eduzhai > Applied Sciences > Engineering >

Learning-Based Real-Time Event Identification Using Rich Real PMU Data

  • Save

... pages left unread,continue reading

Document pages: 8 pages

Abstract: A large-scale deployment of phasor measurement units (PMUs) that reveal theinherent physical laws of power systems from a data perspective enables anenhanced awareness of power system operation. However, the high-granularity andnon-stationary nature of PMU time series and imperfect data quality could bringgreat technical challenges to real-time system event identification. To addressthese issues, this paper proposes a two-stage learning-based framework. At thefirst stage, a Markov transition field (MTF) algorithm is exploited to extractthe latent data features by encoding temporal dependency and transitionstatistics of PMU data in graphs. Then, a spatial pyramid pooling (SPP)-aidedconvolutional neural network (CNN) is established to efficiently and accuratelyidentify operation events. The proposed method fully builds on and is alsotested on a large real dataset from several tens of PMU sources (and thecorresponding event logs), located across the U.S., with a time span of twoconsecutive years. The numerical results validate that our method has highidentification accuracy while showing good robustness against poor dataquality.

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×