Repair and Maintenance of Post-Tensioned Concrete Bridges (2021)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

40 This chapter presents five case examples of post-tensioned bridge repairs and related mainte- nance actions performed by five different public agencies. An attempt was made to gather case examples from a diverse representation of states from different geographic areas. The partici- pating agencies were the Florida Department of Transportation, South Carolina Department of Transportation, Ohio Department of Transportation, Virginia Department of Transportation, and the City of Minneapolis Department of Public Works. The case examples presented in this chapter were solicited from agencies through the survey and through personal contacts of the authors and NCHRP project panel. Interviews were conducted March–July 2020. Case 1: Wonderwood Connector Bridge The Wonderwood Connector Bridge is a 0.7-mile-long post-tensioned I-girder structure that was initially built from 2002 to 2004 for a cost of $36.5 million (see Figure 22). It is located north- east of Jacksonville, Florida, and serves as State Route 116 over the Intracoastal Waterway. It is a critical portion of the Wonderwood Connector, a $140 million project connecting Highway 9A to Mayport, Florida. The bridge opened to the public on July 24, 2004. The structure has a 90 ft 9 in. wide bridge deck, serves both directions of traffic, and is sup- ported by eight lines of drop-in I-girders. Each girder has four bonded, internal tendons with 15 to 19 prestressing strands per tendon. Each tendon has a corrugated metal duct and was injected with a proprietary grout. The bridge is currently undergoing extensive repairs following a series of inspections that revealed some tendon corrosion and severe grouting issues. Construction Several issues with the grouting operations were noted during construction but they were not immediately addressed. The main area of concern was a spliced, three-span, continuous, 644-ft- long section over piers 9 through 12 at the bridge’s high point. The initial design did not provide enough vertical clearance to receive a Coast Guard navigation channel permit, requiring a major mid-construction design change. To facilitate the design change, counterweights were added in the side spans near the haunch segments over piers 10 and 11, and additional post-tensioning strands were added to some tendons. The tendon ducts had originally been sized for 15 strands, but to accommodate the counterweights, some ducts were installed with 18 to 19 strands. As a result, some tendons may have experienced restricted grout flow that contributed to the grout- ing problems detailed in the following section. Grouting operation reports detailed problems including grout cap and hose blow out, leaking grout, anchor head cracking, grout set, and grouting performed from both ends. Post-grouting inspections also noted sections of soft grout. At the time of construction, no remediative efforts were undertaken to address the soft grout. C H A P T E R 3 Case Examples

Case Examples 41   Grout Inspections Subsequent to identifying chloride intrusion problems in tendons in other bridges using the same grout, the bridge owner ordered an invasive inspection of a limited number of PT tendons to obtain some grout samples and to assess the tendons for corrosion. The initial investigation was conducted April–May 2012. Five grout caps were inspected and found to have satisfactory grout. After chipping away the grout to reveal the anchorage, very limited surface corrosion was found on the anchorage, wedges, and strands. Following the grout cap inspections, four exca- vations were conducted along girder 8 between piers 10 and 11 to inspect the galvanized steel duct and tendon grout condition. This section is the haunch and drop-in segment that required the structural modification to add strands. All four of the exposed metal ducts had one or more perforations from corrosion in the top of the duct. A portion of the duct was removed, and grout samples were drilled out of the tendons for chloride analysis and visual characterization. Three of the four exposed tendons had a layer of putty-like grout, and all four had a layer of soft, chalky grout with normal hard grout underneath. Each excavation was repaired according to FDOT repair specifications. On the discovery of corroded ducts and soft grout, the bridge owner initiated a second investigation to develop a better understanding of the corrosion issues. The inspection was conducted July–August 2012. The corrugated metal duct was exposed in an additional 32 loca- tions throughout the haunch and drop-in segments, and at least one excavation was performed in each girder. Of the 32 excavations, 16 exposed ducts had some degree of corrosion. Of those 16 ducts with corrosion, eight had severe corrosion. Fourteen of the excavation sites had hard grout throughout the duct. The rest had some combination of hard grout, putty grout, an unidentified black material, and soft chalky grout. Upon grout removal, the condition of the post-tensioning tendons was also inspected. No corrosion was observed on the strands at 17 of the excavation sites. Minor corrosion was noted at 13 locations, and moderate and severe cor- rosion were noted at one location each. Much of the investigation was focused on the haunch segments, where the worst duct corrosion and grout conditions were found. Chloride analysis on the grout samples taken showed a chloride content of 0.01% to 0.04% by mass of cement, which is below the specification limit of 0.08%. In some locations, grout was found to have moisture contents of 50% to 75%. Comprehensive Tendon Investigation Following the findings of the two initial investigations, the bridge owner ordered a compre- hensive, systematic inspection of every tendon to coordinate a repair strategy. A consultant specializing in nondestructive/minimally invasive inspection was contracted to perform the evaluation. Investigation of every tendon was conducted from September through October 2014. Ground penetrating radar was used to identify the tendon profiles, and impact echo was used to identify Photo credit: Kregg Diemer. Figure 22. Charles E. Bennett Wonderwood Connector Bridge.

