Professional Documents
Culture Documents
(Asce) Be 1943-5592 0000445 PDF
(Asce) Be 1943-5592 0000445 PDF
Abstract: The structures of Santiago Calatrava are noted for their adventurous forms, and Calatrava is heralded as a hybrid architect/engineer
whose works represent an integration of both disciplines. While there have been a number of publications on the structures of Calatrava, nearly
Downloaded from ascelibrary.org by CASA Institution Identity on 04/22/20. Copyright ASCE. For personal use only; all rights reserved.
all have focused on the architecture or artistry of them without addressing how well they fulfill the requirements or represent great works of
structural engineering. Here a thorough structural analysis of one of Calatrava’s most well known bridges is performed to determine how
well it achieves the goals of structural engineering. This provides insight as to how well Calatrava’s works represent engineering as art and what
we can learn from his works in the context of great works of structural art. DOI: 10.1061/(ASCE)BE.1943-5592.0000445. © 2013 American
Society of Civil Engineers.
CE Database subject headings: Cable-stayed bridges; Structural design; Aesthetics; Architecture.
Author keywords: Cable-stayed bridges; Structural design; Aesthetics; Structural art; Architecture.
Santiago Calatrava is currently one of the most celebrated designers Santiago Calatrava was born in Valencia, Spain, in 1951. With formal
of structures in the world. He is lauded not only for his striking works training in drawing and painting through high school, Calatrava en-
but also for integrating architecture and structural engineering, as rolled in the architecture program at Escuela Technica Superior de
he was educated in both fields and practices both disciplines. His Arquitectura de Valencia. Following graduation, he left his homeland
bridges and buildings have been the focus of numerous monographs, of Spain to pursue studies in engineering at the Swiss Federal Institute
papers, and articles. Yet nearly all of these publications focus on of Technology (ETH) in Zurich, Switzerland. ETH-Zurich, whose
the architecture, or sculptural qualities, of Calatrava’s designs rather alumni also include the engineers Robert Maillart, Othmar Ammann,
than an investigation into the structure, or engineering, of them. Thus and Christian Menn, has continuously passed on to its students the
there is little published information allowing for a critique of either great tradition of Swiss bridge design. Menn also continued a tradition
the structural engineering behind his works or how well they rep- of great designers lecturing at ETH (such as Maillart) by returning
resent structures as art. to teach while Calatrava was a student there. Calatrava studied civil
This paper explores the design concept of Calatrava’s 1992 engineering and then earned a Doctorate in Technical Science in the
Alamillo Bridge in Seville, Spain (Fig. 1). This cable-stayed bridge Department of Architecture. In 1981, upon finishing his formal edu-
is one of Calatrava’s most celebrated structures and is emblematic cation, Calatrava opened an architectural office in Zurich (Calatrava
of the dramatic forms for which he is known. It also holds the 2010).
distinction of being one of the few works of Calatrava’s for which Since that time, Calatrava has become one of the most successful
structural details have been published. By studying the engineering designers in the world. His honors include both the Gold Medal from
of the bridge, we seek to critique one of Calatrava’s works of the Institution of Structural Engineers and the Gold Medal from the
structural engineering by the standards of that field. With a thorough American Institute of Architects. His acclaim is also evidenced by
understanding of the structure of the bridge, we may then turn to the exhibition of his work in 2005–2006 at the Metropolitan Mu-
a consideration of if and how it represents a great work of structural seum of Art, “an institution that rarely presents a living architect and
engineering or a work of structural art. obviously endorses this one,” according to the noted architecture
critic Ada Louise Huxtable (2005). The exhibition title, Sculpture
into Architecture, provides insight into Calatrava’s vision of design.
