Jakubiec, J. A. and C. F. Reinhart 1 Submitted to the 2014 Annual Meeting of the Transportation Research Board.
Assessing Disability Glare Potential Due to Reflections from New Constructions: A Case Study Analysis and Recommendations for the Future
J. Alstan Jakubiec(*,♦) and Christoph F. Reinhart(**)
* E-Mail: [email protected] Department of Architecture, Massachusetts Institute of Technology 77 Massachusetts Ave, Rm 5-418 Cambridge, MA 02139, USA Telephone: (912) 247-1086, Fax: (617) 253-6152
** E-Mail: [email protected]
Department of Architecture, Massachusetts Institute of Technology 77 Massachusetts Ave, Rm 5-418 Cambridge, MA 02139, USA Telephone: (617) 253-7714, Fax: (617) 253-6152
♦ Corresponding Author
Jakubiec, J. A. and C. F. Reinhart 2 Submitted to the 2014 Annual Meeting of the Transportation Research Board.
ABSTRACT Disability glare, visual impairment due to extreme brightness or contrast, can be caused by intense
reflections from new constructions nearby existing transportation or building infrastructure. A case study analysis is performed of a disability glare hazard at an airport created by the installation of a large array of photovoltaic panels between an air traffic control tower and the aircraft taxiway. The panels reflect blinding quantities of daylight into the control tower, produce temporary after images and dangerously obscure taxiing aircraft. The existing FAA guidelines for installation of solar technologies are discussed relative their shortcomings in identifying the glare hazard.
High dynamic range photography is used to analyze the glaring situation at the airport, and the authors propose a maximum brightness threshold of 30,000 cd/m2 based on the physiology of human vision and the brightness of tasks necessary for air traffic controllers at the case-study airport. Detailed reflectivity and three-dimensional models of the photovoltaic panels and the airport are created and validated against measured data. Using these models, an annual analysis is performed of the glare hazard. This analysis is displayed temporally using graphs and spatially using images that indicate where the glaring reflections originate. Such information is useful in identifying the potential for disability glare before new constructions are built. Finally, the authors use the new method to analyze designs for remediation of the glare hazard.
INTRODUCTION
New constructions such as photovoltaic (PV) panels or buildings can cause glare due to intense reflections of sunlight from their surfaces. Such reflections can literally occur in blinding quantities, preventing occupants from performing tasks. This effect is known as disability glare. Disability glare measurably impairs vision, reducing the contrast of the retinal image by the presence of a very bright light source in the field of view (1). To remedy such issues after construction is expensive, and those affected must tolerate the glare until the problem is remedied. In order to address the issue of glare hazards from new constructions, this paper presents a general method for analyzing glare hazards based on three-dimensional (3D) models produced during the design phase, measured material properties and physically accurate lighting simulations.
The new method is presented through a case study analysis of glare at an airport caused by intense reflections from a PV array consisting of 2,478 panels located between the airport’s air traffic control tower (ATCT) and an airplane taxiway. Specular reflections from the PV array are so extreme that they prevent the visibility of aircraft on the taxiway and cause temporary after images that impede the viewing of computer monitors inside the ATCT. As such, the reflections at the case-study airport meet the qualifications to be considered disabling.
The PV array was analyzed according to Federal Aviation Administration (FAA) best practice guidelines for the installation of solar technologies prior to its construction; however, the original analysis did not detect the glare hazard. Thus, there is a need to define a clear process through which to conduct glare prediction analysis that is useful during the design of new constructions that have the potential reflect large quantities of daylight. A case study approach is taken to address this need through an analysis of the airport’s glare problem. First, a review of existing methods for analyzing glare from specular surfaces is conducted. Following that, the authors reproduce the problem in a physically-based daylighting simulation software capable of predicting reflections from new designs before they are constructed and outline steps necessary to reproduce our method. The simulations use physically accurate material models that account for specular and diffusing reflective surface properties of PV panels. Side-by-side comparisons of high dynamic range (HDR) photographs and the authors’ physically-based renderings confirm the accuracy of these models. Next, an annual analysis is conducted using a ten-minute simulation interval for every daylit hour in the year. Following this, the likelihood of experiencing disability glare is displayed spatially and temporally in order to understand the time, location and intensity of potential glaring reflections. Finally, the new method is used to analyze proposed remediation strategies of the disability glare problem at the case-study airport.
