Bond Characteristics and Experimental Behavior of Textured Epoxy-coated Rebars Used in Concrete Bridge Decks

The deterioration of bridge decks is a problem typically associated with the corrosion of the reinforcing steel. This issue was partially controlled during the 1970s with the incorporation of the epoxy-coating protection system. However, research later demonstrated that the smooth surface resulting from the epoxy-coating application reduces most of the friction between the rebar and the surrounding concrete. Consequently, forces acting on the rib faces are reconfigured in such a way that the radial components increase, triggering the early development of cracks. To mitigate both the reduction of bonding and the formation of cracks, the Illinois Department of Transportation proposed a new type of coated bars: textured epoxy-coated (TEC) bars. Over the last few years, different projects have been executed to understand and improve the characteristics of TEC rebars. This report is a continuation of research performed at the University of Illinois Urbana-Champaign to evaluate the bond behavior of TEC bars. The experimental program starts by characterizing, qualitatively and quantitatively, the roughness of the TEC rebars. Next, their bond-slip interaction embedded in concrete is evaluated through pull-out tests. Finite element models of these tests are developed to validate the behavior observed as the textured reinforcement loses anchorage with concrete. Based on these results, the experimental program then aims to study the impact of the drying shrinkage, temperature change, and flexural demands on two large-scale bridge deck specimens reinforced, individually, with TEC and standard epoxy-coated bars. The results collected from both specimens using digital image correlation and strain gauges are compared to explore the differences exhibited by the traditional and the new type of reinforcement coatings in terms of stress distribution in bridge decks. Finally, given the specialized equipment and time-consuming procedure needed to calculate the roughness parameters of TEC bars, an empirical, weight-based approach is developed as a rapid method for assessing the rebars’ roughness on-site.

Mapping Ground Penetrating Radar Amplitudes Using Artificial Neural Network and Multiple Regression Analysis Methods

Bridges are aging and deteriorating. Thus, the development of Bridge Management Systems (BMSs) became imperative nowadays. Condition assessment is one of the most critical and vital components of BMSs. Ground Penetrating Radar (GPR) is one of the non-destructive techniques (NDTs) that are used to evaluate the condition of bridge decks which are subjected to the rebar corrosion. The objective of the proposed method is to develop standardized amplitude scale for bridge decks based on a hybrid optimization-decision making model. Shuffled frog leaping algorithm is employed to compute the optimum thresholds. Then, polynomial regression and artificial neural network models are designed to predict the prioritizing index based on a set of multi-criteria decision-making methods. The weibull distribution is utilized to capture the stochastic nature of deterioration of concrete bridge decks. Lastly, a case study is presented to demonstrate the capabilities of the proposed method.

Flutter Derivatives Identification and Uncertainty Quantification for Bridge Decks Based on the Artificial Bee Colony Algorithm and Bootstrap Technique

This paper presents a novel parameter identification and uncertainty quantification method for flutter derivatives estimation of bridge decks. The proposed approach is based on free-decay vibration records of a sectional model in wind tunnel tests, which consists of parameter identification by a heuristic optimization algorithm in the sense of weighted least squares and uncertainty quantification by a bootstrap technique. The novel contributions of the method are on three fronts. Firstly, weighting factors associated with vertical and torsional motion in the objective function are determined more reasonably using an iterative procedure rather than preassigned. Secondly, flutter derivatives are identified using a hybrid heuristic and classical optimization method, which integrates a modified artificial bee colony algorithm with the Powell’s algorithm. Thirdly, a statistical bootstrap technique is used to quantify the uncertainties of flutter derivatives. The advantages of the proposed method with respect to other methods are faster and more accurate achievement of the global optimum, and refined uncertainty quantification in the identified flutter derivatives. The effectiveness and reliability of the proposed method are validated through noisy data of a numerically simulated thin plate and experimental data of a bridge deck sectional model.