Niloy Patari
Published: 2021
Total Pages: 0
Get eBook
Traditionally, distribution power utilities use centralized or local architecture for voltage control and optimization in distribution power systems. However, with increasing penetration of renewable inverter-based resources in distribution systems, voltage control is becoming increasingly challenging. Since converter interfaced distributed energy resources (DERs) also act as voltage control agents in addition to already deployed legacy devices like voltage regulators and line capacitors, voltage optimization is a large-scale optimization problem. Centralized optimization algorithms, though they provide optimal solutions, suffer from scalability issues since computation of setpoints is performed in a centralized entity. Centralized approaches are also prone to single point of failures since these are dependent on secure network-wide communication for data-delivery. Local schemes, however, depend only on local measurements and local communication but provide non-optimal solutions. Distributed approaches combine the advantages of both centralized and local strategies. Distributed optimization (DO) facilitates parallel computation and distributed decision-making while providing near-optimal solutions. Moreover, DO involves communication only among neighbor agents thereby reducing communication overhead resulting in improved robustness to communication failures. This work presents distributed algorithms for optimal voltage control and optimization for multiphase unbalanced distribution feeders.First, a taxonomy of existing distributed algorithms used for optimal control in distribution power grids is developed based on communication infrastructure, database required, and method of implementation. The taxonomy provides two major classes of distributed algorithms- static optimization approaches and dynamic optimization approaches. Second, these two classes have been distinguished by providing two use cases. Static optimization approaches involve optimization problems solved by agents with local optimization subproblems, while dynamic approaches involve agents interacting with the grid as feedback control. Third, a distributed online voltage control algorithm is developed for unbalanced multiphase feeders considering the reactive power capability of DERs. Fourth, another distributed online voltage control algorithm considering both active and reactive power capability of DERs is developed. Both algorithms are validated and are shown to provide superior results for IEEE benchmark unbalanced feeders including the IEEE 13 bus and IEEE 123 bus feeder under communication delays, modeling errors, and asynchronous communication among distributed control agents.