Coordinated Frequency Regulation of Active Distribution Networks Considering Dimension-Augmented Power Flow Constraints

Abstract

Distributed energy resources (DERs) integrated in active distribution networks (ADNs) participating in primary frequency regulation (PFR) service can enhance frequency safety and stability of power systems. However, PFR service can result in power flow (PF) insecurity issues, especially in low and medium-voltage networks without accurate line parameters. To address the problem, this paper proposes a coordinated control architecture based on the Koopman data-driven power flow. The cluster model training layer uses Koopman operator theory to transform the original complex nonlinear PF model into a dimension-augmented linear PF model. The online PFR optimization layer constructs an optimization model of PFR based on the data-driven PF, considering security constraints of ADNs. The local frequency response layer responds to frequency change in real-time and ensures fast frequency support. This method is validated using a modified IEEE 82-node test case, which demonstrates that it has the advantages of fast online solving, and independence on model parameters. The proposed method can fully exploit PFR capability of ADN and achieve the optimal PF profiles while ensuring the aggregate PFR characteristics.