1 Among prominent, contemporary individual architects, Calatrava
Associate Professor, Dept. of Civil Engineering, Johns Hopkins Univ.,
is unique for earning commissions to serve as both the architect and
Baltimore, MD 21218 (corresponding author). E-mail: jkguest@jhu
.edu
the structural engineer. Yet while there are a great number of pub-
2
Assistant Professor, School of Architecture, Planning & Preservation, lished reviews of his architecture and sculpture, there is little in-
Univ. of Maryland, College Park, MD 20742. formation available on the engineering of his works. His Alamillo
3
Gordon Y. S. Wu Professor of Engineering Emeritus, Dept. of Bridge in Seville, Spain, therefore, provides an opportunity to in-
Civil and Environmental Engineering, Princeton Univ., Princeton, NJ vestigate Calatrava’s approach to engineering.
08544.
Note. This manuscript was submitted on December 12, 2011; approved
on September 27, 2012; published online on September 29, 2012. Discus- Alamillo Bridge: Background and Design Concept
sion period open until March 1, 2014; separate discussions must be sub-
mitted for individual papers. This paper is part of the Journal of Bridge The Alamillo Bridge was one of six bridges commissioned and built
Engineering, Vol. 18, No. 10, October 1, 2013. ©ASCE, ISSN 1084-0702/ to improve infrastructure for the 1992 Universal Exposition hosted
2013/10-936–945/$25.00. on La Cartuja, an island located just outside the city of Seville. The
Fig. 2. Elevation view of the Alamillo Bridge; annotations A, B, and C indicate locations of cross sections shown in Fig. 4
Fig. 3. The design concept for the Alamillo Bridge was based on a balance of forces: (a) cable forces were to balance out pylon weights, resolving
to a purely axial force in the pylon; and (b) the resultant pylon axial force was to balance with the deck axial force to create a purely vertical reaction at
the support
member axial force demands, which are typically quite large in of 16 RC caissons. Each caisson is encased in a steel shell that serves
cable-stayed bridges, nor their stress states, which are needed to as permanent formwork and contains an interior stairwell running
gauge component efficiency. This is not a criticism by any means, along the longitudinal axis [Fig. 4(a)]. The average dimension of
but rather explains the motivation for our present study utilizing a caisson is approximately 12 m 3 8 m (39 ft 3 26 ft), and the pylon
a thorough structural analysis to explore and evaluate the efficiency footprint is approximately 110 m2 ð1180 ft2 Þ directly above the
and appropriateness of the design concept. pedestal and approximately 158 m2 ð1700 ft2 Þ at the pylon base. The
We therefore used the available structural data to create a two- deck spans nearly 200 m (656 ft) and is composed of a 4.4 m (14.4 ft)
dimensional FE model. All model input, including structure geome- deep steel box girder constructed from 5 cm (2 in.) thick plates [Figs. 4
try, section properties, material properties, and boundary conditions, is (b and c)]. The girder contains 18 steel T-stiffeners welded along its
provided in Pollalis (1999). Section properties were verified through entire length, with stiffening diaphragms every 3 m (10 ft), and is filled
hand computation by the authors using data contained in the con- with concrete at sections nearest the pylon [for example, Fig. 4(b)].
struction drawings (Guest 2001). The majority of cross-section Load is transferred between the deck and pylon via 13 pairs of cables.
properties were found to be within 5% of the model data reported These pairs range in cross-sectional area from 169 cm2 ð26 in:2 Þ to
in Pollalis. However, a typographical error was noted in Table 4.4 of 124 cm2 ð19 in:2 Þ and are made of 270 grade steel. Detailed cross-
Pollalis (1999), where cable cross-sectional areas are reported as an section drawings can be found in Pollalis (1999).
order of magnitude too small. Discrepancies were also found in the
lower pylon members, where a curved pedestal provides a smooth
Construction Sequence
transition between the pylon and deck (below point A in Fig. 2).
