•  
  •  
 

Keywords

power quality, disturbance recognition, soft threshold function, deep belief network

Abstract

Aiming at the problem that the recognition accuracy of multiple disturbances is not high under noise interference, a new classification method of power quality disturbances based on deep belief network is proposed. Firstly, the stationary wavelet multi?scale transformation is performed on the power quality disturbance signal, and then the soft threshold function is used to process the estimated wavelet coefficients to reconstruct the original signal, thereby realizing the denoising of the power quality disturbance signal. Moreover, it is further proposed to use the deep belief network to classify and identify the reconstructed single disturbance signal and multiple disturbance signals. The calculation example shows that even under the interference of 20 dB noise, the classification accuracy rate is as high as 93%. The results show that the recognition accuracy of the method is high for 7 kinds of single disturbance and 13 kinds of multiple disturbance signals, which verifies that the method has strong anti?noise interference ability.

DOI

10.19781/j.issn.1673-9140.2023.01.020

First Page

171

Last Page

177

Share

COinS