Synthetic data of the paper: “Quantification of feature importance in automatic classification of power quality distortions”
- Citation Author(s):
-
Raul Igual
(Universidad de Zaragoza)
Seyedmahdi Miraftabzadeh (Politecnico di Milano)Federica Foiadelli
(Politecnico di Milano)
Carlos Medrano
(Universidad de Zaragoza)
- Submitted by:
- Raul Igual
- Last updated:
- DOI:
- 10.21227/bt55-nd11
- Data Format:
1291 views
- Categories:
- Keywords:
Abstract
This dataset is related to the paper “Quantification of feature importance in automatic classification of power quality distortions” (IEEE International Conference on Harmonics and Quality of Power, March 2020). It includes the features extracted from synthetic signals with power quality distortions obtained from a public model (doi: 10.1109/ICHQP.2018.8378902).
The dataset includes data associated with 8 features (described in the paper above) obtained from signals processed with the Stockwell Transform. The distortions considered in the paper were: normal signal, sag, swell, interruption, oscillatory transient, harmonics, sag with harmonics, swell with harmonics and flicker. One hundred and fifty samples are included for each distortion and feature.
Different Signal-to-Noise Ratios (SNRs) are included in the dataset: noiseless, 50 dB, 40 dB, 30 dB and 20 dB.
Instructions:
The dataset is organized in 5 folders. Each folder contains data associated with a different SNR: noiseless, 50 dB, 40 dB, 30 dB and 20 dB. Data in different folders are not correlated. They are completely independent.
Each folder contains 9 CSV files. Each CSV file corresponds to a particular type of distortion: normal, sag, swell, interruption, oscillatory transient, harmonics, sag with harmonics, swell with harmonics and flicker.
Each file contains 8 columns. Each column contains data of a different feature: Column 1 is related to Feature 1, Column 2 is related to Feature 2 and so on. The features are numbered and described in the paper cited in the abstract of this dataset.
Each file also contains 150 rows. Each row is a different sample of the disturbances.
Data are not normalized.
-