Research On Partial Discharge Detection Technology Of Portable High-voltage Equipment Based On Zynq

With the development of modern industry and the improvement of the level of urbanization,the scale of the power system is continuously increasing,and operational safety is receiving more and more attention.As an important device in the power grid system,high-voltage electrical equipment’s smooth operation is a reliable guarantee for the normal operation of the power system.The main reason for the failure of high-voltage electrical equipment is insulation degradation and damage.Partial discharge phenomenon can reflect the degree of deterioration of equipment insulation structure and is an important basis for judging the insulation state of equipment.This thesis deeply researched the online partial discharge detection technology to analyze the phenomenon of partial discharge mechanism and characterization parameters,portable design line partial discharge detection system to detect a variety of ways,and made for the lack of traditional pattern recognition in efficiency and accuracy.A new pattern recognition algorithm for partial discharge signals.First,according to the cause and type of partial discharge,the partial discharge process was analyzed and the corresponding characterization parameters were extracted to characterize the severity of partial discharge.According to the cause of partial discharge and the type of damage to the insulation structure,several common types of partial discharge are analyzed,and the formation process of partial discharge is analyzed using the equivalent circuit model of air gap discharge in partial discharge type.According to its formation process,an air gap partial discharge model was built in Simulink for simulation verification,and parameters such as partial discharge amount,discharge energy,average current / power,pulse adjacent discharge time,etc.were extracted to characterize the severity of partial discharge phenomenon and severity.Next,for the detection of various high-voltage electrical equipment on the spot,the partial discharge signals generated and the applicable online detection methods are different,so a portable partial discharge online detection system is designed in combination with multiple online detection methods.Design the hardware system and PCB,and analyze the existing signal integrity problems;build an embedded system on the Zynq processor platform to design the logic circuit and system software in the embedded system;the instrument uses partial discharge Signal acquisition and conversion,data processing,and data transmission finally provide the display of partial discharge signal general parameters,phase resolved partial discharge patterns,and time-domain waveforms on the user graphical interface of the instrument.Finally,the traditional partial discharge type recognition method is mainly recognized by prior knowledge,and there are problems of low efficiency and accuracy.Therefore,a pattern recognition algorithm based on GA-BP neural network is used.The main steps include feature parameter extraction,neural network training and pattern recognition.The algorithm first uses the statistical feature parameter method to extract the feature parameters of the partial discharge signal as the neural network input vector,and then uses the BP neural network for training.Finally,the trained model is subjected to partial discharge signal pattern recognition.Because the BP neural network has the problem of being easily influenced by the initial connection weights and thresholds of the network and due to the poor global search ability,it is easy to fall into the local optimal solution.A genetic algorithm is proposed to optimize the initial weights and thresholds of the BP neural network.Experiments show that after optimization by genetic algorithm,the partial discharge type recognition rate has been improved,verifying the effectiveness and rationality of the algorithm.