Machine Learning-Driven Tool Wear Prediction Using Electromotive Force and AC Motor Slip

Precision manufacturing relies heavily on accurate tool wear prediction systems to maintain quality and optimize production efficiency. However, existing systems are often costly and difficult to integrate, limiting their practical use in industrial environments. In this paper, we present an intelligent Tool Condition Monitoring (TCM) system that addresses these limitations through a novel, low-cost approach. The proposed method combines signal measurements, specifically, the electromotive force (EMF) at the tool-workpiece interface and the slip of an AC induction motor (SIM), to monitor the progression of flank wear. These physical indicators are used as inputs to an AI-driven hybrid machine learning (ML) system, enabling real-time prediction and classification of tool wear states. The architecture is designed for ease of implementation, affordability, and accuracy, making it suitable for both new and retrofitted manufacturing setups. The system categorizes the cutting tool’s condition across various stages, from initial break-in to final failure when tolerance limits are exceeded. Additionally, it can be deployed within an IoT-enabled framework for remote diagnostics and production optimization. Experimental validation demonstrates classification accuracy ranging from 84.1% to 91%.