Available online at www.sciencedirect.

com

ScienceDirect
Solar Energy 110 (2014) 139150
www.elsevier.com/locate/solener

Calculation of the shadow-penumbra relation and its application

on ecient architectural design
Jorge Hernan Salazar Trujillo
Universidad Nacional de Colombia, Faculty of Architecture, Calle 59A No 63 20, Medelln, Colombia
Received 15 January 2014; received in revised form 21 July 2014; accepted 30 August 2014
Available online 29 September 2014
Communicated by: Associate Editor Mario Medina

Abstract
Shadow Dispersion is the eect by which any opaque objects shadow progressively becomes penumbra. This eect originates from a
partial obstruction of the visibility of the solar disk. It allows diminishing solar gain in places where there is high radiation intensity,
facilitating visual ergonomics and energy eciency. Although architecture in the tropics oers a wide array of strategies for creating penumbras, i.e., meshes, lattices, architectural fabrics, openwork walls and pergolas, there is no method for its design. Solar architecture
literature simplies the shadow projection phenomenon and always assumes sunrays as being parallel, but penumbra calculation does
not allow for this simplication. In order to bridge this gap, calculating equations are deduced here and the tables needed to appraise
suitable architectural areas not to block but soften sunlight are included. This paper denes the penumbra rate and shows its application
on a building built in Medellin in 2006, designed for housing exhibitions of orchids, which depend on penumbra to survive. The work
concludes by outlining the future possibilities of incorporating penumbra zones into the architectural design process. Other applications
of the method are also mentioned.
2014 Elsevier Ltd. All rights reserved.

Keywords: Penumbra; Gloom; Shadow; Shading device; Solar control; Shadow Dispersion

1. Introduction
In the tropics, permanent and moderate climatic
conditions enable the construction of open buildings,
highly permeable and able to establish tight bonds with
the place they are in Mesa (2013). Shadow is more signicant in this part of the planet than in temperate zones due
Abbreviations: h, elements height; ld, limit distance; f, solar disk viewing
angle; hz, zenith angle; e, solar elevation angle; d, penumbra zone width; x,
elements size; lh, limit height; lx, limit size; PR, penumbra rate; s,
elements projection perpendicular to the solar rays.
Tel.: +57 4 4449483; fax: +57 4 2604875.
E-mail address: jhsalaza@unal.edu.co
http://dx.doi.org/10.1016/j.solener.2014.08.043
0038-092X/ 2014 Elsevier Ltd. All rights reserved.

to the lack of seasons, and because there is a higher intensity of incident solar radiation (Tzonis et al., 2001). Therefore, for many buildings and outside spaces located in the
tropics, to have enough shadow is sucient to oer human
comfort conditions for many hours a day.
The balance between sunlight and shadow is a fundamental part in building design due to its strong eect on
energy eciency and thermal comfort. In higher scale projects such as urban planning, roads, squares and parks
design, this is also a very important issue. In outdoor spaces
and in any public space project, the energy eciency aspect
is less important since there is usually no air conditioning.
Nevertheless, thermal comfort and visual ergonomics

140

J.H. Salazar Trujillo / Solar Energy 110 (2014) 139150

precede the list of priorities to be fullled (Higueras, 2013).

In turn, in a smaller scale objects design, it is essential to
take shadow calculation into account, which is a prerequisite for the correct dimensioning and distribution of any
shading device (Olgyay, 1968; Salazar, n.d.).
The methods and instruments employed to calculate
shadows are widely used and spread within the professional
environment. In these methods, it is assumed that sunlight
rays have an impact on the terrestrial surface as a set of
parallel rays. The results obtained from these calculations
suggest a drastic change between the shading zone and
the sunny zone, with a clear limit between both regions.
This is something that is never observed in real shadows,
as it can be easily proven. It suces to expose any opaque
object to the sunlight to verify that any shadow is smaller
than the object that projects it, and that the shading region
edge called penumbra zone is clearer as the distance
between the object and its shadow increases. None of those
phenomena are taken into account when shadow calculations are carried out with analysis methodologies in which
sunrays are considered parallel.
A false premise of parallelism does not always lead to an
imprecision with eective consequences on the designed
objects solar performance. If it is a big size object, the
imprecision of not considering the penumbra is unseen
since the shading area is proportionally bigger. Nevertheless, for those cases in which architectural decisions tend
to design elements whose dimension is just of a few centimetres, it is essential to consider sunrays as not being
parallel.
Solar disk viewing angle and solar aureole are important
concepts to improve solar concentration systems. Solar
tracking, mirror design and absorber plane optimization
cannot be done by applying the parallel solar vector
assumption. The solar disk viewing angle has variations during the year, but these variations are not signicant and if it
is considered as a constant value, the accepted limit of the
solar disk is 4.65 mrad (Buie et al., 2003a). The solar aureole
presents higher variations originated in the scattering of the
solar beam through the earths atmosphere depending on
local climate variations (Buie et al., 2003b). The radiant ux
contained within the circumsolar region of the sky, including solar disk and solar aureole, is an important issue dening a pyrheliometers acceptance angle. A 5 eld of view is
dened by the ISO-9060 to capture the sun and part of the
circumsolar region because the total amount of direct radiant ux is found within this angle.
There are subtle evidences of the sunshape distribution
variation on the shadow casting phenomena within architectural elements: during clearer days the shadows will be
sharper and the penumbra zone narrower. In foggy days
and during mornings in warm-humid places, the penumbra
zone becomes wider and the shadow denition becomes
less evident. The 5 magnitude does not have a particular
signicance in architectural design because the extent of
the solar disk provides between 98% (extremely clear skies)
and 90% of the direct insolation (Buie and Monger, 2004).

