Gas-insulated HVDC technologies are generally of high interest for future grid expansion. So far, first recommendations to qualify such equipment are given in CIGRE TB 842 from JWG D1/B3.57. Based on that, TC17 WG5 is working on a new IEC standard 62271-5. An important role during qualification of gas-insulated HVDC equipment have superimposed voltage tests. Such test superimpose an impulse voltage at the device under test (DUT), while the DC voltage is still applied. Two possible coupling elements are knows to generate the desired waveshape: spark gap (SG) and coupling capacitor (CC). Both methods are commonly accepted for testing of gas-insulated HVDC equipment. However, literature reports significant different voltage waveshapes at superimposed tests with SG and CC. Up to now it is uncertain, if the different voltage shapes from SG and CC testing may result in different flashover voltages at the DUT. Because of that, CIGRE TB 842 recommends for SG test circuits, that the voltage shape should not be distorted in a certain time interval. But so far, there is no laboratory data, which proves the sufficiency or necessity of such specifications. The question arises, if this specification is sufficient or even may be neglected. The paper will present first laboratory data as input to the discussion. Withstand voltages of superimposed tests with coupling capacitor and spark gap are determined and compared with each other. The investigation is performed at a gas-insulated model arrangement for 300 kV DC and approx. 1000 kV LI. The model arrangement consists of a commercial HVDC insulator installed in a SF6/N2 filled gas chamber. The overall measuring procedure had to designed to provide reproducible data. Especially discharge effects at the tested insulator, such as traces and degradation, might influence the measurement. The developed procedure, as well as optimizations in the test circuit to reduce such influences will be presented in the paper. Each superimposed voltage measurement was initiated by a DC pre-stress of several days to ensure a DC steady state condition at the insulator surface and in the bulk material. During this period the insulator is also heated at earth potential to simulate high load conditions. LI withstand voltage tests with SG and CC are compared with each other to verify the differences in both methods. Summarized, the contribution will analyse, if the model arrangement reacts differently at superimposed tests with SG and CC, which can be used as input to discuss differences at commercial GIS and GIL.

Superimposed voltage test;spark gap;coupling capacitor