Title: Intelligent tool wear prediction in milling of nickel-based superalloy using advanced signal processing and bi-transformer

Authors: Shailendra Chauhan; Rajeev Trehan; Ravi Pratap Singh; Vishal S. Sharma

Addresses: Department of Industrial and Production Engineering, Dr. B.R. Ambedkar National Institute of Technology, Jalandhar, Punjab, 144011, India ' Department of Industrial and Production Engineering, Dr. B.R. Ambedkar National Institute of Technology, Jalandhar, Punjab, 144011, India ' Department of Mechanical Engineering, National Institute of Technology Kurukshetra, Haryana, 136119, India ' Department of Industrial and Production Engineering, Dr. B.R. Ambedkar National Institute of Technology, Jalandhar, Punjab, 144011, India; School of Mechanical, Industrial and Aeronautical Engineering, University of the Witwatersrand, Johannesburg, 2050, South Africa

Abstract: This study presents an intelligent tool wear prediction approach for milling Nimonic 80, a superalloy used in high-temperature applications. The impact of different cutting-edge radii (0.8 mm and 0.4 mm) on tool wear is examined using sensor fusion techniques to integrate cutting force and vibration signals. Advanced signal processing methods like improved complete ensemble empirical mode decomposition with adaptive noise (ICEEMDAN) and the Largest Lyapunov Exponent (LLE) for feature extraction are utilised. The experimental design employs a response surface methodology (RSM) with a central composite rotatable design and analysis of variance (ANOVA) to evaluate factor significance. A bi-Transformer model processes multi-sensor data, improving prediction accuracy. Results show that the proposed methodology significantly enhances tool wear prediction, offering insights into Nimonic 80 wear mechanisms. The bi-Transformer achieved an R² of 97.61%, a root mean square error (RMSE) of 0.015, a mean absolute error (MAE) of 0.012, and ∣𝑅_95%∣ of 1.49%. In real-time applications, the signal acquisition system continuously feeds data to the bi-Transformer model, which processes this data in near-real-time. The model's high accuracy (R² = 97.61%) ensures it can predict wear progression before significant damage occurs.

Keywords: milling; tool wear; cutting forces; vibrations; signal processing; machine learning; bi-transformer.

DOI: 10.1504/IJMMS.2024.143336

International Journal of Mechatronics and Manufacturing Systems, 2024 Vol.17 No.3, pp.295 - 318

Received: 06 Jul 2024
Accepted: 24 Oct 2024
Published online: 13 Dec 2024 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article