Automatic Identification of Tool Wear Based on Convolutional Neural Network in Face Milling Process

Area of Science:

• Manufacturing Engineering
• Artificial Intelligence
• Materials Science

Background:

• Traditional tool wear monitoring relies on subjective visual inspection, demanding expertise and time.
• Automated methods are needed to enhance accuracy, efficiency, and cost-effectiveness in machining operations.

Purpose of the Study:

• To develop and validate a deep learning model for automatic identification of tool wear types in high-temperature alloy tools during face milling.
• To improve the accuracy of automatic tool wear value detection by integrating identified wear types.

Main Methods:

• A Convolutional Neural Network (CNN) model, named ToolWearnet, was developed using a custom image dataset.
• The network was pre-trained using a Convolutional Autoencoder (CAE) and fine-tuned with Backpropagation (BP) and Stochastic Gradient Descent (SGD) algorithms.
• An experimental system was built to capture tool wear images during the machining process.

Main Results:

• The ToolWearnet model achieved an average recognition precision rate of 96.20% for identifying tool wear types.
• The integrated automatic detection algorithm for tool wear value demonstrated a mean absolute percentage error of 4.76% compared to manual microscopic analysis.
• The experimental system successfully captured wear image information from all inserts during machining.

Conclusions:

• The developed ToolWearnet model offers an effective and practical solution for automated tool wear monitoring in machining.
• The method significantly improves the accuracy and efficiency of tool wear assessment, leading to reduced downtime and costs.
• The study validates the feasibility of using deep learning for real-time tool wear analysis in industrial applications.