From the designers point of view, both conditions can be

considered equivalent because the critical design condition
arises when the sky is clearer, less energy comes from the
aureole and the shadows have a better focus.
Database searching (including Science Direct, One File,
Science Citation Index, Expanded Academic ASAP and
Discovery Service) using penumbra, shadow and architecture as keywords did not nd any relevant paper during
the retrieval made in July 2014. Publications about penumbra in architecture refer to a characteristic of spaces lightened with diused natural light. In computer graphic
research the term soft shadow is also used, and describes
computer techniques to increase realism in rendering algorithms (Kolivand and Shahrizal Sunar, 2013). Classic
books on Bioclimatic Architecture, in which there is always
a chapter completely dedicated to the principles of Solar
Geometry, do not mention the shadow-penumbra relation
either. Arzoumanian (1988), Baruch (1976), Koenigsberger
et al. (1977), Lippsmeier (1969), Lan (1986), Olgyay (1968),
Puppo and Puppo (1972), Szokolay (1977) and Yanez
(1988) works not only do not have the concept of penumbra, but also do not include any discussion about the consequences of assuming sunrays parallelism when making
solar control calculations.
The consequences of this gap are particularly important
in sunlightning studies carried out for the tropics. Meshes,
lattices, architectural fabrics, openwork walls and pergolas,
are part of the formal tradition and repertoire for buildings
in this part of the planet. The architectural enclosures made
with these elements can form surfaces of many square
meters, but they are often composed by fairly small-dimension elements repeatedly installed. They should not be
designed without taking into account the penumbra eects
they produce.
A successful design process is based on analysis and representation methods in which simplications of the applied
model do not distort the essence of the studied phenomenon. To calculate shadows from a large list of architectural
elements, assuming sunrays as parallel, does not have practical consequences or prevent making good design decisions. On the contrary, if rays are considered as divergent,
this would lead to an unnecessary complexity in the calculations since the conclusions would be ultimately the same as
those from using simpler methodologies.
At the detail scale, things are very dierent. For small
size objects, the proportion between shadow and penumbra
reaches a balance. In such cases, assuming sunrays parallelism implies ignoring an important phenomenon called
Shadow Dispersion, which can have a major role in architecture for the tropics, and which oers great possibilities
in the creation of energy ecient and visually comfortable
spaces.
2. Objective
To incorporate penumbra zones into the architectural
design process by developing the technical basis, deducing

J.H. Salazar Trujillo / Solar Energy 110 (2014) 139150

the mathematical functions and creating the tables needed

to calculate architectural devices in which penumbra eect
is more accentuated than shadow eect.
3. Shadow Dispersion
Every shadow is the result of the visibility restriction
established between a point and a radiation source. In
the hypothetical case where the source is a point, there will
be only a straight line linking the source with the studied
point and, therefore, it would be impossible for a partial
visibility of such point to occur. This case represents the
only possibility to have a shadow without having penumbra at the same time.
In the real world, any radiation source is extensive and,
therefore, there is always the possibility to have full or partial restriction of visibility. In the rst instance, when all the
possible straight lines drawn between the source and the
analysed point are intercepted by an opaque object, shadow conditions are given. In the second instance, when
only some of the straight lines are intersected, an incomplete visibility of the radiation source is given and, as a
result, a reduction of incident energy in the analysed point
is noticed.
Every opaque object projects a shadow with an area
close to its own size only when it stands on the shadows
projection plane. As the distance between the object and
the plane gets longer, the shading zone area gets smaller
and the penumbra zone gets bigger until it reaches a distance named limit distance ld where the shadow completely
disappears (Fig. 1). When that occurs, there is not a single
point on the surface experiencing a total obstruction of the
solar disk.
In the interval between distance zero (h = 0) and the limit
distance (h = ld) a progressive replacement of the shadow by
the penumbra takes place. This is perceived as a lack of
focus in the shadows edges and an evident solar shine
reduction inside the edges. Proportion variation between

