Presented By: Zhao Wang, Marquette University
The use of fiber-reinforced polymer (FRP) composites for external bonding has become a popular and widely accepted technique for enhancing the strength of concrete structures due to its excellent mechanical performance, corrosion resistance, and ease of construction. However, premature debonding is a major challenge as it prevents the full capacity of FRP composites from being achieved, resulting in material waste. Recently, grooving the surface of concrete before bonding FRP has emerged as a potential solution to this problem. Several experimental studies have evaluated the bond strength of FRP-to-concrete joints with grooves. To facilitate the practical application of this technique, it is necessary to develop comprehensive reliability-based design guidelines that account for the uncertainty arising from various aspects such as materials, model errors, and loading. A critical factor of such analysis is the calibration of model uncertainty which significantly affects the accuracy of reliability-based design and analysis. The objective of this study was to measure the model uncertainty of the existing prediction model for FRP-to-concrete joint with a longitudinal groove by involving the model factor which is defined as the ratio of observed values from experimental test to calculated values from prediction models. To eliminate the potential correlation from critical parameters, the residual model factor was isolated from model factor by separating the systematic part. The lognormal distribution was found to be the most suitable distribution function to describe the residual model factor, and the mean and variance were determined. With this newfound knowledge, we are better equipped to account for uncertainties in the design and construction of FRP-to-concrete connections with grooves, which will ultimately result in more durable and reliable structural improvements.
5 сен 2024
