Performance measures and availability of European power plants

Performance measures and availability of European power plants | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF data-descriptor Performance measures and availability of European power plants Eric Jahnke, Armin Ardone This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9138908/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Reliable representation of power plant availability is essential for energy system analysis, electricity market modelling and security-of-supply studies, yet openly accessible datasets are scarce and heterogeneous. This data set presents a harmonised set of performance measures for thermal and hydropower units in the ENTSO-E area derived from monthly unavailability reports on the ENTSO-E Transparency Platform. We reconstruct unit-level availability time series from event-based outage records, apply transparent filtering, clustering and interpolation procedures, and enhance the data with technology and commissioning information from merged open power plant databases. From these processed time series we compute hourly measures and a broad range of time-based performance indicators, including availability factors, outage factors and rates, failure and repair rates, and statistics on partial derations, in both capacity-weighted and unweighted form. The resulting indicators are provided by country, season, plant type and technology, and are validated against historical generation records, selected adequacy studies as well as other journal publications. The accompanying open-source Python workflow enables full reproducibility and direct reuse in data processing or modelling. In addition, two example algorithms are included that can be used to simulate outage series for individual generation units or aggregated power plant fleets. Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Figure 10 Figure 11 Figure 12 Figure 13 Background & Summary Availability data of power units is required for realistic models in the energy industry or electrical engineering for various applications e.g., to map supply shortages in electricity market models, to provide the basis for investment decisions based on security of supply in energy system models or to analyse contingency cases in power flow optimizations. On detailed technology base and for specific countries, licensed data is available 1 . However, publicly available data sources are scarce. This data set provides aggregated statistics and performance measures for European electricity generation technologies and fuel types. Information about the performance of units is based on the monthly unavailability reports of the ENTSO-E transparency platform 2 . Methodologically, the data preparation is comparable to other recent publications in this field 3 , 4 . A new feature is a reproducible Python pipeline to generate regular time series per unit, more optional aggregations as well as validation against different studies. The performance measures comply with the current IEEE-762 standard 5 and other KPIs used in practice 6 . The data set can thus be used directly for applications such as probabilistic outage sampling, the simulation of time series in different granularity or as constraints for the aforementioned optimization models. Methods Step 1 - Data source We use monthly CSV file exports of the ENTSO-E transparency platform (“ UnavailabilityOfGenerationUnits_15.1.A_B “). Additional columns of clear names of the ENTSO-E identification codes of bidding zones or control areas were created using mapping tables taken from the Python API for the ENTSO-E transparency data download 7 . The data only contains entries from generation units with a nominal capacity of at least 100 MW and outage reports with a change in capacity of at least 100 MW. Outages of consumption units or transmission lines are not included, as there are other data sources for this on the transparency platform 8 . In November 2025 when this study was written, the ENTSO-E updated its data access from the FTP server and, in the process, also changed the table formats of some data sources 9 . As the code and statistics were created when the unavailability forms were still in change and the final format unclear, an overview and example of the old and currently proposed new table format is provided in Table 1 . The time series for the outage notification is taken from the “StartOutage(UTC)” and “EndOutage(UTC)” columns. The columns “StartTimeSeries(UTC)” and “EndTimeSeries(UTC)” are ignored, as no further information is available for them. Old column name New column name Data type Description Example StartOutage StartOutage(UTC) Timestamp Start time of an outage event 2025-11-12 13:30:00.000 EndOutage EndOutage(UTC) Timestamp End time of an outage event 2025-11-12 23:30:00.000 StartTS StartTimeSeries(UTC) Timestamp Start of a change of the avail. cap. during an outage 2025-11-12 15:30:00.000 EndTS EndTimeSeries(UTC) Timestamp End of a change of the avail. cap. during an outage 2025-11-12 20:30:00.000 TimeZone TimeZone String Local time zone (e.g., “CET”) CET MRID InstanceCode String Unique identifier of outage event 123XYZ Status Status String Current status Cancelled Type Type String “Planned” or “Forced” outage event Forced AreaCode AreaCode String Unique identifier of market area 10YGB———-A AreaTypeCode AreaTypeCode String Bidding zone or control area (“BZN” or “CTA”) CTA AreaName AreaDisplayName String Name of BZN or CTA UK(National Grid) CTA MapCode AreaMapCode String Unique identifier of BZN or CTA GB PowerResourceEIC AssetCode String Unique identifier of the unit 48W0000ABC-15A UnitName AssetName String Name of the unit GBX-2 ProductionType ProductionType String Fuel or tech. type (e.g., “Nuclear”) Fossil Oil InstalledCapacity InstalledCapacity Float Installed capacity 200 AvailableCapacity AvailableCapacity[MW] Float Available capacity during the outage event 100 Version Version Integer Update version of the MRID 2 OldVersion OldVersion Boolean Indicator if the version is still the most current FALSE Reason Reason String Reason for the outage (e.g., “Failure”) Failure UpdateTime UpdateTime(UTC) Timestamp Update time of the outage event 2025-11-12 12:30:00.000 Table 1: Overview about raw monthly CSV file for unavailability of generation units from the old as well as newly updated format 2,8,10 . Step 2 - Data cleaning Not all single entries are valid anymore. If maintenance is planned in the medium term due to a technical problem that has occurred and this can be resolved in the short term, the original notification for future maintenance work is cancelled. As a result, the raw data set is filtered exclusively for reports that are labelled as “Active” and not “Withdrawn” or “Cancelled”. Regardless of their type, outage notifications have an identification code (“MRID”) for clear traceability with, for example, the additional ENTSO-E table for the reasons of the outage. However, as outages for planned maintenance work have to be postponed or ad-hoc messages for unplanned outages have to be readjusted at short notice due to technical problems, there are often duplicates for a specific identification code. Consequently, all entries that show an old version were filtered out. As there were still duplicates in the identifier, all reports were sorted by version and only the line with the highest version number was taken over. In this respect, the clean-up approach of Gjorgiev et al. 3 differs, as the entries with duplicate identifiers were not removed after repeated filtering, but were treated as different reports with more accidentally similar identifiers. In our approach, if there were still duplicates of the identifiers, the entry with the most recent update time was used 1 . All duplicated entries are collected during processing and can be written out as a CSV file. Some countries are missing entirely in the bidding zones (e.g., Germany). Since Germany is also divided into four balancing zone, there are four entries instead of one aggregated. For this reason, we do not yet filter specifically by bidding zone or control areas during data preparation, but instead specify the countries to be filtered by before calculating the performance indicators. There are mainly duplicates here because we did not filter bidding zones or control areas. An overview of the raw and cleaned dataset of reports used for this study is given in Table 2 : N Raw reports 12,042,695 Cancelled 967,496 Withdrawn 301,275 Old version 9,466,610 Active current version 1,307,314 Duplicated rows 287,742 Duplicated MRIDs 69,975 Active unique reports 949,597 Active filtered reports 947,061 2 Duplicated EIC codes 6 Long-ranging single outage clusters 315 Reports used for performance measures 946,740 Table 2: Overview of raw outage reports before and after cleaning and pre-processing Step 3 - Data pre-processing Bridging temporal gaps Sometimes there is a temporal gap of only a few hours between two outage reports, where it can be assumed that the generation unit was not running in normal operation and this gap evolved probably from a reporting error. For this reason, a time interval of τ hours can be defined in which two failure reports are linked to each other ( Figure 1 ). The possibility of defining a bridge parameter is optionally available in the accompanying code, whereby the value of the available capacity as well as the type and reason for the earlier outage are extended until the later outage. Optionally, parameters can also be set to link only short time deltas between reports of the same type and/or reason. A delta of eight hours was assumed for the calculation of the statistics for this data set, but even more hours would have been possible according to sensitivity analysis in other studies 3 . Clustering and Labelling Outage events for a generation unit often overlap in time. The problem here is that not only different available capacities are reported for a specific point in time, but the type and reason for the outage can also differ. In addition, the reports have different creation or update times, which makes it difficult to clearly identify which is the current report in the event of a temporal overlap. The possible reasons for the outage are divided by us into just two categories: “Maintenance” and “Other”. A more specific differentiation would not bring more benefit to the statistical analysis. Regarding the type, all reports distinguish between planned and forced outages. Officially, the report for a planned unavailability must be published to the transparency platform maximum one hour after the official confirmation by the TSO. Updates to this report (a newer version) must be published at the latest an hour after the information is known. Information on unplanned outages must be published no later than one hour after the change in available capacity 8 . In principle, this is in line with other institutions such as NERC, whose declarations between a long-term planned outage, planned maintenance or unplanned maintenance depend on the announcement time before the event and may have to be adjusted afterwards 11–13 . However, for the ENTSO-E unavailability data the difference between notifications for forced and planned is often unclear, as planned outages are only partially published with a long lead time 4 . In some cases, reports of planned maintenances don’t even exist, because every maintenance is reported as ‘forced’ (a more detailed analysis can be found in the supplementary materials ). In reality, the correct classification of outage types as forced or planned, as well as the separation of overlapping reports, depends on the type of power plant and specific background knowledge about the reasons for the outages in terms of expected repair times 11 . The labelling is therefore subject to a level of uncertainty that might bias the statistical analysis. Previous works used different heuristics and methods to cluster and label the reports. Schmitz et al. 14 and Finck 15 aggregate the (un-)availabilities of the reports regardless if it was planned or forced. Gils et al. 16 or Deakin et al. 17 differentiate only between planned and forced, but they ignore the outage reasons and do not explain their methodology for pre-processing the data. Bassini 4 characterises the type in the event of a temporal overlap of 3 notifications based on the majority vote. The procedure for an even number of overlapping events such as two, four or more is not described. We differentiate between the clustering of the overlapping outage reports with regard to the value of the available capacity, the label for the outage type, and the reason. For each time step with overlapping reports, we assign the value of the available capacity by the report with the most recent update timestamp. For the label of the outage type and reason, we assign the values for the earliest outage. For reports with identical outage starts, we again assign the label with the most recent update time. However, the same logic can be implemented for the labels as for the capacity using parameters in the code, so that the report with the latest update timestamp is always decisive. The separation of a non-availability event into separate events is implemented when the state transitions from a planned deration or outage to a forced deration or outage 3 . An example of three overlapping outage reports for a specific day with different update times and types is shown in Figure 2 , visualising the decisions of clustering and labelling the reports. Other cases can be found in the supplementary materials . States A generating unit can be either fully available, derated or fully unavailable. It is fully available if it can potentially produce electricity at its rated power. If a unit is in outage, it cannot generate any electricity at all. Derated means that a unit cannot operate at full capacity, but is not off the grid. Consequently, the possible states could be defined as planned outage, planned derated, unplanned outage, unplanned derated and available 3 . Since we record the outage type separately, we only define the primary states as “out”, “derated” and “available”. Final cleaning Since the outage data for bidding zones and control zones in countries with multiple zones (i.e., Norway and Italy) often contain overlapping entries and also show changes in the affiliation of individual power plants over time, the performance measures were only calculated for each country and duplicate entries of the EIC codes were removed beforehand. These duplicates are exported to a separate CSV file for overview purposes. Similarly, individual outages of the same type (“planned”) lasting longer than 365 days were assumed to be erroneous entries and were therefore removed. However, if two or more outages of different types (a “planned” outage followed immediately by an “unplanned” outage) exceed the defined duration, they are not removed, as they essentially represent two separate clusters 3 . All removed entries in this way are recorded and exported in a CSV file. Step 4 - Meta information for power units On the ENTSO-E Transparency Platform, thermal power plants in particular are only listed by fuel type and general plant types, but not broken down in more detail by technology type such as open-cycle gas turbine and combined-cycle gas turbine. The year of commissioning (or retrofit) of the units is also not available, nor is the information as to whether it is a CHP plant. We add this information from freely available data sources. To do this, we use the open units table from JRC 18 and a power plant list from the PyPSA modelling framework package powerplantmatching 19 . As the JRC table also contains the official EIC codes