Propagation-Based Network Partitioning Strategies for Parallel Power System Restoration With Variable Renewable Generation Resources

Grid resiliency is defined as the ability to prepare for and adapt to changing conditions and withstand and recover rapidly from disruptions. Following large-scale disruptions of the power grid such as a complete blackout, parallel power system restoration accelerates the recovery process by allowing for simultaneous restoration in each island through network partitioning, thus enhancing grid resiliency. However, existing network partitioning strategies have not sufficiently considered variable renewable resources yet, and may not be flexible or computationally efficient to accommodate additional requirements from the evolving bulk power grid. To bridge these gaps, this study proposes a propagation-based optimization approach for power grid network partitioning. The requirements of blackstart resources, power balancing, synchrocheck relays for ties lines, and increased ramp rate requirements due to variable renewable resources are modeled in the constraints. Through propagation on the physical connectivity of grid assets, the network partitioning issue is formulated as a mixed integer linear programming (MILP) problem. The linearity ensures highly computational performance and optimality of the solutions from the proposed approach to identify the cutset edge possibilities for bulk power grid applications. Graph reduction strategies are proposed to further improve the computational performance for online or nearly online applications. Case studies based on two IEEE benchmark systems show that the proposed MILP model is able to determine the optimal cutset for network partitioning in approximately 0.28 seconds, and the proposed network reduction strategy is able to further improve the computational performance by approximately 9%.