42 Repair and Maintenance of Post-Tensioned Concrete Bridges grout irregularities. To characterize the grout irregularity, 87 small holes were drilled at various locations in the tendon profiles to visually characterize the grout with a borescope. The location and type of each grout irregularity was documented to coordinate a repair strategy. The study confirmed the findings of the earlier investigations, revealing large sections of voids and soft grout. One tendon was discovered to have a 20-foot-long void filled with water, attrib- uted to bleed water left from construction grouting operations. Overall, 3% of the 13,555 feet of tendons surveyed had voids or soft grout. Only 30% had completely hard grout, and 67% had a combination of soft and hard grout. The prevalence of the deficient grout in the tendons insti- gated an extensive repair response. Repair Strategy Following the report from the comprehensive tendon investigation, the bridge owner con- tracted a consultant to develop a repair strategy. The consultant studied the bridge’s history from the construction documents, as-built plans, and inspection reports. From these plans, the engineers computed time-dependent stresses to develop the ultimate and service capacity for the bridge assuming no structural irregularities. They then proceeded to evaluate the effects on ultimate and service limits of four tendon failure scenarios. On the basis of the inspection report, the bridge owner adopted a multiphase repair strategy. The first phase had three goals to remediate existing grout problems: 1. Remove the sections of soft grout. 2. Fill voids with new grout using vacuum grouting. 3. Use an impregnation repair method to provide supplementary corrosion protection. The second repair phase proposed adding additional external tendons to the bridge element to ensure that the bridge girders maintain the required prestress and ensure safety. To accomplish Phase 1, the FDOT initiated a research study, conducted by researchers at the University of Florida, to evaluate the best methodology for grout remediation (Torres and Potter 2020). Two methods of grout remediation were evaluated: hydro-demolition to remove soft grout, and grout drying to remediate soft grout without removal. To evaluate grout remedia- tion approaches, a mockup tendon assembly was constructed with nineteen 0.6-inch diameter seven-wire low-relaxation strands in a 4-inch duct. The duct was filled with alternating types of hard and soft grout. For hydro-demolition, inlet and outlet holes were cored into the side of the duct. Then a high-pressure water jet was used to blast the grout surrounding the inlet hole. After clearing the inlet area of grout, a high-pressure water tube was inserted into the hole and fed along the length of the tendon to blast the deficient grout and attempt to remove it from the duct. The study found that hydro-demolition was not effective at removing deficient grout. The study also assessed a method of drying the grout to remove residual moisture in the tendons. Tendon laboratory mock-ups were constructed using layers of soft and hardened grout to simulate the tendons found in the bridge. Dehumidified air was pumped through the tendon to remove moisture from the grout until the relative humidity of grout at the air outlet reached equilibrium with the input air (see Figure 23). Although the methodology proposed effectively dried the grout, the use of atmospheric air caused corrosion of the prestressing steel. Thus, it was recommended to use the grout-drying technique in conjunction with a corrosion inhibitor. As a result of the research project report, FDOT modified the proposed repair strategy to include grout drying instead of grout removal. Phase 1 repairs began in September 2017. The repair was planned as a staged effort and was completed in the summer of 2019. In the first stage, all grout voids were filled using

Case Examples 43   vacuum-assisted grouting. Stage two consisted of the grout-drying process. Grout drying began in July 2018 and continued through January 2019. The criteria for acceptance for drying of the soft grout was set as either the relative humidity from the outlet ports had reached 20% or the moisture content of a sample of soft grout taken from the tendon was 40% or less. The dried air flowed from the center of the span outwards to minimize leakage and maintain air flow throughout the length of the tendon. Air flow rate, air temperature, and relative humidity at the inlet and outlet were monitored throughout the drying process (see Figure 23). During grout drying, while the contractor was obtaining a grout sample to determine the moisture content, a failed strand due to corrosion was discovered in one of the tendons. The discovery of the failed strand validated the decision to include supplementary post-tensioning tendons in Phase 2 repairs. Following grout drying, PT tendons were impregnated with a proprietary corrosion inhibitor—a hydrocarbon and silicon polymer product. Phase 2 is in progress and involves the addition of supplementary external tendons and post- tensioning bars to ensure public safety and longevity of the bridge. A consultant performed the structural analysis to determine the prestressing parameters—a challenge because the bridge already had post-tensioning and is considered structurally indeterminate. The analysis was conducted using various scenarios of tendon failure. It was determined that two tendons would need to be added to each bridge girder. Post-tensioning bars will be added at diaphragms added between the girders at critical locations along the beam profile to ensure the additional prestressing will not exceed the capacity of the bridge. Phase 2 repairs were expected to begin in summer 2019 and continue through May 2020. Repairs are projected to cost $7.2 million. Summary Grouting problems noted during construction may have been the result of late design changes made to add strands to existing ducts, creating strand congestion inside the tendon. Following discoveries of soft grout and tendon issues in similar structures, investigations revealed exten- sive soft grout and tendon corrosion that could lead to tendon loss. To rehabilitate the bridge, the owner • Conducted extensive nondestructive testing to accurately define the bridge’s deficiencies and coordinate a repair strategy, • Relied on significant structural analysis to inform repair decisions, Photo credit: Florida DOT. Figure 23. Grout drying.

