Sparse Distribution System State Estimation: An Approximate Solution Against Low Observability

Abstract

When it comes to distribution system state estimation (DSSE), limited network observability is a major concern, due to limited sensor deployment at practical power distribution systems. To address this issue, this paper proposes a novel DSSE approach, based on sparse recovery for distribution networks with low-observability, where the measurements come from only a handful of distribution-level phasor measurement units. Here, the DSSE problem is formulated over differential synchrophasors, and in form of a least absolute shrinkage and selection operator (Lasso), which is solved using the alternating direction method of multipliers (ADMM). Importantly, our solution method is dynamic, because it uses the state estimation results from the previous time slots in order to update the weights in the instances of the Lasso problem so as to enhance the DSSE performance.