141

both zones is a result of Shadow Dispersion. This is a phenomenon by which a shadow is blurred and extended
until it totally becomes penumbra. It is called this way
due to the fact that, although the quantity of energy blocking the object remains constant, its shadow spreads over an
area bigger than the one occupied by the shading zone.
Once the limit distance is exceeded (h > ld), recognising
in the shadow the objects edges which block the sunrays
becomes impossible. Besides a decrease in solar radiation
intensity under the shadowed area is veried. That is the
reason why penumbra is so appreciated in the tropics; it
constitutes an architectural procedure which allows the
reduction of solar radiation intensity where this is not only
abundant but often annoying because of such abundance
(Fig. 2).
A balanced distribution of both sunny and shading
zones makes part of the basic techniques to increase energy
eciency of an architectural enclosure. In this way, enclosures with a suitable energy performance can be achieved
from both the visual ergonomics and thermal comfort perspectives (OBrien et al., n.d.). However, this technique is
not completely satisfactory since shading and sunny zones
alternation could be unpleasant for carrying out certain
kinds of activities, especially those which are highly-visually demanding. The possibility of preventing this visually
uncomfortable condition by means of the suitable designing and dimensioning of the penumbras projection devices
is what allows obtaining equivalent prots in terms of
energy, but with a more moderate and convenient distribution for human sight.
Incorporating penumbra zones into the architectural
design process entails that the shape, position, and size of
the shading devices composing elements be designed based
on the Shadow Dispersion eect application. This phenomenon is the function of three variables: (1) the solar disk
viewing angle; (2) the size and position of opaque objects
projecting the shadow; and (3) the distance between those
objects and the shadow projection plane. It is evident that

Fig. 1. Variation in the proportion between shadow and penumbra. As an opaque objects height h increases, its shadows size decreases. The size of the
penumbra zone increases too, until it reaches a limit distance ld in which the shadow disappears and totally becomes penumbra.

142

J.H. Salazar Trujillo / Solar Energy 110 (2014) 139150

Fig. 2. Architectural devices for penumbra projection. Left: pavilions blinds. Expo-Sevilla, Spain. Latitude 37N. Centre: traditional markets roof, San
Pedro de Atacama, Chile. Latitude 23S. Right: traditional facade in San Juan de Uraba, Colombia. Latitude 9N. Authors photographs.

the last two variables can be absolutely controlled through

the design process. Due to this fact, it is necessary to formulate the involved trigonometric relations for calculating
tables and diagrams allowing to know the penumbra conditions of an architectural enclosure prior to its construction.
3.1. Solar disk angular size
In the aphelion, the Earth is 152.6 million kilometres
away from the Sun, while during the perihelion its point
in the orbit closest to the Sun that distance is reduced to
147.5 million kilometres (Wilhelm and Dwivedi, 2014).
Since the equatorial radius of the Sun is 0.695 million kilometres, it can be calculated that, from the Earth, the solar
disk is seen in an viewing angle f varying from 0.26095
during the aphelion (which occurs on or about 4th July)
and 0.26997 during the perihelion (which occurs on or
about 3rd January). Variation in the apparent size of the
solar disk is about 4%, a magnitude that needs to be considered for architectural applications in which there is an
interest to take this subject to the limit (Fig. 3). Nevertheless, for most applications, the solar disk viewing angle can
be considered as constant and equal to the aphelion angular opening (f = 0.26095), when the sun is furthest and is
seen with a smaller angle. Under this condition, an objects
shadow will need a longer trajectory before completely fading into penumbra.
The simplication of assuming the apparent solar disk
size as constant allows to not dierentiate between shadows
and penumbras produced at dierent times of the year.
This simplication leads to a maximum error of 4%, which
is equal to specifying 4 additional millimetres in a 0.20 mwide piece installed at 4.0-m height (Fig. 4). This is a short
margin of error for the precision level of a construction. In
the following gures the variation in the results that can be
expected in the case this variation is taken into account is
showed with a dotted line.