This region is mechanically complex, required construction in two
The construction sequence has been discussed extensively in pre-
phases, exhibited cracking during the pylon construction process,
vious works and has been labeled as technically challenging and
and was experimentally monitored for stiffness during construction
expensive as a result of the unconventional design concept (Pollalis
(Aparicio and Casas 1997). For these reasons, and given the con-
1996, 1999; Aparicio and Casas 1997; Lopez and Halpin 2000). We
firmed accuracy of the noncable member calculations, we used
highlight here only a few key points as they relate to the structural
Pollalis’s published values for all pylon and deck members.
engineering and economy of the bridge. First, the bridge was out-
The structural model was analyzed using first-order and second-
fitted with an extensive system of sensors to enable careful moni-
order elastic analysis. Maximum bending moments were found to
toring during construction (Aparicio and Casas 1997; Casas and
increase less than 7% using nonlinear analysis for the governing load
Aparicio 1998). This system was used to detect and mitigate errors
cases. This supports the findings of Aparicio and Casas (1997), who
in geometry, section weights, and resultant forces in the cables
reported that linear and nonlinear analysis yielded comparable
and temporary supports. Second, cantilever construction, typically
results for the Alamillo Bridge. As this increase is relatively minor,
a cost-effective approach to cable-stayed bridge construction, was
linear elastic analysis was deemed sufficient to explore the design
not viewed as a viable option because of the potential for force
concept herein.
imbalance in the pylon or deck. Instead, the river channel was
temporarily dammed and the entire deck was constructed with the
Structural Components use of temporary supports. Individual pylon segments and their
corresponding cables were then installed and tensioned, with cor-
The bridge features three primary structural components. The in- responding deck temporary supports removed along the way [see
clined pylon is 134 m (440 ft) tall, 159 m (520 ft) long, and composed Pollalis (1996, 1999) for detailed sequencing].
Fig. 4. Sample cross sections of the Alamillo Bridge: (a) pylon caisson number 4; (b) largest steel box deck section, which is reinforced with concrete in
the area between the outer plates and the inner rectangular space; and (c) typical hollow steel box deck section; these cross sections are located at points
A, B, and C in Fig. 2; sections are shown without stiffening diaphragms and deck sections without longitudinal stiffeners
Fig. 5. Bending moment diagrams under (a–c) independent dead load cases; (d) design dead load combination; and considering 10% construction
safety factors that maximize (e) positive and (f) negative bending moments in the pylon
under dead load. Although the magnitudes are significantly reduced compared with the uniform traffic load but is often governing in the
from the independent load cases, the bending moment at the base of design of local elements such as the deck slab and beams.
the pylon remains substantial at 802,000 kN×m ð592,000 kip-ftÞ. For The Alamillo Bridge was designed considering a 500-year wind
comparison, symmetric cable-stayed bridges with vertical towers event with maximum design wind speeds estimated at 263 km=h
naturally achieve a (near-) zero-moment state in the pylon under dead ð164 mphÞ and 186 km=h ð116 mphÞ at the top and bottom of the
loads. pylon, respectively. These wind velocities, which are characteristic
The bending moment diagrams for the independent dead load of Category 5 and 3 hurricanes, were converted to static concentrated
cases also provide insight into the consequence of construction errors, nodal loads using drag coefficients of 1.78 for the pylon and 1.30 for
or deviations from the designed weights and cable forces. Such errors the bridge deck [see Pollalis (1999) for additional details].
routinely occur in all construction projects and, because of its un- Thermal loads of 626C were considered for the cables and
conventional design, the Alamillo Bridge was viewed as particularly 623:3 and 613:5C for the deck and pylon members, respectively.
susceptible to them. Construction safety factors of 10% were included Thermal gradients of 21 and 4.5C were also applied to the deck
on the pylon and deck design dead loads to account for any such and pylon members, respectively, to account for the nonuniform
imbalances, and they were considered as compounding (Aparicio and thermal distribution across these very large members (Pollalis 1999).
