State estimator for electrical distribution systems based on a particle filter

This paper presents a state estimator based on the use of a particle filter (PF). Unlike other types of filters, a PF is suitable for both nonlinear systems and non-Gaussian error distributions. The proliferation of distributed energy resources such as distributed generators and controllable loads has been accompanied by a high degree of uncertainty because the lack of sensors necessitates the use of pseudo-measurements rather than real measurements. For this reason, the proposed state estimator was tested using non-accurate measurements. Bus voltages and angles were chosen as state variables. A comparison of the PF with an extended Kalman filter (EKF) on a 5-node distribution system revealed that the PF provides a very high level of performance, superior to that obtained with the EKF. The proposed estimator was further tested on an IEEE 34-node.