Presenter Information

Jim Follum, University of Wyoming

Department

Electrical and Computer Engineering

First Advisor

Dr. John Pierre

Description

Oscillations in the power system known as electromechanical modes are a major concern for stability. If a mode is allowed to become unstable, it can cause widespread blackouts of the power system. For this reason, it is desirable to constantly monitor the stability of the modes by applying signal analysis algorithms to ambient data collected from the power system. The Recursive Maximum Likelihood (RML) algorithm has been shown to be capable of accomplishing this task. To make the RML algorithm better suited for practical applications where missing or corrupt data are a reality, a robustness component was added to the algorithm. This new algorithm, known as Robust RML (RRML), was then tested and compared to the basic RML algorithm and the existing Recursive Least Squares (RLS) and Robust RLS (RRLS) algorithms. Results indicate that the RRML algorithm is able to handle missing and corrupt data better than the RML algorithm, achieves accuracy similar to the RRLS algorithm, and provides real-time statistical information that the RRLS algorithm cannot. From these results, it is clear that the RRML algorithm is a useful tool in monitoring the power system’s stability in real time.

Comments

Oral Presentation, Wyoming NSF EPSCoR

Share

COinS
 

Validation of the Robust Recursive Maximum Likelihood Algorithm for use in Electromechanical Mode Estimation

Oscillations in the power system known as electromechanical modes are a major concern for stability. If a mode is allowed to become unstable, it can cause widespread blackouts of the power system. For this reason, it is desirable to constantly monitor the stability of the modes by applying signal analysis algorithms to ambient data collected from the power system. The Recursive Maximum Likelihood (RML) algorithm has been shown to be capable of accomplishing this task. To make the RML algorithm better suited for practical applications where missing or corrupt data are a reality, a robustness component was added to the algorithm. This new algorithm, known as Robust RML (RRML), was then tested and compared to the basic RML algorithm and the existing Recursive Least Squares (RLS) and Robust RLS (RRLS) algorithms. Results indicate that the RRML algorithm is able to handle missing and corrupt data better than the RML algorithm, achieves accuracy similar to the RRLS algorithm, and provides real-time statistical information that the RRLS algorithm cannot. From these results, it is clear that the RRML algorithm is a useful tool in monitoring the power system’s stability in real time.