44 Repair and Maintenance of Post-Tensioned Concrete Bridges • Commissioned a research study to evaluate efficacy and feasibility of available repair methods, • Added additional external tendons to provide structural redundancy, and • Performed grout remediation and tendon impregnation to inhibit future corrosion. Overall, the project was a large collaborative effort between the bridge owner, consulting engineers, researchers, and the contractors to effectively remediate tendon corrosion in the Wonderwood Connector Bridge. Case 2: James B. Edwards Bridge: I-526 over the Wando River The I-526 bridges over the Wando River are a pair of precast, post-tensioned box-girder structures spanning between Mount Pleasant and Daniel Island, South Carolina. Construction of the structures occurred in 1985–1989. The structures were officially opened to public traffic in 1991, after a period of relative non-use during which they had served as access for another construction project. The structures are owned by the South Carolina Department of Transpor- tation (SCDOT). The eastbound and westbound structures are separate twin structures running parallel to each other for a total abutment-to-abutment length of 7,900 feet. Each structure consists of two symmetrical approach spans—constructed as span-by-span—flanking a main span unit constructed using the balanced cantilever method. Each structure contains 92 tendons—eight accessible, external tendons and 84 internal tendons. A unique pair of structures in the state of South Carolina at the time of construction, the Wando River bridges remain the only post-tensioned box-girder structures in the state’s bridge inventory. There existed little or no in-house post-tensioning expertise at the time— commonplace during this period when post-tensioned structures were still relatively rare and were beginning to be built around the country. Issues encountered in these structures and described in this case example include tendon ruptures (westbound structure), voids in the filler material (eastbound structure), issues with tendon encapsulation and detailing (both structures), and water intrusion (both structures). Construction records and as-built plans for the structures are incomplete or do not exist, further complicating the agency’s ability to respond to issues. Efforts have been undertaken over the years to develop as-built records for the structures, which have been noted as observably dif- ferent from the remaining construction plans. Construction and Early Structure Life In response to public concern regarding alleged poor construction practice by the prime contractor, a state-mandated investigation was undertaken in the 1980s. The investigation found no evidence warranting concern. This effort, however, provided an early review of constitutive materials and construction practices. The general superstructure condition was noted as good in internally conducted inspections during the early 2000s. Inspections during this period reported evidence of efflorescence, map cracking, some holes in the post-tensioning duct, minor diaphragm and deviator cracking, and some vertical cracking. Extensive deck patching was conducted to address spalling with exposed rebar. No major issues of concern specific to the PT system were noted as requiring address. A repeatedly documented issue was the presence of “pigeon droppings” inside the box that were vacuumed out. A summary of inspection reports from this period describes routine inspection

Case Examples 45   findings as containing “limited information due to form only and no corresponding report” (South Carolina Department of Transportation 2018). In 2006, some issues of concern were noted in an inspection of the westbound structure. In addition to fine spalling and map cracking, the following observations were made: • A PT block was exhibiting “crushing.” • Heavy spalling and patching work existed on the underside of two spans. • Large cracks were documented at two spans, one near an expansion joint. In-depth Investigations An asset management contract was initiated in 2010, which began externally managed inspec- tions. Significant water intrusion (running water) into the superstructure during rain events was observed during this period. In August 2010, the first indication of tendon corrosion was noted during a walk-through inspection of both structures. The following items were inspected during this walk-through: external HDPE tendons, anchorage pour backs (PT pour backs), HDPE duct couplers, HDPE duct and diaphragm connections, HDPE duct and deviator connections, grout vents and ports, segment joints, and other components pertinent to the PT system (Theryo 2010). Several issues of concern were noted, including small openings of approximately 1 inch in diameter in many external tendons’ HDPE duct (Figure 24); evidence of uncontrolled grout flow during construc- tion (Figure 25); inadequate corrosion protection of the tendons (Figure 26); geometry conflicts; improper vent connections (Figure 27); and evidence of corrosion at exposed PT bar anchorages (Figure 28). Of particular interest, the inspection team noted white, sometimes wet, material depositing at several locations along the external tendons and appearing to come out of the HDPE duct (Figure 29). These white deposits were noted along free lengths of the duct, at duct couplers, and at pour backs. All of the documented issues were found to occur in both eastbound and westbound structures. Additional routine inspections were performed on both structures on May 28, 2010; at this time, the structures were approximately 21 years old. The first suspicion of a possible loss of tendon force was noted in this inspection report, citing leakage and spalling in one pier of the eastbound structure, though poor joint material/installation was also a provided potential cause. No tendon loss was observed in the eastbound structure (Theryo 2010). Photo credit: Parsons Brinckerhoff. Figure 24. Holes in ducts in free length, assumed to be former grout ports.

Photo credit: Parsons Brinckerhoff. Figure 25. Evidence of leaking grout during construction. Photo credit: Parsons Brinckerhoff. Figure 26. Nonstandard covering of tendon. Photo credit: Parsons Brinckerhoff. Figure 27. Improper vent connections.

Case Examples 47   In light of these findings, a special in-depth, preliminary investigation was undertaken of the external tendons in both the eastbound and westbound structures (October 2011 Preliminary External Tendon Final Report). Inspections included laboratory analysis of samples taken from the structure, including chemical, petrographic, chloride, and water analysis. The investigation findings were multifold: • Testing of the white formations identified the substance as composed of principally calcium carbonate, suggesting that these formations occurred after water had infiltrated the post- tensioning system and leaked out at certain locations. As the water evaporated at the leak points, deposits of calcium carbonate remained in the form of stalagmites and stalactites. The formation of such accumulations likely took many years. • Ongoing water leakage was observed at several locations. In one case, a source of water infil- tration was identified as an improperly filled, deck-level grout vent tube. The report suggested that “many more grout vent tubes [were] not completely filled with grout (or perhaps the grout had degraded) extending directly to the top surface of the deck.” This could provide a direct pathway for air, water, and contaminant (including chlorides from known use of deicing salts) intrusion. Figure 29. White deposits at multiple locations. Photo credit: Parsons Brinckerhoff. Photo credit: Parsons Brinckerhoff. Figure 28. PT bars with corrosion.

