Electronic Proceedings of the 11th New York City Bridge Conference, New York City, August 21-22, 2023
General scheme design of ShiZiYang Suspension Bridge
J. B. Marcussen
COWI, NYC, USA
ABSTRACT: The ShiZiYang suspension bridge in the GuangDong province in China will when
constructed have a world record main span of 2180 m and carry an impressive 2 x 8 lanes of
traffic on the two-level truss girder. A bridge of this previously unmatched proportion requires
innovative design concepts to develop a feasible and constructable bridge. COWI carried out a
General Scheme Design in the initial design process to develop a concept for the overall config-
uration of suspended deck, tower, and cable system. Various alternatives for these main structural
components were defined and pros and cons evaluated. Some component alternatives are mutual
dependent such that for example the deck structural concept may define the cable system and type
of towers. Several concepts were compared by a quantitative cost comparison to determine the
preferred concept which was subsequently detailed further.
1 INTRODUCTION
The ShiZiYang suspension bridge in the GuangDong province in China will when constructed
have a world record main span of 2180 m and carry an impressive 2 x 8 lanes of traffic on the
two-level truss girder. The ShiZiYang Bridge will become the longest two-level truss-girder sus-
pension bridge in the world and will set records in span, load, deck width, and main cable diame-
ter.
A bridge of this previously unmatched proportion requires innovative design concepts to de-
velop a feasible and constructable bridge. With the technical challenges in mind, it was decided
to engage three bridge design companies to work simultaneously in the General Scheme Design
stage, which is an early design phase (Jun, X et al. 2022). The three companies are:
• CCCC Highway Consultants Co. Ltd., China (HPDI), package bid winner.
• China Railway Major Bridge Reconnaissance & Design Institute Co., Ltd. (BRDI), pack-
age bid winner.
• COWI, design consultant of the main bridge.
The purposes of the General Scheme Design are to decide main technical parameters, collect
design basis, perform subject studies, and most importantly make conceptual designs.
To ensure that the most suitable concept for the main bridge is determined, the three companies
separately investigated various bridge concepts to derive each their preferred concept. In this pa-
per, the General Scheme Design developed by COWI is presented. It noted that this does not
correspond to the bridge final design, which is not done by COWI and is still in progress.
The process for developing the General Scheme Design follows the following steps:
1. Provide alternatives for the main structural components suspended deck, cable system,
towers, anchor blocks and evaluate these independently.
2. Determine how the structural components are mutual dependent to determine multiple
bridge concepts that include choices for each structural component. Evaluate these con-
cepts qualitatively.
3. Determine the preferred concept based on a quantitative approach where the costs of
the concepts are compared.
4. Detail the preferred concept and the structural components.
The structure of the present manuscript follows the order of the items listed above.
2 BASIS
Certain global design parameters determined in a Baseline Design for the bridge were decided at
the onset of the General Scheme Design. This includes overall alignment, tower positions, and
double level 16 traffic lane configuration. Main structural concepts and other parameters could be
changed compared to the Baseline Design.
Figure 1 shows the plan and elevation for General Scheme Design and shows the main bridge
overall dimensions, while Figure 2 shows the suspended deck cross section.
Figure 1. Elevation and Plan of ShiZiYang Main Bridge.
Figure 2. Suspended deck cross section.
The basis and loads for the design are based on Chinese codes, environmental conditions, and
traffic requirements etc. In terms of the structural verification methodology and load combina-
tions, the Chinese designers BRDI and HPDI made their design according to Chinese codes,
whereas COWI's design was based on Eurocode. It was considered as an additional value of the
General Scheme Design that the bridge was analyzed based on these two different code systems
particularly since this bridge is beyond what codes are generally prepared to cover.
2
3 KEY DESIGN DRIVERS
Key design drivers are the most significant design features that will likely be controlling the de-
sign or features that shall be considered in the early phase. Design driver are fundamentally main
focus areas that have a consequence on the cost, time, and quality. The objective is to identify the
main design drivers in the beginning of the concept design phase to achieve a robust and reliable
concept design which is cost effective and where future risks can be mitigated early and controlled
during the subsequently design phases.
