irr-uncertainty 1.3

Irradiance uncertainty package

irr_uncertainty

This repo enables to generate irradiance (horizontal and plane-of-array) 95%-intervals anywhere in Europe with latitude and longitude between [35°,60°] and [-20°,40°] respectively, excluding atypical locations such as high-altitude snowy locations.

The intervals are generated with the basis of satellite (CAMS) data.

Package

This repo can be used as a package by running the following commands.

pip install irr_uncertainty

Example

The following lines generate Monte Carlo simulations to obtain the 95% interval for one orientation.

import pandas as pd

from irr_uncertainty.data.irr_data import cams_data_pvlib
from irr_uncertainty.data.solar_data import solarpos
from irr_uncertainty.models.uncertainty_model import irrh_scenarios, transpo_scenarios

start = pd.to_datetime("20220812").tz_localize("UTC")
end = pd.to_datetime("20220816").tz_localize("UTC")
n_scenarios = 1000

# Grenoble, FRANCE
lat = 45.16
long = 5.72
alt = 212

# Installation plan
tilt = 25
azimuth = 180

# Fetch CAMS data and get hourly solar position (with same convention)
sat_data = cams_data_pvlib(lat, long, alt, start, end)
solar_position = solarpos(sat_data.index, lat, long, alt).shift(-1)  

# Compute Monte Carlo simulations for horizontal plans
ghi_scns, dhi_scns, bhi_scns = irrh_scenarios(lat, long, alt, solar_position, sat_data["ghi"],
                                              n_scenarios=n_scenarios)
											  

# Generate Monte Carlo simulations for tilted plans
poa_scns_s, _, _, _ = \
    transpo_scenarios(tilt, azimuth, lat, long, alt, solar_position, ghi_scns, dhi_scns, n_scenarios=n_scenarios)

# Compute 95% interval bounds
q_95 = poa_scns_s.quantile([0.025, 0.975], axis=1).T

Then, typical intervals can be computed with the quantiles as in the Figure below.

Mandatory credentials

A secret file "secret.ini" should be placed in the "data/" folder with the credentials for:

• Soda-pro: for the CAMS data - to generate 95% interval for any location.
• BSRN with a email request - to access BSRN station data and recreate the methodology Figures.

Command files

Two command files ease the creation of virtual environment and the execution of jupyter notebooks:

• create_env.cmd: Create a virtual environment and installed all the required packages
• notebook_start.cmd: Create a kernel to link with the virtual environment in order to use within the notebook and open the jupyter notebook

Setup

Python 3.9 python docs

Contact

Created by @Alex - feel free to contact me for any concerns and happy to receive feedbacks !