diff --git a/doc/configtables/load.csv b/doc/configtables/load.csv index 6e98f881..ac666947 100644 --- a/doc/configtables/load.csv +++ b/doc/configtables/load.csv @@ -1,5 +1,4 @@ ,Unit,Values,Description -power_statistics,bool,"{true, false}",Whether to load the electricity consumption data of the ENTSOE power statistics (only for files from 2019 and before) or from the ENTSOE transparency data (only has load data from 2015 onwards). interpolate_limit,hours,integer,"Maximum gap size (consecutive nans) which interpolated linearly." time_shift_for_large_gaps,string,string,"Periods which are used for copying time-slices in order to fill large gaps of nans. Have to be valid ``pandas`` period strings." manual_adjustments,bool,"{true, false}","Whether to adjust the load data manually according to the function in :func:`manual_adjustment`." diff --git a/doc/release_notes.rst b/doc/release_notes.rst index ee7bd64b..fce4ae1b 100644 --- a/doc/release_notes.rst +++ b/doc/release_notes.rst @@ -10,6 +10,9 @@ Release Notes Upcoming Release ================ +* Merged two OPSD time series data versions into such that the option ``load: + power_statistics:`` becomes superfluous and was hence removed. + * Add new default to overdimension heating in individual buildings. This allows them to cover heat demand peaks e.g. 10% higher than those in the data. The disadvantage of manipulating the costs is that the capacity is then not quite diff --git a/doc/retrieve.rst b/doc/retrieve.rst index 06a07441..e4800fd2 100644 --- a/doc/retrieve.rst +++ b/doc/retrieve.rst @@ -91,7 +91,7 @@ None. **Outputs** -- ``resources/load_raw.csv`` +- ``resources/electricity_demand.csv`` Rule ``retrieve_cost_data`` diff --git a/rules/build_electricity.smk b/rules/build_electricity.smk index 89f4f736..f05e18c0 100644 --- a/rules/build_electricity.smk +++ b/rules/build_electricity.smk @@ -24,9 +24,9 @@ rule build_electricity_demand: countries=config["countries"], load=config["load"], input: - ancient(RESOURCES + "load_raw.csv"), + ancient("data/electricity_demand_raw.csv"), output: - RESOURCES + "load.csv", + RESOURCES + "electricity_demand.csv", log: LOGS + "build_electricity_demand.log", resources: @@ -417,7 +417,7 @@ rule add_electricity: if config["conventional"]["dynamic_fuel_price"] else [] ), - load=RESOURCES + "load.csv", + load=RESOURCES + "electricity_demand.csv", nuts3_shapes=RESOURCES + "nuts3_shapes.geojson", ua_md_gdp="data/GDP_PPP_30arcsec_v3_mapped_default.csv", output: diff --git a/rules/retrieve.smk b/rules/retrieve.smk index 2980583f..acb8a1c1 100644 --- a/rules/retrieve.smk +++ b/rules/retrieve.smk @@ -188,27 +188,17 @@ if config["enable"]["retrieve"]: if config["enable"]["retrieve"]: rule retrieve_electricity_demand: - input: - HTTP.remote( - "data.open-power-system-data.org/time_series/{version}/time_series_60min_singleindex.csv".format( - version=( - "2019-06-05" - if config["snapshots"]["end"] < "2019" - else "2020-10-06" - ) - ), - keep_local=True, - static=True, - ), + params: + versions=["2019-06-05", "2020-10-06"], output: - RESOURCES + "load_raw.csv", + "data/electricity_demand_raw.csv", log: LOGS + "retrieve_electricity_demand.log", resources: mem_mb=5000, retries: 2 - run: - move(input[0], output[0]) + script: + "../scripts/retrieve_electricity_demand.py" if config["enable"]["retrieve"]: diff --git a/scripts/add_electricity.py b/scripts/add_electricity.py index ab97dcd0..614e3330 100755 --- a/scripts/add_electricity.py +++ b/scripts/add_electricity.py @@ -52,7 +52,7 @@ Inputs :scale: 34 % - ``data/geth2015_hydro_capacities.csv``: alternative to capacities above; not currently used! -- ``resources/load.csv`` Hourly per-country load profiles. +- ``resources/electricity_demand.csv`` Hourly per-country electricity demand