Renewable Energy Integration through Coalition Formation for P2P Energy Trading

Renewable energy sources are highly unreliable; hence prosumers connected to renewable energy sources find unreliable energy surplus and demands which should be managed frequently within neighbourhoods. Peer-to-peer(P2P) energy trading has emerged as the next-generation energy management mechanism to manage unreliability concerns within local communities. This process needs a prosumer with energy demand to join with a set of prosumers with energy surpluses to fulfil the energy need. Finding the optimal set of prosumers with energy surpluses is a challenge that is not yet duly addressed in literature considering the unreliability of energy sources and constraints in local communities. This paper presents a game-theoretic approach, applies coalition formation game theory supported by a clustering-based approach and solves an optimisation problem to find optimal coalitions. The optimal coalitions are the winning coalitions found from our coalition formation game. This game considers frequent change of energy surpluses and demands, distance to prosumers, the quantity of energy sale, and dynamic clustering on potential coalitions. The payoff will calculate the quantity of energy sale. We used a multi-threading and parallel computing approach to find the optimal size of the cluster that gives winning coalitions. We implemented our clustering-based coalition formation algorithm in Java. We tested for the efficiency and success rate using a residential PV energy production and consumption data set from California state in the USA. We compared the number of successful coalitions resulted from our algorithm with the hierarchical coalition formation algorithm(HCF) in literature. Our experimental results showed that both the sizes of the neighbourhood and sizes of the cluster have a significant impact on the number of successful coalitions and the proposed algorithm has a higher success rate compared to the HCF algorithm.