Choosing the ptimum Fenestration in Commercial uildings

This paper presents a simple method of comparing the life-cycle cost of fenestration constructions incommercial buildings. The method is based on energy regression equations developed from DOE-2computer simulations. The equations take account of window wall ratio (WWR), thermal transmittance(V-factor), shading coefficient (SC) and visible light transmission (Tvis). Coefficients are developed forhigh internal gain commercial buildings, low internal gain commercial buildings and continuouslyoperated residential buildings or hotels. Climate variables are heating and cooling degree days at base65°P (18.3°P). Fenestration options can be evaluated for the whole building (assuming uniform fenestra­tion distribution) or separately for each of the major orientations.

The energy models consider daylighting benefits associated with fenestration. Required inputs are lightingpower density and design inumination in the perimeter zones of the building. The daylighting equationcan also be used to give weight to the intangible benefits of fenestration, such as views, connections withthe out-of-doors, and enhanced property value.

The energy use results from the are integrated with simple economic models so thatthe life-cycle cost of fenestration constructions can be easily compared. The cost effective fenestrationchoice for several climates and building conditions is presented to illustrate the use of the model.

Introduction Methodology

where

(1)

,11~'-(:V(~S~ cost is used to select the optimum fenestrationconstruction. The cost methodology and theassociated energy in this section.

+ $ Therms j

For a set of climate conditions, the optimum fenes­tration construction is the one with the lowest relative11T~,-cvele cost-Relative cost is used because it isonly necessary to account for those elements that changebetween one construction and the next. The relative life-

cost of a fenestration construction is calculated withbQU,ltlOn (1).

== Relative life-cycle cost of the ith fenestrationconstruction per square foot of wall area($/ft2)$

== The construction cost of the ith fenestrationconstruction per square foot of glass ($/ff).

WWR Window waH ratio (unitless).

The choice of fenestration in commercial buildings isoften driven code requirements or by initial construc­tion costsz If energy costs were considered in the selectionof alternative choices result in Sl1l;nltilCaJrlX

energy savings. accurate energy cost data arenot readily available and generally time-consumingCO]mplutt~r simulations skilled pr()re:SSl~:>n::us.

This paper a forfenestration It takes account of window area,S~nt"f"11l"1In power, desired inumination and building occu-pancy When these factors areequations may be used to evaluate fenestration"""JJl.,.itv.U.";'. For each fenestration under consideration,it is necessary to know the thermal transmittance

the coefficient the visible lighttransmission and the incremental initial cost It is notnecessary to know the absolute initial construction cost ofeach but the relative cost differencebetween the fenestration options)z The result of the proce­dure is of alternatives based on relative

1~-f~V(~I!F. cost.

c.;nOO,SlfJ1_{J the UlJ''llnrJUJ71 Fenestration in Commercialtsu'IIoJrnfJrs - 1~45

In comparing fenestration constructions in a given climate,the coefficients 110, hi' co, C1 and C2 are not significantbecause the associated terms cancel each other when twoor more fenestration constructions are compared. Only thesignificant coefficients ~, h3, c3 and c4) are presented inthis paper (see Table 1). These coefficients are presentedfor three building use patterns: a high gain commercialbuilding with 3.5 W/ft2 (38 W/m2

) total lighting andequipment load, a low gain commercial building with1.75 W/ft2 (19 W/m2

) , and a residential type occupancywith 1.65 W/ft2 (18 W/m2

) which is continuouslyoperated. Both the high and low gain commercial build­ings are assumed to be operated about 12 hours per dayMonday through Friday and a half a day on Saturday.

kWhlght = Lighting energy in the perimeter zone.Discussed in the next section.

cn , ~ = Coefficients determined through regressionanalysise

The energy use coefficients are calculated throughregression analysis of DOE2.1D computer simulationresults. The building model used for these runs is a simplefive-zone building shown in Figure 1. The computer runsare performed for 36 climates (see Table 2). For eachclimate, the performance of seven fenestration construc­tions, with V-values ranging from 1.21 Btu/(h·ft2°F)(6.76 W/(m2 0C» to 0.48 (2.73) and shading coefficientsranging from 0.81 to 0.16, are simulated. Assumptionsabout the remainder of the building envelope include anR-19 insulated roof, R-11 insulated metal frame walls andR-11 insulation under the exposed portion of the floor. Nodaylighting controls are assumed in these simulations; theelectricity savings due to daylighting are calculatedseparately, as described in the next section.

