Construction of the Salt River Bridge and Mitigation of Unforeseen Subsurface Conditions

Transcription

1 Construction of the Salt River Bridge and Mitigation of Unforeseen Subsurface Conditions ABSTRACT Duncan Paterson PE PhD, HDR Inc. (principal author) Steve Lorek, PE, HDR Inc. Doug Voegele, PE, HDR Inc. Eric Reuschling, PE, CSX Transportation The 740 Salt River Bridge in Katharyn, KY was originally constructed in 1905 with a 338ft centerpinned swing truss carrying a single track. The swing truss was fixed in place in the 1940 s. The bridge was selected for replacement in 2005 due to significant deterioration of the bridge, pin connections, and members. Several replacement alternates were evaluated and the selected alternate was four 85ft span deck girders. The design was chosen based on factors including geometric site constraints, yearly river level fluctuations, and mobility concerns. This design adds two concrete piers in the Salt River waterway. Construction was first completed successfully on the new southern pier. While attempting to install the northern pier pile foundation, however, a debris field consisting of stones, concrete, and other materials was discovered. After evaluating the results of a preliminary investigation and excavation, it is believed that the original pivot pier had partially collapsed during a flood event shortly after construction and was abandoned to the river. A new pier was then built on top of the remains. A reassessment of the alternates and additional geotechnical exploration indicated that the traditional pile foundation should be replaced with a micro-pile system that penetrates the debris field. Thus, the original pier location and superstructure layout were maintained. Presented herein is the evaluation of the original pin truss for replacement and demolition, the debris field discovery, reevaluation of superstructure alternatives, the foundation redesign including impacts of obstructions, a revised construction plan, and final construction of the bridge. In addition, a three dimensional representation of the construction phasing and lifting plan will supplement the construction portion of the presentation. All analysis and design were completed according to the current applicable edition of the American Railway Engineering and Maintenance-of-Way Association, Chapters 8 - Concrete Structures, and Chapter 15 - Steel Structures (1).

2 THE SALT RIVER BRIDGE The 740 Salt River Bridge in Katharyn, KY was originally constructed in 1905 with a 338 center-pinned swing truss, carrying a single track (Figure 1). The truss is located over the Salt River, a tributary to the Ohio River, Southwest of Louisville, KY (Figure 2). The effective construction period at normal pool elevation is from June through September. Each truss has seven 24ft panels of Figure 1. Swing Span varying depth out from the main pier and is connected at the crown by tension bars that were loaded during the stress reversal from opening the span. With the exception of the tension bars, the truss, floorbeams, and stringers are all built up sections fabricated from plates, angles, and lattice bars. Additionally, the main truss members all terminate at pin-connection joints. The North approach to the truss consists of one 40ft and two 60ft deck girder spans, whereas the South approach consists of twenty 12ft timber spans on a 2 curve. The truss was fixed in place in 1940s and is no longer permitted to swing open. The existing substructure for the truss spans are N.T.S Figure 2. Bridge Location of masonry construction. The truss ends are supported by two approximately 6ft deep end support seats, whereas the pivot point is supported by a massive 20ft diameter center swing pier. Pier heights vary from 51ft to 64ft measuring from top of cap to top of spread footing. Due to significant deterioration of the bridge, pin connections and members, as well as evidence of impact damage from debris and derailments, the structure was selected for replacement in 2005 by its owner, CSX Transportation.

3 BRIDGE REPLACEMENT Alternates for bridge replacement were evaluated, including repairing the existing structure, replacement of entire bridge on new alignment, replacement with a new 340ft truss, replacement with two new 170ft through plate girders, and different variations of deck girder spans. Each evaluation included corresponding estimates of probable construction costs. Due to multiple factors, including geometric site constraints, yearly river level fluctuations, and mobility concerns, the recommended alternate was a four span deck girder with 85ft spans. The design consisted of a timber ballasted deck supported by four 65in depth deck girders. The recommended alternate would reuse the existing center pier, and add two additional piers within the Salt River waterway. Since the new piers would be located in a navigable portion of the Salt River, approval was obtained from multiple agencies including, the United States Coast Guard, the Army Corps of Engineer s Louisville District, the Kentucky Department of Environmental Protection (Division of Water), Bullitt County, Hardin County, and the City of West Point. Historic records indicated that the Salt River was once used for transportation but is now primarily used for recreational purposes. Modifications were made to the existing piers to accommodate the new spans. An approximately 4ft tall pedestal was added to the center pier to raise the cap to the new required elevation. Portions of the existing truss span end piers were removed to accept the new depth of the deck girders and associated bearings. Each pier was determined to be more than adequate for the load patterns for the new span configuration. Foundation Construction & Obstruction Discovery Construction activities commenced as planned in June of 2006 with pile driving at the new southern pier location. Once the pile driving was completed, the footer concrete was poured concurrently with modification work to the existing piers to accommodate the new spans. After completion of the southern pier and modification work, construction staging was set for the northern pier in June of It was quickly discovered, however, during driving of the initial HP 14x89 piles that the subsurface condition at

