Robust Linear State Estimation For Large Multi-area Power Grids

This paper focuses on state estimation in very large systems with several control areas where there are sufficient phasor measurement unit (PMU) measurements to allow implementation of a linear state estimator. The paper exploits the natural partitioning of the grid by control areas and develops an estimation framework based on the well-known Dantzig-Wolfe decomposition principle for linear programming problems. The objective of this approach is to fully utilize PMU measurements, achieve robustness against bad data irrespective of their locations and develop a solution algorithm whose computational performance will remain insensitive to the number of control areas in a large multi-area power grid. Problem formulation, derivation of the solution algorithm and sample results illustrating its performance are presented.