4.430 Daylighting
Christoph Reinhart
4.430 Visual Comfort & Occupant Behavior
Massachusetts Institute of Technology
Department of Architecture
Building Technology Program
Daylighting/Lighting in LEED
Indoor Environmental Quality section:
credits for:
 (a) glazing factor (daylight factor) of 2%
(b) Between 250 and 5000 lux under equinox
clear sky at 9AM and 3PM.
 view to the outside
Compliance via spreadsheet method.
MIT 4.430 Daylighting, Instructor C
Reinhart
1
1
LEED 2.2 Glazing Factor Formula
How many of you have used
this formula?
External obstructions are not considered.
View to the Outside in LEED I Credit 8.2 Views for 90% of Spaces Achieve
direct line of sight to vision glazing for
building occupants in 90% of all regularly
occupied spaces. Examples for exceptions
copy rooms, storage areas, mechanical,
laundry and other low occupancy support
areas.
MIT 4.430 Daylighting, Instructor C
Reinhart
2
3
View to the Outside in LEED II
Image by MIT OpenCourseWare.
View to the Outside
• Size and content matter
• Information rich views with natural
elements provide satisfaction and
health benefits
MIT 4.430 Daylighting, Instructor C
Reinhart
3
4
Daylighting/Lighting in LEED
Sustainable Sites:
Light Pollution Credit “eliminate light
trespass from the building and site, improve
night sky access and reduce development
impact on nocturnal environments”.
Occupant Behavior
MIT 4.430 Daylighting, Instructor C
Reinhart
4
5
Occupant Behavior
100
blinds al
ways up
daylight autonomy [%]
80
US
60
40
ER
B
blin
ds
alw
ays
20
0
EH
A
d ow
VIO
UR
?
n, s
lats
!
at 4
5o
blinds always fully closed
0.5
1
1.5
2
2.5
3.3
4.5
distance to facade [m]
5.5
6.5
7.5
Monitoring User Behavior
architecture: Meier-Weinbrenner-Single, Nürtingen
Paper: Reinhart C F, Voss K, “Monitoring manual control of electric lighting and
blinds.” Lighting Research & Technology, 35:3 pp. 243-260, 2003.
MIT 4.430 Daylighting, Instructor C
Reinhart
5
6
Monitoring Setup in the Offices
Illuminance
Temperature
occupancy
HOBO data logger
Monitoring Blind Usage
video surveillance camera
receiver
2414.5 MHz
Blind setting
MIT 4.430 Daylighting, Instructor C
Reinhart
data acquisition
EIB system
6
7
Example Picture
MIT 4.430 Daylighting, Instructor C
Reinhart
7
8
switch-on probability at arrival
Switch-On Probability (I)
11
type 1
0.75
0.75
0.5
0.5
0.25
0.25
type 2
00
00
100
100
200
200
300
400
500
minimum work plane illuminance [lux]
minimum
600
Jim Lo
Lov
ve, Univ
Universit
ersity
y of Calgary
Switch-On Probability (II)
switch-on probability at arrival
1
Hunt 1978
Lamparter 2000
0.75
0.5
0.25
0
0
MIT 4.430 Daylighting, Instructor C
Reinhart
100
200
300
400
500
minimum work plane illuminance [lux]
8
600
9
People are Consistent but Different
switch-on probability at arrival
1
0.8
0.6
0.4
0.2
0
0
100
200
300
400
minimum work plane illuminance [lux]
500
switch-offprobability
probabilitywhen
when
leaving
switch-off
probability
when
leaving
switch-off
probability
leaving
switch-off
when
leaving
Switch-Off Probability
11
1
0.75
0.75
0.75
no controls
no controls
dimmed
system
no controls
occupancy
occupancysensor
sensor
0.5
0.5
0.5
0.5
0.25
0.25
0.25
0.25
000
0
minutes
1-2
hours
4-12
hours
hours
<30minutes
minutes
1-2
hours
<30
1-2
hours
4-12
hours
24+
hours
4-12
hours
24+24+
hours
<30
minutes
1-2
hours
4-12
hours
24+
hours
30-59
minutes
2-4
hours
12-24
hours
30-59
minutes
2-4
hours
12-24
hours
30-59
minutes
2-4
hours
12-24
hours
30-59 minutes
2-4 hours
12-24 hours
Pigg et al. University of Wisconsin
behavioral patterns change in the presence of
automated controls
MIT 4.430 Daylighting, Instructor C
Reinhart
9
10
intermediate switch-on probability
Intermediate Switch-On Probability
0.03
0.02
0.01
0
0
100
200
300
400
500
600
700
minimum work plane illuminance [lux]
Manual Lighting Control Algorithm
blinds occlusion [%]
100
80
Lamparter
60
40
20
Inoue
0
0
1
2
solar penetration depth [m]
3
 Blinds get lowered to avoid direct sunlight falling on the
work plane.
