On statistical models of the amplitude and the duration of pulsed electromagnetic interference in automatic-control and telemechanic channels of subway lines

Abstract

On the basis of the results of experimental studies of the interference from the traction current of subway trains, the distribution type of this interference has been established. Approximation of experimental data by an analytical expression based on the Pearson curve family underlies the proposed method. The approximation is performed by calculating the first four distribution moments. The common array of the initial data on the electromagnetic interference is divided into two parts, depending on the type of the rolling stock—either with the switched-on-off tracking regenerative-rheostatic braking system or without it. The 7‒10-min recordings of the data on the traction current for trains with or without regenerative braking have been processed. To normalize the experimental data, the mathematical expectation estimate is subtracted from the former and the obtained values are divided by the estimate of the standard deviation. The resulting new data array has normalized properties, i.e., the zero mathematical expectation and a unit standard deviation. Then, from the normalized data array, the first four sampling moments are found, on the basis of which the corresponding Pearson curve is selected. The method underlies the estimation of the distribution density of the pulsed interference amplitude and the duration of the electromagnetic interference pulses. It has been shown that, in all 12 frequency ranges from 50 to 780 Hz that are used in subway automatic control and telemechanical devices, the distribution density of the pulsed interference amplitude and the interference pulse duration can be described by the analytical expression of the β-distribution of the first and second kind.