Evaluation of breakdown test procedures for self-restoring insulation types

Summary

For the evaluation of the breakdown strength of self-restoring insulation it is necessary to perform a large number of breakdown experiments. For example, for the development of a mathematical model, with which the breakdown voltage of GIS can be predicted, a large amount of verification breakdown tests is required. The investigation into the breakdown strength of alternative gases also requires a relatively large amount of breakdown tests. In most cases, one of the breakdown test procedures from IEC standard 60060-1, procedure D, is selected. The available procedures in this standard (class 1, class 2 or class 3 tests) differ significantly in time usage. Furthermore, these procedures are designed in such a way that the 50% resp. 10% breakdown voltage or breakdown strength can be obtained from the test results by means of a simple calculation which can be performed with a simple calculator. Complex computational algorithms are thus not required. Because of this, separate tests are required to find the 10% breakdown voltage or the 50% breakdown voltage. Especially the 10% breakdown voltage requires a relatively lengthy breakdown test. To increase the efficiency of research related to the breakdown voltage or breakdown strength of gas insulation systems or other self-restoring insulation types it would be favourable to find a fast but accurate breakdown test procedure. Therefore, in this paper the available breakdown test procedures are evaluated with respect to the required testing time and the accuracy of the test results for a case study of a rod-plane electrode configuration in dry air. For the evaluation of the test procedures, the 50% resp. 10% breakdown voltage is obtained with the application of statistical distribution fitting software, such as Weibull++. Our evaluation has shown that the obtained 50% resp. 10% breakdown voltage, including the accuracy, is comparable among the different test procedures when the test results are fitted to a suitable statistical distribution. Therefore, we conclude that it is not necessary to apply the separate test procedures for the 50% resp. 10% breakdown voltage when statistical distribution fitting is applied. Moreover, with statistical distribution fitting software it is possible to select the fastest test procedure while maintaining the same accuracy in the determination of the 50% resp. 10% breakdown voltage resulting in a significant reduction of the required testing time.

Additional informations

Authors

YAMAMOTO

Keywords

GIS, self-restoring insulation, test procedure, breakdown voltage, gas insulation