(2)

Cooling

@W'WR®+

= The present value of a kilowatt-hour peryear of electricity used over the life of thebuilding ($/kWh).

= The annual electricity use of the buildingwith the ith fenestration construction persquare foot of wall area. See Equa­tion (3). (kWh/ft2).

= The present value of a therm of D'lturalgas used each year over the life of thebuilding ($/kWh).

= The annual gas use of the building withthe ith fenestration construction per squarefoot of wall area. See Equation (2).(therms/ft2

) •

The present value terms, PVe andPVg, used in the resultssection of this paper are $O.96/kWh and $6. 72/therm.These represent energy costs of $O.08/kWh and$O.56/therm along with a real discount rate of 3 % (with­out inflation) and building life of 15 years. Other valuesmay easily be substituted into the equationl

•

and natural gas use are calculated withthat take into account not only the

fenestration performance, but also the heating and coolingfor a particular location (see Equations (2)

These equations are shown below.

where

+ c1

1$

+ Cz 0

+ C3

~ W'WR ®

C4

@ W'WR @

+

Each of the five zones is assumed to be served by its ownconstant volume packaged gas/electric HVAC system,probably the most common system type in commercialconstruction. The system includes a non-integratedeconomizer with a 60 0 P (15.6°C) setpoint. The heatingthermostat setpoint is 70°F (21.1°C), and the coolingsetpoint is 75°P (23.9°C). Single zone systems arenecessary in the computer modeling in order to tabulateresults by orientation.

WWR - Window wall ratio 11l1l,..., .. i"l",~ ....n\

== The U-value of the i th fenestration construction

:::: The shading coefficient of the ith fenestrationconstruction (unitless).

== degree days at base 65°P (18.3°C)for a particular climate zone.......... 'W"J£.Bl.L,lll:;. degree days at base 65°F (18.3for a particular climate zone.

While the building envelope and the HVAC system repre­sent a single building configuration, the results of thisanalysis should be useful to analyze fenestration choices ina wide range of building types. This life-cycle costmethodology produces a ranking of fenestration construc-tions based on relative cost-effectivenesse The

1a46 -

(3)

..... «

Commercial -9. 15E-QS

.;.1

N ·1. l1E-04

.........• ••...

.... ....

W 4. 3~ lSE-04 -2.

Annual used for in theperimeter zone per square foot of wall area(kWh/(y·ft~).

The choice of fenestration can also have an impact on theenergy required for electric lighting, and this is accountedfor in the energy model. Consideration of thedaylighting benefits gives an advantage to fenestrationwith a high visible light transmission, an other thingsbeing equal. The electric energy requirement for lightingcan be calculated with Equation (4). This is added to theelectricity use predicted Equation (3).

(4)kWh1ght :; -------

where

of HVAC and charac-teristics on the relative ....on .....,Rr .............. is much smaller than onthe absolute energy of a DlIJ .....'....... "'0"'" blLulc11n1!!.

2 and 3 compare the results of (2) andto the DOE-2 results. If there were agree-

ment, aU the would occur the lines.Points above the line are cases where the gas or

the isthan that DOE-2. Points below the

line cases when the regression is lessthan the DOE-2 results. The plots are presented for thelow commercial building, the graphs arevery similar for each of the three types. These

are for the whole building regression equations.The fit for the orientation specific regression equations isnot as with R-squared values ranging between 0.96to 0.99 for heating and 0.85 to 0.95 for cooling. (Theresults improve if a weather term accounting forsolar radiation were added to the equation).

