled with each day or even each hour. • Advanced glazing system 61 The glazing system provided with R-value or transmittance can be used for energy consumption and visual lighting consumption. This research used an SPD curve to describe how much light with a certain range of wavelength can get through the glazing system. Based on that, more research should focus on a glazing system which can pass a certain range of light to maximize circadian lighting effects, while retaining the visual lighting quality (photopic range of light). • Glare issues evaluation The glare issue is the component that should be taken into consideration in daylight designs. There are researches about comparison of glare evaluation methods but none can address the glare issue for the whole year (Suk, Schiler, and Kensek 2017) (Osterhaus 2009). Further studies on glare evaluations for a time period can contribute to bringing “glare” into proposed workflow. • Advanced sensor-control system This research used a simulation-based approach to design artificial lighting as supplements. However, daylight conditions change all the time that weather data cannot describe it precisely. To handle this, an advanced sensor system should be installed indoors, monitoring the EML at eye level. Lighting control systems can react based on EML data to provide circadian lighting as a supplement. • Consideration of lighting fixture installation The lighting control system assumes artificial lights can map directly to eye-level as stimulus. However, the reality is more complex. Especially when the lighting fixture hangs above the occupant, all the light that the eyes received is reflected by the surface. In this case, even though the lighting fixtures are set to full output, there is no light from lighting fixtures that can directly get into eyes at the vertical surface. More research should be done based on the amount of light can received at eye level. 7.3. Summary The circadian lighting effect is affecting occupants and they might have no idea about that. A lot of architecture firms haven’t persuaded their clients to apply circadian lighting design strategies in their projects. As researchers and architects, they should study more about addressing the importance of circadian lighting effects and provide a cost-effective way to apply a circadian lighting design strategy. 62 8. Reference “3.2 Space Planning.” n.d. Accessed October 27, 2017. https://www.gsa.gov/node/82314. “6 Circadian Rhythm Sleep Disorders That May Be Disrupting Your Sleep.” n.d. Accessed October 15, 2017. http://www.alaskasleep.com/blog/circadian-rhythm-sleep-disorders-disrupting-your-sleep. Andersen, Marilyne, Sharon J. Gochenour, and Steven W. Lockley. 2013. “Modelling ‘Non-Visual’ Effects of Daylighting in a Residential Environment.” Building and Environment 70. Elsevier Ltd:138–49. https://doi.org/10.1016/j.buildenv.2013.08.018. “Appendix C: Tables | WELL Feature Library.” n.d. Accessed October 15, 2017. http://standard.wellcertified.com/tables#melanopicRatio. Beck, Andreas, Werner Korner, Oliver Gross, and Jochen Frick. 1999. “Making Better Use of Natural Light With A Light-Redirecting Double-Glazing System.” Solar Energy 66 (3):215–21. https://doi.org/10.1016/S0038- 092X(99)00022-5. “Berkeley Lab — Lawrence Berkeley National Laboratory.” n.d. Accessed October 27, 2017. http://www.lbl.gov/. Berson, D. M. 2002. “Phototransduction by Retinal Ganglion Cells That Set the Circadian Clock.” Science 295 (5557):1070–73. https://doi.org/10.1126/science.1067262. Chang, Anne Marie, Frank A J L Scheer, and Charles A. Czeisler. 2011. “The Human Circadian System Adapts to Prior Photic History.” Journal of Physiology 589 (5):1095–1102. https://doi.org/10.1113/jphysiol.2010.201194. Crawley, Drury B. 1998. “Which Weather Data Should You Use for Energy Simulations of Commercial Buildings?” ASHRAE Transactions 104 (2):498–515. Culver, Christine. n.d. “Portland Monthly Photoshoot.” Accessed March 4, 2018. https://www.glumac.com//content/uploads/2016/05/Working-By-Daylight-How-Circadian-Lighting-Increases- Productivity_.pdf. Curtis, Leland, and Richard Mistrick. n.d. “Parametric Daylighting Tools Parametric Daylighting Tools.” “Daysim.” n.d. Accessed October 27, 2017. http://daysim.ning.com/. DiLaura, David L., Kevin W. Houser, Richard G. Mistrick, and Gary R. Steffy. 