Research On Finite Volume Method For Numerical Computation Of Electric And Flow Field In SF₆ High Voltage Circuit Breaker

Finite Volume Method (FVM), is a new generation calculation method holding advantages of Finite Differential Method (FDM) and Finite Element Method (FEM). In this thesis, the research on FVM for numerical computation of electric and flow field have been proceeded.In electric field calculation: electric field in circuit breaker is described by cylinder Laplace equation. Curved quadrangle grids are generated in physical area by differential equation, and the Laplace equation is discrete by FVM, forming the numerical calculation model. Furthermore, in order to accelerate the computation speed, a novel computation method, named of improved spot iteration is applied for solving the numerical calculation model. Considering the complex structure of chamber in 550kV SF₆ high voltage circuit breaker, domain decomposition method is used on both mesh and numerical computation.In flow field calculation, the flow field problem in the SF₆ circuit breaker is a kind of flow with time-dependent, compressible, supersonic, non-liner, and the complex flow path. A set of differential equations named Navier-Stocker is used to describe this complicated problem. The numerical calculation models are set based on the theory of FVM. In order to get better computation precision, interlaced meshing method is applied to form different grid. And dispersed equations are educed by FVM based on interlaced mesh. Line-boxing iteration method is used to solve this numerical calculation model, which accelerate the computation speed a lot.At the end of this thesis, based on above theories about electric field and flow field, a series of programs are made: meshing program, electric field calculation program and flow field calculation program. Besides, 550kV SF₆ high voltage circuit breaker is taken as an example, numerical computation of electric field and flow field in the arc quenching chamber during the whole interrupting process are carried out. The result shows that the theories and computation method presented in this thesis are feasible and correct.