Machine Learning-based Tool Wear Prediction Using Bolt Type Piezo Sensors
Summary
No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.
Summarise with VIRAT Free to try with your Science Catalogs account.Abstract
This study proposes a method for predicting cutting tool flank wear using bolt-type piezo sensors (Piezobolts) and validates its performance through comparison with a three-axis dynamometer, the laboratory standard for cutting-force measurement. During machining, reaction forces transmitted through the workpiece clamping interface are measured by Piezobolts, converted into approximate three-axis cutting forces, and used to estimate tool wear. Results show that wear progression trends derived from Piezobolts closely match those obtained from the dynamometer throughout the full wear process. A support vector regression (SVR) model preserving temporal data order successfully predicts later-stage tool wear using only early-stage training data. Using features extracted solely from Piezobolt signals, the model captures both gradual wear progression and accelerated wear near the end of tool life with high consistency. These findings demonstrate the feasibility of Piezobolts as surrogate sensors for tool-wear monitoring without direct cutting-force measurement or machine-tool modification. Although validation was limited to a single tool, workpiece material, machining operation, and cutting condition, this study confirms experimental feasibility and provides a practical foundation for extending Piezobolt-based tool condition monitoring to broader machining environments.
Fields of Study & Subjects
Advanced machining processes and optimization
Engineering Technology and Methodologies
Machine Fault Diagnosis Techniques
Publication details
- Journal / source
- International Journal of Precision Engineering and Manufacturing-Smart Technology
- Volume
- 4
- Issue
- 2
- Pages
- 85–94
- Publication date
- 31 March 2026
- Language
- EN