Submitted for publication to Building Simulation 2011 in Sydney (November 2011)
A METHODOLOGY FOR DESIGNING CONTEMPORARY HIGH PERFORMANCE SHADING SCREEN-THE INTEGRATION OF ‘FORM’ AND THE DIVA SIMULATION TOOL Azadeh Omidfar Harvard University Graduate School of Design
[email protected] ABSTRACT This project is an attempt to integrate and evaluate the formal and ornamental desires of contemporary architecture with the pressing need to create designs that optimize energy and daylight performance. With increased sophistication of digital tools to assess daylight and energy in buildings, a great potential exists to optimize the performance of contemporary building façades. While pre-modern and modernist architecture have often employed light shelves, overhangs, and other passive shading techniques, such forms are not easily applied to highly articulated forms of contemporary architecture. This research study proposes a process for applying daylighting and energy analysis software to optimize the performance of a sunshading screen based on Sculptor Erwin Hauer's design. Through an iterative analysis of the south facing screen, daylight and energy performance of an interior test space is optimized through parametrically manipulating the opening size and module depth of the screen. Results are then compared to baseline spaces with 30% and 100% Window-to-Wall ratios (WWR) to understand the screen's relative performance gains. This analysis shows that the optimized screen manages to reduce annual energy use by 35 percent and 42 percent visà-vis the two baseline cases.
INTRODUCTION The progression of style from traditional to modern architecture and now contemporary architecture has resulted in a deficiency of incorporating sustainable techniques in building design. Consequently, according to the U.S Energy Information Administration (EIA), lighting, cooling and heating are now responsible for the greatest amounts of energy consumption in the building sector; followed by water heating, ventilation, etc. (EIA, 2003). As architecture continues to reinvent itself in the way form is created, the application of a building’s skin has shifted from pre-Modern notions of applied ornamentation, to the deliberate removal of ornamentation in Modernism (~1920s), and has reemerged in Postmodernism (~1960) and present day architecture. The birth of the digital era has brought about tools that allow for parametric design of both organically- and geometrically-derived building skins. These skins however, largely tend to be an
aesthetic expression of form (or ornament), and rarely address environmental conditions such as effective radiation control, daylighting or thermal comfort. To that end, the ever present climatic crisis coupled with increased environmental regulations has created the necessity for increased integration of formal design with sustainable practices. As a result, the building skin, as both the building’s primary protection from the elements and as the Architect's statement, must be addressed and better implemented into design to address increasingly complex sets of needs. The building skin must perform the role of an environmental filter between interior and exterior conditions, while addressing and resolving a wide spectrum of issues such as technical performance, visual appearance, ventilation, assembly, etc. (Lovell, 2010). However, controlling the sun’s radiation and providing adequate daylight levels and views are a few of many important tasks that the skin must also perform. These issues, among others, are the focal point of this research study. A survey of 177 design practitioners in 2007 indicated that the designers defined daylighting as “the interplay of natural light and building form to provide a visually stimulating, healthful and productive interior environment” while, engineers defined daylighting as “the use of fenestration systems and responsive electric lighting controls to reduce overall building energy requirements (heating, cooling, lighting)” (Galasiu & Reinhart, 2007). This study seeks to show that providing adequate daylight, creating a visually stimulating and healthful interior environment, as well as reducing overall energy consumption of buildings can be achieved by incorporating a well-designed and performative building skin. With new complex building forms gaining popularity, semi-standardized shading devices such as louvers or overhangs are often disregarded due to the difficulty of integration of these forms with highly articulated contemporary building design. This poses the question: Is it possible to design an ornamental building skin that is simultaneously a performative shading device that also provides acceptable views to the outside?
