Principal Component Analysis of Multiple Sensor Signals for Tool Wear Identification during Composite Materials Machining

Composite materials are among the most interesting engineering materials because their complex structure allows for very high strength and stiffness properties per unit weight, i.e. specific strength and stiffness, in comparison with the specific properties of metal alloys. These properties are highly desirable in a wide variety of advanced engineering applications, ranging from commercial aircrafts to sports equipment, prompting the extensive employment of composite materials in many different industrial sectors. To achieve the effective optimisation of composite materials machining processes, a profound understanding of the phenomena concerned with material removal modes and tool wear development is fundamental. Sensor monitoring of machining processes has been extensively utilised for optimal selection of cutting parameters, tool conditions classification, chip form control, cut surface integrity assessment, chatter detection, etc. In this work, the identification of tool wear state was carried out through detection and analysis of acoustic emission and cutting force signals generated during orthogonal cutting of different types of composite materials. Sensor fusion of signals provided by sensors of different nature was performed through signal data processing based on the Principal Component Analysis (PCA), in order to reduce the high dimensionality of sensor signals by extracting significant signal features to utilise for pattern recognition of tool wear state.