4 the northern Pier 2 was vastly different than that previously encountered during subsurface exploration, or at the southern Pier. The piles were refusing at a depth of only 15 to 20 feet below pool level. Exploratory probing by the contractor suggested that the piles were refusing on a rather large anomaly within the riverbed. An initial underwater investigation led to the discovery of a debris field consisting of masonry stones, concrete, and other materials including metals. As a result, the swing span was left in service while the pier construction was immediately halted to investigate the obstruction. Divers experienced in bridge inspection explored the extents of the obstruction by direct measurement of the object where it was exposed, and by probing beneath the mud-line. In addition, geotechnical borings were advanced through the obstruction to obtain an estimation of the obstruction thickness as well as to help ascertain the extent of the debris field, particularly given that much of the obstruction Figure 3. Original Gear was below the mud-line. Portions of the debris field were excavated, as well. Among the debris field was the original rack and gearing for the swing span which was salvaged for recycling (Figure 3). Additionally as a part of the investigation, local historians were contacted about the possible origin of the obstruction. The combined historical background and excavation of portions of the debris field lead to the conclusion that the original pivot pier had partially collapsed during a flood event shortly after its initial construction. Although unclearly documented in the full project file, the original center pier appears to have been abandoned to the river while a new pier was built on top of the remains. The extent of the debris field incorporated an approximate volume of 15ft depth x 35ft width x 25ft length, but was of oblong shape (Figure 4). The debris field was positioned such that piling could not be shifted and the new pier (as designed) would have to be relocated. Alternately, the debris could be excavated or demolished. It was quickly determined that demolition of the obstruction using explosives would not be a viable option due to the proximity of a gas main line running through the river near the bridge. Furthermore, other methods such as the use of a hydraulic impact hammer would be overly time

5 Obstruction 15ft x 12ft Pier Footing Figure 4. Approximate Extents of Obstruction consuming given the lateral and vertical extent of the obstruction (over 20 feet thick in certain locations). Therefore, other alternates needed to be considered. Two additional considerations were necessary for discussing design options. First, Salt River has large fluctuations in river levels and has demonstrated scour on the existing piers. Second, the obstruction is presumably resting on alluvial sediments, rather than rock. Should the foundation for the new pier penetrate the debris field, it is conceivable that it would need to demonstrate an ability to carry any loads introduced by the obstruction, including lateral, vertical and moment loads. Thus, the obstruction s weight and potential effects on the substructure needed to be estimated. If exposed by scour, it was estimated that the obstruction would add nearly 2000 kips of axial load and an additional 3800kip-ft of overturning moment. These loads assume that the current location of the pier remains in the same location and that any load created by the settlement would be carried solely by the foundation. Design options A final complicating factor in consideration of the options was that the 85ft steel girder spans had already been fabricated while the southern pier was being constructed. The following represents an abbreviated list of options that were considered as alternates for the replacement bridge design once the obstruction was discovered.

