Introducing guard smart meters: voltage predictions and its implications for smart LV grid operation

With the advent of new loads and generation on the low voltage grid, voltage fluctuation has increased, especially in active distribution grids with a high penetration of distributed resources and a large deployment of electric vehicles. All this leads to greater uncertainty in future consumption patterns, giving greater importance to predictive models capable of adapting in real time to unexpected variations. Given the current measuring and communication infrastructure of smart meters in Spain, it is not feasible to request real-time data from all consumers. However, it is possible to communicate on-line with a few, which we call guard (or sentinel) smart meters. The results present (I) a novel methodology to select the sentinel meters and (II) their integration in predictive models based on neural networks, improving their ability to adapt and anticipate future states. For the analyses, it has been used (i) data from three groups of meters belonging to the living lab of a DSO, (ii) data from the head-end feeders of the transformer substation and (iii) data from the nearest weather station.