2011. The Lighting Handbook: Reference and Application. Drake, Christopher L, Timothy Roehrs, Gary Richardson, James K Walsh, and Thomas Roth. 2004. “Shift Work Sleep Disorder: Prevalence and Consequences beyond that of Symptomatic Day Workers.” Sleep 27 (8):1453–62. https://doi.org/10.4088/JCP.10062ah1. Dubois, Marie-claude. n.d. “Impact of Shading Devices on Daylight Quality in Offices Simulations with Radiance.” Ecker, Jennifer L., Olivia N. Dumitrescu, Kwoon Y. Wong, Nazia M. Alam, Shih Kuo Chen, Tara LeGates, Jordan M. Renna, Glen T. Prusky, David M. Berson, and Samer Hattar. 2010. “Melanopsin-Expressing Retinal Ganglion- Cell Photoreceptors: Cellular Diversity and Role in Pattern Vision.” Neuron 67 (1). Elsevier Ltd:49–60. https://doi.org/10.1016/j.neuron.2010.05.023. “EIA - Annual Energy Outlook 2018.” n.d. Accessed March 28, 2018. https://www.eia.gov/outlooks/aeo/. 63 Figueiro, M.G., J.D. Bullough, M.S. Rea, and Rensselaer Polytechnic Institute. Lighting Research Center. 2007. “Light Isn’t Just for Vision Anymore: Implications for Transportation Safety.” Security 7120 (May 2014). http://204.68.195.52/dotrc/hfct/Shared Research/Light Isn?t Just for Vision Anymore - Implications for Transportation Safety (Circadian Responses to Light).pdf. Figueiro, Mariana G., Kassandra Gonzales, and David Pedler. 2016. “Designing with Circadian Stimulus.” Ld+a, no. October:31–33. http://www.lrc.rpi.edu/resources/newsroom/LD+A_CircadianStimulus_Oct2016.pdf. “FLEXLAB® | FLEXLAB.” n.d. Accessed October 27, 2017. https://flexlab.lbl.gov/. Foley, Daniel J., Andrew A. Monjan, S. Lori Brown, Eleanor M. Simonsick, Robert B. Wallace, and Dan G. Blazer. 1995. “Sleep Complaints Among Elderly Persons: An Epidemiologic Study of Three Communities.” Sleep 18 (6):425–32. https://doi.org/10.1093/sleep/18.6.425. “Frequently Asked Questions about ASHRAE IWEC2 Weather Data and White Box Technologies.” n.d. Accessed October 15, 2017. http://ashrae.whiteboxtechnologies.com/faq. Gooley, J J, J Lu, T C Chou, T E Scammell, and C B Saper. 2001. “Melanopsin in Cells of Origin of the Retinohypothalamic Tract.” Nat Neurosci 4 (12):1165. https://doi.org/10.1038/nn768. “Grasshopper - Algorithmic Modeling for Rhino.” n.d. Accessed November 1, 2017. http://www.grasshopper3d.com/. Hattar, S. 2002. “Melanopsin-Containing Retinal Ganglion Cells: Architecture, Projections, and Intrinsic Photosensitivity.” Science 295 (5557):1065–70. https://doi.org/10.1126/science.1069609. Hopkinson, R. G. 1972. “Glare from Daylighting in Buildings.” Applied Ergonomics 3 (4):206–15. https://doi.org/10.1016/0003-6870(72)90102-0. “IGDB | Windows and Daylighting.” n.d. Accessed February 25, 2018. https://windows.lbl.gov/software/igdb. Illuminating Engineering society. 2013. IES Spatial Daylight Autonomy (sDA)and Annual Sunlight Exposure (ASE). IES. https://infoscience.epfl.ch/record/196436. Inanici, Mehlika, Martin Brennan, and Edward Clark. 2008. “SPECTRAL DAYLIGHTING SIMULATIONS : COMPUTING CIRCADIAN LIGHT University of Washington , Department of Architecture , Seattle , WA , USA ZGF Architects , Seattle , WA , USA.” International WELL building institute. 2015. “The WELL Building Standard, V 1.0.” Jakubiec, JA, and CF Reinhart. 2012. “The ‘adaptive Zone’ – A Concept for Assessing Discomfort Glare throughout Daylit Spaces.” Lighting Research & Technology 44 (2):149–70. https://doi.org/10.1177/1477153511420097. Jens Christoffersen. 2005. “No Title.” In Towards a New Daylight Glare Rating, 157–61. Berlin: Ludwig Erhard Haus Industrie- und Handelskammer zu Berlin. Kantermann, Thomas, Myriam Juda, Martha Merrow, and Till Roenneberg. 2007. “The Human Circadian Clock’s Seasonal Adjustment Is Disrupted by Daylight Saving Time.” Current Biology 17 (22):1996–2000. https://doi.org/10.1016/j.cub.2007.10.025. Khachiyants, Nina, David Trinkle, Sang Joon Son, and Kye Y. Kim. 2011. “Sundown Syndrome in Persons with Dementia: An Update.” Psychiatry Investigation 8 (4):275–87. https://doi.org/10.4306/pi.2011.8.4.275. 