of ENTSO-E for power plant locations and the individual units, we can join the power plant lists from JRC and PyPSA together with ENTSO-E using these columns. A geographical overview of plant types, technologies and installed capacities can be seen in Figure 3 . Descriptive statistics on the meta-information of the power plants can be found in the supplementary materials . An example is given with Table 3 . Since the JRC and PyPSA power plant lists do not contain the relevant technology types or commissioning data for all units in the ENTSO-E tables, the characteristics of individual technology types are unfortunately sometimes very limited and less reliable. This applies in particular to thermal power plants that use natural gas, coal, or oil as fuel. For this reason, the results also include the values for NAs in the technologies. Column name Data type Description Example unit_eic string Unique identifier of the gen. unit 48W0000ABC-15A plant_eic string Unique identifier of the gen. plant (a plant can consist of multiple units) 48W0000ABC unit_name string Name of the unit GBX-2 plant_name string Name of the plant GBX plant_tech string Technology type CCGT fuel_type string Fuel type Fossil gas fuel_type_code string ENTSO-E code of the fuel type B04 set string Technology group CHP year_commissioned integer Year of first operation (or retrofit) 1995 year_commissioned integer Year of decommissioning 2040 unit_installed_capacity float Installed capacity (unit) 400 MW plant_installed_capacity float Installed capacity (plant) 800 MW country string Country location GB nuts2 string NUTS 2 code of the unit location GB5 lat Float Geographical latitude 51.50 lon float Geographical longitude -0.12 eic_match_source string Source column to merge PyPSA on JRC power plant databases plant_eic Table 3: Overview about the merged power plant list to enhance the statistical analysis. Data Records Repository Code and datasets are hosted online and can be downloaded from Zenodo . The code already contains implementations for calculating more aggregated KPIs by e.g., age, size, or CHP type of the units. The performance measures originate from the IEEE 762 standard 5 as well as frequently used statistics in practice and academic studies 6 . They primarily relate to the time-based availability of power generation units. As the publication of data of individual non-anonymised generation units of the transparency platform is not permitted, capacity-weighted aggregated performance factors were calculated. Although these do not allow the same level of detail as plant-specific data, they are still sufficient enough for simulations of power plant outages or for a rough estimate of the availability. The files are: ‘kpis_tech_ALL.csv’ ‘kpis_tech_MOY.csv’ ‘kpis_plant_ALL.csv’ ‘kpis_plant_MOY.csv’ The four files contain average annual and monthly performance measures for ENTSO-E plant types or technology types and cover the ENTSO-E area with the following countries: Albania (AL), Austria (AT), Belgium (BE), Bosnia-Herzegovina (BA), Bulgaria (BG), Switzerland (CH), Czechia (CZ), Germany (DE), Denmark (DK), Estonia (EE), Spain (ES), France (FR), Finland (FI), Great Britain (GB) 3 , Greece (GR), Croatia (HR), Hungary (HU), Ireland (IE), Italy (IT), Lithuania (LT), Latvia (LV), Luxembourg (LU), Moldova (MD), Montenegro (ME), North Macedonia (MK), Netherlands (NL), Norway (NO), Poland (PL), Portugal (PT), Romania (RO), Serbia (RS), Sweden (SE), Slovenia (SI), Slovakia (SK) and Kosovo (XK). The data is calculated as annual averages and seasonally disaggregated by month, among other things, because seasonal factors such as outside temperature, cooling water temperature, etc. change the likelihood of unplanned outages and therefore also TSOs schedule their planned maintenance according to the season 20–24 . The repository additionally contains the Python and R scripts for preparing, analysing, evaluating and visualising the data, as well as an example script for simulating further time series: ‘availability_data.py’ ‘availability_statistics.py’ ‘availability_simulation.py’ ‘powerplants_merge.py’ ‘plot_outages_vs_gen.py’ ‘validate_kpis.R’ ‘descriptive_stats_jrc_pypsa.R’ ‘plot_outage_statistics.R’ Tables The performance measures described above contain can be found with various additional prefixes or letter combinations as shown in Table 4 : Abbreviation Example w/u Weighted vs. Unweighted (E) Equivalent F/P/M/U/S Abbreviation for outage type O/D Total outage vs. derated Actual measure Rate, Factor, etc. Table 4: Structure of the name of a measure. Outage types were taken into account according to their type and reason 5 : U nplanned - all forced outage events including technical failures as well as maintenance work. S cheduled - all planned outage events combining maintenance work with other planned outages. M aintenance - calculated accordingly from events with the reason ‘maintenance’, whereby the type can be planned or forced. P lanned - calculated from reports containing planned outages where the reason is not maintenance. F orced - calculated from reports containing forced outages excluding maintenance. This non-standard with regard to the IEEE-762 standard summary of types and reasons was chosen because it simplifies further use in simulations. Aggregations of the performance indicators were calculated in two different ways: capacity-weighted (w) and unweighted (u). In capacity weighting, the figures are weighted according to the rated capacity of the respective generation unit. Since unplanned outages often only involve a reduction in output but do not require the power plant to be shut down completely, the equivalent values, which take into account both total outages and derations, are particularly useful here, in addition to the rates for total outages. For this reason all three variants have been calculated for factors and rates. The actual measures are briefly described in Table 5 , but a detailed overview can be found in the supplementary materials : Performance measure 4 Abbreviation Description Availability factor AF Fraction of time without outages Outage hours OH Cumulated hours in outages Outage factor OF Fraction of time in outages Outage rate OR Share of time in outage relative to operating time Repair rate RR Average time from start to end of an outage event Failure rate FR Event rate of an outage event Event counts N Number of single outage events Mean deration / Mean deration of partial outage events Standard deviation of derations / Standard deviation of derations of partial outage events Table 5: Overview about calculated performance measures independent of type and reason. Factors and rates for the specific outage types are the usual measures used to describe performance, but also for further use in simulations 25 . The repair time can be interpreted as the average duration of an event, i.e., how long maintenance takes on average in the event of an unplanned outage. The failure rate can be interpreted as the average duration between the occurrence of two outage events of a given type. Unless otherwise specified, all units describing time duration are calculated in hours. Rates and factors are relative and can be converted into percentages. Although the name suggests otherwise, the repair rate is also calculated in hours. The number of events N describes the frequency of individual events that have occurred, e.g., of type “forced”. The mean value and standard deviation of a performance limitation describe the first two statistical moments of the distribution in the event of partial outages of the respective outage type and are calculated in MW. Total outages were not taken into account, as these always correspond to the nominal power and would skew the value upwards 12 . With regard to the rates for planned or unplanned outages, etc., a simplification had to be made: according to the standard, the service times of a power plant are actually required to calculate rates such as the forced outage rate (FOR). These are determined from the active times during which the power plant is active minus reserve shutdowns. Those shutdowns describe the unavailability of a power plant, e.g., due to negative electricity prices, insufficient demand, or the impossibility of operating the power plant economically and are therefore also called ‘economical shutdowns’ 5 . Since these reserve shutdowns are not included in the outage time series and would have to be derived with uncertainty from e.g., historical market prices, they are not taken into account. Instead, service hours were calculated as active hours minus hours with total scheduled outages from (SOH). We therefore assume that a power plant was always potentially in service when it was not scheduled to be out of service and could theoretically produce electricity. The entire table for, e.g., ‘kpis_tech_ALL.csv’ contains 140 columns. An example excerpt can be seen in Table 6 : Column Abbreviation Description Country country ISO-A2 name of the country (e.g., “NO”, “DE”) ENTSO-E plant type plant_type Plant type of the ENTSO-E domain (e.g., “Biomass”) Active hours ACTH Cumulated number of hours of entries for a power plant (typically a whole year) Service hours SH Active time minus hours in total scheduled unavailability Observed hours DH Active time minus hours in total planned unavailability Forced outage hours FOH Cumulated number of hours during forced outages. Equivalent availability factor wEAF Capacity-weighted fraction of time without total outages or derations Forced outage factor uSOF Unweighted fractions of time with scheduled outages Unplanned deration factor wUDF Capacity-weighted fractions of time in unplanned deration Equivalent forced outage rate uEFOR Unweighted shares of time in forced outage relative to service hours Repair rate of forced outages wRR_FO_h Capacity-weighted average time from start to end of a forced outage event (in hours) Failure rate of maintenance outages uFR_MO_h Unweighted event rate of a maintenance outage event (in hours) Event counts for planned outages N_PO Cumulated number of single outage events for planned outages Mean deration for unplanned outages wEUD_mw_mean Capacity-weighted mean deration of unplanned partial outage events (in MW) Standard deviation of derations uEFD_mw_std Unweighted standard deviation of derations of forced partial outage events (in MW) Table 6: Overview about calculated performance measures for the data set. Technical Validation Historical generation To verify the aggregated generator outages, Figures 4-6 exemplarily shows curves of generation and available capacity relative to installed capacity for some countries of the dataset with their thermal and/or hydraulic plant types from ENTSO-E Transparency platform and available data for generation (“AggregatedGenerationPerType_16.1.B_C_r3”) as well as installed capacity (“InstalledGenerationCapacityAggregated_14.1.A_r3”) 26,27 . All data were aggregated to daily averages. The information on installed capacities is published annually and is therefore sometimes very inaccurate, especially for RES, because capacity is constantly being expanded during the year. Due to the restriction that only generation units with a minimum rated power of 100 MW should be included, there should be no outage reports for individual RES plants actually. However, the graph also shows restrictions for wind offshore. If generation is greater than the specified available capacity, there may be several reasons for this. On the one hand, the reported generation or the reported outage level could be incorrect, but on the other hand, there could also be an error in the clustering and processing of the outage reports for the specific case. The obvious caps on hard coal and natural gas in particular suggest errors in the raw data. Results for the remaining countries can be found in the supplementary materials . Other publications and studies The data set was compared with various other studies: the European Resource Adequacy Assessment ( ERAA ) 2023 28 and Ten Years Network Development Plan ( TYNDP ) 2024 29 by ENTSO-E, as well as the journal article by Gjorgiev et al 3 . A comparison with availability factors for German plant types from Bassini 4 can be found in the supplementary materials . Data sets are compared using absolute values of the KPIs. The ERAA data set was chosen because there is data modelled for the current year 2025. From the TYNDP study we took earliest data for 2030. With ERAA , we can compare forced outage rates unweighted and capacity-weighted by plant type from the ENTSO-E domain. Since the methodological reports did not provide any information on how the data was specifically determined, we compare it with FOR and UOR of our dataset, also taking forced maintenance into account. Regarding TYNDP , we can also compare planned outage days by converting our hourly results into days in addition to the FOR. Since it is not clear from the TYNDP data on planned unavailability in days why the shutdown is taking place, we also compare this with POD, SOD, and MOD. For us, MOD also includes forced outages due to maintenance, but in principle it reflects the meaning of POD in the TYNDP study. For most countries and power plants, it can be seen that ERAA therefore assumes higher FORs ( Figure 7 ). The same applies to the TYNDP data in Figure 8 . One reason for the heterogeneous results of the FOR with ERAA as well as TYNDP is probably the problem described above with the derivation of service hours. Our rates tend to underestimate reality due to the simplified assumptions of economic shutdowns. The PODs are a KPI per power plant in TYNDP . To make this value comparable with our aggregated annual average for all power plants of a given type, we calculate all values per plant. Since it is also unclear whether a failure day in TYNDP refers to a total failure or only a deration, we additionally compare the equivalent performance indicators from total failure plus derations. Because the earliest capacity expansion year from the TYNDP is 2030 and some power plants may have been decommissioned or newly connected to the grid by then, small differences from our historical values may slightly distort the results. It is striking that, especially for gas-fired power plants, the number of estimated days of planned outages in our data set is significantly underestimated in almost all countries compared to TYNDP ( Figure 9 ). One reason for this could be the clustering and labelling approach (Methods – Step 3), where, in the event of a time overlap between planned and unplanned outages, the unplanned outage is selected as the label. As a result, many power plants tend to record forced outages for longer than planned outages. The differences illustrate the sometimes varying reporting practices of the TSOs with regard to the type and reason for outages: sometimes the reports with all maintenance are closer to the TYNDP value (e.g.: RO, RS or CZ) and sometimes the planned maintenance with various reasons (e.g.: FR, BE or HU). Since the methodology used to derive forced and planned outages in the TYNDP 2024 does not reveal the underlying key indicators or assumptions, the expected total unavailability of a power plant within a year was also compared with the available data set. As before, all indicators are computed per single power unit of a plant type. The results can be seen in Figure 10 . The most comprehensive comparison can be made against the data from Gjorgiev et al., because they provide similar empirical performance measures based on the same raw data. The results for ( Figure 11 ) or availability factors ( Figure 12 ) are very homogeneous for most countries, with minor differences, except for individual power plant