Adversarial FDI Attack against AC State Estimation with ANN
This addresses a security problem for smart grid operators by exposing a critical vulnerability in ANN-based state estimation, though it is incremental as it builds on known adversarial attack methods applied to a new domain.
The paper tackles the vulnerability of artificial neural networks (ANN) used for AC state estimation in smart grids to adversarial false data injection (FDI) attacks, showing that proposed algorithms, particularly DE, can degrade ANN accuracy with high probability in simulations on IEEE bus systems.
Artificial neural network (ANN) provides superior accuracy for nonlinear alternating current (AC) state estimation (SE) in smart grid over traditional methods. However, research has discovered that ANN could be easily fooled by adversarial examples. In this paper, we initiate a new study of adversarial false data injection (FDI) attack against AC SE with ANN: by injecting a deliberate attack vector into measurements, the attacker can degrade the accuracy of ANN SE while remaining undetected. We propose a population-based algorithm and a gradient-based algorithm to generate attack vectors. The performance of these algorithms is evaluated through simulations on IEEE 9-bus, 14-bus and 30-bus systems under various attack scenarios. Simulation results show that DE is more effective than SLSQP on all simulation cases. The attack examples generated by DE algorithm successfully degrade the ANN SE accuracy with high probability.