Subsurface defect detection in thin steel plates using thermoelastic stress analysis

Abstract

Thermoelastic Stress Analysis (TSA) is a non-contact full-field measurement technique that relates cyclic elastic stress to small temperature changes on the surface of a material. While TSA is well established for surface stress measurement and crack detection, its application for detecting subsurface defects remains limited, particularly in thin metallic components. Conventional TSA is typically performed under near-adiabatic conditions to relate surface temperature changes directly to cyclic stress; in this work, loading frequency is deliberately varied to relax these conditions and assess whether thermal diffusion can enhance subsurface defect visibility at the observed surface. This study investigates the capability of TSA to detect and characterise subsurface defects in 3mm thick GR250 steel plates under cyclic tensile loading. A controlled experimental programme was carried out using plain specimens, specimens containing flat-bottom subsurface defects of different depths, and notched specimens for validation. The study focused on 3 mm thick GR250 steel plates containing circular flat-bottom subsurface defects of 5 mm diameter with depths of 1 mm and 2 mm, tested under cyclic loading frequencies between 4–20 Hz. TSA measurements were performed at multiple loading frequencies using an infrared camera system, and thermoelastic stress maps and extracted stress profiles were obtained through lock-in processing and signal stacking. Backside imaging was employed to assess the influence of defect depth and remaining ligament thickness on the measured surface response. The results demonstrate that TSA can detect subsurface defects through their influence on surface stress redistribution, provided that the defects are sufficiently close to the observed surface. Defects associated with smaller remaining ligament thickness produce clearer and more localised thermoelastic signatures, while defects separated from the observed surface by a thicker ligament generate weaker and more diffuse responses. Stress profile analysis shows that defect depth influences not only the magnitude of the response but also the nature of the surface stress redistribution. The frequency investigation shows that loading frequency primarily affects signal stability and repeatability, while also confirming that defect-related surface signatures persist across the tested range. A supporting finite element modelling study was conducted to provide mechanical interpretation of the experimentally observed stress redistribution patterns. The numerical results show consistent depth-dependent stress transfer behaviour and support the experimental interpretation of remaining ligament thickness as a key controlling factor in subsurface defect detectability. Overall, this study demonstrates the potential and limitations of TSA for subsurface defect detection in thin steel plates and provides experimental and numerical insight into the stress redistribution mechanisms governing defect visibility. The findings contribute to the understanding of TSA sensitivity to subsurface defects and provide a basis for further development of TSA-based subsurface inspection methodologies.