SHED

In order to enable high frequency NILM algorithm evaluation, we release a synthetic dataset called SHED whose purpose is to evaluate the disaggregation performence of NILM algorithm. This dataset contains simulated current and voltage measurements for X buildings. Each building contains the total current and voltage measurements on top of individuals consumptions: the power (building x, y, z) or the current (building a, b, c). The building features (number of categories, devices class) are summarized in the metadata file. For further information on the simulator, the data used or the evaluation of the dataset please refere to the Publication page.

The current version of the dataset is 1.0.

This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.

To receive the password needed to unzip the files, please fill in this form.

Individual building archives containing csv files:

• Building 1
• Building 2
• Building 3
• Building 4
• Building 5
• Building 6
• Building 7
• Building 8

Metadata:

• Metadata

Complete HDF5 file:

• SHED This HDF5 files has been created using Python (h5py, numpy).

To use the SHED dataset in publications, please cite the following paper:

Simon Henriet, Umut Simsekli, Benoit Fuentes, and Gaël Richard, “A generative model for non-intrusive load monitoring in commercial buildings,” Energy and Buildings, vol. 177, pp. 268 – 278, 2018. doi.