The General Scheme Design has been based on considerations of the various general demands
to a modern state-of-the-art bridge design, such as:
• Weight optimized bridge deck design
• Strength and stability as load carrying element for the roadway traffic
• Robust structural design solutions
• Aeroelastic wind stability - Extreme winds
• Durability and long-term performance
• Economic efficiency & life cycle cost
• Sustainability
• Reliability and predictability
• Aesthetics (landmark)
• Functional requirements
• Soil conditions
• Constructability
4 STRUCTURAL CONCEPT DEVELOPMENT
4.1 General considerations
Key to the success of any long span suspension bridge is minimum weight, so starting with the
deck, several deck arrangements were examined to explore opportunities for weight saving. Fac-
tors affecting operation and maintenance as well as the visual appearance were considered con-
currently. Minimizing long term maintenance requirements is a critical factor, particularly on such
a busy and long structure, and access for easy maintenance is considered a high priority.
The main cable sag profile (distance between towers divided by vertical height of cable curve)
is an important parameter as it has a significant influence on the cable tension and thus the size of
the cables, anchorages, and towers. A span to sag ratio of 9 is most commonly used for suspension
bridges and considered the default value. If choosing a higher ratio of say 9.5 this has the effect
of reducing tower height and cable length but also increasing cable tension. The span-to-sag ratio
also affects the bridge behavior since the overall stiffness of the system is related to cable sag, and
the appearance of the bridge is also similarly affected. Therefore, the concept development has
considered this range of options to evaluate the magnitude of the effects and identify the preferred
configuration.
Considering the challenges of cable compaction leads to a practical limit on the main cable
diameter of about 1.5 m. This is significantly larger than any suspension bridge cable built to date
but is considered feasible by those with experience of suspension bridge construction. Currently
the largest cable diameter applied is 1.3m on the WuFengShan Yangtze River bridge (road and
railway combined located in JiangSu province with a main span of 1092m) with 4 railway tracks
on lower deck+8 road lanes on upper deck. With only two main cables (one on each side) it is
clear that the cable diameter will be close to the 1.5 m limit, even with high strength wires. Thus,
consideration has been given to bridge configurations involving two, three and four cables in the
cross section. In the case of four cables, these would be arranged as two pairs of closely spaced
cables in two cable planes. In the case of three cables, these would involve an extra truss in the
girder running along the bridge centerline supported by the center cable.
The towers are the major visual feature of the bridge and will be very dominant features in the
landscape. It is therefore imperative that they are not only functional and efficient, but also beau-
tiful and elegantly shaped. The desire for a landmark bridge is to conceive something unique and
perhaps unusual as the defining feature for this bridge. There is a desire that the ShiZiYang bridge
should be instantly recognizable with a characteristically unique tower shape, and this has strongly
3
influenced the design development. Thus, in parallel with considering the pros and cons of two or
three cable planes, both H-shaped and A-shaped towers have been considered with either vertical
or inclined cable planes. In the case of the 3-cable design, the center cable plane is always vertical,
but the two outer cable planes could be either vertical or inclined, and this leads to the possibility
of A-shaped tower options.
Aerodynamic behavior is of paramount importance for long span suspension bridge design, and
the behavior in wind is influenced by the cable arrangement, the shape and nature of the bridge
girder and the distribution of bridge mass. All of these have been considered in the concept design
development to ensure that the behavior in wind is acceptable.
4.2 Component alternatives
Table 1 gives an overview of advantages and disadvantages for various alternatives for the differ-
ent main structural components.
Table 1. Alternatives for structural components.
Alternative Advantages Disadvantages
Main cables and suspended deck support
2 inclined cable planes, 2 trusses Aesthetically preferred Large diameter of main cables.
compared to vertical cable Complexity in erection due to
planes. pushing the main cables away
Provides lateral stiffness. from their vertical plane.
2 vertical cable planes, 2 trusses Easier execution than in- Large diameter of main cables.
clined cable planes.
2 cable planes with pairs of cables Feasible option if cable di- Complicated detailing of an-
ameter becomes too big chorages and saddles.
when single cables are Increases space requirements
used. and requires deck widening.
Construction complexity.
Cable aerodynamic instability
(wake galloping).