MIT 4.430 Daylighting, Instructor C
Reinhart
10
11
Model Overview
Lightswitch 2002
annual occupancy
profiles
annual illuminance
profiles
Lightswitch Algorithm
(stochastic)
el. lighting/blinds profile
Paper: Reinhart C F, "Lightswitch 2002: A model for manual control of electric lighting
and blinds", Solar Energy, 77:1 pp. 15-28, 2004
Lightswitch - Manual Lighting Control
NO
NO
Does occupant arrive?
YES
NO
stochasticprocess:
process:
stochastic
switch
switchon
onprobability
probability
(F ig. 7-5)
NO
YES
Are lights switched on?
Does occupant arr ive?
1
switch-on probability
NO
YES
Is work place
already occupied?
YES
Does occupan t leav e?
Are lights switched on?
YES
NO
YES
Does occupant leave?
Hunt
0.8
Lamparter
NO
0.6
NO
Is there an
occupancy sensor?
0.4
NO
YES
stochastic proces s:
intermediate sw itch on
probability
(F ig. 7 - 7)
NO
stochastic process:
Pigg’s switch off probability
with or without occupancy
sensor (Fig. 7-5)
YES
0.2
0
NO0
YES
100room been
200
300
400
Has the
minimum
w ork plane illumiance [lux]
deserted for longer
than sensor
delay time?
NO
500
600
YES
YES
YES
Switch lights on
Switch lights off
MIT 4.430 Daylighting, Instructor C
Reinhart
Switch lights on
11
Switch lights off
12
Lightswitch - Manual Blind Control
))))))))
))))
Lightswitch - Manual Blind Control
window with
blinds
work plane
sensors
 Define work plane sensors that define where the occupants are usually located.
 Associate sensors with shading groups. A shading group consists of a (set of) blinds
that are opened and lowered at the same time.
 Check when direct sunlight (>50Wm-2) is incident on a work plan sensor.
 Close shading device if yes until occupant is away for more than an hour.
MIT 4.430 Daylighting, Instructor C
Reinhart
12
13
Blind Use in New York City Classrooms
Question 14: How often do you adjust the shading device(s)?
MDesS thesis, Jennifer Sze 2009
5%
8%
17%
18%
31%
21%
Multiple times a day
Once or twice a week
Once or twice a month
Never
Other
N/A
Image by MIT OpenCourseWare.
 183 teacher surveys, 9 participating schools
DIVA Demo: Occupant Behavior
MIT 4.430 Daylighting, Instructor C
Reinhart
13
14
Modeling Occupant Behavior
No Blinds
With Blinds
Detecting Glare
MIT 4.430 Daylighting, Instructor C
Reinhart
14
15
What is glare?
 Glare is a subjective human sensation that describes ‘light within the
field of vision that is brighter than the brightness to which the eyes are
adapted’ (HarperCollins 2002).
Illustration by ZStardust on Wikimedia Commons.
Glare Indices
 A glare index is a numerical
evaluation of high dynamic range
images using a mathematical formula
that has been derived from human
subject studies.
UGR  8 log10
0.25 n L2s  s

Lb i 1 P 2
CIE Unified Glare Rating
MIT 4.430 Daylighting, Instructor C
Reinhart
 Example indices include the unified
glare rating (UGR) and the daylight
glare index (DGI). All of these
equations
were
derived
from
experiments
with
artificial
glare
sources – none of them under real
daylight conditions.
The reason for this is that until
recently it has been next to
impossible to collect high dynamic
range images of daylit scenes under
continuously changing lighting levels.
