A Solar Resource Classification Algorithm for Global Horizontal Irradiance Time Series Based on Frequency Domain Analysis

(DOI: 10.1063/5.0045032)
A solar resource classification method based on frequency domain theory and the clearness index.

Carmen Lewis, Johann M Strauss, Arnold Johan Rix (Department of Electrical and Electronic Engineering, Stellenbosch University, South Africa)

Code

This classification method is based on minute resolution data.
May be used for either timezone-naive datasets such as SAURAN or timezone-aware datasets such as BSRN.

main.py: Main function for 15-minute period solar resource classification method.

• Set LOCATION parameters (latitude, longitude, timezone, altitude, name) via pvlib.location.Location
• Set FILEPATH to .csv/.tab file
• Set DATASOURCE to either 'BSRN' or 'SAURAN' to select pre-set date formats for .csv or .tab
• Set YEAR, MONTH, DAY_F for plots

fft_sd.py: Function for calculating the fast Fourier Transform and sample standard deviation.

kt.py: Function for calculating the clearness index.

plots.py: Plot functions for 15-minute period solar resource classification method.

Notes

pvlib and pandas python libraries are used. Please see library specific links for more information on converting timezones, calculating solar position and general dataframe manipulation.

Output Examples

Stellenbosch University, Western Cape, South Africa (SAURAN)

De Aar, Northern Cape, South Africa (BSRN)