NON-LINEAR STATE ESTIMATION IN POWER SYSTEMS UNDER MODEL UNCERTAINTY

This paper considers the problem of non-linear state estimation in power systems when the system model is not known with certainty due to lack of complete information about the model or possible disruptions in the network. Specifically, this paper focuses on the settings in which the true model might deviate from the nominal model to a group of alternative models. Such uncertainty in the true model adds another dimension to the system state estimation. Specifically, the state estimator must also detect if the system model has deviated from the nominal model, and then isolate the true model. The estimation and detection/isolation decisions are intertwined as the estimation performance is linked with the detection/isolation decisions, but isolation of the true model is never perfect due to noisy measurements. This paper establishes this fundamental interplay between model isolation and state estimation, and characterizes the optimal state estimator and model detector.