Study On Time-frequency Characteristic And Regression Model Of Off-line Arc Conducted Interference Of Pantograph-catenary System

With the rapid development of the electrified railway,the electromagnetic noise which is generated by the catenary-pantograph arc is increasingly prominent in the high-speed rail system.The jamming signal poses a threat to the reliability of power supply system and affects the locomotive control signals and the correct transmission of the communication signal along the railway,and it does harm to health of the residents along the railway.In order to do the further study on the formation mechanism and suppression methods of the electromagnetic noise,we use the the self-made experiment platform and collection and save the signal of voltage,current,speed and pressure mainly.The simulation experiments were conducted under different conditions(contact pressure,running speed and loop current)and the waveform and data of different conditions were obtained.We use the wavelet algorithm as a method to analyze the experimental data and study the characteristic of conducted interference.The results show that there are many harmonics including odd harmonics and even harmonics during the off-line state of the catenary-pantograph system and the DC component can cause the magnetic saturation of the transformer.After analyzing much experimental data,we got the distortion situation of current and calculated the total harmonic distortion rate(THD)of current.By analyzing the characteristic curves of THD and working conditions,we found that THD decreases with the increase of contact pressure,the increase of running speed,and decreases slowly with the increase of loop current.Finally,the support vector machine(SVM)and the particle swarm optimization(PSO)were used to build the model of THD which can be used to make the prediction of conducted interference of the catenary-pantograph system.