Jakubiec, JSubmitted to
QUALIFY Disability
FIGUR
Tlevels fromphotographwas taken. indicated inupon captuare compoluminance scenes andimage formMinolta LsPV panel i
Ththeir job ofsunlight, shthese refer10,900 cd/however, i
Thlooking towto an effecopening aran extremediscomfort
J. A. and C. F. the 2014 Annual M
YING DISABI
glare at the ca
RE 1 High dy
o document thm the case-studh and can be usFigure 1 illust
n units of lumiuring several exsited and calibdata with calib
d 10.1% for daymat, a lossless fs-110 spot lumnstallation. hese ultra-brigf seeing aircrafhown in Figureence interior sum2. The sun, et is not directlyhe very bright ward the taxiwt known as for
rea when reflecely bright light t or visual disab
Reinhart Meeting of the Tra
ILITY GLAR
ase-study airp
ynamic range ph
e disability glady airport’s ATsed to calculatetrates one suchinance (candelaxposures at knorated to physicbrated instrumeylit interior sceformat where e
minance meter to
ght reflections cft on the taxiwae 1 have measuurfaces. At theven partially oy in the field of(> 250,000 cd/
way and are murward scatteringcted due to roug
source such asbility is likely
nsportation Resea
RE DUE TO SP
port
hotograph of g
are problem, thCT. Calibratede the likelihood
h photograph wa per square mown shutter spcal values usingents. Inanici foenes (2). The reeach pixel corro verify that va
cause disabilityay. Most indooured a luminane time of observobscured by thef view when lo/m2) reflectionch larger than g. Forward scaghness and subs the sun is sprto occur.
arch Board.
PECULAR R
glare problem t
he authors took d HDR photogrd of discomfor
with spot brightmeter, cd/m2). T
eeds, lens aperg a camera-speound this methoesulting HDR presponds to a phalues above 25
y glare, prevenor surfaces, witce below 3,000vation, outdoore shading systeooking toward tns from the PV the sun. The gr
attering is whenb-surface scatteread across an a
REFLECTION
taken on Augu
k HDR photogrraphs record thrt or disability gtness values me
The method usertures and expoecific calibratiood to have an aphotograph is shysical value.
50,000 cd/m2 w
nting the air trath the exceptio0 cd/m2. Outdor surface lumin
em, is very brigthe taxiway. panels are locrowth in size cn a direct light ering propertiearea due to for
NS
st 30th, 2013 at
raphs calibratedhe brightness oglare at the timeasured at the red to create sucosure times. Afon function basaverage error ostored in the RThe authors us
were observed r
affic controllerson of areas illumoor objects are nances were beght and over 6,
ated in the fielcompared to th
source is spreaes of the reflectrward scattering
t 7:22 solar tim
d to physical liof each pixel inme the photograred squares andch images reliefterwards the psed on measureof 5.8% for out
Radiance RGBEsed a Konica reflected from
s from performminated by direoften brighter
etween 2,475 a000,000 cd/m2
ld of view wheat of the sun isad out in an anting material. Wg, extreme
3
me.
ight n a aph d
es photos ed tdoor E
the
ming ect than
and 2;
n s due ngular When
Jakubiec, J. A. and C. F. Reinhart 4 Submitted to the 2014 Annual Meeting of the Transportation Research Board.
Dynamic Range of Human Vision To assess the impact bright specular reflections have on perception, considering the physiology of human
vision is critical. The eye is a logarithmic sensor capable of perceiving 12 log units of luminance; however, human vision is only capable of resolving a portion of the luminous scale at any given time based on the visible luminous environment (3). Specifically, the adaptation luminance is the brightness level the eye is adapted to that determines what range of the luminous scale is perceivable.
Ferwerda, et al. suggest that the adaptation luminance should be half the value of the highest perceived luminance, which will vary depending on the outside brightness, solar position, and reflectivity of glare-inducing surfaces (4); however, this works primarily for normal viewing conditions where the sun and glaring reflections are not present. In the case of disability glare analysis where the task is known, it makes sense to choose the adaptation luminance relative to the desired tasks. From Figure 1, focusing on the monitor (255 cd/m2), is the lowest task luminance level, which is reasonable considering the luminous range of a typical computer monitor is between 10 and 300 cd/m2. The highest task luminance level is focusing on the taxiway (8,901 cd/m2).
