1. Real-time dynamic security assessment of power system using strategic PMU measurements and control technique
- Author
-
Ensaf, Mohammad and Ensaf, Mohammad
- Abstract
This research presents a comprehensive exploration of adaptive control strategies and data-driven anomaly detection within power systems. At the core of the study lies the investigation of adaptive finite-time tracking control for strict-feedback nonlinear continuous-time systems. These systems, when influenced by full-state constraints and dead zones, present significant challenges. By harnessing the principles of finite-time stability theory combined with barrier Lyapunov functions, the study introduces a groundbreaking adaptive tracking control strategy. Coupled with the adaptive backstepping method, this approach guarantees that the closed-loop system’s signals remain bounded. Furthermore, it ensures that outputs adeptly track reference signals, while all system states are confined within predefined compact sets, enhancing system reliability and performance. In parallel, the research unveils an innovative approach to tackle the intricate challenge of anomaly detection in Phasor Measurement Unit (PMU) data. Recognizing the high-dimensional nature of PMU data, an ensemble model, synthesizing the strengths of Gaussian Process Regression (GPR) and Autoencoders, is proposed. This ensemble not only boasts superior data reconstruction fidelity but also features a Bayesian optimization-driven threshold determination. Such a methodology fosters an adaptive, data-driven anomaly detection process, resulting in heightened specificity and sensitivity. Validation tests conducted on a synthetic dataset, infused with 84 frequency events, attest to the ensemble model’s superior capability in discerning nuanced anomalies. This superiority is evident both visually and through rigorous quantitative metrics, underscoring the ensemble’s edge over traditional models.
- Published
- 2024