mind. This basic configuration will

mind. This basic configuration will be considered the
basecase.
The second system has the same characteristics as the
first one, except that a lightshelf of highly reflective
material (r ¼ 0:80) has been placed according to the
dimensions of Fig. 4(b) in the view window area. The
third system is presented in Fig. 4(c); for this case
the device is an anidolic concentrator [14].
3.2. Testing conditions
The systems were modelled using Desktop Radiance,
and added to a prototype office of dimensions 800
(width)
1200 (depth)
270 (height) cm with the reference
working plane at 80 cm. The prototype is supposed
to be in the third floor of an office building placed at a
height of about 10m over the ground which has a
reflectance value r ¼ 0:20. The location studied is Tel
Aviv (32N 35E, standard meridian 30E) due to the high
concentration of office buildings found there.
Simulation dates presented here are 21 June and 21
December even though the experiment also included 21
March and September. Testing hours are 10, 12, 14 and
16 h. The four main orientations are tested, with skytype
CIE intermediate with sun, as mentioned in the
previous section.
This study aims at integrating evaluations for light
quality and quantity, since presenting only illuminance
data can be misleading if human factors are not
incorporated. It is a known fact that users will react to
correct the amount of light entering a space in order to
achieve visual comfort [15]. A first standard approach in
the simulations to potential human behaviour is the use
of automatic horizontal venetian blinds. They are
assumed to be lowered when the Radiance simulation
picture shows direct solar penetration (sunspot) over the
office’s floor beyond 1m from the view window. This
blinds system does not cover the clerestory part of the
devices.
A second approach to the qualitative issue of human
visual comfort is made through glare analysis. Glare
from large natural sources needs more research to
understand it; but a reasonable assessment can be given
with what is known today. For this study, the daylight
glare index (DGI) is used. This index was derived from
research for glare produced by artificial sources, but
tries to reconcile the fact that users can stand better
glare from natural sources [16].
Radiance provides routines to evaluate the luminance
of a given scene and assigns to it different glare indexes
as requested by the user. The evaluation of electronic
pictures to determine the presence of glare, through
methods unrelated to Radiance, has also been proposed
by Schiler to analyse through digitized video images,
real-life conditions in test offices and is based on
statistical analysis of pixel luminances [17]. For the
simulations of this study, the view point is 100 cm from
the window and faces it directly. It is an unfavourable
position, yet this proved to be a limiting condition from
presently accepted glare formulas.
3.3. Results and analysis
The results of the simulations are presented in Figs.
5–7. Even though hundreds of parametric simulations
were done in the original study, here the most
representative data of south orientation are shown.
The minimum acceptable horizontal illuminance for the
workplane is 300 lx, as required by Israeli standards [18],
while the upper limit illuminance is 4000 lx [19].
Fig. 5 presents year-round histograms for the
compared systems (basecase with and without blinds,
lightshelf and anidolic concentrator) taken at 200, 600
and 1200 cm from the view window. The histogram at
200 cm has its range of data from 0 to +4501 lx, while
those at 600 and 1200 cm range from 0 to +901 lx. From
them, it can be seen that the effectiveness of the systems
decreases significantly after 6m from the view window,
yet only the anidolic concentrator could keep illuminance
levels above the minimum limit at 1200 cm.
Fig. 6 presents illuminance line graphs for 21
December and 21 June for south orientation. Dashed
lines indicate the use of blinds; the window is presented
on the left-hand side of every graph. In them it can be
seen that the proposed geometry of the devices
accomplishes the goals for highly luminous climates of
providing shading at the front part of the room to
reduce the contrast between front and back of the room
(therefore, lowering the chances for glare). They also
confirm that the effect of the devices decreases after
600 cm. The sudden peaks observed are ‘‘sunspots’’ that
in reality would be less apparent.
Qualitative analysis is given through glare graphs in
Fig. 7, which also shows Radiance-generated pictures
for the basecase without blinds. Parametric simulations
were done for glare, but only few produced conclusive
results. In the graphs the view angle is on the x-axis, and
the DGI index on the y-axis. The grey area indicates the
values where DGI is acceptable. At noon the glare
pattern is nearly ‘‘symmetric’’ and at 16 h ‘‘asymmetric’’,
due to the asymmetric luminance distribution of the sky.
This could also be seen on the view inside the room,
depending on solar penetration (striking the floor or one
of the walls, respectively).
3.4. Discussion
This study focuses not on proposing novel systems
but on evaluating their potential application to highly
luminous climates. The quantitative evaluation in terms
of illuminance shows that the anidolic concentrator

ผลลัพธ์ (ไทย) 3:

Of illuminance shows that the anidolic concentrator.