Statistical Characterization of Vacuum Surface Flashover for Positive 45-Degree Geometries

Journal Title
Journal ISSN
Volume Title

Large-scale pulsed power systems often cannot operate at their maximum potential due to failure points throughout the system. One of the first points of failure is when the electric stress between an anode and cathode separated by an insulator at a vacuum interface becomes too high, and vacuum surface flashover occurs. This flashover shorts the anode to the cathode and limits the maximum energy transfer to the intended load. Flashover can initiate from either the cathode or the anode, depending on the angle of the insulator. For +45° degree insulators seen in many pulsed power machines, anode initiation is believed to be more common than cathode initiation. A moderate scale test stand is used to statistically characterize flashover on +45° insulator geometries. Two different cross-linked polystyrene (Rexolite) insulator geometries are characterized using the system. The first is a wedge geometry that separates a hemispherical anode from a planar cathode by a distance of 2.5 mm (Testbed I). The second is a frustum-style insulator that separates a flat cathode from two distinct anode geometries (Testbed II and Testbed II.b). An 8-stage Marx generator capable of delivering voltages up to 264 kV is the voltage source for the system. Voltage and current measurements are collected using capacitive voltage dividers (CVD) and a current viewing resistor (CVR). Flashover events for Testbed I are characterized by flashover type (anode or cathode-initiated) using time-integrated and post-test imaging and peak voltage just before flashover occurs. Tree-like branching structures observed from time-integrated and post-test imaging are used to characterize a flashover event as anode-initiated. The observed branching structure is consistent with Anderson’s description of anode-initiated flashover. Testbed II and II.b are characterized by peak voltage just before the flashover event and the effective time it takes for the flashover to occur (teff ). It has been shown statistically that the different anode geometries used on Testbed II and II.b with the same size insulator produce different values of teff.

Embargo status: Restricted until 09/2173. To request the author grant access, click on the PDF link to the left.

Flashover, Vacuum Breakdown