Research On UHF Sensor And Signal Recognition Of Partial Discharge Detection In Switch Cabinet

Partial discharge fault of switch cabinet is one of the main causes of high voltage equipment fault.Rapid and accurate detection of partial discharge fault plays an important role in ensuring normal operation of power equipment.Ultra-high frequency(UHF)detection method is to determine the discharge fault type and hazard level by detecting the300MHZ~3GHZ ultra-high frequency electromagnetic wave signal generated at the discharge end by the ultra-high frequency sensor.Moreover,the UHF sensor is easy to install,has low interference to power equipment,strong anti-interference ability and high sensitivity.Therefore,the UHF detection method has become an important research direction in this field.In this paper,UHF detection sensor,namely UHF antenna,is taken as the research object,and the miniaturization technology and performance parameter optimization method of UHF sensor are mainly studied.;The effects of different physical structure dimensions of the Archimedes spiral antenna and the fourth-order Hilbert fractal antenna on the performance of the sensor are studied.The performance of the two types of UHF antennas under the optimal size parameters is compared and analyzed;Through theoretical analysis and simulation,the miniaturization design of UHF sensor based on Archimedes spiral antenna was determined,and the optimal size of sinusoidal antenna was determined by HFSS simulation analysis;The miniaturization techniques of the self-similar sinusoidal antenna,the improved sinusoidal antenna,and the improved sinusoidal antenna embedded in a ring are studied respectively,And made a miniaturized actual UHF sensor;the partial discharge detection platform and the partial discharge fault type physical model are built,and the applicability of the miniaturized UHF sensor to different fault types is studied;The partial discharge fault signal is processed by FIR filter,and its eigenvalues are extracted by wavelet packet decomposition and triple envelope envelope.The two features compound to form the image matrix eigenvalue.The data reduction and fault classification are realized by convolution neural network(CNN),and the pattern recognition algorithm of partial discharge fault is constructed.The research on the miniaturization of UHF antennas shows that the size of the partial discharge UHF sensor can be further reduced,and the performance of the small size sensor can be improved,which is conducive to improving the convenience of sensor installation and use in engineering applications.