Fig. 3. Annual variation in the apparent size of the solar disk. During the
perihelion (January), the Earth is closest to the sun; at that time, shadows
are smaller and penumbras are wider than what they will be six months
later. The radius of the arcs represents the proportion of such variation in
real magnitude. The solar disk viewing angle f can be calculated from
the relation between the Radio of the Sun and the distance between the
Earth and the Sun.

3.2. Penumbra rate

Penumbra calculation is relevant when designing objects
whose visual obstruction angle is similar to the angle with

Fig. 4. Shadow size annual variation. When sunrays occur perpendicularly, a 0.20 m object located at 4.0 m height will have an annual variation
of 4 mm in its shadows width.

J.H. Salazar Trujillo / Solar Energy 110 (2014) 139150

which the solar disk is seen from the earths surface, a

moment when penumbra zone edges on both sides get close
enough and the reduction in shadow size becomes evident.
This occurs with objects that are very distant from the shadow projection plane (i.e.: an airplane ying by) or with
small objects such as prismatic elements of just some centimetres wide which make up a pergola.
Using the accepted solar nomenclature (Fig. 5)
Blanco-Muriel et al. (2001), the penumbra zone width d,
going from point B to point C (Fig. 6), can be calculated
from triangles ABA0 and ACA0 , which are formed from
solar elevation angle e, adding or subtracting angle f/2.
Preparing equations for denite shapes and orientations
would imply latitude and solar geometry generalization.
To give architectural freedom to the innite possible shapes
to be designed, it was preferred to develop a method of calculation based upon only ve basic equations valid for
both hemispheres instead of a set of complex equations
with a narrower location or shape validity. That is why
the following equations start from a solar elevation angle
that must be previously calculated according to the latitude
of the location under study.
The expression for calculating the penumbra zone width
d that is projected by an opaque objects edge hanging at h
height with respect to the shadow projection plane is as
follows:
d h= tane  f =2  h= tane f =2

For an opaque object with width x, limit height lh

(Fig. 7) can be calculated from the point where its shadow
will not reach the ground, according with the following
equation:
lh x tane f =2 tane  f =2=tane f =2
 tane  f =2

143

limit size lx would be reached, from which its shadow

can no longer be projected on the ground (Fig. 8). As it
is shown in Figs. 7 and 8, this occurs when the objects
width is equal to its penumbras width d and, therefore, it
is also dened by Eq. (1), shown with grouped variables
as follows:
lx htane f =2  tane  f =2=tane f =2
 tane  f =2

The diagrams derived from those last two equations are

sucient for the dimensioning of some architectural elements, but in neither equation is the objects thickness or
the geometry of its section taken into account. Some additional precisions will be necessary for the calculation
method generalization so that variation in the direction
of the sunrays could be considered. This will allow to consider the eect of the elements proximity and the variation
in their apparent section based on solar elevation angle e.
These subjects will be addressed in Section 3.3.
At the interval between B and C (Fig. 9) there is a transition from the sunny zone (with 100% visibility of the solar
disk) to the shading zone (0% visibility). This gradation in
shadow density, which is called penumbra rate PR, can be
calculated for every point in the segment going from the
beginning to the end of the penumbra zone. To do so,
the distance between the calculation point and the beginning of the penumbra (distance y) will be needed to calculate de g angle, corresponding to the circular sector visible
from point Q (Eq. (4)).
g arccos1  2y=d

Finally, to compute the area of the circular segment that

becomes visible from every point (Fig. 10) and transform
the result in a percentage value, another expression (Eq.
(5)) will be enough to obtain the penumbra rate value PR
from a g value obtained in degrees.

If the objects height h became constant and the width of

the object projecting the shadow were progressively
reduced, the analysis sequence would be similar and the

PR% 2g=360  sing cosg=p

Fig. 5. Solar nomenclature used in this paper, corresponding with BlancoMuriel et al. (2001). The solar elevation angle e is measured from the
horizon line, the standard reference in architectural drawings.

Studying penumbras near an object whose dimensions

are signicantly larger than the penumbra zones width is
not of much practical interest in architecture. For example,
for a 10-m-high building with e = 60, the width of this
zone is just 0.12 m. Nevertheless, when penumbras created
on every side of the object begin to interact and to overlap
because of their proximity, their study is most relevant.
When limit distance is exceeded, an object can no longer
completely block the solar disk visibility from any point on
the studied surface. The shadow has already disappeared
and penumbras on both sides of the object begin to overlap
in an additive way. In order to calculate the penumbra rate
from a point under these conditions, it suces to calculate
partial visibilities of the solar disk segments on both sides
of the object. When graphing the solar radiation attenuating eect that produces a series of opaque elements whose
penumbra zones are overlapped, a sector becomes evident
in which solar radiation takes values notably lower than

144

J.H. Salazar Trujillo / Solar Energy 110 (2014) 139150

Fig. 6. Penumbra zone width d. For a solid element at a height h, penumbra zone width d can be calculated. Between points B and C there is a transition
between the sunny sector and the shading sector. The results for two solar elevation angles e are graphed on the right. Results obtained for the aphelion are
dierentiated with solid lines.

