Classification of Power Quality Events Using Deep Learning on Event Images

Balouji E., SALOR DURNA Ö.

3rd International Conference on Pattern Analysis and Image Analysis (IPRIA), Shahr-e Kord, Iran, 19 - 20 April 2017, pp.216-221 identifier identifier

  • Publication Type: Conference Paper / Full Text
  • Volume:
  • Doi Number: 10.1109/pria.2017.7983049
  • City: Shahr-e Kord
  • Country: Iran
  • Page Numbers: pp.216-221
  • Keywords: Data classification, deep learning, multi-layer convolution neural network (MLCNN), power quality (PQ)


In this paper, a new method for the classification of power quality (PQ) events of the electricity networks based on deep learning approach is presented. In contrast with the existing PQ event data analysis techniques, sampled voltage data of the PQ events are not used, but image files of the three-phase PQ event data are analyzed by taking the advantage of the success of the deep leaning approach on image-file-classification. Therefore, the novelty of the proposed approach is that, image files of the voltage waveforms of the three phases of the power grid are classified. PQ events obtained from four transformer substations of the electricity transmission system for a year are used for training and testing the proposed classification method. DIGITS deep learning platform of NVIDIA is employed for the application of the deep learning algorithm on PQ event data images. It is shown that the test data can be classified with 100% accuracy. The proposed work is believed to serve the needs of the future smart grid applications, which are fast and automatic analysis of the electricity grid and taking automatic countermeasures against potential PQ events.