S.Sengupta, Director, Span Consultants Pvt. Ltd. Bangalore.
Zuari is one of the major rivers in India located near the western coast of the country very close to the sea. The river separates the state capital town Panaji from the airport Dabolim which provides access to the tourist state of Goa from across the world. A two–lane prestressed concrete cantilever box girder bridge 807m long was built in 1983 across this river as a part of national highway no. 17 connecting both sides. Subsequently, the vehicular traffic on this corridor increased manifold and the existing bridge had shown signs of serious structural distress. This prompted the authorities to plan for a four-lane new bridge on the river adjacent to the existing one designed on expressway standard. There exists navigational requirement for the crossing and tourism being one of the most important activities of the state, authorities have taken decision to construct a landmark bridge at this location in the form of a long span cable stayed bridge.
Site Investigations & Alignment OptionsPrior to any planning/design activities, detailed site investigations were carried out which include following:
- Topographic survey using computerized Total Station, Satellite Communication equipment (GPS) and river soundings.
- Subsoil investigation in the river bed, and on land.
- Traffic survey for economic and financial viability analysis.
Choice of Main Span
It was considered prudent to locate the main pylon foundations substantially away from the existing bridge foundations P3 & P6 in order to avoid any confusion in the movement of the vessels underneath the bridge since a series of well foundations close to navigational channel may lead to unintentional impact for any straying ship during fog/mist etc and cause major damage to the existing and proposed bridge foundations.
In the draft feasibility report, the central span of the cable stayed part had been proposed as 470m with side spans of 190m each.
Subsequently, during the detailed project preparation, it was felt that increase in the main span by 30m with marginal shifting of the pylon foundation towards Panaji side will solve the problem of tourist access to this pylon on top of which a revolving restaurant will be constructed. This has resulted in a central span of 500m and provision of a short length access jetty from the shore through which tourists can easily reach the high speed elevators leading to pylon top.
The span configuration finally adopted is 200m + 500m + 200m = 900 m. There are viaduct approach spans on each end of the cable stayed portion. Construction of high embankment in place of approach viaduct spans had to be ruled out in view of trechorous subsoil of black cotton or soft clay strata susceptible to long term settlement. The viaduct constitutes 4-lane bridge deck of 300 m length on each side with 10 spans of 30 m continuous PSC box girder. Thus the total length of crossing works out to 1500 m. Figure 1 shows the general layout plan and Figure 2 shows the elevation of the crossing.
Presently, this is the longest span cable stayed bridge designed in India. The previously completed longest span bridge was the 2nd Hooghly cable stayed bridge in Calcutta with a central span of 457m.
Deck Cross–Section & Post–Tensioning
Two alternative cross–sections of the cable stayed deck have been considered. One with an open cross–section with edge beams housing the stressing end of the stay cables with in-situ diaphragms and deck slab with footpath on each side, refer Figure 3. A similar cross–section was adopted in another recently constructed cable stayed bridge in Bangalore where the author was again the team leader of the design group. Footpath traffic is segregated from the vehicular traffic by reserving 1.0 m wide cable anchorage zone on each edge beam throughout the bridge length. To reduce the buffeting effect of wind on the vertical face of edge beam, a small wind fairing nose has been added which also acts as the fascia.
The spacing of the stay cables at the deck level is 10 m c/c while the diaphragms are spaced at 5 m c/c. The deck girder has been provided with post-tensioning cables near the central zone of the 500 m central span where the normal forces due to stay tension is not available. All the in-situ diaphragms are transversely post-tensioned with 2 nos. 1905 (19 x 0.5" dia HT strands) tendons.
Wind Tunnel TestWind tunnel test was conducted on a static section model in order to validate the static co-efficients of lift, drag and moment considered in the detailed design of the bridge. The test was conducted at the Wind Tunnel facilities of Indian Institute of Technology, (IITR) Roorkee.
Similar test was earlier carried out at the same facility for the deck section of 2nd Hooghly Cable Stayed bridge in Calcutta with a central span of 457m. A representative length of the deck of the bridge scaled at 1:50 was placed in the open circuit boundary layer wind tunnel. Forces and moments were measured on the model with load cells in smooth low-turbulence flow (turbulence intensity of 1%). A smoke generator was used to get a visual picture of the flow around the deck. The model was provided with tilting facility to fix the deck at the desired angle of attack of the incoming wind. Complete set of instrumentation was provided for measurement of wind velocity, drag force and moment, lift force and digital integrators with data acquisition system were used in recording of test data .
Force CoefficientsThe static force co-efficients for drag, lift and moment are given by 
CD = 0.5p V2B2L / FD
CL = 0.5p V2DL / FL
Cm= 0.5p V2BL / M
Á = mass density of air
V = mean wind velocity, m/s
L = length of the model, m
FD = measured drag force in N
FL = measured lift force in N
M = measured moment in N-m
Aerodynamic ConsiderationsThe bridge deck under consideration is having an aerodynamically streamlined section. The ratio of torsional to vertical bending frequency in this case is 1.82 which indicates a favorable separation of the torsional frequency from vertical bending frequency for the geometrical shape selected. The flutter and vortex oscillations were found to exist at only large angles of attack and at wind velocity much above the design velocity.
A similar study was earlier conducted in IITR wind tunnel facility on a sectional model of cable stayed bridge with a central span of 457m over river Hooghly in Calcutta, India. It was observed that turbulence tended to stabilise the flutter characteristics of the bridge. Also, as is well established, the wind incidence angle influenced the behavior and at angles greater than + 3°, the aerodynamic stability of the section in torsion was seen to improve marginally.
ObservationsThe recommended values of the force coefficients and the design forces obtained using these coefficients for a deck level wind speed of 150 kmph are given in Table 1.
Computer Simulation Carried Out and Comparison of results with Model TestAn analytical computerized wind simulation study of the Zuari bridge has also been carried out using a proprietary software DVMFLOW (developed by Cowi Consult of Denmark) and used in other cable stayed bridge projects. This gives very interesting results for design of future long span cable stayed bridge as highlighted below.
The crash barriers are assumed permeable for the flow and so in the DVMFLOW section model, they are represented by the horizontal and continuous top part only, as seen in Figure 5.
Wind Load CoefficientsThe expressions of the wind load coefficients in the model tests are as defined earlier. The full-scale width and depth of the bridge cross section are 25.5 m and 2.1 m respectively. The depth (D) including crash barriers is 3.05m.
In the model test, angles of attack varied between -6° and +6° with 1° spacing (only exception is that-5° does not exist). Lift coefficients are given in between -4° and +4°. Since the objective here is to perform a preliminary check on the results of model tests, only three angles of attack were modelled with DVMFLOW viz -3°, 0° and +3°. The results are shown in Figure 6. The drag coefficient calculated by DVMFLOW is normalized with width B, so the DVMFLOW values have been multiplied with B/D = 25.5m/3.05 m in order to be comparable to the model test values.
The drag coefficient calculated with DVMFLOW is approximately 10-15% lower than the model test results. The lift coefficients are smaller with DVMFLOW than in the tests. The lift slope for the model ests is dCL/d±=3.84, compared to dCL/d±= 4.76 for DVMFLOW (based on linear least-squares fits to the data points). The moment slopes are dCM/d±= 0.602 and 0.504 for model tests and DVMFLOW respectively.
Results from Wind Tunnel Test
- The flow pattern is similar to that for sharp-edged bodies and therefore, remains unaffected by the Reynolds number.
- The effect of turbulence on the force coefficients also appears to be insignificant. Test was also conducted on a separate model of bridge deck with fully closed soffit. The effect of the force co-efficient values are found to be negligible for both options.
- The ratio of the torsional to the vertical natural frequency of vibration of the bridge deck has been found to be 1.82 by analysis, which is reasonably high to ensure no adverse affect on the aerodynamic response of the bridge deck selected.
- The detailed design for the wind loading has been made based on the CL, CD & CM values as mentioned in  which is on a conservative side.
Pylon HeightPylon height decides the cable inclinations directly and it has been estimated through approximate analysis that the cable weight and girder deflection (and hence also the girder moment) are related to the function of (1/sin± cos””) where”” is the typical cable inclination to the horizontal.
To minimize this function, ± should be maintained approximately between 25° to 75°. Pylons in most of such cases generally have height between 1/5 and 1/7 of main span in a 3-span configuration. Raising of pylon height would lead to reduction in girder moments, in normal forces and in cable forces. However, it would also lead to increase in cost due to increased pylon height and longer lengths of stays. In the present case, after a few trials the cable inclination with horizontal has been maintained between 24° to 83° and pylon height above deck level approx. 1/5 of the main span.
Pylon Cross–SectionTwo alternative pylon configurations have been considered. In the first alternative, the cable planes are truly vertical and the legs have a gentle slope from top tie girder level (+ 71.8m) to deck level (+ 19.1m) and then converging in a diamond configuration below deck level towards well cap. In this configuration, the double–D shaped well foundation below the pylon legs will be of lesser size. The inclination of the legs below deck level however, works out little steeper due to the restricted height between the deck level and well cap.
In the second alternative, the pylon takes the shape of an inverted “Y.” The cable planes have a higher degree of inclination towards the top and spacing of the pylon legs at top of well cap level will be larger resulting in two separate circular well foundations connected by a heavy post-tensioned transverse girder at well cap level. After detailed technical evaluation and costs, this alternative was adopted.
The dead ends of the stay cable are housed in pylon top and stressing ends at girder soffit level. The pylon leg cross section all along the height is a hollow section in order to achieve optimality in the design and to provide uninterrupted inspection access inside pylon shaft from base level right upto the top anchorage zone. The design caters for detailed dynamic (seismic) modal analysis of the entire bridge according to seismic design code of Bureau of Indian Standard. The detailed design also caters for entire construction stage analysis of the pylon.
The governing load case for the design of pylons is the construction phase just before the closing pour of the deck. The pylon width in the longitudinal direction has been kept uniform throughout the height. A number of post-tensioned Mcalloy bars of 40/50 mm dia with UTS of 1030 MPa have been provided in the pylon top cable anchorage zone at various levels in both X and Y directions to take care of the splitting forces due to tension in the stay cables.
Design of Stay CablesThe maximum stress levels in the cables due to permanent loading is limited to 0.45 UTS as per international practice. At this stress level, relaxation effects can be neglected and the cable may be considered to behave elastically upto the level of design forces. The bridge has also been designed for the condition that any of the stay cables may get snapped (or inoperative) under full live load condition or due to possible replacement of any stay cable in future. Under such condition, the maximum allowable stress in the stay cables has been increased by 25%. The anchorage details for the stay cables in the pylon and deck have been carefully worked out after detailed review of the various available international stay cable systems in India.
Cost analysis and parametric study (stay spacing, size, length, costs etc) carried out indicates that increase in cable sizes beyond a certain limit does not either decrease girder moment nor the combined cable and girder costs. In fact, if the influence of cable size on its sag/length ratio is considered, it will be seen that increase in cable size beyond a certain limit may become counter -productive. Furthermore, in the context of very high unit cost of stay cables in India compared to other items, it is more important to optimize the sizes of all these stay cables.
Another important factor for this bridge is that it is located in one of the most corrosive climates in India, being very close to the sea. After detailed review of the entire international scenario, the stay cable system adopted is parallel strands of dia 15.2 mm (0.6") low relaxation grade, galvanized, greased and PE coated, housed in UV-resistant HDPE ducts, conforming to . The length of the various stay cables ranges from 85.3 m to 262.6m. Size varies from 26x0.6" strand to 65x0.6" strand (UTS respectively 7254 kN and 18,135 kN), HDPE size varying from 160 mm to 250 mm outer dia.
Selection of Stay Cable SystemSite assembled stays with galvanized, PE coated, greased strands have a high protection against corrosion and a good fatigue resistance. These are particularly feasible for long and large size cables.
Factory made stays with parallel wires are also suitable for long and large size cables, but they require very heavy equipment for installation/stressing compared to strand by strand installation with site assembled stays. Also, this system is supplied by only single propreitory agency in India and abroad and are relatively costlier. Corrosion protection system will be either cumbersome (e.g., cement grout) or too costly (e.g., Polyeurathene/Polybutadene). Very high accuracy in fabrication to predetermined length is also required. Site fabrication is not possible for major cable stayed bridges thereby rendering the transportation cost from factory to site very expensive.
Factory made stay with Carbon Fibre (non-metallic) material are suitable in extremely aggressive environment and has very high fatigue resistance. However, these cables are still prohibitively costly (at least 6 to 8 times of standard metallic stays). There are only very few cable stayed bridges of shorter lengths constructed with these types of stays.
Helical strands/locked coil ropes: experience has shown that in highly corrosive climate, these type of cables do not provide a long term durable solution and future replacement of stays poses a major problem. Also, they have low fatigue resistance due to conventional zinc socketting and have lower young’s modulus compared to parallel strand or parallel wire cables. Hence, these are also not considered for the proposed project in the extremely corrosive climate.
Considering all the above factors, parallel strand stays with PE coated strands conforming to guidelines in  have been selected for this bridge.
Special Aspects of Stay Cables
It is a standard practice to utilize high density polyethylene (HDPE) ducts as stay cable sheathing because of their UV resistance. For black HDPE pipes, 2% carbon black are normally used in the granules for the production of the duct.
However, nowadays, UV stabilized and colored co-extruded HDPE ducts are available which satisfy long-term durability and are also aesthetically attractive. Such ducts can be produced in any color, however the UV-resistance depends on the tone chosen. The following duct options have been examined alongside the black HDPE pipes.
- Coextruded pipes, whereby the outer 1.5-2.0 mm can be of any colour desired, the internal ring bearing the grouting pressure, however is made of black polyethylene.
- Fully colored HDPE pipes, whereby the entire cross-section consists of colored material. Caution must be taken with certain colors e.g., shiny red which are not suitable for long term UV exposure.
Cable VibrationsA problem associated with longer span stay-cable bridges is the wind-rain induced cable vibration, which may cause distress to the stay-cables. Investigations have shown that longitudinally or spirally ribbed and dimpled ducts reduce the vibration considerably. The neoprene damping device provided in the ring space between the cable and steel exit pipe of pylon and deck anchorages as well as the use of profiled ducts may not be always sufficient to suppress this wind-rain induced vibrations. Therefore provisions have been made in the design to enable implementation of complementary vibration suppressing measures viz attachment of friction/viscous dampers to the anchorage ends at deck level.
Design of Well FoundationsSubsoil borings carried out at the proposed locations of the two pylon foundations upto a depth of 50 m below river bed level reveals presence of sandy and clayey strata with soft laterite or weathered rock Amygdamoidal, hard soil with fractured rock at depth varying from (-) 29.5 m to (-) 43.0 m below bed level. The well foundations designed for the worst combination of all forces and maximum scour conditions have founding levels of (-) 39.00 m for well at PM and (-) 50.00 m for well at PP. The well dia are 14.00 m & 12.00 m respectively.
Both these foundations have been designed for a possible impact force from a barge with maximum DWT of 25,000 KN moving at a speed of 6 knots at a level of 1.0 m above High Tide Level. As additional protection, fender islands (riprap) around well foundations have been provided.
The results of detailed seismic (dynamic) analysis has also been taken into account in the design. However, finally the design is guided by static seismic condition. The grip lengths of the well foundations and the foundation levels caters for adequate factors of safety as per Indian Codal guidelines.
Design of Anchor Piers
For transfer of the uplift force from the backstays to the deck and from deck to the anchor pier, numbers of vertical post-tensioned tie down cables are provided in addition to the counterweight mobilized by ballasting inside the anchor chambers. The full counterweight against uplift force shall be available through a combined system of tie down cables of size 43 Ø 0.6" HT strands (with UTS of 11,997 kN), ballast weight, dead weight of anchor foundation and through a number of in-situ friction piles provided under the pier foundations. Figure 7 shows the structural arrangement. This ensures an adequate level of factor of safety against uplift for all possible loading conditions during the service life of the bridge.
Corrosion Protection to Post–Tensioned Elements
- PE coated strands for stay cable manufacture shall meet the following test requirement.
- Chemical resistance test as per ASTM G20 standard
- Chloride permeability test as per FHWA standard
- Impact test as per ASTM-G14 standard
- Abrasion resistance test as per ASTM-D968 standard
- Salt spray (fog) test as per ASTM-B117 standard.
- Epoxy coated strands used in stay cables conforms to ASTM A-882 “Standard specification for epoxy coated 7-write prestressing strand.” These strands shall be of the type in which the interstices of the strand are filled with epoxy and the strand shall be weldless, low relaxation grade.
- HDPE pipe: The UV resistance of the colored PE pipe shall be equal to the black ones. Light colour PE pipe having very high UV radiation resistance has been already developed and used in cable stayed bridges in Europe. The acceptance test for this item shall be as per ASTM–D3350 standard.
- Barrier 1 : galvanizing and greasing for entire length
- Barrier 2 : PE–coating for entire length
- Barrier 3 : external HDPE tube
Corrosion Protection to Post Tensioning Cables in SuperstructureExperience shows that metallic sheathing (in marine climate) is completely vulnerable to corrosion. In order to prevent this situation nd in accordance with the latest international practices, all the sheathing for post tensioning cables have been considered to be of corrugated HDPE duct which is available indigenously.
Additionally, anticorrosive admixtures are also proposed to be used in the cement grouting. These admixtures ensure anticorrosive properties of the cement grout and protects the steel from corrosion when used with cement, reduce the permeability of concrete and ensures long term durability.
Construction Method & Time FrameThe total construction of the project is divided into the following major items of work.
Construction of Well FoundationWell foundations are founded on rock level at approx. (-) 42m on Panaji side and approximately (-) 30m on Margao side. The bed level is at (-) 5m on Panaji side and (-) 0.5 m on Margao side. The subsoil strata consists of greyish black clay near bed level, followed by yellowish grey clay, boulders, grey white sand, yellowish white sand and greyish white rock at approx. (-) 40m to (-) 55m in case of Panaji side and (-) 28 m in case of Margao side. Jackdown method of sinking with prestressed rock anchors shall be made alongwith simultaneous grabbing through the dredge hole.