Fig. 7. Minimum size needed to project shadow at dierent heights. For an object with a dimension x there is a limit height lh from which its shadow will
not reach the oor.

Fig. 8. Maximum elements height to obtain shadow for dierent elements sizes. For a height h, it is possible to calculate the minimum size an object must
have for its shadow to reach the oor. The design point of the Orchids Exhibition Centre of Medellin is highlighted: 4.6 cm wide elements installed 14.0 m
above the oor. If a stripped oor was desired during middays, it would be necessary to specify elements at least 13 cm wide. Keeping the strips visible
when the solar elevation angle is lower (e = 45) would require elements wider than 25 cm.

those in the sunny zone, but very similar to each other

(Fig. 11). The oscillations in PR values under a surface
for penumbra projection vary according to the distance
that separates the opaque elements: it will become longer
as the distance increases, and it will approach zero when
the elements are too close to each other. If distance is excessive, the direct lightning bands will become visible again

(PR = 1), but the median value would be equal to the surfaces drilling percentage.
3.3. Variation in the origin direction of solar radiation
When the elements forming a surface have a circular section, limit distance ld and limit size ls values are constant

J.H. Salazar Trujillo / Solar Energy 110 (2014) 139150

Fig. 9. Penumbra rate calculation, part one. From the beginning of the
penumbra zone (identied with point B) to the end of the transitional
condition (point C), the edge of a solid object obstructs part of the solar
disk. The expression to calculate the obstruction rate starts from angle g
calculations, according to the distance between the point under study Q
and the beginning of the penumbra zone (y value).

Fig. 10. Penumbra rate calculation, part two. The visible solar disk area
[a] from point Q can be calculated as the subtraction of two areas: the
circular sector area [acs] dened by point Q0 minus the inscribed triangle
area [at]. The resulting circular segment area is expressed in percentage
according to the angle g, obtained from the distance y that separates the
point Q from point B.

for all the solar elevation angles since, regardless of the

direction the sunrays come from, the object projections
width will be always the same. The minimum height at
which a circular section object with a diameter x should

145

be installed, for it to not project shadow on a horizontal

plane, according to the solar elevation angle variation,
can be calculated from Eq. (2) (Fig. 12).
These diagrams provide enough information to specify
the minimum or maximum dimensions of a shading element with circular section. When the elements have a dierent section, it will be necessary to reiteratively make the
limit distance and/or size calculations for dierent incidence angles. This procedure implies calculating the element projections width perpendicularly to the sunrays
for every case and then, taking the limit distance value ld
corresponding to each projections size s. The results ld
diagrams for a particular section allow to demonstrate
how objects often used in the construction of pergolas
and other shading devices regularly have asymmetrical
behaviours on distances from which they do not project
shadows on the ground (Fig. 13).
The continuous variation in the direction in which solar
radiation comes from, demands considering also the proximity eect between adjacent elements to identify the angle
from which each elements shadow begins to project itself
on the adjacent element. In this way, the moment can be
determined when a discontinuous surface, composed of a
repetition of similar elements, begins to function as a solid
object (Fig. 14).
Limit distance diagrams allow to include in the analysis
the variety of distances that can appear between an opaque
element and the surface or surfaces where shadow and penumbra occur. If it becomes interesting to establish variable
dimensioning and spacing criteria according to the proximity relationship of the surfaces set which compose an enclosure, limit distant variations for every direction will need to
be considered.
Sunrays are not always annoying. Even in the tropics,
there are times and places where shadow patterns recognition can be bearable or even favourable. Following the
optimization path, it is advisable to start the design of a
surface for penumbra projection by identifying the intervals and places where such condition will be desirable
and, subsequently, calculate the corresponding solar trajectories. This will allow the monitoring of sectors where
shadows will or will not be recognisable as well as the intervals where such conditions will be present.
There are two ways of applying the concept and diagrams for limit distance calculation: spatial and temporal.
Spatial application consists in calculating the intersection
between a limit distance diagram and the surfaces marking
o a space. This procedure allows to identify zones that can
possibly receive shadow or penumbra. On the other hand,
temporal application consists of the projection of these
very intersections towards the celestial canopy, which
allows to mark o the intervals (in dates and hours) when
specic penumbra conditions could appear (Fig. 15). In
both applications, associated angles and distances are
issues that must be treated in a tridimensional way by
applying shade covers and solid angles, even though bidimensional angles are suitable enough for most situations.