6 Alternative Options: Straddle the obstruction with a structural footing Since the extents of the obstruction were investigated but not fully known, the width of the required extension would have been large and possibly eccentric (i.e. centerline pier would not be in line with centerline of footing). This created two undesirable effects, the increased possibility structural torsion movement and excessive cost for construction. Drilled Shafts Four drilled shafts would be used to replace the piles. The shafts would be advanced by coring through the debris field. There were three problems for this option. First, the required diameter of the shafts was too large for the equipment already mobilized and a coring bit of sufficient diameter was not available. Second, metal pieces of the former swing span gearing were discovered in multiple locations during the underwater inspection and in the additional geotechnical borings. Thus, there was no guarantee that coring equipment would be able to penetrate the debris field. Third, there were concerns about equipment losses or damage due to the steeply sloping sides of the obstruction. Modify Superstructure Design - Change the span length, or eliminate the second pier. Several options were considered (Figure 5): o o Revise span lengths to one 55ft span and one 115ft span to shift the pier location Design a new 170ft steel span as either a deck girder, Through Plate Girder (TPG), or a new truss span. o Rehabilitate the truss pins at the approximately 20 locations of severe deterioration and elongation of the tension bars. Each of these options proved to be excessively expensive, especially considering that the two new 85ft spans were already fabricated and were on sight.

7 Deck Girder Option (115ft) TPG Option (170ft) Deck Girder Option (170ft) Truss Option (170ft) Figure 5. Superstructure Alternates Pin Replacement Option Micropile Foundation Micropile foundation would be used to replace the piles. This option provided the greatest potential. From a construction standpoint, the relatively small diameter of the piles ( 7 inches) would be easier to install. From a design standpoint, multiple micro-piles could be installed within the same planned footprint for the pier, thereby providing adequate redundancy should one or more piles not fully penetrate the debris field. That is, the micropile system could be designed such that the whole system would still function properly without the risk of overstressing the micropiles if the obstruction could not be penetrated at one or more locations. Conversely, there is inherent risk for this approach. For example, an excessive number of micropiles may be unable to penetrate the obstruction. Or, a large number of nonfunctioning micropiles could introduce excessive lateral or torsional forces to the group. In other words, the system could still fail.

8 After considering the concerns of the contractor, the designer, and the owner, the micropile system was selected since it had the greatest opportunity for success and had the potential to limit project losses, both monetary and schedule related. FOUNDATION REDESIGN The original foundation redesign was performed by Teasley Services Group (Teasley) of Nashville, Tennessee as a subconsultant to Nicholson Construction Company. Utilizing the structural design loads and geotechnical design information provided by HDR Engineering, Teasley developed a micropile foundation system consisting of twenty, 7-inch diameter micropiles extending through the obstruction and socketed 15 feet into the underlying shale bedrock. Reinforcing of the micropile consisted of a inch outside diameter (O.D.) casing extending to the top of rock and a full length No. 18 (75 ksi) all thread center bar. In addition, a second No. 18 all thread bar would be located in the bond zone (lower 20 feet) of the pile. Teasley also designed a structural bracing system to fully transfer all axial and lateral loads from the pile cap to the mud-line, with the micropiles designed for the full loads transferred to the mud-line by the bracing system. The bracing system was further designed to serve as the template for the micropile installation. A rendering of the bracing system is shown in Figure 6. Approval of the design was provided after an engineer s review. Complications arose, however, after the contract was rebid by CSXT. Figure 6. Micropile Rendering The foundation subcontractor listed by the successful general contractor did not have the required equipment to install the micropile foundation system. As designed, a concentric underreamer was required to enlarge the borehole once the bit had passed through the bracing system as designed. Therefore a smaller diameter casing was necessary to accommodate the new contractor. As such, a reduced casing diameter of inches and a single 2.5-

9 inch diameter 150 ksi bar for the entire length of the micropile was proposed. This revised system made allowances for the reduced diameter of the rock socket. The rock socket also needed to be extended an additional 7 feet, to a minimum length of 22 feet. One of the main complicating issues with the redesign was that the bracing system was designed specifically for a inch O.D. casing, and the system needed to be reevaluated to ensure that the load transfer would occur as intended. The proposed equivalent micropile foundation system was approved after it was determined that the redesign satisfied the design intent of the original micropile design. Proposed Truss Removal A particular challenge of railroad work on active rail lines is the time period in which projects can be completed. The cost of train rerouting, train stoppage and other effects can greatly affect the profitability of the rail line. As such, it is imperative that all demolition and preparations for new construction that can safely be accomplished prior to halting trains (down-time) be finished while the rail line is still active. For the truss structures, this included removal of all members not critical to the operation of the bridge in order to reduce the weight of truss components and expedite demolition of the remaining structure. This maximizes the amount of the bridge that can be removed in one crane pick. Understanding that the original span was designed as a swing truss, the truss spans were originally proposed to be separated into its two swing span portions. The two spans (North and South of the center swing pier) were proposed to be removed and new spans placed in two 96 hour work windows. As such it was necessary to analyze the truss for load from multiple crane pick points, and with a portion of its secondary members removed, particularly the lateral bracing system and the center top-chord tension bars. A 3-D model was initially created to analyze the trusses from multiple construction angles with both fixed and pinned connections. The model included the truss main members, floorbeams, stringers, and bottom chord members with or without the lateral bracing. Initially, the model was run with the complete swing span as constructed. Next, the model was reduced to one span and the lateral bracing was

