Railway Suspension System Fault Diagnosis using Cubature Kalman Filter Techniques

Abstract The railway industry is constantly looking for methods and technologies to significantly reduce the life cycle costs of their vehicles. Currently present fixed inspection-intervals lead to correspondingly high costs, and therefore, an adequate condition-triggered maintenance is needed. In this paper, a Cubature Kalman filter is proposed as a solution for fault diagnosis in the most safety critical part of the railway vehicle, i.e., its suspension system. The performance of the filter is assessed using different faulty scenarios based on validated multibody simulation model data, and both high sensitivity to faults and accuracy of estimation are shown.