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 (R 2 = 0.99 and 0.98). Their accuracy is lower when predicting internal system variables such as installed capacities of PV and batteries (R 2 = 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. • A novel methodology to create surrogate models for energy planning is presented. • Gaussian regression is compared to multi-variable regression. • The surrogate models for Economic indicators have a high prediction accuracy. • PV and Batteries installed capacity have lower prediction accuracy. • Surrogate models are created and coupled to a rural energy planning tool. [ABSTRACT FROM AUTHOR]

Efforts towards ensuring clean and affordable electricity for all have been progressing slowly in rural, off grid areas of developing countries. In this context, hybrid microgrids may offer reliable and potentially clean electricity for isolated locations. Nevertheless, the process of planning and operation of these systems faces several challenges, often due to the uncertainties related to the renewable resources and to the stochastic nature of electricity consumption in rural contexts. This paper tackles this problem and contributes to the literature in bridging the gap between field practices and two-stage stochastic modeling approaches by identifying an open-source modeling framework which is then applied to real local data. As reference case-study, we consider a microgrid built in 2015 in Bolivia. Overall, the optimal system results from a compromise between the Net Present Cost, the peak capacity installed and the flexibility (to balance variable generation). Different approaches to size isolated microgrids are tested, with the conclusion that methods accounting for the uncertainty in both demand and renewable generation may lead to a more robust configuration with little impacts on the final cost for the community. • The historical monitoring data from a remote PV-diesel microgrid in Bolivia for rural electrification is presented. • A novel two-stage optimization framework is developed for the robust design of such systems under uncertainty. • Components models and performance curves are derived from the monitoring data. • Among different model formulations, stochastic MILP models generate the most robust system configuration. • Recommendations are formulated regarding the optimal configuration with respect to the existing system. [ABSTRACT FROM AUTHOR]