False data injection attack detection based on interval affine state estimation

Abstract

With the development and application of information communication technology, it makes the potential threat of cyber-attacks prominent increasingly that the power information network and physical network are becoming deeply integrated. In the energy management system (EMS), the state estimation module is the main target of hackers' for false data injection attacks (FDIAs) on power systems deliberately. Therefore, from the perspective of redundancy of system measurement data, it proposes a pseudo-measurement model based on a new neural network, convolutional long short-term memory (CLSTM) neural network, which is a combination of different networks. And using interval number and affine number to analyze and quantitatively describe the measurement uncertainty, this paper proposes a FDIAs detection model based on interval affine state estimation. It provides technicians with the upper and lower limit information of the systems state, which makes it easier to judge whether the measured data is attacked and thus exceeds the normal fluctuation range. This paper has tested IEEE 33-bus system and the results show that this method has obvious advantages in term of calculation speed and accuracy.