The Cost of Ignoring Uncertainty: How Stochastic Optimization Enhances Modelling of Energy Communities' Outcomes

Mariana Mira Guedes

Key information

Authors:

Mariana Mira Guedes (Mariana Mira Guedes)

Supervisors:

Ian James Scott; Diana Pereira Neves (Diana Pereira Neves)

Published in

June 24, 2025

Abstract

The increasing complexity of modern energy markets, driven by the transition to renewable energy sources and the rise of decentralized energy production, requires innovative approaches to modelling and optimizing energy communities. This thesis aims to investigate how uncertainties in load, solar generation, and electricity prices influence the outputs of peer-to-peer (P2P) energy communities. A centralized optimization model for a P2P energy community was developed, capturing the interactions between multiple consumers, prosumers, and shared storage systems within a community-based trading network. Using real-world data from Portugal, five optimization approaches were implemented, accounting for uncertainty and attitude towards risk: deterministic optimization with a single average scenario and with multiple scenarios, stochastic optimization, least regret, and robust optimization. The findings demonstrate that incorporating uncertainty into optimization models significantly enhances operational robustness and economic performance. Stochastic methods provide greater confidence in consistently attaining higher positive welfare outcomes, particularly in high-uncertainty scenarios. Expected welfare gains from stochastic methods can reach approximately 6 times higher than deterministic approaches in winter scenarios, underscoring the essential role of uncertainty modelling for resilient energy communities. Each method exhibited distinct trade-offs between performance and reliability, offering practical insights for selecting optimization strategies tailored to the specific objectives of each energy community. The results provide evidence-based guidance for energy community modelling and policymakers, emphasizing the critical importance of uncertainty modelling for sustainable and economically efficient energy communities.