Generally, the human eye can recognize between two and three orders of magnitude of luminance variation at any time (3). Taking the mean brightness of the two task luminances, approximately two orders of magnitude of luminous difference can be visualized (3, 5). This means that in order to view planes on the taxiway while maintaining the ability to read the monitor, luminance levels in the line of sight would need to be less than 30,000 cd/m2 observing these two orders of magnitude as a rule. Therefore the authors propose a brightness of 30,000 cd/m2 as a threshold at which the probability of experiencing disability glare is likely. The measured value of 250,000 cd/m2 from the PV panels in Figure 1 is well above this threshold. Air traffic controllers attempting to view the runway experience after images and are prevented from visually identifying aircraft on the taxiway. Although the sky is very bright and often above 30,000 cd/m2, the operable shades present in the ATCT can be used to block the brightness of the sky; however, they cannot be lowered to obstruct the PV array without also blocking the taxiway. Until the glare problem is remedied, the airport is covering the PV array with tarps, which prevent disability glare but have the side effect of also preventing the generation of electricity.
Existing Disability Glare Metrics from Specular Reflections
The FAA defines interference from solar panels concerning airspace penetration, reflectivity, and communication systems interference (6). As this analysis relates to visual disability, the authors are concerned with the regulations dealing with reflectivity. All solar installations at airports must include an “assessment of reflectivity including time periods when reflection may contact [the ATCT] and aircraft.” The FAA provides three methods of analyzing potential glare problems:
1. A qualitative analysis of potential impact in consultation with [air traffic control staff], pilots, and airport officials,
2. a demonstration field test with solar panels at the proposed site in coordination with FAA Tower personnel or,
3. a geometric analysis to determine days and times when an impact is predicted. Method 1 is potentially inadequate as the involved parties may not have the experience or the ability to
assess the potential glare hazards involved with the proposed PV system; however, it is probably adequate when PV panels will be installed far away from any critical visual areas. Method 2 is inadequate as the field test is performed during a single moment; however, the sun changes its position in the sky throughout the entire year. A test at a single time may miss glare hazards during other parts of the year. Method 3, geometric analysis, could analyze potential glare hazards throughout the year; however, there is currently no guidance on its implementation beyond two example figures that portray a perfectly specular reflection at four times during the year as if from a mirror. Such an analysis ignores potential glaring reflections during other times of the year. More importantly, a purely geometric analysis neglects the true behavior of reflected light from PV panels. Such an analysis could miss identifying potential glare hazards where an analysis that considers forward-scattering when daylight is reflected would not.
Ho et. al propose a geometric analysis method that accounts for diffuse and specular reflective properties as well as the form of a reflecting surface in order to assess glare hazards (7). The result of their geometric calculation is evaluated based on irradiance at the retina and the size of the glaring source. It is based on previous research on the physiology of the human eye to experience after images and retinal burning. The method is useful for detecting problems and quickly iterating through solutions; however, it does not provide spatial feedback regarding where the glare originates from at any given time.
Jakubiec, JSubmitted to
METHODIn
simulation the physicaRadiance imaterials b
Umethod forwould be tshow that mespecially the authorsand HDR ppercentagespectrum; hspecular reforward-scto the panewere transldescriptionmeasured Pdepending respectivel
Fi
software anpixel correconstitutesbased rendblue areas similar clea
J. A. and C. F. the 2014 Annual M
DOLOGY n order to analy
model of the Aally-based backs a reasonable
based on opticaUnfortunately, dr resolving the o use costly gomeasured refleat shallow incis took measurephotography ase of diffuse andhowever, the in
eflections fromcattering properel and at an obllated into the An is useful as it PV panel data. on the inciden
ly.
igure 2 portraynd the author’s
esponds to a phs a way to assesdering. For examreflect approxiar sky conditio
Reinhart Meeting of the Tra
yze the current ATCT, PV arrakward raytracechoice as it inc
al measuremendetailed measucomplex, angu
oniophotometeections can diffident angles ofements of the ins documented id specular reflenstrument is un
m the panel at dirties based on Hlique angle (76Ashikhmin-Shi
allows separatOur measurem
nt angle and are
TA
ys an example Hs custom materhysical luminanss large luminomple, yellow aimately 10,000ons. Visually, t
nsportation Resea
glare situationay, taxiway ander Radiance, decludes a series
nts (9) and has bred reflection dular-dependentr measurement
fer from standaf sunlight (15).nstalled PV panin Table 1. Theection from thenable to resolviffering incidenHDR photogra
6 degrees). Thisirley material tyte diffuse and s
ments found a tea-weighted di
ABLE 1 Measu
M
A
HDR photograrials (2b) rendence value in cdous ranges visuareas in Figure 0 cd/m2. The phthe behavior of
arch Board.
