COMPARISON OF PREDICTED VERSUS OBSERVED STRUCTURAL DISPLACEMENTS OF EXISTING STRUCTURES AT THE PORT OF MIAMI

Transcription

1 COMPARISON OF PREDICTED VERSUS OBSERVED STRUCTURAL DISPLACEMENTS OF EXISTING STRUCTURES AT THE PORT OF MIAMI Anthony Bauer Gall Zeidler Consultants Vojtech Gall Gall Zeidler Consultants Philippe Bourdon Bouygues Civil Works Florida ABSTRACT The Port of Miami Tunnel Project is currently being constructed near downtown Miami, Florida to relieve congestion downtown due to port related traffic. The project consists of twin bored tunnels excavated by a hybrid Tunnel Boring Machine (TBM) within complex mixed face conditions beneath existing surface structures located on Dodge Island. Several structures were identified as critical within the zone of influence and were considered to be sensitive to ground settlements induced by the tunneling operation. Complex three-dimensional and two-dimensional finite element analyses were conducted incorporating the twin-tbm tunnels and existing Dodge Island structures. The effects of the TBM tunneling and the behavior of the structures due to the tunneling-induced ground settlements were explicitly modeled. This provided structural displacements which were compared to the respective capacities of the elements to evaluate the potential of damage to the structures. Accordingly, building elements were highlighted which were most sensitive to ground movements and were closely monitored. The results of the numerical modeling assured the owner that tunneling could be undertaken without the use of additional measures such as ground improvement or underpinning of the structures prior to the TBM reaching Dodge Island. Extensive instrumentation was installed on the existing structures and data was evaluated daily throughout the tunneling to ensure the structures were performing as predicted. A thorough comparison of observed structural displacements to predicted movements shows that the Dodge Island structures performed as or better than expected and validated that the structures were only affected within the range of permissible limits. INTRODUCTION The Port of Miami Tunnel project is currently under construction and consists of twin bored tunnels constructed between Watson Island and Dodge Island near downtown Miami (see Figure 1). Once completed, port related traffic will be diverted away from downtown city streets and have direct access to I-395 and I-95. To complete the project, a Public-Private Partnership (PPP) was established between the Florida Department of Transportation, Miami-Dade County, the City of Miami, Meridiam Infrastructure Finance, and Bouygues Travaux Publics as part of the design-build-finance-operateand-maintain (DBFOM) contract. Bouygues Civil Works Florida (BCWF) acted as the prime contractor for the project. To construct the tunnels, a specially designed hybrid Tunnel Boring Machine (TBM) was built by Herrenknecht in Germany and is currently the largest diameter soft ground TBM in the United States (see Figure 2). Adding additional challenges, no tunnel project of a similar scale has been attempted in the South Florida region. The 382

2 Comparison of Predicted Versus Observed Structural Displacements 383 Figure 1. Port of Miami Tunnel project overview Figure 2. Port of Miami Tunnel TBM cutter head (left) and trailing gear (right) TBM was launched on November 11, 2011 from Watson Island to begin the Eastbound tunnel boring and broke through at Watson Island on July 31, 2012 to complete the Eastbound tunnel construction. The machine was repositioned and launched again on October 29, 2012 to begin the Westbound tunnel with construction expected to be completed by the Spring of On the northern shore of Dodge Island, the TBM was required to pass beneath several existing structures currently in use by the various cruise lines that operate out of the Port of Miami. As the cruise lines are so influential in the local economy, the tunneling operations were not permitted to negatively impact the structures in any way. Several critical structures were identified as posing risk of damage due to tunnelinginduced settlements described below and shown in Figure 3 and Figure 4. Seawall & Bulkhead Corrugated steel sheet piling seawall, reinforced concrete pile cap, and tieback to sheet piling dead man anchor system. Originally constructed in the late 1950s and exhibits corrosion and deterioration of concrete and steel due to seawater exposure. Pedestrian Bridge Elevated steel structure providing access to loading gantries for cruise ship embarking. Supported by reinforced concrete columns on drilled pile foundations.