64 Kim, Gon, Hong Soo Lim, Tae Sub Lim, Laura Schaefer, and Jeong Tai Kim. 2012. “Comparative Advantage of an Exterior Shading Device in Thermal Performance for Residential Buildings.” Energy and Buildings 46:105–11. https://doi.org/10.1016/j.enbuild.2011.10.040. Konak, Abdullah, David W. Coit, and Alice E. Smith. 2006. “Multi-Objective Optimization Using Genetic Algorithms: A Tutorial.” Reliability Engineering and System Safety 91 (9):992–1007. https://doi.org/10.1016/j.ress.2005.11.018. Konis, Kyle. 2013. “Evaluating Daylighting Effectiveness and Occupant Visual Comfort in a Side-Lit Open-Plan Office Building in San Francisco, California.” Building and Environment 59. Elsevier Ltd:662–77. https://doi.org/10.1016/j.buildenv.2012.09.017. ———. 2017. “A Novel Circadian Daylight Metric for Building Design and Evaluation.” Building and Environment 113. Elsevier Ltd:22–38. https://doi.org/10.1016/j.buildenv.2016.11.025. ———. 2018. “Field Evaluation of the Circadian Stimulus Potential of Daylit and Non-Daylit Spaces in Dementia Care Facilities.” Building and Environment 135 (March):112–23. https://doi.org/10.1016/j.buildenv.2018.03.007. Konis, Kyle, Alejandro Gamas, and Karen Kensek. 2016. “Passive Performance and Building Form: An Optimization Framework for Early-Stage Design Support.” Solar Energy 125 (February):161–79. https://doi.org/10.1016/j.solener.2015.12.020. “Ladybug Tools - Grasshopper.” n.d. Accessed November 1, 2017. http://www.grasshopper3d.com/group/ladybug. Lange, Tanja, Stoyan Dimitrov, and Jan Born. 2010. “Effects of Sleep and Circadian Rhythm on the Human Immune System.” Annals of the New York Academy of Sciences 1193 (1):48–59. https://doi.org/10.1111/j.1749- 6632.2009.05300.x. “Lark Spectral Lighting.” n.d. Accessed February 11, 2018. http://faculty.washington.edu/inanici/Lark/Lark_home_page.html. Lewy;, Robert L. Sack; Richard W. Brandes; Adam R. Kendall; Alfred J. 2000. “Entrainment of Free-Running Circadian Rhythms by Melatonin in Blind People.” http://www.nejm.org/doi/pdf/10.1056/NEJM200010123431503. “Light and Health | Research Programs | LRC.” 2017. 2017. http://www.lrc.rpi.edu/programs/lightHealth/. “Lighting Design Software by Lighting Analysts.” n.d. Accessed January 12, 2018. https://www.agi32.com/. “Lighting Research Center | Education | Learning | Terminology | Spectral Power Distribution.” n.d. Accessed October 14, 2017. http://www.lrc.rpi.edu/education/learning/terminology/spectralpowerdistribution.asp. Lucas, Robert J., Stuart N. Peirson, David M. Berson, Timothy M. Brown, Howard M. Cooper, Charles A. Czeisler, Mariana G. Figueiro, et al. 2014. “Measuring and Using Light in the Melanopsin Age.” Trends in Neurosciences 37 (1). Elsevier Ltd:1–9. https://doi.org/10.1016/j.tins.2013.10.004. “Luminance.” n.d. Lighting Design Glossary. Accessed January 18, 2018. http://www.schorsch.com/kbase/glossary/luminance.html. Maltais, Louis-Gabriel, and Louis Gosselin. 2017. “Daylighting ‘energy and Comfort’ Performance in Office Buildings: Sensitivity Analysis, Metamodel and Pareto Front.” Journal of Building Engineering 14 (September). Elsevier Ltd:61–72. https://doi.org/10.1016/j.jobe.2017.09.012. 65 Maurice M. Ohayon, Mary A. Carskadon, Christian Guilleminault, and Michael V. Vitiello. 2004. “Meta-Analysis of Quantitative Sleep Parameters From Childhood to Old Age in Healthy Individuals: Developing Normative Sleep Values Across the Human Lifespan.” Sleep 27 (7):1255–73. https://doi.org/10.1080/16506073.2015.1026386. McCluney, Ross. 1994. Introduction to Radiometry and Photometry. Artech House. “Measuring and Calculating Lux Values, Part 2.” n.d. Accessed October 16, 2017. https://www.allaboutcircuits.com/technical-articles/measuring-and-calculating-lux-values-part-2/. Miguel, A. De, and J. Bilbao. 2005. “Test Reference Year Generation from Meteorological and Simulated Solar Radiation Data.” Solar Energy 78 (6):695–703. https://doi.org/10.1016/j.solener.2004.09.015. Mistlberger, Ralph E. 2005. “Circadian Regulation of Sleep in Mammals: Role of the Suprachiasmatic Nucleus.” Brain Research Reviews 49 (3):429–54. https://doi.org/10.1016/j.brainresrev.2005.01.005. Morrissey, J., T. Moore, and R. E. Horne. 2011. “Affordable Passive Solar Design in a Temperate Climate: An Experiment in Residential Building Orientation.” Renewable Energy 36 (2). Elsevier Ltd:568–77. https://doi.org/10.1016/j.renene.2010.08.013. “Optics | Windows and Daylighting.” n.d. Accessed February 25, 2018. https://windows.lbl.gov/software/optics. Osterhaus, Werner. 