ChOO.SlflIQ the OlJ'tinrJUJ71 Fenestration in Commercial BuildJrnfJrs

-- .......

.r/ .......".,,~

'....#<'" .......

,."

""".r/.r/ ......

....... .........; ......

"" 0:/'....... ./

...... J..... .,F'

...". ...,/...... ./

..... ./'

'" ,I"

"J ./"

Plenum space above rooms30% exposed to outside air70% adjacent to floor "above u

13 ft floor to floor height

Adiabatic wallsat corners

AIlIIIIIIIlIIII~_ Strip window withvarying height

North orientation oncompass points

100 x 100 ttInterior Zone

~ Ten 10 x 15 roomson each facade

Floor Construction15% slab on grade15% exposed to outside air70% adjacent to plenum below

1" Five Zone <,;olnpl-1tter Model Used in DOE-2 An,UVSlS

n.>

.

• •••

Heating Regression OFF - R-Squared = 0.975

100:

: ,: i··················r ··············l··················r· ······r······m ·····r .

-g ~ 1 ··················1···················t··········;;····r················,··················r··················

~ ),e:t-

10 15

DOE-2 Results(~)

2~ COlnmercial J5U,ual:nJ!.

to DOE-2 siJnulation results. Units in thousands o..f therms ofnatural gas.L'Oln1J'art~S njret1lrll-~f1 results using J..:A.-/jr"f,Ut,l,Ufl..

The coefficients for the rt~'li.T11l'Yht"'lno equation are listed inTable 3.

= Coefficients listed in Table 3. From1988).

esults

The thermal of reduced lighting energy is notaccounted for in the equations. as lightingenergy is reduced through daylighting, the cooling loadsare also reduced heating loads Thethermal benefit of daylighting can result in cooling savings

to 10% to 30 % of the direct lighting savings(depending on climate conditions). This interaction is notaccounted for in the equations, but could be a subject forfurther development. Consideration would increase thebenefits associated with fenestration constructions with ahigh visible light transmission since cooling is moresignificant than heating in most commercial buildings.

When the power and illumination levels areknown (or when reasonable assumptions can be made), themodel can be used to evaluate fenestration options and tofmd the optimal construction at different window wall

with no

== .a-I»..f....A.A~,I..A.»..,.... power in the nO'l"1II1f"1f"&:I&t,::lo1l" zone per squarefoot of wall area (W/ft~.

== Annual hours of lighting operationconsideration of daylighting controls

== Daylight fraction from(unitless).

\Vith continuous thelight savings fraction in a space can beeXl)reSSe~a in terms of the window waH the visible

transmission of the fenestration and theillumination. rrhe between these variables isdocumented in (Sullivan 1988) as:

x(5)

[1 -

where

:::: Visible light transmission of the i th

fenestration construction (unitless).WWR = Window wall ratio (unitless).C == Design Illumination (footcandles).

c;nOO,Slnlg the OlJliinr'JUln Fenestration in Commercial BurJldJfrJfjrS .. 1M49

Cooling Regression OFF .. R-Squared =0.973

. . . ......._ [ ! 1' ··r··············m.

. . . .~ i ! ~ II: : : :. a

........................_ ~ _ · · ~ · · ···· i··..·· ·· ..: ; : _s: m~ : : ~ em:; 1 ; II !

······················[······················1········ r-~ i" .

l i- ~ i............................_ ~............................ .. ~ e _ ~ ..

l.,. I I..........................................~ e • t · ·..· · · i· · · i ·.. e ..

Im I I IDOE-2 Results

(Thousands)

3~ High-Gain Commercial Building. Compares predicted results using Equation (3)DOE-2 simulation results. Units in thousands of kWh of electricity.

<PI -" ~

P2«-/iF Ali "'11 v.... ......