3 384 Geotechnical Instrumentation Settlement Control Figure 3: Pedestrian Bridge (left) and Shed #2 (right) on Dodge Island Figure 4. TBM orientation to existing structures on north shore of Dodge Island Shed #2 Open-air reinforced concrete frame structure currently used for cruise ship supply storage. Originally constructed in 1967, the building shows extensive superficial concrete deterioration and spalling due to environmental exposure. Seaman s Center Swimming Pool Recreational swimming pool within the influence zone. The owner was concerned with potential concrete cracking and water loss due to ground settlements. Additionally, the TBM was required to pass close beside drilled piles supporting a two-span bridge abutment. Other construction activities associated with the project required that the Mechanically Stabilized Earth (MSE) ramp structure be demolished prior to the TBM reaching the abutment location, resulting in the structure seen in Figure 5. While the TBM did not undermine the pile tips, lateral ground movements acting on the piles resulting in instability of the abutment were a concern. GEOLOGICAL CONDITIONS The geologic history of the Miami area provided challenging geotechnical conditions with respect to tunneling. A specially designed hybrid Tunnel Boring Machine

4 Comparison of Predicted Versus Observed Structural Displacements 385 Figure 5. Bridge abutment structure (left) and orientation to TBM tunnels (right) (TBM) was selected for the project which allowed for tunneling beneath the water table and channel through difficult mixed face conditions. Tunneling generally occurred through four geologic formations: Fort Thompson Formation Anastasia Formation Key Largo Formation Tamiami Formation The Key Largo Limestone, the Anastasia and Fort Thompson formations, and the Miami Limestone present on Dodge Island were formed from expansive coral reefs which covered the Florida Platform at the end of the Pliocene Epoch roughly 2.6 million years ago. Sediments from the eroding Appalachian Mountains to the north were deposited into natural grooves present in the coral reefs, which created localized areas of sands, silts and clays. The geotechnical investigation program conducted for the Port of Miami Tunnel Project identified eight (8) distinct ground strata in the area, with most of the tunneling being within Strata 6 through 8. Dodge Island is a man-made island formed of reclaimed land which was dredged from Biscayne Bay and deposited during the deepening of the Port of Miami. These upper fill layers generally consist of sand, silty sand, and silt and overlay the rock formations. Beneath, the Fort Thompson Formation is characterized by a pale orange to yellowish-grey fossil-rich wackestone/packstone containing corals, bryozoans and mollusks (Stratum 6). Underlying the Fort Thompson Formation is the porous, coquina and coquinoid limestone of the Anastasia Formation (a grainstone) and the fossil-rich, coralline Key Largo Limestone (a boundstone), which contains coral heads, bryozoans and mollusks encased in calcarenite (Stratum 7, Figure 6). This unit also contains isolated zones of loose, uncemented sands and silts much weaker than the surrounding limestone. The Anastasia and Key Largo formations may occur as interfingered lenses and layers within the basal Fort Thompson Formation. The ground investigation terminated in the Tamiami Formation (Stratum 8), a grey, porous grainstone with layers and lenses of shelly sands and sands interbedded with clays and silts. While the Fort Thompson and Tamiami Formation generally exhibit high degrees of cementation resulting in a strong competent limestone, the interbedded Anastasia and Key Largo Formations present grooves of uncemented sands and silts between the more competent rock material. Consequently, geotechnical parameters derived during the ground investigation of this stratum varied widely depending on whether the borehole penetrated the rock or the uncemented sands/silts. Modeling of this material proved to be difficult and predictions of surface settlements from the finite element analysis were highly dependent on the geotechnical parameters chosen for this