2009. “Design Guidelines for Glare-Free Daylit Work Environments.” Proceedings of the 11th European Conference on Lighting Lux Europa 2009. http://scholar.google.com/scholar?hl=en&btnG=Search&q=intitle:Design+Guidelines+for+Glare- free+Daylit+Work+Environments#0. Panda, S. 2005. “Illumination of the Melanopsin Signaling Pathway.” Science 307 (5709):600–604. https://doi.org/10.1126/science.1105121. Parakou, B E N, B F A Dori, K E N Garissa, K E N Kisumu, K E N Kitale, K E N Lodwar, K E N Makindu, et al. n.d. “Appendix A . Complete Listing of the 3012 IWEC2 Weather Files” 1 (29):70–88. Pierson, Clotilde, Jan Wienold, and Magali Bodart. 2017. “Discomfort Glare Perception in Daylighting: Influencing Factors.” Energy Procedia 122 (Supplement C). Elsevier B.V.:331–36. https://doi.org/https://doi.org/10.1016/j.egypro.2017.07.332. “Radiance — Radsite.” n.d. Accessed October 27, 2017. https://www.radiance-online.org/. Rea, Mark S, Mariana G Figueiro, Andrew Bierman, and John D Bullough. 2010. “Circadian Light.” Journal of Circadian Rhythms 8 (1):2. https://doi.org/10.1186/1740-3391-8-2. “Rhinoceros.” n.d. Accessed November 1, 2017. https://www.rhino3d.com/. Roenneberg, Till, and Martha Merrow. 2016. “The Circadian Clock and Human Health.” Current Biology 26 (10). Elsevier Ltd:R432–43. https://doi.org/10.1016/j.cub.2016.04.011. Sack, Robert L, Dennis Auckley, R Robert Auger, Mary a Carskadon, Kenneth P Wright, Michael V Vitiello, and Irina V Zhdanova. 2007. “Circadian Rhythm Sleep Disorders: Part I, Basic Principles, Shift Work and Jet Lag Disorders. An American Academy of Sleep Medicine Review.” Sleep 30 (11):1460–83. https://doi.org/10.1136/bmj.e6664. Savvidis, Christos, and Michael Koutsilieris. 2012. “<Title/>.” Molecular Medicine 18 (9):1. https://doi.org/10.2119/molmed.2012.00077. 66 Suk, Jae Yong, Marc Schiler, and Karen Kensek. 2017. “Investigation of Existing Discomfort Glare Indices Using Human Subject Study Data.” Building and Environment 113. Elsevier Ltd:121–30. https://doi.org/10.1016/j.buildenv.2016.09.018. “The Human Suprachiasmatic Nucleus | HHMI BioInteractive.” n.d. Accessed October 27, 2017. http://www.hhmi.org/biointeractive/human-suprachiasmatic-nucleus. Tresguerres, J a, C Ariznavarreta, B Granados, M Martín, M a Villanúa, D a Golombek, and D P Cardinali. 2001. “Circadian Urinary 6-Sulphatoxymelatonin, Cortisol Excretion and Locomotor Activity in Airline Pilots during Transmeridian Flights.” Journal of Pineal Research 31 (1):16–22. https://doi.org/10.1034/j.1600- 079X.2001.310103.x. USGBC. 2016. “LEED v 4 for NEIGHBORHOOD DEVELOPMENT,” 161. http://www.usgbc.org/sites/default/files/LEED v4 BDC_04.05.16_current.pdf. Vitaterna, Martha Hotz, Joseph S Takahashi, and Fred W Turek. 2001. “Overview of Circadian Rhythms.” Alcohol Research & Health : The Journal of the National Institute on Alcohol Abuse and Alcoholism 25 (2):85–93. Volicer, Ladislav, David G. Harper, Barbara C. Manning, Rachel Goldstein, and Andrew Satlin. 2001. “Sundowning and Circadian Rhythms in Alzheimer’s Disease.” American Journal of Psychiatry 158 (5):704–11. https://doi.org/10.1176/appi.ajp.158.5.704. “What Is Correlated Color Temperature? | Light Sources and Color | Lighting Answers | NLPIP.” n.d. Accessed October 27, 2017. http://www.lrc.rpi.edu/programs/nlpip/lightinganswers/lightsources/whatisCCT.asp. “What Is Glare? | Light Pollution | Lighting Answers | NLPIP.” n.d. Accessed October 29, 2017. http://www.lrc.rpi.edu/programs/nlpip/lightinganswers/lightpollution/glare.asp. Wienold, Jan, and Jens Christoffersen. 2006. “Evaluation Methods and Development of a New Glare Prediction Model for Daylight Environments with the Use of CCD Cameras.” Energy and Buildings 38 (7):743–57. https://doi.org/10.1016/j.enbuild.2006.03.017. “WINDOW | Windows and Daylighting.” n.d. Accessed February 25, 2018. https://windows.lbl.gov/software/window. Wong, Ing Liang. 2017. “A Review of Daylighting Design and Implementation in Buildings.” Renewable and Sustainable Energy Reviews 74 (March). Elsevier Ltd:959–68. https://doi.org/10.1016/j.rser.2017.03.061.