.........·A

\)'1... .............\

•••••••

<.

ratios. Results of the can be presented inone is to show the cost effective glazing

choice for each window waH ratio. This is illustrated in4 through 8 for several climates.

l'able 4 shows a subset of the fenestration constructionsconsidered in the optimization. Only those constructions

that are optimal for one or more of the conditionsdescribed in Figures 4 through 8 are included. Table 4gives the performance characteristics of the optimalglazing constnlctions and the relative cost. All costs arerelative to single clear glass in a metal frame which isassumed to have a cost of zero. The costs are based on asurvey of glazing contractors and include a 30% generalcontractor markup. The V-factor, shading coefficient andvisible light transmittance for each of the fenestrationconstructions is calculated using the Window 3.1 program

1988).

A total of 104 fenestration constructions were consideredin the life-cycle cost analysis, and they include combina­tions of the following: (1) standard metal framing, metalframing with a thermal break and vinyl frames, (2) singleand double glass, (3) clear, bronze, green and highperformance tinted glass for the outer lite, (4) clear, greenand high performance tinted glass for the inner lite,(5) low-emissivity coating on the second surface, the thirdsurface or both, (6) low-emissivity coated mylar filmsuspended between double panes of clear glass and

medium and high performance reflective coatings onthe second surface.

1~50 -

'il~s ...~..:thrOugh.····

:::::.

12·,}8 O~~5

S·.SB ·-0·.81 O~7S -·0.58

6.S8 ·0.57- O~66:· 0.58.

9.7fj O~:39 O~59

frmne

0.18 .... (tiS

4

9

10

11

12

13

14

15

16..:.

17

:::.. :... :.

double ·pane9'lear with suspended mylar

double pane clear ~ metaltnermai:br:eak. frame.

double pane greenouter/celarinner; metal thermal brOOkframe

double pane high performance tint·outer/ci~rinner1low-eonf/.2 $utface1 metal thernull·break;: frame

double pane green outerlclear inner, low-e. on #2 surface;:: -. 8~ 8:3:".metal thermal

0.49 0.62

i ...•••••••••••••0,11

O~45

Several assumptions were made in order toreduce the number of glazing constructions considered inthe analysise These assumptions were used in developingperformance and cost data. All of these simplifications arebelieved to represent typical conditions for nonresidential

include: (1) all glass is assumed to be1/4 inch windows are assumed to all be48 inches 72 inches for both costing and performancecalculations and (3) double glass is assumed to have atotal thickness of one inch and is constructed of two 1/4inch panes of glass.

Figures 4 through 8 each show three building types: (1) aRESidential occupancy with 1.2 W/ft2 of lighting powerand a design illumination of 30 fc, (2) a LOW gaincommercial building with 1.2 W/ft2 of lighting power anda design illumination of 30 fc and (3) a HIGH gaincommercial building with 2.0 W/ft2 of lighting power anda design illumination of 75 fc. The lighting systems areassumed to operate 2,700 hours per year with no daylight­ing contribution.

G§100,SJ/1IQ the Optimum Fenestration in Commercial Buildings ... 1,,51

Res WRes ERes SRes N

Low WLow ELow SLow N

·.·.;-:.;t

T"

~~i\5<<is::

High WHigh EHigh SHigh N

0.1 0.15 0.2 0.25 0.3 0.35Window Wall Ratio

0.4 0.45 0.5

4.. Optilnal Fenestration Constructions for Madison, WIfenestration constructions listed in Table 4.

== = 7466). Numbers correspond to

Res WRes ERes SRes N

Low WLow ELow SLow N

High SN

All

. '::":;·8'·'

0.1 0.15 0.2 0.25 0.3 0.35Window Wall Ratio

0.4 0.45 0.5

5& Fenestration Constructions forfenestration constructions listed in Table 4.

