A condition monitoring system for rock drill bits
Date of Award
Master of Engineering Science
School of Engineering
The use of acoustic emission analysis to detect problems in mechanical parts has gained increasing attention in recent years. In large industrial plants, real-time condition monitoring of mechanical parts can help to avoid sudden shutdowns and failures of the machine. As a result, a condition monitoring system can benefit the industrial sector in terms of profit and time when machines fail. In this research, the main goal is to develop a condition monitoring system that can analyse and detect the signatures of acoustic signals emitting from mining rock drill bits, and use this information to predict the operating/health condition of the drill bits. This is a prerequisite for the design and implementation of intelligent unmanned drilling systems.
In this work, data is collected from a custom-built test rig that mimics the process of drilling in mine sites. The data is then pre-processed and stored into a dataset. Finally, the dataset is tested using three types of deep learning models: multilayer perceptron neural network (MLP), convolutional neural network (CNN), and, recurrent neural network (RNN)
The proposed models demonstrate the ability to distinguish between drill bit failures based on spectrogram features derived from audio samples captured from drilling operation. These models also show an improved ability to detect faults, by almost 10%, compared to traditional methods such as k-nearest neighbours and support vector machines.
Access to this thesis is embargoed until 8th December 2023.
Qutaish, M. Q. (2022). A condition monitoring system for rock drill bits. https://ro.ecu.edu.au/theses/2607