Research On Anti-Electromagnetic Interference Technology For High-Frequency Current Partial Discharge Detection

Partial discharge detection is an effective method to evaluate the insulation status of electrical equipment.The high-frequency current method has been widely used due to its advantages such as convenient sensor installation and wide detection frequency band.However,the complex electromagnetic interference existing on the test site to detect the high-frequency PD signal will have an important impact on the detection results and even cause serious problems such as wrong alarming and failing to alarm.Therefore,obtaining accurate PD high frequency signal is the key to making a correct judgment on the insulation performance of electrical equipment.Based on the analysis of domestic and foreign PD detection de-noising techno logy,the de-noising technology for high-frequency current PD detection was studied in this paper.The main contents are as follows:The representative substations and test points were selected,and the oscilloscope and high-frequency partial discharge detection device were used to systematically study the time-frequency and statistical characteristics of electromagnetic interference in substations In addition,the location of the interference signal sources was performed.The results showed that the periodic narrowband interference and the external discharge of high voltage equipments in substations are the main forms of electromagnetic interference to high frequency PD detection.The maximum amplitude of narrowband interference can reach more than 50mV,and the frequency spectrum is mainly distributed in the range of 300-1000 kHz;over 70%of the substations exist discharge-type interference signals,and the frequency spectrum is usually below 30MHz.According to the typical high-frequency electromagnetic interference characteristics of the test site,a combined interference suppression strategy based on wavelet transform and pulse clustering separation was proposed.Four evaluation indicators including signal-to-noise ratio,mean square error,energy difference and correlation coefficient were introduced to quantify the de-noising effect of wavelet decomposition to find the optimal wavelet parameter combination suitable for on-site PD detection.Partial discharge online monitoring requires the detection system to be able to process the collected data in real time.In order to reduce the computational complexity and time consumption of the de-noising process,a noise reduction plus clustering method based on wavelet decomposition was proposed,and its noise reduction effect was tested and verified.Finally,the proposed interference suppression strategy was used to optimize the anti-interference performance of the high-frequency partial discharge detection device developed by the laboratory and applied to the field partial discharge detection.The partial discharge defects of the equipment under test were successfully detected,and the cause of the failure was analyzed also.