Temperature normalisation and moisture assessment of oil-paper insulation based on dielectric response in extremely cold conditions

Abstract

Frequency domain spectroscopy (FDS) is widely used for assessing the condition of oil-paper insulated electrical equipment. However, in low temperatures, measurement results often deviate from expected values. This study investigates the low-temperature dielectric response characteristics of oil-paper insulation under large temperature differences (−60°C to 30°C), wide frequency ranges (1 mHz–5 kHz) and varying moisture contents (0.41%–3.91%). The mechanisms of low-temperature effects on oil-paper insulation relaxation polarisation are revealed. Further, a novel low-temperature normalisation method and a moisture evaluation method are proposed based on the improved Havriliak–Negami model. Compared to traditional methods, the new approach significantly enhances the accuracy. The goodness of fit R² for temperature normalisation improves from 0.9526 to 0.9757, and the error in moisture content evaluation has reached 0.498%. Finally, the novel approach is applied to diagnose the condition of a damped bushing model, demonstrating high potential for practical applications. This study enables condition diagnosis of oil-impregnated electrical equipment in extremely cold environments, filling a critical gap in the field.