types in some countries (i.e., Biomass in Belgium or France as well as Gas in Slovakia or Bulgaria). The tendency toward higher availability in the present data set compared to the comparative data can be explained, among other things, by the calculation of a power plant's active hours, since Gjorgiev et al. counted the hours between outage reports and thus potentially did not take into account the times without outages before the first report and after the last report of their investigation horizon 3 . In Figure 13 we also compare planned and unplanned days in outage. Since we do not know the number of power unit per plant type used by Gjorgiev et al. in their study, we sum up the values for hours in planned or forced outages (vice versa for the equivalent measures) and divide them by 24 to have an overall counter of days in planned or forced outage for a plant type. Our measures for all scheduled as well as unplanned outages or derations (including maintenance) have smaller deviations compared to just planned or forced KPIs, why we only plot them for comparison. Usage Notes The data can be directly used in energy system models, electricity market models, power plant dispatch models, and optimal power flow calculations, or as input in simulations to generate time series for the above-mentioned models. Average availability factors for power plant technologies or fuel types can implicitly reflect unplanned as well as maintenance outages and limit the supply side. If the attached scripts are used to determine the availability indicators of individual power plants, restrictions can also be mapped at the grid node level, allowing changes in power flow in contingency cases to be examined. The calculation of seasonal factors allows the use of availability profiles. This simplifies the modelling of hydropower technologies in particular, as the power restriction caused by outages at individual power plants is complex to implement due to cascading interdependencies of the water inflow. Furthermore, the script ‘outages_simulation.py’ provides an example of code that can be used to simulate total failure events of individual power plants based on different statistical distributions and the respective failure or repair rates. In addition, partial failures can be simulated in two variants by limiting power output. The statistical models are based on e.g. the Antares Simulator 30 . An example of one algorithm of the script can be found in the supplementary materials . Furthermore, a static or continuous multi-state Markov-Chain model using state transition probabilities could be used for the simulations 3 . Abbreviations Example w/u Weighted vs. Unweighted (E) Equivalent F/P/M/U/S Abbreviation for outage type O/D Total outage vs. derated Actual measure Rate, Factor, etc. Declarations Data Availability The created dataset as well as other data from different sources are publicly available. Different licences and terms of use may apply to the underlying input data from other sources than our dataset: Version 1.0 of the data set can be retrieved via the Zenodo repository. This link will also point to future updates: https://doi.org/10.5281/zenodo.18998099 Version 0.7.1 of the open-source power plant list from powerplantmatching (PyPSA) can be found on Zenodo: https://zenodo.org/records/14785651 Version 1.0 of the open-source power plant database from JRC can be found on Zenodo: https://zenodo.org/record/3266807 Code Availability The code for processing the raw data from the ENTSO-E Transparency Platform, merging power plant databases, calculating the performance measures of the attached data set, simulating outage time series, and reproducing all figures is available as open-source code in Zenodo . It can be freely used and adapted, for example, to calculate additional outage KPIs or grouped aggregations based on other power plant characteristics: Version 1.0 of the scripts can be retrieved via the Zenodo repository. This link will also point to future updates: https://doi.org/10.5281/ zenodo.18998099 Acknowledgements This work was supported in part by the Helmholtz Association under the project “ Helmholtz platform for the design of robust energy systems and raw material supply” (RESUR) [grant number: 37.12.02 ] as well as the Helmholtz Program Energy System Design (ESD) [grant number: 37.12.03 ]. Author contributions E.J.: Writing – original draft, Visualization, Validation, Software, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. A.A.: Conceptualization, Writing – review & editing, Supervision. Competing interests The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. References VGBE-TW-103V. Availability of Power Plants 2014 – 2023 . (vgbe energy service GmbH, Essen, 2024). ENTSO-E Transparency Platform. UnavailabilityOfGenerationUnits_15.1.A_B. Transparency Platform HelpDesk https://transparencyplatform.zendesk.com/hc/en-us/articles/40477405498257-UnavailabilityOfGenerationUnits-15-1-A-B (2025). Gjorgiev, B., Stankovski, A., Wengler, J., Sencan, S. & Sansavini, G. Availability of the European power system assets: What we learn from data? Reliab. Eng. Syst. Saf. 258 , (2025). Bassini, C. F. Seasonal and weather influences in the outages of German thermal generators. ACM SIGEnergy Energy Inform. Rev. 5 , (2025). IEEE SA. 762-2023 - IEEE Standard Definitions for Use in Reporting Electric Generating Unit Reliability, Availability, and Productivity . (IEEE, New York, NY, 2023). N-SIDE. Study on the Outage Parameters of Generation Units and DC Links . (2022). Pecinovsky, J. & Boerman, F. entsoe-py. (2025). ENTSO-E Transparency Platform. Planned Unavailability & Changes in Actual Availability of Generation & Production Units [15.1.A] & [15.1.B] & [15.1.C] & [15.1.D]. Transparency Platform HelpDesk https://transparencyplatform.zendesk.com/hc/en-us/articles/16652173943828-Planned-Unavailability-Changes-in-Actual-Availability-of-Generation-Production-Units-15-1-A-15-1-B-15-1-C-15-1-D (2025). ENTSO-E Transparency Platform. New Transparency Platform Website Go Live. News https://transparency.entsoe.eu/news (2025). ENTSO-E Transparency Platform. Unavailability of Production and Generation Units [15.1.A&B&C&D]. Transparency Platform HelpDesk https://transparencyplatform.zendesk.com/hc/en-us/articles/32492060810513-UnavailabilityOfProductionAndGenerationUnits-15-1-A-B-C-D-r3 (2025). NERC. Generating Availability Data System - Data Reporting Instructions . (2025). North American Electric Reliability Corporation (NERC). Derate Event Reporting - Data Reporting Instructions - Section III. (2025). North American Electric Reliability Corporation (NERC). Outage Event Reporting - Data Reporting Instructions - Section III. (2025). Schmitz, R., Frischmuth, F., Braun, M. & Härtel, P. Coping with Risk Factors in Energy System Transformations - Climate Change Impacts on Nuclear Power Plant Availability in Europe. in 2024 20th International Conference on the European Energy Market (EEM) (IEEE, Istanbul, Turkiye, 2024). doi:10.1109/EEM60825.2024.10608936. Finck, J. R. Techno-economic assessment of market coupling regimes in future electricity systems. (Karlsruher Institut für Technologie, Karlsruhe, Deutschland, 2024). Gils, H. C., Bothor, S., Genoese, M. & Cao, K.-K. Future security of power supply in Germany - The role of stochastic power plant outages and intermittent generation. Int. J. Energy Res. 42 , (2018). Deakin, M., Greenwood, D., Brayshaw, D. J. & Bloomfield, H. Comparing Generator Unavailability Models with Empirical Distributions from Open Energy Datasets. in 2022 17th International Conference on Probabilistic Methods Applied to Power Systems (PMAPS) (IEEE, Manchester, United Kingdom, 2022). doi:10.1109/PMAPS53380.2022.9810629. Kanellopoulos, K., De Felice, M., Hidalgo Gonzalez, I. & Bocin, A. JRC Open Power Plants Database (JRC-PPDB-OPEN). Zenodo https://doi.org/10.5281/ZENODO.3266807 (2019). Gotzens, F., Heinrichs, H., Hörsch, J. & Hofmann, F. Performing energy modelling exercises in a transparent way - The issue of data quality in power plant databases. Zenodo https://doi.org/10.5281/ZENODO.14785651 (2025). Sergio, A. & Colelli, F. P. Weather-induced power plant outages: Empirical evidence from hydro and thermal generators in Europe. Energy Econ. 148 , 108549 (2025). Murphy, S., Sowell, F. & Apt, J. A time-dependent model of generator failures and recoveries captures correlated events and quantifies temperature dependence. Appl. Energy 253 , (2019). Ersayin, E. & Ozgener, L. Performance analysis of combined cycle power plants: A case study. Renew. Sustain. Energy Rev. 43 , 832–842 (2015). Guénand, Y. et al. Climate change impact on nuclear power outages - Part I: A methodology to estimate hydro-thermic environmental constraints on power generation. Energy 307 , (2024). Collet, L. et al. Future nuclear power outages in a changing climate - A case study on two contrasted French power plants. Energy 320 , (2025). Curley, G. M. Reliability Analysis of Power Plant Unit Outage Problems. (2013). ENTSO-E Transparency Platform. Actual Generation per Production Type [16.1.B&C]. Transparency Platform HelpDesk https://transparencyplatform.zendesk.com/hc/en-us/articles/16648290299284-Actual-Generation-per-Production-Type-16-1-B-C (2025). ENTSO-E Transparency Platform. Installed Generation Capacity Aggregated [14.1.A]. Transparency Platform HelpDesk https://transparencyplatform.zendesk.com/hc/en-us/articles/16648300912916-Installed-Generation-Capacity-Aggregated-14-1-A#:~:text=A%5D,-2%20months%20ago&text=The%20sum%20of%20installed%20Net,generation%20capacity%2C%20per%20production%20type. (2025). ENTSO-E. European Resource Adequacy Assessment 2023 Edition - Annex 1: Input Data and Assumptions . https://www.entsoe.eu/eraa/2023/report/ERAA_2023_Annex_1_Assumptions.pdf (2024). ENTSO-E. TYNDP 2024 Scenarios Methodology Report - Final Version . https://2024.entsos-tyndp-scenarios.eu (2025). Antares-Simulator Team. Antares-Simulator. Antares Simulator Documentation https://antares-simulator.readthedocs.io/en/latest/user-guide/ts-generator/05-algorithm/ (2025). Footnotes As it turned out that there were still duplicates of the identification codes, the line with the lowest available capacity was selected to cover the worst-case scenario. The difference of 2,536 entries is due to the fact that additional countries were filtered and the start and end dates of the report must be between January 2015 and October 2025. Northern Ireland is included in Great Britain. Iceland, Malta, Ukraine, the Azores, the Canary Islands, the Balearic Islands, and other overseas territories of France or Great Britain were not included. The term “outage” in this list includes not only total outages, but also performance limitations and equivalents consisting of total and partial outages. Additional Declarations No competing interests reported. Supplementary Files SupplementaryMaterials.docx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9138908","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"data-descriptor","associatedPublications":[],"authors":[{"id":609922243,"identity":"e0a09c5f-84f0-4d9c-b3e2-6b5979a87c5c","order_by":0,"name":"Eric Jahnke","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA30lEQVRIiWNgGAWjYFACxgYGBh6GBCBmfNhgABJJIF4Ls2GDgQExWiAApIVNsoGBCC380ocbPzDI2OWZ8xw+Vjmj4A8DP3uOAV4tkn2JzRIMPMnFlr1taTc3AB0m2fMGvxaDM4xtQL8wJ244z2N28wFQi8ENArbYQ7TUA7XwfysEabEnpMWAB6zlcOKGsz1sjCCHGUgQ0CJxhrFZIoHneOKGM8eMJWcYGPNInHlWgFcLfw/7ww8fe6qBWpIffuz5IyfH3568Aa8WMEjsQbB5CCsHgx9EqhsFo2AUjIKRCQC4O0OCLIggRQAAAABJRU5ErkJggg==","orcid":"","institution":"Karlsruhe Institute of Technology","correspondingAuthor":true,"prefix":"","firstName":"Eric","middleName":"","lastName":"Jahnke","suffix":""},{"id":609922244,"identity":"3feb10a2-e503-4ec4-b450-df8d4a2d2a08","order_by":1,"name":"Armin Ardone","email":"","orcid":"","institution":"Karlsruhe Institute of Technology","correspondingAuthor":false,"prefix":"","firstName":"Armin","middleName":"","lastName":"Ardone","suffix":""}],"badges":[],"createdAt":"2026-03-16 14:10:07","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9138908/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9138908/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":105354211,"identity":"2a69a94e-b9b0-4ced-9eb7-1881fc8c0768","added_by":"auto","created_at":"2026-03-25 06:27:27","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":13153,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eMethod to bridge gaps between outage reports using a defined maximum temporal difference.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-9138908/v1/7b4696983a40a9d35ba3966d.png"},{"id":105354171,"identity":"88a211f6-63ec-45d7-945c-68ac5d00ddab","added_by":"auto","created_at":"2026-03-25 06:27:16","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":36436,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eClustering and labelling of multiple different overlapping outage reports.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-9138908/v1/81a44dee70a52f4f949aa276.png"},{"id":105354170,"identity":"5a8041e8-652d-44ee-86da-1d0115cb2f1b","added_by":"auto","created_at":"2026-03-25 06:27:16","extension":"jpeg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":1728267,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eMap of the power plant list used for the statistical analysis merged by databases from ENTSO-E, JRC and PyPSA.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"floatimage3.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-9138908/v1/10e135d7e7f5be8e540957e0.jpeg"},{"id":105354219,"identity":"6f341c4e-eb9a-426f-b18f-2519aa0561ca","added_by":"auto","created_at":"2026-03-25 06:27:28","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":305772,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eComparison of aggregated available capacity vs daily generation relative to installed capacity for Spain.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-9138908/v1/19fcc30be1bc8fd9aba91dba.png"},{"id":105354290,"identity":"4967d12b-3172-4bb0-8882-38465c8d3e2b","added_by":"auto","created_at":"2026-03-25 06:27:40","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":308533,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eComparison of aggregated available capacity vs daily generation relative to installed capacity for France.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-9138908/v1/f01fd38970f2946c12f4c9f7.png"},{"id":105354201,"identity":"c1ef546e-c83c-4685-a67d-d2a09b9cc637","added_by":"auto","created_at":"2026-03-25 06:27:25","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":263175,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eComparison of aggregated available capacity vs daily generation relative to installed capacity for Italy.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-9138908/v1/17bd7bd7c670804abd387ed6.png"},{"id":105354177,"identity":"07aa182f-9601-47b4-928d-bbb6cbeeb8c8","added_by":"auto","created_at":"2026-03-25 06:27:18","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":526295,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eComparison of FOR from ERAA 2023 against FOR and UOR from our dataset\u003c/em\u003e.\u003c/p\u003e","description":"","filename":"floatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-9138908/v1/bb97b88da81055b4c1a2a700.png"},{"id":105354262,"identity":"e30327fb-0175-48c9-8c54-38b50ba86769","added_by":"auto","created_at":"2026-03-25 06:27:37","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":432311,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eComparison of FOR from TYNDP 2024 against FOR and UOR from our dataset\u003c/em\u003e.