Wear Parameter Diagnostics of Industrial Milling Machine with Support Vector Regression

Modern industrial machine applications often contain data collection functions through automation systems or external sensors. Yet, while the different mechanisms might be effortless to construct, it is advised have a well-balanced consideration of possible inputs based on characteristics, usage, and operational environment. Prior collected parameters reduces risk excessive data, yet another challenge remains distinguish meaningful features significant for purpose. This research illustrates peripheral milling pre-processing approach diagnose relevant blade wear. The experiences gained from this encourage conducting pre-categorisation purpose, those being manual setup programmable logic controller (PLC) system calculated parameters, measured under study. Further, results raw phase performed with Pearson Correlation Coefficient permutation feature importance methods indicate that most dominant correlation recognised wear characteristics in case context perceived vibration excitation monitoring. root mean square (RMS) signal further predicted by using support vector regression (SVR) algorithm test SVR’s overall suitability asset’s health index (HI) approximation. It was found SVR has sufficient parameter behaviour forecast capabilities used prognostic process its development. Gaussian radial basis function (RBF) kernel receives highest scoring metrics; therefore, outperforming linear polynomial kernels compared as part