Distributed Dynamic Clustering Algorithm for Formation of Heterogeneous Virtual Power Plants Based on Power Requirements

• Published in: IEEE transactions on smart grid Vol. 12; no. 1; pp. 192 - 204
• Main Authors: Zhang, Runfan, Hredzak, Branislav
• Format: Journal Article
• Language: English
• Published: Piscataway IEEE 01.01.2021; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Agglomeration; Algorithms; bulk power supply; Clustering; clustering algorithm; Clustering algorithms; Communication networks; Control systems; Demand; distributed communication network; Distributed generation; Electric power distribution; Electric power generation; Energy; Energy storage; frequency regulation; Heterogeneous virtual power plants; Heuristic algorithms; photovoltaic; Power demand; Power generation; Power plants; Provisioning; Real time; state of charge balancing; Storage systems; Virtual power plants; voltage restoration
• ISSN: 1949-3053; 1949-3061
• DOI: 10.1109/TSG.2020.3020163

Real-time distributed clustering algorithm for aggregation of distributed energy storage systems into heterogeneous virtual power plants is proposed. Two types of virtual power plants are formed: one for provisioning the bulk (low-frequency) power demand and one for provisioning the high-frequency power demand. The proposed distributed clustering algorithm determines the virtual power plants' memberships, for each distributed energy storage system, based on each energy storage system's capacity and owner's willingness to participate in one of the virtual power plants. The proposed distributed secondary level control system regulates each energy storage system according to each virtual power plant's operational objectives. Specifically, a balanced state of charge of all energy storage systems inside each virtual power plant is maintained. One of the virtual power plants is responsible for the frequency and voltage regulation by providing the required high-frequency power, while the other one provides the required bulk (low-frequency) power demand. In addition, the proposed clustering algorithm enables to meet a required energy capacity of the bulk virtual power plant by automatically tuning the clustering algorithm parameters. RTDS real-time technique verifies the proposed clustering algorithm and control systems on the IEEE 13 nodes power system with distributed energy storage systems and photovoltaic sources. The presented results demonstrate dynamic aggregation of energy storage systems into heterogeneous virtual power plants based on power demand while all regulation requirements are met.