WA == 106, == NUlnbers correspond to

lM52 - Jr.. and KD,ftJe,rUD

Res W '::':::::::~:... :""".:.. ,: '::': ::·.··a::··.: ':" '.:" '::. ':.' : :.' .... :11 ·<:9 '.' '"Res E ".:.::::::::;~:.: : :;:;::.;:>:::::: : :::·:·::·::·::\:·/·:J3·· :.... : ::-: :' .-::'.: :1-1' '.;: :. >?9: : . :':.,:.:. :::';:\Res S ':::':':: ..,.::::.; : . :' ;.:". ·:;<::<8 -: "::.:. . . ···..11::·: ' '-: "::9.'· .;:':::.. ::::-: ::::

Res N ;'. >\ :·5::<:':,"::·. .:'::.:'..:'.. : '<: .. ". . . . .' . :6 . :"'::': .' "':.' ::-:-:.; :-. <:,::::::::.:.".':: ..,Res All . ..:;:-. :;::':;:::::<' :;'7>'" ' :..;:/8.':' .;<':' :..:.. :: .::::.. . :..... .:-: "::::'

Low WLow ELow SLow N ::)\~5 . :'

. ........ :::..:.:- ....::..­ I]

......'.:-'. ::.:. . ..

':".::::6/>"· :..>:- ::..: :::... ..:::..···,'···1·1 ..

0.45 0.50.40.25 0.3 0.35Window Wall Ratio

0.2

.':.:: {:f\\ ~»)<><~~:\»."}::<~:/).. ·:·::{:8..';'::})\:::8"':;"';':"

0.15

High WHigh EHigh SHigh N

0.1

60 Optimal Fenestration Constructionsfor St. Louis, MO (CDD65 =: 1467,to fenestration constructions listed in Table 4.

=: 4860). Numbers correspond

Res WRes ERes S

NAll

High SN

All

0.1 0.15 0.2 0.25 0.3 0.35Window Wall Ratio

0.4 0.45 0.5

7" Optimal Fenestration Constructionsfor Los Angeles, CA (CDD65 =:

to fenestration constructions listed in Table 4.= 1494). Numbers correspond

c;nOO,Slng the (Jo:timum Fenestration in CommerciallSu;r/OIJ1f/S ... 1M 53

Res WRes ERes SRes N

Low WLow E-Low SLow N

High WHigh EHigh SHigh N

0.1

•0.15 0.2 0.25 0.3 0.35

Window Wall Ratio0.4 0.45 0.5

Figure 8@ Optimal Fenestration Constructions for Miami, FL (CDD65 = 4045, HDD65 == 185). Numbers correspond tofenestration constructions listed in Table 4.

For each building type, the cost effective fenestrationconstruction is shown for each orientation and for a rangeof window wall ratios. For instance, consider the highgain commercial building in Madison (Figure 4). Theoptimum glazing construction on the east, south and westis a high performance tinted glass with a low-e coating (#8and #14). A thermal break frame is cost effective on theeast and west but not the south. On the east and west, thisfenestration construction is optimum up to a window wallratio of about 45 %. After that, clear double glass with asuspended mylar film (#16) is the cost effective choice.The cost effective choice changes to double clear with amylar film (#11) at about a 35% WWR on the south. Onthe north side of the high gain commercial building,however, clear double glass (#12) is optimum up to aWWR of about 28 %. After that, green double glass (#13)is the cost effective choice. A thermal break frame is costeffective on the north for all constructions. The optimalchoice for a uniform glazing on all orientations is the highperformance tinted glass with a low-e coating (#14) up toa WWR of 33 %, then clear double glass with a suspendedmylar film (#16) is the optimal choice. Thermal breakfranles are cost-effective in both cases.

1,,54 ... Eley Jr" and Kolderup

iscussion

In general the optimum fenestration construction changesto one with a lower shading coefficient (this usually meansthat the light transmission is also lower) as the windowwall ratio is increased. A lower light transmission will stillachieve dayiighting saturation at higher window waHratios.

While it is not likely that automatic daylighting controlswill be installed in all buildings, it may still be valid toassume daylighting controls for the foHowing reasons.

