End-effector tools wear prediction: machine and interaction modeling, system identification based on the EKF approach.

This thesis work is part of the project MOREPRO, an industrial program owned by BrainTechnologies, whose main goal is the realisation of a predictive monitoring system for the tool’s wear and for the state of health of the machinery in real-time. The whole project is carried out in teamwork; in particular, the team’s partition is the following: •modelling team •prediction team •requirements team. My initial role was within the modelling team with the aim of finding a kinematic and dynamic model of the system and create a simulation environment for the robot considered. Afterwards, I continued the modelling work for the prediction team in order to find a model for the interaction between the end-effector of the machine and the workpiece. The crucial parameter considered in the interaction model is the friction coefficient because it has a strong impact on the tool’s wear. Different models of friction coefficient were studied and once found the final interaction formulation, this is added to the CNC machine model previously carried out by the prediction team, in order to achieve a better degree of accuracy and detail of the system. During this work, I also had the opportunity to become more familiar with state observers, because an Extended Kalman Filter is used in order to perform the system identification, able to estimate the unknown parameters. The whole work performed can be divided as follow: •machine modeling from a Kinematics point of view; •study and implementation of the interaction between the end-effector and the work-piece; •system identification solution using an Extended Kalman Filter approach for parameters estimation. The whole project was developed and implemented on Mathworks environment, bothMatlab and Simulink have been employed.