5 386 Geotechnical Instrumentation Settlement Control Figure 6. Samples of Stratum 7 material Key Largo Limestone stratum; however, due to the large variation of observed data, conservative estimations for compressive strength and Elastic Modulus were necessitated. North Shore Structures FINITE ELEMENT ANALYSIS To evaluate the influence of the TBM excavation on the existing structures, an independent three-dimensional ground structure interaction analysis was performed for the northern shore of Dodge Island. Shed #2 and the Seawall & Bulkhead were modeled explicitly to assess the behavior of the structures due to the tunneling-induced ground settlements. Additionally, structural element loadings were calculated which were compared to the design capacity to determine whether the structure was still within acceptable loading limits. The model was constructed using the Midas GTS 2012, v.1.1 software utilizing the DIANA solver by TNO DIANA BV for nonlinear analyses. Ground materials were modeled with three-dimensional, 4-node tetrahedral elements with a Mohr-Coulomb failure criterion. The extents of the modeling geometry and various ground strata are shown in Figure 7. Shed #2 is an open-air reinforced concrete frame structure consisting of prefabricated Double Tee roof beams resting on S-shaped support elements on drilled pile foundations (see Figure 3). The structure is nearly perpendicular to the tunnel alignment but does not provide structural connecting elements laterally between the frames besides the roof Double Tee s. Therefore, the structure behaves flexibly during the tunnel construction and is tolerant to differential displacements between the frames. In the numerical model, the structure was approximated as one-dimensional elastic beam elements and two-dimensional elastic plate elements (see Figure 7). Horizontal displacements of the Seawall towards the respective TBM face resulting in global stability issues of the Seawall and dead man system were of concern. Therefore, the Seawall, dead man, and tie backs were explicitly modeled with twodimensional elastic plate elements and one-dimensional tension-only elastic rod elements (see Figure 7). Bridge Abutment A two-dimensional finite element model was developed to analyze the bridge abutment behavior as the tunnels passed beside the drilled piles. In the area near the bridge abutment, soil material above the tunnel crowns and below the abutment ramp was

6 Comparison of Predicted Versus Observed Structural Displacements 387 Figure 7. Overall model geometry (left) and existing structures discretization (right) Figure 8. Overall geometry for bridge abutment model improved using Shallow Soil Mixing (SSM) techniques and was included in the model. Abutment structural elements and drilled piles were modeled by one-dimensional elastic beam elements, while ground materials were modeled using 3-node triangular and 4-note quadrilateral elements with a Mohr-Coulomb failure criterion (Figure 8). Assessing the TBM Excavation To simulate the TBM excavation, the Elastic Modulus of the material within the excavated volume along the alignment was reduced by a Softening Factor. This process models movement of ground material into the excavated area and into the face prior to installation of the tunnel lining and mobilizes the strength of the surrounding ground. While the softening factor is generally obtained from research or empirical relationships; in this case monitoring data was available for the Watson Island segment of the alignment and was used to calibrate the softening factor used in the model. A lower and upper bound Softening Factor of 30% and 50% were selected to incorporate the variety of ground settlements observed at Watson Island during the beginning of the TBM drive. INSTRUMENTATION OF EXISTING STRUCTURES Extensive monitoring installed on the existing structures provided the opportunity to observe the structural displacements during tunneling and compare to movements predicted by the finite element analysis. The Seawall & Bulkhead was monitored by

7 388 Geotechnical Instrumentation Settlement Control Figure 9. Instrumentation in Shed #2 (left) and automated total stations used for data collection (right) mounted three-dimensional optical prisms surveyed by an automated electronic theodolite total station. Horizontal movements of the sheet pile toe were observed with the use of multi-point borehole extensometers (MPBX s) installed behind the seawall. Mounted three-dimensional optical prisms were also installed throughout Shed #2 and the Passenger Bridge and were recorded twice per day by an automated total station when the TBM was within the vicinity. Rotations of the S-frame supports were monitored by two-dimensional electronic tilt meters, while existing concrete cracks and expansion joints were monitored by several crack meters (Figure 9). PREDICTED VERSUS OBSERVED STRUCTURAL DISPLACEMENTS North Shore Structures As of the publication deadline the Westbound TBM has not reached the northern shore structures and therefore, data is only available for the Eastbound tunnel drive. The finite element analysis predicted a relatively wide settlement trough due to the more competent rock material above the tunnel crown acting to disperse the surface settlements laterally across the structure, as shown in Figure 10. This behavior was confirmed by instrumentation installed on Dodge Island and led to only minor differential displacements and rotations of the structures as shown by the vertical structural settlements in Figure 11. As expected, settlements observed at the Seawall & Bulkhead structure were less than those observed at Shed #2 due to the stiffening effect of the steel sheet piling. Vertical settlements exhibited a wide, shallow trough resulting in minimal differential displacements as shown in Figure 12. Bridge Abutment The model predicted a maximum vertical displacement of 0.35 in after the Eastbound TBM has passed, whereas approximately 0.3 in were observed from monitoring data and closely followed the global behavior of the abutment structure reacting to the passing TBM (see Figure 13). As the Westbound TBM passed the structure, only slight deformations were actually observed by the installed instrumentation. The model predicted slightly more displacements by the structures as the second drive was constructed due to the interaction between the SSM improved ground and the non-improved ground.

