Temperature correction
Please refer to the module documentation for information on resource URLs and UUIDs
Temperature during the year can influence general power consumption and is dependent on the scenario during simulation. Divergence from the average temperature could have a correlation to higher or lower load due to power used for heating. This connection can be expressed in the Trident system in two ways:
Preprocessing the effect from the temperature and supplying scenario dependent load series, prebaking the effect. This can also apply to other effects such as wind and photovoltaic, so make sure to track the changes to the load series. The LTMv10 model loader features loading pre-processed temperature corrections via the preprocess_firm_load_with_temperature_correction option. This option requires the file Dellast_data.h5 to be present.
This option does not use this module. Using this option and loading the pre-processed load series is recommended should the option be available.
Calculating the effect on the load series during simulation. This option uses input data for the scenario dependent temperature series for each area, in addition to a temperature-change function in the form of a XY time series. Can be loaded from a ltmv10 data set from .ARCH files and historical.h5.
In order to use this option, this module must be added to the run config. Add the module to the module list when running Trident. See module documentation for UUID
Input to the module
When using the module certain input data must be supplied. Scenario dependent time series for temperature, temperature function for correction and load parts set up for temperature corrections. See module documentation for resource URLs.
Temperature series, usually connected to a single area. In degrees celsius and must span the set of scenarios used. Temperature series may be shared between load parts. Each load part may have more than one correction series.
Curves for corrections of load. XY time series detailing a relationship between percentage of correction per degree difference for the Y-axis, and the average temperature (centigrade) at that time for the X-axis.
Weight of the series used. Factor between 0 and 1 that is multiplied by the calculated correction amount. Weights is how to use more than one temperature series with different contribution to the temperature at the load part.
Ceiling in degrees celsius. This sets the upper ceiling for temperatures when performing the comparison between average and scenario temperature.
How the module works:
Average temperature calculation
The temperature module perform temperature correction of loads depending on the difference between scenario temperature and average temperature per area.
The average temperature is calculated from all available scenario temperatures, using the historical time series for temperature. For each area, for each series within the area, for each sample of the temperature series, the value is added to a vector of length 365. The index is calculated from the ISO day of the year based on the timestamp of the sampled temperature.
The accumulated sampled temperature is then averaged based on the number of samples. The temperature series is saved per area with timestamps from the beginning of 1970 as a default. See module documentation for output URL of this time series.
Should an area have no associated temperature series, the area is skipped and no calculations done.
Leap years
For years that are leap years, the 59th day of the year is skipped. After that, each day is indexed as the day before, skipping the leap year temperature.
For each year, index 366 is sampled from the 1st of the next year. Should the scenario year be a leap year, December 31st will be sampled from this 366st value.
Correction of load
During simulation, the load parts for each area is modified per week. Each load part may have N load correction sets. Each set is composed of the following:
Temperature series
Correction curve
Weight
For each temperature series, the corresponding temperature and weight values are used for the load corrections. For each time point, the historical tempearture series is shifted by year so that the start of the year of the historical temperature series correspond to the start of the simulation year.
The correction curve is sampled using the average temperature, and used as a factor to find the correction. This factor is \( S(T_{avg}) \)
Where LoadDiff is the difference in load, the correction, in GWh. \( T_{avg} \) is the average temperature, \( T_{scenario, i} \) the scenario temperature.
In the load module, this correction is subtracted from the right hand side of the power balance. This correction is stored to the result file, see module documentation for the resource URL.
Should either the average or the scenario temperature exceed the ceiling value they are clipped to the ceiling, meaning that should both values be above the ceiling the difference would be 0.