Outlier-Removal Memory-Event-Triggered Proportional-Integral State Estimation for Wind Turbine Systems Under Multimodal Deception Attacks

This paper studies the outlier-removal memory-event-triggered proportional-integral state estimation for variable-speed wind turbine systems subject to multimodal deception attacks. An innovative outlier-removal memory-event-triggered mechanism using historic data and comparison error upper bound is proposed to enhance the utilization of the limited network resource. The first merit of this mechanism is that the average value of some historical data is employed to smooth the measured outputs with stochastic fluctuations, which is able to reduce the unnecessary triggering events caused by the fluctuations. The second one is that the triggering error term’s upper bound is introduced to avoid the false transmissions induced by measurement outliers. A multimodal deception attack model with different occurrence probabilities for different modes is considered, which is more general and practical than the existing single-modal model. In order to estimate the system state, a fuzzy proportional-integral estimator is presented. With the introduction of an extra integral term, the presented estimator has the potential to improve the estimation performance compared to the traditional proportional estimator without the integral information. Then, the closed-loop estimating error system for variable-speed wind turbine is established as a T-S fuzzy system. Resorting to the Lyapunov theory and linear matrix inequality technique, a set of sufficient conditions for co-solving the estimator gains and triggering matrices are produced. Ultimately, the effectiveness of the developed method is verified by some simulation results. Note to Practitioners—This paper investigates the T-S fuzzy-model-based proportional-integral state estimation for variable-speed wind turbine systems, addressing challenges raised from practical constrained network bandwidth, sampling outliers and multimodal deception attacks. The novel outlier-removal memory-event-triggered mechanism improves network utilization, and the fuzzy proportional-integral estimator enhances the state estimation accuracy and ensures the asymptotic stability for wind turbine systems with nonlinearities and multimodal deception attacks. Note that the investigated method can be applied to deal with the above practical issues in real-world systems, which can provide technique support for practitioners in real-world implementation. In addition, based on the estimated state, various state-feedback control methods can be realized to improve the control performance of wind turbine systems.