An Improved Model Predictive Fast Frequency Control for Power System Stability Against Unknown Time-Delay Switch Attack

A modern power grid is a cyber-physical system, which are vulnerable to cyber attacks. A recently found attack, the time-delay switch attack (TDSA), is made by inserting time delays into communication channels. A TDSA can be highly destructive to a power system as it can lead to instability. This paper presents a novel model predictive control (MPC) for fast frequency controller in a power system which can effectively mitigate the unknown TDSA. The MPC recently has received great attentions to be applied as FFC in a power system. Most of the MPC design are based on discrete-time model, whose future plant behaviour is calculated through iteration, rather than convolution. Nevertheless, one crucial step in the derivation of discrete-time MPC (DTMPC) is to capture the control trajectory over a finite prediction horizon. This imposes a challenge in designing a DTMPC to counteract the time-delay with unknown time length. Thus, a continuous-time MPC (CTMPC) is proposed to deal with TDSA. To overcome the unknown time delay, we synthesize an accurate time-delay estimator and sequential state predictor (SSP), are designed to accurately estimate and effectively counteract the unknown and random TDSA. All presented case studies are based on a real Taipower system and justification of the effectiveness of the proposed method was verified.