A DIGITAL TWIN BASED SELF-CALIBRATION TOOL FOR FAULT PREDICTION OF FDM ADDITIVE MANUFACTURING SYSTEMS

Abstract

Among the advantages of introducing digital twins on production systems, there is the ability to identify their eventual critical state and to enable predictive maintenance policies. The failure of a manufacturing process, in general, can be anticipated in phase of simulation, if tied up to wrong settings, or in phase of operation, if tied up to environmental variables. In both cases, knowing the conditions that could cause the failure allows to intervene in a timely and effective manner. Here a method is proposed to explore the system operating parameters in a systematic way: the system is able to process signals collected in real time by machine's sensors and to reproduce both the trajectories of the moving parts and the material deposition process. This also makes possible to predict manufacturing tolerances that will be obtained. On a FDM Cartesian 3D printer a self-calibration procedure is used to find the maximum torque that can be delivered by the drives at different speeds in an automatic and repeatable way to find the maximum speed and acceleration at which the machine can operate safely. Additional accelerometers were installed on the machine to validate the adopted procedure: tests results demonstrate the effectiveness of the system.