Angular Reference Models for Optimization-based Unbalanced Distribution System State Estimation

The importance of state estimation has surged for distribution networks, given their increased digitalization and measurement availability. However, distribution system state estimation (DSSE) implementations still encounter several practical challenges. One of them is defining unbalance at the reference bus. Existing three-phase DSSE methods usually assume reference buses with equally spaced voltage phase angles. This assumption does not hold in general in low voltage (LV) distribution networks, due to their imbalance.Although recent publications proposed alternatives to the balanced reference bus assumption, several aspects of such alternatives have not been explored. This paper benchmarks DSSE using the balanced angle approach and two recently proposed alternatives. Our DSSE implementation leverages optimization concepts that are not featured in conventional DSSE. These allow to model unbalanced reference angles, while overcoming the numerical challenges that were reported in prior literature. Accuracy trade-offs with the different reference angle models are demonstrated on a realistic LV feeder across various scenarios.