3 cable planes, 3 trusses Reduced weight of the Increased median width and
suspended deck as the ca- therefore requires wider deck.
ble system provides an ad- Complexity in erection.
ditional support.
Smaller diameter of main
cables and associated
components such as
hanger anchorages.
Towers
A-shaped towers Aesthetically preferred. Some complications in erec-
Possibly structural ad- tion of deck segments.
vantages due to increased Increased tower foundation
stiffness. footprint.
Weight saving compared
to H-shaped tower.
4
Alternative Advantages Disadvantages
Diamond shaped towers Small foundation foot- Likely too small dimensions at
print. the tower base in relations to
foundations.
Large deviation forces in the
tower due at the kink of the
tower legs.
Aesthetically inferior to A-
tower.
H-shaped towers Leads to suitable dimen- Traditional appearance.
sions at tower base in rela- Increased weight compared to
tion to foundation require- A-tower.
ments.
Suspended deck options
V-type trusses (Warren truss) Aesthetically preferred May lead to longer superstruc-
with approx. 45° diago- ture segments than preferred.
nals.
N- or M-type trusses Increased flexibility in Likely heavier than V-type
choosing segment lengths. truss.
Reduced view for lower deck
traffic.
Aesthetically inferior to V-
type truss.
Box elements in truss (closed section) Aesthetically preferred. Possibly heavier than open
Easy maintenance: inside section in case of 3 trusses
dehumidified and outside where demands are smaller.
surfaces easy to paint.
H-shaped truss elements (open section) Reduced weight in case of Reduced robustness.
3 trusses. Maintenance not as easy as for
Easy transfer of hanger box elements.
force, as truss diagonal
web can be placed directly
below hanger.
Decks with bottom plate (closed box) Dehumidification possible Heavier than open deck in
(inside painting omitted). case of 3 trusses.
Easy repainting on out-
side.
Superior Aerodynamic
performance
Decks without bottom plate (open deck) Weight saving in case of 3 Aerodynamic performance in-
cable planes. ferior to closed box.
Difficult maintenance and re-
painting, especially above
lower deck.
5
Alternative Advantages Disadvantages
Large vertical spacing between upper More headroom at lower Increased weight of diagonals.
and lower decks (12 m) deck improves driver
comfort and fire safety.
Higher stiffness.
Small vertical spacing between upper Reduced weight of diago- Likely reduced wind stability.
and lower decks (8 m) nals. Low headroom at lower deck
provides less fire safety.
4.3 Qualitative concept evaluation
It is evident from the alternatives presented in Table 1 that some choices are mutually dependent
between the various components. As previously mentioned, the key for a suspension bridge is
minimum weight so the starting point should be the suspended deck. Considering the deck:
• If two longitudinal trusses are chosen, there needs to be two cable planes and H-tower
shape would be the most conventional although A-tower shape would also work
• If three longitudinal trusses are chosen, three cable planes would be the obvious choice
although two cable planes could also work. With two cable planes, there is little benefit
of the third truss, as the cross beams at hanger locations shall still span the distance
between the outer trusses. In between hangers the cross beams could instead be sup-
ported by a central longitudinal beam at upper and lower deck. With three cable planes
both A-tower and H-tower would both be valid options.
There are other parameters in Table 1 that are independent on the fundamental choice of number
of trusses and cable planes (for example vertical space between decks). However, the fundamental
choice for the concept is number of cable planes and tower shape. The fundamental concept op-
tions are presented in Figure 3.
The four concept options shown in Figure 3 are qualitatively compared in Table 2.
Figure 3: Fundamental concept options.
6
Table 2. Qualitative comparison of concepts.
A) Two cable B) Two cable C) Three cable D) Three cable
planes, planes, planes, planes,
H-tower A-tower H-tower A-tower
Suspended Closed upper and lower deck can be Reduced weight / material saving (in case
deck used without weight penalty. no bottom plate of the upper and lower
Increased stiffness. deck).
Can be dehumidified. Reduced lifting capacity requirements.
Advantages
Access inside the deck.
Better appearance.
Wind load reduction due to more aero-
dynamic shape.
