Technical Articles
Adjacent Precast Box-Beam Bridges: Longitudinal Shear-Key Joint Cracking and Leakage
Understand why longitudinal shear-key joints between adjacent precast box beams crack and leak, and how bridge owners organize connection investigations.
Adjacent precast concrete box-beam bridges are common; a 2015 feature in The Construction Specifier reported that two-thirds of state transportation departments used them. Valued for rapid erection and shallow depth, these bridges often develop longitudinal joint cracking between adjacent beams.
When shear keys crack, joint leakage can accelerate reinforcement corrosion and diminish load transfer between beams. FHWA research notes that beams in a failed connection will not deflect equally under live load — one possible consequence of a degraded joint, not an automatic one.
This guide reviews structural, thermal, and erection mechanisms from research and offers an observation register for inspecting affected superstructures. Structural evaluations, load ratings, and repairs remain the responsibility of the designated bridge engineer.
Experimental and field evaluations of thermal gradients, grouts, and bearing seating
Steinberg, Semendary, and Walsh's 2015 feature attributes recurring longitudinal cracks mainly to poor transverse connection of the shear keys. FHWA research (FHWA-HRT-17-094) separately found that in its conventionally grouted test connections, the grout-to-concrete interface was the weak link.
In one full-scale outdoor test from that feature, a mid-depth shear key cracked within a single week. The cause traced to the temperature difference between the top and bottom of the beam, not the surrounding air: daily warming and cooling flexed the beam enough to pry some joints open while squeezing others shut, and because the beams did not bear evenly on the abutments, that flexing concentrated enough strain to crack the key.
Epoxy was also used as a grouting material in that study's shear key, and no cracks appeared from temperature or loading. However, the authors considered epoxy undesirable because of the large difference in thermal expansion between epoxy and concrete, without generalizing to all polymeric grouts.
The same authors describe a separate Michigan investigation of in-service box-beam bridges; the inspected sample developed reflective cracks regardless of age. On one bridge inspected during construction, reflected cracks appeared within three days of grouting, before post-tensioning, and continued afterward; by three weeks, every shear key had cracked, attributed to debonding between the key and the adjacent beams.
Transverse post-tensioning limits and erection sequence conflicts
FHWA (FHWA-HRT-17-094) noted that neither 2002 AASHTO Standard Specifications nor 2010 AASHTO LRFD Construction Specifications give specific guidance for adjacent box-beam connections; the 2013 AASHTO LRFD Bridge Design Specifications suggest only 0.25 ksi minimum transverse prestress. Russell's 2011 PCI Journal paper, reporting DOT survey findings, found no consensus on tie count or post-tensioning magnitude.
FHWA strain data showed post-tensioning force only effectively compressed a small area around tie locations; compression dissipated quickly, leaving locations a few feet away with little confinement.
Erection sequence also affects joint stresses: Russell reported that about half of states grout the keyway before post-tensioning and half after. Tensioning first concentrates contact stress because bearing area is smaller, leaving the grout mainly as a filler carrying compressive stress. Grouting first precompresses the interface, giving a higher moment capacity before the precompression is overcome.
FHWA cautioned that if post-tensioning preceded grouting, loss of that force after casting could develop transverse tensile forces, cracking the connection if the beams had enough relative sweep in their fabricated shape. Russell reported half of survey respondents experienced uneven seating, more often with full-width supports. In FHWA testing of cracked connections, cyclic loading propagated cracks along the seam, with deflection and crack propagation increasing as post-tensioning fell.
Visual inspection symptoms and condition assessment limits
Russell's 2011 PCI Journal paper reported that, per survey responses, visual inspection was the practice used at the time to document beam condition — checking for water dripping from the joints during rain, icicles, or staining from water wicking along the soffit.
Where concrete overlays are used, joint distress can reflect upward; Russell reported that responding states observed reflective cracking in the topping.
