Measurement noise propagation in distribution-system state estimation

The distribution power networks are growing in complexity due to the increasing penetration of the distributed power generation such as from wind and solar plants, and distributed resources such as electric vehicles and batteries, which are making the operation of the system more and more challenging. The first step towards more efficient operational capabilities is to introduce an observability of the system. Power system is considered observable, when all its system variables are known, in other words, system variables are chosen in such a way that when known, they allow computation of all other physical quantities in the system. In this paper a State Estimation (SE) software that considers bus voltage magnitudes and phase angles as state variables is presented. To deduce the state of the system from measurements a non-linear relation between measurements and state variables is used. The system is also over-determined, i.e. there are more measurements then state variables. A nonlinear Weighted Least Square (WLS) approach is used to solve the problem at hand with the final goal to demonstrate how quality of measurements affects the quality of power system state awareness. The results are presented in terms of statistical information about the state variables estimated with presented state estimator using generated measurements with added zero-mean Gaussian noise. We show that noise propagation through the SE algorithm is greatly influenced by the selection and placement of measurement devices.