Simulation-Based Optimization of Peer-to-Peer Energy Management in Residential Households

This master's thesis presents a simulation-based approach to peer-to-peer (P2P) renewable energy integration, focusing on household-level energy management in Iceland. The study addresses the limitations of coordinating prosumers and consumers with shared battery storage while maintaining grid connectivity. It reviews existing P2P energy management systems, identifies gaps in the literature, and outlines a conceptual framework for the proposed model. User energy demand is analyzed using time series techniques, progressing from linear models on limited data to Temporal Convolutional Networks with feature engineering to improve demand prediction accuracy. A mathematical optimization model for energy exchange is developed, capturing intra- and inter-group trading scenarios and interactions with the grid and shared batteries for both prosumers and consumers. Using data from Icelandic households demonstrate that intra-group P2P trading significantly reduces energy costs, while inter-group trading provides additional, though marginal, savings. The study also evaluates the impact of prediction errors and battery depreciation costs, highlighting the importance of accurate forecasting and fair cost allocation. Jain's fairness index is used to evaluate the fairness of the model. The findings confirm the reduced total cost of the proposed model in optimizing distributed energy resources and provide actionable insights for implementing P2P renewable energy systems in practice.