\u003c/p\u003e","description":"","filename":"floatimage8.png","url":"https://assets-eu.researchsquare.com/files/rs-9138908/v1/de0acb16b39ce47c49f79773.png"},{"id":105354223,"identity":"124de924-2336-4c6f-ba70-c4cb8b3eb050","added_by":"auto","created_at":"2026-03-25 06:27:28","extension":"png","order_by":9,"title":"Figure 9","display":"","copyAsset":false,"role":"figure","size":731388,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eComparison of POD from TYNDP 2024 against planned and maintenance outages as well as derations from our dataset\u003c/em\u003e.\u003c/p\u003e","description":"","filename":"floatimage9.png","url":"https://assets-eu.researchsquare.com/files/rs-9138908/v1/2fdd38f0378045c5187fb112.png"},{"id":105354179,"identity":"2ae8dce2-a53b-4348-8ea1-7a3fc92163f0","added_by":"auto","created_at":"2026-03-25 06:27:18","extension":"png","order_by":10,"title":"Figure 10","display":"","copyAsset":false,"role":"figure","size":671456,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eComparison of expected total unavailability during a year per power unit of a given type from TYNDP 2024 against our dataset\u003c/em\u003e.\u003c/p\u003e","description":"","filename":"floatimage10.png","url":"https://assets-eu.researchsquare.com/files/rs-9138908/v1/38b89daa4a5b250c8623404e.png"},{"id":105354247,"identity":"4f925f01-5344-48d2-a035-129e948ce53d","added_by":"auto","created_at":"2026-03-25 06:27:32","extension":"png","order_by":11,"title":"Figure 11","display":"","copyAsset":false,"role":"figure","size":883939,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eComparison of FOR and UOR with Gjorgiev at al. 2025\u003c/em\u003e\u003c/p\u003e","description":"","filename":"floatimage11.png","url":"https://assets-eu.researchsquare.com/files/rs-9138908/v1/fd904946b075b8dd689d7226.png"},{"id":105354221,"identity":"46e3d521-36a6-4ab5-8dd5-8d8eb9dd3207","added_by":"auto","created_at":"2026-03-25 06:27:28","extension":"png","order_by":12,"title":"Figure 12","display":"","copyAsset":false,"role":"figure","size":531046,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eComparison of availability factors with Gjorgiev et al. 2025\u003c/em\u003e\u003c/p\u003e","description":"","filename":"floatimage12.png","url":"https://assets-eu.researchsquare.com/files/rs-9138908/v1/b2b79ec463de98ed38225089.png"},{"id":105354202,"identity":"e94bce38-8fa7-4622-afc5-054ee6df23af","added_by":"auto","created_at":"2026-03-25 06:27:25","extension":"png","order_by":13,"title":"Figure 13","display":"","copyAsset":false,"role":"figure","size":900132,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eComparison of planned as well as unplanned outage and derations days with Gjorgiev et al. 2025\u003c/em\u003e\u003c/p\u003e","description":"","filename":"floatimage13.png","url":"https://assets-eu.researchsquare.com/files/rs-9138908/v1/4ff554e0c59ed030150ab3e2.png"},{"id":105569490,"identity":"4609806a-6b72-4a0c-8b8d-a2c76b066d7c","added_by":"auto","created_at":"2026-03-27 13:12:43","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":9455933,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9138908/v1/38763d41-ffad-4a1a-beeb-83dcdb93b44b.pdf"},{"id":105565268,"identity":"f7d30f57-529b-4ff9-905f-50b4a4ad0af0","added_by":"auto","created_at":"2026-03-27 12:52:40","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":13990205,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryMaterials.docx","url":"https://assets-eu.researchsquare.com/files/rs-9138908/v1/ce0077d08e543a1b1f30f1f8.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Performance measures and availability of European power plants","fulltext":[{"header":"Background \u0026 Summary","content":"\u003cp\u003eAvailability data of power units is required for realistic models in the energy industry or electrical engineering for various applications e.g., to map supply shortages in electricity market models, to provide the basis for investment decisions based on security of supply in energy system models or to analyse contingency cases in power flow optimizations. On detailed technology base and for specific countries, licensed data is available\u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u003c/sup\u003e. However, publicly available data sources are scarce. This data set provides aggregated statistics and performance measures for European electricity generation technologies and fuel types. Information about the performance of units is based on the monthly unavailability reports of the ENTSO-E transparency platform\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e. Methodologically, the data preparation is comparable to other recent publications in this field\u003csup\u003e\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e,\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u003c/sup\u003e. A new feature is a reproducible Python pipeline to generate regular time series per unit, more optional aggregations as well as validation against different studies. The performance measures comply with the current IEEE-762 standard\u003csup\u003e\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u003c/sup\u003e and other KPIs used in practice\u003csup\u003e\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u003c/sup\u003e. The data set can thus be used directly for applications such as probabilistic outage sampling, the simulation of time series in different granularity or as constraints for the aforementioned optimization models.\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003e\u003cstrong\u003eStep 1 - Data source\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe use monthly CSV file exports of the ENTSO-E transparency platform (\u0026ldquo;\u003cem\u003eUnavailabilityOfGenerationUnits_15.1.A_B\u003c/em\u003e\u0026ldquo;). Additional columns of clear names of the ENTSO-E identification codes of bidding zones or control areas were created using mapping tables taken from the Python API for the ENTSO-E transparency data download\u003csup\u003e7\u003c/sup\u003e. The data only contains entries from generation units with a nominal capacity of at least 100 MW and outage reports with a change in capacity of at least 100 MW. Outages of consumption units or transmission lines are not included, as there are other data sources for this on the transparency platform\u003csup\u003e8\u003c/sup\u003e. In November 2025 when this study was written, the ENTSO-E updated its data access from the FTP server and, in the process, also changed the table formats of some data sources\u003csup\u003e9\u003c/sup\u003e. As the code and statistics were created when the unavailability forms were still in change and the final format unclear, an overview and example of the old and currently proposed new table format is provided in \u003cstrong\u003eTable 1\u003c/strong\u003e. The time series for the outage notification is taken from the \u0026ldquo;StartOutage(UTC)\u0026rdquo; and \u0026ldquo;EndOutage(UTC)\u0026rdquo; columns. The columns \u0026ldquo;StartTimeSeries(UTC)\u0026rdquo; and \u0026ldquo;EndTimeSeries(UTC)\u0026rdquo; are ignored, as no further information is available for them.\u003c/p\u003e\n\u003ctable\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eOld column name\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eNew column name\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eData type\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eDescription\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eExample\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eStartOutage\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eStartOutage(UTC)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cem\u003eTimestamp\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eStart time of an outage event\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e2025-11-12 13:30:00.000\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eEndOutage\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eEndOutage(UTC)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cem\u003eTimestamp\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eEnd time of an outage event\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e2025-11-12 23:30:00.000\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eStartTS\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eStartTimeSeries(UTC)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cem\u003eTimestamp\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eStart of a change of the avail. cap. during an outage\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e2025-11-12 15:30:00.000\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eEndTS\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eEndTimeSeries(UTC)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cem\u003eTimestamp\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eEnd of a change of the avail. cap. during an outage\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e2025-11-12 20:30:00.000\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eTimeZone\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eTimeZone\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cem\u003eString\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eLocal time zone (e.g., \u0026ldquo;CET\u0026rdquo;)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eCET\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eMRID\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eInstanceCode\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cem\u003eString\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eUnique identifier of outage event\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e123XYZ\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eStatus\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eStatus\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cem\u003eString\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eCurrent status\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eCancelled\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eType\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eType\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cem\u003eString\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026ldquo;Planned\u0026rdquo; or \u0026ldquo;Forced\u0026rdquo; outage event\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eForced\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eAreaCode\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eAreaCode\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cem\u003eString\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eUnique identifier of market area\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e10YGB\u0026mdash;\u0026mdash;\u0026mdash;-A\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eAreaTypeCode\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eAreaTypeCode\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cem\u003eString\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eBidding zone or control area (\u0026ldquo;BZN\u0026rdquo; or \u0026ldquo;CTA\u0026rdquo;)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eCTA\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eAreaName\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eAreaDisplayName\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cem\u003eString\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eName of BZN or CTA\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eUK(National Grid) CTA\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eMapCode\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eAreaMapCode\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cem\u003eString\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eUnique identifier of BZN or CTA\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eGB\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003ePowerResourceEIC\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eAssetCode\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cem\u003eString\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eUnique identifier of the unit\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e48W0000ABC-15A\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eUnitName\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eAssetName\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cem\u003eString\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eName of the unit\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eGBX-2\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eProductionType\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eProductionType\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cem\u003eString\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eFuel or tech. type (e.g., \u0026ldquo;Nuclear\u0026rdquo;)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eFossil Oil\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eInstalledCapacity\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eInstalledCapacity\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cem\u003eFloat\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eInstalled capacity\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e200\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eAvailableCapacity\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eAvailableCapacity[MW]\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cem\u003eFloat\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eAvailable capacity during the outage event\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e100\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eVersion\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eVersion\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cem\u003eInteger\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eUpdate version of the MRID\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eOldVersion\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eOldVersion\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cem\u003eBoolean\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eIndicator if the version is still the most current\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eFALSE\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eReason\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eReason\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cem\u003eString\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eReason for the outage (e.g., \u0026ldquo;Failure\u0026rdquo;)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eFailure\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eUpdateTime\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eUpdateTime(UTC)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cem\u003eTimestamp\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eUpdate time of the outage event\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e2025-11-12 12:30:00.000\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cem\u003eTable 1: Overview about raw monthly CSV file for unavailability of generation units from the old as well as newly updated format\u003c/em\u003e\u003csup\u003e2,8,10\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eStep 2 - Data cleaning\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot all single entries are valid anymore. If maintenance is planned in the medium term due to a technical problem that has occurred and this can be resolved in the short term, the original notification for future maintenance work is cancelled. As a result, the raw data set is filtered exclusively for reports that are labelled as \u0026ldquo;Active\u0026rdquo; and not \u0026ldquo;Withdrawn\u0026rdquo; or \u0026ldquo;Cancelled\u0026rdquo;. Regardless of their type, outage notifications have an identification code (\u0026ldquo;MRID\u0026rdquo;) for clear traceability with, for example, the additional ENTSO-E table for the reasons of the outage. However, as outages for planned maintenance work have to be postponed or ad-hoc messages for unplanned outages have to be readjusted at short notice due to technical problems, there are often duplicates for a specific identification code. Consequently, all entries that show an old version were filtered out. As there were still duplicates in the identifier, all reports were sorted by version and only the line with the highest version number was taken over. In this respect, the clean-up approach of Gjorgiev et al.