add distribution grid losses

[lisazeyen]

Describe the feature you'd like to see

Please give a clear and concise description and provide context why the feature would be useful. Also, we'd appreciate any implementation ideas and references you already have.

[fneum]

duplicate of #559 ?

[lisazeyen]

@Parisra can you add some sources

[Parisra]

Based on the ENTOSE 2022 report, World Bank data, and the 2nd CEER 2020 report, the power loss in the distribution grid for EU countries is between 2-14%, with an average of 7-8%: So, assuming a 5-7% power loss (assuming improvements in the future) for the distribution grid links would be the simplest way to model these losses. However, there could be some advantages in having a unique efficiency for each country. The most important effect of including distribution losses when modeling the European energy system is increasing the cost-efficiency of distributed generation sources like rooftop PV [This is investigated in detail in this paper]. This means that if we model Germany with a 4% power loss and Poland with 7%, rooftop PV would be more cost-efficient in Poland than Germany. This result might be negligible though, if the rooftop PV capacity is small compared to other system components, meaning the simple approach would be preferred.
I can implement the country-specific distribution grid efficiency and do a PR if this is the preferred solution. This could be a constant efficiency, or one that is improved as time passes (either at a constant rate or a rate calculated based on each country's historical data).

[lisazeyen]

@Parisra If there is data for country specific losses this would be great and then adding a switch in the config where one can choose to set homogenous losses in all countries or of the country-specific once. Maybe you could add as a first step the constant efficiency per country and then one could think about a second PR to add improvement of the efficiencies.

[fneum]

# 2nd CEER Report on Power Losses; https://www.ceer.eu/documents/104400/-/-/fd4178b4-ed00-6d06-5f4b-8b87d630b060; Table 22. Losses in distribution in %
Country,2010,2011,2012,2013,2014,2015,2016,2017,2018
AT,3.92,3.76,3.44,3.35,3.26,3.22,3.00,2.95,2.88
BA,13.33,11.85,12.41,11.42,10.68,10.45,10.16,9.79,9.26
BE,4.54,4.36,4.58,4.41,4.23,4.36,4.35,4.40,4.27
CY,-,-,-,-,-,-,-,2.93,2.11
CZ,6.35,6.15,5.72,5.62,5.31,5.29,5.25,5.27,5.22
DE,10.57,-,-,-,-,-,-,-,3.15
DK,-,-,-,-,3.76,3.77,3.63,3.38,3.53
EE,6.72,5.93,5.78,5.35,5.65,5.03,4.42,4.95,4.55
EL,5.39,6.91,7.10,5.88,8.13,8.80,9.75,9.33,-
ES,7.51,7.54,8.33,8.67,8.70,8.63,8.48,8.10,8.47
FI,2.15,2.59,2.60,2.56,2.55,2.66,2.92,3.27,3.32
FR,-,-,-,-,-,5.99,5.98,6.02,6.01
GB,-,-,-,-,-,-,-,-,6.70
GE,-,10.31,8.62,7.84,7.38,7.76,7.71,7.39,6.80
HR,8.30,7.83,8.37,8.79,8.14,8.06,7.64,8.04,7.69
HU,9.05,8.81,8.75,8.69,8.52,8.36,8.08,7.84,7.36
IE,-,-,6.84,6.77,6.73,6.62,6.62,6.70,6.70
IT,-,-,-,-,-,-,-,5.23,-
KS*,41.22,38.15,36.69,35.54,33.50,31.82,29.69,29.30,27.91
LT,9.06,8.17,7.76,7.42,7.31,6.76,6.49,6.14,5.96
LU,-,-,-,-,-,-,10.48,8.87,8.37
LV,6.36,6.36,5.91,4.95,4.77,4.58,4.91,4.63,4.43
MD,-,-,-,10.46,9.35,8.42,8.34,8.30,8.31
ME,19.99,19.19,20.84,18.96,17.65,17.11,15.61,14.96,13.83
MK,17.16,18.11,17.38,16.42,15.50,14.76,14.67,14.44,14.18
MT,11.79,11.82,13.54,13.08,9.50,6.74,5.08,4.34,4.68
NL,4.97,4.59,4.47,4.75,4.48,4.26,4.17,4.43,4.38
NO,-,-,-,2.80,2.62,2.68,2.87,2.62,2.59
PL,6.56,5.74,5.75,5.71,5.46,5.32,4.73,4.56,4.45
PT,7.47,7.19,8.37,10.09,9.35,8.90,8.66,9.08,8.74
RS,16.28,15.51,15.16,14.91,14.36,14.43,12.99,12.96,12.20
SE,4.19,4.03,3.86,3.77,3.67,3.64,3.63,3.59,3.49
SI,5.83,4.94,5.25,4.94,4.76,4.74,4.56,4.28,4.18
SK,-,-,-,-,-,-,6.05,5.64,5.52
UA,10.73,10.44,10.54,10.17,10.08,9.79,9.90,9.89,9.84

[fneum]

xref #559

The option to insert electricity distribution network capacities and associated expansion costs were added in https://github.com/PyPSA/pypsa-eur-sec/commit/9b326ba493c27191896e57bd6fdf930b3b354b20. Dealing with the losses is another matter, since I think they're already incorporated in the existing electricity load, so would have to be deducted and then added back.

So the electricity load we take from ENTSO-E statistics should be reduced since they already include the losses:

https://nbviewer.org/github/Open-Power-System-Data/datapackage_timeseries/blob/2020-10-06/main.ipynb

[fneum]

I implemented a simplified version of the distribution grid losses which we can refine in 7ce15931.

• It assumes 3% distribution grid losses (on the lower end of the spectrum)
• It is not yet country-specific.
• It does not include progressions by planning horizon.