METHODOLOGY To integrate all aspects of architectural design, a comprehensive methodology is required to consider
Submittedd for publicatiion to Buildinng Simulation 2011 in Sydnney (Novembeer 2011)
how a buuilding looks and works; hoow it sustainss the physical comfort of itss occupants; how h it fits wiithin its contexxt and climatee; and what thhe design sign nifies to its ooccupants an nd communitty. Due to the hypothetiical environm ment of this ressearch study, only local clim mate and the formal f languagge of design have h been evalluated in detaiil. Case Stu udy: Intercirclles To underrstand the lim mitations andd possibilities of transform ming an ornam mental design into a functiional one, the Intercircles project by Erwin E Hauer was chosen aas a case studdy (Figure 1). Erwin Haueer is best knnown for his multivaalent geometric architectuural screens, developed beetween 1950 and 1959, innspired by the t concept of infinity and continuouus surfaces. This T study soought to anaalyze this parrticular screeen through a process of experimeentation and parametric modulation. The typology was developped into noveel mutations that ginal maintaineed the aesthetiics and languaage of the orig prototypee while prroposing neew performaative functionss. Intercircles allowed for the t opportunitty to explore tthe possibilities of modullating the sccreen walls to optimize o poro osity to controol the quality and quantity of daylight transmissionn. Through this o units were developed whhose exploratioon, a system of geometricc distortions produce varrying degreess of porosity, and thereforee daylight transmission.
Figgure 2: Rotatioon of the smalll circles in thee Z-axis
Figure 3: Constructing C t unit geometry the
Figgure 4: Creatiing a larger screen by mirrroring 1/8 off the surface geometry g Figuree 1: Intercirclles designed byy Erwin Haueer Creation n of Units: By B closely exxamining Hau uer’s complex screen, its pattern was reconstruucted geometriccally through the manipulation of diffeerent sized circcles. The moodule consistss of four smaller circles rootated along the Z-axis, which w are plaaced around thhe perimeter of a larger circle c (Figuree 2). The unit module was created by coonstructing 1//8 of the surfaace and then mirroring annd reflecting it 8 times aroound its edgess (Figure 3). E Each module was mirrored along its ed dge to createe a larger sccreen (Figure 4).
Thee initial exploration of the geometry g wass executed in the Rhinocerros CAD envvironment. Rhinoceros R mmercial NUR RBS-based (Rhhino) is a stannd-alone, com 3-D D modeling toool, developed by Robert McNeel and Asssociates (MccNeel, 2010). The realizzation of mulltiple parametters that control the surface geometry – such s as the size s of the circles, c rotatioon of the circcles, the placement of the t smaller ccircles in relaation to the bigger circlee, etc. – reqquired the geoometry to be constructed in Grasshoppper as a meaans to conttrol each vaariable param metrically. Graasshopper is a graphical algorithm pllug-in for Rhiino which allows for parrametric moddeling and scriipting (McNeeel, 2010). Althhough multip ple parameterss exist that coontrol and varyy the geomeetry, two weere chosen foor further expploration that affect apertture size andd module deppth. To undeerstand theirr effect bothh on the geoometry and their t perform mance as meaasured by radiiation control and availablee daylight, thee size and
Submittedd for publicatiion to Buildinng Simulation 2011 in Sydnney (Novembeer 2011)
the rotatiion of the smaaller circles were w manipulaated, as illustraated in Figuree 5. To undersstand the moduules' effectivenness as a shadding device, the t next step was to studyy how the variations v woould strategiccally respond tto specific envvironmental coonditions.
Figure 7: 7 42 variationn of the modulle Figurre 5: Circle siize and rotatioonal variationn Environm mental Conditions A building's skin muust respond to the particcular mental condittions of itss location. For environm purposes of this case study, Bostoon, MA (42.33°N, 71.1°W) was chosen as The a the testing environment. e Boston-L Logan Intl AP P 725090 (TM MY3) weatherr file was usedd for the lightting and energgy simulation ns. A south faccing, double glazed, g open-plan office sppace was assuumed as thee test space. The test sppace measuredd 15’-0”(4.5m m) wide, 35’-0””(10.6m) deep p, by 15’-0”(4.5m) high, and was asssumed to have h M. In studyingg the occupied hours from 8 AM to 5 PM internal ggains of an open o office space s modeled in DesignBuuilder V4[2], a shading deevice is necesssary to block most of the unwanted u raddiation, preferrably from earrly-May to laate-October (Figure ( 6) while w maintainiing a desired d luminance level l of 500 lux, recommeended by IESN NA (IESNA, 2000). 2
Figure 8: The 10 seleected moduless Vieew: View is defined d here as a the ability to have a visuual connectioon to the outtside. The viisual field from m a locationn in the room m, that is noot entirely imppinged upon by b solid objectts. Duee to angle rotation, eaach module provided diffferent viewinng conditionns. Shallow w angles provvided a frontaal direct view w, whereas largger angles provvided a more slanted indireect view (Figuure 9).
