Computational framework for studying charge transport in high-voltage gas-insulated systems

Licentiate thesis, 2015

Increasing demands in electric energy stimulates a shift in functional requirements to power distribution networks and pushes equipment manufacturers to develop highly optimized products where high rated parameters (e.g., energy, voltages) are combined with reduced geometrical dimensions of components. Thus, newly developed gas-insulated switchgears (GIS) for distribution networks operate at enhanced electric fields, which should be withstood by the insulation. In this case, new additional challenges are imposed due to strong requirements concerning environmental safety that call for replacement of SF6-gas used as insulating medium in GIS by more environmentally friendly substitutes. Ideally, future “green” insulation is to be based on natural gases (e.g. synthetic air), which, however, are characterized by much lower electrical performance and thus need to be strengthened, e.g., by using solid insulating elements. To provide reliable design criteria for hybrid gas-solid insulation, there is a need in detailed understanding of basic physical phenomena taking place due to its exposure to strong electric fields. These include initiation and propagation of electrical discharges in gas phase, interactions of produced gas discharge plasma with solid dielectric surfaces, charge transport though solid material, etc. The research presented in the thesis focuses on analyzing field induced phenomena in gas and aims at developing computer simulation framework incorporating physical processes associated with the motion of charged species in gas medium under electric fields. The model of electrical discharges in air developed in the present study comprises a set of coupled non-linear partial differential equations (PDE) describing time dependent drift and diffusion of charge carriers associated with reactions between them (ionization, recombination, etc.) ; dynamics of electric fields in discharge volume affected by accumulated space charges; and intensity of photo-ionization in gas. The reactions rate coefficients for the charge transport equations are derived from numerical solution of Boltzmann’s equation for electrons energy distribution function in N2:O2 (80:20) mixture. The set of PDEs is solved numerically utilizing commercial software based on finite element method. Numerical challenges and details of the implementation are discussed in the thesis; in particular, logarithmic formulation of transport equations, approaches for numerical stabilization of the solution, adaptive mesh refinement, etc. The developed computational framework is utilized for analyzing several study cases including propagation of streamer discharges in air in different 2d geometrical arrangements and streamer branching in fully 3d representation. The obtained results are discussed and compared with results of experimental and theoretical studies available in the literature.