Statistical analysis of moisture's effect on AC breakdown strength of TiO 2 nanofluids

Transformer oil-based nanofluids are widely investigated and exhibit an outstanding dielectric breakdown strength. Traditionally, the modifying effect of nanoparticles on the insulating properties of transformer oil is evaluated by average breakdown voltages. Little attention, however, was paid to the data dispersion of breakdown voltage of nanofluids. In this work, three types of statistical analysis methods including Normal, Gumbel and Weibull distributions were employed to analyze the AC breakdown performance of both mineral oil based TiO2 nanofluids and pure mineral oils with different levels of moisture contents. And the applicability of all the three distribution methods for estimating the low probability breakdowns of the nanofluids was discussed. The three parameters Weibull was found to give a best overall distribution fit. The weakest link theory was used to explain the effects of TiO2 nanoparticles on AC breakdown strength of mineral oil. TiO2 nanoparticles can absorb dissolved water in oil, relieving the distortion of electric field and preventing the formation of conductive bridge in oil. Therefore, the resistance to moisture degradation of TiO2 nanofluids was increased.