Casas 1997; Pollalis 1999). The effect of a 10% heavier pylon and
10% lighter deck, for example, will nearly double the bending mo-
ment demand on the pylon to 1,480,000 kN×m ð1,090,000 kip-ftÞ Live Load Analysis
[Fig. 5(e)]. Reversing the dead load variations leads to maximum
negative moment of 2439,000 kN×m ð2324,000 kip-ftÞ [Fig. 5(f)]. The full suite of possible load combinations, including alternating
Dead load design amplification factors would ultimately increase construction safety factors, wind directions, and thermal directions,
these magnitudes another 50%. Deviations of 10% were also con- were considered. We present here only the cases and results that
sidered for the cable posttensioning forces independently of deck and produce the maximum bending moments in the pylon and deck.
pylon dead load errors. Such deviations would have approximately the
same impact on pylon bending moments as deviations in deck dead Governing Load Cases
loads. It is interesting to note that the potential for 10% errors in all
Examining each load independently helps determine the combina-
three loads cases was not considered. These may be readily estimated
tions that will create the largest moments in the pylon. As summa-
from Fig. 5 and would have added 202,000 kN×m ð149,000 kip-ftÞ
rized in Table 1, applied loads will cause the pylon to rotate either
and 2159,000 kN×m ð2117,000 kip-ftÞ to the pylon maximum
clockwise or counterclockwise. The load combinations producing
moments in Figs. 5(e and f), respectively.
maximum positive and negative bending moments in the deck and
The dead load analysis results are in good agreement with Casas
pylon may then readily be identified using this classification, with
and Aparicio (1998) and Pollalis (1999). Deviations are likely due
construction safety factors applied accordingly. These combinations
to assumptions about the boundary conditions, connections, and
and corresponding load factors are defined in Table 2.
loads. For example, we have fully released bending moments in the
It should be noted that the load cases in Table 2 and associated
cable elements of our model to prevent shear and bending moment
results include the effects of thermal loads. Their contribution is
transfer, while the deflected shapes presented in Pollalis (1999)
relatively minor—for example, thermal loads contribute approxi-
appear to suggest bending in elements representing the cables. Casas
mately 2 and 3% to pylon bending moment magnitudes for cases A
and Aparicio (1998) collected experimental data that may have been
and B, respectively. Although minor here, we note that such loads
used to update their models.
are typically neglected in ultimate strength design under the
Although the overall design objective is to minimize the bending
moment envelope, the extreme value of pylon bending moment
under dead loads alone indicates the inefficiency of the no-backstay
design concept. The large bending moments require large cross Table 1. Breakdown of Load Effects on Pylon Rotation (Orientation as
sections, which in turn increase structure weight and the impact of Shown in Fig. 2)
imbalances. Inducing clockwise rotation Inducing counterclockwise rotation
Deck dead load Pylon dead load
Cable post-tensioning Deck thermal gradient
Live Loads
Eastward pylon thermal gradient Westward pylon thermal gradient
Cold thermal loads Warm thermal loads
The structural model was also subjected to the full suite of design
Eastward wind Westward wind
load combinations to explore the bending moment demand envelope.
Traffic load —
These load combinations were dictated by the Spanish design code
ing moment. Although the angle appears relatively small, the vector the behavior of this structure. The presence of backstays without
magnitudes and pylon length are quite large, leading to extremely posttensioning would not in itself affect the bending moment dis-
large magnitude bending moments. The resultant vectors swing tribution under counterclockwise rotation of the pylon [Fig. 7(a)].
more dramatically clockwise under combination B, although the They would, however, provide a restoring force under clockwise
unsupported pylon weights at the base now reduce the effect. This rotation, leading to significant reduction in bending moment mag-
will be shown to translate into double curvature in the pylon. nitudes in Fig. 7(b). This would allow pylon weight to be reduced,
The design concept is relatively successful at reducing horizontal leading to a recentering and further reduction in magnitude of the
thrust at the support as motivated in Fig. 3(b). The horizontal re- bending moment envelope, and further reduction in member sizes.
