Multi-sensor heterogeneous data-based online tool health monitoring in milling of IN718 superalloy using OGM (1, N) model and SVM

Cutting tool health monitoring (THM) is of great practical significance to estimate the tool life to enhance productivity, machining efficiency and reduce machine tool downtime. Therefore, the present study aimed to develop an online THM system using an optimization grey model (OGM (1, N)) and support vector machine (SVM) by integrating multiple sensor signals (MSS) in situ. The MSS was acquired using a vibration sensor and a complementary metal-oxidesemiconductor (CMOS) camera at optimum cutting parameters during Inconel718 milling. The valuable features of tool vibration are extracted in time, frequency, and time-frequency domains, and machined surface texture characteristics are extracted using a three-dimensional grey level co-occurrence matrix (3DGLCM). Finally, an OGM (1, N) and SVM models are employed to realise the preceding methodology, which predicts flank wear in real-time based on extracted features of multiple sensor heterogeneous data with an average error of 4.07% and 7.38% compared to the experimental results.