A sensorless cyberattacks mitigation technique based on braided lyapunov state observer

Abstract

The transferred data in microgrids is susceptible to cyberattacks. This paper presents a primary and secondary control loops integrated with units for detecting and mitigating cyberattacks for each distributed generator in a DC microgrid. Most of the literature focused on mitigating cyberattacks using supervised learning techniques. On the contrary, this paper introduces reliable state observer-based mitigation units for voltage and current cyberattacks. A braided Lyapunov-based state observer is proposed to generate accurate estimate of terminal voltage and current, simultaneously, without the need for external measurements. Afterwards, cyberattack mitigation controller is triggered to develop the appropriate corrective actions. Fast control action is essential for effective mitigation of cyberattacks, along with precise state estimation. Therefore, two adaptive PI controllers are introduced in the cyberattack mitigation units. Different case studies are provided to assess the performance of the suggested sensorless cyberattack detection and mitigation strategy. To prove the superiority of the proposed strategy, comparative studies are carried out with the scheme of mitigating cyberattacks based on artificial neural networks (ANNs). The Lyapunov based state observer showed a lower settling time of 0.4 s, lower voltage maximum peak overshoot of 0.16 %, almost zero voltage and current sharing errors, when compared to ANNs.