Optimal sizing and operation of isolated microgrids for developing countries: hedging uncertainties with Monte Carlo techniques

Achieving universal electrification is a challenging goal for governments in developing countries; the traditional approach of extending the national grid to rural areas has been usually pursued but since many communities are located far from the national grid, this approach may be sub-optimal. Hybrid microgrids, composed by renewables, batteries and a backup fuel-fired genset, are very promising solutions, but socio-political and geographical concerns undermine the optimal design of the system and the economic viability of projects. Recent advanced approaches use Monte Carlo techniques to size the system accounting for uncertainties regarding the load and renewable production; however, few papers have investigated the optimal number of Monte Carlo scenarios to evaluate for a proper sizing. In this study, we investigate the benefits and drawbacks of using Monte Carlo techniques to cope with uncertainties involved in the optimal design and operation of hybrid microgrids, proposing new sizing criteria and highlighting best practices. A case study using real data from a minigrid in Kenya is proposed.