The breakdown mechanism of compressed SF6 in gas insulation is known to be controlled by stepped leader propagation. This process is reasonably well understood for strongly non-uniform insulation gaps (‘point-to-plane’) and in the absence of pre-breakdown discharge activity (‘corona stabilization’). Open questions still remain for weakly non-uniform insulation gaps with small electrode protrusions (particles, surface roughness), in which pre-breakdown partial discharge (PD) activity is present. This paper presents a first attempt to derive a consistent picture under these conditions, which are characteristic for practical gas insulation systems. Experiments were carried out in a uniform field gap with a short protrusion on one electrode. This configuration was studied at various pressures from 0.1 to 0.5 MPa and both polarities using electrical and optical diagnostics. The results are interpreted using a quantitative model and order-of-magnitude estimates. The emerging picture allows prediction of most of the technically relevant aspects of the discharge processes and their main parameter dependences. It comprises statistical time lags, formative time lags including pre-breakdown PD activity and breakdown fields as a function of gas pressure, protrusion length and polarity.

No abstract

The breakdown mechanism of compressed SF6 in gas insulation is known to be controlled by stepped leader propagation. This process is still not well understood in uniform and weakly non-uniform background fields with small electrode protrusions, such as particles or surface roughness. In a previous publication an investigation of partial discharges and breakdown in uniform background fields that focused on streamer and leader inception mechanisms was presented (Seeger et al 2008 J. Phys. D: Appl. Phys. 41 185204). In this paper we present for the first time a physical leader propagation model that consistently describes the observed phenomena in uniform background fields in SF6. The model explains two different types of leader breakdown; these can be associated with the precursor and the stem mechanisms. It also yields the parameters of stepped leader propagation, which include step lengths, associated step charges, step times and fields and temperatures in the leader channel. Further, it explains the features of arrested leaders in uniform background fields. The model predicts the range of parameters under which arrested and breakdown leaders occur in good agreement with the experimental data.

No abstract

No abstract

The average surface ablation of poly-tetra-fluro-ethylene (PTFE) nozzles in high voltage circuit-breakers was investigated experimentally by means of the specific ablation. Experiments were done with a test device in the current and arcing time range of 2–40 kArms and 5–17 ms, respectively. This allowed determining the dependences of the specific PTFE ablation on current and arcing time. Additionally, the specific PTFE nozzle ablation in commercial high voltage circuit-breakers was analysed in a broad range of current amplitudes and arcing times. For understanding the results more in detail, experiments with a small scale test device were performed. The different experiments yielded consistent values for the PTFE ablation. The ablation depends mainly on the total arc energy. The total ablation is given by the ablation of walls surrounding the arc and by the ablation in regions where plasma from the arc cools down. Different specific ablation values can be observed for the axial blown arc mode and the ablation controlled arc mode.

No abstract

A simple integral arc model for ablation arcs in tubes is presented. The model predicts the arc temperature in the axis, the average electrical field strength, the pressure generation and the mass ablation rate at the tube wall. The model is entirely based on first principles and does not require fit parameters. This is achieved with the help of detailed numerical arc simulations which use a combination of a radiative energy transfer code and computational fluid dynamics. This allows a physically correct and consistent averaging over the arc when deriving the integral model. New experiments for validation of the model were done. The results of the model are in satisfactory agreement with the older and new experimental data. The dependences of the predicted arc parameters are discussed for tube radii and lengths in the parameter range R = 2-8 mm and L = 40-80 mm, respectively.