Deep learning for big power system data

New advanced signal processing methods, that ho under the name "deep learning", will be applied to large amounts of measurement data ("big data") of voltage and current in electrical power systems. The project will develop methods for the analysis of electric power data with the help of deep-learning techniques, where the focus will be on two specific applications: long time series with high time resolution of voltage and current in the vicinity of solar and wind power installations; large numbers of power quality events. In both applications, the analysis will be aimed towards the cause of the disturbance. The project is a collaboration between signal processing and power engineering researchers and is a continuation of a previous, very successful cooperation.

Partner organizations

  • Luleå University of Technology (Academic, Sweden)
Start date 01/01/2017
End date The project is closed: 31/12/2020

Page manager Published: Thu 03 Jan 2019.