\u003csup\u003e3\u003c/sup\u003e differs, as the entries with duplicate identifiers were not removed after repeated filtering, but were treated as different reports with more accidentally similar identifiers. In our approach, if there were still duplicates of the identifiers, the entry with the most recent update time was used\u003ca href=\"#_ftn1\" name=\"_ftnref1\" title=\"\"\u003e\u003c/a\u003e\u003csup\u003e1\u003c/sup\u003e. All duplicated entries are collected during processing and can be written out as a CSV file. Some countries are missing entirely in the bidding zones (e.g., Germany). Since Germany is also divided into four balancing zone, there are four entries instead of one aggregated. For this reason, we do not yet filter specifically by bidding zone or control areas during data preparation, but instead specify the countries to be filtered by before calculating the performance indicators. There are mainly duplicates here because we did not filter bidding zones or control areas. An overview of the raw and cleaned dataset of reports used for this study is given in \u003cstrong\u003eTable 2\u003c/strong\u003e:\u003c/p\u003e\n\u003ctable style=\"margin-right: calc(53%); width: 47%;\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eN\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eRaw reports\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e12,042,695\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eCancelled\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e967,496\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eWithdrawn\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e301,275\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eOld version\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e9,466,610\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eActive current version\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e1,307,314\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eDuplicated rows\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e287,742\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eDuplicated MRIDs\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e69,975\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eActive unique reports\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e949,597\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eActive filtered reports\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003e947,061\u003c/strong\u003e\u003ca href=\"#_ftn2\" name=\"_ftnref2\" title=\"\"\u003e\u003c/a\u003e\u003csup\u003e2\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eDuplicated EIC codes\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e6\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eLong-ranging single outage clusters\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e315\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eReports used for performance measures\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003e946,740\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cem\u003eTable 2: Overview of raw outage reports before and after cleaning and pre-processing\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eStep 3 - Data pre-processing\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eBridging temporal gaps\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eSometimes there is a temporal gap of only a few hours between two outage reports, where it can be assumed that the generation unit was not running in normal operation and this gap evolved probably from a reporting error. For this reason, a time interval of \u003cem\u003e\u0026tau;\u003c/em\u003e hours can be defined in which two failure reports are linked to each other (\u003cstrong\u003eFigure 1\u003c/strong\u003e). The possibility of defining a bridge parameter is optionally available in the accompanying code, whereby the value of the available capacity as well as the type and reason for the earlier outage are extended until the later outage. Optionally, parameters can also be set to link only short time deltas between reports of the same type and/or reason. A delta of eight hours was assumed for the calculation of the statistics for this data set, but even more hours would have been possible according to sensitivity analysis in other studies\u003csup\u003e3\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eClustering and Labelling\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eOutage events for a generation unit often overlap in time. The problem here is that not only different available capacities are reported for a specific point in time, but the type and reason for the outage can also differ. In addition, the reports have different creation or update times, which makes it difficult to clearly identify which is the current report in the event of a temporal overlap. The possible reasons for the outage are divided by us into just two categories: \u0026ldquo;Maintenance\u0026rdquo; and \u0026ldquo;Other\u0026rdquo;. A more specific differentiation would not bring more benefit to the statistical analysis. Regarding the type, all reports distinguish between planned and forced outages. Officially, the report for a planned unavailability must be published to the transparency platform maximum one hour after the official confirmation by the TSO. Updates to this report (a newer version) must be published at the latest an hour after the information is known. Information on unplanned outages must be published no later than one hour after the change in available capacity\u003csup\u003e8\u003c/sup\u003e. In principle, this is in line with other institutions such as NERC, whose declarations between a long-term planned outage, planned maintenance or unplanned maintenance depend on the announcement time before the event and may have to be adjusted afterwards\u003csup\u003e11\u0026ndash;13\u003c/sup\u003e. However, for the ENTSO-E unavailability data the difference between notifications for forced and planned is often unclear, as planned outages are only partially published with a long lead time\u003csup\u003e4\u003c/sup\u003e. In some cases, reports of planned maintenances don\u0026rsquo;t even exist, because every maintenance is reported as \u0026lsquo;forced\u0026rsquo; (a more detailed analysis can be found in the \u003cstrong\u003esupplementary materials\u003c/strong\u003e). In reality, the correct classification of outage types as forced or planned, as well as the separation of overlapping reports, depends on the type of power plant and specific background knowledge about the reasons for the outages in terms of expected repair times\u003csup\u003e11\u003c/sup\u003e. The labelling is therefore subject to a level of uncertainty that might bias the statistical analysis.\u003c/p\u003e\n\u003cp\u003ePrevious works used different heuristics and methods to cluster and label the reports. Schmitz et al.\u003csup\u003e14\u003c/sup\u003e and Finck\u003csup\u003e15\u003c/sup\u003e aggregate the (un-)availabilities of the reports regardless if it was planned or forced. Gils et al.\u003csup\u003e16\u003c/sup\u003e or Deakin et al.\u003csup\u003e17\u003c/sup\u003e differentiate only between planned and forced, but they ignore the outage reasons and do not explain their methodology for pre-processing the data. Bassini\u003csup\u003e4\u003c/sup\u003e characterises the type in the event of a temporal overlap of 3 notifications based on the majority vote. The procedure for an even number of overlapping events such as two, four or more is not described. We differentiate between the clustering of the overlapping outage reports with regard to the value of the available capacity, the label for the outage type, and the reason. For each time step with overlapping reports, we assign the value of the available capacity by the report with the most recent update timestamp. For the label of the outage type and reason, we assign the values for the earliest outage. For reports with identical outage starts, we again assign the label with the most recent update time. However, the same logic can be implemented for the labels as for the capacity using parameters in the code, so that the report with the latest update timestamp is always decisive. \u0026nbsp;The separation of a non-availability event into separate events is implemented when the state transitions from a planned deration or outage to a forced deration or outage\u003csup\u003e3\u003c/sup\u003e. An example of three overlapping outage reports for a specific day with different update times and types is shown in \u003cstrong\u003eFigure 2\u003c/strong\u003e, visualising the decisions of clustering and labelling the reports. Other cases can be found in the \u003cstrong\u003esupplementary materials\u003c/strong\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eStates\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eA generating unit can be either fully available, derated or fully unavailable. It is fully available if it can potentially produce electricity at its rated power. If a unit is in outage, it cannot generate any electricity at all. Derated means that a unit cannot operate at full capacity, but is not off the grid. Consequently, the possible states could be defined as planned outage, planned derated, unplanned outage, unplanned derated and available\u003csup\u003e3\u003c/sup\u003e. Since we record the outage type separately, we only define the primary states as \u0026ldquo;out\u0026rdquo;, \u0026ldquo;derated\u0026rdquo; and \u0026ldquo;available\u0026rdquo;.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eFinal cleaning\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eSince the outage data for bidding zones and control zones in countries with multiple zones (i.e., Norway and Italy) often contain overlapping entries and also show changes in the affiliation of individual power plants over time, the performance measures were only calculated for each country and duplicate entries of the EIC codes were removed beforehand. These duplicates are exported to a separate CSV file for overview purposes. Similarly, individual outages of the same type (\u0026ldquo;planned\u0026rdquo;) lasting longer than 365 days were assumed to be erroneous entries and were therefore removed. However, if two or more outages of different types (a \u0026ldquo;planned\u0026rdquo; outage followed immediately by an \u0026ldquo;unplanned\u0026rdquo; outage) exceed the defined duration, they are not removed, as they essentially represent two separate clusters\u003csup\u003e3\u003c/sup\u003e. All removed entries in this way are recorded and exported in a CSV file.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eStep 4 - Meta information for power units\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eOn the ENTSO-E Transparency Platform, thermal power plants in particular are only listed by fuel type and general plant types, but not broken down in more detail by technology type such as open-cycle gas turbine and combined-cycle gas turbine. The year of commissioning (or retrofit) of the units is also not available, nor is the information as to whether it is a CHP plant. We add this information from freely available data sources. To do this, we use the open units table from JRC\u003csup\u003e18\u003c/sup\u003e and a power plant list from the PyPSA modelling framework package \u003cem\u003epowerplantmatching\u003c/em\u003e\u003csup\u003e19\u003c/sup\u003e. As the JRC table also contains the official EIC codes of ENTSO-E for power plant locations and the individual units, we can join the power plant lists from JRC and PyPSA together with ENTSO-E using these columns. A geographical overview of plant types, technologies and installed capacities can be seen in \u003cstrong\u003eFigure 3\u003c/strong\u003e. Descriptive statistics on the meta-information of the power plants can be found in the \u003cstrong\u003esupplementary materials\u003c/strong\u003e. An example is given with \u003cstrong\u003eTable 3\u003c/strong\u003e. Since the JRC and PyPSA power plant lists do not contain the relevant technology types or commissioning data for all units in the ENTSO-E tables, the characteristics of individual technology types are unfortunately sometimes very limited and less reliable. This applies in particular to thermal power plants that use natural gas, coal, or oil as fuel. For this reason, the results also include the values for NAs in the technologies.\u003c/p\u003e\n\u003ctable style=\"width: 4.2e+2pt\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eColumn name\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eData type\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eDescription\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eExample\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eunit_eic\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cem\u003estring\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eUnique identifier of the gen. unit\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e48W0000ABC-15A\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eplant_eic\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cem\u003estring\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eUnique identifier of the gen. plant (a plant can consist of multiple units)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e48W0000ABC\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eunit_name\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cem\u003estring\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eName of the unit\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eGBX-2\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eplant_name\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cem\u003estring\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eName of the plant\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eGBX\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eplant_tech\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cem\u003estring\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eTechnology type\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eCCGT\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003efuel_type\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cem\u003estring\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eFuel type\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eFossil gas\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003efuel_type_code\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cem\u003estring\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eENTSO-E code of the fuel type\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eB04\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eset\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cem\u003estring\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eTechnology group\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eCHP\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eyear_commissioned\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cem\u003einteger\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eYear