An algorithm for improved performance of railway condition monitoring equipment: Alternating-current point machine case study

This paper develops a new approach for fault detection and diagnosis utilising parameters collected from low-cost and accessible sensors. An electrical railway point machine within a railway junction is used as a case study. The paper shows that electrical active power collected from electrical current and electrical voltage sensors can be used for condition monitoring systems. The methodology proposed in this paper utilises Wavelet Transforms and Support Vector Machines. It was found that together these methods can detect and diagnose misalignment faults of electrical railway point machine to a high degree of accuracy. Furthermore, it was proved that the approach can provide an indication of the severity of the faults. This work was carried out in collaboration between the University of Birmingham and Central Japan Railway Company.