Fault-Tolerant Optimization of Thrust Bearing Shutdown Alarms Using PT100 Sensor Data with Transformer Algorithm

The reliability of thrust bearings is critical for the safe operation of rotating machinery, and excessive temperatures can lead to catastrophic failures. PT100 temperature sensors are widely used for monitoring, but sensor malfunctions can cause false shutdown alarms, disrupting operations and increasing costs. This study proposes a fault-tolerant approach utilizing the Transformer algorithm to detect and manage sensor anomalies in multi-sensor thrust bearing systems. By identifying abnormal sensor readings and distinguishing them from genuine high-temperature conditions, the method effectively reduces false alarms. The results demonstrate enhanced system reliability and operational stability, making it suitable for various industrial temperature monitoring applications.