Abstract (if available)

Abstract This research proposed a new daylighting design workflow for architects to develop parametric designs for commercial buildings based on multiple building performance indicators. Conventional daylighting design mainly focuses on building performance like visual lighting quality and energy consumption. However, with a deeper understanding of daylight and its effect on human beings’ wellness, more factors need to be considered, such as its effect on people’s “circadian rhythms”. Several circadian lighting evaluation methods have been introduced but not integrated into a daylighting design workflow. The proposed workflow adapted parametric design including window wall ratio, ceiling height, building orientation and glazing type. The parametric design aimed at seeking for a better design solution based on three building performance indicators which are visual lighting quality, circadian lighting quality, and energy consumption. In this research, visual lighting quality took spatial Daylight Autonomy (sDA) as the metric while energy consumption took Energy Use Intensity (EUI) as the metric. Similar to sDA, this research proposed a new metric named Circadian Effective Percentage (CEP) and adapted it as the metric for circadian lighting quality. Also, the parametric design has another parameter which is interior finish. Different than other parameters, interior finish color only affects circadian lighting quality. Besides, this workflow ended with a detailed circadian lighting evaluation of daylighting. The evaluation is aimed at presenting circadian lighting effect shifts in both time and space so that lighting designers can layout lighting fixtures and use pre-programed lighting control systems as supplements. The application of this workflow is qualified for Title 24 (Section 140.6: General Illumination Level) and WELL building standard (Section 54: Circadian Lighting Design). Architects can adapt the proposed workflow for a design solution that maximizes daylight as a resource for circadian stimulus and entrainment. A case study took FLEX LAB in Lawrence Berkeley National Laboratory as a prototype to demonstrate the viability of the proposed workflow.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Daylight and health: exploring the relationship between established daylighting metrics for green building compliance and new metrics for human health

