Title:

Computer Simulation and Visualisation of Complex Systems: Arcs and Hot Gas Flow in Autoexpansion Circuit Breakers

Autoexpansion circuit breakers (also known as the selfblast circuit breakers) are an
advanced generation of high voltage switchgear. This type of circuit breaker uses the
arc's energy to generate a high pressure SOurce in the expansion volume (also known
as the heating chamber) to create the required thermal and aerodynamic conditions
for interrupting the circuit at a current zero. Ablation of the arc confining nozzle at
high current is the main mechanism for pressurisation of the expansion volume. The
operation of such a breaker is extremely complex and its performance depends on
the whole arcing history as well as a number of geometrical factors characterising
the geometry of the interrupter. It is a complex system with interrelated mechanical
parts (moving pistons, valves, ablating nozzles, and moving contact) and physical
processes (radiation, thermodynamics, aerodynamics, turbulence and ablation of solid
materials). The relationship between all the constituent parts and physical processes
gives rise to the collective behaviour of the whole system.
Computer simulation of the arcing process in such a breaker has been carried out
in the present work. Because of the very high power level encountered in such a
breaker the arc interacts intensely with its surroundings. The objective of the work
is to establish a computer model to simulate the whole arcing process, validate the
model and then perform an extensive analysis of system behaviour to extract useful
information for the optimisation design of such devices.
The history of circuit breaker development, fundamental aspects related to the operation
of high voltage circuit breakers, and history of computer modelling of switching
arcs are first reviewed in chapter I, which provides an overall background picture
for the present work. The mathematical description of the important physical processes
is then given in chapter 2 which includes the governing equations for arc flow,
the modelling of radiation and turbulence, the calculation of nozzle ablation, and the computation of electrical and magnetic fields for Lorentz force and Ohmic heating.
The temperature and pressure encountered in computer simulation of the autoexpansion
circuit breaker arc covers a wide range, from 300K to possibly 40, OOOK and
from atmospheric pressure to 100 bar. The material and transport properties of the
mixture of the working gas and ablated nozzle vapour are highly nonlinear functions
of plasma parameters. Thus a robust computational fluid dynamics (CFO) solver is
essential. In the present work, a commercial CFO package, PHOENICS, is used for
the simulation. The practically important issues, such as the implementation of the
arc model with input of material properties into the solver, the specification of initial
and boundary conditions, the approximate of the circuit breaker geometry, and choice
of time step and control of convergence, are discussed in chapter 3.
In the operation of an ABB autoexpansion circuit breaker, there are a number of
mechanical parts that move with time during an operation process. The operation of
over pressure valves, with one of them attached to the moving piston, has to be correctly
modelled. This is detailed in chapter 4 where validation of the numerical meth~
'
ods is provided by comparing the prediction with analytical results from isentropic
compression and also with measurement from ABB. Results show that the proposed
numerical scheme can satisfactorily model the valve operation and the piston movement.
Typical results of the gas flow in such a circuit breaker without the presence of
an arc (noload operation) are presented and discussed.
In chapter 5 the operation of the ABB breaker under specified arcing current is
then simulated for almost a whole arcing period. Results indicate that Lorentz force
has a profound effect on the flow field as well as the arc shape. Detailed energy and
mass balance calculation is performed for the arcing space and also for the expansion
volume, which clearly shows the importance of radiation transfer, convection at different
nozzle exits and the change of energy and mass storage at different instants in
the arcing process. It is also shown that the pressurisation of the expansion volume
is due to the influx of thermal energy, not the mass influx. The predicted arc voltage
overally agrees with the test results within 15% for all three cases simulated with different
breaker geometry. The predicted pressure at current zero is within 10% of the
test results. On the whole the prediction is considered satisfactory in consideration of
the approximations that have been introduced in the geometry and radiation model. It has been found that for the autoexpansion circuit breaker the pressure in the
arcing space can fluctuate rapidly in the period shortly before the thermal recovery
period. Pressure fluctuation with several bars around the current zero period results
in a scatter of thermal interruption and dielectric recovery performances. Large pressure
variation is therefore not desirable. Optimisation of design parameters is necessary
in order to avoid pressure variation and to ensure maximum pressure and lowest
temperature possible in the arcing voll!me. A systematic study of the mechanisms responsible
for the pressure fluctuation is therefore carried out in chapter 6. It has been
found that the evolution of pressure and temperature fields in the arcing space around
current zero depend on the supply rate of gas from the expansion volume and the exhaustion
rate at the nozzle exits. Thus, an optimum design is directly linked with the
design of the expansion volume and the link channel between the arcing space and the
expansion volume. A systematic study of the influence of various design parameters is
also carried out to identify the most influencing parameter, which is the dimension of
the channel link. Based on the knowledge and understanding derived from this study
a new design has been simulated which produces very promising result in smoothing
the pressure fluctuation in the arcing space. Pressure and temperature fields at current
zero depend on the whole arcing history as well as the contact movement which
determines the gas exhaust passage. Arcing processes with different arcing time (altogether
three cases with different arcing times) are finally performed to assess the
efficiency of the new design. In all cases it has been shown that with the addition of
a buffer volume the pressure smoothly changes in the period approaching the final
current zero.
In summary, the three objectives stated in chapter 1 have all been achieved by the
work presented in chapters 2 to 6. Nevertheless, there are still several aspects of the
model that need to be improved. This is discussed in the final chapter.
