Tricone bit health monitoring using wavelet packet decomposed vibration signal

Abstract

This paper presents part of a research work developing an automated bit failure prediction system for drilling applications. The approach relies on analysis of vibration signals generated as a result of bit-ground interactions. Extensive full-scale in-situ tests have been accomplished in participating Canadian mine sites. A high-frequency data acquisition unit was installed in the control cabin of a blasthole drill unit to collect the vibration signal from several accelerometers placed on different spots of the rig. The vibration signals collected from the drill mast were analyzed in time and frequency domains and the frequency pattern produced by worn bit is achieved. The frequency ranges of vibration signal those are affected by bit wear are investigated. An Artificial Intelligence (AI) classifier is being designed to perform the automated bit condition classification based on selected signal features. For this purpose, a time-frequency representation of the vibration signal is achieved by application of Wavelet Packet Decomposition (WPD). This research final outcomes will enable mining operations to improve drilling performance by detection of bit wear status and prediction of the bit catastrophic failure to avoid additional costs and delays for the production.