Modelling and design optimisation of renewable energy communities to support the energy transition aspirations of rural and islanded populations

Abstract

The market for distributed Renewable Energy Systems has increased considerably in recent decades, driven by the necessity for a reduction in global carbon emissions in an effort to combat climate change. While the focus on decarbonising the energy sector in Europe has been successful in recent years, it has disproportionately benefited urban population centres. Those who live in built-up environments will likely have better access to newer, greener technology, with islanded communities often relying on a weaker grid with fossil fuel reliant infrastructure. These communities are therefore at high risk of being left behind in the energy transition towards net-zero emissions. This study presents a novel solution to the problem of decarbonising remote, islanded populations by means of Renewable Energy Communities (RECs). The test location, Formentera, was chosen due to its unique set of challenges and opportunities regarding energy security and access to clean energy. A generalised, modular model was developed in Python, allowing the integration of generation (wind and solar), storage (battery and hydrogen), and real-world data from the test location. The model simulates the dynamic dispatch of the system over hourly increments to evaluate the annual performance. The system is optimised using the Non-dominated Sorting Genetic Algorithm (NSGA-II), which identified an inherent trade-off relationship between cost reduction and decarbonisation of the REC. Results show that the deployment in the case study location can deliver improvements in both cost and emissions relative to a grid-only scenario. A comparison of storage configurations shows a considerable benefit to co-locating batteries and a regenerative hydrogen storage system due to the latter’s ability to act as a seasonal storage buffer. Findings suggest that a ’friendly’ local trading policy outperforms a market-based regime on cost savings, and ensures better energy equity between members. The analysis incorporates Monte Carlo simulations of estimated assumption ranges and a variance-based Sobol sensitivity analysis. These methods reveal the range of variability in the result arising from uncertainty in the input assumptions, including those which most impact performance, thus identifying high-risk areas for project monitoring and intervention. These can not only support the design stage of the REC but also contribute to risk-aware planning and policy development. The model’s development in Python allows for a scalable foundation on which future research can be built, and contribute to the commercialisation of an REC-focused planning tool. The outcome of this work provides a novel, quantitative guide for energy developers, government entities, and network operators on REC development. The model framework can be used to trade-off system cost and emissions reduction, design for and navigate potential future energy policy, assess energy equity, and ensure a clearer route to realising the net-zero aspirations of rural, islanded communities.