Figgure 9: Analyzzing the view based b on the rotation r of the angle
Figure 6: 6 Energy behhavior of an oppen-space offfice Changing the Parameeters The two chosen paraameters produuced 42 diffeerent module variations v (Fiigure 7). How wever, not all 42 The iterationss were viable for further exploration. e eliminatioon process considered isoolating condittions of vectorrs extending beyond the given bound dary, irregular surface behaavior, as well as geometriccally closed module m aperturres obstructingg occupant vieews. Out of thhe 42 variatioons, 10 moduules were seleected for furthher exploratio on under the following three t performaance criteria: view, v control of radiation, and interior illluminance lev vels (Figure 8).
Con ntrol of Radiation: To undderstand the behavior of eachh unit, all 10 modules weree analyzed forr radiation pennetration throough modulee apertures, using a griddded nodal suurface placed d directly behhind each moddule in a so outh facing test space within w the sim mulation env vironment. This T simulattion was perfformed using the DIVA plu ugin for the Rhinoceros R CA AD environmeent as a meaans to depict levels of radiiation. The DIVA plug-iin supports a series of perfformance evaaluations by using validaated tools inclluding Radian nce and Daysiim (Reinhart, Jakubiec, Laggios, & Niemaasz, 2011). DIIVA was chossen so that all modeling and daylighting g simulations could be carrried out withinn Rhino. Thee radiation levvel of the testt surface withhin the test space without thhe screen oveer the entire year was meaasure at 847kwhrs/m². Across all 10 modules, radiiation levels within w the tesst space over the entire yeaar ranged froom 299-507kkwhrs/m². Thhis study produced data for f total radiaation levels as a well as
Submittedd for publicatiion to Buildinng Simulation 2011 in Sydnney (Novembeer 2011)
high and low levels off radiation conncentration accross the year ((Figure 10).
Figuure 10: Radiattion map of thhe entire year Daylightt Penetration n: To furtherr understand the available daylight inn the test space, Dayllight my (DA), wass measured by b calculating g the Autonom 'climate based b metric' in DIVA, whhich uses Dayysim for the bback-end calculation. DA is defined ass the “percentaage of the occcupied hours of o the year where w a minimuum illuminancce threshold is met by dayllight alone (R Reinhart & Walkenhorst, W 2001).” Clim matebased meetric is the prrediction of various v radiannt or luminouss quantities using u sun andd sky condittions derived from f standardd meteorological datasets; the results arre dependentt both on thee locale and the building orientation, in addition to t the buildiing’s composittion and configguration (Marrdaljevic, 20099). The DA ccalculation measured the am mount of dayllight that pennetrated eachh unit moduule falling on o a horizontaal surface whiich was placeed directly behhind the south facing screen n at desk levell (Figure 11). This exercise showed how w module deppth had a diirect correlatioon to light qu uantities on the t node surfface. Shallow modules ressulting from low angless of rotation aallowed signifficant amountts of light closse to the apertuure, whereas the t deeper moodules with higgher angles off rotation allow wed deeper peenetration of light l into the rroom. These modules m effecctively acted as a light shellf. Therefore,, to achieve adequate a dayllight penetratioon into the space, s the moodules requireed a minimum m depth comppared to the baseline dessign. Thus, thhe results off lighting leevels were more m successfuul through rottational variattions as meanns to create a nnecessary deptth, rather thann through aperrture size variaation.
