Bad data detection, identification and correction in distribution system state estimation based on PMUs

This paper presents a novel approach for bad data correction in state estimation (SE) for three-phase distribution systems. Based on an optimization model, a SE technique is presented considering branch currents as state variables to be estimated in regular time intervals. In this work, the presence of gross errors is detected by a comparative analysis of the objective function with a threshold value determined by Monte Carlo simulations assuming different load scenarios and Gaussian aleatory errors associated with the measurements gathered from the network. A novel index is proposed for identifying the corrupted measurements based on their corresponding largest residuals. For bad data correction, a new procedure is presented based on statistical analysis of the measurements variation along the time. Computational simulations are carried out using the IEEE 33-bus test system in order to prove the efficiency of the proposed methodology. The main contribution of this paper is the development of a gross error correction technique for SE assuming a limited number of phasor measurement units allocated along the feeders considering all the peculiarities of unbalanced distribution networks. This feature ensures that estimation errors lower than 1% are provided for network operators eliminating the effect of bad data in the SE process.