n at the airport,d surrounding eveloped at Law of material mobeen validated data for PV pant reflections frots (14). Goniopard material mo
As such measunels using a Koe spectrophotome PV panel surfve information nt angles of sunaphy taken withs results in beaype in Radiancspecular reflectotal beam spreffuse and spec
ured Properties
Angle fromsolar incide
12 de76 de
Measured arePV Cells D
PV Cells SWhite AreasWhite Areas Silver StripsSilver Strips
Frame DFrame Sp
Area-WeightArea-Weighte
aph of the PV pered in falsecold/m2 as indicateually as they va2 reflect great
hotograph and f the simulation
, the authors crbuildings. Thewrence Berkelodifiers that al
d by numerous nels is largely om semi-specuphotometer meodels by over turement deviconica Minolta meter measureface sampled aabout the amou
unlight. The auth the sun at a n
am spread anglce (16, 17). Thction parameteread angle of becular reflectivit
s of PV Panel
m normal ence (θ)
eg. eg. ea and typeDiffuse Specular s Diffuse s Specular s Diffuse Specular iffuse
pecular ted Diffuse ed Specular
panel (2a) and lor. In a falsecoed by the colorary across a calter than or equathe rendering w
n closely mimi
reated a high-qe simulation eney National Lallow users to sestudies (10, 11unavailable. T
ular objects suceasurements hatwo orders of mces were not av2600d portable
ements charactat every 10 nm unt of forwardthors thereforenear-normal incles described in
he Ashikhman-Srs, which vary etween 2.0 andties of 8.67 and
Half beam spread angle
(β/2) 1.0 deg. 4.6 deg.
Reflectance 2.56% 4.12%
45.13% 4.52%
18.63% 36.40% 61.23% 3.97% 8.67% 5.58%
a rendering usiolor rendering,rbar scale. Suchlibrated photogal to 100,000 cwere captured ics that of the a
quality daylightnvironment useaboratory (8). et up custom 1, 12, 13). The best practicch as PV panelsave been used tmagnitude, vailable for thise spectrophotoerize the overaof the visible
d-scattering of e quantified thecidence (12 den Table 1, whicShirley materiasignificantly in
d 9.2 degrees d 5.58 percent
AshikhmShirley spe
exponen3500175
Relative a84.80%
5.62%
4.47%
5.11%
ing the Radian, the color of eah falsecolor imgraph or physiccd/m2 while dar
and rendered uactual panel.
5
t ed is
ce s to
s study, ometer all
e egrees) ch al n the
man-ecular nt
area %
%
%
%
nce ach
magery cally-rk under
Jakubiec, JSubmitted to
(a)
A
reproduce the US DepUsing the 3probabilityRadiance rtaken at 7:2displayed ssame time than 30,00from the P
take
J. A. and C. F. the 2014 Annual M
High dynamictaken on AugFIGURE 2
Apart from apprthe observed gpartment of Tr3D model and y of disability grendering using22 solar time osuch that yellowis shown in 3b0 cd/m2. UsingV panels at any
(a) HDR en on 8-30-201
Reinhart Meeting of the Tra
c range photogrgust 31st, 2012.
Photographed
ropriate materiglaring conditioransportation’s the detailed m
glare for every g the same dayon the morningw areas corresp
b. Careful compg this calibratedy time during t
photograph 12 at 7:22 solar
FIGURE
nsportation Resea
raph . and rendered f
at near-n
al descriptionsons in the RadiVolpe center aaterial definitiodaylit hour in lit condition ar
g of August 30th
pond to brightnparison of Figud simulation mthe year in a fe
r time 3 Imagery of
arch Board.
(b) PhysicCIE c
falsecolor imagnormal inciden
s, a geometricaiance simulatioand was translaons described pthe year. As a re compared inh, 2012 from thnesses greater ures 3a and 3b
model, it is possew seconds.
(b) Phsimulated f
f glare issues se
cally-based renclear sky conditges of a single nce.
ally precise 3Don environmentated by the autpreviously, it iproof of conce
n Figure 3. 3a ihe ATCT viewthan 30,000 cdshows a very g
sible to simulat
hysically-basefor 8-30-2012 aeen from the A
ndering under itions with sun.PV panel with
model is necet. A 3D modelthors into the Ris possible to pept, a HDR phois a HDR lumin
wing the taxiwad/m2. A rendergood agreemente the intense s
ed rendering at 7:22 solar ti
ATCT.
ideal .
h the sun
ssary to accura was provided
Radiance formapredict the otograph and anance photograay. The image iring simulated nt of areas greasolar reflection
me
6
ately by
at.
a aph is at the
ater ns
Jakubiec, JSubmitted to
Tdistributionyields the tand does nconditions
RESULTS
Inthe glare odisplay is uFigure 4 illten minuteminute inteorder to assolid angleby the PV of 0.18 str-portray a staxiway.
