The report deals with the activities carried out to evaluate the possibility of using the Power Quality data recorded by Italian monitoring systems in an “alternative way” for diagnostic and prognostic purposes for the components of the electricity networks. First, a summary was made of the type of problems that prognostics must resolve by field work.

For transmission and distribution networks. Starting from this reference framework, a bibliographic investigation was conducted on all the artificial intelligence techniques that have been applied to Power Quality data to try to promptly identify, before the definitive intervention of the protections, any faults (the so-called “incipient faults”) to the network components and possibly locate them.

The proposed techniques range from pure Signal Processing to those techniques that combine Machine and Deep Learning models with it (mixed techniques) up to Fuzzy Logic or purely statistical methods. The evaluation of the possibility of applying the techniques proposed in the literature to the voltage waveforms recorded by the Italian monitoring systems was then made.

Most of the proposed techniques cannot be applied to the monitoring system available to RSE, due to the present unavailability of the current waveforms. On the other end, the techniques involving exclusively voltage waveforms need some improvements to traditional monitoring systems (sampling frequency, recording time etc.).

The Report is available on the Italian site