Representative feeders for spatial scaling of stochastic PV hosting capacity

Abstract

The photovoltaics (PV) hosting capacity (HC) of the power system infrastructure is an important planning problem. The energy transition is happening now, resulting in the addition of new load types and generation in the low voltage distribution network. With these new loads and generation devices connected to the distribution network, a new planning approach is required that considers their stochastic nature. Computing the stochastic HC for all individual low voltage distribution feeders is challenging, as a small service area can have hundreds of feeders. This work aims to capture appropriate clustering schemes and the most relevant features of low voltage distribution feeders so that an accurate estimate of the hosting capacity of the full service area can be calculated by scaling up the hosting capacity of a small number of representative feeders. A case study using actual feeders from suburban Spain showed that representative feeders obtained from feature reduction and using appropriate cluster size enables to scale the stochastic PV HC of an extensive service area. The case study showed that using 3% of the total feeders enables to estimate the PV HC of the LVDS feeders in the large service area.