Random forests based classification of tool wear using vibration signals and wear area estimation from tool image data

In precision manufacturing, tool condition monitoring is critical for improving surface finish, increasing efficiency, and lowering manufacturing costs. The present work discusses a complete workflow to accurately predict the tool condition based on vibration data obtained during the turning operati...

Full description

Saved in:

Bibliographic Details
Published in:	International journal of advanced manufacturing technology Vol. 126; no. 7-8; pp. 3069 - 3081
Main Authors:	Cardoz, Basil, Shaikh, Haris Naiyer E Azam, Mulani, Shoaib Munir, Kumar, Ashwani, Rajasekharan, Sabareesh Geetha
Format:	Journal Article
Language:	English
Published:	London Springer London 01-06-2023 Springer Nature B.V
Subjects:	Acoustics Advanced manufacturing technologies Algorithms Artificial intelligence CAE) and Design Classification Computer-Aided Engineering (CAD Condition monitoring Costs Cutting tools Engineering Image processing Industrial and Production Engineering Internet of Things Machine learning Manufacturing Measurement techniques Mechanical Engineering Media Management Neural networks Original Article Production costs Productivity Sensors Surface finish Tool wear Turning (machining) Vibration Workflow Condition monitoring Image processing Random forests Tool wear Machine learning
Online Access:	Get full text
Tags:	Add Tag No Tags, Be the first to tag this record!

Description
Summary:	In precision manufacturing, tool condition monitoring is critical for improving surface finish, increasing efficiency, and lowering manufacturing costs. The present work discusses a complete workflow to accurately predict the tool condition based on vibration data obtained during the turning operation performed on a lathe. An image processing methodology is applied to compute the tool wear area. A specialized misclassification cost matrix is used to train the random forests algorithm to improve the classification of tool conditions. This model can correctly classify tool condition from vibration signals of 0.5 s with an accuracy of 97%. Furthermore, this investigation can be modified to suit the real-world classification of the tool condition based on tool wear requirements.
ISSN:	0268-3768 1433-3015
DOI:	10.1007/s00170-023-11173-y