Towards Data Acquisition for Predictive Maintenance of Industrial Robots

Abstract

Predictive Maintenance of industrial robots offers the potential to increase productivity and cut costs in highly automated production systems. The success of such maintenance strategies is highly dependent on the data acquisition strategy used to monitor the robot’s health state. In this publication, we first describe a methodology for deriving a suitable data acquisition strategy. Second, we apply this methodology to shape a data acquisition strategy for articulated robots. This strategy defines the robot components for which data is acquired, the robot trajectories used for the data acquisition and the frequency that measurements are taken. To conclude, we discuss the methodology’s limitations.