48 Repair and Maintenance of Post-Tensioned Concrete Bridges Actions were taken in 2012–2013 to identify water penetration pathways and to address water intrusion, including injection of improperly filled vent ports, partial depth penetration patches to the deck, and application of methacrylate coating to the top of the deck. Attempts (multiple over the life of the bridge) to identify and address water intrusion pathways have only been partially effective. Communication (of air, water, contaminants, for example) between tendons remains an ongoing issue, with water penetration of the structure still occurring. Extensive nighttime inspections revealed hundreds of improperly constructed vent locations. Inspections to identify vents in need of address were difficult. In some cases, it was not easy to distinguish between polished aggregate and a vent location in need of remediation. And in some cases, vents permitted clear transmission from the deck level to tendon ducts(s). In visual inspections made with a borescope, sunlight and traffic were identifiable through vents that had not been fully grouted and sealed. In other cases, vents were found to contain poor-quality grout; it was suspected that improper grouting practices (i.e., not waiting until uniform consistency grout was outflowing before closing the vent), were used (M. K. Turner, interview, 2020). Incident 1: Tendon Rupture In 2016, a tendon rupture occurred in tendon “M1” of the westbound bridge. Analytical checks were performed, and a truck acceleration lane was closed, maintaining the two primary traffic lanes during the subsequent repair. SCDOT consulted with post-tensioning experts in both private firms and public agencies, including the Florida DOT, to inform its response to the situation. Hydro-demolition was performed to remove the tendon from the box girder. The damaged tendon was salvaged for inspection and testing. Corrosion of the M1 was evident, with pencil- point reductions apparent in some wires of the prestressing strand. Inspection of the grout within the tendon found some sections with visually observable, multiple colors, leading the agency to believe that some carbonation had occurred. The M1 tendon was replaced. Repairs took approximately four months to complete and cost approximately $1.8 million (Hall 2018). A replacement tendon was inserted in the same loca- tion, tensioned, and filled with a proprietary thixotropic nonshrink grout. The replacement tendon was installed to vent into the inside of the box girder in an effort to protect the new vents by using the physical barrier of the structure itself. Several testing and inspection efforts were undertaken to assess bridge health, including magnetic flux, corrosion potential testing, extensive borescope inspection, and destructive inspections during which tendons were cut open for assessment. An estimated 12 to 14 miles of tendon free lengths were inspected through various means. Inspections recorded voids, evidence of poor-quality grout, and other issues of concern. In September 2017, testing was conducted on the external tendons in the main spans of both the eastbound and westbound structures. Testing was conducted on all seven of the original tendons in the main span of the westbound structure. The eighth tendon was not tested because it had just been replaced earlier that year. On the basis of preliminary information, corrosion was found on two of the seven tendons of the westbound structure. When testing was nearly complete, a second tendon failure was identified (Hall 2018). Incident 2: Tendon Failure On May 14, 2018, a second tendon failure was identified during a weekly bridge inspection; the time of the tendon failure remains unknown. The weekly inspections—initiated in response

Case Examples 49   to the rupture of M1—included monitoring of the M4 tendon; this monitoring was important in identifying the tendon loss because no other indication had revealed its occurrence. The inspector observed the tendon move a noticeable, unexpected amount under manual excitation by the inspector. Both the M1 tendon that ruptured in October 2016 and the May 2018 M4 tendon failure occurred at the 7-foot-thick concrete diaphragm that the tendon passes through between the individual box girders. The response upon discovering the second failed tendon was immediate and included in-depth inspection, structural analysis checks, replacement of the affected tendon, and installation of redundancy tendons. Initial efforts to identify the cause of the tendon failure did not identify an obvious cause. An attempt to borescope the tendon had only limited success, encountering a blockage assumed to be grout. The tendon was drilled within 6 in. of the failure location, but no conclu- sions could be drawn. Slight evidence of corrosion was found; it remained unclear if this was the only mechanism affecting the tendon. Given the uncertainty, the agency elected to close the impacted structure to traffic to ensure the public’s safety. The affected tendon was replaced, tensioned, and grouted. In addition, two additional, supplementary tendons—initially detailed for the addition of future lanes—were installed for redundancy, with requisite analysis checks performed. Technical support was provided by multiple outfits to investigate issues, address concerns, provide on-site support, and ensure safety prior to the bridge reopening. Supporting enti- ties included FHWA South Carolina Division staff; FHWA structural engineering specialists deployed from Washington, D.C.; consultants with experience in evaluating PT issues; and staff from other state departments of transportation. Summary Following circumspect reviews of documentation in investigations subsequent to the second tendon failure in 2018, it was noted that these bridges had been “problematic almost from the beginning, especially with regards to water intrusion” (Hall 2018), a finding at the time con- sistent with other states’ reported issues with similar structures (from Hall 2018). The agency’s general approach to addressing issues with this structure has been multifold. In the absence, in some instances, of identifiable causes, the approach has focused on • Significant structural analysis, • Development of additional structural redundancy, and • Development and deployment of a rapid repair framework in-system, installed and ready to respond. Case 3: Veterans’ Glass Skyway Owned and maintained by the Ohio Department of Transportation (ODOT), the Veterans’ Glass City Skyway (Figure 30) over the Maumee River in Toledo, Ohio, serves as a critical section of I-280. Built in 2002–2007 at a total project cost of $237 million, it was the largest project undertaken at the time by ODOT. Among the many innovative features incorporated into the bridge design, the bridge wearing surface was cast integrally into the bridge segments. The girder was cast with a 2-inch-thick integral wearing course and 2.5 inches of concrete clear cover.

50 Repair and Maintenance of Post-Tensioned Concrete Bridges Construction During construction, two custom gantry cranes were used to facilitate the erection of each bridge span; each 150-ft span could be erected in approximately one week using the cranes. The cranes supported the spans during post-tensioning operations. On February 16, 2004, while erecting the 12th span, one crane fell, killing four workers and injuring four others. The failure and subsequent problems with the other gantry crane caused a 16-month pause in bridge con- struction, with ramifications to the post-tensioning operation. Because of the construction delay, many different shipments of Sika grout were used for post-tensioning operations. The bridge opened to the public June 24, 2007, one year after its scheduled opening. Tendon Investigation Following completion of construction, some of the grout used in the structure was identified by its manufacturer as potentially containing high chloride contents beyond the project speci- fications and the PTI guide specifications (limit of 0.08%). A technical memorandum, HIBT-1, issued by the Federal Highway Administration on November 23, 2011, identified 34 projects in 18 states affected by the manufacturer’s product (Gee 2011). In response, a 2012 study was conducted on the Veterans’ Glass Skyway Bridge to evaluate the potential for elevated chloride levels in the PT tendons. As part of the investigation, 148 longitu- dinal external tendons were opened to sample for chloride content. Several tendons were found to contain acid-soluble chloride exceeding the 0.08% limit; the maximum identified content was 0.116% in some cases. Further inspection of the post-tensioning tendons found only one strand with minor surface corrosion, which could have possibly been present at installation. No extensive repairs were required. Incident 1: Truck Fire Reveals Tendons On November 6, 2015, a semitruck had a tire blow out and it caught fire after it struck a guard rail. The truck stopped near the midpoint of the bridge. The ensuing fire damaged the surround- ing concrete, causing pop-outs and spalling of the deck. Approximately 100 feet of the concrete Photo credit: Ohio Department of Transportation. Figure 30. Veterans’ Glass Skyway.

Case Examples 51   deck underwent grinding in preparation for repair. Unexpectedly, grinding the deck exposed some of the post-tensioning ducts. Response: Where the ducts were broken and exposed, the tendons were repaired by pressure washing out the remaining soft grout and replacing with epoxy. The wet epoxy was covered with sand, and the epoxy surface was to be scratched to provide an adequate bonding surface for the concrete deck pour. Incident 2: Bridge Resurfacing Resurfacing of the bridge was scheduled to occur in 2020; the project is ongoing. The project replaces the bridge deck by milling 1 inch off of the concrete overlay and hydro-demolishing the next ½ inch to provide a roughened surface for resurfacing. As the contractor attempted demolition, however, the process exposed transverse post-tensioning tendons in the top flange of the box girder in more than 70 locations. Because the ducts were specified to have at least 4.5 inches of cover, it is unclear exactly how the ducts were exposed by milling. Several ducts have been identified in the wrong locations. Hydro-demolition has also broken open some of the ducts (see Figure 31). All locations where the ducts have been broken are coincident with voids or soft grout. Ducts filled with apparently good grout have not broken during hydro-demolition. It is hypothesized that the voids resulting from improper duct filling may be contributing to the duct damage during the hydro-demolition process. Locations of duct damage are undergoing repair. Repair is a multistep process: water blast- ing to remove loose debris, epoxy to seal, sand to provide a rough surface, and a final concrete cover layer. Lacking an alternative method of removing the concrete deck, the bridge owner’s milling, hydro-demolition, and repair approach is being employed for the entire bridge length where necessary. The repair strategy has led to concern that the bridge deck will have to be repaired again in 15 to 20 years. To facilitate later resurfacing and repair efforts, the bridge owner is cataloging thorough documentation of damage locations. (a) (b) (c) Photo credit: Ohio Department of Transportation. Figure 31. Exposed tendons during resurfacing: (a) severely damaged, (b) moderately damaged, and (c) undamaged.

52 Repair and Maintenance of Post-Tensioned Concrete Bridges Summary The Veterans’ Glass Skyway in Toledo, Ohio, serves a critical section of I-280 over the Maumee River. It features several innovative design concepts and a significant amount of PT. The following are pertinent details related to the repair and maintenance actions for the PT system of this structure: • Construction delays led to the use of a wide variety of grout batches and lots. • An innovative integrally cast concrete slab minimized the clear cover over tendons in the top slab. • Proactive invasive investigations revealed minimal chloride intrusion and corrosion in the tendons. • Bridge deck resurfacing exposed tendons in the top slab and revealed areas of soft grout. • The agency has developed a standard repair policy, although it will not prevent the issue from occurring in future repairs of this structure. Case 4: Varina-Enon Bridge: I-295 over the James River Owned and maintained by the Virginia Department of Transportation (VDOT), the Varina- Enon Bridge carries I-295 over the James River near Richmond, Virginia (see Figure  32). It was constructed in the late 1980s and opened to traffic in 1990. The structure is approx- imately 4,686 feet long with two approach spans flanking a main span. The main span, sup- ported by cable-stayed pylons, is composed of two single-cell precast box girders connected by a precast delta frame to form one structural unit. Each of the approach spans is composed of two separate single-cell precast box girders, with each girder carrying one direction of traffic. The specification at the time of construction called for a grout filler material for the post- tensioning tendons to be mixed on-site, composed of conventional cement, water, and expan- sive admixtures (Parsons Brinckerhoff 2010). Given the location and climate of the structure, the structure is expected to routinely experience moisture intrusion and condensation throughout its life (Parsons Brinckerhoff 2013). Early Structure Life Early inspections were conducted routinely and in accordance with NBIS requirements. Inspections in 2001 uncovered voids in a large quantity of tendons, identified primarily through Photo credit: Virginia Department of Transportation. Figure 32. Varina-Enon Bridge.

