Partial Discharge Detection of High Voltage Electrical Equipment using Multiclass Support Vector Machine

In recent times, condition-based monitoring of power systems has grown in significance to guarantee the dependable and effective functioning of various types of High Voltage Equipment (HVE). However, the degree of Partial discharge detection of high voltage electrical equipment can vary from mild to severe. As a result, this research has presented Multi-class Support Vector Machine (MSVM) technique that provides a dependable means of electrical system monitoring for minimizing the probability of failure. Initially, when high-voltage equipment works, data gathering devices record signals from the equipment in real time. Ultrasonic testing is discover to select high-frequency sound waves that are brought on by protecting fractures. It is a flexible instrument for identifying different PD issues in intricate electrical equipment because of its components’ capacity to handle a wide range of classes. The discharge signals are classified using the designated MSVM in order to correctly identify various types of faults. According to the results, it illustrated that when compared to existing models which includes Multilayer Multifunction Electrical (MME), Stacking Ensemble Neural Network (SENN), and Partial Discharge of Switchgear Detection (PDSD) the proposed methodology achieved better accuracy of $\mathbf{99. 61 \%}$ respectively.