146

J.H. Salazar Trujillo / Solar Energy 110 (2014) 139150

Fig. 11. Subtractive overlapping of penumbra rates. To calculate the solar disks obstruction level that creates a succession of opaque elements it suces
to subtract the penumbra rates PR of adjacent elements. On the penumbra zone edges, there will be an increase from the condition of complete solar
exposure until reaching the sector in which penumbra rate value oscillates corresponding with opaque elements and solar disk rhythmical interactions.

Fig. 12. Calculation of several limit distances ld according to the elements section. When the element section is circular, the limit distance ld is constant for
every solar elevation angle e. When the element section is not circular, it will be enough to calculate the elements projection s in the direction of a plane
perpendicular to the sunrays, and calculate the limit distance corresponding to every solar elevation angle.

J.H. Salazar Trujillo / Solar Energy 110 (2014) 139150

147

Fig. 13. Limit distance diagrams for dierent proles. Every prole has a particular limit distance diagram. The separation between neighbouring elements
(identied as step in the gure) restricts the possibility of having partial visibilities of the solar disk in certain directions, trimming some diagrams
portions. Intersections with the oor identify the sectors where shading conditions can appear and projections to the sky dome identify the time intervals
when such conditions could be observed.

Fig. 14. Application of limit distance diagrams. The proximity of an opaque object to the surfaces of a place can cause intersections with its limit distance
diagram. This makes possible to identify the sunrays directions (hours or months) that will project shadow and those that will project penumbra only.

148

J.H. Salazar Trujillo / Solar Energy 110 (2014) 139150

4. Case of implementation. Orchids exhibition centre of

Medellin, Colombia
Since Feria de las Flores (Flowers Festival) has
become an event of international renown, the Botanical
Gardens of Medellin, Colombia (located at 6.25North latitude, 1460 masl, MeanT 22 C MinMeanT 16 C, MaxMeanT 29 C (Atlas Climatologico de Colombia, 2005)
demanded a space destined to the orchids annual exhibition. The building should also favour the carrying out of
dierent activities such as weddings, concerts, parades,
fairs, and conferences for the rest of the year (Fig. 16).
The challenge was to guarantee solar light conditions
intense enough for the plants to nourish, but faint enough
for visitors to be comfortable.
This project, with a 4000 m2 covered area, was designed
in 2006 by Plan:b Arquitectos in association with JPRCR
Arquitectos (Mesa et al., 2014). As a design principle, the
exhibition zone should guarantee visual ergonomic conditions without high contrast situations despite the fact that
its roof was 50% made of translucent polycarbonate. As an
additional requirement, any place on the oor and up to
4.0 m high could not contain sunny zones due to the solar

Fig. 15. Duration intervals of a penumbra condition. Intervals where a

shadow condition begins and ends allow to calculate in the direction of
the sky dome the intervals in dates and hours on which penumbra
conditions can appear.

radiation coming in between the wood strips that congure

a 14.0 m high pergola. This design demand was clearly stated: the shadow of the elements forming the pergola must
be totally indistinguishable at any hour of the day or period of the year.
Orchids are extremely fragile, and it is unacceptable to
have a plant exhibited under a spot of sun, even for a
few minutes. In this place, orchids from all over the world
collectible pieces, rare, and sometimes endangered specimens are exhibited. The Orchids Exhibition Centre
Orquideorama in Spanish is responsible for protecting
this valuable collection from sunrays, which quickly damage some of the ower species. The canopy the upper part
of the foliage in a tropical forest creates the suitable
microclimate for orchids to grow and develop since that
is their natural habitat. The building forms an articial
canopy that produces a penumbra level similar to the one
of a native forest and also gives continuity to the set of
trees surrounding the building. Thanks to the transition
from the foliage of the existing trees and the articial canopy formed by the pergola, the Orchids Exhibition Centre
does not have any lateral enclosures.
The botanical experts suggested maintaining the insolation level inside the Exhibition Centre below the 40% limit
to have an insolation exposure value appropriate to the
orchids health. Budget and structural reasons lead to a
clear sheet polycarbonate (g-value = 0.68) as the selected
roof material. To reach the required shadow level of 0.4
it would be necessary to specify a darker polycarbonate
sheet that would have modied the orchids colours. As part
of the bioclimatic counselling work made by PVG Arquitectos an alternative solution was proposed: a secondary
wooden shadow layer to maintain the natural light spectrum and provide the 0.28 missing value, giving origin to
the 40% fullness and 60% emptiness pergola ratio.
Innite size-distance combinations can provide a 40
60% pergola. The elements limit size was obtained from
Eq. (3) using the following values: solar elevation angle
e = 90, element height h = 10 m and solar viewing angle
in aphelion f = 0.26095. It must be noticed that the
designed height of the pergola was not 10.0 m, but using
a 14.0 m value would have led to a oor without shadow
strips accompanied by shadow patterns on visitors,

