Abstract

Quality control, commissioning and diagnostic tests on insulation systems of electrical apparatus are necessary to reduce the failure risk under operation, eliminate infant mortality failures and reduce the impact of ageing on insulation system reliability. This study tries to establish a path which, based on the statistical–phenomenological modelling of the hazard rate, or bathtub curve, drives in the choice of test overvoltage and procedures. The choice of the test voltage and its waveform should be made carefully to improve reliability (reduce failure hazard) in any section of the bathtub curve. This may require, however, different choices depending on the section which corresponds to the planned test. For example, large overvoltages could be acceptable to remove the infant mortality to ensure the highest reliability of newly commissioned apparatus, but they can incept highly energetic, and, thus, fast degradation mechanisms on apparatus operating on the field for a period of time, that can worsen, rather than improving, the insulation system reliability. Resorting to partial discharge measurement during any overvoltage test seems to be the optimal solution to find the most appropriate overvoltage for any type of testing need, leveraging the Latin concept that ‘in medio stat virtus’.