Event-triggered model-free adaptive predictive control for multi-area power systems under deception attacks

Abstract

This paper focuses on the event-triggered model-free adaptive predictive load frequency control (LFC) problem for multi-area power systems under false data injection (FDI) attacks. First, the power system model is assumed to be unknown, and the nonlinear power system is converted into an equivalent linear data model using the input and output data of the powerline system. The FDI attacks on the power system is also modeled, and a Bernoulli stochastic process is used to represent whether the data is transmitted successfully or not. Second, the equivalent linear data model to predicts upcoming events, and a new event-triggered model-free adaptive predictive control scheme is designed via the predictions. And a RBF neural network disturbance estimator is designed to estimate and compensate for the effects caused by disturbances, and then power system stability is analyzed. An event-triggered scheme is also developed in the design of the LFC in order to save communication and computational burden. The results show that the designed control algorithm is data-driven independent of the power system structure and does not require the measurement of any system state signals. Finally, the effectiveness and correctness of the scheme is verified by a three-area numerical example and simulation results.

Keywords

• Model free adaptive predictive control
• Load frequency control
• Event-triggered
• Deception attacks
• Data driven