Estimating crack size and location in a steel plate using ultrasonic signals and CFBP Neural Networks

Abstract

This paper presents a novel approach to estimate crack size and crack location in a steel plate using ultrasonic signals and Artificial Neural Networks (ANN). The feature indicators extracted from collected ultrasonic signals include the peak amplitude, energy of the signal, and the time of flight of ultrasonic echo signals. We develop a cascade feed forward back propagation (CFBP) neural network model to estimate both crack size and crack location simultaneously. The obtained results are compared with the conventional feed forward back propagation (FFBP) neural network. Our data indicate that the CFBP model performs better than the FFBP model.