Study On Short Circuit Characteristics Of DC Traction Power Supply System For Urban Rail

Reliable DC protection of urban rail transit can ensure the safe operation of urban rail transit and the personal safety of passengers,but DDL protection may be misoperated or refused to move.In order to ensure the safe and reliable operation of urban rail transit,the protection of DC power supply system of urban rail transit must be studied in depth.Firstly,the existing protection methods of DC traction power supply system for urban rail are summarized.Then the load current of the feeder of the traction substation is obtained by establishing the simulation model.The simulation results are compared with the measured current of a train in Shanghai.Through the analysis,the causes of over-section current in train are obtained,and the measured current of DDL protection maloperation caused by a subway in Wuhan proves that over-section current has the characteristics that may cause DDL protection maloperation.By establishing a rail model considering skin effect,the far-end short-circuit current at the feeder is obtained.The Pearson coefficient of simulated and measured far-end short-circuit current is 0.999,which verifies the rationality of the model.By comparing the over-section current with the far-end short-circuit current,the current increment and the current change rate of the over-section current and the short-circuit current are similar,but the high frequency component of the short-circuit current is more than the high frequency component of the over-section current.In view of the difference in frequency between over-section current and far-end short-circuit current,the empirical wavelet(EWT)theory proposed by Gilles is introduced in this paper.EWT has a complete mathematical derivation process,good adaptability,stable decomposition results,and its calculation is more rapid and simple.However,EWT may have "over-segmentation" and fail to solve the problem of how to determine the number of spectrum partitions M.In order to overcome the shortcomings of EWT,an improved EWT algorithm based on Gauss filtering is applied to effectively avoid the phenomenon of over-segmentation.Then,an improved EWT algorithm based on one-dimensional maximum entropy theory is used to update the number of spectral partitions M,and the number of spectral partitions M and the partition boundary are obtained adaptively.The improved EWT algorithm is combined with the energy weight to extract the characteristic frequency f*,and the implementation process is given.The analysis of over-section current and short-circuit current shows that the improved EWT can extract effective characteristic frequencies,while the unmodified EWT can not obtain effective characteristic frequencies.The characteristic frequencies of several groups of measured and simulated data are analyzed.It is found that the characteristic frequencies of short-circuit fault current are between 1000 Hz and 5000 Hz,and the characteristic frequencies of over-section current are lower than 100 Hz.The difference between them is obvious.Finally,the algorithm of extracting characteristic frequency is compared with DDL protection.It is found that the correct rate of fault identification of DDL protection is 9% higher than that of the algorithm of extracting characteristic frequency of DDL protection.It proves that the method of extracting characteristic frequency of short-circuit current by improved EWT proposed in this paper can distinguish short-circuit fault current from train over-section current better.