Figgure 10: DA of o each 10 moddules of the enntire year Parrametric Varia ation of the Depth: D The moodule was
recoonstructed in n Grasshoppper for a parametric p alteeration of thee depth onlyy. The rotatioon of the smaaller circles was w parametriccally adjusted from 0 to 85 degrees. Thee Diva Pluginn for Grasshoopper was used to measuree radiation levvels at 6 deggree steps, usinng Septemberr 21st as the teest day. As thhe angle of rotaation increassed, the dep pth of the modules exppanded, whichh resulted in loower levels off radiation pennetrating onto the test surfacce; four iteratiions of the paraametric variattion are illustrrated in Figurre 12. The radiiation level on n the test surfface without the t screen wass 8kwh/m². This T level waas reduced to 4kwh/m² withh the screen at its lowest angle of rotaation, and furtther reduced to t 2 kwh/m² at the highest angle of rotaation. In thiss research study, four moddules were thenn selected to be b tested furtther: 20, 40, 60, 6 and 80 deggrees.
F Figure 11: Mea asurement of radiation pennetration thrrough each un nit module Chooosing the Ap ppropriate Modules M Thee goal of this research wass to design annd test an ornamental shad ding system thhat would both control the radiation and a light leevels and allow a the inhaabitants to have h a view w to the outside. The moddules with thee highest degrrees of rotatioon allowed
Submittedd for publicatiion to Buildinng Simulation 2011 in Sydnney (Novembeer 2011)
for deepeest penetratioon of light innto the spacee by redirectinng daylight, as well as bloccking undesirrable radiation during summ mer months. However, duue to their deppth, these mo odules obstruccted views when w standing close to the wall. w Moreoveer, the redirecction of light, although goood for dayllight penetrattion, could ppotentially result r in ddistracting glare. Thereforee, it was neecessary to place p the deeeper modules above six feeet above finissh floor (AFF F) to minimizee glare and maximize m lightt penetration into the spacee. The modulees with low aangles of rotaation performed similarly too a thin screenn, and were ideal i mizing viewss while mainttaining a leveel of for maxim daylight control. The T aperturre sizes were w homogennous which allowed a for the t eye to focus fo beyond thhe screen planne allowing foor a uniform view v out, and therefore werre better optioons for the lo ower part of thhe screen wall (Figure 13).
secoond variant was w created with w a 30 perceent WWR apeerture evenly split s into two openings – a clerestory apeerture near thee ceiling for light l penetratiion, and a secoond centered on the walll as a visionn window. Thiird, 100 perceent glazed sou uth façade was w shaded by the t test screenn, and compaared to a spacee with the fourrth variant; an n un-shaded 10 00 percent glaazed south façaade (Figure 15 5).
Figure 12: This imagge illustrates the t benefit off the Hybrid SScreen redireecting the liight, as welll as providingg views to the outside Looking at the originaal picture of “Intercircles,” “ ” the circles apppear to havee a low rotatioonal angle. Iff the entire scrreen consists of all low angles a of rotaation then theree are no surfaaces to redirecct and reflect light l deeper innto the space. However, whhen the screenn is a hybrid m mix of mod dules – wheere the mod dules graduallyy deepen tow wards the ceeiling, the sppace receives approximately y 500 lux 1000 percent off the occupanccy time in a yeear (Figure 14).
