Study on the Maintenance of Subway Equipment Based on Data Mining Techniques

摘要

Attempting to avoid severe malfunction, save cost and reduce risk which could lead to serious impact on an operating subway system, applying proper maintenance modes would be critical for all installed equipment. Literature review about maintenance strategy shows that application of Data Mining algorithm and computer assisting system would be a good way for improving maintenance efficiency. The article presents a measure employing Data Mining techniques, which, by analyzing historical maintenance record, aims to create structured datasets for devices of subway system for monitoring their daily status and possible trend of failure development and trying to apply predictive maintenance before any device having an actual breakdown. Besides, a BP neural network model is introduced for diagnosing primary sub-system on safety, such as brake shoes mounted on trains. Experiment results show that the method could be promising for preventive maintenance in subway.