Sensitivity Analysis of Controller Performance in Decentralized Resilient Load Frequency Control for Cyber-Physical Power Systems under DoS Attacks

Cyber-physical power systems (CPPSs) integrate control devices, sensor networks, and communication with traditional power systems, which complicates load frequency control (LFC) due to issues like packet dropouts and delays. The increasing threat of cyber-attacks, such as energy-limited denial-of-service (DoS), has amplified the need for resilient control strategies, attracting research interest. This study performs a comprehensive sensitivity analysis for a decentralized resilient H<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">∞</inf> LFC scheme using state-feedback control for a single-area CPPSs within a network-based control framework. Sampled measurements transmitted through communication networks, vulnerable to DoS (characterized by maximum count of continuous data losses), are considered. The closed-loop LFC system under DoS is modeled as an aperiodic sampled-data control system. The work presented here makes a comprehensive investigation of how heavily the robust control performance depends on suitable choices of these hyper-parameters involved in the controller design and how impactful are choices of these parameters with significant variations in attack scenarios and sampling intervals. Performance indices computed in terms of integral absolute error (IAE), used to measure frequency deviation responses under two different attack conditions, firmly establish that an accurate understanding of sensitivity of these hyper-parameters indeed help in achieving superior performances compared to the latest state-of-the-art controllers reported in literature.