profiles. - ``resources/regions_onshore.geojson``: confer :ref:`busregions` - ``resources/nuts3_shapes.geojson``: confer :ref:`shapes` - ``resources/powerplants.csv``: confer :ref:`powerplants` diff --git a/scripts/build_electricity_demand.py b/scripts/build_electricity_demand.py index a08055ba..5d013065 100755 --- a/scripts/build_electricity_demand.py +++ b/scripts/build_electricity_demand.py @@ -1,15 +1,13 @@ # -*- coding: utf-8 -*- -# SPDX-FileCopyrightText: : 2020 @JanFrederickUnnewehr, The PyPSA-Eur Authors +# SPDX-FileCopyrightText: : 2020 @JanFrederickUnnewehr, 2020-2024 The PyPSA-Eur Authors # # SPDX-License-Identifier: MIT """ -This rule downloads the load data from `Open Power System Data Time series. - +This rule downloads the load data from `Open Power System Data Time series `_. For all countries in -the network, the per country load timeseries with suffix -``_load_actual_entsoe_transparency`` are extracted from the dataset. After -filling small gaps linearly and large gaps by copying time-slice of a given -period, the load data is exported to a ``.csv`` file. +the network, the per country load timeseries are extracted from the dataset. +After filling small gaps linearly and large gaps by copying time-slice of a +given period, the load data is exported to a ``.csv`` file. Relevant Settings ----------------- @@ -19,9 +17,7 @@ Relevant Settings snapshots: load: - interpolate_limit: - time_shift_for_large_gaps: - manual_adjustments: + interpolate_limit: time_shift_for_large_gaps: manual_adjustments: .. seealso:: @@ -31,12 +27,12 @@ Relevant Settings Inputs ------ -- ``resources/load_raw.csv``: +- ``data/electricity_demand_raw.csv``: Outputs ------- -- ``resources/load.csv``: +- ``resources/electricity_demand.csv``: """ import logging @@ -49,7 +45,7 @@ from pandas import Timedelta as Delta logger = logging.getLogger(__name__) -def load_timeseries(fn, years, countries, powerstatistics=True): +def load_timeseries(fn, years, countries): """ Read load data from OPSD time-series package version 2020-10-06. @@ -62,29 +58,15 @@ def load_timeseries(fn, years, countries, powerstatistics=True): File name or url location (file format .csv) countries : listlike Countries for which to read load data. - powerstatistics: bool - Whether the electricity consumption data of the ENTSOE power - statistics (if true) or of the ENTSOE transparency map (if false) - should be parsed. Returns ------- load : pd.DataFrame Load time-series with UTC timestamps x ISO-2 countries """ - logger.info(f"Retrieving load data from '{fn}'.") - - pattern = "power_statistics" if powerstatistics else "transparency" - pattern = f"_load_actual_entsoe_{pattern}" - - def rename(s): - return s[: -len(pattern)] - return ( pd.read_csv(fn, index_col=0, parse_dates=[0], date_format="%Y-%m-%dT%H:%M:%SZ") .tz_localize(None) - .filter(like=pattern) - .rename(columns=rename) .dropna(how="all", axis=0) .rename(columns={"GB_UKM": "GB"}) .filter(items=countries) @@ -149,17 +131,18 @@ def copy_timeslice(load, cntry, start, stop, delta, fn_load=None): ].values elif fn_load is not None: duration = pd.date_range(freq="h", start=start - delta, end=stop - delta) - load_raw = load_timeseries(fn_load, duration, [cntry], powerstatistics) + load_raw = load_timeseries(fn_load, duration, [cntry]) load.loc[start:stop, cntry] = load_raw.loc[ start - delta : stop - delta, cntry ].values -def manual_adjustment(load, fn_load, powerstatistics, countries): +def manual_adjustment(load, fn_load, countries): """ Adjust gaps manual for load data from OPSD time-series package. - 1. For the ENTSOE power statistics load data (if powerstatistics is True) + 1. For years later than 2015 for which the load data is mainly taken from the + ENTSOE power statistics Kosovo (KV) and