The basic model forecasting renewable power generation was based primarily on weather forecasts at locations with the highest concentration of wind and photovoltaic farms. On the demand side, on the other hand, we have projected power demand, which also depends to some extent on the weather. Information on holidays is also important because then it […]
The basic model forecasting renewable power generation was based primarily on weather forecasts at locations with the highest concentration of wind and photovoltaic farms.
On the demand side, on the other hand, we have projected power demand, which also depends to some extent on the weather. Information on holidays is also important because then it „knows” why the demand for power will fall (e.g., due to a reduction in power consumption by companies). For some locations in Poland as well as abroad, it is also useful to add separately some weather data.
On the basis of the above-mentioned variables, a basic model was created to forecast the spot price of energy in Poland (TGEBase index forecast).
Among other things, the basic model has the advantage of being quite easy to interpret. The table below lists the explanatory variables of the basic model along with their impact on the energy price forecast. The model’s response [%] tells how much the energy price will change if the explanatory variable in the list changes by 1[%] and the other variables remain unchanged.
| Model variables : ID[59614] Model: RES vs TGEBase | Model response[%] |
| PSE PROGNOSIS of power generation from wind and photovoltaic sources Poland [MW] smoothed dz | -0.492850779 |
| Gdańsk – temperature min | -0.125198939 |
| Gdańsk – temperature max | -0.042660698 |
| Amsterdam – temperature min | -0.042007839 |
| TIME_Days_free_also_on_weekends_Poland | -0.00012186 |
| PSE PRONOZA Projected network demand | 0.076150017 |
| TGEBase course smoothed out | 0.35766362 |
Extending the basic model with several additional explanatory variables improved the accuracy of the model.
The additional variables took into account more information on, for example, the fossil fuel segment, or the costs associated with the need for power producers to purchase CO2 emission rights, as well as planned power unit shutdowns (total or partial). Thus, we had a forecast of the basic model and to it we attached as additional explanatory variables the aforementioned data series.
The next step was to forecast all the explanatory variables and use the resulting predictions in the energy price forecasting model. The model currently uses the following predictions:
Forecast generation from wind and photovoltaic sources (total generation):
Fossil generation forecast:
Forecasting the share of fossil sources in the energy mix:
Projected power demand:
TTF Natural Gas Price Forecast:
ARA coal price forecast:
CO2 emission rights price forecast:
The result is a TGEBase rate forecast based on the predictions mentioned above:
On average, between 2018 and 2022, a 1[%] increase in the share of RES generation in Poland’s energy mix resulted in a 0.229[%] decrease in energy price. An increase in the share of RES in the mix supports decreases in energy prices. The higher the share of RES in the energy mix, the greater its impact on the price of energy. Intuitively, increasing the share of RES in the mix supports declines in electricity prices. This is due to a reduction in the cost of energy production.
