Online power system state estimation using alternating direction method of multipliers

The convex relaxation approaches for power system state estimation (PSSE) offer robust alternatives to the conventional PSSE algorithms, by avoiding local optima and providing guaranteed convergence, critical especially when the states deviate significantly from the nominal conditions. On the other hand, the associated semidefinite programming problem may be computationally demanding. In this work, a variable splitting technique called alternating direction method of multipliers is employed to reduce the complexity, and also efficiently accommodate a regularizer promoting desired low-rank matrix solutions. Both static and online formulations are developed. Numerical tests verify the efficacy of the proposed techniques.