action at the base of the pylon is approximately 7,000 kN (1,600 kip) Figs. 7(c and d) indicate that the deck deflection exhibits double
for load combination A, significantly less than the vertical reaction curvature under bending, with the inflection point occurring east-
of 339,000 kN (76,200 kip) at this location. ward of the third cable pair. For load combinations that amplify
clockwise rotation of the structure, the deck behaves very similarly
to a cantilever beam over the clear span region and simply supported
beam where the deck is supported by cables [Figs. 7(b and d)]. In this
Table 2. Factored Load Cases Producing Maximum Bending Moments in case, bending introduces tension in the top plates of the steel box in
Pylon and Deck the cantilever zone and in the bottom plates in the simply supported
Case Load combination zone. The force couple is reversed for combinations that amplify
counterclockwise rotation of the pylon [Figs. 7(a and c)].
A. Maximum positive 0:9 ½1:5 ð1:1 PDL 1 0:9 DDL 1 1:0 CFÞ
The extremely large bending moment demands necessitate
pylon bending moment 1 1:5 westward wind 1 warm thermal load
the use of deep members to reduce the corresponding longitudinal
1 westward thermal gradient
stresses. This increases structure weight and amplifies axial force
B. Maximum negative 0:9 ½1:5 ð0:9 PDL 1 1:1 DDL 1 1:0 CFÞ
demand, which is the primary internal force developing in typical
pylon bending moment 1 1:5 traffic load 1 1:5 eastward wind
cable-stayed bridges. Each cable introduces a compressive axial
1 cold thermal load 1 eastward thermal gradient
force into the deck and pylon, increasing axial force as we move
C. Maximum positive 0:9 ½1:5 ð1:1 PDL 1 0:9 DDL 1 1:0 CFÞ
westward along the deck and downward along the pylon. This trend
deck bending moment 1 1:5 westward wind 1 cold thermal load
is clearly illustrated in Fig. 8 for pylon dead load and the live load
1 eastward thermal gradient
combination A. The locations of maximum axial force and maxi-
D. Maximum negative 0:9 ½1:5 ð0:9 PDL 1 1:1 DDL 1 1:0 CFÞ
mum bending moment are coincident in the pylon and deck
deck bending moment 1 1:5 traffic load 1 1:5 eastward wind
members, leading to a cumulative effect on the compressive stress
1 warm thermal load
demand. In the case of the deck, the steel box girder is almost
Fig. 6. Resultant vectors at each pylon node for comparison with Fig. 2(a) for load cases: (a) design (no errors) dead loads; (b) governing positive
pylon bending moment; and (c) governing negative pylon bending moment; all vectors are normalized against the maximum resultant in (a)
Fig. 7. Bending moment diagrams for load combinations producing (a) maximum positive moment in the pylon; (b) maximum negative moment in the
pylon; (c) maximum positive moment in the deck; and (d) maximum negative moment in the deck
Fig. 8. Axial force diagram under (a) design pylon dead loads and (b) the pylon bending moment governing load case [Fig. 7(a)]
completely filled with concrete westward of the first cable pair As indicated in the preceding figures, the axial force diagram
[Fig. 4(b)] and transitions to a larger concrete deck as it joins the [Fig. 8(b)] is nearly linear, while the bending moment diagram
pedestal. An interesting point here is that the axial forces developing [Fig. 7(a)] increases dramatically as we approach the pylon base.
