Answers are grounded in this paper and cite their sources.
Dielectric loss angle (DLA) is a critical indicator for assessing the insulation condition of high-voltage capacitive equipment. However, rapid and accurate DLA estimation within short observation windows remains challenging due to spectral leakage, the picket fence effect, and measurement noise, particularly in harmonic-rich environments. This paper presents an enhanced DLA estimation method based on a windowed interpolated discrete Fourier transform framework. A three-term maximum sidelobe decay window is adopted to suppress harmonic leakage, and a closed-form phase estimator accounting for both positive- and negative-frequency components of the fundamental is derived using the corresponding window spectrum. Furthermore, the independently estimated voltage and current frequencies are averaged and used for both phase calculations to improve noise robustness. Simulations evaluate the method under frequency deviations, harmonic interference, and varying noise levels, observation window lengths, and sampling rates. Experimental results demonstrate maximum relative errors of 1.87% and 1.32% using two and three nominal cycles, respectively. Compared with the Hann-window-based estimator, the proposed method achieves lower experimental estimation errors across the tested frequency range of 49.5–50.5 Hz. These results support its potential for short-window online insulation monitoring.
No takes yet. Share an insight, caveat, or question.
Song et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: