You are here

ECE PhD Defense: "Robust Dynamic State Estimation In Power Systems," Alireza Rouhani


442 Dana Research Center

February 23, 2017 11:00 am
February 23, 2017 11:00 am

The widespread deployment of Phasor Measurement Units (PMUs) on power systems has facilitated the real-time monitoring of the power systems dynamics. Dynamic State Estimators (DSEs) are used by the investigators to estimate and identify the state variables and parameters of the nonlinear dynamic models within the power systems by using the measurements which are mainly provided by the PMUs. This dissertation addresses fundamental research on dynamic state estimation of the power systems and presents innovative and robust dynamic state estimation approaches for estimation/identification of the state variables/parameters associated with the nonlinear dynamic models within the power systems.

The first part of this dissertation focuses on real-time parameter identification of the nonlinear dynamic load models. For this purpose an Unscented Kalman Filter (UKF) bases DSE is developed to identify the unknown parameters of an exponential dynamic load model in real-time.

As a next step this work presents a two-stage distributed dynamic state estimation approach which remains robust under the occurrence of the bad-data associated with the measurements that are used for the dynamic state estimation. The first stage of the proposed approach utilizes a Least Absolute Value (LAV) linear phasor estimator and the second stage of the proposed approach uses UKF as an efficient DSE.

Observability analysis of the nonlinear dynamic models within the power systems is another topic that is investigated in this dissertation. A Lie-derivative based observability analysis approach is presented in this work which allows us to evaluate the level of the observability for a given measurement associated with a nonlinear dynamic model such as dynamic model of the synchronous generator and load.

This dissertation introduces an UKF based DSE which is named Constrained Iterated Unscented Kalman Filter (CIUKF). One of the main features of the proposed DSE is that it is capable to identify the unknown parameters of the synchronous generators such as inertia constants and transient reactances while estimating the dynamic state variables of the synchronous generator.

This work presents a dynamic state estimation method which remains robust under occurrence of the synchronous generators excitation system failure.

Moreover, the proposed approach informs the systems operator about the occurrence of this failure in a timely fashion without any need to locally evaluate the performance of the excitation system by using already existing costly and time-consuming approaches.

Finally in this work a stand-alone robust DSE algorithm will be presented.

One of the most important advantages of the proposed approach is that it is capable to evaluate to quality of the measurements provided by the local PMUs in real-time and therefore detect the occurrence of the bad-data associated with measurements.

All of the aforementioned methods in this work will be implemented on the WECC and NPCC test systems and the associated results will be also presented.


  • Prof. Ali Abur (Advisor)
  • Prof. Hanoch Lev-Ari
  • Prof. Bahram Shafai