Wind Farm Electrical Asset Components: Insulation Diagnostics through Partial Discharge Measurements

The reliability of asset components of wind and photovoltaic plants has been an issue, especially regarding premature failures of electrical insulation systems. Extensive testing of cables is often carried out, generally off-line, to assess partial discharge activity and level of dissipation factor, and this occurs, to lower extent, also other components, as generators and transformers. However, they may be not effective in predicting health conditions, which are related to the dynamic of operation, the real electrothermal stresses acting with load variations and abnormal electrical stresses associated to the often non-sinusoidal power supply. Another issue about diagnostic measurements is that they require often data processing and interpretation made by experts, which does not favor, for a matter of cost and timing, the diffusion of on-line monitoring systems. This paper presents an approach to partial discharge monitoring addressed towards an automated and unsupervised technology, able to provide traffic light alerts to maintenance manager upon reaching critical conditions for wind-farm asset components. Examples of applications to rotating machines controlled by power electronics, transformers, switchgears and cables are presented, illustrating how the new automatic approach works for noise and partial discharge recognition, noise rejection, partial discharge harmfulness identification and dynamic health index build up.