Transient Stability Assessment

James J. Q. Yu, David J. Hill, Albert Y. S. Lam, Jiatao Gu, Victor O. K. Li
The University of Hong Kong

Transient stability refers to the capability of a power system to maintain its synchronism subject to large disturbances, which may lead to significant system failures or power blackouts. In order to prevent such situations, system operators need to assess the stability condition of the grid and, when necessary, plan a collection of remedial control actions to retain the stability. The gradual adoption of synchrophasor measurement facilities, e.g., phasor measurement units (PMUs), makes it feasible and efficient to utilize real-time system parameters for transient stability assessment (TSA) decision making. Our research focuses on establishing an intelligent system to address TSA problem, specifically by employing machine learning techniques. Utilizing the advantages of these methodologies, the proposed system is able to make pre-mature assessments at the earliest possible time and revise the predictions when more information is available, rendering early warnings on transient stability issue available for modern power systems. Comparing with existing work on TSA, our proposed framework comprises the following:

• We propose a system for early TSA, where the temporal data dependencies within system parameters are extracted and utilized for better accuracies with minimal response time.
• We propose a delay sensitive TSA scheme to address the impact of communication delay on system parameters in the process of TSA, which is critical for making fast response system, but has not been considered in the previous literature.
• Our proposed system is based on model-free learning techniques, which significantly reduce the requirement on analytical knowledge of the power system infrastructure.
• Our proposed system is robust under high and unstable transmission delay with significant data noise.

Our approach can be extended and employed to solve a wide range of data-driven power system applications.