Statistical development of microgrid resilience during islanding operations

This study develops a statistical framework to quantify resilience of grid-connected microgrids to ensure critical loads are served during islanding scenarios. The primary metric of resilience is microgrid survivability, and is expressed as the probability for a microgrid to meet critical load requirements during an islanding event. Probabilities are evaluated using a Markov chain that describes microgrid operating states each hour for a 7-day time horizon. The system-level resilience metric uses asset-level reliability data of prominent failure modes (i.e. up-time, failure to start, failure to run). A military microgrid with 5250 kW diesel generation is contrasted to a hybrid microgrid that contains 2250 kW diesel generation, 3450 kW/13800 kWh storage, and 16,479 kW solar photovoltaics. An optimized control algorithm for island operations of microgrids is developed and sensitivity analyses identify control parameters that improve survivability, reduce fuel usage, and minimize curtailed critical load. The hybrid microgrid uses 47.80% less fuel than the generator-only microgrid under normal islanding operations. The hybrid microgrid also provides 99.70% survivability at the end of a 7-day islanding event compared to 95.03% for the generator-only microgrid. If solar photovoltaic generation decreases by 50%, the survivability of the hybrid microgrid falls below the generator-only microgrid. Increasing critical load by 30% above the design value decreases autonomy and survivability by 58.76% and 79.07%, respectively, for the hybrid microgrid. At the onset of an islanding event, the battery starting state-of-charge must be at least 70% to exceed the survivability of a generator-only microgrid and serve the full critical load without curtailment over the duration of the islanding event. Beginning the islanding event with a lower state-of-charge diminishes the survivability of the microgrid because there is insufficient capacity in storage to offset the potential failure of a generator or battery stack. Methods and findings from this study can inform design and operation of resilient microgrids for critical grid-connected loads or off-grid applications.