Knowing the response of railway vehicles running on a track (such as the derailment quotient, reductions in the wheel load and roll/pitch) enables the provision of maintenance services on a priority basis to track segments where railway vehicles tend to vibrate or where their safety may be compromised. This reduces the volume of maintenance work necessary without compromising the level of safety. To understand wheel loads, lateral pressure, vibration acceleration and other data related to vehicle response, the accepted method has been to simulate rolling stock motion, but this can be difficult to perform at maintenance depots and other field units.
   The RTRI has therefore developed a method of using the track maintenance control database system(LABOCS) for calculations to simulate rolling stock motion. The computer used in making these calculations creates a model based on vehicle dimensions to simulate rolling stock motion, and then creates a statistical model (i.e. a linear digital filter) approximately equivalent to the simulation model. The statistical model predicts the response of rolling stock based on track inspection data (Fig. 1). As an example, the method predicts wheel load changes in several seconds for a route length of 100km, with an error margin of about 1kN in terms of standard deviation both for straight and curved sections, as compared to simulation results.
   Application of this method to the LABOCS of JR companies will make it possible to calculate the response of rolling stock at track maintenance depots (Fig. 2), thereby improving the efficiency of track maintenance work.