Albania (AL) do not exist in the data set. Kosovo gets the same load curve as Serbia and Albania the same as Macdedonia, both scaled @@ -167,7 +150,8 @@ def manual_adjustment(load, fn_load, powerstatistics, countries): IEA Data browser [0] for the year 2013. - 2. For the ENTSOE transparency load data (if powerstatistics is False) + 2. For years earlier than 2015 for which the load data is mainly taken from the + ENTSOE transparency platforms Albania (AL) and Macedonia (MK) do not exist in the data set. Both get the same load curve as Montenegro, scaled by the corresponding ratio of total energy @@ -183,9 +167,6 @@ def manual_adjustment(load, fn_load, powerstatistics, countries): ---------- load : pd.DataFrame Load time-series with UTC timestamps x ISO-2 countries - powerstatistics: bool - Whether argument load comprises the electricity consumption data of - the ENTSOE power statistics or of the ENTSOE transparency map load_fn: str File name or url location (file format .csv) @@ -195,88 +176,72 @@ def manual_adjustment(load, fn_load, powerstatistics, countries): Manual adjusted and interpolated load time-series with UTC timestamps x ISO-2 countries """ - if powerstatistics: - if "MK" in load.columns: - if "AL" not in load.columns or load.AL.isnull().values.all(): - load["AL"] = load["MK"] * (4.1 / 7.4) - if "RS" in load.columns: - if "KV" not in load.columns or load.KV.isnull().values.all(): - load["KV"] = load["RS"] * (4.8 / 27.0) - copy_timeslice( - load, "GR", "2015-08-11 21:00", "2015-08-15 20:00", Delta(weeks=1) - ) - copy_timeslice( - load, "AT", "2018-12-31 22:00", "2019-01-01 22:00", Delta(days=2) - ) - copy_timeslice( - load, "CH", "2010-01-19 07:00", "2010-01-19 22:00", Delta(days=1) - ) - copy_timeslice( - load, "CH", "2010-03-28 00:00", "2010-03-28 21:00", Delta(days=1) - ) - # is a WE, so take WE before - copy_timeslice( - load, "CH", "2010-10-08 13:00", "2010-10-10 21:00", Delta(weeks=1) - ) - copy_timeslice( - load, "CH", "2010-11-04 04:00", "2010-11-04 22:00", Delta(days=1) - ) - copy_timeslice( - load, "NO", "2010-12-09 11:00", "2010-12-09 18:00", Delta(days=1) - ) - # whole january missing - copy_timeslice( - load, - "GB", - "2010-01-01 00:00", - "2010-01-31 23:00", - Delta(days=-365), - fn_load, - ) - # 1.1. at midnight gets special treatment - copy_timeslice( - load, - "IE", - "2016-01-01 00:00", - "2016-01-01 01:00", - Delta(days=-366), - fn_load, - ) - copy_timeslice( - load, - "PT", - "2016-01-01 00:00", - "2016-01-01 01:00", - Delta(days=-366), - fn_load, - ) - copy_timeslice( - load, - "GB", - "2016-01-01 00:00", - "2016-01-01 01:00", - Delta(days=-366), - fn_load, - ) - - else: + if "AL" not in load and "AL" in countries: if "ME" in load: - if "AL" not in load and "AL" in countries: - load["AL"] = load.ME * (5.7 / 2.9) - if "MK" not in load and "MK" in countries: + load["AL"] = load.ME * (5.7 / 2.9) + elif "MK" in load: + load["AL"] = load["MK"] * (4.1 / 7.4) + + if "MK" in countries: + if "MK" not in load or load.MK.isnull().sum() > len(load) / 2: + if "ME" in load: load["MK"] = load.ME * (6.7 / 2.9) - if "BA" not in load and "BA" in countries: - load["BA"] = load.HR * (11.0 / 16.2) - copy_timeslice( - load, "BG", "2018-10-27 21:00", "2018-10-28 22:00", Delta(weeks=1) - ) - copy_timeslice( - load, "LU", "2019-01-02 11:00", "2019-01-05 05:00", Delta(weeks=-1) - ) - copy_timeslice( - load, "LU", "2019-02-05 20:00", "2019-02-06 19:00", Delta(weeks=-1) - ) + + if "BA" not in load and "BA" in countries: + if "ME" in load: + load["BA"] = load.HR * (11.0 / 16.2) + + if "KV" not in load or load.KV.isnull().values.all(): + if "RS" in load: + load["KV"] = load["RS"] * (4.8 / 27.0) + + copy_timeslice(load, "GR", "2015-08-11 21:00", "2015-08-15 20:00", Delta(weeks=1)) + copy_timeslice(load, "AT", "2018-12-31 