MBTR voltage sensitivity coefficient dataset

This data has been used to carry out one of the experiment presented in Lorenzo Nespoli and Vasco Medici (2020). Multivariate Boosted Trees and Applications to Forecasting and Control arXiv

The MBTR python library is accessible here, while the repository containing all the code for the experiments carried out in the paper (including the one generating the figures) is accessible here.  

This dataset contains 41 days of simulated P, Q and voltage data for a 3-phases low voltage grid located in Switzerland. The grid topology, along with parameters for the grid’s cables, were retrieved from the local DSO. Power profiles of uncontrollable loads were generated with the LoadProfileGenerator; power profiles of photovoltaic roof-mounted power plants were obtained through the PVlib python library, while the electrical loads due to heat pumps was retrieved simulating domestic heating systems and buildings thermal dynamics, modelling them starting from building’s metadata. The grid was then simulated with KrangPower, an OpenDSS python wrapper, and the 3 phases voltages, power and currents retrieved for all the QP nodes of the grid, with a 1 minute sampling time.

The data can be imported in python with:

import pickle as pk

with open('vsc_data.pk', 'rb') as f:
data = pk.load(f)

This project is carried out within the frame of the Swiss Centre for Competence in Energy Research on the Future Swiss Electrical Infrastructure (SCCER-FURIES) with the financial support of the Swiss Innovation Agency (Innosuisse - SCCER program) and of the Swiss Federal Office of Energy with the project SI/501523.