|Home | About | Journals | Submit | Contact Us | Français|
This article presents an open data repository, the methodology to generate it and the associated data processing software developed to consolidate an hourly snapshot historical data set for the year 2015 to an equivalent Nordic power grid model (aka Nordic 44), the consolidation was achieved by matching the model׳s physical response w.r.t historical power flow records in the bidding regions of the Nordic grid that are available from the Nordic electricity market agent, Nord Pool.
The model is made available in the form of CIM v14, Modelica and PSS/E (Siemens PTI) files. The Nordic 44 model in Modelica and PSS/E were first presented in the paper titled “iTesla Power Systems Library (iPSL): A Modelica library for phasor time-domain simulations” (Vanfretti et al., 2016)  for a single snapshot. In the digital repository being made available with the submission of this paper (SmarTSLab_Nordic44 Repository at Github, 2016) , a total of 8760 snapshots (for the year 2015) that can be used to initialize and execute dynamic simulations using tools compatible with CIM v14, the Modelica language and the proprietary PSS/E tool are provided. The Python scripts to generate the snapshots (processed data) are also available with all the data in the GitHub repository (SmarTSLab_Nordic44 Repository at Github, 2016) .
This Nordic 44 equivalent model was also used in iTesla project (iTesla)  to carry out simulations within a dynamic security assessment toolset (iTesla, 2016) , and has been further enhanced during the ITEA3 OpenCPS project (iTEA3) . The raw, processed data and output models utilized within the iTesla platform (iTesla, 2016)  are also available in the repository. The CIM and Modelica snapshots of the “Nordic 44” model for the year 2015 are available in a Zenodo repository.
Value of the data
Raw data from Nord Pool was consolidated with a physical model representation of the Nordic grid by matching the measurement records to the model׳s power flow results (processed data). This has resulted in thousands of representations of the “Nordic 44” model for the year 2015 that are made available in the GitHub repository with the submission of this article.
These snapshots are provided in the form of CIM v14, Modelica and PSS/E (Siemens PTI) files. The Python scripts (i.e. software toolset) used to generate these snapshots (CIM v14, Modelica and PSS/E) are also made available in the repository.
The pre-processed (raw) data was downloaded from Nord Pool webpage. For each hour, the records contain the active power production and consumption data in the bidding regions of the Nordic grid and the active power exchange between them. These records were matched to an electrical grid model steady-state response, as explained next.
To match and consolidate the historical market data to a physical description of the power network, the Nordic 44 model was developed. Note that the aim here was to set a “base case”, from which multiple snapshots of actual measurements records could be mapped to the quantitative response from computations on a physical model of the grid that included both steady-state and dynamic analysis features.
The development of this source model/“data” consisted of (a) obtaining editable files in the PSS/E form to, (b) extend the model to represent better the Norwegian portions of the Nordic grid and to adapt it to modeling limitations in both the iTesla platform and the iPSL library (i.e. lack of HVDC link models).
The model developed in Step (b) is used throughout the historical data matching and consolidation process to create snapshots of the actual operational conditions of the Nordic Grid for 2015. This base case can be found in the repository in the GitHub repository at./nordic44/models/ and archived in Models.zip in respective Zenodo repositories.
The pre-processed (raw) data was downloaded from Nord Pool webpage as MS-Excel files (Consumption_xx.xlsx, Exchange_xx.xlsx, and Production_xx.xlsx) were saved in a folder with corresponding date (e.g. N44_20150401 refers to the folder containing the hourly snapshots of the 1st April 2015). The MS-Excel files (Production_xx.xlsx and Consumption_xx.xlsx) contain the active power production and consumption data in the bidding regions of the Nordic grid for every hour as shown in Fig. 1. The MS-Excel file (Exchange_xx.xlsx) contains the active power exchange data between the bidding regions of the Nordic grid for every hour. Now, the toolset was updated to automatically download the data from the Nord Pool ftp server, inspired by the work in . The raw data is stored in a Python dictionary allowing for easy data manipulation in Python and integration with PSS/e through psspy. It is still possible to save the data to excel and to load in excel files to the dictionary ensuring both forward and backward compatibility.
