False Data Injection Attacks With Incomplete Network Topology Information in Smart Grid

False data injection (FDI) attack causes a disadvantage to the safety of a power system. In the past, the adversary launching FDI attack mostly needed the complete network topology information of the power system and a large amount of measurements. However, the reality is that the adversary has limited resources and can hardly manipulate massive measurements without being detected. This paper proposes a new method to form attack vector in the condition of incomplete system information. The proposed method first maps the restricted data into a new Jacobian matrix by a kernel-independent component analysis, then constructs the undetectable attack model, and finally, designs a less costly attack vector to accomplish the FDI attack. The experimental results on different IEEE buses with Matpower tool illustrate the validity of the proposed method. Additionally, the proposed method shows a relatively high success rate in the cases of different degrees of incomplete information and various quantities of manipulated data in different power systems.