On-line Detection And Identification Of DC Series Arc Faults

DC series arc fault in DC system is easy to cause equipment and personnel safety problems.With the rapid development of DC power system in photovoltaic systems,electric aircraft and electric vehicles,detection technology of DC series arc fault has received wide attention.Because of the complexity of arc physical process and the lack of zero-crossing characteristics,there is a lack of detection technology to judge DC series arc fault in the world.In this paper,DC motor,switching power supply and resistance,are used to simulate three different types of loads including inductive,capacitive and resistive loads in DC system,and an arc fault experimental platform is built.The time and frequency domain characteristics of DC arc and the detection methods of DC series arc fault will be deeply explored so as to realize online identification and detection of DC series arc fault.First,we design the mechanical structure of the arc generator and built a DC series arc fault experimental platform.The grating ruler is used to transmit the position signal of the motion mechanism of the arc generator to the PLC.The PLC controls the rotation of the stepper motor and the mechanical structure transforms it into the lateral movement of the guide rail,which promotes the separation of the contacts and generates the arc,thus realizing the accurate control of the separation speed of the contacts and the gap size of the contacts.According to the experimental voltage and current conditions,110 V DC power supply,DC motor,switching power supply and high-power resistance are selected and we finally build a DC series arc fault experimental platform.Secondly,the effects of contact separation speed,contact clearance,circuit current and different load types on arc current waveform are considered comprehensively.We constructed a large number of experimental databases by using the control variable method and qualitatively analyzed the influence of different factors on the arc current waveform based on the time domain waveform.Then,comparing several arc fault detection algorithms based on time domain and frequency domain,a phase frequency domain ratio decision method is proposed for DC series arc fault diagnosis.Based on time domain,the maximum difference of arc current is analyzed,and the corresponding eigenvalues are extracted to distinguish the arc fault when the loop current is small.The frequency standard deviation of arc current is analyzed based on frequency domain.After comparing the advantages and disadvantages of different algorithms,a phase frequency domain ratio judgment method based on FFT is proposed to diagnose dc series arc faults under different conditions,and the phase frequency domain maximum judgment method is adopted to eliminate interference signals and reduce misjudgment in the detection process.Finally,the dc series arc fault diagnosis module is made to realize the fault diagnosis algorithm.STM32F405RGT6 microcontroller is used as the main control chip of fault diagnosis module.The current data are transmitted and stored by using DMA function of the single chip microcomputer.And the arithmetic speed is improved by using MCU DSP floating point arithmetic function.Statistics of dc series arc fault detection accuracy and misjudgment rate.Experiments show that the algorithm can effectively identify arc faults in the circuit and has high anti-interference performance.