Distributed optimization of energy profiles to improve photovoltaic self-consumption on a local energy community

Abstract

The development of local energy communities and collective self-consumption framework at a large scale requires new control methods that take into account users preferences. This article presents a model of such a community, with diverse actors (photovoltaic generators, electric vehicles, storage system and tertiary buildings). Game theory is used to model the preferences of each user and to build a mathematical framework where each user optimizes individually his power profile according to these preferences. An ADMM distributed algorithm (Alternating Direction of Method of Multipliers) is employed for practical implementation. Thus, a central agent is no longer needed to reach the system equilibrium, in which all users are satisfied while ensuring that the local energy community does not import more power from the grid than allowed. The simulations performed on real data for different scenarios representing diverse users behaviors show that the developed approach converges to a stable state, and leads to a maximization of local energy exchanges.