Research and Application Analysis on the Real-time Axle Temperature Monitoring and Diagnosis Logic of High-speed Multiple Units

Abstract

The conventional diagnostic logics for hotbox warning/alarms are mainly based on predefined absolute temperature thresholds and temperature rising slope thresholds. They often fall short in effective analysis and judgments when encountering fluctuating measured temperature values. Abnormal jumps and missing of axle temperature measurements frequently occur on the hotbox monitoring system for high-speed trains due to external factors such as inadequate hardware circuit contact or electromagnetic interference, potentially triggering false warning/alarms. To solve these, this paper proposes a logic framework for removing abnormal temperature curves. Drawing insights from an extensive analysis of field data on axle temperature measurements, this approach enables extracting features that accurately depict variations in temperature curves, to display fluctuations as well as rising trends of temperature data, thereby offering a basis for identifying abnormal temperature variations. Simulation results showcased the effectiveness and viability of the proposed scheme in identifying abnormal jumps in temperature data. Furthermore, the methodology has been assessed through long-time track operation, with a notable reduction in the false warning rate from 78.3% to 26.5%.

Keywords

• high-speed multiple units
• axle temperature monitoring logic
• false warning/alarm
• temperature abnormal jump