No abstract

No abstract

Microgrids are an increasingly relevant technology for integrating renewable energy sources into electricity systems. Based on a microgrid implementation in California, we investigate microgrid operation under real-world conditions. These conditions encompass energy charges, demand charges, export limits, as well as uncertainty about future electricity demand and generation within the microgrid and have not yet been considered in this combination. Under these conditions, we evaluate the performance of two frequently applied strategies for microgrid operation. The first is a reactive greedy strategy that makes operational decisions based exclusively on the current state of the microgrid and does not require a centralized control architecture. The second strategy is proactive and optimizes decisions based on forecasts of future electricity generation and demand. We evaluate the performance of the strategies under varying operational parameters, forecast accuracies, and configurations of the microgrid.We thereby provide detailed guidance for research and practice on what kind of operational strategy is advantageous in a variety of settings, well beyond our Californian showcase. In addition, the interplay between real-world conditions and operational strategies reveals several novel insights for research on microgrid operation. For instance, escalating negative interactions between forecast errors and demand charges make proactive strategies benefit from longer control horizons. This result is contrary to existing best practice, which promotes short control horizons to minimize the impact of uncertainty.

The increasing prevalence of distributed photovoltaic (PV) units raises stress on distribution grids and necessitates increased grid planning efforts. We present a decision support system (DSS) based on integer programming that is able to determine cost-optimal grid reinforcements at the level of individual grid segments. The functionality of the DSS is demonstrated in a scenario analysis of a rising adoption of PV units relying on 1,000 simulation runs in a real-world grid. Based on the results, we provide guidelines for operative grid planning and illustrate how the system assists in the evaluation of reinforcement technologies as well as in long-term investment planning. Furthermore, thanks to segment-specific optimization, the DSS shows that at constant adoption levels, reinforcement cost can vary largely depending on the location of the PV units in the grid. Therefore, a high amount of uncertainty seems unavoidable in long-term prognoses on the effects of solar power on distribution grids

No abstract