PDF

Integrating non-visual effects of lighting in the evaluation of electrical lighting designs for commercial offices

PDF

The intelligent control strategy of kinetic façades for daylight and energy performance: evaluating the daylight effect of adaptive systems based on parametric workflow

PDF

Development of circadian-effective toplighting strategies: using multiple daylighting performance goals for dementia care communities

PDF

Simulation-based electric lighting control algorithm: integrating daylight simulation with daylight-linked lighting control

PDF

Effective light shelf and form finding: development of a light shelf design assistant tool using parametric methods

PDF

Performative shading design: parametric based measurement of shading system configuration effectiveness and trends

PDF

Exploring the influence of scheduling on daylighting performance evaluation in assisted living: testing two new metrics based on sDA and ASE

PDF

Environmentally responsive buildings: multi-objective optimization workflow for daylight and thermal quality

PDF

Double skin façades performance: effects on daylight and visual comfort in office spaces

PDF

Comparative evaluation of lighting design software programs for daylighting in buildings

PDF

ctrl+z: exploring the effects of undoing retrofits to pre-war buildings in Los Angeles

PDF

Designing‐in performance: energy simulation feedback for early stage design decision making

PDF

Evaluation and development of solar control screens: using daylight simulation to improve the performance of facade solar control screens

PDF

Daylighting in Riyadh and Los Angeles: comparison of cultural factors in potential market penetration

PDF

Visualizing architectural lighting: creating and reviewing workflows based on virtual reality platforms

PDF

Daylighting study of a LEED platinum laboratory building: a post-occupancy evaluation comparing performance in use to design intent

PDF

Impact of occupants in building performance: extracting information from building data

PDF

Multi-domain assessment of a kinetic facade: determining the control strategy of a kinetic façade using BIM based on energy performance, daylighting, and occupants’ preferences; Multi-domain asse...

PDF

A parametric study of the thermal performance of green roofs in different climates through energy modeling

Asset Metadata

Creator Qiu, Geli (author)

Core Title Evaluation of daylighting circadian effects: Integrating non-visual effects of lighting in the evaluation of daylighting designs

School School of Architecture

Degree Master of Building Science

Degree Program Building Science

Publication Date 07/02/2018

Defense Date 06/28/2018

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag circadian lighting,daylighting,genetic algorithm,OAI-PMH Harvest,optimization,parametric design

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Konis, Kyle (committee chair), Regnier, Victor (committee member), Schiler, Marc (committee member)

Creator Email gqiu@usc.edu,gqiu0513@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c40-509062

Unique identifier UC11267408

Identifier etd-QiuGeli-6361.pdf (filename),usctheses-c40-509062 (legacy record id)

Legacy Identifier etd-QiuGeli-6361.pdf

Dmrecord 509062

Document Type Thesis

Format application/pdf (imt)

Rights Qiu, Geli

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA