«Professional paper UDC 550.8 Investigation of the behavior of the cable-stayed bridge under test load Rastislav Mandi} 1, Gordana Had`i-Nikovi} 2 and ...»
GEOFIZIKA VOL. 28 2011
Investigation of the behavior of the cable-stayed
bridge under test load
Rastislav Mandi} 1, Gordana Had`i-Nikovi} 2 and Stanko ]ori} 1
1 Faculty of Civil Engineering, University of Belgrade, Serbia
2 Faculty of Mining and Geology, University of Belgrade, Serbia
Received 25 March 2011, in final form 18 May 2011
A field load testing is an effective method for understanding the behavior
and fundamental characteristics of cable-stayed bridges. This paper presents results of the behavior of the reconstructed cable-stayed bridge over river Dan- ube in Novi Sad under test load. The bridge was built between 1976 and 1981 and in 1999 was heavily damaged by two Tomahawk missiles. In 2003 process of the reconstruction began and at the end of 2005 Faculty of Civil Engi- neering, University of Belgrade was invited to test reconstructed bridge struc- ture. The results of static load testing presented herein include displacements, rotations and strains of the 351 m long main span. Vibrations of the bridge were obtained by impact load produced by heavy truck passing the bridge and the time history response of vertical accelerations was recorded. The frequency content of the signals was determined using Fourier transformation and five flexural natural periods were extracted. The geological structure and engineer- ing properties of the soil are also given. A finite element model of the bridge was made and a good agreement is achieved between the experimental and an- alytical results. These results have shown that the bridge is in the elastic state under the code-specified serviceability load, which indicates that the bridge has adequate load-carrying capacity and can be put safely into service.
Keywords: cable-stayed bridge, experimental analysis, numerical analysis
1. Introduction There are many papers related to the investigation of behavior of cable stayed bridges (Gattulli et al., 2002; Weon et al., 2007). Although sophisticated finite element methods are currently available for this purpose, successful analysis is strongly dependent on the experimental verification of the results (Cunha et al., 2001; Santos and Min, 2007; Wei-Xin et al., 2007). A field load test is an essential way to understand real behavior of such kind of bridges.
Test results of long span cable stayed bridges are very important and valuable data, which could be used by the other authors to check and improve their an- alytical methods for calculation the behavior of cable stayed bridges.
146 RASTISLAV MANDI] ET AL.: INVESTIGATION OF THE BEHAVIOR OF THE CABLE-STAYED...
Figure 1. Damaged bridge structure.
In this paper static and dynamic behavior of the reconstructed Sloboda Bridge over the Danube River in Novi Sad, under test load is presented. The tests were performed by the University of Belgrade to experimentally identify the most relevant bridge parameters and to correlate them with the corresponding parameters provided by the numerical finite element model.
The Sloboda Bridge was designed by Professor Nikola Hajdin and the Kirilo Savi} Institute from Belgrade. The main contractor was Mostogradnja (Belgrade). The bridge was built between 1975 and 1981. On April 4th, 1999 two Tomahawk missiles launched by NATO forces heavily damaged the bridge structure.
Initial assessments of the reusability of the demolished bridge were made in 1999 and 2000. In 2003, the process of reconstruction began, and it was completed in 2005. The details of how the bridge was dismantled and reconstructed have been published previously (Stipani}, editor, 2004; Hajdin et al., 2002). In September, 2005, The Faculty of Civil Engineering, University of Belgrade, was invited to test the reconstructed bridge structures using trial static and dynamic load. The main aim of the bridge testing was to certify the safety of the structure prior to putting it into operation (]ori}, 2005).
2. Description of the bridge
Detailed descriptions of the bridge have been published previously (Hajdin, 1979; Hajdin, 1983); only the most important data will be given here. The bridge is composed of a main cable bridge and approach structures on the left and right Danube banks. The main cable-stayed steel structure has a 60 + 60 + 351 + 60 + 60 m span arrangement. The pylons (60 m in height) are embedded into the deck structure. The cable stay system consists of six groups of stay cables, each composed of four stays positioned in the central vertical plane. At the time of the construction, the 351 m main span was the largest span in the world for a cable stayed bridge with stays in a single vertical plane.
147 GEOFIZIKA, VOL. 28, NO. 1, 2011, 145–160
Figure 2. Reconstructed Sloboda Bridge.
The deck structure consists of a three cell closed box girder with side cantilevers (Figure 3). The depth of the girder structure is approximately 3.8 m.
The total width of the orthotropic deck is 27.68 m. Two internal vertical webs enable a transfer of internal forces from the deck structure to the cables and then to the pylons.
Figure 3. Deck structure of the bridge.
The approach structures on the left Danube bank consist of two continuous pre-stressed girders with a total length of 301 m and four composite beams of 4 ´ 60 m. On the right river bank, there are three composite approach beams with spans of 3 ´ 60 m.
The properties of the soil in the area around the bridge foundation were investigated before and during construction for different purposes by the following groups: Geosonda of Belgrade, Industroproject of Zagreb, Civil Engineering Institute of Croatia, Zagreb, The Water Resources Development Institute Jaroslav ^erni from Belgrade, Hidrozavod DTD of Novi Sad and The Institute for Testing Materials of Serbia (IMS).
These investigations included stereoscopic analysis of aerial photographs, interpretation of satellite images, a quantitative geomorphologic study, detailed engineering geological mapping, exploratory drilling and trenching, cone 148 RASTISLAV MANDI] ET AL.: INVESTIGATION OF THE BEHAVIOR OF THE CABLE-STAYED...
penetration tests, hydrogeological and hydrometeorological observations, installation and monitoring of piezometers and inclinometers, pumping tests for soil permeability in-situ and laboratory soil mechanical tests (Sakovi}, 1993).
The bridge is located on the Danube bank slope. The heterogeneous lithology of the surface has been formed by both old and recent endogenic and exogenic geological processes, as well as by human impacts. The following effects of the exogenic processes are prevailing: surface rock weathering, slope erosion, and both eolian and fluvial processes.
There are three principal landforms, spacious plateau, varied slope and alluvion. Windblown loess on the plateau has greatly subdued the relief-toheight difference from 139 to 143 m.
The bridge area was built of Pliocene and Quaternary deposits; the oldest materials (to the exploration depth) are the Pliocene shallow lake sediments represented by grey and grey-yellow clayey marl and silt with lenses and intercalations of sand and clay. Smaller proportions of black soft clay, reddish marl and lignite seams also exist. All of these deposits are stratified and inclined to the Danube at angles from 2° to 5°. Upper portions of the deposits are altered by weathering and deteriorated over time. In addition, they are enriched with iron and manganese oxides and calcium carbonate.
The sediments composing the terrain to the explored depth are varied in composition and age. In addition, they were deposited under different sedimentation conditions (deep water, shallow water, swamp, dry land) and subsequently were exposed to intensive engineering geological processes; the result was high heterogeneity and anisotropy of the engineering-geological and hydrogeological properties in the given area.
Rocks are classified on the basis of engineering geological mapping and
core examination into the following seven geotechnical units, (Figure 4):
• Fresh lake deposits (1) – Marl and clay (1/1) – Sand and silt (1/2)
• Altered, deteriorated deposits (2) – Weathered marl (2/1) – Soft clay (2/2)
• Clayey sand (3) Quaternary • “Srem” sediments (4)
• Loess deposits (5)
• Recent slope erosion products (6)
• Alluvial deposit (7).
149 GEOFIZIKA, VOL. 28, NO. 1, 2011, 145–160
Figure 4. Engineering geological cross section.
The numbers in parentheses designate respective geotechnical units listed in the table of the mechanical properties of soil (Table 1).
Classification parameters, as well as values of unconfined compressive strength and shear strength by drain tests, were derived on the basis of statistical analysis of results of laboratory testing. The performed tri-axial compression and odometer tests were inadequate for proper statistical processing;
therefore additional test data were analyzed.
The highly heterogeneous hydraulic properties of sedimentary deposits and the activity of the currently evolving geologic processes have given rise to a varied hydraulic conductivity ratio of sediments, even within the same lithologic unit. Loess is a soil of high vertical hydraulic conductivity and appreciable porosity, which enables atmospheric precipitation to infiltrate without being retained. Srem series soils have varied hydrologic properties, from those of an aquifer, to an aquitard, to a local groundwater barrier. These soils form two aquifers at depth intervals 21.3–25.4 m and 30.8–35.5 m. The groundwater drainage in morasses and at springs indicates water movement from the loess plateau inland to the Danube bank slope in the bridge area. A sand-gravel series was deposited in shallow water after the Pliocene hiatus, and it act as sub-artesian aquifers. Three major sand aquifers have been found in the depth intervals 45.8–47.0 m, 57.2–59.2 m and 75.9–85.0 m. An excess water pressure was measured by piezometers in each aquifer. The aquifers are separated by thick impermeable marl and clay marl.
Geodynamic processes that have evolved in the bridge area are surface weathering (area of weathered clayey marl and marl), slidings, planar and linear erosion, wind erosion (loess deposition), fluvial erosion (the Danube right bank, consequent river channel shift to the south and deposition of sediment load transported by traction), and delluvial erosion.
A wider bridge area on the right side of the valley is designated as unstable, associated with a slide almost 40 m deep and of complex descent and differential block movement. The sliding process of variable intensity is evolving Table 1. Mechanical properties of the soil for different geotechnical units.
across a zone of 400–600 m. In the immediate bridge area, a block about 200 m wide is affected by sliding. The sliding body has a front at a steep, sub-vertical scarp, which is the boundary between the unstable and stable soils of the broad loess plateau in the bedrock. The slide bottom enters deep under the Danube bed and ascends at a low angle some 100–150 m from the right riverbank. Soil slides along the geological contact between fresh and weathered sediments. The shear surface has a slope angle of about 3° and it is located within the layer of saturated soft clay of high plasticity with low-strength parameters.
In the area affected by sliding, piers S17 and S18 through S21 are founded on piles. Soil consolidation involved a complex drainage system of two wells that are 6.5 m in diameter and 21–24 m deep, eight horizontal drains that are
0.47 m in diameter and 205 m in total length, and a tunnel outlet that is 1.5 m in diameter and 139 m long (Jelisavac et al., 2001).
Mechanical properties of the soil for different geotechnical units are presented in Table 1.
The foundation system is composed of reinforced bored piles with diameters Æ = 1200 mm Æ = 1500 mm and connected using large RC pile caps. For the foundations of the main bridge (between piers 14 and 19, Figure 2) the following arrangement was used: 10 Æ 1200 (pier 14), 8 Æ 1200 (pier 15), 32 Æ 1500 (pier 16 – pylon pier), 24 Æ 1500 (pier 17 – pylon pier on the right river bank), 8 Æ 1200 (pier 18), 10 Æ 1200 (pier 19). The length of piles varies between 20 m on the left Danube bank and 35 m on the right bank.
Bellow the piers and pile caps, the piles with large diameters were modeled with vertical beams loaded at the top and supported by deformable medium characterized by horizontal and vertical moduli of subgrade reaction. The soil under the bottom of the piles is simulated by vertical springs and the soil around the pile shafts is simulated by series of horizontal springs. The moduli of subgrade reaction were estimated using solution proposed by Vesi} (1961).
Modulus of elasticity of soil were determined from the results of unconfined sompression tests of soil. On the basis of that, for modeling of substructure soil, for static analysis was assumed E = 80 MPa. With regard to the type of structure, for dynamic analyses was used E = 120 MPa.
4. Numerical analysis
During the investigation studies, which were performed during 2000 by the Kirilo Savi} Institute, numerical simulations were performed to assess any damage to the foundation structure. Analyses of the nonlinear response of the piles and the foundation structures of the heavily damaged piers (S14 and S15 on the left river bank) were performed using geotechnical data (presented in this paper) and assumed collapse mechanism (Hajdin et al., 2001).
The numerical analysis of the rebuilt bridge under the test load was performed by the authors of this paper. It should be mentioned that the results of 152 RASTISLAV MANDI] ET AL.: INVESTIGATION OF THE BEHAVIOR OF THE CABLE-STAYED...
Figure 5. Finite element mesh of the bridge superstructure.