Fig. 16. Orchids Exhibition Centre in the Botanical Gardens of Medellin on a sunny day. Courtesy: Jardn Botanico de Medelln.

J.H. Salazar Trujillo / Solar Energy 110 (2014) 139150

149

Fig. 17. Architectural section of the building. The size and spacing between the elements that compose the pergola of the Orchids Exhibition Centre is
highlighted. Having the sun in the zenith is the critical design condition for orchids exhibition. The region free of shades during that moment is showed.
Courtesy: PVG Arquitectos.

collections and possible publicity elements within the vision

eld. To prevent this annoying situation, a 4.0 m height
allowance was initially reserved to keep that volume free
of shadow patterns, resulting in a 94 mm-limit size value.
Wood optimization reasons nally led to specify
46 mm  40 mm pieces. The 66 mm distance it is the
needed value to guarantee the required shadow ratio during mid-days. This pattern has a 5.04 m minimum height
value and as a result, in the Orchid Exhibition Centre there
is an 8.96 m height zone free of shadow patterns (Fig. 17).
The design point of the above-mentioned elements is
indicated with an icon (Fig. 8). The performance observed
during eight years of operation demonstrates the advantage
of using penumbra as a strategy for public space creation in
the tropics. The Orchids Exhibition Centre manages a full
programming agenda all year round and, regardless of how
clear the sky in Medellin is at a certain time, this buildings
interior always maintains a gentle penumbra environment.
5. Discussion
Surfaces for creating penumbra can be calculated from
the proportion between the full area and the empty area.
Those numbers are subsequently transformed into something similar to a shadow coecient, as if it were a translucent material. It is an imprecise method; therefore,
prototype construction was often necessary (Salazar,
2007). For those who design and construct in the tropics
it is evident that there are dierent shading qualities. By
applying penumbra rates, the geometrical characteristics
of these elements can be dened from required penumbra
conditions in a simpler and more precise way.

The calculation method has been applied in the design

of discontinuing surfaces composed by opaque elements,
but it is also being used to design double skin facades. In
this second type of application, the exterior facade consists
of surfaces with calculated drillings following the same
Shadow Dispersion principles. The maximum width of
the partition wall and the drillings silhouette represent here
the design variables to be dened, considering that the distance between the drilling and the shading plane is variable.
This requires a progressive reduction in the partition walls
width as the distance between drillings and the work plane
decreases.
Any set of drillings made in a repeated and homogeneous way causes interference patterns, which can be visually uncomfortable and even incapacitating. Drillings made
with random shape, position and size as long as they are
within the correct dimensional interval are the simplest
way to prevent interference patterns. Nevertheless, even
with randomly distributed drillings it is not possible to
obtain a completely homogenous penumbra. It has been
empirically veried that, from a certain homogeneity level,
PR variations between neighbouring regions on the same
work plane are not annoying. Penumbra equations will
be used to continue studying this subject.
Architectural enclosures for penumbra projection have a
lot of potential in proting from natural light resources in
buildings. Drillings and distances between opaque elements
do not imply absorption or light transmission phenomena;
that is the reason why they oer great possibilities as natural light proting devices. Their performance in skylights
and windows in which visual conditions are not important
will be superior to the performance of glasses with high

