Flexible Connection to Accelerate Load Interconnection in Partially Measured Distribution Grids

Flexible connection enables rapid energization of fast-emerging loads such as data centers and electric vehicle charging stations by allowing limited real-time interventions during periods of grid stress. Existing hosting capacity analyses under this paradigm, however, typically assume full observability of load realizations, an assumption that rarely holds in distribution networks. We propose a stochastic flexibility-aware hosting capacity analysis framework that jointly determines the planning-stage maximum admissible load capacity and an operation-stage intervention policy under limited load observability, subject to constraints on the probability of intervention. We show that optimal policies admit compact characterizations based on load distribution information, yielding tractable and efficient solution algorithms. Numerical studies on an IEEE 123 feeder with realistic load data quantify the impact of metering coverage and placement on hosting capacity and demonstrate that targeted metering placement can substantially enlarge the hosting capacity unlocked by flexible connection.