15
16
Daylight Glare Probability (DGP)
Room 1: CCD Camera
Room 2: Human Subject
 DGP is a recently proposed
discomfort
glare
index that was derived by
Courtesy
of Elsevier. Used
with permission.
Wienold and Christoffersen from laboratory studies in daylit spaces using 72
 test
Weak
correlation
DGI and CGI (cventional galre
subjects
in Denmarkbetween
and Germany.
 Two identical,
side-by-sideevaluations.
test rooms were used. In Room 1 a CCD camera
metrics)
and occupant
based luminance mapping technology was installed at the exact position and
orientation as the head of the human subject in Room 2.
Paper: Wienold & Christoffersen, " Evaluation methods and development of a new glare prediction model for daylight
environments with the use of CCD cameras ", Energy & Buildings 2006.
Image-based Glare Source Detection using
the Radiance evalglare Program
Paper: Wienold & Christoffersen, " Evaluation methods and development of a new glare prediction model for daylight
environments with the use of CCD cameras ", Energy & Buildings 2006.
MIT 4.430 Daylighting, Instructor C
Reinhart
17
18
In Search of a Glare Metric…
Courtesy of Elsevier. Used with permission.
 Weak correlation between DGI and CGI (conventional glare
metrics) and occupant evaluations.
Paper: Wienold & Christoffersen, " Evaluation methods and development of a new glare prediction model for daylight
environments with the use of CCD cameras ", Energy & Buildings 2006.
100%
100%
Percentage of
disturbed persons
Percentage of
disturbed persons
In Search of a Glare Metric…
80%
60%
40%
20%
0%
2
R =0.12
0
2000
4000
6000
80%
60%
40%
20%
8000
0%
2
R =0.56
0
10
Window Luminance [cd/m2]
40
50
60
DGI
+ Standard deviation
-
100%
100%
Percentage of
disturbed persons
Percentage of
disturbed persons
30
Daylight glare index DGI
Window Luminance
+ Standard deviation of DGP
-
80%
60%
40%
20%
0%
20
2
R =0.56
0
10
20
30
40
50
CGI
Total responses: 349
Number of responses per
illuminance- class:29
80%
60%
40%
20%
0%
2
R =0.77
0
2000
4000
6000
8000
10000
Vertical Eye Illuminance [lux]
CGI
+ Standard deviation
-
Vertical Eye Illuminance
+ Standard deviation
-
Image by MIT OpenCourseWare.
 Vertical eye illuminance promising for the first term.
MIT 4.430 Daylighting, Instructor C
Reinhart
17
18
Final Daylight Glare Probability Metric
100%
Total responses: 349
Number of responses per DGP - class: 29
Probability of disturbed
persons
90%
80%
70%
60%
50%
40%
30%
20%
R2 = 0.94
10%
0%
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
Daylight glare probability DGP
DGP
+
_ Standard deviation of DGP
+
_ Standard deviation of binomial deviation
Image by MIT OpenCourseWare.
Paper: Wienold & Christoffersen, " Evaluation methods and development of a new glare
prediction model for daylight environments with the use of CCD cameras ", Energy &
Buildings 2006.
DGP allows users to go back and forth between
simulation and reality through HDR photography
HDR Image (Digital Camera)
HDR Image (Radiance)
evalglare program
instantaneous daylight glare probability
MIT 4.430 Daylighting, Instructor C
Reinhart
18
19
DGP Comfort Ranges
DGP < 35%
imperceptible
35% < DPP < 40%
perceptible
40% <DGP < 45%
disturbing
DGP > 45%
intolerable
Example DGP Calculation
MIT 4.430 Daylighting, Instructor C
Reinhart
19
20
DIVA Demo: DGP Point in Time Calculation
Towards Annual Glare Calculations
Generated visualizations for illuminance calculation removed due to copyright restrictions.
 Wienold developed a process through which the annual glare calculation is
split into a regular Daysim illuminance calculation and an ab=0 contrast image.
Paper: J Wienold, “Dynamic Daylight Glare Evaluation”, Building Simulation 2009,
Glasgow Scotland.
