Wayside detectors are nowadays used to not only preserve the infrastructure but also to detect abnormal behavior or condition of railway vehicles or locomotives. Typical way side detectors are hot box, hot wheel, wheel impact load detectors, wheel profile detectors and bearing acoustic detectors, which provide insights on the condition of wheels and bearings.

In this paper, a novel approach is presented to detect and analyze deviating bogie performance on railway vehicles using data from multiple wheel impact load detectors over time. The analysis of the wheel impact load time series can be used to identify bogies with deviating performance, which in turn, results in increased wear on wheels and bearings. An analytics scheme has been developed based on the load distribution deviations to detect bogies with performance issues on a fleet. The analytics scheme is then evaluated on a fleet operating in Scandinavia and deviating bogies have been inspected and further assessed to determine root causes for the deviating behavior.

It is concluded that railway vehicles with deviating behavior can be detected using the analytics scheme and that these deviating bogies exhibited errors requiring repair. The detection is possible well before the bogie problem would render an unplanned stop or pose a hazard for operation.