150

J.H. Salazar Trujillo / Solar Energy 110 (2014) 139150

shadow coecient, which often alter the spectral composition of light and cause changes in colours.
Nowadays, the use of discontinuing and drilled surfaces
for penumbra projection is not proportional to its possibilities. The diculties in construction to ensure a good nishing with small and repeated elements, added to the obstacles
to know in advance the sunrays softening eect, could
explain this situation. With the marked tendency to introduce digital production and parametric design techniques
(e.g., laser cut sheets, tridimensional plotter, or digitally
printed glass, among other techniques), a large number of
architectural applications of the equations and methodology of analysis presented here are foreseen in the short-term.
6. Conclusions
Penumbra rates allow to design and specify architectural
devices for softening solar radiation starting from a
required penumbra condition. Their application allows to
carry out quantitative comparisons and predict the penumbra level of many frameworks of habitual use in tropical
zones. It also allows to establish the geometrical characteristics of the elements composing the surface for solar control, this time taking into account sunrays divergence.
Interdependency relationship between limit size lx and
limit distance ld prevents from giving a unique prescription,
from which the Shadow Dispersion eect will be evident.
By way of generalisation, when designing shading devices
at a height of 4.0 m from the ground (the typical height
for urban shading elements) composed by elements with
a dimension equal to or lower than 0.075 m, Shadow Dispersion phenomenon must be considered. When designing
bigger elements, the eect would be less signicant and calculations could be made assuming sunrays as being parallel. For elements installed at any other height, it would
be sucient to make linear proportionality calculations.
It is necessary to determine what the visual tolerance is
to the changes in penumbra rates between neighbouring
points. This will allow to include human factors in the
method exposed. So far, tolerance based on visual demand
level, kind of task, lighting level and age group is unknown.
In the short-term, advantage will be taken of the possibility
of quantifying the intensity of a penumbra to continue
studying this subject.

References
Arzoumanian V. Sol y Arquitectura, Ed. Gustavo Gili, Barcelona , 1988.
IDEAM. Atlas Climatologico de Colombia, Bogota, 2005.
Baruch, G., 1976. Man, Climate and Architecture, second ed. Applied
Science Publishers, London.
Blanco-Muriel, M., Alarcon-Padilla, D., Lopez-Moratalla, T., LaraCoira, M., 2001. Computing the solar vector. Sol. Energy 70 (5), 431
441.
Buie, D., Monger, A., 2004. The eect of circumsolar radiation on a solar
concentrating system. Sol. Energy 13, 181.
Buie, D., Dey, C., Bosi, S., 2003a. The eective size of the solar cone for
solar concentrating systems. Sol. Energy 5, 417.
Buie, D., Monger, A., Dey, C., 2003b. Sunshape distributions for
terrestrial solar simulations. Sol. Energy 2, 113.
Higueras, E., 2013. Urbanismo Bioclimatico. Coleccion Arquitectura y
Diseno + Ecologa. Editorial Gustavo Gili, Madrid.
Koenigsberger, O.H., Ingersoll, O.H., Mayhew, T.G., 1977. Viviendas y
Edicios en zonas calidas y tropicales. Editorial Paraninfo, Madrid.
Kolivand, H., Shahrizal Sunar, M.A., 2013. Survey of shadow volume
algorithms in computer graphics. IETE Technical Review, vol. 30(1),
Medknow Publications & Media Pvt. Ltd., pp. 3846.
Lan, W.M.C., 1986. Sunlightning as Formgiver in Architecture. Van
Nostrand Reinhold Company, New York.
Lippsmeier G. Tropenbau, Building in the Tropics, Ed. Callwey Verlag,
Munich, 1969.
Mesa, F., 2013. Permeabilidad. Mesa Editores, Medelln.
Mesa, C., Mesa, M., Toledo, R., 2014. Medelln Architecture Guide 1994
2014. Mesa Editores, Medelln.
OBrien, W., Kapsis, K., Athienitis, A.K., n.d. Manually-operated
window shade patterns in oce buildings: a critical review. Build.
Environ., pp. 60319338. doi: 10.1016/j.buildenv.2012.10.003.
Olgyay, V., 1968. Clima y Arquitectura en Colombia. Editorial Universidad del Valle, Cali.
Puppo, E., Puppo, G.A., 1972. Sol y Diseno. Editorial Marcombo,
Barcelona.
Salazar, J.H., 2007. Diseno de Sombras. Universidad Nacional de
Colombia, Medelln.
Salazar, J.H., n.d. Solar performance and shadow behaviour in buildings.
Case study with computer modelling of a building in Loranca, Spain.
Building and Environment, pp. 33117130. doi: 10. 1016/S03601323(97)00027-9.
Szokolay, S.V., 1977. Solar Energy and Building. Architectural Press,
London.
Tzonis, A., Stagno, B., Lefaivre, L., 2001. In: Alexander Tzonis, Liane
Lefaivre and Bruno Stagno (Eds.), Tropical Architecture: Critical
Regionalism in the Age of Globalization, Wiley-Academic, Chichester,
New York.
Wilhelm, K., Dwivedi, B., 2014. Secular perihelion advances of the inner
planets and asteroid Icarus. New Astronomy, February 18, 2014.
Yanez, G., 1988. Arquitectura Solar. MOPU, Madrid.