Research on the Timing of Replacing Worn Milling Cutters by Using Wear Transition Percentage Constructed Based on Spindle Current Clutter Signals

Area of Science:

• Manufacturing Engineering
• Mechanical Engineering
• Artificial Intelligence in Manufacturing

Background:

• Worn cutting tools degrade machining accuracy and increase surface roughness, necessitating effective tool replacement strategies.
• Traditional methods for determining tool wear require frequent operational shutdowns for measurements, hindering practical industrial application.
• Existing wear regression studies are complex and difficult to implement in real-time production environments.

Purpose of the Study:

• To develop a novel, non-intrusive method for determining the optimal timing for replacing worn cutting tools.
• To eliminate the need for direct tool wear measurement, simplifying operational procedures.
• To create a method adaptable to various machining parameters and conditions.

Main Methods:

• Utilized the Vold-Kalman filter to extract spindle current clutter signals (SCCS) by removing rotation frequency components.
• Employed convolutional neural networks (CNN) to analyze SCCS features, distinguishing between normal and severe tool wear stages.
• Developed a binary classification CNN model to identify wear states from full-life SCCS data across different cutting parameters.

Main Results:

• Successfully developed a CNN model capable of classifying tool wear stages based on processed spindle current data.
• Introduced the 'wear transition percentage' calculated from the model's output to indicate optimal tool replacement timing.
• Validated the method's effectiveness by comparing its predictions with measured flank wear data, showing high correlation.

Conclusions:

• The proposed wear transition percentage method accurately predicts optimal tool replacement timing without direct wear measurement.
• This approach significantly reduces operational complexity and is suitable for diverse machining parameters.
• The findings offer a practical solution for improving machining efficiency and product quality in industrial settings.