of first operation (or retrofit)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e1995\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eyear_commissioned\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cem\u003einteger\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eYear of decommissioning\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e2040\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eunit_installed_capacity\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cem\u003efloat\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eInstalled capacity (unit)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e400 MW\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eplant_installed_capacity\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cem\u003efloat\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eInstalled capacity (plant)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e800 MW\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003ecountry\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cem\u003estring\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eCountry location\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eGB\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003enuts2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cem\u003estring\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eNUTS 2 code of the unit location\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eGB5\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003elat\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cem\u003eFloat\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eGeographical latitude\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e51.50\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003elon\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cem\u003efloat\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eGeographical longitude\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e-0.12\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eeic_match_source\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cem\u003estring\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eSource column to merge PyPSA on JRC power plant databases\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eplant_eic\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cem\u003eTable 3: Overview about the merged power plant list to enhance the statistical analysis.\u003c/em\u003e\u003c/p\u003e"},{"header":"Data Records ","content":"\u003cp\u003e\u003cstrong\u003eRepository\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eCode and datasets are hosted online and can be downloaded from \u003cstrong\u003eZenodo\u003c/strong\u003e. The code already contains implementations for calculating more aggregated KPIs by e.g., age, size, or CHP type of the units. The performance measures originate from the IEEE 762 standard\u003csup\u003e5\u003c/sup\u003e as well as frequently used statistics in practice and academic studies\u003csup\u003e6\u003c/sup\u003e. They primarily relate to the time-based availability of power generation units. As the publication of data of individual non-anonymised generation units of the transparency platform is not permitted, capacity-weighted aggregated performance factors were calculated. Although these do not allow the same level of detail as plant-specific data, they are still sufficient enough for simulations of power plant outages or for a rough estimate of the availability. The files are:\u003c/p\u003e\n\u003cul\u003e\n \u003cli\u003e\u0026lsquo;kpis_tech_ALL.csv\u0026rsquo;\u003c/li\u003e\n \u003cli\u003e\u0026lsquo;kpis_tech_MOY.csv\u0026rsquo;\u003c/li\u003e\n \u003cli\u003e\u0026lsquo;kpis_plant_ALL.csv\u0026rsquo;\u003c/li\u003e\n \u003cli\u003e\u0026lsquo;kpis_plant_MOY.csv\u0026rsquo;\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003eThe four files contain average annual and monthly performance measures for ENTSO-E plant types or technology types and cover the ENTSO-E area with the following countries: Albania (AL), Austria (AT), Belgium (BE), Bosnia-Herzegovina (BA), Bulgaria (BG), Switzerland (CH), Czechia (CZ), \u0026nbsp;Germany (DE), Denmark (DK), Estonia (EE), Spain (ES), France (FR), Finland (FI), Great Britain (GB)\u003ca href=\"#_ftn1\" name=\"_ftnref1\" title=\"\"\u003e\u003c/a\u003e\u003csup\u003e3\u003c/sup\u003e, Greece (GR), \u0026nbsp;Croatia (HR), Hungary (HU), Ireland (IE), Italy (IT), Lithuania (LT), Latvia (LV), Luxembourg (LU), Moldova (MD), Montenegro (ME), North Macedonia (MK), Netherlands (NL), Norway (NO), Poland (PL), Portugal (PT), Romania (RO), Serbia (RS), Sweden (SE), Slovenia (SI), Slovakia (SK) and Kosovo (XK). The data is calculated as annual averages and seasonally disaggregated by month, among other things, because seasonal factors such as outside temperature, cooling water temperature, etc. change the likelihood of unplanned outages and therefore also TSOs schedule their planned maintenance according to the season\u003csup\u003e20\u0026ndash;24\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eThe repository additionally contains the Python and R scripts for preparing, analysing, evaluating and visualising the data, as well as an example script for simulating further time series:\u003c/p\u003e\n\u003cul\u003e\n \u003cli\u003e\u0026lsquo;availability_data.py\u0026rsquo;\u003c/li\u003e\n \u003cli\u003e\u0026lsquo;availability_statistics.py\u0026rsquo;\u003c/li\u003e\n \u003cli\u003e\u0026lsquo;availability_simulation.py\u0026rsquo;\u003c/li\u003e\n \u003cli\u003e\u0026lsquo;powerplants_merge.py\u0026rsquo;\u003c/li\u003e\n \u003cli\u003e\u0026lsquo;plot_outages_vs_gen.py\u0026rsquo;\u003c/li\u003e\n \u003cli\u003e\u0026lsquo;validate_kpis.R\u0026rsquo;\u003c/li\u003e\n \u003cli\u003e\u0026lsquo;descriptive_stats_jrc_pypsa.R\u0026rsquo;\u003c/li\u003e\n \u003cli\u003e\u0026lsquo;plot_outage_statistics.R\u0026rsquo;\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003e\u003cstrong\u003eTables\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe performance measures described above contain can be found with various additional prefixes or letter combinations as shown in \u003cstrong\u003eTable 4\u003c/strong\u003e:\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 39.0769%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eAbbreviation\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 60.9231%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eExample\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 39.0769%;\"\u003e\n \u003cp\u003ew/u\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 60.9231%;\"\u003e\n \u003cp\u003eWeighted vs. Unweighted\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 39.0769%;\"\u003e\n \u003cp\u003e(E)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 60.9231%;\"\u003e\n \u003cp\u003eEquivalent\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 39.0769%;\"\u003e\n \u003cp\u003eF/P/M/U/S\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 60.9231%;\"\u003e\n \u003cp\u003eAbbreviation for outage type\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 39.0769%;\"\u003e\n \u003cp\u003eO/D\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 60.9231%;\"\u003e\n \u003cp\u003eTotal outage vs. derated\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 39.0769%;\"\u003e\n \u003cp\u003eActual measure\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 60.9231%;\"\u003e\n \u003cp\u003eRate, Factor, etc.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cem\u003eTable 4: Structure of the name of a measure.\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eOutage types were taken into account according to their type and reason\u003csup\u003e5\u003c/sup\u003e:\u003c/p\u003e\n\u003cul\u003e\n \u003cli\u003e\u003cstrong\u003eU\u003c/strong\u003enplanned - \u003cu\u003eall\u003c/u\u003e forced outage events including technical failures as well as maintenance work.\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003eS\u003c/strong\u003echeduled - \u003cu\u003eall\u003c/u\u003e planned outage events combining maintenance work with other planned outages.\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003eM\u003c/strong\u003eaintenance - calculated accordingly from events with the reason \u0026lsquo;maintenance\u0026rsquo;, whereby the type can be planned \u003cu\u003eor\u003c/u\u003e forced.\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003eP\u003c/strong\u003elanned - calculated from reports containing planned outages where the reason is \u003cu\u003enot\u003c/u\u003e maintenance.\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003eF\u003c/strong\u003eorced - calculated from reports containing forced outages \u003cu\u003eexcluding\u003c/u\u003e maintenance.\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003eThis non-standard with regard to the IEEE-762 standard summary of types and reasons was chosen because it simplifies further use in simulations.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAggregations of the performance indicators were calculated in two different ways: capacity-weighted (w) and unweighted (u). In capacity weighting, the figures are weighted according to the rated capacity of the respective generation unit.\u003c/p\u003e\n\u003cp\u003eSince unplanned outages often only involve a reduction in output but do not require the power plant to be shut down completely, the equivalent values, which take into account both total outages and derations, are particularly useful here, in addition to the rates for total outages. For this reason all three variants have been calculated for factors and rates.\u003c/p\u003e\n\u003cp\u003eThe actual measures are briefly described in \u003cstrong\u003eTable 5\u003c/strong\u003e, but a detailed overview can be found in the \u003cstrong\u003esupplementary materials\u003c/strong\u003e:\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"557\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 25.3597%;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePerformance measure\u003ca href=\"#_ftn2\" name=\"_ftnref2\" title=\"\"\u003e\u003c/a\u003e\u003csup\u003e4\u003c/sup\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.9065%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eAbbreviation\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 57.7338%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eDescription\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 25.3597%;\"\u003e\n \u003cp\u003eAvailability factor\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.9065%;\"\u003e\n \u003cp\u003eAF\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 57.7338%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eFraction of time without outages\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 25.3597%;\"\u003e\n \u003cp\u003eOutage hours\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.9065%;\"\u003e\n \u003cp\u003eOH\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 57.7338%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eCumulated hours in outages\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 25.3597%;\"\u003e\n \u003cp\u003eOutage factor\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.9065%;\"\u003e\n \u003cp\u003eOF\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 57.7338%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eFraction of time in outages\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 25.3597%;\"\u003e\n \u003cp\u003eOutage rate\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.9065%;\"\u003e\n \u003cp\u003eOR\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 57.7338%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eShare of time in outage relative to operating time\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 25.3597%;\"\u003e\n \u003cp\u003eRepair rate\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.9065%;\"\u003e\n \u003cp\u003eRR\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 57.7338%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eAverage time from start to end of an outage event\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 25.3597%;\"\u003e\n \u003cp\u003eFailure rate\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.9065%;\"\u003e\n \u003cp\u003eFR\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 57.7338%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eEvent rate of an outage event\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 25.3597%;\"\u003e\n \u003cp\u003eEvent counts\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.9065%;\"\u003e\n \u003cp\u003eN\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 57.7338%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eNumber of single outage events\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 25.3597%;\"\u003e\n \u003cp\u003eMean deration\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.9065%;\"\u003e\n \u003cp\u003e/\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 57.7338%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eMean deration of partial outage events\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 25.3597%;\"\u003e\n \u003cp\u003eStandard deviation of derations\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.9065%;\"\u003e\n \u003cp\u003e/\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 57.7338%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eStandard deviation of derations of partial outage events\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cem\u003eTable 5: Overview about calculated performance measures independent of type and reason.\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eFactors and rates for the specific outage types are the usual measures used to describe performance, but also for further use in simulations\u003csup\u003e25\u003c/sup\u003e. \u0026nbsp;The repair time can be interpreted as the average duration of an event, i.e., how long maintenance takes on average in the event of an unplanned outage. The failure rate can be interpreted as the average duration between the occurrence of two outage events of a given type. Unless otherwise specified, all units describing time duration are calculated in hours. Rates and factors are relative and can be converted into percentages. Although the name suggests otherwise, the repair rate is also calculated in hours. \u0026nbsp;The number of events N describes the frequency of individual events that have occurred, e.g., of type \u0026ldquo;forced\u0026rdquo;. The mean value and standard deviation of a performance limitation describe the first two statistical moments of the distribution in the event of partial outages of the respective outage type and are calculated in MW. Total outages were not taken into account, as these always correspond to the nominal power and would skew the value upwards\u003csup\u003e12\u003c/sup\u003e.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWith regard to the rates for planned or unplanned outages, etc., a simplification had to be made: according to the standard, the service times of a power plant are actually required to calculate rates such as the forced outage rate (FOR). These are determined from the active times during which the power plant is active minus reserve shutdowns. Those shutdowns describe the unavailability of a power plant, e.g., due to negative electricity prices, insufficient demand, or the impossibility of operating the power plant economically and are therefore also called \u0026lsquo;economical shutdowns\u0026rsquo;\u003csup\u003e5\u003c/sup\u003e. Since these reserve shutdowns are not included in the outage time series and would have to be derived with uncertainty from e.g., historical market prices, they are not taken into account. Instead, service hours were calculated as active hours minus hours with total scheduled outages from (SOH). We therefore assume that a power plant was always potentially in service when it was not scheduled to be out of service and could theoretically produce electricity.