Where to Install Distribution Phasor Measurement Units to Obtain Accurate State Estimation Results?

With the rapid expansion of distributed energy resources in the distribution network, system operators need accurate and real-time estimations of system states in order to actively manage the distribution system. However, existing metering infrastructure does not provide adequate support for real-time monitoring in the distribution system. In this paper, we propose a real-time, data-driven approach to estimate the nodal voltage magnitudes in the distribution system. In this approach, we first train a neural network that maps historical measurements from distribution phasor measurement units (DPMUs) to smart meter data. The trained neural network is then used to estimate real-time nodal voltage magnitudes based on streaming DPMU data. In addition, we design an approach to identify the optimal locations to install DPMUs based on the facility location function selection method. Numerical study results on an IEEE test feeder show that our proposed method produces accurate nodal voltage magnitude estimates. The facility location function selection method also recommends DPMU candidate locations that yield lower voltage magnitude estimation error.