INTRODUCTION
Atria have become popular factors in Architecture because they
provide several simultaneous values. They resolve thermal problems
in exterior walls. In a ddition to their characteristics in cooling
a space, they act like private spaces isolated from the exterior
noise. Daylighting, however, represents one of the esthetic values
of an atrium. Due to a lack of good prediction tools, today's
atria are overlit with excessive solar gains.
Most daylighting calculations do not work well for atrium buildings. At present, the most sophisticated method used is still the physical scale model (Atif, Boyer & Degelman, 1994). Physical models are used in this research to measure the illumination level inside an atrium and in the adjacent spaces of an atrium building of four floors.
BACKGROUND
The aim of this study is to provide insight on the relationship
between light and form. For example, in some cases light is less
desirable and heat gain is extremely undesirable. Therefore, the
form of that space is different than when much light is needed.
In spaces adjacent to atria, the study of the behavior of light
is more interesting than when there is sidelighting opened directly
to outside because several variables play major factors as to
how the light behaves. There are direct sunlight, diffuse light
from the sky, reflected light from the ground and the walls of
the atrium, and reflected light from the walls, the ceiling and
the ground of the adjacent spaces.
DESCRIPTION OF METHOD
A physical model of a scale of 1"=1'-0" (1:12) was used
to simulate the daylight distribution inside the atrium and the
adjacent spaces. An atrium of 20' width, 20' length, and 40' depth
was modeled, (four-story) atrium. Shallower atria and courtyards
have been modelled in other studies, but for the thermal considerations
in hot arid climates, deeper atria are more appropriate. Adjacent
spaces were modeled, as well, of 20' length, 20' width, and 10'
height The critical issue in such situations is not the light
in the atrium (which is usually sufficient), but the light available
within the adjacent spaces. The physical model was built of foamcore
panels which was covered with black sheets from the back to avoid
any light leak.
Reflectance materials for the atrium were used as follows:
The reflectance of the ground = 30%
The net reflectance of the walls = 50%
Reflectance materials for the adjacent spaces were used as follows:
The net reflectance of the ground = 30%
The net reflectance of the walls = 50%
The net reflectance of the ceiling = 70%
Since one side of the atrium was modeled with adjacent spaces,
the other three sides of the atrium should simiulate adjacent
spaces, as well. Therefore, clear specular surfaces were added
to those walls wherever windows would have been located, and an
absorptive surface were placed behind these window surfaces. This
approximated the absorptance of a full space behind the windows,
but still allowed the specularly reflected component to be included
whenever sun angles caused excessive reflection from the glass
surfaces. Since there might actually be some slight contribution
from the adjacent space back to the atrium, this was a conservative
approximation. The difference becomes negligible (Al-Turki, 1994).
Actual wall surfaces were the same on all four sides. A glazing
area of 40% for all four floors was used (California Energy Commission,
1992).
A grid pattern of one foot by one foot was drawn on the ground
floor of the model to mark at which point measurements were to
be taken. One of the four panels is removable to facilitate the
process of placing the measuring devices.
The physical model was fixed to a base made of plywood with a
small 6" x 6" piece of plywood attached to the bottom
of the base. The small piece of plywood was attached to a tripod
to facilitate orienting the model in all directions (Al-Turki,
1994).
Licor LI-210-S illuminance meters, which are designed to measure
the illumination in foot-candles, were connected to a portable
computer to read the illumination level from each licor using
a computer program called "DATALIT", which was written
at University of California, Los Angeles (UCLA), (Milne, 1986).
MEASURING THE RESULTS
4.1. The model was tested under exclusively clear sky conditions
in Los Angeles, California, 34 N Latitude, where the tests took
place.
4.2. Seven reference points were carefully selected and distributed on the ground floor of the atrium (Figure 1).
Figure 1.Placing light sensors inside the atrium of 20' x 20'
x 40' to measure the illumination level from the physical model.
4.3. Eighteen reference points were selected at the workplane level (3 feet) and located inside the adjacent spaces of 20' x 20' x 10' as follows (Figure 2):
4.3.1. Two rows of six sensors each along the sides of the adjacent spaces with three feet distance between every two sensors, and one foot from the side walls.
4.3.2. One row of six sensors at the center of each space and
along the depth with three feet distance between each two sensors
(Figure 2.)
4.4. One light sensor was located completely outside the atrium to measure the total horizontal illumination.
Figure 2. Locating light sensors inside the adjacent spaces of 20' x 20' x 10' of a four-floor atrium building to measure the illumination level in all orientations on 21 of each month.
4.5. The illumination level in the adjacent spaces was measured
at the hours 9:00 a.m., noon, and 3:00 p.m. on the 21 of each
month. On the other hand, the illumination level inside the atrium
was measured every hour from 9:00 a.m. to 3:00 p.m. using the
sundial diagram for the same latitude, and using the tilting device
mentioned above.
4.6. The adjacent spaces were photographed using a 28 mm camera,
and ASA 400 daylight film to provide a permanent record of daylighting
conditions inside the model (Schiler, 1990).
