The EMSx dataset: historical photovoltaic and load scenarios and forecasts for 70 industrial sites

The EMSx dataset gathers a large collection of photovoltaic and load profiles collected by the company Schneider Electric on 70 anonymized industrial sites. This dataset is released as a component of the EMSx benchmark for energy management systems introduced in a sister publication DOI: 10.1007/s12667-020-00417-5 . Please refer to this related journal article for a detailed description of the dataset and if mentioning the EMSx dataset in your work.

This dataset contains compressed files with a site ID ranging from 1 to 70. For these files, each line represents a timestep, gathering historical photovoltaic energy production and energy consumption data over the last 15 minutes, together with historical forecasts computed by Schneider Electric. The column naming follows the rules: 

• load_XX in kWh: For XX from 00 to 95, a forecast for the consumption where each subsequent value is 15 minutes later. load_00 is a forecast for the load in the next 15 minutes. These values are forecasts, and so will not exactly match the actual consumption at that timestep.

• pv_XX in kWh: For XX from 00 to 95, a forecast for the pv produced on site where each subsequent value is 15 minutes later. pv_00 is a forecast for the pv production during the next 15 minutes. These values are forecasts, and so will not exactly match the actual_pv at that timestep.

Besides files per sites, one file contains the unique historical photovoltaic production profile and forecasts employed in the data of all sites, after being rescaled appropriately. In this file, we have rescaled power values to [0,1] so that this data  can serve the modeling of a photovoltaic unit beyond the scope of the EMSx benchmark. Finally, this dataset contains a metadata file providing battery storage capacities in kWh, battery flow capacities for 15 minutes in kWh, and battery efficiency coefficients, designed for the EMSx benchmark.

This dataset contains information from Schneider's Microgrid Energy Management Benchmark, which is made available under the Open Database License (ODbL).