Investigation into the Application of Smart Aggregates in Monitoring Composite Precast Concrete Beams Made Continuous

Abstract

Ensuring the long-term reliability of ageing concrete infrastructure has increased the demand for effective Structural Health Monitoring methods capable of detecting internal damage at early stages. One promising technique involves the use of Smart Aggregates (SA) to monitor changes within the structure non-destructively. This study focuses on their application in composite precast concrete beams made continuous, particularly at the concrete-to-concrete interface between the precast inverted T-beam and the cast-in-situ topping layer. This interface is vulnerable to degradation due to shear forces and stress concentrations, which can result in internal defects such as microcracking or interface separation—forms of damage that are important to monitor, yet difficult to assess in practical in-situ conditions.

The aim of this study is to evaluate how SA-based ultrasonic monitoring can be effectively utilised to detect, localise, and characterise interface delamination in composite precast concrete beams made continuous. Focus is placed on assessing the complementary capabilities of Ultrasonic Pulse Velocity (UPV), in terms of spatial localisation, and Coda Wave Interferometry (CWI)-derived indicators, in terms of sensitivity to early-stage damage. Additionally, the study evaluates how ray-path and tomographic visualisation approaches support the interpretation of these ultrasonic techniques.

UPV is applied by analysing changes in wave arrival time to determine variations in propagation velocity, enabling detection of discontinuities along defined transmission paths. In contrast, CWI utilises the sensitivity of microcracking and stress redistribution to detect incremental changes within the material. From these methods, three ultrasonic indicators are derived: relative velocity from UPV, and the correlation coefficient (CC) and relative velocity change (ε) from CWI. These indicators are analysed through ray-path representations, which preserve path-based physical meaning. Additionally, UPV results are reconstructed into a tomographic visualisation to provide a spatial representation of internal structural changes, from which the interface time interference indicator is derived to assess interface delamination.

The experimental program consists of large-scale precast girders made continuous, embedded with SA and subjected to staged loading. The experimental work was conducted as part of a broader research campaign at TU Delft, in which the physical experiments were carried out by M. Ibrahim. Ultrasonic measurements are collected and processed using the described methods, and the resulting indicators are compared with Digital Image Correlation measurements for validation. This approach enables assessment of the capability of ultrasonic indicators to detect, localise, and quantify damage progression at the interface.

The results show that the ultrasonic indicators provide distinct yet complementary insight into structural behaviour. CWI-based indicators demonstrate high sensitivity to early-stage changes: ε responds to early disturbances prior to visible damage, while CC provides a clearer and more consistent indication of crack initiation. In contrast, UPV-derived relative velocity correlates strongly with developed cracking and provides reliable localisation along transmission paths, particularly for sensor pairs oriented to capture flexural and shear cracks. However, sensor pairs crossing the interface show reduced capability in distinguishing specific crack types, indicating limitations in isolating interface delamination independently.

Ray-path results show strong agreement with DIC observations in terms of crack initiation, localisation, enabling direct interpretation along propagation paths. Tomographic results provide a spatial overview of damage distribution and indicate potential for identifying interface disturbances through the dt indicator at early load stages; however, reconstruction limitations and numerical sensitivities reduce reliability in consistently representing damage magnitude and progression.