\u003c/p\u003e\n\u003cp\u003eThe entire table for, e.g., \u0026lsquo;kpis_tech_ALL.csv\u0026rsquo; contains 140 columns. An example excerpt can be seen in \u003cstrong\u003eTable 6\u003c/strong\u003e:\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"557\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 25.3597%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eColumn\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.9065%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eAbbreviation\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 57.7338%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eDescription\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 25.3597%;\"\u003e\n \u003cp\u003eCountry\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.9065%;\"\u003e\n \u003cp\u003ecountry\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 57.7338%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eISO-A2 name of the country (e.g., \u0026ldquo;NO\u0026rdquo;, \u0026ldquo;DE\u0026rdquo;)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 25.3597%;\"\u003e\n \u003cp\u003eENTSO-E plant type\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.9065%;\"\u003e\n \u003cp\u003eplant_type\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 57.7338%;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePlant type of the ENTSO-E domain (e.g., \u0026ldquo;Biomass\u0026rdquo;)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 25.3597%;\"\u003e\n \u003cp\u003eActive hours\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.9065%;\"\u003e\n \u003cp\u003eACTH\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 57.7338%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eCumulated number of hours of entries for a power plant (typically a whole year)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 25.3597%;\"\u003e\n \u003cp\u003eService hours\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.9065%;\"\u003e\n \u003cp\u003eSH\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 57.7338%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eActive time minus hours in total scheduled unavailability\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 25.3597%;\"\u003e\n \u003cp\u003eObserved hours\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.9065%;\"\u003e\n \u003cp\u003eDH\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 57.7338%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eActive time minus hours in total planned unavailability\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 25.3597%;\"\u003e\n \u003cp\u003eForced outage hours\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.9065%;\"\u003e\n \u003cp\u003eFOH\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 57.7338%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eCumulated number of hours during forced outages.\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 25.3597%;\"\u003e\n \u003cp\u003eEquivalent availability factor\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.9065%;\"\u003e\n \u003cp\u003ewEAF\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 57.7338%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eCapacity-weighted fraction of time without total outages or derations\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 25.3597%;\"\u003e\n \u003cp\u003eForced outage factor\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.9065%;\"\u003e\n \u003cp\u003euSOF\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 57.7338%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eUnweighted fractions of time with scheduled outages\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 25.3597%;\"\u003e\n \u003cp\u003eUnplanned deration factor\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.9065%;\"\u003e\n \u003cp\u003ewUDF\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 57.7338%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eCapacity-weighted fractions of time in unplanned deration \u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 25.3597%;\"\u003e\n \u003cp\u003eEquivalent forced outage rate\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.9065%;\"\u003e\n \u003cp\u003euEFOR\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 57.7338%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eUnweighted shares of time in forced outage relative to service hours\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 25.3597%;\"\u003e\n \u003cp\u003eRepair rate of forced outages\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.9065%;\"\u003e\n \u003cp\u003ewRR_FO_h\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 57.7338%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eCapacity-weighted average time from start to end of a forced outage event (in hours)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 25.3597%;\"\u003e\n \u003cp\u003eFailure rate of maintenance outages\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.9065%;\"\u003e\n \u003cp\u003euFR_MO_h\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 57.7338%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eUnweighted event rate of a maintenance outage event (in hours)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 25.3597%;\"\u003e\n \u003cp\u003eEvent counts for planned outages\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.9065%;\"\u003e\n \u003cp\u003eN_PO\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 57.7338%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eCumulated number of single outage events for planned outages\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 25.3597%;\"\u003e\n \u003cp\u003eMean deration for unplanned outages\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.9065%;\"\u003e\n \u003cp\u003ewEUD_mw_mean\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 57.7338%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eCapacity-weighted mean deration of unplanned partial outage events (in MW)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 25.3597%;\"\u003e\n \u003cp\u003eStandard deviation of derations\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16.9065%;\"\u003e\n \u003cp\u003euEFD_mw_std\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 57.7338%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eUnweighted standard deviation of derations of forced partial outage events (in MW)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cem\u003eTable 6: Overview about calculated performance measures for the data set.\u003c/em\u003e\u003c/p\u003e"},{"header":"Technical Validation","content":"\u003cp\u003e\u003cstrong\u003eHistorical generation\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo verify the aggregated generator outages, \u003cstrong\u003eFigures 4-6\u003c/strong\u003e exemplarily shows curves of generation and available capacity relative to installed capacity for some countries of the dataset with their thermal and/or hydraulic plant types from ENTSO-E Transparency platform and available data for generation (\u0026ldquo;AggregatedGenerationPerType_16.1.B_C_r3\u0026rdquo;) as well as installed capacity (\u0026ldquo;InstalledGenerationCapacityAggregated_14.1.A_r3\u0026rdquo;)\u003csup\u003e26,27\u003c/sup\u003e. All data were aggregated to daily averages. The information on installed capacities is published annually and is therefore sometimes very inaccurate, especially for RES, because capacity is constantly being expanded during the year. Due to the restriction that only generation units with a minimum rated power of 100 MW should be included, there should be no outage reports for individual RES plants actually. However, the graph also shows restrictions for wind offshore. If generation is greater than the specified available capacity, there may be several reasons for this. On the one hand, the reported generation or the reported outage level could be incorrect, but on the other hand, there could also be an error in the clustering and processing of the outage reports for the specific case. The obvious caps on hard coal and natural gas in particular suggest errors in the raw data. Results for the remaining countries can be found in the \u003cstrong\u003esupplementary materials\u003c/strong\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eOther publications and studies\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe data set was compared with various other studies: the European Resource Adequacy Assessment (\u003cem\u003eERAA\u003c/em\u003e) 2023\u003csup\u003e28\u003c/sup\u003e and Ten Years Network Development Plan (\u003cem\u003eTYNDP\u003c/em\u003e) 2024\u003csup\u003e29\u003c/sup\u003e by ENTSO-E, as well as the journal article by Gjorgiev et al\u003csup\u003e3\u003c/sup\u003e. A comparison with availability factors for German plant types from Bassini\u003csup\u003e4\u003c/sup\u003e can be found in the \u003cstrong\u003esupplementary materials\u003c/strong\u003e. Data sets are compared using absolute values of the KPIs. The \u003cem\u003eERAA\u003c/em\u003e data set was chosen because there is data modelled for the current year 2025. From the \u003cem\u003eTYNDP\u003c/em\u003e study we took earliest data for 2030. With \u003cem\u003eERAA\u003c/em\u003e, we can compare forced outage rates unweighted and capacity-weighted by plant type from the ENTSO-E domain. Since the methodological reports did not provide any information on how the data was specifically determined, we compare it with FOR and UOR of our dataset, also taking forced maintenance into account. Regarding \u003cem\u003eTYNDP\u003c/em\u003e, we can also compare planned outage days by converting our hourly results into days in addition to the FOR. Since it is not clear from the \u003cem\u003eTYNDP\u003c/em\u003e data on planned unavailability in days why the shutdown is taking place, we also compare this with POD, SOD, and MOD. For us, MOD also includes forced outages due to maintenance, but in principle it reflects the meaning of POD in the \u003cem\u003eTYNDP\u003c/em\u003e study.\u003c/p\u003e\n\u003cp\u003eFor most countries and power plants, it can be seen that \u003cem\u003eERAA\u003c/em\u003e therefore assumes higher FORs (\u003cstrong\u003eFigure 7\u003c/strong\u003e). The same applies to the \u003cem\u003eTYNDP\u003c/em\u003e data in \u003cstrong\u003eFigure 8\u003c/strong\u003e. One reason for the heterogeneous results of the FOR with \u003cem\u003eERAA\u003c/em\u003e as well as \u003cem\u003eTYNDP\u003c/em\u003e is probably the problem described above with the derivation of service hours. Our rates tend to underestimate reality due to the simplified assumptions of economic shutdowns.\u003c/p\u003e\n\u003cp\u003eThe PODs are a KPI per power plant in \u003cem\u003eTYNDP\u003c/em\u003e. To make this value comparable with our aggregated annual average for all power plants of a given type, we calculate all values per plant. \u0026nbsp;Since it is also unclear whether a failure day in \u003cem\u003eTYNDP\u003c/em\u003e refers to a total failure or only a deration, we additionally compare the equivalent performance indicators from total failure plus derations. Because the earliest capacity expansion year from the \u003cem\u003eTYNDP\u003c/em\u003e is 2030 and some power plants may have been decommissioned or newly connected to the grid by then, small differences from our historical values may slightly distort the results. It is striking that, especially for gas-fired power plants, the number of estimated days of planned outages in our data set is significantly underestimated in almost all countries compared to \u003cem\u003eTYNDP\u0026nbsp;\u003c/em\u003e(\u003cstrong\u003eFigure 9\u003c/strong\u003e). One reason for this could be the clustering and labelling approach (Methods \u0026ndash; Step 3), where, in the event of a time overlap between planned and unplanned outages, the unplanned outage is selected as the label. As a result, many power plants tend to record forced outages for longer than planned outages. The differences illustrate the sometimes varying reporting practices of the TSOs with regard to the type and reason for outages: sometimes the reports with all maintenance are closer to the TYNDP value (e.g.: RO, RS or CZ) and sometimes the planned maintenance with various reasons (e.g.: FR, BE or HU).\u003c/p\u003e\n\u003cp\u003eSince the methodology used to derive forced and planned outages in the \u003cem\u003eTYNDP\u003c/em\u003e 2024 does not reveal the underlying key indicators or assumptions, the expected total unavailability of a power plant within a year was also compared with the available data set. As before, all indicators are computed per single power unit of a plant type. The results can be seen in \u003cstrong\u003eFigure 10\u003c/strong\u003e.\u003c/p\u003e\n\u003cp\u003eThe most comprehensive comparison can be made against the data from Gjorgiev et al., because they provide similar empirical performance measures based on the same raw data. The results for (\u003cstrong\u003eFigure 11\u003c/strong\u003e) or availability factors (\u003cstrong\u003eFigure 12\u003c/strong\u003e) are very homogeneous for most countries, with minor differences, except for individual power plant types in some countries (i.e., Biomass in Belgium or France as well as Gas in Slovakia or Bulgaria). The tendency toward higher availability in the present data set compared to the comparative data can be explained, among other things, by the calculation of a power plant\u0026apos;s active hours, since Gjorgiev et al. counted the hours between outage reports and thus potentially did not take into account the times without outages before the first report and after the last report \u0026nbsp; of their investigation horizon\u003csup\u003e3\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eIn \u003cstrong\u003eFigure 13\u003c/strong\u003e we also compare planned and unplanned days in outage. Since we do not know the number of power unit per plant type used by Gjorgiev et al. in their study, we sum up the values for hours in planned or forced outages (vice versa for the equivalent measures) and divide them by 24 to have an overall counter of days in planned or forced outage for a plant type. Our measures for all scheduled as well as unplanned outages or derations (including maintenance) have smaller deviations compared to just planned or forced KPIs, why we only plot them for comparison.\u003c/p\u003e"},{"header":"Usage Notes ","content":"\u003cp\u003eThe data can be directly used in energy system models, electricity market models, power plant dispatch models, and optimal power flow calculations, or as input in simulations to generate time series for the above-mentioned models. Average availability factors for power plant technologies or fuel types can implicitly reflect unplanned as well as maintenance outages and limit the supply side. If the attached scripts are used to determine the availability indicators of individual power plants, restrictions can also be mapped at the grid node level, allowing changes in power flow in contingency cases to be examined. The calculation of seasonal factors allows the use of availability profiles. This simplifies the modelling of hydropower technologies in particular, as the power restriction caused by outages at individual power plants is complex to implement due to cascading interdependencies of the water inflow.