Case Examples 53   borescope inspection via vent tubes that had been incompletely filled with grout or that were completely empty. Evidence of bleed water and grout subsidence was noted. While it was unclear whether expansive grout materials (though permitted) were used by the contractor, later testing revealed the original grout material to be of poor quality, with a suspected high water-to-cement ratio beyond the maximum specified. Two efforts in 2003 and 2004 were undertaken to fill identified voids via vacuum grouting, though only approximately 50% were accessible. Repairs were made with three different pro- prietary grouts. A 2005 inspection identified one tendon “bent out of plane.” A 2007 inspection identified one failed tendon and another tendon of concern. Additional inspection and hammer sound- ing of all accessible tendons was performed. An area of concern was identified where a drain plug had been blocked with grout during construction, allowing water to collect within the structure and to submerge a tendon in acid water. The duct of the adjacent tendon was opened, revealing corrosion. The tendon was carefully de-tensioned and replaced. A total of two tendons were replaced. Of particular note, post-mortem inspections were performed on the removed tendons. In the failed tendon, corrosion was noted at the interface between the old grout material and the new material injected in 2003–2004. The use of dissimilar grouting materials was suspected of creating a macrocell corrosion site. In 2007, a research effort with the University of Wisconsin used magnetic flux to inspect all of the longitudinal tendons. Section loss identified by magnetic flux was confirmed in two tendons. The locations with section loss were opened and the duct was replaced with trans- parent duct so further corrosion could be observed. Eight locations were visually monitored over time. The rate of corrosion appeared to be very low. Additional observations indicated worsening structural integrity, including advanced degradation of the deck wearing surface (perhaps influencing the structure’s water-tightness), the presence of diagonal cracking in the girder webs, and significant cracking/separation at diaphragms and top deck. Conditions and specific issues were observed to deteriorate between inspections in 2007 and 2010, with a number of issues identified with potential structural and durability impli- cations to the post-tensioning system, including, in part • Shifting of a box segment at one pier, • Diagonal cracking in box segment webs, • Suspected strand slippage in external tendons, and • Suspected defects in additional tendons. In-depth Inspection (2010) These findings motivated the bridge owner, VDOT, to authorize an additional in-depth study by a consultant in 2010, to include recommendations for remediation. The intent of this effort was to provide expert opinion on a course of action on the basis of the existing inspection reports (primarily those from 2007 and 2010) and visual inspections of the identified critical deficiency items. In a parallel effort in 2010, the Federal Highway Administration conducted testing at Varina- Enon in an early effort to use magnetic flux leakage (MFL)—a technique evaluated for its effectiveness at the Turner Fairbank Highway Research Center in early 1990—to evaluate in a nondestructive manner the external tendons inside the box girders for evidence of defect. Locations of suspected defect were cut open for additional testing and inspection, grout defects

54 Repair and Maintenance of Post-Tensioned Concrete Bridges were repaired, and the duct was resealed with heat-shrink sleeves. The bridge owner is positive of the efficacy of MFL as a defect detection technique, commenting that all identified locations by MFL did reveal defect. The possibility of defects not detected by MFL was not evaluated. Inspectors noted that the external tendons were in fairly good condition, though there was evidence of PT duct connection repair occurring both during and after construction (see Figure 33 and Figure 34). The connection repairs also exhibited evidence of grout leaking during its injection, which may have necessitated the repairs in the first place. Unlike current standard practice, the inspectors noted that most of the duct couplers were not waterproof. Inspectors surmised that grout leaking during construction may be associated with void presence in the hardened grout. It is important to note that connections between lengths of HDPE duct have been greatly improved and standardized since this time. Inspectors noted that several strands had been exposed and exhibited some corrosion. In addition, some tendons contained a length of clear duct material, allowing direct observation of the tendon’s strands and evident strand corrosion (see Figure 35). Photo credit: Parsons Brinckerhoff. Figure 33. Repair connection and coupler at diaphragm. Photo credit: Parsons Brinckerhoff. Figure 34. Mid-length repaired tendon coupler.

Case Examples 55   Several tendons had been identified in the first 2010 inspection report as suspected of slip; these tendons had been suspected because of changes in the repair material, with an evident change in a crack width in the epoxy coating the connection, between inspections occurring in 2007 and 2010. These tendons were carefully inspected during the subsequent in-depth inspection, but slip- page could not be verified as the cause by inspecting the tendon mid-length because the crack in the repair material could have been caused by thermal changes within the structure. Inspection of the anchor (and wedges) and the trumpet was recommended. Though not inspected because of access issues, the vertical tendons were indicated as impor- tant for later consideration, given that the older generation of grout was likely used at the time of their construction, and given evidence of excessive bleed water in such tendons with these materials. Box Girder Cracking Structural member cracking of box-girder webs and diaphragms was also inspected in 2010. The structure contains two types of diaphragms over the piers: 1. Diaphragms over sliding bearings, and 2. Diaphragms integrally connected to the pier (with PT bars). The inspection report concluded that there existed no evidence that unwanted movement of the bridge structure had compromised structural integrity, but it recommended repair of the external tendon connections, as well as restoration of the corrosion protection system. No direct evidence of water intrusion or deficient grout was noted during this inspection (Varina-Enon Bridge I-295 over James River Final Report: Stage I Study Report for Evaluation of Bridge Defi- ciencies and Recommendations 2010). In-depth Investigation (2012) Grout voids, corrosion, and wire breaks observed and documented in regular and in-depth inspections from 1999 to 2012 triggered an additional assessment of the post-tensioning system. A field evaluation was performed in 2012 by Parsons Brinckerhoff, in collaboration with an Photo credit: Parsons Brinckerhoff. Figure 35. Clear duct for tendon observation.