MIT 4.430 Daylighting, Instructor C
Reinhart
20
21
Comparison of Simplified DGP and hour -byhour method
Graph of vertical illuminance removed due to copyright restrictions.
Paper: J Wienold, “Dynamic Daylight Glare Evaluation”, Building Simulation 2009,
Glasgow Scotland, 2009.
Annual Glare Map
MIT 4.430 Daylighting, Instructor C
Reinhart
21
22
Number of working hours
above DGP
DGP Comfort Ranges
1500
DGPintolerable = 39%
1200
1012 hours
900
600
300
0
30
35
40
45
50
55
60
Daylight glare probability [%]
Imperceptible
Disturbing
Perceptible
Intolerable
Image by MIT OpenCourseWare.
Paper: C F Reinhart, Simulation-based Daylight Performance Predictions", Book
chapter in Building Performance Simulation for Design and Operation, Editors J Hensen
and R Lamberts, Taylor & Francis, 2011.
100
No blinds
1500
No blinds
(avoid direct sunlight)
1200
900
Active blind use
(avoid discomfort glare)
600
Blinds always
300
0
20
lowered
25
30
35
40
45
50
55
60
Disturbing
Perceptible
Intolerable
Active blind use
(avoid discomfort
glare)
60
40
Blinds always
lowered
20
0
0
1
2
3
4
5
6
7
Distance to facade [m]
Daylight glare probability [%]
Imperceptible
Active blind use
(avoid direct sunlight)
80
Active blind use
Daylight Autonomy [%]
Number of working hours with DGP above
Expanded Blind Control Model:
Close Blinds when DGP > 40%
Image by MIT OpenCourseWare.
22
The formula makes sense. How plausible are
DGP results compared to other glare indices?
Multidirectional Time-Lapse Simulation
The image shows a cylindrical 360o view of a work space in Gund Hall. The color
coded lines at the bottom show the predictions of different glare indices (DGP, DGI,
UGI, CGI and VCP) whether discomfort glare will be experienced in a particular
direction at different times of the day (Green=Imperceptible Glare;
Yellow=Perceptible Glare; Orange=Disturbing Glare; Red=Intolerable
Glare).
Design: Jeff
Niema
Paper
J A Jakubiec, C F Reinhart, “The Use of Glare Metrics in the Design of Daylit Spaces: Recommendations for Practice", submittedsz
to Lighting
Research and Technology 2011.
MIT 4.430 Daylighting, Instructor C
Reinhart
23
24
Multidirectional Time-Lapse Simulation
 DGP yields most plausible results in these spaces.
Design: Jeff
Niema
Paper
sz
J A Jakubiec, C F Reinhart, “The Use of Glare Metrics in the Design of Daylit Spaces: Recommendations for Practice",
submitted to Lighting Research and Technology 2011.
How to analyze for visual discomfort?
Paper: J A Jakubiec, C F Reinhart, 2010, “The Use of Glare Metrics in the Design of Daylit
Spaces: Recommendations for Practice", Lighting Research and Technology, 2011.
MIT 4.430 Daylighting, Instructor C
Reinhart
24
25
Multidirectional Time-Lapse Simulation
 DGP yields most plausible results in these spaces.
Design: Jeff
Niema
Paper
sz
J A Jakubiec, C F Reinhart, “The Use of Glare Metrics in the Design of Daylit Spaces: Recommendations for Practice",
submitted to Lighting Research and Technology 2011.
Concept of the Adaptive Zone
Annual DG Calculation: Fixed view looking forward
Annual DG Calculation: +/‐ 45 degrees rotational freedom
 The concept helps to quantify the benefits of flexible furniture
settings etc.
Design: Jeff
Niema
Paper
sz
J A Jakubiec, C F Reinhart, “The Use of Glare Metrics in the Design of Daylit Spaces: Recommendations for Practice",
submitted to Lighting Research and Technology 2011.
MIT 4.430 Daylighting, Instructor C
Reinhart
25
26
The Ultimate Adaptive Space
Christoph Reinhart
Associate Professor
MIT 4.430 Daylighting, Instructor C
Reinhart
email: creinhart@mit.edu
26
27
MIT OpenCourseWare
http://ocw.mit.edu
4.430 Daylighting
Spring 2012
For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.