Partial Discharge Analysis and Monitoring in HVDC Gas Insulated Substations Roland Piccin, Armando Rodrigo Mor, Peter Morshuis Delft University of Technology Delft, The Netherlands Gian Carlo Montanari University of Bologna and Techimp SpA Bologna, Italy Alain Girodet Alstom Grid Villeurbanne, France Abstract— Partial discharge (PD) detection is a widely used tool for monitoring and diagnostics of high voltage gas insulated substations (GIS). Despite the fast growing demand of high voltage direct current (HVDC) interconnections, little research has been dedicated to PD monitoring systems for HVDC GIS. In this paper, conventional and unconventional PD detection systems are employed for the detection of three different defects common in GIS, and relevant PD analysis methods are presented. Finally, the analysis methods and detection systems are compared to point out their potential and limitations in monitoring PD in HVDC GIS. Keywords—partial discharges, IEC60270, UHF, GIS, HVDC. I. INTRODUCTION Nowadays, gas insulated substations (GIS) are vital nodes of the transmission network. The use of SF 6 as insulating medium permits compact dimensions, but at the same time high reliability is guaranteed. Nonetheless, the strict requirements of safety in operation and continuous power supply require the development of tools to keep the reliability at the required level and of appropriate maintenance strategies for GIS. Indeed, reliability can be affected by defects, such as undesired metal particles, which generate partial discharges (PD). PD are responsible for harmful degradation mechanisms, and eventually failure, of GIS. To mention only few of them, a free moving particle approaching the conductor may trigger a flashover, and if it lies on a spacer, can lead to the local carbonization of the latter. Although PD monitoring is recognized as a fundamental tool for alternate- current (AC) GIS diagnostics, it has been little investigated for direct-current (DC) applications. In fact, the growing demand of high-voltage direct-current (HVDC) transmission brings up the issue of maintainability of HVDC apparatus. Regardless the voltage waveform under consideration, either AC or DC, a PD monitoring system is composed by two subsystems, that is, hardware and software. The set of sensors, splitters, coaxial cables and detection unit constitutes the hardware subsystem. The detection subsystem communicates unidirectionally with the software subsystem which takes care of data, feature extraction and classification. The hardware subsystem would not change considerably whether PD occurs under AC or DC, since the electromagnetic signal is of the same nature, but the software must be modified considerably, due to the different phenomenology of PD under AC or DC. This paper presents the results of a research project carried out in the HV Laboratory of Delft University of Technology which aims at investigating the applicability of ultra-high frequency (UHF) and wide band (WB) methods to the detection of PD under HVDC. In particular, the focus is on the PD analysis strategies of extracting valuable information from the data in order to recognize the type of defect causing PD and reject the noise. II. EXPERIMENTAL SET UP A. Test object and defects Most incipient dielectric failures in GIS are ascribable to defects that are detectable by PD diagnostic. Among those, this study investigates: a protrusion at the HV conductor (HV protrusion), a protrusion at the enclosure (LV protrusion) and a free moving particle on the enclosure surface. These defects are placed in a section of a 380 kV single phase test-GIS. The outer/inner radius ratio is 150/35 mm and the length of the aluminum conductor is 510 mm. The GIS has a dielectric window where an internal UHF antenna is installed. The protrusions are tested in the GIS first filled with humid air at 100 kPa and then with SF 6 at 100 kPa, while for the free moving particle the GIS is filled with CO 2 at 400 kPa. 488 2014 Electrical Insulation Conference, Philadelphia, Pennsylvania, USA, 8 to 11 June 2014 978-1-4799-2789-0/14/$31.00 ©2014 IEEE