FIG Th
authors creprovide spexample ingeometric the size of dimensionaand is itsEquation 2( ,W/m2)visual condEquation 1Finally, the179 lm/W Ho, et al.’s
J. A. and C. F. the 2014 Annual M
The authors anan as a ‘worst catime in the yea
not account for are stored in th
S n order for glarriginates from.useful to assesslustrates a falses in the year. Terval, and the csess this chart,
e of the sun viepanels in the li-hrs annually. Tolid angle of 9
GURE 4 Annu
he metrics of Heate. The benefatial feedback
n Figure 3, it ismethod wouldthe glaring soual angle in radis equivalent in
2. Finally, the rand the diame
ditions of Figu1 renders an ane mean irradianluminous effic
s metric, this ir
Reinhart Meeting of the Tra
alyzed the glarase’ scenario w
ar during whichovercast or clohe Radiance R
re analysis to b. To accompliss the extent, loecolor plot sho
The horizontal colors correspo, it is useful to ewed from Eartine of sight to tThe reflections
9×10−3 str, roug
ual temporal m
Ho, et al. (6) cafit of renderingthat reveal the
s evident that dd only report thurce in our imaians. This is ac
n radians. The dretinal irradianceter of the pupiure 3 as an examgle of 1.1x10-1
nce at the corncacy. Plugging rradiance and r
nsportation Resea
e potential for when the sun ish glare could boudy days. Res
RGBE image fo
e meaningful, sh this, the analcation, and tim
owing the perceaxis contains o
ond to the percenote that the soth is roughly 6the taxiway. Ths account for a ghly 15 times la
ap showing the
an be calculateg physically-ba causes of disa
disabling glare oat there is glar
ages must be cochieved using Ediameter of thece ( ) is a funil when exposemple, there is a1 rad, which eqea is 940 W/mthese values inetinal projectio
arch Board.
every ten minus present and the caused by dirsults recording ormat, the same
it is necessary lysis results are
me of occurrenceived size of glone column foreived size of tholid angle of an×10−5 str. Annuhis can be comnearly 20-fold
arger than that
e perceived siz
d using input fased images ratability glare whoriginates frome, not its locationverted from Equation 1 whee image projectnction of a 0.5 ted to daylight (0a solid angle ofquates to a 1.9xm2 as derived fronto Equation 3 on is likely to c
utes in the yearhe sky is free orect sunlight bthe intensity a
e format as the
to know at whe displayed temce of disabilitylaring reflectior every day, thehe glare sourcen idealized vieually there are
mpared to the vd increase. Thet of the sun, bu
ze of reflection
from rendered ther than usinghile geometric m the center of ion or intensitya three-dimen
ere Ω is the solted on the retintransmission fa0.002 m) as shf 9.5x10-3 str brx10-3 m diameteom the authors gives a retinalcause after-ima
r assuming a Cof clouds. The aeing reflected
and spatial locae authors’ HDR
hat time the glamporally and spy glare from speons over 30,000e vertical axis e in solid angleewing hemisph
3.57 str-hrs ofvisibility of the e photo and renut located in the
ns greater than 3
images such asg geometric rescalculations al
f the PV array; y. To calculate sional solid anlid angle of thena ( , m) can bfactor, the irradhown in Equatirighter than 30er projected ons’ simulations ul irradiance of ages (6).
CIE clear sky (1analysis thereffrom the PV pation of glaringR photographs.
are occurs and patially. This decular reflectio0 cd/m2 for evea row for each
e steradian (str)here is 2π str, anf glaring area csun from the A
ndering in Figue line of sight t
30,000 cd/m2.
