Streamer inception mechanisms in air - Polarity and pressure dependence in strongly non-uniform fields

Gas insulation is used in several different components in the power grid, and SF₆ is the most commonly used insulation gas. Due to its high global warming potential, the EU will ban the use of SF₆. Air-based insulation systems are one of the alternatives, especially for medium voltage switchgear. To ensure a safe transition away from SF₆, it is important to have a good understanding of the breakdown and pre-breakdown phenomena in air and other SF₆-alternatives. The aim of this work has been to study streamer inception for non-uniform fields in alternative gas mixtures for SF₆, focusing on air. Short (≈ 50mm) needle-plane gaps were used throughout. High-speed imaging, photomultipliers, and a current sensor were used to detect streamer inception. The experimental results were compared with simulations. Streamer inception in a non-uniform field in air was found to be polarity dependent, with positive streamers having a higher inception probability than negative streamers at 1 bar (absolute pressure) when the gap was stressed by an impulse voltage. When a DC step voltage was used, the polarity dependence was observed in the time lags instead, with positive streamers having shorter time lags than negative streamers at 1 bar. At 3 bar, the polarity dependence of the time lags was reversed, and it was negligible at 5 bar. The polarity dependence of the streamer inception probability when using the impulse voltage was found to be because detachment from negative ions was the main source to start electrons for both polarities in that setup. For negative streamers, the negative ions were electrostatically removed from the inception region which resulted in this polarity requiring higher voltages to incept. For the time lag experiments, the simulations indicate that detachment from negative ions was the main source to free electrons for positive polarity. For negative polarity, the simulations indicate that neither cold field emission nor detachment inside the inception regionwas the dominating electron source. This indicates that the simulation model was missing the description for the dominating electron source for negative streamers under those conditions. It is speculated that this missing description might be related to secondary electron emission, but more research is needed to verify this. The results presented in this thesis contribute to a better understanding of streamer inception in air.