Derailment predictor detection system, controller, derailment predictor detection method and derailment predictor detection program

摘要

Wavelet analysis is applied to both a pitch angular velocity θ (t) and a roll angular velocity φ (t) output from an angular velocity sensor (35) installed in a train car and a wavelet coefficient (14) of the pitch angular velocity and a wavelet coefficient (15) of the roll angular velocity are calculated. Each of the two wavelet coefficients (14, 15) that change in chronological order is compared to a wavelet coefficient threshold (16) and a derailment predictor is detected when both coefficients exceed the threshold. Wavelet coefficients are used, which are calculated for a low frequency range of, for example, 0.5 to 100 Hz. Two types of derailment prediction detection algorithms, one involving a frequency domain and the other involving a time domain, are combined to increase accuracy of detection of a derailment predictor. Real-time processing of the frequency domain is made possible and prevention of derailment is achieved by means of wavelet analysis.