The paper discusses the preconditions of the diagnosing system methodology development of dynamic impact a rolling stock on the basis of processing and analysis data, which were obtained in operation on the results of measurement parameters that characterize the dynamic vibration processes of the mechanical system "rolling stock - track". On the basis using the processing methods of time series and stochastic processes was established the relationship between these dynamic processes and wheel defects, and were designed experimental data processing alghoritms, which will be an interal part of the intellectual decision making systems when assessing the impact of the rolling stock on the track. Some results of the experimental and theoretical investigation of rail acceleration data, which were registerd by passing the train, are presented.  The computational aghoritm of the spezialized pre-processing of the multidimensional signal that is recorded by this system is described. As processing results of rail accelerations were set the growth parameters of all statistical indicators this data with increasing of the train speed. It was detected that the level of indicators for the accelerations of the rails in the vertical direction is two times higher than the corresponding parameters in the horizontal direction.

This work is supported by: KEGA