Structural health monitoring of concrete bridges using infrared thermography

Abstract

The research in this thesis addresses some of the basic challenges facing the practical implementation of Infrared thermography (IRT) as a non-destructive technique for the detection, localisation, and quantification of hidden subsurface defects in concrete components of bridges. It thereby contributes to the prospects of structural health monitoring (SHM). Issues such as the possibility of early-stage damage detection due to corrosion initiation, the earliest possible stage for detection of subsurface defects by IRT, and the differentiation of true concrete defects and other environmental and surface effects, as well as the effect of thermal excitation mechanism, and estimation of heat input or output for minimum detectable thermal contrast, are among the challenges investigated in this thesis. The main aspects of this thesis are the literature review and experimental and numerical programme. The review of literature presents the scopes and merits of SHM and the channels through which IRT contributes to it. After the description of the main elements of an IRT scenario and the research conducted for subsurface defect detection in concrete, the research gaps and the difficulties facing the practical implementation of IRT are explained. Therefore, an experimental programme of several phases was planned and conducted to cover the gaps and address the challenges. IRT monitoring of reinforced concrete (RC) blocks with corrosion-induced defects, and slabs with simulated subsurface voids at 5-25mm depths at 5mm increments was conducted to fill the gap of lack of research with convection heat exchange as thermal excitation. Moreover, quantitative, and objective analysis and interpretation of the thermograms deliver a basic knowledge of the types of corrosion-induced defects detectable by IRT, and the earliest possible stage for their detection. This finding as result of IRT monitoring of the test blocks with progressive corrosion-induced defects contributes to the differentiation of the true bridge defects and the other surface or environmental effects. Quantification of subsurface defects of random shape, objective thresholds as a measure for degradation of concrete affected by corrosion, variation of thermal contrast and surface temperature distribution with depth of defect, as well as the performance of IRT in cool down and heat up transient processes were among the other achievements of tests with convection heat exchange as thermal excitation. The comparative tests using three thermal excitation mechanisms (convection, step-heating using an IR heater and solar irradiance) reveal the significance of the excitation mechanism, on surface thermal contrast. In addition, as IR images were collected in regular intervals, it was possible to develop time-dependent and non-time-dependent metrics based on the variation of thermal contrast and its derivatives. The sign analysis of thermal contrast and its derivatives led to the characterisation of stages of heat transition mechanism in defective concrete. The formation of lateral heat flow around the defective concrete was realised as a requirement for the detection of thermal contrast from subsurface defects. The outputs of comparative tests show that the thermal excitation mechanism that directly correlates with the heat input or output significantly affects the development of thermal contrast of detectable and distinguishable magnitude on the surface. As such, the numerical programme complemented the conclusions of the comparative experiments, the need for quantitative measure for the total heat input and output for detection of defects, by finite element analysis (FEA) of validated finite element models. The results of FEA delineate the reasons for hardships in the detection of deep defects by quantitative measures. In this thesis, the total required heat input or output is estimated by FEA according to several criteria such as formation of lateral heat flow or minimum detectable thermal contrast of 10 to 20 times noise equivalent temperature difference (NETD) of the camera, 0.5°C based on ASTM D4788-03 standard, or 1°C. As corrosion-induced defects often initiate at the rebar depth, the estimation of total heat required for the minimum safe detectable thermal contrast corroborates decision-making about the viability of early-stage damage detection by IRT, selection of thermal camera, data collection time and condition, and thermal excitation mechanism. The conclusions of this thesis lead to new directions for further future research to improve IRT for SHM of concrete bridges. Based on the estimation of the total heat input or output required for safe detectable thermal contrast, future research could focus on the selection of the appropriate thermal excitation mechanism for the detection of defects at rebar depth based on FEA estimations. In addition, in the case of using ambient weather exciters, FEA estimations of required energy could lead to a basis for the selection of the appropriate timeframe for data collection by analysis of weather forecast parameters. Moreover, the current thesis contributes to the differentiation of true subsurface defects and other environmental, operational, and surface effects by unsupervised clustering analysis and thresholding of the surface thermograms. Replicating the procedures of current studies in future research, with necessary modifications, could facilitate the examination of how humidity level, carbonation, chloride content, and other operational and environmental factors impact the evolution of detectable thermal contrast. The observations of such studies, in conjunction with the findings of the current study, can then be employed to train predictive models aimed at distinguishing thermal contrast resulting from corrosion defects and other factors. Such research efforts will lay a foundation for benchmarking various machine-learning techniques, ultimately leading to the development of robust solutions for intelligent monitoring of concrete bridges.

Date of Award	Jul 2024
Original language	English
Awarding Institution	Queen's University Belfast
Sponsors	US-Ireland R&D Partnership Programme
Supervisor	Su Taylor (Supervisor) & Gerard Hamill (Supervisor)