Vibration-Based Cutting Tool Condition Monitoring in Face Milling Process Using Wavelet Features and Machine Learning Algorithms A Comparative Study B S A Crescent Institute of Science and Technology

Format:

Book

Conference/Event

Author/Creator:

D, Pradeep Kumar, author.

Contributor:

S, Ravikumar

Syed, Shaul

V, Muralidharan

Conference Name:

Advances in Design, Materials, Manufacturing and Surface Engineering for Mobility (ADMMS'25) (2025-02-07 : Chennai, India)

Language:

English

Physical Description:

1 online resource cm

Place of Publication:

Warrendale, PA SAE International 2025

Summary:

In automotive applications, most of the engineering components come across the material removal process in manufacturing. Face milling is one of the prominent material removal processes wherein a multi-point cutter is used to machine the flat workpiece to bring it to its required dimension. In the material removal process, the cost of the cutting tool occupies the major part of the total manufacturing cost of a product. Also, the continuous usage of the cutting tool results in tool wear. The usage of the cutting tool after the threshold value of the tool wear deteriorates the surface finish of the workpiece which leads to product rejection. Hence, optimal tool usage is inevitable. The continuous monitoring of the cutting tool condition will ensure optimal tool usage. In the present work, four real-time tool conditions are considered, namely, fresh tool (G), tool flank wear (FW), tool flaking on rake surface (FL) and tool with broken tip (B). Vibration signals are acquired while milling mild steel workpiece with cutting tools of considered conditions. From the vibration signal, Discrete Wavelet Transform (DWT) features are extracted and the top-ranked wavelet member in terms of classification accuracy of the tool condition is selected using the Decision Tree (DT) algorithm. The Mean Squared Energy (MSE) of the detailed coefficients of the selected wavelet member is computed and that forms the features set. Then the classifying ability of Machine Learning (ML) algorithms such as Support Vector Machine (SVM), and Naïve Bayes (NB) are analysed using the feature set. The results show that the NB outperformed the SVM with the MSE of selected wavelet members derived from DWT

Notes:

Vendor supplied data

Publisher Number:

2025-28-0151

Access Restriction:

Restricted for use by site license