For each hour, an equivalent Nordic power system model (Nordic 44) was created, matched and consolidated with the load generation balance in the bidding regions and the active power exchange between the bidding regions. The buses in the Nordic 44 model were named according to the closest city or town corresponding to both the geographical location and the detailed grid structure available to Statnett SF. The work flow used to create the PSS/E snapshots is shown below in Fig. 2.
The major steps of the data processing workflow include:
At the end of this work flow, MS-Excel files (PSSE_in_out.xlsx) are generated by the Python script for every snapshot with raw data from Nord Pool and the results from PSS/E. These MS-Excel files (PSSE_in_out.xlsx) include limit checking messages (branch overloading, bus voltage out of limits and generator overloading). A screenshot of the created Excel file is shown in Fig. 3.
The PSS/E snapshots for each hour before solving the power flow (hx_before_PF.raw, unmatched processed data) and after solving the power flow (hx_after_PF.raw, matched processed data) were also made available in the repository. The Nordic 44 model and the PSS/E snapshots can be accessed from the repository as shown in Fig. 4.
CIM v14 and Modelica snapshots were generated from the matched PSS/E solved power flow snapshots (processed data). The PSS/E snapshots and PSS/E dynamic model parameters data (.dyr) files were used by the Python script to generate the CIM v14 snapshots as shown in Fig. 5. The Python script uses the Application Programming Interface (API) of the Operational Database Management System (PSS/ODMS) software  to generate the snapshots. These generated CIM snapshots can be used for information exchange according to CIM and to perform analysis in CIM compliant tools , .
The generated CIM snapshots were placed in the folder corresponding to the day they refer to (e.g. N44_20150401 refers to the 1st of April 2015). In each folder there are three files that define individual CIM snapshots for each hour (N44_hx_EQ.xml, N44_hx_SV.xml and N44_hx_TP.xml). N44_noOL_RDFIDMAP.xml is the file with IDs mapping of those cases with fixed overloading problems. N44_RDFIDMAP_2015-1.xml and N44_RDFIDMAP_2015-2.xml are the files with IDs mapping of the remaining snapshots from 2015. The screenshot of the generated CIM snapshots in the GitHub repository is shown in Fig. 6.
PSS/E snapshots and PSS/E dynamic model parameters data (.dyr) files were used by the Python script “Raw2Record” (./…/SmarTS-Lab/Raw2Record) to generate the associated Modelica snapshots as shown in Fig. 7. These generated Modelica snapshots together with the OpenIPSL library can be used for simulation in the Modelica compliant tools. The generation of the record files using the python script is illustrated with an example (./examples/multiple_data_sets_from_nordpool.py /data_set_from_nordpool.py) and is provided in the GitHub repository.
During the iTesla project, another methodology to generate Modelica model snapshots using the iTesla platform  was attempted. Note that these snapshots were created using a different methodology reported in . The resulting snapshots available for this approach cover only from April 1, 2015 to July 31, 2015 and are archived in iTesla_Platform.zip in the Zenodo repository. This archive contains both the snapshot models together with the appropriate version of the iPSL library.
The data records are stored in the /ModelicaSnapshots/ sub-folder of the Zenodo archive. They require the Modelica model that was manually implemented for Nordic 44 reported in  with the record structures corresponding to the PSS/E snapshots. The Python script (“torecord”) used to generate Modelica snapshots from PSS/E snapshots as described in Fig. 7. The OpenIPSL library used in generation of these snapshots can be found in./…/SmarTS-Lab/OpenIPSL (Fig. 8).
The CIM and Modelica snapshots (processed data) of the “Nordic 44” model for the year 2015 are available the aforementioned Zenodo repository (see ).
The Nordic44 data and model reported in this paper were used in the iTesla project (iTesla)  to carry out simulations within a dynamic security assessment toolset (iTesla, 2016) . The raw, processed data and output models utilized within the iTesla platform . The Modelica modeling has been further enhanced during the ITEA3 OpenCPS project (iTEA3) . iTesla (2016)  are also available in the repository. The CIM and Modelica snapshots of the “Nordic 44” model for the year 2015 are available in a Zenodo repository (see ).
Therefore, the authors would like to acknowledge the contributions of Emil Hillberg of STRI that during the FP7 iTesla project by provided the source files for the Nordic 44 model and his suggested modifications to it as described in Section 1.2, step (a). The financial support from following funding bodies and projects is gratefully acknowledged:
Transparency documentTransparency data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.dib.2017.02.021.