Columbia University

Synthetic Power Grid Data Set

  • WI
  • SERC
  • FERC

Enhancing power grids performance and resilience has been one of the greatest challenges in science and engineering over the past decade. This challenge spans numerous aspects of the power systems such as incorporating renewable resources, efficient and robust grid monitoring, transmission expansion planning, grid vulnerability analysis and control, cyber security, and energy market design.

Addressing most of these challenges requires real grid topologies with real geographical coordinates. However, in order to avoid exposing vulnerabilities, topologies of the power transmission networks and particularly the locations of the substations and the lines are usually not publicly available or are hard to obtain.  Motivated by this need, we introduced the Network Imitating Method Based on LEarning (NIMBLE) for generating synthetic networks with similar structural and spatial properties. A synthetic network generated by NIMBLE which is based on the Western Interconnection (WI) is available in this page.

Data Description

The following data set is generated based on the statistical properties of the real power grid. Detailed description of the generation procedure and an evaluation of the generated network is provided in:

  • S. Soltan, A. Loh, G. Zussman, “A learning-based method for generating synthetic power grids,” IEEE Systems Journal, vol. 13, no. 1, pp. 625–634, Mar. 2019. [download]

We would appreciate it if you cite this paper when publishing results that use the provided data.

The data set is also available via DR POWER:

Please contact Saleh Soltan with any questions.

Data Files

This data set consists of four tables, each contained in a single comma-delimited CSV file. The files contains the data regarding the buses and lines in the synthetic network generated based on the topology of the Western Interconnection (WI) power grid as described in the paper above. The used parameters  are c = 55, η = 0.5, β = −2.5, and γ = 1.5. The data is organized as follows:

  • Gen_WI_Bus_Locations.csv provides the geographical location of each bus in the generated network. Each row of the corresponding table contains of three fields: the bus number (labeled as Bus Number), the longitude of the bus in degrees (labeled as Lon), and the latitude of the bus in degrees (labeled as Lat).
  • Gen_WI_Lines.csv provides the list of the lines in the generated network. Each row contains four fields: the line number (labeled as Line Number), the bus number at one end of the line (labeled as Bus 1), the bus number at the other end of the line (labeled as Bus 2), and the length of the line (labeled as x) that can be used as the reactance value of that line.
  • Gen_WI_Supply_Values.csv provides the supply at each node in the generated network.  Each row contains two fields: the bus number (labeled as Bus Number) and the supply value in megawatts (labeled as Supply).
  • Gen_WI_Demand_Values.csv provides the demand at each node in the generated network. Each row contains two fields: the bus number (labeled as Bus Number) and the demand value in megawatts (labeled as Demand).

file_downloadDownload Zip File

Additional Publications

  • S. Soltan and G. Zussman, “Generation of synthetic spatially embedded power grid networks,” in Proc. IEEE PES-GM’16, July 2016. [download]
  • S. Soltan and G. Zussman, “Generation of synthetic spatially embedded power grid networks,” in arXiv:1508.04447 [cs.SY], Aug. 2015. [download]

Related Programs and Data Sets