[1C3] Application of ultrasonic imaging and pattern recognition for classification of corrosion severity in concrete structures
M Kumar1 and A Ganguli2
1Indian Institute of Technology, India
2Liverpool John Moores University, UK
The degradation and declining safety of reinforced concrete structures owing to rebar corrosion is a matter of global concern. Corrosion causes rebar diameter reduction, interfacial bond loss, cracking and spalling of the concrete cover. From the standpoint of safety and durability, systematic and periodic assessment of concrete infrastructure through advanced non-destructive testing (NDT) is a critical necessity for sustainable concrete infrastructure. In this context, the potential of statistical learning algorithms and ultrasonic imaging-based NDT for classification of the cracking risk of concrete is presented in this study. A laboratory set-up consisting of an ultrasonic transducer pair for scanning the condition of rebars embedded in a slab subjected to various levels of accelerated corrosion is developed. The synthetic aperture focusing technique (SAFT) is used to create images of the rebar. Features are extracted from the images and are fed into a Gaussian mixture model (GMM)-based pattern recognition algorithm to identify clusters that would categorise the severity of corrosion-induced cracking risk. The suggested technique is highly promising for providing useful inputs for pre-emptive maintenance and efficient management of concrete structural assets.
Keywords: SAFT, Gaussian mixture model, ultrasonic imaging, image-based pattern recognition.
Keywords: SAFT, Gaussian mixture model, ultrasonic imaging, image-based pattern recognition.
