This study has demonstrated the effectiveness of the Internet of Things (IoT)-based technology using ANNs for localising AE sources in rail sections, providing a promising avenue for future research and practical applications.
The paper’s main focus is to develop an IoT based, energy efficient and smart sensor-based system that can detect AE sources accurately and effectively. In this study, the AE sensor is attached to an Arduino board for wireless data transmission. As the AE simulation process in this paper is an non-destructive testing (NDT) technique, pencil lead break (PLB) is done on the top flange (TF), side of top flange (STF), web and bottom flange (BF) to simulate artificial AE sources in the rail section. The generated AE signal due to PLB is collected by mounting the AE sensor over the rail web portion. It is found from the literature that optimal placement of the AE sensor is at the web portion of the rail section. Therefore, the good-quality signal from every segment (i.e. TF, web and BF) of the rail section can be captured from the web part. After capturing the signals wirelessly, the AE features like amplitude, energy, duration, etc. are used to find the AE source location using an artificial neural network (ANN). The developed ANN model is giving very promising results in terms of localisation of the AE source. The main challenge of this paper is transmitting AE signals wirelessly to DACs. As the ANN model is running perfectly, it can be said the Arduino can transmit the full packet AE signal to DAC through the Internet.
The experiment conducted to localise AE sources in a rail section using the developed ANN model has yielded very promising results. PLB is applied at 10 mm intervals up to 1,200 mm on the top flange, side of the top flange, web and bottom flange of the rail section. The MATLAB NNTOOL was used to develop the ANN model, which accurately detected AE sources in the rail section. The AE sensor mounted at the web provided excellent localisation of AE sources, with PLB at the TF, STF, web and BF detected without error. The percentage of error was found to be less than 1%, which is a highly promising result. Consequently, the developed ANN model has significant potential for use in detecting and localising damage in railway systems, which could ultimately improve the safety and reliability of railway transportation.
This study has shown that the use of ANNs for damage detection and localisation and the use of IoT in railway systems is more practical and can lead to significant improvements in efficiency and safety. With further research and development, the developed ANN model could become an essential tool for maintaining the rail section’s infrastructure, ultimately enhancing the railway system’s overall performance and reliability.
