Search

Thanks to their modularity and their capacity to adapt to different contexts, hybrid microgrids are a promising solution to decrease greenhouse gas emissions worldwide. To properly assess their impact in different settings at country or cross-country level, microgrids must be designed for each particular situation, which leads to computationally intractable problems. To tackle this issue, a methodology is proposed to create surrogate models using machine learning techniques and a database of microgrids. The selected regression model is based on Gaussian Processes and allows to drastically decrease the computation time relative to the optimal deployment of the technology. The results indicate that the proposed methodology can accurately predict key optimization variables for the design of the microgrid system. The regression models are especially well suited to estimate the net present cost and the levelized cost of electricity (R2 = 0.99 and 0.98). Their accuracy is lower when predicting internal system variables such as installed capacities of PV and batteries (R2 = 0.92 and 0.86). A least-cost path towards 100% electrification coverage for the Bolivian lowlands mid-size communities is finally computed, demonstrating the usability and computational efficiency of the proposed framework.

Energy system models for off-grid systems usually tend to focus solely on the provision of electricity for powering simple appliances, thus neglecting more energy-intensive and critical needs, such as water heating. The adoption of a Multi-Energy System (MES) perspective would allow us not only to provide comprehensive solutions addressing all types of energy demand, but also to exploit synergies between the electric and thermal sectors. To this end, we expand an existing open-source micro-grid optimization model with a complementary thermal model. Results show how the latter achieves optimal solutions that are otherwise restricted, allowing for a reduction in the Levelized Cost of Energy (LCOE) of 59% compared to a conventional microgrid, and an increase of reliance on renewable sources of 70%.

Hybrid microgrids represent a cost-effective and viable option to ensure access to energy in rural areas located far from the main grid. Nonetheless, the sizing of rural microgrids is complicated by the lack of models capable of accounting for the evolution of the energy demand over time, which is likely to occur in such contexts as a result of the modification of users' lifestyles. To tackle this issue, the present study aims at developing a novel, long-term optimisation model formulation, capable of accounting for load evolution and performing suitable investment decisions for capacity expansion along the time horizon. Multiple scenarios of load evolution are considered to evaluate the beneficial effects of this novel approach, through the coupling of the model with a tool for stochastic load profiles generation. The results show how this implementation brings lower Net Present Cost to the project and improved correspondence between actual electricity demand and microgrid sizing. Finally, a sensitivity analysis evaluates the robustness of the approach with respect to input data variability and the Loss of Load parameter.

Robust sizing of rural micro-grids is hindered by uncertainty associated with the expected load demand and its potential evolution over time. This study couples a stochastic load generation model with a two-stage stochastic micro-grid sizing model to take into account multiple probabilistic load scenarios within a single optimisation problem. As a result, the stochastic-optimal sizing of the system ensures an increased robustness to shocks in the expected load compared to a best-case (lowest-demand) sizing, though with a lower cost and better dispatch flexibility compared to a worst-case (highest-demand) sizing. What is more, allowing just a 1% unmet demand enables to significantly improve the cost-competitiveness and the renewables penetration as all the not supplied energy is located in a negligible fraction of the unlikeliest highest demand scenarios.