Reducing label dependence in vibration-based drill-bit condition monitoring with masked feature pretraining

Key Points

• Improved condition monitoring accuracy can be achieved by reducing label dependence in drill-bits, enhancing operational efficiency.
• The study demonstrates a significant improvement in monitoring precision with a numeric accuracy increase of 15% using new features.
• This analysis utilized masked feature pretraining to refine machine learning algorithms for better signal interpretation.
• Enhancing condition monitoring may enable more proactive maintenance strategies, minimizing downtime and costs.