Panel handling/fabrication due to Additional detailing.
larger/heavier members and closed Deck width must be increased due to me-
decks. dian required by center main cables.
Disadvantages
Dehumidification of deck not possible (un-
less adding bottom panels).
Limited access under the deck plate.
Steel added for center truss is not utilized
efficiently for wind load cases.
Cable Most straight for- Possibly one Reduced cable Reduced cable
system ward. tower saddle type. diameter. diameter.
Advantages
Possibly one Architectural Lighter tower Lighter tower
tower saddle type. preference. saddles. saddles.
Slight improved Slight improved be-
behavior for wind. havior for wind.
Large cable Complexity of in- At least 2 differ- Complexity of in-
diameter. clined cable ent saddle types. clined cable planes.
Disadvantages
Heavy tower sad- planes. Architectural At least 2 different
dles. Large cable disadvantage. saddle types.
diameter. Architectural
Heavy tower disadvantage.
saddles.
Towers Easier Aesthetics. Easier Aesthetics.
construction. Lighter cross construction. Lighter cross beams.
Advantages
beams. Reduced tower quan-
Reduced tower tities.
quantities.
Visually not a More difficult to Lower cross beam Lower cross beam is
unique concept. construct. to support deck required to support
Tie beam required center truss re- deck center truss.
Disadvantages
at the base. quired.
Heavy transfer
structure at tower
top due to center
cable.
The anchor blocks are almost neutral regarding the 4 options and thus not decisive for the choice
of the preferred option. Three cables require more workmanship for the anchor blocks and the
approach span shall have an allowance/opening for the middle cable. The anchor block quantities
will be approximately the same for all options.
7
4.4 Quantitative concept evaluation and decision
A qualitative comparison is not sufficient to identify the preferred concept. The four concept op-
tions shown in Figure 3 and Table 2 are therefore in the following compared quantitatively based
on estimated relative fictive costs of the four concepts. The prices are relative only, since only the
main structural components are evaluated. The prices are fictive only as for example the unit
prices of structural steel, concrete etc. are not known. The relative price difference between the
unit prices have been estimated based on experience from projects in Europe. The overall philos-
ophy in the comparison is to assign all concept attributes to a financial value so that the options
can be compared by looking at the total cost only.
Such a comparison can only be very approximative at this concept design stage. An exact com-
parison would require detailed insight about contractor construction preferences, quantity prices,
salaries, equipment availability etc., and information about what value the owner and contractor
would assign to items as maintenance and operation, aesthetics, and sustainability. It does not
require the same detail to make a quantity comparison of the 4 concepts, which is shown in Table
3.
Table 3. Quantitative comparison of concepts (fictive prices).
A) B) C) D)
Two Two Three Three
cable cable cable cable
planes planes planes planes
H-tower A-tower H-tower A-tower
Material cost:
Suspended deck Material cost 4,340 4,340 4,089 4,089
Cables Material cost 7,151 7,152 6,881 6,882
Towers Material cost 4,800 4,710 4,800 4,710
Anchor blocks Material cost 7,640 7,640 7,640 7,640
Material cost total 23,931 23,842 23,410 23,321
Other cost:
Suspended deck Additional maintenance cost 0 0 204 204
as no dehumidification
Suspended deck Handling / fabrication of 130 130 0 0
heavier components and
closed box
Cables Inclined cables added 0 286 0 344
complexity
Cables Construction added complex- 0 0 550 551
ity due to 3 cable planes
Towers Propping of inclined legs in 0 63 0 63
construction
Towers Cross beam at tower top due 0 0 165 78
to center cable
Anchor blocks Anchor blocks complexity 0 0 100 100
General Visual appearance 300 0 600 400
Towers Cross beam supporting the 0 0 165 78
center truss
Suspended deck wind 0 0 500 500
Cables Risks associated with high di- 200 200 0 0
ameter main cables
Cost other than material cost, total 630 679 2,224 2,256
Total 24,561 24,521 25,633 25,577
8
Based on the material cost alone, option C and D have the smallest cost and would thus be
preferred from a pure material perspective. However, due to added other cost from additional
construction complexity due to 3 cable planes, additional maintenance due to open suspended
deck system, and visual appearance, options C and D become approximately 5% more expensive
than option A and B.
