Predictive Maintenance of Industrial Robots Based on Real-Time Data Collection

摘要

Modern industrial enterprises face many challenges, including intense competition, increasing consumer demands, and stricter standards. To maintain competitiveness, it is critical to minimize production losses, including equipment downtime, which can lead to significant economic losses. One of the effective approaches to reducing equipment downtime is the implementation of predictive maintenance, which allows real-time monitoring of equipment conditions and forecasting possible failures. This paper proposes a predictive maintenance algorithm that uses the Palmgren-Miner rule to estimate the accumulated fatigue damage of robot joint components. The algorithm calculates contact stresses and accumulated fatigue damage using real-time data on joint moments and angles. The method considers dynamic loads and robot operating modes, providing accurate forecasts of the components' condition. The analysis is based on real data collected during a production operation. It demonstrates the potential of the proposed approach to reduce downtime, optimize maintenance costs, and increase the service life of a robotic system.