The smaller Italian islands are a paradigmatic example of systems not connected to the national electricity grid, where energy security largely depends on generators based on the use of fossil fuels and for which hybrid energy systems—combining diesel generators with renewable sources and storage technologies (including the accumulation of drinking water produced by desalination)—are an increasingly experimented solution for producing clean energy at lower costs. Since the producibility of renewable sources depends strongly on the technologies that can be used and on the availability of sources, in terms of planning, it is interesting to evaluate how the uncertainty associated with climate variables and technological parameters influences the sustainability of the system, as well as its optimal design. Therefore, a methodology has been developed consisting of four phases.

The first phase involves the generation of scenarios to explore the uncertainty in the climate variables and parameters describing the different technologies. The second phase involves the generation of alternatives to be analyzed, i.e. the definition of different configurations of the hybrid energy system. The third phase involves a sensitivity analysis to quantify the effects of different climate and technological scenarios on the system performance, with the ultimate goal of identifying the key forcing factors that most influence the economic and environmental sustainability of such systems. The fourth phase involves a robustness analysis that identifies the most robust alternatives with respect to the uncertainty in the external forcing factors, using different metrics that reflect different levels of risk aversion of the decision maker.

The results show that the performance of hybrid energy systems in island settings is strongly influenced by the uncertainty in the main climate and technological forcings, especially for high wind potential configurations, and the identification of the optimal configuration strongly depends on how such uncertainty is treated and which sustainability indicator is considered. In particular, in a decision-making context, if RES penetration is considered as the only indicator, the most robust solution is always the one with the highest installed capacity.

If instead the net present cost is considered, the most robust solutions change significantly in relation to the robustness metric used to deal with uncertainty, i.e. the level of risk aversion of the decision maker. The results also show that the inclusion of the dynamic modeling of the desalination plant leads to a further improvement in the robustness of the solutions.