Influence of charged micrometer-sized particles on the breakdown voltage of air gaps

Abstract Micrometer-sized particles are ubiquitously present and can reduce the breakdown voltage of high-voltage gas insulations by several tens of percent. Despite this influence, a quantitative and mechanistic understanding of how particles determine both the probability and severity of a breakdown is missing. To address this, the breakdown voltage reduction caused by micrometer-sized dielectric $$\hbox {SnO}_{2}$$ particles in a uniform DC ambient air gap is measured and modeled. The experiments cover diameters ranging from $${1}$$ to $${170}\,\upmu \text {m}$$ . High-speed macro imaging reveals that the breakdown initiation coincides with a single charged particle approaching an electrode to within approximately $${20}$$ to $${110}\,\upmu \text {m}$$ . This occurs for positive and negative particle charges with magnitudes between 130 and 1170 fC. Across the diameters, the breakdown probability increases from 80 to 100%. The mean breakdown voltage reduction increases linearly with the particle diameter from approximately 5 to 28% at approximately $${-19.4} \,\hbox {V }{\upmu }\hbox {m}^{-1}$$ . Outliers are observed to double the breakdown voltage reduction. Increasing the particle number from one to higher quantities raises the likelihood of a breakdown but does not significantly change the worst-case breakdown voltages. Based on a weakest-link distribution scaling, the influence of higher particle quantities is shown to be predictable. Thereby, the system-specific critical particle diameter falls in the range between $${15}$$ and $${54}\,\upmu \text {m}$$ . The analytical field distortion calculations demonstrate that the charge dominates the magnitude and spatial extent of the field enhancement. Based on this, a developed streamer integral-based model accounting for the measured charge magnitude, polarity, and particle position during the initiation is evaluated to determine if it can reproduce the experiments. This shows that the particle chargemagnitude cannot be directly mapped to the breakdown voltage reduction without considering the particle-to-electrode distance during the breakdown initiation. Additionally, the significance of the spatial position of the charge on the particle surface is demonstrated based on assumed distributions. Collectively, this provides statistical and charge-resolved breakdown data within a mechanistic framework that can serve as a baseline for quantifying the potential influence of particles on gas insulation systems