Corrosion Quietly Rewrites How Precast Bridge Joints Fail Under Cyclic Loads

Grouted sleeve connections are the invisible workhorses of modern prefabricated bridge construction. These compact steel couplers, filled with high-strength cementitious grout and threaded onto the ends of reinforcing bars, allow entire bridge piers to be cast in a factory, shipped to site, and assembled in days rather than months. Their industrial speed and ecological appeal have made them a cornerstone of prefabricated concrete bridges worldwide. Yet a new experimental study published in Case Studies in Construction Materials reveals a sobering vulnerability: when chloride-induced corrosion eats away at the sleeve wall, the way these joints fail under repeated tensile loading changes fundamentally, and the thresholds at which they stop meeting design codes are now far more precisely defined.

The research team, led by Wenting Yuan and colleagues including Yiwei Zuo, Long Ma, Xiangtong Wu and Anxin Guo, set out to close a critical gap in the engineering literature. Previous studies had characterized how corroded sleeve splices behave under monotonic, steadily increasing tension, but almost nothing was known about their performance under cyclic loading, the kind of repeated stress that earthquakes and traffic impose on real structures. Because the seismic performance of an entire precast pier depends on the cyclic stress-strain relationship of its sleeve connections, this missing constitutive knowledge left engineers without a reliable tool for assessing aging precast bridges in corrosive environments, particularly coastal and marine settings where chloride ions steadily penetrate the concrete cover.

To anchor their experiments in realistic service conditions, the researchers first developed a theoretical framework for predicting how deeply a sleeve corrodes over time. Using Fick’s second law of diffusion, they modeled how long chlorides take to breach the concrete cover and reach a critical concentration at the sleeve surface, then applied Faraday-based relationships to track the time-variant corrosion depth of the steel wall. The calculations produced striking numbers: increasing the concrete cover from 30 to 70 millimeters delays corrosion initiation from just 3.3 years to 47.4 years, while at a 100-year service life the predicted sleeve corrosion depth ranges from 1.8 to 3.4 millimeters. Because the sleeve is a hollow section, such losses can approach full wall penetration, a scenario the experiments were designed to probe.

The experimental program involved twenty-four commercial sleeve specimens, each 310 millimeters long with a 41.4 millimeter outer diameter and a 3.2 millimeter wall, connecting 16 millimeter reinforcing bars with an effective anchorage length of roughly 150 millimeters. The sleeves themselves yielded at 381 megapascals and ultimately carried 669 megapascals, while the grout achieved a compressive strength of 70.8 megapascals. Six target corrosion ratios were chosen, from 0 to 50 percent, with four replicate specimens at each level. To corrode the sleeve without corroding the rebars, the team sealed the exposed bars and immersed the specimens in a 3.5 percent sodium chloride solution, applying an accelerated current density of 800 microamperes per square centimeter and calculating the required exposure time from Faraday’s law of electrolysis.

After corrosion, the specimens were scanned with a high-resolution 3D scanner, and the reconstructed models revealed a crucial and previously underappreciated feature: corrosion was markedly non-uniform, both around the circumference and along the length of each sleeve. The most severe damage concentrated at the sleeve ends, a pattern the authors attribute to residual welding stresses at the ends and to the shorter, more accessible transport pathways for chloride ions and dissolved oxygen near the vent openings. Maximum local corrosion depths reached approximately 4 millimeters, indicating localized wall penetration in the most aggressive cases. The team is careful to note that accelerated corrosion at such high current densities does not perfectly replicate decades of natural exposure, but it remains an effective method for reproducing the non-uniform deterioration that governs mechanical response.

Under cyclic tensile loading, the specimens told a clear story of thresholds. Below roughly 30 percent corrosion, the connections behaved almost identically to intact ones, sustaining peak loads near 120 kilonewtons and ultimate displacements of about 100 millimeters, with failure dominated by rebar fracture, the desirable ductile mode. As corrosion deepened and sleeve walls began to penetrate, the confinement the sleeve provides to the internal grout weakened and the effective anchorage length shortened, shifting failure toward grout crushing, rebar pull-out and grout pull-out. Strength ratios held near 1.4 for corrosion below 30 percent, then collapsed in the heaviest group, where three specimens dropped to values of 1.16, 0.35 and 0.89, all below the 1.25 threshold that design codes require for adequate bearing capacity.

Deformation capacity degraded along a parallel track. Intact specimens achieved an average deformation ratio of 9.4, which barely changed at 30 percent corrosion but fell to 4.6 at 50 percent. Two heavily corroded specimens failed before even yielding, their deformation ratios falling below 1.0. The hysteretic loops painted the same picture from an energy perspective: below 30 percent corrosion the loading-unloading cycles remained stable with little degradation, at around 40 percent the loops grew erratic and post-peak strength loss intensified, and at 50 percent premature pull-out produced narrow, incompletely developed loops signaling a substantial collapse in energy dissipation, precisely the property that protects structures during earthquakes.

From the full dataset, the researchers distilled an equivalent stress-strain constitutive model for corroded sleeve splices, expressing yield stress, peak stress, yield strain, peak strain and ultimate strain as functions of the measured corrosion ratio. The trends are distinctive: yield and peak stresses hold steady until corrosion exceeds roughly 33.6 percent and then decline, peak and ultimate strains rise before falling, and yield strain climbs continuously as interfacial slippage accumulates, collectively eroding the connection’s effective elastic modulus. To prove the model’s usefulness beyond the coupon scale, the team embedded it in a finite element simulation of corroded precast bridge piers previously tested under cyclic loading, using the OpenSees platform with fiber-based cross-sections and bond-slip elements. The predicted hysteretic curves matched the experiments well, with peak force errors of 10 to 11 percent for one pier and an average of 5 percent for the more heavily corroded one.

The implications reach well beyond the laboratory. Prefabricated bridges in coastal and deicing-salt environments are aging into precisely the corrosion regimes this study maps, and seismic assessments of such structures have until now relied on constitutive relationships that ignore sleeve deterioration. By quantifying where the safe plateau ends, below 30 percent corrosion for deformation capacity and below 50 percent for strength, and by supplying a validated stress-strain model that captures the transition, the researchers have given engineers a quantitative bridge between field corrosion surveys and structural-level seismic evaluation. The authors caution that their experiments corroded the sleeve in isolation, whereas real sleeves and rebars may deteriorate concurrently, and that validation so far rests on two pier specimens. Future work will address combined corrosion, natural exposure, additional sleeve geometries and variable-amplitude loading, but the present findings already mark a decisive step toward keeping the joints that hold prefabricated cities together honestly accounted for in the face of rust.

Subject of Research: Effect of chloride-induced corrosion damage on the cyclic tensile mechanical behavior and constitutive modeling of grouted sleeve connections in prefabricated concrete bridges

Article Title: Effect of corrosion damage on cyclic tensile mechanical behavior of grouted sleeve connections

Article References: Yuan, W., Zuo, Y., Ma, L., Wu, X., & Guo, A. (2026). Effect of corrosion damage on cyclic tensile mechanical behavior of grouted sleeve connections. Case Studies in Construction Materials, 25, Article e06620. https://doi.org/10.1016/j.cscm.2026.e06620

Image Credits: AI Generated

DOI: Not provided

Keywords: grouted sleeve connections, corrosion, prefabricated bridges, cyclic tensile loading, chloride-induced corrosion, stress-strain model, seismic performance, precast concrete, bond degradation, accelerated corrosion testing, finite element validation, structural durability