python

Solar Power Forecasting

Two scripts can be executed to get an estimation of the solar power that will be produced tomorrow in the province of Liège:

• solar_panelwise.py: estimate tomorrow's production by computing a power curve for each panel installation detected by our panel enumerator.
• solar_provincial.py: estimate tomorrow's production by computing a power curve directly for the whole province, using Elia's production measurements for the past 7 days to build a posterior predictive model.

To evaluate the performance of the models on a given period, one can run misc/solar_panelwise_eval.py as well as misc/solar_provincial_eval.py.

Both caches present in the computational flow can be updated using the following scripts:

Cache	Script to use
SOLAR POWER MEASURES	cache/power.py
IRRADIANCE MEASURES	cache/solar_weather.py
IRRADIANCE FORECASTS	cache/solar_weather.py

Wind Power Forecasting

The interface is composed of:

• wind_eval.py [-h] [--forecast]: evaluate the current cached model on the validation set, or the test set if the --forecast option is used.
• wind_forecast.py [-h] [--display]: forecast the wind power production for today and tomorrow, using live weather prediction.

The different caches of the computation flow can be updated using the following scripts:

Cache	Script to use
FARMS LOCATIONS	cache/wind_farms.py
WIND POWER MEASURES	cache/power.py
WIND WEATHER MEASURES	cache/wind_weather.py [--start START] [--end END]
WIND WEATHER FORECAST	cache/wind_weather.py --forecast
LEARNING SET	cache/wind_ls.py
TEST SET	cache/wind_ls.py --test
MODEL	cache/wind_model.py [--train-set-only]

The wind power forecast computation flow is display hereafter.