Model-Data-Driven Approach for Achieving Decoupled Power Flow and Its Application in Asymmetric Bipolar DC Distribution Networks

Abstract

Bipolar DC distribution networks (Bi-DCDNs) offer a promising alternative to medium and low voltage distribution networks by enhancing both the loadability and the access capability of renewable energy sources. However, coupled power flow in asymmetric Bi-DCDNs poses challenges for system-level optimal operation problems. Hence, this paper proposes a model-date-driven decoupling framework and employs it to construct the static voltage stability region (SVSR), a representative operational challenge in asymmetric Bi-DCDNs. More specifically, the model-driven approach defines the decoupling coefficient and derives its analytical expression through branch flow analysis. The expression denotes a parametric equation governing pole voltage, functioning as a posteriori indicator of the state of Bi-DCDNs. This equation manifests as a highly nonlinear expression, which can be further determined through a data-driven approach. Various operational scenarios of Bi-DCDNs are simulated using Monte Carlo sampling, without making assumptions about the distribution of loads. The distribution of the decoupling coefficient is derived from power flow calculations, with the decoupling coefficient determined as the expected value within an acceptable confidence interval. Subsequently, the optimal power flow problem for decoupled Bi-DCDNs is formulated, serving as the basis for constructing the SVSR of Bi-DCDNs. The numerical results indicate that the proposed decoupling framework achieves both computational efficiency and accuracy. Furthermore, it exhibits advantageous applications for asymmetric operational Bi-DCDNs.