Diagnostika provozu robotických pracovišť pro prediktivní údržbu

Abstract

The thesis discusses the topic of monitoring robotic workstations and the design of a robot data processing system for the implementation of predictive maintenance. The definition of predictive maintenance and its comparison with other types of maintenance is introduced. Based on the methods and procedures of the predictive maintenance concept, a robot operation data monitoring system is designed. The purpose is to detect the failure or degradation of the mechanical condition of the robot. The complex designed system includes data collection, recording, analysis, and archiving, including visualization of the analysis results. The visualization includes an integrated alarm system that alerts the robot when the robot's condition changes. In the practical part, the proposed system is implemented and integrated into the robot workstation. The functionality of the robot health monitoring system is verified using degradation simulation. This is simulated by using weights of different masses to represent the added load on the actuators. As a result of the work, the system is capable of detecting changes of the monitored operating variables in a repetitive robot sequence and determine a KPI indicator representing the robot state by using algorithms.