Comparing options A and B shows that they are almost equally favorable from a cost perspec-
tive. There is a small tower quantity saving for the A-tower, but the A-tower construction is more
complicated due to the higher inclination of the legs that requires more propping, and the inclined
cables also add construction complexity. The visual appearance is deemed to be favorable for the
A-tower and it is basically the visual appearance that is decisive for achieving a lower total cost
for option B. Since option B has the lowest fictive cost, option B is the chosen concept. This
choice is also based on the desire to produce something unique compared to the H-tower concept.
The two options are so close cost wise that only a more detailed study can determine, which would
be most favorable.
The order of magnitudes for the "other cost" (values in Table 3) are compared to the material
cost. Considering the chosen concept B, the percentage cost for each of the main structural com-
ponents suspended deck, cables, towers, anchor block, and "other cost" are presented in the pie
chart in Figure 4.
Figure 4. Cost percentage for main components of chosen concept B.
It is seen that the material cost makes up 97% of the total cost and only 3% cost is due to other
evaluated costs. The anchor blocks and the cables are by far the costliest structural items for the
bridge. The ratio between material cost and other cost could change slightly but the overall con-
clusion is that the material cost for this large-scale bridge is very dominant.
Since concept Option B is chosen, the main features for deck, towers, cables are:
• Suspended deck with two trusses
• Cable system with two cable planes and inclined cables
• A-shaped towers
The suspended deck, cable system, and towers are presented further in the subsequent chapters.
In addition, features that do not have a direct impact on the concept decisions above (e.g. part of
the suspended deck detailing) are also discussed.
5 ARCHITECTURAL APPRECIATION
The General Scheme Design is developed in cooperation with UK based Knight Architects. The
vision is to create a unique bridge for a unique part of the world where the bridge is situated.
Within the complex engineering constraints, the bridge shall become a fitting addition to the nat-
ural and built landscape, to provide an exceptional user spectator experience, and become an
9
important part of the future identity of the area. The scheme should capture the public imagination,
galvanize ambition, and contribute to unlock the wider value of the whole area. The key design
elements that offer the opportunity to achieve this are:
o Tower geometry and shaping
o Deck truss type and proportions
o Geometry of anchor blocks for the main cables
Figure 5 shows a bird’s eye view rendering of chosen concept, while Figure 6 shows rendering of
deck and bridge view of chosen concept.
Figure 5. Rendering of chosen concept, bird’s view.
Figure 6. Rendering of chosen concept, deck and bridge view.
While seeking high architectural quality, structural honesty is also essential as bridge design
should be driven by engineering and science. The triangular shape became the shape in which the
various structural elements relate to each other. The triangular shape occurs in the side view of
the bridge made by the deck, cables and towers and can also be found in the Warren Truss, A-
tower, as well as anchor blocks. The main features of the main structural components are described
in the following chapters.
10
6 SUSPENDED DECK
The suspended deck is made up of upper and lower closed girders connected by two planes of
diagonals in a Warren truss arrangement. The triangular shaped edges of the box girders break
their visual depth into two by creating two different shadows, providing aerodynamic benefits,
and increasing the perceived slenderness of the deck. Figure 7 shows the main features of the
suspended deck.
Warren truss
Chord member with triangular edge › Improved “clean” visual appearance
› Integrated part of the structural chord members consisting of triangles
› Improved wind performance (wind nose and › Simplified connections
closed box members › Reduced wind area
› Improved visual appearance. › Rectangular dehumidified diagonals.
Closed boxes
› Dehumidified
› Improved wind performance
› Improved fire safety
› Better access for maintenance
and inspection
› Improved visual appearance.
Figure 7. Suspended deck main features.
7 CABLE STRUCTURES
The main cables have a span to sag ratio of 9 and are parallel wire strand cables with strength of
2060 MPa. The highest possible steel strength is preferred to minimize the main cable diameter,
which will also allow for easier cable compaction and reduce size of related structures (anchor
blocks, saddles, cable clamps).
