PMU deployment for optimal state estimation performance

The future `smart´ grid will see increasing deployments of intelligent electronic devices (IED), that sense the grid state variables so as to support enhanced, real-time monitoring and control within a service area for achieving increased operational efficiency, reliability and security. Increasing penetration of Phasor Measurement Units (PMUs) like devices within the distribution grid are anticipated; however, due to the high cost of PMU installation, their deployment will continue to be selective for the foreseeable future. Previous work has largely focused on achieving full topological observability with (appropriately placed) minimal number of PMUs for a given network. We observe that there typically exist multiple minimal-PMU sets that achieve full observability, affording an additional degree of freedom to select an optimal choice among this set. We define the desired solution as the placement that also achieves best state estimation performance. Accordingly, we propose two approaches - a greedy algorithm and an integer programming method to determine the PMU locations that minimize the mean-squared state estimation error.