in the deck under pylon weight alone account for 26% of the axial Although the moment of inertia also increases in this region, the
force demand under load combination A [Fig. 8(a) amplified by resulting tensile bending stresses near the base are much larger than
a 1.485 load factor]. For comparison, vertical towers carry their own the compressive axial stresses, leading to concrete cracking. The
self-weight and do not induce axial forces in the deck. tensile stresses in the outermost (eastern) face of the steel caisson are
estimated to approach 400 MPa (58 ksi), larger than the yield
strength but less than the ultimate strength of the A52d structural
Pylon Stress Distributions
steel. Compressive longitudinal steel stresses are significantly less
Given the bending moment and axial force demands, we can esti- on the western side of the pylon, ranging from approximately
mate the longitudinal stresses that develop in the pylon concrete and 70 MPa (10 ksi) to 140 MPa (20 ksi) over the bottom three steel
steel. This is a complicated endeavor given the complexity of the caissons. The compressive stress in the concrete is estimated to be
cross-section geometry (and reinforcement schedule), and its be- 12 MPa (2 ksi) to 26 MPa (4 ksi) over this same region. These stress
havior at ultimate limit state. The primary motivation for this stress magnitudes are below 35 MPa (5 ksi), the ultimate strength of the
analysis, however, is not to model structural failure mechanisms but concrete (Pollalis 1999), although the maximum is slightly larger
rather to determine if the extremely large pylon sections are over- than the factored ultimate strength of 23.5 MPa (3.4 ksi).
designed at the component level or are the result of an inefficient The stresses are significantly lower in the pylon’s upper caissons.
structural form. At a height of approximately 80 m (260 ft), the compressive axial
sections at these locations are appropriately sized considering the Previous designers who have successfully produced works that
large internal force demands that develop for this structural system. integrate aesthetics and structure demonstrated a focus on low cost
One could then argue that any deemed inefficiencies are due to by also integrating a thoughtful or innovative approach to con-
system-level, and not component-level, design. struction in their works. Robert Maillart, Pier Luigi Nervi, and Felix
Candela all either served as contractors as well as designers or
collaborated closely with builders. Nervi has been described as “a
Integration of Aesthetics and Structural Engineering hero of Calatrava’s” (Dunlap 2003). Calatrava even studied under
(and Construction) Candela for a time (D. Candela, personal communication, 2005).
Calatrava’s origins, education, and accolades would appear to
A recent graduate civil engineering thesis investigating modern make him an ideal inheritor of this tradition of the master builder.
pedestrian bridge design revealed severely high costs and com- However, it is difficult to see an integration of constructability and
plicated construction required of bridges designed by Calatrava structural rationality in works such as the Alamillo Bridge. Clearly
(Woodruff 2005). This has also been documented for the Alamillo striking, the bridge nonetheless created many difficulties to build,
Bridge (Aparicio and Casas 1997) and the Sundial Bridge in reflecting his “preference for bravura effect at the expense of function”
Redding, California, of similar form (Melnick 2004), as well as (Filler 2005). Unlike his predecessors, Calatrava does not integrate the
Calatrava’s most recent structures (Newcomb 2011). Simplified role of contractor or a sense of construction into his designs. As noted
analyses in the thesis also echo the lack of cohesion between the form in a compilation of his works, “Calatrava shows himself more in-
and structure of Calatrava’s bridges; this includes reference to slack terested in getting his work built rather than how it is built” and also
cables apparent during a visit to Calatrava’s Campo Volantin “exposes an impatience with the vagaries of construction” (Frampton
footbridge in Bilbao, Spain. et al. 1996). A result of this disconnect between design and con-
From our investigation of his Alamillo Bridge, we are able to struction is that Calatrava’s works frequently prove quite costly to
discern a mark of Calatrava’s work. His forms are undoubtedly build. Seville public officials recognized such construction and
aesthetically adventurous, signifying a daring approach to archi- budgetary challenges as the Alamillo Bridge concept took shape and
