Distrbution System State Estimation

This dataset comprises training and test input and output data for a deep neural network model designed to estimate the states of a distribution system in real-time using measurements reported from sparsely placed micro-phasor measurement units (micro-PMUs). There are two distinct sets of data corresponding to two different distribution systems: a 240-node US Midwest distribution system and a 34-node feeder.
The input data include voltage and current phasors (arranged column-wise) for various operating conditions of the power system (arranged row-wise) based on a specific hour of the day. The order of columns is voltage magnitude (per unit), voltage phase angle (radians), current magnitude (Amp), and current phase angle (degrees).
Micro-PMUs are strategically placed at 5 nodes for the 240-node system and at 3 nodes for the 34-node system. It's important to note that the input data, representing micro-PMU measurements, have been subject to corruption by a non-Gaussian noise model, adhering to a 1% total vector error as per PMU standards.
The generation of synthetic micro-PMU measurements involved using hourly smart meter data provided for the 240-node US Midwest and Pecan Street Dataport for the 34-node feeder. This process was carried out by solving power flows for various operating conditions.
Note that there are three distributed generation units (PVs) having ratings of 135, 60, and 60 kW placed at 3 different nodes in the 34-node feeder.

Reference:
B. Azimian, R. S. Biswas, S. Moshtagh, A. Pal, L. Tong, and G. Dasarathy, ?State and topology estimation for unobservable distribution systems using deep neural networks,? IEEE Transactions on Instrumentation and Measurement, vol. 71, pp. 1?14, 2022.