Resilient load frequency control of multi-area power systems with packet-loss mitigation over lossy communication links

Abstract The growing dependence on communication networks in modern load frequency control (LFC) systems makes frequency control susceptible to packet-loss and network-induced uncertainty. Most LFC methods assume reliable communication or asymptotic estimation, so handling stochastic and bursty packet-losses remains a key challenge. This research introduces a robust LFC architecture for power systems operating over lossy communication channels by utilizing nonlinear finite-time active disturbance rejection control (NFT-ADRC). A finite-time extended state observer (FT-ESO) is utilized to estimate system states and aggregate disturbances resulting from load variations, and parameter uncertainties. The finite-time convergence property guarantees that estimation errors are driven to zero in finite time rather than asymptotically, resulting in an improved transient response and enhanced performance in frequency regulation. To address communication imperfections, packet-loss is modeled using a Gilbert-Elliott process, which accurately represents both stochastic and bursty loss characteristics. A hold-last-value (HLV) based mitigation strategy is implemented at the actuator side to ensure control signal continuity during packet dropouts, while preserving the integrity of the controller structure. The proposed approach is evaluated on a multi-area hybrid power system for various realistic operating conditions including packet-loss, communication delays, random load perturbation (RLP), and uncertainty in system parameters to assess robustness against practical disturbances, modeling inaccuracies and communication network induced uncertainties. The simulation results show that the NFT-ADRC significantly enhances the LFC performance through accurate disturbance estimation. The HLV based packet-loss mitigation strategy effectively maintains control signal continuity during packet dropouts, minimizing performance deterioration in lossy network conditions. As a result the proposed method ensures resilient and robust load frequency control in the presence of actual disturbances, uncertainties, communication delays, and lossy communication conditions.