A Round-Robin ADMM algorithm for identifying data-manipulators in power system estimation

Abstract

In this paper we develop an algorithm that can detect the identity of false data-injection attackers in distributed optimization loops for estimating oscillation modes in large power system models. The fundamental set-up for this distributed optimization is based on Alternating Direction Multiplier Method (ADMM). The power system is divided into multiple non-overlapping areas, each equipped with a local estimator. These local estimators use local sensor measurements to carry out a local regression algorithm for generating a local estimate for the characteristic polynomial for the system transfer function, and, thereafter, communicate this estimate to a central supervisor. The supervisor computes the average of all estimates, and broadcasts this average or consensus variable back to each local estimator to be used in the next round of regression. However, if one of the local estimators is compromised by a malicious attacker that may send corrupted values of its local estimate to the central supervisor, then it is difficult to detect the identity of this attacked estimator from the algorithm stated above. In this paper we propose an alternative algorithm where the central supervisor, instead of computing the average, employs a Round-Robin technique to generate the consensus variable, and show that by tracking the evolution of this consensus variable it is possible to identify which estimator is malicious. We analyze the convergence properties of this modified ADMM algorithm, and illustrate its effectiveness using simulation results.