s the ones the ults is that imalone do not. Fohowever, a Ho, et al.’s me
ngle to a two-e glaring reflecbe calculated u
diance at the coion 3. Using th0,000 cd/m2. Unto the retina. using Radiance522.8 W/m2. U
7
18) fore anels
g
where dual ons. ery
h ten ). In nd the caused ATCT ure 3 to the
ages or
etric,
ction using ornea e
Using
e’s Using
Jakubiec, JSubmitted to
Su
reflections colored arephotographadding 1 fointervals wlabels for enot cause gSuch an im
A
prevent disthroughoutoccurance methodolowith tarps;in order to clearly porexample, ifcause glari
DISCUSS
Induring the representatspatially. Istrategies. to replace tfaces awaya PV arrayparking strof a vertica
J. A. and C. F. the 2014 Annual M
upplementing tthroughout the
eas cause no glh in the Radianor every timest
within an hour –each row in theglare; therefore
mage would hav
FIGURE
As discussed earsability glare frt the year until of glaring refle
ogy to that of F however, in Nmaximize PV
rtrays that doinf a field test waing reflections
ION n the results secdesign of a PVtions and a comt is therefore uThe authors evthe PV panel w
y from the ATCy design procesructure, were nal shading devi
Reinhart Meeting of the Tra
the temporal me year in hourslare, and yellownce RGBE imatep when a pixe– spatial represe PV array are e, it is likely thve provided va
E 5 Annual hou
rlier, the airporrom occuring. remediation m
ections for halfigure 5 but usi
November it is electric produ
ng a spot field tas done in Julyin September,
ction, it is showV system basedmplete 3D moduseful to emplovaluated two rewith a less-refleCT. These two ss. Other optionnot considered dice on the park
nsportation Resea
2 cos
0
0.5
map of Figure 4s above 30,000w-colored areage format has el is over 30,00sentations of thapplied from Aat if the PV arr
aluable feedbac
urs of predicted
rt is covering pThe author’s m
methods can bef of the year frong monthly intonly necessaryction as there itest of PV paney, no glare wouOctober and N
wn that the newd on sky modeldel. Furthermoroy the method temediation optective panel. Ooptions consti
ns, including codue to regulato
king structure).
arch Board.
1Ω2
0.017
0.002
4, Figure 5 show cd/m2 with the
as cause 60 or ga physical lum00 cd/m2 and thhe glare problemA1 to C3. The bray was locatedck during the d
d glare under C
parts of the PVmethod was use
taken. Figure om July 1st to tervals. For exay to cover row is no likelihoodels as possible uld be detectedNovember.
w image-based ls that accuratere, the occurrento detect potentions developed
Option 2 was totute typical maonstructing extory pressures an
ws in falsecoloe scale ranginggreater hours o
minance value athen dividing bm are quickly bottom cornerd further north
design of the PV
CIE clear sky c
V array that caued to plan when6 illustrates mDecember 31sample, in July A1. In Decembd of glare fromin the FAA’s
d for panel row
d analysis methely resolve the nce of glare ca
ntial new glare hd in concert wi
o rotate the entiaterial and formterior shading
and their potent
or the location g between 0 anof glare annuallassociated withby 6 – the numband easily creaof row C2 and
h, no glare probV array.
conditions with
use glare with tn each row sho
monthly falsecost generated usrows B2 – C2 ber all rows ca
m specular refleguidelines cans A1-B1; howe
hod can identifysolar position,
an be displayedhazards in propith the airport sire array 90 degmal changes thsystems on thetial size (five s
of glaring nd 60 hours. Puly. As each ren
h each pixel, byber of ten minuated. In the figud all of row C3blems would oc
h sun.
tarps in order toould be coverelor hours of
sing an identicashould be cov
an be left uncovections. Figure n be insufficienever, those row
y glare problemappropriate m
d temporally anposed remediastaff. Option 1 grees such that
hat may occur de ATCT or the tories tall in th
8
(1)
(2)
(3)
urple-ndered y ute ure, do ccur.
o d
al vered vered 6 also
nt. For ws
ms material nd ation
was t it during
he case
Jakubiec, JSubmitted to
FIGURE Option 1:
Aresulted in 2.97%, whThe new sienough to photograph125,000 cdcd/m2. Theis substant
Thbright as nbecause ducause disab
Option 2:
Arotating theway, the opsimulationface away
Binvestigatemaximum analyzed afeet above hemispherevisible soli
J. A. and C. F. the 2014 Annual M
Jul
Septem
NovemE 6 Selected m
A Less ReflecA potential repla
an area-weighhich means the imulation and asolve the glareh in Figure 1, wd/m2, a substanerefore, in the cially less forwahe less reflectioted above; ho
uring clear sky bility glare.
Rotate the PaA geometric sol
e panels 90 degperators can nos using an idenfrom the ATCefore assessing
e the potential fpotential for th
at every 100 versea level (just e. For referencid angle sizes o
Reinhart Meeting of the Tra
y
mber
mber monthly image
ctive PV Panelacement panel
hted diffuse refproposed paneanalysis proces
e problem. Befowhere the PV pntial reduction; case of Figure ard-scattering. ve PV panels d
owever, the annconditions eve
anels 90-degreution to rotate grees counter-co longer see thentical process tT, a new glare g disability glafor glare. It is phe size of a reflrtical feet of debefore landing
ce this is 2% ofof the proposed
nsportation Resea
s showing hou
l was measured
flectance value el reflects 45.3ss was used to ore describingpanels had a mehowever, it is 1, a less-reflec
do cause a percnual glaring areen the reduced
ees to the Eastthe panels suc
clockwise, theye reflective facto those shownhazard may be
are potential forpossible to deteflected glare soescent. The mag) with a visiblf the glare size d redesign of th
arch Board.
