Home Electrical Appliances Recordings for NILM
- Citation Author(s):
-
Sarra Houidi (IBISC - Univ. Evry)Dominique Fourer
(IBISC - Univ. Evry)
Francois Auger
Houda Ben Attia SethomLaurence Miegeville - Submitted by:
- Dominique Fourer
- Last updated:
- DOI:
- 10.21227/ww76-d733
- Data Format:
1474 views
- Categories:
- Keywords:
Abstract
We introduce a novel dataset containing a total of 61 distinct HEAs. The proposed appliances (e.g. fans, fridges, washers, etc.) are of different kinds, ages, brands and power
levels. They have been recorded in steady-state conditions in a French 50 Hz electrical grid. The measurement setup consists of an AC current probe (E3N Chauvin Arnoux) with a 10 mV/A sensitivity and a differential voltage probe with
a 1/100 attenuation. Voltage and current waveforms were captured by an 8-bit resolution digital oscilloscope (RIGOL DS1104Z) each corresponding to an acquisition of 8 periods of the utility frequency. The sampling rate was set at Fe=250 kHz for some of the recordings and Fe = 50 kHz for the other part. Each HEA is described by 8 periods of the current and voltage steady-state waveforms, resulting consequently in a set of n = 8 * 61 = 488 distinct individuals.
Instructions:
Each HEA is stored both in the .csv and .mat file formats (8 measurements per HEA). Each .mat file associated to a unique HEA measurement containing the current and the voltage signals stored in a vector of length N = 2000 samples.
It also contains the related information about the sampling frequency Fe. A total of 90 electrical features are also stored in a real-valued vector that can be used for machine learning simulations applied to non-intrusive load monitoring scenarios.