Construction of PylonOverall height of the pylon from top of well cap is 125.80 m. It is in the shape of inverted “Y”.
All the four inclined legs shall be constructed simultaneously with jump form shutters with lift of 3 to 4 m. The cross–section of each leg is a cellular box of overall size 6 m x 6 m at base. After joining at a height of 83.10 m above base, the pylon has a vertical leg (cellular section) of overall size 6 m x 9 m. Access is available inside the legs right upto the top where all the dead anchorages are housed. Temporary steel struts with horizontal hydraulic jacks shall be erected at various levels for geometric control while constructing the inclined legs. One of these pylons houses a revolving restaurant at its top, accessible by a high speed elevator from its base.
Construction of Cable Stayed Deck Portion
Stage–2Main span and side span erection will start simultaneously from each pylon in a cantilevering construction procedure. Each stage consists of in-situ segments of approx. 10m each along with PSC cross girders. Cantilevering Form Travellers (CFT) to be assembled on deck, already built.
Stage–4Erect back stays on both pylons.
Stage–5Proceed erection of deck with CFT’s with simultaneous erection and stressing of cables towards midspan and PP1/PM1 simultan– eously checking the deck profile.
Stage–6Closing pour at midspan.
Stage–7Finish deck with footpath, crash barrier, railing, WC etc. and fine tune all stay cable tensions to finished vertical profile.
Aesthetics Aspects of Design
For a pleasing appearance of such cable-stayed bridges, it is aesthetically important that the fascia runs undisturbed throughout the length on the outside the deck. For the adopted deck cross - section, the fascia comprises of the thickened edge of the deck slab of trapezoidal box section. Care has been taken to ensure that in the viaduct portion also, the fascia shall integrate with the above inorder to impart an uninterrupted fascia line allthrough the bridge length. The stressing anchorheads will be visible from outside, but again in the backdrop of the main girder depth, this will have insignificant visual impact.
Careful proportioning in the spans have been ensured in order to obtain harmony. The side spans which end with the backstay cables should be less than half of the main span which in this project is 0.40. This not only leads to visual harmony but also keep the stress changes in the backstays within allowable limits.
The inverted “Y” shaped tower with stay cables in a three dimensional plane impart a sense of stability and confidence in the vertical as well as horizontal plane. This will be particularly important for a large central span as in this project. The structural dimensions of the pylon legs and the well foundations have been judiciously worked out to ensure a feeling of robustness and a deliberate visual contrast to the thin stay cable network.
In order to avoid a dull aesthetic situation by providing all the exposed HDPE ducts as black tubes, it has been proposed to have brightly coloured HDPE ducts which are co-extruded with UV resistance. Large numbers of recent cable-stayed bridges have been built with such coloured ducts which provides an impressive contrast against the backdrop of the blue sky.
This has been considered in the design after detailed interaction with international suppliers for such items to ensure their availability for this project.
Illumination of such long span cable stayed bridge plays a very important role in highlighting the structure and creating a location of tourist interest. A major part of the energy required has been proposed to be obtained by installing permanent solar panels as has been done in a number of major bridge projects in Japan. Also it has been considered to have different illumination schemes during different seasons (rainbow effect). The detailed design of the computerised illumination scheme has considered above aspects.
Protection Against Barge Impact
The proposed pylon locations are close to the shoreline where depth of water is shallow (maximum depth of water varying form 6m to 8m). The latest international practice is to provide fender protection islands around the pylon foundation rather than attaching separate structural units (fenders) so that any barge going astray will run aground on the islands before reaching the bridge structure itself. In the present case, in addition to designing the well foundations for the barge impact forces, provision has been kept for construction of fender island (riprap) around the foundations as an additional safeguard.
The riprap built around the well foundations will have a core of ‘quarry run’ stone and two sizes of rock armor in PVC netted crates, an all-over layer to prevent the core from washing away and a layer of larger materials (e.g., stones) around the upper section to protect against wave action.
- Section Model Tests, “Aerodynamics of large bridge,” A. Larsen, Balkema, Rotterdam.
- IS: 875 (Part 3) - 1987: “Design Criteria for Wind Loads on Structures,” Bureau of Indian Standards, New Delhi.
- Model Force Moment & Pressure Measurements, “Low Speed Tunnel Testing,” John Wiley & Sons, Inc, New York
- “Acceptance of stay cable systems using prestressing steels,” fib bulletin 30, January 2005
- “Ship Collision with Bridges,” Structural Engineering, O.D. Larsen documents 4 of IABSE, 1993.