Three-Phase State Estimator Using Dependent Weighted Least Squares

Classical Power System State Estimation (PSSE) assumes random, independent measurement errors. However, this assumption may not hold, as sensor dependencies introduce correlations between measurements, potentially compromising estimation accuracy. This paper proposes a three-phase Dependent Weighted Least Squares (DWLS) state estimator for distribution systems, extending the single-phase DWLS approach to account for network asymmetries and unbalanced loads using two-port based modelling. Monte Carlo simulations were performed on the IEEE 4-bus system to evaluate its robustness to Gaussian noise and bad data. Compared to three other estimators, the proposed approach demonstrated superior performance, particularly in the presence of bad data processing, where it allows the correction of affected measurements despite its exclusion. The improved results over its single-phase counterpart highlight the importance of accurate three-phase modelling in addressing the unique characteristics of unbalanced systems.