ANALYSIS
5.1. The Atrium
Preliminary data indicated that the primary variation of the Daylight Factor (DF) for every reference point inside the atrium occurred due to the change in time and month. (See figures 4 and 5.)
The geometry of the atrium was plotted showing the DF for each
reference point. The DF was calculated by measuring the total
horizontal illumination and the indoor illumination.
5.2. The Adjacent Spaces
All the output data was plotted in graphs showing the geometry of each space. The x-axis and the y-axis show the geometry (width and length), and the z-axis shows the DF in that space at a particular time of the year.
Figure 3. South facing wall of the adjacent spaces showing the
illumination level for the four floors in June 21 at noon. Each
graph shows the geometry of each space. Figure 3.A shows the fourth
floor, figure 3.B the third floor, figure 3.C the second floor,
and figure 3.D the first floor. The DF at the points two feet
from the rear wall (2.3%) is higher than DF five feet from the
rear wall (2.1% ) (Figure 3.D).
CONCLUSIONS
For this atrium with a ratio of 1:1:2, we observed several behaviors
which were expected, and further observed consistent ratios beyond
what we had expected.
6.1. Expected behaviors included:
6.1.1. The adjacent spaces for the South facing wall received
sufficient quality and quantity of light in all four floors, both
summer and winter (Figure 3), (Al-Turki, 1994).
6.1.2. The adjacent spaces for the North facing wall received
good quality and quantity of light, mostly reflected from the
opposite wall, only at Summer noon for all floors (Al-Turki, 1994).
6.1.3. The East and West facing adjacent spaces act exactly the
same at noon, in terms of quality and quantity of light. They
act exactly the opposite at 9:00 a.m. and 3:00 p.m. (Al-Turki,
1994).
6.2. Unexpected behaviors included:
6.2.1. At 9:00 a.m. in winter, the floor of the atrium on the
East side receives more light than the floor of the atrium on
the West side. This is counter intuitive. The direct beam sunlight
is striking the West wall. However, most of the light received
on the atrium floor at that particular time is reflected light
bouncing from the upper portion of the opposite wall. The same
observation was found for the West side of the atrium at 3:00
p.m. (Figure 4), (Al-Turki, 1994). This does not appear to be
true for the corners.
Figure 4. The floor of the atrium on the East side receives more light (6% DF) than the floor on the West side (4% DF) on December 21 at 9:00 a.m. because of the low altitude angle of the sun in December. The same observation does not occur in July or March, when there are higher sun angles.
Figure 5. The floor of the atrium on the South side receives more
light (7% DF) than the floor of the atrium on the North side (5%
DF) on December 21 at noon. The low altitude angle causes that.
However, on July 21 the direct sun beams hit the sensors that
are one foot from the North wall.
6.2.2. At noon, in winter, the South side of the atrium the (North
facing wall) received more illumination than the North side for
the same reason as mentioned in 6.2.1 (Figure 5), (Al-Turki, 1994).
6.2.3. Unlike all other measurement methods, for interior adjacent spaces, the physical model showed that the points immediately adjacent to the rear wall of the space (two feet from the rear wall) received more light than the point five feet from the rear wall, which is the traditional U.S. measurement station. This is from the light reflected from the rear wall (=50%) in a sidelighted condition (Figures 3, 6 & 7), (Al-Turki, 1994).
Figure 6. Daylighting Factor (DF) curve using the lumen method calculations. Notice that only three reference points are calculated (MAX, MID & MIN), where the MIN point is five feet from the rear wall.
Figure 7. Daylight Factor (DF) measurements, which obtained from the tested physical model, proved that the points five feet from the rear wall are not the lowest illumination. Points two feet from the rear wall are lower.
RECOMMENDATIONS
7.1. Graphing the daylight factor in the spaces adjacent to atria
shows that some spaces do not get sufficient illumination level
and part of these spaces need artificial lighting. From an architectural
perspective, there is a strong relationship between light and
form in a space; therefore, the adjacent spaces could be formed
upon the available functional light in each floor (see Al-Turki,
1994 for suggested form modifications). This allows us to arrange
the function in a space without using any artificial light which
resultantly saves a lot of energy.
7.2. To improve the quantity of light in the first and second
floors in winter, the glazing area could be increased. At the
same time, the glazing area in the third and fourth floors could
be decreased. (Figure 3 shows sufficient light on the fourth floor
in June, and the fourth floor receives an increase from low angle
sun in December, see Al-Turki, 1994). However, the total glazing
area for all floors should remain the same so that the function
of the cooling and heating systems will remain the same (Al-Turki,
1994).
7.3. Another way of increasing the quantity of light is using
light reflectors to supply more light to the bottom two floors.
The reflectors could be of any high reflectance and specular materials.
They should be tilted toward the areas which require more light.
The depth of the reflectors should be carefully chosen so that
they do not block the light coming to the other sides. (Al-Turki,
1994).