Turbulent switching arcs and influence of design parameters on performance of auto-expansion circuit breakers

Turbulence enhanced momentum and energy transport is an important mechanism in shaping the flow and temperature field of low current turbulent SF6 (Sulfur-hexafluoride) arcs burning in a nozzle. It is a factor that largely determines the prediction accuracy of the interruption capability of high voltage circuit breakers. Previous study [81] has shown that for SF6 arcs the parameters in the Prandtl mixing length model and the two equation k-ε model must be adjusted to bring agreement between prediction and measurement. The experimental observation [109] of steep temperature or density gradient exists at the arc edge, where turbulence is usually strongest. The density gradient can be regarded as a result of temperature gradient when the pressure difference across the low current arc column is small. In previous studies of turbulence models, the effects of large density or temperature gradient have not been considered. In the present work the k-ε model is modified to take into account the effect of the presence of steep local temperature gradient. The model is first applied to the steady state SF6 nozzle arcs in the current range from 100 A to 1800 A. The performance of the modified turbulence model is assessed by a comparison of the predicted and measured radial temperature profiles at different currents and its behaviours with another two most commonly used turbulence models, the Prandtl mixing length model and the k-ε model. The relevant turbulence parameters of the Prandtl mixing length model are adjusted according to the different nozzle shapes with different values and it has been found that its applicability is limited. The modified k-ε model, which is modified to take into account the effect of large temperature gradient with all default coefficients, can make reasonable prediction for turbulent arcs in the Aachen nozzle [67, 108] under direct current conditions. The model is then applied to the transient nozzle arcs in a GE nozzle [86] and the Campbell nozzle [85]. Finally, a real puffer-type circuit breaker of 252 kV has been used to verify this model. The predicted arc voltage and pressure agree reasonably well with the measurements at both high and low current levels, justifying the applicability of the modified k-ε turbulence model. The auto-expansion circuit breaker is a relatively new interruption technique, which creates the required fast gas flow at current zero using the energy dissipated by the arc itself. A design study has been carried out for a prototype 245 kV auto-expansion circuit breaker at 50 Hz, in order to investigate the influence of key design parameters in terms of arcing conditions at the current zero phase and the critical RRRV (Rate of Rise of Recovery Voltage). Auxiliary nozzle with different lengths, and the severity of leakage from the expansion volume have been implemented for testing the influence of design parameters in an auto-expansion circuit breaker. PC-based arc modelling taking account the effect of nozzle ablation and the mixing process of PTFE vapour with SF6 has been made for auxiliary nozzle investigation. For the effect of leakage, various sizes of a leakage hole are specified on the valve of the expansion volume. The critical RRRV values for the Ref. Case and the its modifications are used for comparison of its interruption capability in order to estimate the influence of these key design parameters. The solution of all of the equations in the arc models is based on a commercial computational fluid dynamics package, PHOENICS. PHOENICS has been extensively used at the University of Liverpool to model the arc behaviours since 1992. Before using PHOENICS to simulate the arc behaviour in this thesis some conditions need to be resolved. These are related with the choice of turbulence model, the computational domain size for electric field, and the influence of the Lorentz force. The differential models for reported works in the thesis is implemented into PHOENICS version 3.6.1. All boundary conditions which are treated as sources terms are discussed in Section 2.