Abstract The sensitivity of partial discharge detection is compromised by electromagnetic interference, leading to inaccuracies and inefficiencies in quantitative detection and pattern recognition of partial discharges. To address this issue, a method based on a large number of sample data is proposed to simulate partial discharge pulse generation using electronic circuits while controlling pulse phase and amplitude, thereby outputting discharge characteristics. A reconfigurable hybrid digital-analog signal generation system is designed and developed. The system is capable of generating pulse voltages with a rise time of less than 1 ns, controlled precisely by a microcontroller and an upper computer. The original PRPD spectrum of partial discharges showed high similarity with the system-generated PRPD spectrum, demonstrating that this system can effectively output different types of partial discharge characteristics. This system provides a robust foundation for machine learning-driven insulation diagnostics while substantially reducing calibration complexity.
Shi et al. (Fri,) studied this question.