8 Comparison of Predicted Versus Observed Structural Displacements 389 Figure 10. Vertical structural displacements predicted by 30% softening model of Shed #2 0.2 Shed #2 Ver cal Se lement Se lement, inches Horizontal Distance From EB TBM Center Line (feet) Observed Data - EB Passed 30% So ening Model 50% So ening Model -100 Figure 11. Comparison of predicted to observed structural settlements for Shed #2 0.1 Seawall Pile Cap Ver cal Se lement 0.0 Se lement, inches Horizontal Distance From EB TBM Center Line, feet Observed Data - EB Passed 30% So ening Model 50% So ening Model -150 Figure 12. Comparison of predicted to observed structural settlements for the Seawall & Bulkhead

9 390 Geotechnical Instrumentation Settlement Control Ver cal Displacement, inches Horizontal Distance Along Abutment, inches Original Loca on EB Passed Loca on WB Passed Loca on Observed Data - EB Passed Observed Data - WB Passed Note: Horizontal Displacements Scaled by 100 Figure 13. Comparison of predicted to observed structural displacements of bridge abutment CONCLUSIONS While the settlements observed from instrumentation of the structures proved to be less in magnitude than those predicted by the analysis, the settlement troughs displayed the wide and shallow trough anticipated by the three-dimensional finite element model. The two-dimensional analysis performed for the bridge abutment closely projected the observed behavior of the structure due to the TBM excavation adjacent to the drilled piles. This would suggest that the Softening Factor selected during the calibration exercise was relatively accurate; however, the geotechnical parameters chosen for Stratum 7 were conservative. As a conservative assumption, lower bound parameters were selected for this material; however, it has been shown at the Port of Miami Tunnel Project that Stratum 7 is more competent than was observed at the localized zones of silt and sand. The composite structure of the interbedded Anastasia and Key Largo Formation can provide competent support for the overlying strata and has not exhibited the tendency for high volume loss potential. Potential of damage to the existing Dodge Island structures was successfully evaluated using the structural displacements calculated from the numerical analyses, in spite of the challenging geotechnical conditions observed at the project site and lack of local related projects from which to draw experience. Due to the highly three-dimensional nature of the interaction between the tunneling-induced settlements and the movements of the northern shore structures, a three-dimensional numerical analysis was considered critical. By explicitly modeling the twin tunnels and the existing structures, the structural deformations caused by the underground operations on Shed #2 and the Seawall could be assessed. By undertaking the numerical analyses, the project owner was assured that the risk of damage to the structures was at a minimum. To ensure the structures would not be negatively impacted as a result of tunneling, extensive instrumentation was installed. During construction, monitoring data was evaluated daily to validate the results of the modeling and further reassure all parties involved of acceptable structural deformations. REFERENCES Attewell, P.B., Woodman, J.P Predicting the dynamics of ground settlement and its derivatives caused by tunnelling in soil. Ground Engineering, Boscardin, M.D., Cording, E.J Building Response to Excavation-Induced Settlement. Journal of Geotechnical Engineering, Mair, R.J., Taylor, R.N., Burland, J.B Prediction of ground movements and assessment of risk of building damage due to bored tunneling. International Symposium

10 Comparison of Predicted Versus Observed Structural Displacements 391 on Geotechnical Aspects of Underground Construction in Soft Ground, Rotterdam: Balkema. Midas GTS 2012 (v1.1) by MIDAS Information Technology Co., Ltd. Build: December 28, Solver: DIANA Solver by TNO DIANA BV. O Reilly, M.P., New, B.M Settlements above tunnels in the United Kingdom their magnitude and prediction. Tunneling 82, London: Institute of Mining & Metallurgy. Skempton, A.W., MacDonald, D.H The allowable settlement of buildings. Proceedings Institution of Civil Engineers Part III 5,