Surface observations have limits: soffit staining shows possible water passage, not confirmed leakage, and cannot identify thermal gradients, grout bond failure, or girder sweep as the cause. Corrosion is harder to see — Russell notes only two visible signs during a survey, rust bleeding through or a spall exposing reinforcement, and by the time either shows, the steel has already corroded for some time. Neither sign confirms the strands underneath are sound.
Historical recommendations and experimental connection details
Russell's 2011 PCI Journal paper recommended full-depth shear keys that can be grouted easily, a composite deck with 4,000 psi strength and 5-inch minimum thickness, and seven days of wet curing — historical recommendations, not design figures for a particular bridge. He also noted about half of states sandblasted keyways before erection, recommending precasters sandblast surfaces before shipping and contractors clean them with compressed air or water before erection.
FHWA's November 2017 TechBrief recommended a minimum interface bond strength of 150 psi for grout material, to avoid interface cracking from eccentrically applied external loads in its test setup; this excludes thermal and shrinkage forces and is not a field-acceptance criterion for any bridge. It also reports the 2011 third-edition PCI manual's 0.02-inch differential-deflection limit for spans up to 100 ft — relayed from FHWA's account, not the manual itself, and not a project's acceptance criterion.
The same TechBrief reports related research by De la Varga et al. on UHPC-to-concrete connections with an exposed-aggregate finish, finding failure likely to occur in the concrete rather than the interface — distinct from the Fayette County trial below.
Steinberg, Semendary, and Walsh instrumented a 2014 Fayette County, Ohio bridge using UHPC shear keys with a similar exposed-aggregate surface. No cracks were visible at form removal; the authors later reported that after load testing and months of monitoring, cracks still had not appeared, unlike early cracking seen elsewhere in non-shrink-grout joints. As of that 2015 report, the bridge remained under instrumented monitoring for further evaluation. This single bridge does not establish that UHPC resists debonding elsewhere; matching its geometry and surface preparation is the engineer's judgment.
Superstructure joint observation register
When evaluating a box-beam bridge showing joint distress, inspectors should compile a structured observation register:
| Field observation | Documented condition | Engineering question | Evaluation decision and authority |
|---|---|---|---|
| Water dripping or winter icicles along longitudinal joint seams | Visual leak mapping during or after rain events | Has debonding penetrated full keyway depth or reached beam flanges? | Owner agency or engineer determines whether to schedule close-up access, leak mapping, or drainage evaluation |
| Efflorescence or moisture staining along beam soffits | Photographic log of mineral deposits along bottom flanges and adjacent edges | Does soffit staining indicate active chloride migration from deck deicing? | Engineer evaluates whether conditions warrant sampling, petrographic examination, or chloride analysis |
| Reflective cracking along deck wearing surface or overlay | Crack pattern survey and alignment relative to underlying joints | Are reflective cracks permitting water intrusion through the topping slab? | Agency evaluates whether to monitor crack progression, reseal overlay surfaces, or inspect joint interfaces |
| Rust staining, localized spalling, or exposed reinforcement | Documentation of rust trails, delamination boundaries, and spall depths | Has corrosion reduced internal shear stirrups or prestressing strands? | Designated bridge engineer determines scope of delamination sounding, nondestructive testing, load-rating review, or repair design |
What to assemble for an engineering evaluation
Before commissioning an evaluation or repair design for leaking box-beam joints, bridge owners should compile a documentation package:
- Structural drawings: girder cross sections, keyway geometry, bearing seats, tie locations, and overlay specifications.
- Precaster fabrication records: camber measurements, sweep logs, break records, and keyway surface preparation.
- Erection records: bearing pad placement, shim adjustments, tie tensioning sequence, grout submittals, and curing logs.
- Inspection reports: underside photos, crack progression maps, sounding surveys, and drainage maintenance records.
For assistance organizing project records into a clear technical summary, submit your documentation to the PRECAST.WORLD technical desk. Practitioners can share product details, drawing status, standards, and photos to organize their information into a structured technical question.