urs of predicted
d using the samof 4.73% and and 46.8 percinvestigate whthe results, it ieasured brightnstill much high
ctive panel will
ceptible glare dea decreases onspecular reflec
, Away from tch that they facy will face soutce of the PV pan in Figure 4. Oe created for pir each positionermine the visiurce. The case
aximum visibille solid angle oobserved on A
he PV array are
August
October
December d glare under C
me process as dan area-weighent less light re
hether this subsis worthwhile tness of over 25her than our del not correct th
decrease. The rnly 8.0% to 3.2ctions of the im
the ATCT ce away from thutheast and awaanels. This was
One concern wailots.
n along the landible size of the e-study airport lity of the solarof 1.8x10-4 sr oAugust 30th 20e shown in Fig
CIE clear sky co
documented in Thted specular reespectively thastantial reflectito consider the50,000 cd/m2. Hefined comfort e glaring situat
reflections wou29 str-hrs. Thismproved panel
he ATCT was ay from air trafs confirmed usias that by rotat
ding paths, it is PV panels whhas four landinr panels is from
or 0.0029% of t12, depicted in
gure 7a from th
onditions with
Table 1, whicheflectance valuan the installedive reduction is case of the Half of this vat threshold of 3tion unless the
uld always be ls meager decreare bright enou
explored. By ffic operators. Iing annual ting the PV pan
s worthwhile tohich is exactly tng paths whichm runway 35 atthe visual n Figure 3. Thehe ATCT and a
9
sun.
h ue of d panel. s
lue is 30,000 panel
less ease is ugh to
In this
nels to
o the h were t 300
e along
Jakubiec, JSubmitted to
the variousof the PV afrom the aito experienonly the ba
CONCLU
Thconstructioappropriatebuildings wthe behavioluminance resolve itemapplied to
Afor Evaluataccount forthroughoutproblem lik
Ain practicespecular mmeasuremethat may b
J. A. and C. F. the 2014 Annual M
s landing pathsarray from the ircraft landing nce glare. Furthack sides of the
(b) Vis
FIGURE 7
USIONS his case study
ons that use spee for the specuwith a large amor of physical rvalue of 30,00
ms beyond thathe analysis of
An issue identifting Selected Sr the physical rt the year. Hadkely would hav
A significant ba. This makes it
materials. In ordents (14) need e used in critic
Reinhart Meeting of the Tra
s. To reinforce ATCT (5x10-2
paths to the exher, the potentie PV panels.
(a) Visibility
sibility of rotatthe ori
Solid angle of
illustrates the pecular material
ular surfaces of mount of glazinreflections whe00 cd/m2 in theat threshold whf an airport withfied in this studSolar Technoloreality of reflec
d the analysis mve been identif
arrier to such ant nearly impossder for meaningto be performe
cal situations.
nsportation Resea
what this 2% m2 sr). When comxisting visibilityial for air traffi
y of rotated PV
ed PV panels fginal visibilityvisible area of
potential for phls in critical sitf PV panels; hong or bodies of en using appro
e line of sight hhen also viewinh a known disa
dy is that the anogies on Airporctions from PV
method proposefied before connalysis is that msible to compargful analysis toed and made av
arch Board.
means, Figure mparing PV viy of PVs from ic controllers to
V panels from a
from airplane ly of PV panels f PV in steradia
hysically-baseduations such a
owever, it is alswater. The aut
opriately calibrahas been proposng a typical comability glare prnalysis guidelinrts” document
V panels or thated in this studynstruction. measured matere the benefits o be made in thvailable for a w
7b adds an extisibility, the mathe ATCT, the
o experience g
airplane landin
landing paths cfrom the ATCan from the AT
d renderings tos at airports. Tso appropriate thors have showated material msed based on thmputer screen.roblem. nes proposed inare vague and t does not acco
y or that of Ho,
erial data from of choosing be
he future, detaiwide variety of
tra data point, taximum possibere is very littl
glare is now zer
ng paths.
compared againCT. TCT and landin
o analyze propoThe new method
for other constwn that simula
models. A newhe physical abiThis upper lum
n the FAA’s “Tallow analysis
ount for differin, et al. been uti
PV panels is laetween differeniled angular-def potential cons
the current visible glaring refle potential for ro as they only
nst
ng paths.
osed new d of analysis istructions such ations closely m
w maximum visility of the eyeminance limit w
Technical Guids that does not ng solar positiolized, the glare
argely not avaint PV panels o
ependent materstruction mater
10
ibility lection, pilots
y see
s as
mimic ible to was
dance
ons e
ilable or rial rials
Jakubiec, J. A. and C. F. Reinhart 11 Submitted to the 2014 Annual Meeting of the Transportation Research Board.
