Paper 1 · Second author
Machine Learning-Based Predictive Maintenance for Mechanical Systems: Tool Wear and Bearing Fault Detection Using Vibration and Acoustic Emission Signals
Md Zakir Hossen (first author), M Abidur Rahman Sakib (second author), Md Ikbal Hasan.
EJASET, volume 4, issue 5. Published 16 October 2026.
We built a deep learning framework that fuses vibration and acoustic emission signals. One network finds bearing faults and predicts tool wear. It reaches 99.28% fault accuracy and a flank wear error of 5.14 µm, with an R squared of 0.982. We tested it on the PHM 2010 milling data and the Paderborn bearing data.
+
99.28%
bearing fault accuracy
+
5.14 µm
flank wear error (RMSE)
+
R² 0.982
for tool wear prediction
Read the paper
DOI 10.59324/ejaset.2026.4(5).17
