The corrosion of steel reinforcement in bridge structures can create uncertainty with structural capacity. With that in mind, a new prediction tool developed by researchers from the Minnesota Department of Transportation (MnDOT) (St. Paul, Minnesota, USA) and its partners is designed to help estimate rebar section loss, which can allow the department to more accurately plan its bridge maintenance repairs.
“We’ve developed and calibrated a set of models with Minnesota-specific conditions for more accurate estimates of reinforcement section loss,” says Behrouz Shafei, associate professor at Iowa State University (Ames, Iowa, USA). “Such estimates are instrumental in ensuring the safety and performance of bridge structures in service.”
While bridges are generally designed for at least a 75-year life span, corrosion damage can manifest after just a few years of exposure, MnDOT explains. Fluctuations in temperature and moisture, as well as exposure to deicing salts, can cause material deterioration such as the corrosion of steel rebars embedded in these concrete bridge structures.
This rebar corrosion creates rust and cross-section loss, described as a weakening of and decrease in steel content. Because rust has a higher volume than the steel rebar, the rusting process also causes expansion, which leads to concrete cracking and potentially spalling. Whether the rebar is still providing structural reinforcement at the design capacity depends on how much of the steel bar’s cross-sectional area remains intact.
Conventionally, bridge inspectors visually assess bridge structures to look for cracks while using hammers to find delaminated concrete and visible rebar. There is, however, no industry standard to measure the section loss of the reinforcing steel rebar, MnDOT explains. While an inspector may use visual judgment or a caliper to measure the remaining diameter if the rebar is exposed, there is little guidance to assess the section loss if steel reinforcement is still within the concrete.
Furthermore, visual judgments and spot measurements can be inaccurate due to the irregular nature of corrosion pitting and general corrosion. Even the highest quality observations may be inadequate, given the difficulty in accurately measuring a corroded steel bar and estimating the lost steel.
Thus, understanding the level of reinforcement deterioration is critical to determining its structural capacity. Without guidance, bridge inspectors may be gathering less accurate data for engineering evaluations.
With that in mind, MnDOT sought guidance on accurately estimating rebar cross-section loss to inform its repair strategies and efficiently plan preventive and corrective bridge maintenance. According to the department, better tailoring these repair strategies can reduce material and labor costs while minimizing bridge closures and traffic disruptions.
As such, the project’s goal was to develop guidance and aids to enhance visual and other nondestructive methods of estimating reinforcement section loss. The guidance and aids were developed specifically for standardized 2-inch (50.8 mm) concrete cover over reinforcement with 4,000 psi (27.6 MPa) strength concrete.
Researchers reviewed available literature on steel reinforcement corrosion and identified mathematical equations to best correlate the progression of reinforcement corrosion to concrete cracking. They examined both destructive and nondestructive methods to evaluate deterioration.
From the literature reviews, they identified two formulations that could best envelope rebar corrosion. In situation 1, reinforcement could not be directly observed, whereas in situation 2, reinforcement was exposed. For situation 1, concrete crack measurement was the means to estimate rebar section loss, because increased rust volume is a causal factor of concrete cracking. In situation 2, where direct visual examination could be possible, the high variability of rebar section loss led the researchers to choose a mathematical formula based on exposure time.
For both situations, field data collection was necessary to correlate with the predictive methods. Many piers of a bridge near Minneapolis were scheduled for concrete repair due to reinforcement corrosion, providing an opportunity to document observed conditions before the repair and discover actual rebar section loss during the repair. Before performing any concrete repairs, researchers conducted extensive photo documentation and mapping of the delaminated concrete, which they say was almost always due to rust formation and the corrosion of internal steel reinforcement.
During the concrete repair project, researchers collected data including photo documentation and visual assessments of rebar section loss. In addition, they extracted select steel reinforcement samples that were taken to the laboratory for more exact measurements and testing. In the laboratory, researchers then removed the rust and measured the remaining steel. They used 3-D (three-dimensional) scanning to plot cross-sectional areas of the bars for calculating section loss, and they also performed mechanical tests to determine the strength and failure points of corroded rebars.
Finally, they recalibrated the predictive models for two types of rebar section loss: concrete that had maintained its overall integrity but with visible cracks in its surface (situation 1), and concrete that had partially or completely delaminated to expose embedded steel rebars (situation 2).
Based on empirical and theoretical research validated by this field data, the researchers say they were able to develop a guidance tool to predict rebar reinforcement section loss, which can help MnDOT better manage its bridge inventory while informing maintenance and replacement decisions.
Field investigations showed that the range of locations, as well as the magnitudes of corrosion-induced concrete damage, were very large once the reinforcement had spalled concrete. Visual assessments were not sufficiently accurate, particularly when the bar was rust-covered or partially embedded.
The 3-D scanning method, however, accurately estimated section loss. The results from 3-D scanning were comparable to visual assessments when the section loss was relatively large (above 75%), but they were not for smaller losses. Mass measurements were consistent with the 3-D scanning.
“The upper bound of section loss estimated by the model for cracked concrete will be helpful to avoid being overly conservative by closing a bridge when we don’t have to,” says Paul Pilarski, bridge construction and scoping engineer for MnDOT’s bridge office. “We will, however, need more bridge rebar samples to refine the model for delaminated concrete.”
By combining these results with tensile tests, researchers concluded that visual-only assessments of section loss can be conservative for very low actual losses in steel reinforcement. However, they can be unconservative for larger section losses.
A conservative estimate would result in underestimating the structural capacity and could cause an engineer to program unnecessary and costly strengthening. On the other hand, an unconservative assessment would result in overpredicting structural capacity and reducing the factor of safety.
Using the field results, researchers developed section loss guidance tables. In the cracked concrete model, section loss—including upper and lower bounds—is based on the crack width from a photograph or field measurement. In the delaminated concrete model (situation 2), where the bridge support suffers from advanced corrosion-induced damage, the section loss is based on the age of the steel reinforcement. The generality of rebar age was due to a lack of correlation with any other factors.
The models and guidance tables developed have large ranges, and going forward, MnDOT believes the accuracy can be improved with data from additional bridges. This would improve the number of samples in the data set while also representing a greater diversity of ages, locations, and exposure conditions.
MnDOT says it plans to collect additional samples during major bridge repairs.
Source: MnDOT, www.mntransportationresearch.org.