(1) Even if daylighting controls are not installed when thebuilding is initially constructed, tenants change incommercial buildings every five to six years and therewill be additional opportunities during the building life toinstall daylighting controls and realize daylight savings. Itmakes good poHcy to design all buildings for some day­lighting potential, even if it is not realized in thebeginning.

(2) Crediting daylighting is a way to weight tofenestration benefits that are difficult to inmonetary terms, including views, connection to the out-of­doors and enhanced property values. If it were possible toquantify these intangible benefits, it is that the

benefits would exhibit the same exponential decay as day­lighting savings when window area or effective aperture isincreased. (Effective aperture is defined as the product ofwindow waH ratio, WWR, and visible light transmittance,Tvis ·)

It is not the intent of this paper to show that providing aneffective aperture of a certain size will assure successfuldaylighting. While effective aperture is important, thesuccess of daylighting also depends on a number of otherfactors such as ceiling height, surface reflectances withinthe room, brightness contrast and glare and the layout ofwork stations and/or tasks within the space.

The architect must weigh many considerations whenselecting a glazing material, and energy performance isonly one. Another important consideration is color andreflectivity. Fenestration materials with a low visibletransmittance or with a high exterior reflectivity provide amore uniform exterior appearance to the building. It is notas easy to see inside where there can be a variety ofinterior designs or even window treatments (draperies inone place, blinds in another). While a uniform exteriorappearance may be desirable to some, a visually openfacade may be desirable to others. For instance, windowswith a high transmission (generally clear glass) arehighly desirable for ground level retail shops so thatpassers can see the merchandise on display. Somecommunities such as San Francisco have planning policiesthat encourage transmission glass so that buildingfacades do not appear so opaque (the opposite of"1"I'''\'1711,f'tn''110' a uniform appearance).

onclusion

This research shows that if daylighting is not considered,the optimum fenestration construction is generally the onewith the lowest coefficient (usually high ?"\,Q>,lI..T ..... "tf~'2"ll"'l_

ance thus or eliminating day-for energy savings from

daylighting is a way to value to the intangiblebenefits of windows such as views and a sense of connec-tion with the it may be reasonableto assume benefits even if automatic controlsare not installed.

The results of this methodology show fenestration featuressuch as tinted glass and low-e coatingsto be the lowest life-cycle cost option in many cases.

Suspended mylar films are also found to be the optimalchoice under several conditions. Reflective glass, with lowvisible light transmittance, is found to be the best choiceonly in very warm climates or with large window areas inmoderate climates.

Acknowledgments

This research was conducted by Eley Associates for thePrimary Glass Manufacturers Council (PGMC) and theAmerican Architectural Manufacturers Association(AAMA). Valuable assistance and data was provided bythe PGMC Management Committee and the PGMeTechnical Committee.

Endnotes

1. This method does not account for time-of-use utilityrates since they would add significantly to thecomplexity of the analysis. In addition, Equation (1)does not include an extra term for peak electricitydemand charges or for savings due to HVAC systemdown-sizing. These factors are not considered for sakeof simplicity, and their impact on the relative rankingof fenestration constructions could be a subject forfurther study.

References

Eley Associates. 1990. Proposed Addendum to ASHRAEStandard 90.1, Section 8 Building Envelope, Phase IIIResearch Report. Prepared for the Primary GlassManufacturers Council, Topeka, KN.

Sullivan, R; Ds Arasteh; C. Papamichael; J. J. Kim; R.Johnson; S. Selkowitz; R. McCluney. 1988. "An IndicesApproach for Evaluating the Performance of FenestrationSystems in Non-Residential Buildings" s ASHRAE Trans­actions, Vol. 94, Part 2.

Windows and Daylighting Group (W&DG)s 1988.Window 3.1, A PC Program for Analyzing WindowThermal Performance, Lawrence Berkeley uboratory.

c;noo.slna the Ul.J~tlnr;UJn Fenestration in Commercial Buildings ... 1,,55