After using a HDR photographic method to derive angular-dependent material reflection data, the authors undertook an annual analysis of the case-study airport with a known occurrence of disability glare. The new method produces charts (Figure 4) that illustrate the time and intensity of glaring reflections and images (Figures 5 and 6) that show the location of glaring reflections. Such results allow intuitive design changes based on observations. For example, Figure 5 suggests that a more northern site for the PV array would have been beneficial. The authors’ new method was used to analyze two proposed remediation strategies. This analysis found that a material solution using a less-reflective PV panel was not viable in this case, but that a geometric solution to rotate the PV panels would remedy the glare hazard.
REFERENCES
1. Commission Internationale de l’Eclairage. Discomfort Glare in Interior Lighting. CIE Publication 117, 1995.
2. Inanici, M. N. Evaluation of High Dynamic Range Photography as a Luminance Data Acquisition System. Lighting Research and Technology, Vol. 38, No. 2, 2006, pp. 123-34.
3. Boyce, P. R. Human Factors in Lighting, Second Edition. Taylor & Francis Inc., New York, 2003. 4. Ferwerda, J. A, S. N. Pattanaik, P. Shirley, and D. Greenberg. A Model of Visual Adaptation for Realistic
Image Synthesis. Proceedings of ACM SIGGRAPH '96, pp. 249-58. 5. Hopkinson, R. G. and J. B. Collings. The Ergonomics of Lighting. Macdonald & Co., 1970. 6. Federal Aviation Administration, Office of Airports, Office of Airport Planning and Programming,
Airport Planning and Environmental Division (APP-400). Technical Guidance for Evaluating Selected Solar Technologies on Airports. http://www.faa.gov/airports/environmental/policy_guidance/media/airport_solar_guide.pdf. Accessed May 9, 2013.
7. Ho, C. K., C. M. Ghanbari and R. B. Diver. Methodology to Assess Potential Glint and Glare Hazards from Concentrating Solar Power Plants: Analytical Models and Experimental Validation. Journal of Solar Energy Engineering, Vol. 133, No. 3, 2011.
8. Ward, G. J. The RADIANCE Lighting Simulation and Rendering System. Proceedings of the 21st Annual Conference on Computer Graphics and Interactive Techniques, 1994, pp. 459-72.
9. Reinhart, C. F. and M. Andersen. Development and Validation of a Radiance Model for a Translucent Panel. Energy and Buildings, Vol. 38, No. 7, 2006, pp. 890-904.
10. Mardaljevic, J. Validation of a Lighting Simulation Program Under Real Sky Conditions. Lighting Research and Technology, Vol. 27, No. 4, 1995, pp. 181-8.
11. Reinhart, C. F. and O. Walkenhorst. Validation of Dynamic RADIANCE-based Daylight Simulations for a Test Office with External Blinds. Energy and Buildings, Vol. 33, No. 7, 2001, pp. 683-97.
12. Reinhart, C. F. and P. F. Breton. Experimental Validation of Autodesk® 3ds Max® Design 2009 and Daysim 3.0. Leukos, Vol. 6, No. 1, 2009.
13. Ibarra, D. and Reinhart, C. F. Solar Availability: A Comparison Study of Irradiation Distribution Methods. Proceedings of Building Simulation 2011: 12th Conference of International Building Performance Simulation Association, 2011, pp. 2627-34.
14. Apian-Bennewitz, P. and J. von der Hardt. Enhancing and Calibrating a Goniophotometer. Solar Energy Materials and Solar Cells, Vol. 54, No. 1, 1998, pp. 309-22.
15. PAB Advanced Technologies, Ltd. BME BSDF Data. http://www.pab.eu/gonio-photometer/demodata/bme/ward_metal_ps.anofol606.37.385/index.html. Accessed July 1, 2013.
16. Ashikhmin, M. and P. Shirley. An Anisotropic Phong BRDF Model. Journal of Graphics Tools, Vol. 5, No. 2, 2000, pp. 25-32.
17. Ward, G. What’s New in Radiance for 2013. http://www.radiance-online.org/community/workshops/2012-copenhagen/Day2/Ward/RadianceImprovements.pdf. Accessed January 15, 2013.
18. Commission Internationale de l’Eclairage. Spatial Distribution of Daylight - CIE Standard Overcast Sky and Clear Sky. CIE Publication S003, 1996.