(a)
(b)
o the test spacce shaded withh Figuree 14: (a) DA of shallow w modules, (b)DA of the sam me space shadded w the Ornaamental Hybriid Screen with Daylightt Calculation of the Screen n To underrstand the peerformance off the Ornameental Hybrid SScreen, the 155'x35'x15' (4.5mx10.6mx4..5m) test spacee was equipped with four variants of south façade gllazing: First, a 30 percent Window-to-W Wall Ratio (W WWR) was crreated with a single openning. Because daylight peneetrates deeperr into the spacce if openings are position ned closer too the ceilingg, a
Figure 13:: Four differennt tested variaants Eneergy Calculattion of the Sccreen Ballancing usefull daylight witth harmful heeat gain is critical in the design of approopriate shadingg devices, i greatly as the energy demand of a building is inflluenced by solar radiaation levels through apeertures. Throough compaarative analyysis, this reseearch study sought to understand how to incoorporate thesee criteria intoo the design process, p to testt the Ornam mental Hybrid d Screen's aability to mitigate solar heat h gain aggainst the otther three s variiants describbed above. However, simulating eneergy behavior of the test sppace posed a problem: the limits of diggital construcction for the screen in eneergy simulatioon software suuch as Desiggn Builder andd EnergyPlus. Directly creaating a compleex surface geoometry is inffeasible usinng the currennt energy sim mulation tools. In order to model m the efffect of the Ornnamental Hybrid Hy Screeen on thee energy connsumption off the space, specific perrformance criteria of the Screen had to t be measuured using VA, includingg the screen's shading coeffficient and DIV a yeearly electric lighting scheddule based onn DA. The resuults were thhen synchron nized with a thermal anaalysis perform med in Desiggn Builder linnked with Eneergy Plus (Reiinhart, 2010). Thee shading perfformance of thhe screen was simulated withh DIVA/Dayysim every hour h over thee year by placcing a verticcal test surfacce directly behind b the screeen. In this study it was w assumed that the occcupants had the ability to open thhe screen (dessigned similaar to a hinged d shutter) in the t winter
Submittedd for publicatiion to Buildinng Simulation 2011 in Sydnney (Novembeer 2011)
months ((preferably frrom Novembeer to April). The simulatioon measured light l falling on o the test surrface with and without the screen s as a shaading device. The ratio of bboth values resulted r in ann hourly shading coefficiennt of the Orna amental Hybrrid Screen (Figure 16). To accurately measure m the electric lighhting demand oof the test space, four senssors were loccated on the miid-line of the room r in a lineear fashion spaaced 9’-0” (2.77m) apart, 2'6 6”(0.8m) AFF F (Figure 18). The sensors w were place in the occuupant work area, a starting aapproximatelyy 8’-0” from the glazing. To optimize the work environment, the initial 8’0”(2.4m) buffer zone is designed foor circulation and mming to avvoid other noon-task relatted program shadows being cast onn computers annd work surfaaces. Each sennsor’s data gennerated by DIVA/Daysim was averaged to make an ellectric lightingg schedule forr the year. For the thermal simuulations, a seecond model was created in i Design Builder B and loaded with the analysis rresults listed above. The model m was teested with the aforementionned Boston-Loogan weather file, with a soouthern orienttation, and exxported as an IDF file for fuurther analysiss into Energy plus. All therrmal simulatioons were set up u in DesignB Builder V4[2] and run usingg EnergyPlus version v 6.0 [00]. It was assu umed that the space was bordered onn five sides by w analogouus spaces. As a consequencce, interior walls, floor andd the ceiling were w modeled adiabatically. a The exterior wall w has a U value v of 0.0444 Btu/h-ft²-F (0.25 W/m²-K). Heating and a cooling set points and setbacks were 20°C/15°C, and a 23°C/26 6°C, respectiveely. The glazzing was moddeled as a double pane glasss with Low-E E coating andd has a U valu ue of 0.317 Btu/h-ft²-F(1.7 B 79 W/m²-K)). All therrmal simulatioons were modeeled in Designn Builder andd run using Eneergy Plus.
Figure 14: Temporall map of the Sccreen’s Shadiing Coefficient C Consistennt with prior tests, thhe Energy Plus simulatioon was measuured four tim mes: first withh 30 percent W WWR; seconnd with 30 perrcent split WW WR; third wiith 100 perccent WWR shaded by the Ornamenntal Hybrid Screen; S and lastly with 100 percent WWR W withoutt any shading device. d
RESUL LTS Daylight Autonomy: the t 30 percennt WWR receeived 24 percennt DA, the 300 percent WW WR Split receeived 26 perceent DA, the Ornamentall Hybrid Screen
receeived 60 perccent DA, and the 100 perceent WWR receeived 71 perccent DA (Figgure 17). Acccording to thesse results, th he DA of the t screened space is signnificantly higgher than thee 30% WWR R spaces, alm most as high ass that of the un-shaded u 1000% WWR space.
o the space with w different W WWR and Figgure 17: DA of the Screened Space Bassed on these results, electricc lighting usee would be the highest in all a zones of the 30 perceent WWR spaces due to lower levelss of natural daylight, especially fartheer back in the space. Figure F 18 illustrates the antticipated electtric lighting energy e use of each e space.
