SciEPublish

Geometric Characterization and Prediction of Tool Wear in Ultrasonic Machining Using Reverse Engineering and Machine Learning

Issue 2, Volume 3
Part of Special Issue Selected Papers from 10th International and 31st All India Manufacturing Technology, Design & Research (AIMTDR 2025)
Download PDF
Cite This Article
Contents

Article Open Access

Geometric Characterization and Prediction of Tool Wear in Ultrasonic Machining Using Reverse Engineering and Machine Learning

Arindam Chakraborty 1 Kailash Roy 2 Sudip Santra 1,* Biswanath Doloi 1 Somnath Das 3 Jitendra Sarkar 4

Author Information
Other Information
1
Production Engineering Department, Jadavpur University, Kolkata 700032, India
2
Mechanical Engineering Department, IIT Kanpur, Kanpur 208016, India
3
Department of Mechanical Engineering, Swami Vivekananda Institute of Science & Technology, Kolkata 700145, India
4
Mechanical Engineering Department, Brainware University, Kolkata 700125, India
*
Authors to whom correspondence should be addressed.

Received: 01 October 2026 Revised: 12 October 2026 Accepted: 05 November 2026 Published: 09 November 2026

Creative Commons

© 2026 The authors. This is an open access article under the Creative Commons Attribution 4.0 International License (https://creativecommons.org/licenses/by/4.0/).

Views:137
Downloads:46
Intell. Sustain. Manuf. 2026, 3(2), 10029; DOI: 10.70322/ism.2026.10029
ABSTRACT: The ultrasonic machining product depends heavily on tool shape, as the tool replicates the workpiece surface. Because of this, tool wear during ultrasonic machining (USM) needs in-depth analysis, as it changes tool geometry and can affect the dimensions of the machined cavity. This study examines an image-based reverse-engineering workflow for characterizing changes in USM tool geometry over cumulative machining time. Images captured using a digital microscope were processed to obtain three-dimensional point-cloud data, from which tool height, width, and volume were evaluated. Linear Regression (LR), Random Forest Regression (RFR), and Extreme Gradient Boosting (XGBoost) were used to estimate the measured geometric responses. The recorded measurements showed an overall reduction in tool dimensions and volume during the experimental sequence. The model estimates followed the broad trends in the data, although their agreement varied by response and machining stage. The work demonstrates the use of microscope-based image capture and point-cloud processing for tracking tool geometry and provides an initial comparison of regression approaches for the recorded sequence.
Keywords: Tool wear; Ultrasonic machining; Reverse engineering; CAD model; Machine learning (ML)
TOP