Automated vulnerability analysis of AC state estimation under constrained false data injection in electric power systems

We introduce new methods for the automatic vulnerability analysis of power grids under false data injection attacks against nonlinear (AC) state estimation. We encode the analysis problems as logical decision problems that can be solved automatically by SMT solvers. To do so, we propose an analysis technique named “symbolic propagation,” which is inspired by symbolic execution methods for finding bugs and exploits in software programs. We show that the proposed methods can successfully analyze vulnerability of AC state estimation in realistic power grid models. Our approach is generalizable towards many other applications such as power flow analysis and state estimation.