Figgure 15: Electtric lighting consumption of the four teest spaces illusstrating the measured valuee of each sensor After importingg the DIVA A/Daysim ressults into Dessign Builder, the energy siimulation perrformed in Eneergy Plus reevealed that the cooling energy connsumption is the lowest in i the space with the Ornnamental Hybrid Screen. Th his was due too both the screeen’s ability to block thee undesirable radiation during summer months, andd to the redduction of elecctric lighting g needed throughout t t the year. (Figgure19).
Submittedd for publicatiion to Buildinng Simulation 2011 in Sydnney (Novembeer 2011)
DIS SCUSSION N: By studying the results of the simulation annd testing, w concluded d that the Ornamental Hybrrid Screen it was connsisting of shaallow and deepp unit modulees resulted in the t most desiirable lighting g conditions. Although the solution was the best design in terms of perfformance, it was imperattive to underrstand the connstructability of o the screen to t understand economic feassibility. To construct c the screen, a method m for connnecting multiiple modules with differennt rotation anggles was needded. This was solved byy creating inteermediate un nits that mediated m betw ween the chaanging angled modules (Figgure 21).
Figuure 19: Energyy loads of the 4 test spaces To underrstand the overall energy consumption,, the sum of source energ gy requiremeents for lightting, cooling aand heating weere calculatedd for each scennario (Bourgeoois, Reinhart,, and Macdoonald 2006). To analyze tthe total sourcce energy use of each scenaario, it was asssumed that thee test space ussed natural gas for heating annd grid-purchhased electricitty for coolingg and lighting. The EPA sourrce-site ratio of o 1.047 and 3.34 were usedd to calculatee the total souurce energy usse of each scennario (EPA, 20011). In this study, s an assu umed overall coooling coefficcient of perform mance (CoP) of 3 and heatiing efficiencyy of 0.85 perccent were useed to convert space heating g and coolinng load into site electrical and naturaal gas energgy consumpttion. Source-siite ratios wherre then used to t convert thee site electrical consumptionn for lighting and cooling and gas consuumption for heeating into souurce energy. The screeened space consumed approximately a y 35 percent lless source energy compaared to the sppace with 30 ppercent WWR R, and approxim mately 42 perrcent less source energy com mpared to thee space with 100 percent W WWR (Figure 20).
Figure 21: All A Modules ha ave uniform center c ggeometry. Con nnector moduules (blue and green) mediatte between chaanging angless Theere are numerous ways off constructingg the unit moddules in physical p spacce. Two means m of connstruction werre chosen foor testing: caasting and robotic milling. Casting is a relatively siimple and ecoonomically feaasible process. Due to the complexity of the t surface geometry, g a tw wo-part moldd made of rubbber (Smooth--On Oomoo-330) was consttructed for ¼ of the surfacce geometry. Once the mold m was creaated from 3dd printed prrototype, mulltiple unit moddules were produced p out of high denssity liquid plasstic (Smooth--On smooth-C Cast 300Q) inn less than 20 minutes m per piiece (Figure 22). 2 Thee second methhod of constru uction exploreed the use of a 5-axis robottic arm with subtractive s miilling. The moddule was mod deled in Masteercam softwarre and cut from m high-densityy foam (Figurre 23).
Figure F 22: Casst Model
Figure 220: Source energy consumpption of the 4 test t space
Submittedd for publicatiion to Buildinng Simulation 2011 in Sydnney (Novembeer 2011)
Figure 23: 2 Milled Moodel In the auuthor’s opinionn, a minimum m of two diffeerent unit moddules and a maximum of o four diffeerent modules are needed too achieve dessired lighting and energy reesults preferreed in typical office buildiings. The addeed complexityy of multiple module m variannts is worth the extra efforrt, when it reesults in a better daylit spaace with lowerr energy use. Clearly, additional research r and design proocess u screen performaance would bee required to understand on non-soouth facing faaçades, as welll as to effectiively deploy thhis screen on a building at a scale, including issues off maintenancee, relationshipp to glazing and structure..
