Development of Algorithms and Criteria for Continuous Condition Monitoring of Railroad Bearings

This thesis summarizes research conducted at the University of Texas-Pan American to investigate the implementation of condition monitoring algorithms for use with railway roller bearings equipped with wireless temperature sensor nodes. Current temperature based condition monitoring techniques utilize wayside detection equipment intermittently located along the track to garner sporadic bearing cup temperatures. Such devices are not capable of detecting brief transient temperature trending events. The developed algorithms were tested using a fleet of ten railway freight cars instrumented with wireless sensor nodes and operated by an Australian freight transportation company. Eighteen months of continuous monitoring combined with the inspection of three setout bearings demonstrate the effectiveness and accuracy of the developed algorithms as a continuous bearing health monitoring system