Real-time Train Control under Strong Wind Risk - Methodology and Tools

摘要

This paper introduces our developed strong-wind prediction system of WINDAS, which can be applied for real-time train operation control under strong wind risks. As mathematical procedure, a statistical approach, namely a time-series analysis, is employed to predict wind velocities some ten minutes ahead. To hold an adequate accuracy of prediction and a high robust for calculating system, multiple prediction models are examined with sets of parameters estimated with several lengths of observed values in time series. Using the results from WINDAS against the observed wind velocities at past events, the efficiency of wind prediction for train regulation is confirmed in comparison with a conventional train regulation for strong winds. For a practice, user interface is considered especially targeting train control operators. Additionally, draft systems for a real-time data communication are also constructed with applicability for present on-active cable communication.