Joint Sparse Estimation of Sensitivity Distribution Factors and Power Flows in Low-Observable Power Distribution Systems.

Sensitivity distribution factors (SDFs) have diverse applications in power system operation. In particular, when SDFs are obtained in a data-driven fashion based on measurements, they can eliminate the need to repeatedly solve the computationally complex non-linear power flow equations. However, in a typical low-observable power distribution system, where the measurements are not sufficient to achieve full-observability, it is a major challenge to estimate SDFs using the available measurements. This challenge is addressed in this paper. Specifically, a new method is proposed for joint estimation of SDF and power flows in power distribution systems that lack full-observability. The proposed method requires measuring the nodal injection power and line power flow at only a few locations on the power distribution feeder. The proposed method is physics-aware. It is built upon extracting physics-based sparsity features of power distribution feeders. In this regard, the aforementioned joint estimation problem is formulated as a sparse matrix completion problem. The advantages of the proposed method in comparison with the existing methods are investigated via numerical results.