CONCL LUSION The prooliferation off digital toools has alloowed architectss powerful freedom to envision novel n complex geometries, but b too often at the expensse of climatic performance measures. In light of this growing pproblem, this research projject illustratess the potential for the conttemporary dessign processees to fundamenntally integ grate buildinng performaance simulatioon with the arrchitectural deesign process. As proved heere, a building's shading deevice can be both b a compleex geometry that satisfies the immaterial aesthetic desires for f ornamennt, while also fundamenntally addresssing building performance and interior user u comfort.
ACKNO OWLEDGE EMTS I would liike to thank my m advisor, Prrofessor Reinhhart for his treemendous suppport and guiddance. A speciial thank youu to my facultty advisers: Leeyre Villoria, Erik Olsen, annd Matthias Ruudolph. I wouuld also like to o thank thee following stuudents at the Graduate G Schoool of Designn for their help p and support: Alstan Jackubiecc, Diego Ibarrra, Dan Weissm man, Maria Galustiann, Carl Koepckke, Manuel Diaz, Jungon Kim, K and Liween Zhang.
REFER RENCES Bourgeoiis, Denis, Reinhart, Christoph, & Macddonald, Ian. (2006). Adding A advannced behaavioral modells in whole building energy simuulation: a study y on the total energy impacct of manuual and autom mated lightingg control. Energy and Buildings, 38(7), Retrieved R frrom: 10.10016/j.enbuild..2006.03.002
EIA A, ."Overview w of Commerciial Buildings." U.S. Energy Inforrmation Admin nistration . DO OE, 2003. Retrieved froom: http://www.eeia.doe.gov/em meu/cbecs/. EPA A, Initials. U.S S Environmenntal Protectionn Agency, DOE. (2011)). Energy starr performancee ratings, methodologyy for incorporrating source energy e use EPA: Reetrieved from: http://www.eenergystar.govv/ia/business/evaluate_ performancee/site_source.p pdf IES SNA. IESNA A lighting han ndbook. 9th ed. New York, NY: Illluminating Engineering E Society of North N Americaa, ISBN 0-879995-150-8; 2000. Lovvell, Jenny. (2010). Buillding enveloppes : an N York: Princeton integrated approach. New Architecturaal Press. Marrdaljevic, Joh hn. (2009). Climate-based C d daylight analysis for residential buuildings. the daylighting d Retriev ved site, from: http://www.tthedaylightsitee.com/showarrticle.asp? id=166&tp= =6 McN Neel, R. 2010. Rhinoceros: Nurbs Moddeling for S Rhino. (Verssion 4.0 SR 8) [Computer Software]. Robert NcN Neel and Assoociates. Retrieeved from www.rhino33d.com Gallasiu, A. andd C. F. Reinnhart (2007). "Current Daylighting Design Practice: P A Survey." Building Reesearch and Innformation 366(2): 159174. Reiinhart, Christo oph, & Wein nhold, Jan. (2010). The daylighting dashboard - a simulattion-based design analyysis for daylit spaces. Buiilding and Environment, Reetrieved from: http://www.ggsd.harvard.eddu/research/gssdsquare/ Publications/HolisticDayllightingDesignnEvaluati ons.pdf Reiinhart, Christooph, Jakubiec, Alstan, Laggios, Kera, & Niemasz, Jeff. (20111). Harvard university graduate scchool of dessign (gsd)2. Retrieved from: http://www.ggsd.harvard.eddu/research/gssdsquare/ ABPS.html Reiinhart, C. F. F and O. Walkenhorstt (2001). "Dynamic CE-based RADIANC Daylight Simulations for a full-scale Test Offfice with outer Venettian Blinds." Energy and Buildings 33(7): 683-6697