Diagnosis and Prognosis of a Cartesian Robot’s Drive Belt Looseness

Abstract

Maintenance cost is among the highest operational expenses for manufacturing firms. Proper scheduling of maintenance intervention results in optimized equipment life utilization, higher product quality, and reduced costs. For Cartesian Robot's accuracy and precision it is important that the belt is well calibrated. Nonetheless, the manual assessment of calibration requires to stop the robot, which in turns causes the stop of the production with related consequences. In this work we are going to develop a Machine Learning based Classification Model, able to use cycle current consumption data of a Cartesian Robot in order to understand if the drive belt is calibrated or not. The trained model will be tested on completely new data whose label is known, to ensure its reliability.