56 Repair and Maintenance of Post-Tensioned Concrete Bridges NDE and corrosion specialist, Siva Corrosion Services, Inc., to assess the grout and prestressing strand in the external tendons and the vertical tendons. The evaluation examined 13 external tendon locations and 18 vertical PT bars. The inspected locations represent only a handful of the total 480 external tendons and 360 PT bars in the structure and were selected on the basis of previous inspections. Test locations were selected by VDOT on the basis of their perceived corrosion risk, focusing on external tendon high points and PT bar tendons in which the grout vent tubes were observed to be empty. External tendons were manually opened, removing the plastic duct at test locations. The con- ditions of the grout and strand were observed, noting section loss or wire breaks, if any. With the tendon open, the team performed field tests on the grout for resistivity and alkalinity. Samples were gathered for later laboratory testing of moisture, chloride, sulfate, and gypsum contents, as well as petrographic assessment. Exposed strands were tested for rate of corrosion (via polariza- tion resistance testing) and corrosion potential. In addition, evidence of corrosion was noted using the PTI corrosion classifications, observed by visual means. PT bars were inspected via the open grout vent tubes using a borescope. The general condition of the grout, including the presence of voids, was documented and water samples were taken, if applicable. At all but two external tendon locations, the strands were observed to have Class 2 corrosion (by visual classification), or light surface rust. In one location, a tendon was observed to have one broken wire (see Figure 36); another location contained four broken wires (Figure 37). Section loss was also exhibited by adjacent wires in both locations. These two locations were the same locations where broken wires had been identified in a 2007 National Transportation Research Center visual survey; since 2007, the number of broken wires and the percentage of section loss had progressed (Parsons Brinckerhoff 2013). In general, the rate of corrosion testing for the strands, however, did not corroborate visual findings. These tests assessed the corrosion rates for tested tendons to be between low to mod- erate, with greatest corrosion damage encountered in the tendons with the highest moisture content. Rate of corrosion testing via polarization resistance testing has some limitations that are important to keep in mind. The test can only determine the rate of corrosion for strands embedded in grout. Strands that are uncovered or exposed because of a grout void with obvious section loss (such as those encountered in the Varina-Enon Bridge), cannot be evaluated with this method. Also, corrosion potentials were measured, but these also provided no corroborat- ing evidence to suggest active corrosion. The tendons have not been replaced. Photo credit: Parsons Brinckerhoff. Figure 36. One broken wire and heavy section loss.

Case Examples 57   Considering the multitude of evaluation methods used to assess the grout, the quality of the grout tested in this inspection was generally good. Though no strands were found to be embedded in such locations, grout near the top of the duct was found to be visually poor (white and chalky) and to exhibit lower than desirable pH. Since 2012 Additional maintenance, repair, and inspection efforts have been performed by the bridge owner since 2012 to address a handful of concerns and to facilitate assessment of the structure’s integrity. Hairline and map cracking has recently been identified in the grout pour backs on the transverse PT in the deck. No information is available regarding the surface preparation or pour-back materials used in these details. During a regular bridge safety inspection, an external transverse PT tendon, part of the delta frame, was found ejected from the structure. Further inspection revealed that this tendon had never been grouted. The tendon was replaced. This finding precipitated an increase in inspection frequency. An additional inspection was undertaken of the transverse and vertical tendons. Access to vertical tendons was not always possible; of those inspected, approximately 80% of the vertical tendons were found to have a void of the top 8 ft of the tendon. A wire break was also found. The field work for this inspection concluded in 2019. The agency is currently developing its response. Considering the structure’s long history of issues, the bridge owner has taken several actions to facilitate monitoring. These actions include increased inspection frequency of greater scrutiny, installation of monitoring instrumentation and supplementary equipment, and the categorization and management of the Varina-Enon as a “special structure.” In 2013, VDOT installed interior lighting in boxes throughout each span. Although this has no direct effect on the PT system, the agency is largely positive on the effort, citing that it has greatly helped with subsequent inspection efforts. In 2014, four of the tendons with monitoring locations were impregnated with a corrosion inhibiting liquid and the monitoring was discontinued. The FHWA magnetic flux monitoring occurred several times after 2014. In 2018, in response to concrete cracking in the pylons, additional PT was installed, using flexible filler materials. VDOT has implemented a long-term monitoring plan with researchers at Virginia Tech, with instrumentation to gather information on thermal effects and effective prestressing. Photo credit: Parsons Brinckerhoff. Figure 37. Four broken wires and heavy section loss.