tecture. But with a thorough structural investigation of a signature canceled plans for a second, sister bridge across Cartuja Island
work of his such as the Alamillo Bridge, we can now address the (Pollalis 1999). Piers were not prohibited in the waterway at this
merits of his structural engineering. The Alamillo Bridge, while location and thus officials instead approved a more traditional mul-
dramatic, is shown to have an inefficient structural form. A clear tispan beam form. As for actual cost, the Alamillo Bridge is estimated
indication of this is the extreme weight of the pylon member to have cost $5,470 per m2 of deck according to Pollalis (1999). Cost
compared with other cable-stayed bridges. The Erasmus Bridge in comparisons between bridges are always a delicate matter, as bridges
Rotterdam, for example, serves as a good comparison because it rarely have similar design conditions and itemized costs are not
likewise features an inclined pylon but uses a single pair of backstays typically available. The Erasmus Bridge, for example, includes
to provide a restoring force under changing loads. The main span of several side spans, one of which is a bascule span, making comparison
the Erasmus Bridge is 284 m (932 ft), 40% longer than the main span with the Alamillo nonviable. For comparison, we therefore consider
of the Alamillo Bridge, and it carries a slightly smaller traffic load the cable-stayed East Huntington Bridge, an elegant bridge without
(two lanes each of auto, tram, pedestrian, and bicycle). The pylon of significant additional cost (Billington and Nazmy 1991). The East
the Erasmus Bridge is a steel box, yet at 1,800 t (1,984 tons) (Reusink Huntington Bridge features a vertical delta tower with main and back
and Kuijpers 1998), it uses only 10% more structural steel than the spans of 274 m (900 ft) and 185 m (608 ft), respectively (Grant 1987).
pylon of the Alamillo Bridge, a RC pylon encased in a steel shell. Despite a significantly longer main span, the cost of the East Hun-
Considering the total mass, the pylon of the Alamillo Bridge is tington Bridge is estimated at a much lower (inflation-adjusted)
18,860 t (20,794 tons), over 10 times the mass of the Erasmus Bridge $2,960 per m2 (Buchwalter 1994).
pylon. This is remarkably high, particularly given the significantly To be fair, Calatrava does not subscribe to this way of thinking,
longer span of the Erasmus Bridge. Dividing these total masses by nor agree with such criticism: “A current problem is that in the world
the surface area of the main span deck gives a deck-adjusted pylon of criticism certain people are developing moral criteria about en-
mass measure of 2:947 t=m2 ð0:302 ton=ft2 Þ for the Alamillo gineering. The laws of statics say that there are things you can do and
Bridge, 16 times greater than the Erasmus pylon at 0:281 t=m2 can’t do. But the new moralists say there are things you shouldn’t
ð0:029 ton=ft2 Þ. do. . . . Why is it not permissible when it is possible? If you asked me
This enormous material requirement appears to arise directly out why, for example, I designed columns in the shape of my hand, I
of the structural form. Our results show the bending moment en- would say because it is possible. That is what engineering is: the art
velope is balanced reasonably well about zero, suggesting that cable of the possible” (Lyall 1992).
posttensioning forces were selected appropriately. Longitudinal Communities, owners, or developers may at times seek a distinc-
stress estimates also indicate reasonable component sizing for the tive structure where cost is of little or no concern. And this freedom
Huxtable, A. L. (2005). “Too much of a good thing?” Wall Street Journal, Pollalis, S. N. (1999). What is a bridge? The making of Calatrava’s Bridge
Dec. 8, D8. in Seville, MIT Press, Cambridge, MA.
Lopez, R. A. Y., and Halpin, D. W. (2000). “A computer simulation of Reusink, J., and Kuijpers, M. (1998). “Designing the Erasmus Bridge,
the construction of the Alamillo Bridge in Seville, Spain.” Proc., Rotterdam.” Struct. Eng. Int., 8(4), 275–277.
Construction Congress VI, ASCE, Reston, VA, 121–130. Saint, A. (2007). Architect and engineer: A study in sibling rivalry, Yale
Lyall, S. (1992). Santiago Calatrava, D. Sharp, ed., E&FN Spon, London. University Press, New Haven, CT.
Melnick, S. L. (2004). “Sun sculpture.” Modern Steel Construction, Woodruff, S. J. (2005). “A critical analysis of recently built pedestrian
24(10), 24–28. bridges.” M.S.E. thesis, Princeton Univ., Princeton, NJ.