\u003c/p\u003e\n\u003cp\u003eFurthermore, the script \u0026lsquo;outages_simulation.py\u0026rsquo; provides an example of code that can be used to simulate total failure events of individual power plants based on different statistical distributions and the respective failure or repair rates. In addition, partial failures can be simulated in two variants by limiting power output. The statistical models are based on e.g. the \u003cem\u003eAntares Simulator\u003c/em\u003e\u003csup\u003e30\u003c/sup\u003e. An example of one algorithm of the script can be found in the \u003cstrong\u003esupplementary materials\u003c/strong\u003e. Furthermore, a static or continuous multi-state Markov-Chain model using state transition probabilities could be used for the simulations\u003csup\u003e3\u003c/sup\u003e.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cdiv class=\"DefinitionList\"\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003e\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eExample\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003ew/u\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eWeighted vs. Unweighted\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003e(E)\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eEquivalent\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eF/P/M/U/S\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eAbbreviation for outage type\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eO/D\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eTotal outage vs. derated\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eActual measure\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eRate, Factor, etc.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003c/div\u003e"},{"header":"Declarations","content":"\u003ch3\u003eData Availability\u003c/h3\u003e\n\u003cp\u003eThe created dataset as well as other data from different sources are publicly available. Different licences and terms of use may apply to the underlying input data from other sources than our dataset:\u003c/p\u003e\n\u003cul class=\"decimal_type\"\u003e\n\u003cli\u003eVersion 1.0 of the data set can be retrieved via the Zenodo repository. This link will also point to future updates: \u003cstrong\u003ehttps://doi.org/10.5281/zenodo.18998099\u003c/strong\u003e\u003c/li\u003e\n\u003cli\u003eVersion 0.7.1 of the open-source power plant list from \u003cem\u003epowerplantmatching\u003c/em\u003e (PyPSA) can be found on Zenodo: \u003cstrong\u003ehttps://zenodo.org/records/14785651\u003c/strong\u003e \u003c/li\u003e\n\u003cli\u003eVersion 1.0 of the open-source power plant database from JRC can be found on Zenodo: \u003cstrong\u003ehttps://zenodo.org/record/3266807\u003c/strong\u003e \u003c/li\u003e\n\u003c/ul\u003e\n\n\u003ch3\u003eCode Availability\u003c/h3\u003e\n\u003cp\u003eThe code for processing the raw data from the ENTSO-E Transparency Platform, merging power plant databases, calculating the performance measures of the attached data set, simulating outage time series, and reproducing all figures is available as open-source code in \u003cstrong\u003eZenodo\u003c/strong\u003e. It can be freely used and adapted, for example, to calculate additional outage KPIs or grouped aggregations based on other power plant characteristics:\u003c/p\u003e\n\u003cul\u003e\n\u003cli\u003eVersion 1.0 of the scripts can be retrieved via the Zenodo repository. This link will also point to future updates: \u003cstrong\u003ehttps://doi.org/10.5281/\u003c/strong\u003e\u003cstrong\u003ezenodo.18998099\u003c/strong\u003e\u003c/li\u003e\n\u003c/ul\u003e\n\n\u003ch3\u003eAcknowledgements\u003c/h3\u003e\n\u003cp\u003eThis work was supported in part by the Helmholtz Association under the project \u0026ldquo;\u003cstrong\u003eHelmholtz platform for the design of robust energy systems and raw material supply\u0026rdquo; \u003c/strong\u003e(RESUR) [grant number:\u003cstrong\u003e37.12.02\u003c/strong\u003e] as well as the \u003cstrong\u003eHelmholtz Program Energy System Design \u003c/strong\u003e(ESD) [grant number:\u003cstrong\u003e37.12.03\u003c/strong\u003e].\u003c/p\u003e\n\n\u003ch3\u003eAuthor contributions\u003c/h3\u003e\n\u003cp\u003e\u003cstrong\u003eE.J.:\u003c/strong\u003e Writing \u0026ndash; original draft, Visualization, Validation, Software, Methodology, Investigation, Formal analysis, Data curation, Conceptualization.\u003cstrong\u003e A.A.:\u003c/strong\u003e Conceptualization, Writing \u0026ndash; review \u0026amp; editing, Supervision.\u003c/p\u003e\n\n\u003ch3\u003eCompeting interests\u003c/h3\u003e\n\u003cp\u003eThe authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eVGBE-TW-103V. \u003cem\u003eAvailability of Power Plants 2014 \u0026ndash; 2023\u003c/em\u003e. (vgbe energy service GmbH, Essen, 2024).\u003c/li\u003e\n\u003cli\u003eENTSO-E Transparency Platform. UnavailabilityOfGenerationUnits_15.1.A_B. \u003cem\u003eTransparency Platform HelpDesk\u003c/em\u003e https://transparencyplatform.zendesk.com/hc/en-us/articles/40477405498257-UnavailabilityOfGenerationUnits-15-1-A-B (2025).\u003c/li\u003e\n\u003cli\u003eGjorgiev, B., Stankovski, A., Wengler, J., Sencan, S. \u0026amp; Sansavini, G. Availability of the European power system assets: What we learn from data? \u003cem\u003eReliab. Eng. Syst. Saf.\u003c/em\u003e \u003cstrong\u003e258\u003c/strong\u003e, (2025).\u003c/li\u003e\n\u003cli\u003eBassini, C. F. Seasonal and weather influences in the outages of German thermal generators. \u003cem\u003eACM SIGEnergy Energy Inform. Rev.\u003c/em\u003e \u003cstrong\u003e5\u003c/strong\u003e, (2025).\u003c/li\u003e\n\u003cli\u003eIEEE SA. \u003cem\u003e762-2023 - IEEE Standard Definitions for Use in Reporting Electric Generating Unit Reliability, Availability, and Productivity\u003c/em\u003e. (IEEE, New York, NY, 2023).\u003c/li\u003e\n\u003cli\u003eN-SIDE. \u003cem\u003eStudy on the Outage Parameters of Generation Units and DC Links\u003c/em\u003e. (2022).\u003c/li\u003e\n\u003cli\u003ePecinovsky, J. \u0026amp; Boerman, F. entsoe-py. (2025).\u003c/li\u003e\n\u003cli\u003eENTSO-E Transparency Platform. Planned Unavailability \u0026amp; Changes in Actual Availability of Generation \u0026amp; Production Units [15.1.A] \u0026amp; [15.1.B] \u0026amp; [15.1.C] \u0026amp; [15.1.D]. \u003cem\u003eTransparency Platform HelpDesk\u003c/em\u003e https://transparencyplatform.zendesk.com/hc/en-us/articles/16652173943828-Planned-Unavailability-Changes-in-Actual-Availability-of-Generation-Production-Units-15-1-A-15-1-B-15-1-C-15-1-D (2025).\u003c/li\u003e\n\u003cli\u003eENTSO-E Transparency Platform. New Transparency Platform Website Go Live. \u003cem\u003eNews\u003c/em\u003e https://transparency.entsoe.eu/news (2025).\u003c/li\u003e\n\u003cli\u003eENTSO-E Transparency Platform. Unavailability of Production and Generation Units [15.1.A\u0026amp;B\u0026amp;C\u0026amp;D]. \u003cem\u003eTransparency Platform HelpDesk\u003c/em\u003e https://transparencyplatform.zendesk.com/hc/en-us/articles/32492060810513-UnavailabilityOfProductionAndGenerationUnits-15-1-A-B-C-D-r3 (2025).\u003c/li\u003e\n\u003cli\u003eNERC. \u003cem\u003eGenerating Availability Data System - Data Reporting Instructions\u003c/em\u003e. (2025).\u003c/li\u003e\n\u003cli\u003eNorth American Electric Reliability Corporation (NERC). Derate Event Reporting - Data Reporting Instructions - Section III. (2025).\u003c/li\u003e\n\u003cli\u003eNorth American Electric Reliability Corporation (NERC). Outage Event Reporting - Data Reporting Instructions - Section III. (2025).\u003c/li\u003e\n\u003cli\u003eSchmitz, R., Frischmuth, F., Braun, M. \u0026amp; H\u0026auml;rtel, P. Coping with Risk Factors in Energy System Transformations - Climate Change Impacts on Nuclear Power Plant Availability in Europe. in \u003cem\u003e2024 20th International Conference on the European Energy Market (EEM)\u003c/em\u003e (IEEE, Istanbul, Turkiye, 2024). doi:10.1109/EEM60825.2024.10608936.\u003c/li\u003e\n\u003cli\u003eFinck, J. R. Techno-economic assessment of market coupling regimes in future electricity systems. (Karlsruher Institut f\u0026uuml;r Technologie, Karlsruhe, Deutschland, 2024).\u003c/li\u003e\n\u003cli\u003eGils, H. C., Bothor, S., Genoese, M. \u0026amp; Cao, K.-K. Future security of power supply in Germany - The role of stochastic power plant outages and intermittent generation. \u003cem\u003eInt. J. Energy Res.\u003c/em\u003e \u003cstrong\u003e42\u003c/strong\u003e, (2018).\u003c/li\u003e\n\u003cli\u003eDeakin, M., Greenwood, D., Brayshaw, D. J. \u0026amp; Bloomfield, H. Comparing Generator Unavailability Models with Empirical Distributions from Open Energy Datasets. in \u003cem\u003e2022 17th International Conference on Probabilistic Methods Applied to Power Systems (PMAPS)\u003c/em\u003e (IEEE, Manchester, United Kingdom, 2022). doi:10.1109/PMAPS53380.2022.9810629.\u003c/li\u003e\n\u003cli\u003eKanellopoulos, K., De Felice, M., Hidalgo Gonzalez, I. \u0026amp; Bocin, A. JRC Open Power Plants Database (JRC-PPDB-OPEN). Zenodo https://doi.org/10.5281/ZENODO.3266807 (2019).\u003c/li\u003e\n\u003cli\u003eGotzens, F., Heinrichs, H., H\u0026ouml;rsch, J. \u0026amp; Hofmann, F. Performing energy modelling exercises in a transparent way - The issue of data quality in power plant databases. Zenodo https://doi.org/10.5281/ZENODO.14785651 (2025).\u003c/li\u003e\n\u003cli\u003eSergio, A. \u0026amp; Colelli, F. P. Weather-induced power plant outages: Empirical evidence from hydro and thermal generators in Europe. \u003cem\u003eEnergy Econ.\u003c/em\u003e \u003cstrong\u003e148\u003c/strong\u003e, 108549 (2025).\u003c/li\u003e\n\u003cli\u003eMurphy, S., Sowell, F. \u0026amp; Apt, J. A time-dependent model of generator failures and recoveries captures correlated events and quantifies temperature dependence. \u003cem\u003eAppl. Energy\u003c/em\u003e \u003cstrong\u003e253\u003c/strong\u003e, (2019).\u003c/li\u003e\n\u003cli\u003eErsayin, E. \u0026amp; Ozgener, L. Performance analysis of combined cycle power plants: A case study. \u003cem\u003eRenew. Sustain. Energy Rev.\u003c/em\u003e \u003cstrong\u003e43\u003c/strong\u003e, 832\u0026ndash;842 (2015).\u003c/li\u003e\n\u003cli\u003eGu\u0026eacute;nand, Y. \u003cem\u003eet al.\u003c/em\u003e Climate change impact on nuclear power outages - Part I: A methodology to estimate hydro-thermic environmental constraints on power generation. \u003cem\u003eEnergy\u003c/em\u003e \u003cstrong\u003e307\u003c/strong\u003e, (2024).\u003c/li\u003e\n\u003cli\u003eCollet, L. \u003cem\u003eet al.\u003c/em\u003e Future nuclear power outages in a changing climate - A case study on two contrasted French power plants. \u003cem\u003eEnergy\u003c/em\u003e \u003cstrong\u003e320\u003c/strong\u003e, (2025).\u003c/li\u003e\n\u003cli\u003eCurley, G. M. Reliability Analysis of Power Plant Unit Outage Problems. (2013).\u003c/li\u003e\n\u003cli\u003eENTSO-E Transparency Platform. Actual Generation per Production Type [16.1.B\u0026amp;C]. \u003cem\u003eTransparency Platform HelpDesk\u003c/em\u003e https://transparencyplatform.zendesk.com/hc/en-us/articles/16648290299284-Actual-Generation-per-Production-Type-16-1-B-C (2025).\u003c/li\u003e\n\u003cli\u003eENTSO-E Transparency Platform. Installed Generation Capacity Aggregated [14.1.A]. \u003cem\u003eTransparency Platform HelpDesk\u003c/em\u003e https://transparencyplatform.zendesk.com/hc/en-us/articles/16648300912916-Installed-Generation-Capacity-Aggregated-14-1-A#:~:text=A%5D,-2%20months%20ago\u0026amp;text=The%20sum%20of%20installed%20Net,generation%20capacity%2C%20per%20production%20type. (2025).\u003c/li\u003e\n\u003cli\u003eENTSO-E. \u003cem\u003eEuropean Resource Adequacy Assessment 2023 Edition - Annex 1: Input Data and Assumptions\u003c/em\u003e. https://www.entsoe.eu/eraa/2023/report/ERAA_2023_Annex_1_Assumptions.pdf (2024).\u003c/li\u003e\n\u003cli\u003eENTSO-E. \u003cem\u003eTYNDP 2024 Scenarios Methodology Report - Final Version\u003c/em\u003e. https://2024.entsos-tyndp-scenarios.eu (2025).\u003c/li\u003e\n\u003cli\u003eAntares-Simulator Team. Antares-Simulator. \u003cem\u003eAntares Simulator Documentation\u003c/em\u003e https://antares-simulator.readthedocs.io/en/latest/user-guide/ts-generator/05-algorithm/ (2025). \u003c/li\u003e\n\u003c/ol\u003e"},{"header":"Footnotes","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003e As it turned out that there were still duplicates of the identification codes, the line with the lowest available capacity was selected to cover the worst-case scenario.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003e The difference of 2,536 entries is due to the fact that additional countries were filtered and the start and end dates of the report must be between January 2015 and October 2025.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003e Northern Ireland is included in Great Britain. Iceland, Malta, Ukraine, the Azores, the Canary Islands, the Balearic Islands, and other overseas territories of France or Great Britain were not included.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003e The term \u0026ldquo;outage\u0026rdquo; in this list includes not only total outages, but also performance limitations and equivalents consisting of total and partial outages.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":true,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-9138908/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9138908/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eReliable representation of power plant availability is essential for energy system analysis, electricity market modelling and security-of-supply studies, yet openly accessible datasets are scarce and heterogeneous. This data set presents a harmonised set of performance measures for thermal and hydropower units in the ENTSO-E area derived from monthly unavailability reports on the ENTSO-E Transparency Platform. We reconstruct unit-level availability time series from event-based outage records, apply transparent filtering, clustering and interpolation procedures, and enhance the data with technology and commissioning information from merged open power plant databases. From these processed time series we compute hourly measures and a broad range of time-based performance indicators, including availability factors, outage factors and rates, failure and repair rates, and statistics on partial derations, in both capacity-weighted and unweighted form. The resulting indicators are provided by country, season, plant type and technology, and are validated against historical generation records, selected adequacy studies as well as other journal publications. The accompanying open-source Python workflow enables full reproducibility and direct reuse in data processing or modelling. In addition, two example algorithms are included that can be used to simulate outage series for individual generation units or aggregated power plant fleets.\u003c/p\u003e","manuscriptTitle":"Performance measures and availability of European power plants","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-03-25 06:26:25","doi":"10.21203/rs.3.rs-9138908/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"949881f6-9fca-4538-ac7e-b947a9a27e55","owner":[],"postedDate":"March 25th, 2026","published":true,"recentEditorialEvents":[{"type":"editorInvitedReview","content":"","date":"2026-05-18T18:01:23+00:00","index":26,"fulltext":""}],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2026-03-30T18:53:25+00:00","versionOfRecord":[],"versionCreatedAt":"2026-03-25 06:26:25","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9138908","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9138908","identity":"rs-9138908","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}