Making informed decisions in cutting tool maintenance in milling: A KNN-based model agnostic approach
This addresses the need for transparent decision-making in machining maintenance for manufacturers, though it is incremental as it builds on existing machine learning methods in TCM.
The study tackled the problem of limited interpretability in Tool Condition Monitoring (TCM) by developing a KNN-based white-box model that analyzes real-time force signals to detect tool wear and provide insights into decision-making, enabling manufacturers to make informed maintenance choices.
Tool Condition Monitoring (TCM) is vital for maintaining productivity and product quality in machining. This study leverages machine learning to analyze real-time force signals collected from experiments under various tool wear conditions. Statistical analysis and feature selection using decision trees were followed by classification using a K-Nearest Neighbors (KNN) algorithm, with hyperparameter tuning to enhance performance. While machine learning has been widely applied in TCM, interpretability remains limited. This work introduces a KNN-based white-box model that enhances transparency in decision-making by revealing how features influence classification. The model not only detects tool wear but also provides insights into the reasoning behind each decision, enabling manufacturers to make informed maintenance choices.