Urban-scale distributed power generation — Forecast methods for the estimation of electricity exchange profiles for grid-connected solar photovoltaic (PV) systems

Photovoltaic powered distributed generation (PV-DG) is increasingly being embedded into electricity distribution networks worldwide with fast penetration rates. Among various elements which can limit a substantial dissemination of the PV-DG, one is the paucity of reliable forecasting methods for estimating production and exchange profiles of non-programmable sources with particular reference to solar photovoltaic generators. In facts, forecasting PV power allows to optimize user´s energy management, for example scheduling electric loads and operating energy storages. Thus, the presented research work has been focused on the development of a methodology aimed at forecasting the electricity exchange of grid-connected PV systems up to 48 hours ahead. Different forecast techniques have been analyzed and, subsequently, a methodology based on physical models is proposed and tested on a real case study, represented by a residential nearly zero energy building (NZEB) equipped with an electric ground source heat pump (GSHP) and demand-side management solutions.