Reliable Control Based on Improved Adaptive Observer for Cyber-Physical Power Systems with Power Load Anomalies

Power systems are indispensable infrastructure in modern society, and reliability and robustness are crucial for stable and sustainable power delivery. This paper investigates the reliable control of cyber-physical power systems (CPPSs) subject to aggregated load anomalies at load buses. Such load anomalies are generally triggered by load fluctuations and dispatching imbalance. Considering that only partial system states can be measured in practical CPPSs, an output-feedback control framework is adopted in this work. To mitigate adverse consequences induced by load anomalies, including circuit overheating, insulation aging, grid faults and even structural damage of power grids, an improved observer is constructed to achieve accurate real-time estimation of unknown load anomalies and provide feedforward compensation for the control loop. Different from the existing literature, the restrictive norm-bounded assumption on load anomalies is eliminated in the proposed design. Based on the estimated load anomaly information and measurable system outputs, a novel improved observer-based sliding mode control law is developed. By utilizing Lyapunov stability theory and matrix inequality techniques, rigorous sufficient conditions are derived to guarantee the bounded stability of the closed-loop CPPS. Meanwhile, the boundaries of system states and load anomaly estimation errors are explicitly obtained. Finally, the feasibility and effectiveness of the proposed method are verified on a test power system with three generators and six buses.