10 removed (Figure 7). As expected, the lateral bending moment of the bottom chords was Figure 7. Idealized Model of Lateral Bracing greatly increased. The increase in out-of-plane bending, however, was still far below the allowable stress of the members even with full loading applied to the truss (operational load values). Additionally, stability from the increased out-of-plane deflection was not a great concern since the bottom chords act as tension members (in the closed position). COMPLETING CONSTRUCTION Installation of the micropile system was completed as planned (Figure 8). Three of the twenty micropiles were initially unable to penetrate the debris field but after reconfiguration of the drilling equipment, all piles were drilled per plan. After completion of the micropile system in August of 2008, normal construction resumed, with completion of the northern pier in September of Demolition of the swing span and erection of the new girder spans were completed in September of Thus, the bridge was completed a little over two years after construction was halted due to discovery of the Figure 8. Micropile Installation obstruction.

11 Figure 9. Construction Demolition and Erection Plan (Advanced Consulting Technology, Inc.)

12 One of the key elements in expediting the demolition of the truss and construction of the new spans was a 3-Dimnesional (3D) construction plan to help visualize the procedure. Advanced Consulting Technology, Inc. prepared an alternate construction erection plan for Figure 10. Truss Removal Hall Contracting of Kentucky. This included removing the truss in three sections while balancing the center pivot area as the end spans were removed (Figure 9). Hall Contracting proposed full demolition of the existing structure and erection of the new spans in a 7 day track outage (Figure 10). Aside from the substructure construction and modifications to the existing piers, the actual time from train stoppage to reopening the bridge to train traffic was approximately 6 days, including inclement weather and unanticipated construction delays. The completed project is shown in Figure 11. Figure 11. Completed Structure CONCLUSIONS AND HINDSIGHT With the benefit of perspective, would the design and construction of this bridge gone forward in the same manner? One argument is to insist borings are completed at all foundation locations. This is a customary recommendation of geotechnical engineers, but is also commonly not done due to the large expense, drilling at every foundation location for every bridge, over the long term. It is possible that the boring would have hit a portion of the masonry or metal and lead to increased scrutiny. The counter

13 argument is that there was no guarantee that a boring at the pier location would have been given special notice. That is, the boring might have missed the masonry, or metal. Furthermore, even if some of the debris had been cored, there is the possibility it would not have been treated as out of the ordinary. Routinely consulting a local historian is another consideration. In this instance it did lead to information indicating an abandoned pier is the likely cause of the obstruction. The futility of researching arcane information from local historians for each construction project, however, is readily apparent. That is not to say that historical information is not beneficial for particular projects with known concerns. Weighing the cost/benefit analysis of boring at each location and gathering historic information, therefore, still rests with the bridge owner. In this instance, it could be argued that not taking these steps cost the owner considerable cost and delays, including two separate mobilizations. But, it can not be stated conclusively that the additional information would have changed the outcome.

14 REFERENCES (1) Manual for Railway Engineering (2010), American Railway Engineering and Maintenance-of-Way Association (AREMA).

15 LIST OF FIGURES Figure 1. Swing Span (Photo) Figure 2. Bridge Location Figure 3. Original Gear (Photo) Figure 4. Approximate Extents of Obstruction Figure 5. Superstructure Alternates Figure 6. Micropile Rendering Figure 7. Idealized Model of Lateral Bracing Figure 8. Micropile Installation (Photo) Figure 9. Construction Demolition and Erection Plan (Advanced Consulting Technology, Inc.) Figure 10. Truss Removal Figure 11. Completed Structure