58 Repair and Maintenance of Post-Tensioned Concrete Bridges The effort is expected to wrap up in 2021 and will assist in both improving the accuracy of load-rating and assessing the structures’ durability. Additional efforts, including investment in an acoustic emissions system, and further measurements using magnetic flux leakage are anticipated to be used in conjunction with these efforts to assist the agency’s future decision making. Inspection frequency has been increased. Regular inspections occur every 2 years, with hands-on, in-depth inspections occurring every 4 years. Baseline data on the structure’s vibra- tion characteristics and geometry, including pylon elevations, are currently being gathered to facilitate later assessment. Summary The Varina-Enon Bridge near Richmond, Virginia, has undergone two external tendon replacements in the past 20 years. Attempts to use nondestructive methods (corrosion poten- tial, rate of corrosion) of tendon assessment were not successful in confirming known tendon corrosion issues. Testing grout quality at isolated test locations found the grout quality to be good, and the cause of the tendon corrosion could not be attributed to the grout itself. Instead, locations of identified strand corrosion and wire breaks were all in areas of voided grout. VDOT completed the following efforts to preserve the structure: • Vacuumed grouted grout voids, but noted grout discontinuities could contribute to the formation of macrocell corrosion. • De-tensioned and replaced multiple broken tendons. • Successfully used the magnetic flux leakage method of NDE. • Increased visual and in-depth bridge inspection frequency. • Installed instrumentation to monitor the structure’s condition. • Designated the structure as a Special Structure to ensure long-term surveillance. Case 5: Plymouth Avenue Bridge Owned and maintained by the City of Minneapolis Public Works Department, the Plymouth Avenue Bridge crosses the Mississippi River in downtown Minneapolis, Minnesota. Opened in 1983, the bridge is a 934-ft-long pair of post-tensioned segmental structures. Four lanes of traffic are carried by parallel, single-cell, concrete box girders in two directions, westbound and eastbound. Cast-in-place, the main span was constructed with form travelers and cantilever construction; the two approach spans were cast-in-place on falsework. In-depth Inspection (2010) Biennial inspections were performed regularly on the Plymouth Avenue Bridge from its opening in 1983 to 2008. During this period, inspectors did not note anything to trigger an in-depth inspection. In 2010, an inspector walking through the interior of the segmental box noticed light penetrating the interior of the box, emanating from the bottom slab of the box. With manual effort and a heavy rod, the inspector was able to remove large chunks of concrete, revealing heavily corroded tendons (Figure 38). Subsequent inspections conducted in 2010 revealed five heavily corroded tendons in one span of the eastbound structure. The identified tendons were located in the bottom slab con- crete of the box adjacent to drains (Figure 39). At least two of the identified tendons were found to have lost all prestressing force. Further examination revealed that water had been penetrating the interior of the box for many years.

Case Examples 59   Photo credit: Corven Engineering. Figure 39. Evident water intrusion around drainage system. Photo credit: Corven Engineering. Figure 38. Corroded bottom slab tendons identified in initial inspection.

60 Repair and Maintenance of Post-Tensioned Concrete Bridges Subsequent in-depth inspections were conducted by a consultant; the inspection aimed to further visually inspect, with invasive techniques. The following were targeted for this in-depth inspection: • Previously identified, corroded tendons, • Continuity tendons in the bottom slab haunch in the same general vicinity, • Continuity and draped tendons in the same general vicinity of the bridge, • Selected cantilever tendons, and • Selected anchorages of the bottom slab continuity tendons. In-depth invasive inspections involved careful drilling through the concrete superstructure at assumed tendon locations for visual inspection. Observation was made of the galvanized duct condition, relative grout fill in the tendon, grout condition, and the appearance of the pre- stressing strand. Damage in localized regions was severe (complete PT loss in Span 3) to mod- erate (wire breaks in other locations). In areas adjacent to regions of damage, grout appeared to fill the ducts and be of good quality, and galvanized ducts and prestressing strand did not exhibit corrosion. Invasive inspection of selected anchorages was performed on bottom slab continuity tendons; the two pour backs selected for this investigation were chosen on the basis of observed signs of potential distress (see Figure 40). One pour back had exposed rusting reinforcement (Span 3 westbound), one pour back had observable water staining (Span 3 eastbound). Pour backs were removed with pneumatic air hammers. Grout vents in the anchorages were drilled as far as geo- metrically possible; grout appeared to fill the anchor. Despite the presence of water, the Span 3 eastbound anchorage was in good condition, with what was assumed to be a thin epoxy coating more than 85% of the anchorage; a light surface rust was observed on the remainder of the anchorage. The Span 3 westbound anchorage was covered in a heavier rust on the wedge and anchorage plate, but no deep pitting or section loss was noted (Corven Engineering 2010). Damage was attributed to a long period of wet/dry cycling, during which run-off and wind- blown water from the drainage system saturated the lower slab of the box girders. The presence of moisture, combined with seasonal freeze/thaw, caused local cracking and spalling of the concrete, eventually allowing exposure of the post-tensioning tendons in the box girders’ bottom slab. Significant damage was identified by the inspections, mainly in Span 3; tendons in this span were found to have lost all post-tensioning force. Photo credit: Corven Engineering. Figure 40. Exposed bottom slab continuity tendon anchorages.

Case Examples 61   Less severe damage was identified in adjacent spans; wire breaks were identified, but force was found to have been recovered at some distance from the breakage through friction and grout interlock. Identified damage was severe enough to require bridge closure until the con- clusion of the repairs (and the satisfaction of the engineer). Repair recommendations were formulated with multiple goals: to restore prestress force lost by damaged tendons, to improve the structure’s water-tightness, and to provide strength redundancy to compensate for potential future issues. Response Actions taken to address identified concerns were aimed at restoring original design capacity with replacement tendons, providing additional corrosion protection, and removing the source of moisture intrusion. Following the investigations, specific repair actions were taken in multiple spans, including the following: • For Span 3, tendons that had lost post-tensioning force were replaced. • Tendons in Spans 1, 2, 4, and 5 with evident wire breaks were exposed, sealed, and entombed in a corrosion protection material. All tendons in these spans were assumed to retain effective post-tensioning force. • Post-tensioning pour backs were repaired with an epoxy grout. • A surface seal coat was applied to the interior of the box girder as an additional measure of protection. • Repaired identified concrete cracking with injections. • An existing drainage system was permanently sealed at road-level, and drainage pipes inside the structure were removed. Summary The Plymouth Avenue Bridge in Minneapolis, Minnesota, underwent routine inspections until 2010, when an inspector noted light penetrating the interior of the box for the bridge’s underside (bottom slab of box). Despite reporting routine findings in past inspections, an in-depth inspection revealed that the bridge’s drainage system—piping water from the top of the deck and passing through the box—had been the source of moisture intrusion for many years. Remediation actions include tendon replacement to re-achieve original design capacity, re-entombment of exposed tendons, repair of pour backs, crack injection, and removal of the drainage system.