Application of Deep Learning and Symmetrized Dot Pattern to Detect Surge Arrester Status

Surge arresters primarily restrain lightning and switch surges in the power system to avoid damaging power equipment. When a surge arrester fails, it leads to huge damage to the power equipment. Therefore, this study proposed the application of a convolutional neural network (CNN) combined with a symmetrized dot pattern (SDP) to detect the state of the surge arrester. First, four typical fault types were constructed for the 18 kV surge arrester, including its normal state, aging of the internal valve, internal humidity, and salt damage to the insulation. Then, the partial discharge signal was measured and extracted using a high-speed data acquisition (DAQ) card, while a snowflake map was established by SDP for the features of each fault type. Finally, CNN was used to detect the status of the surge arrester. This study also used a histogram of oriented gradient (HOG) with support vendor machine (SVM), backpropagation neural network (BPNN), and k-nearest neighbors (KNN) for image feature extraction and identification. The result shows that the proposed method had the highest accuracy at 97.9%, followed by 95% for HOG + SVM, 94.6% for HOG + BPNN, and 91.2% for HOG + KNN. Therefore, the proposed method can effectively detect the fault status of surge arresters.

An Optimized Pulse-Counting Method for Compensation Vector Calculation in the Automatic Balancer

The accuracy of the counterweight positions in an automatic balancing system deeply affects dynamic balancing. Compensation vector is synthesized by the two counterweights located in the electromagnetic dual-weight automatic balancer. Therefore, if the position of the counterweight is inaccurate, it may result in a wrong adjustment and a larger imbalance of the rotor system. In this paper, an optimized pulse counting method for compensation vector calculation in an electromagnetic dual-weight balancing system is proposed based on a programmable logic controller (PLC). A propeller automatic balancing simulation test bench is used to verify the effect of the method by obtaining the positions of the counterweights and synthesizing the compensation vector in the working mode. The error is less than 1/80 which means that it does not exceed one step in the 80-position-balancer at 1200 rpm. The proposed control system can work without computers or high-speed data acquisition equipment, which improves the stability and flexibility of the control system, facilitates the design of the automatic balancing system, and shows excellent potential for industrial applications.

Characteristics of the Partial Discharge in the Development of Conductive Particle-Initiated Flashover of a GIS Insulator

Conductive particles are one of the most important defects which can greatly degrade the performance of gas-insulated metal-enclosed switchgear (GIS). Many efforts have been made to clarify the influence on the withstand voltage, understand the flashover mechanism, and build a comprehensive model to describe the particle-initiated flashover. In this study, a partial discharge (PD) signal detected through a photomultiplier (PMT) and recorded by a high-speed data acquisition (DAQ) system was used to analyze the discharge development of a conductive particle-contaminated GIS insulator under constant high AC voltage. An additional PMT was used as a reference to eliminate the dark count of the PMT and the data collection method of a DAQ system was optimized to capture the pulse waveform of each PD to obtain detailed physical information. Spectra of the PD pulse amplitude over pulse width, PD counts within various amplitude ranges over time and phase resolved partial discharge (PRPD) patterns of the PDs in different stages are obtained through the captured PD waveforms. Characteristics of the PDs from the application of the high AC voltage up to the flashover of the insulator were then analyzed, and it was found that the features of the PDs in the near-flashover stage were significantly different to the previous stages.