Characteristics And Suppression Methods Of High Frequency Partial Discharge Electromagnetic Interference In 35kV Collecting Cables Of New Energy Stations

With the promotion of energy transition in China,the proportion of new energy generation in the power system is constantly increasing.35kV power cables are widely used as collecting lines in new energy stations,and their reliable operation is closely related to the safety of personnel and equipment in the station and the benefits of power generation.High-frequency partial discharge online detection is an important means to discover potential insulation defects in cables and evaluate their operating status.For a long time,complex electromagnetic interference at substation sites has been a bottleneck for online partial discharge detection.Compared with traditional power stations,new energy stations have a large number of power electronic devices,and their interference is more severe and complex.If not handled properly,it will seriously hinder the effect of online partial discharge monitoring and lead to mistakes and failure of reporting.Based on the engineering practice of high-frequency partial discharge online monitoring of cables in Ulanqab New Energy Station,this paper conducts in-depth research on the characteristics of high-frequency electromagnetic interference in the station.For a typical strong electromagnetic interference signal with both high frequency and high amplitude characteristics unique to new energy stations,a proprietary interference suppression method is proposed.The main contents are as follows:The characteristics of the equipment working conditions in the new energy station and their impact on high-frequency partial discharge online monitoring are analyzed.The frequent switching of power electronic device switches in new energy stations causes dv/dt and di/dt to interact with capacitance and inductance in the circuit to form switch transient interference.The parasitic parameters inside the equipment provide coupling paths for interference and are coupled by the HFCT on the ground lead of collecting cable,thereby introducing a large number of pulse-type interference into the cable high-frequency partial discharge monitoring device,seriously affecting the normal operation of the partial discharge monitoring device.A field test of high-frequency electromagnetic interference(EMI)signals was performed at multiple cable measurement points at a new energy plant.Test points were chosen based on the equipment layout,topology,and zoning characteristics within the plant.The timing waveform,frequency spectrum,and phase statistical characteristics of high-frequency current signals were systematically studied.It was discovered that an intermittent pulse-like interference signal occurred frequently with a high amplitude on the order of microseconds/time which heavily influenced the monitoring system’s effectiveness by complicating discharge pulse extraction and multi-source signal separation.This interference type represents the most significant threat to fieldrelated operations.The waveform of this interference signal exhibited highly similar features,forming multiple clusters on the PRPD map and resembling a Vshaped pattern.Additionally,the pulse amplitude and polarity were closely related to three-phase power signals and repeated periodically across different power frequencies.The spectrum analysis indicated that the measured pulse signals had a concentration distribution below 5MHz.These findings provide a data foundation for developing targeted anti-interference strategies.In response to the above-mentioned strong electromagnetic interference problem with high frequency and high amplitude,a pulse extraction method based on the maximum inter-class variance method(OTSU)and local energy threshold is proposed according to the characteristics of the interference signal.The method is verified using on-site data and the results show that it has a good pulse extraction effect.For the typical repetitive interference pulses existing in most on-site measurement points,a repetitive pulse interference separation method based on spectral clustering is proposed.A similarity matrix is constructed based on the similarity between pulses,and the silhouette coefficient is introduced to determine the number of clusters.Several similar pulses were successfully separated from the on-site data.On the basis of spectral clustering,combined with the template test method,a repetitive pulse interference suppression strategy is proposed.It is verified with on-site data and the interference suppression rate reaches 93.67%,proving that this strategy is feasible.