22:00", "2019-01-01 22:00", Delta(days=2)) + copy_timeslice(load, "CH", "2010-01-19 07:00", "2010-01-19 22:00", Delta(days=1)) + copy_timeslice(load, "CH", "2010-03-28 00:00", "2010-03-28 21:00", Delta(days=1)) + # is a WE, so take WE before + copy_timeslice(load, "CH", "2010-10-08 13:00", "2010-10-10 21:00", Delta(weeks=1)) + copy_timeslice(load, "CH", "2010-11-04 04:00", "2010-11-04 22:00", Delta(days=1)) + copy_timeslice(load, "NO", "2010-12-09 11:00", "2010-12-09 18:00", Delta(days=1)) + # whole january missing + copy_timeslice( + load, + "GB", + "2010-01-01 00:00", + "2010-01-31 23:00", + Delta(days=-365), + fn_load, + ) + # 1.1. at midnight gets special treatment + copy_timeslice( + load, + "IE", + "2016-01-01 00:00", + "2016-01-01 01:00", + Delta(days=-366), + fn_load, + ) + copy_timeslice( + load, + "PT", + "2016-01-01 00:00", + "2016-01-01 01:00", + Delta(days=-366), + fn_load, + ) + copy_timeslice( + load, + "GB", + "2016-01-01 00:00", + "2016-01-01 01:00", + Delta(days=-366), + fn_load, + ) + + copy_timeslice(load, "BG", "2018-10-27 21:00", "2018-10-28 22:00", Delta(weeks=1)) + copy_timeslice(load, "LU", "2019-01-02 11:00", "2019-01-05 05:00", Delta(weeks=-1)) + copy_timeslice(load, "LU", "2019-02-05 20:00", "2019-02-06 19:00", Delta(weeks=-1)) if "UA" in countries: copy_timeslice( @@ -297,14 +262,13 @@ if __name__ == "__main__": configure_logging(snakemake) - powerstatistics = snakemake.params.load["power_statistics"] interpolate_limit = snakemake.params.load["interpolate_limit"] countries = snakemake.params.countries snapshots = pd.date_range(freq="h", **snakemake.params.snapshots) years = slice(snapshots[0], snapshots[-1]) time_shift = snakemake.params.load["time_shift_for_large_gaps"] - load = load_timeseries(snakemake.input[0], years, countries, powerstatistics) + load = load_timeseries(snakemake.input[0], years, countries) if "UA" in countries: # attach load of UA (best data only for entsoe transparency) @@ -321,7 +285,7 @@ if __name__ == "__main__": load["MD"] = 6.2e6 * (load_ua / load_ua.sum()) if snakemake.params.load["manual_adjustments"]: - load = manual_adjustment(load, snakemake.input[0], powerstatistics, countries) + load = manual_adjustment(load, snakemake.input[0], countries) if load.empty: logger.warning("Build electricity demand time series is empty.") diff --git a/scripts/retrieve_electricity_demand.py b/scripts/retrieve_electricity_demand.py new file mode 100644 index 00000000..a8a44b68 --- /dev/null +++ b/scripts/retrieve_electricity_demand.py @@ -0,0 +1,46 @@ +# -*- coding: utf-8 -*- +# SPDX-FileCopyrightText: 2023-2024 The PyPSA-Eur Authors +# +# SPDX-License-Identifier: MIT +""" +Retrieve electricity prices from OPSD. +""" + +import logging + +import pandas as pd + +logger = logging.getLogger(__name__) + +from _helpers import configure_logging + +if __name__ == "__main__": + if "snakemake" not in globals(): + from _helpers import mock_snakemake + + snakemake = mock_snakemake("retrieve_electricity_demand") + rootpath = ".." + else: + rootpath = "." + configure_logging(snakemake) + + url = "https://data.open-power-system-data.org/time_series/{version}/time_series_60min_singleindex.csv" + + df1, df2 = [ + pd.read_csv(url.format(version=version), index_col=0) + for version in snakemake.params.versions + ] + combined = pd.concat([df1, df2[df2.index > df1.index[-1]]]) + + pattern = "_load_actual_entsoe_transparency" + transparency = combined.filter(like=pattern).rename( + columns=lambda x: x.replace(pattern, "") + ) + pattern = "_load_actual_entsoe_power_statistics" + powerstatistics = combined.filter(like=pattern).rename( + columns=lambda x: x.replace(pattern, "") + ) + + res = transparency.fillna(powerstatistics) + + res.to_csv(snakemake.output[0])