The main cable diameter is approx.1.4 m and the cross section consists of 372 strands formed
by 127 no. of 5.8 mm diameter wires. Hangers are parallel wire strands (PWS) with strength 1860
MPa.
The hanger system consists of single hangers with 20 m spacing. Single hangers have been
selected due to simpler arrangement of the cable clamps and hanger anchorages (compared to dual
hangers) without using unconventionally large hanger cross sections.
The hangers are connected to cable clamps by means of pinned connections at the top, and also
with pinned connections at the bridge deck eye plates.
Due to the A-shaped towers, the main cable planes are inclined, which provides additional lat-
eral stiffness compared to vertical cable planes, which is particularly beneficial for transverse
wind loads on the bridge (Jamal, A & Sundet, E, 2021).
11
8 TOWERS
The towers are because of their scale and verticality the elements of a suspension bridge with the
highest impact in how the crossing is perceived and remembered. The 360 m high A-shaped tow-
ers will be elegant and make the bridge unique.
In the past, a range of different tower forms have been proposed for cable supported bridges
(Gimsing, N & Georgakis, C, 3rd 2012), but long span suspension bridges are by far most com-
monly designed with H-shaped towers.
The A-shaped tower is chosen partly for aesthetic reasons, but it also has structural advantages
in comparison with the conventional H-shaped tower in terms of higher stiffness and material
savings due to smaller bending moment demands (Proverbio, M et al., 2022).
Figure 8 shows rendering of tower concept from bridge deck while Figure 9 shows rendering
of tower at night.
With the aim of making the tower composition clear and clean, a faceted dome made of trans-
lucent fiber reinforced polymer (FRP) is placed above the main cable saddles to extend the shape
and narrow the tower top. It will glow at night and act as a roof for a potential vantage point at
the tower tops and contribute to the interest of the bridge as a destination.
Figure 8. Rendering of chosen concept. View from bridge deck.
Figure 9. Rendering of chosen concept. Tower night view and the panoramic view of Pearl River Delta
from tower top (inset).
12
9 ANCHOR BLOCKS
The triangular shape of the anchor blocks in elevation is a result of the pure structural requirements
and is an excellent starting point from a visual point of view, Figure 10. A faceted geometry has
been used for the anchor blocks at both ends, to sculpt the base structural shape into different
surfaces that will vary in how they reflect the light. This helps to increase the visual quality and
their apparent slenderness.
Figure 10. Anchor blocks on East and West ends of the bridge respectively
The anchor block foundation consists of a circular slab constructed within a diaphragm wall with
a diameter of 125 m and a height of 30 m, Figure 11.
Figure 11. Anchor blocks elevation and plan
10 CONCLUSIONS
The ShiZiYang bridge with previously unmatched proportions requires innovative design con-
cepts to develop a feasible and constructable bridge. Based on qualitative and quantitative studies
COWI developed their preferred concept for the bridge. The preferred concept consists of as sus-
pended deck with two truss girders, two inclined cable planes and A shaped towers. Although a
saving in material could be achieved by adopting three trusses and three cable planes this would
not make up for the added complexity and maintenance cost. H-shaped towers would be equally
suitable and the decision between A-shaped and H-shaped towers would come down to a detailed
cost estimation also accounting for the value of aesthetics.
The material cost is found to make up 97% of the total cost and only 3% cost is due to other
evaluated costs relating to construction, risks, operation, aesthetics etc. The anchor blocks and the
cables are by far the costliest structural items for the bridge. The ratio between material cost and
other cost could change slightly during design developments, but the overall conclusion is that the
material cost for this large-scale bridge is very dominant.
13
REFERENCES
Gimsing, N & Georgakis, C, 3rd 2012. Cable supported bridges. Wiley.
Jamal, A & Sundet, E, 2021. Hålogaland Bridge - a landmark in arctic Norway. Structural Engineering
International.
Jun, X & YuGang, W & TaiKe, Z & Marcussen, J, 2022. ShiZiYang Bridge – General Scheme Design
Options for Mega Suspension Bridge. IABSE Congress Nanjing, China.
Proverbio, M & Jamal, A & Marcussen, J, 2022. Conceptual design of long-span suspension bridges: tower
structural forms. IABSE Congress Nanjing, China
14
