Comparison of time-dependent variation of surface charge accumulation in SF<inf>6</inf>/N<inf>2</inf> and pure SF<inf>6</inf>
Han Wang,Jian-Yi Xue,Bao-Hong Guo,Jun-Bo Deng,Guan-Jun Zhang
DOI: https://doi.org/10.1109/icpadm.2018.8401232
2018-01-01
Abstract:SF
<sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">6</sub>
/N
<sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub>
Mixture has good insulation properties and reduces the use of SF
<sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">6</sub>
, thus it has drawn more and more attention. In addition, Surface charge accumulation on insulating materials is a major factor that greatly distorts the electrical field and reduces the insulation level. Therefore, it is necessary to research the characteristics of surface charge accumulation in SF
<sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">6</sub>
/N
<sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub>
mixture. In this paper, the difference between the time-dependent variation of surface accumulated charge of spacer in 5atm pure SF
<sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">6</sub>
and that in 5atm 20% SF
<sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">6</sub>
/N
<sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub>
is researched. The downsized truncated cone-type spacer made of 70% Al
<sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub>
O
<sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sub>
filled epoxy resin and coaxial cylinder electrodes are used to model a GIL. In the experiment, ±30kV and ±40kV DC voltage are applied to the central electrode for 30-1080 min, then the surface potential of spacer is measured by a Kelvin vibrating electrostatic probe, and the corresponding surface charge density is further calculated by a calibration method. The results reveal that in both gases, the surface accumulated charge exhibits a time-dependent variation characteristic that the local fluctuates and the whole tends to saturate. Under the DC voltage of same amplitude and same applied time, the total amount of surface charge in 20% SF
<sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">6</sub>
/N
<sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub>
is far more than that in pure SF
<sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">6</sub>
. However, the maximum of surface charge density in 20% SF
<sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">6</sub>
/N
<sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub>
is not much different from that in pure SF
<sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">6</sub>
under the same condition of DC voltage. Moreover, in pure SF
<sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">6</sub>
, the total amount of surface charge under negative DC voltage is far more than that under positive DC voltage, which means there is polar effect in pure SF
<sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">6</sub>
. However, in 20% SF
<sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">6</sub>
/N
<sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub>
the polar effect does not exist.