Sparse representation-based residual current waveform automatically identifying method

A residual current, sparse representation technology, applied in character and pattern recognition, instruments, calculations, etc., can solve problems such as protector malfunction, waveform misrecognition, refusal to operate, etc., to achieve good recognition effect and avoid misrecognition effect.

CN107578016AActive Publication Date: 2018-01-12CHONGQING UNIV
6 Cites 2 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Publication Date
2018-01-12

Smart Images

  • Figure 1
    Figure 1
  • Figure 2
    Figure 2
  • Figure 3
    Figure 3
Patent Text Reader

Abstract

The invention discloses a sparse representation-based residual current waveform automatically identifying method. The sparse representation-based residual current waveform automatically identifying method comprises collecting residual current signals of N different devices during leakage fault through a residual current device (RCD), performing de-noising treatment to acquire N preprocessed signals as a sample set; extracting time-frequency domain characters of signals inside the sample set to acquire character vectors, and performing normalization treatment; combining the normalized charactervectors into a character matrix of all the signals as an over-complete dictionary of the N residual current signals; collecting residual current signals during leakage fault through the RCD, performing de-noising treatment to obtain a sample to be tested, and extracting the character vectors of the sample to be tested; performing sparse representation on the extracted character vectors through the over-complete dictionary; inputting the sparse representation of every signal into a sparse representation classifier to acquire an identification result of the type of residual current signals to be tested. The sparse representation-based residual current waveform automatically identifying method can improve effectiveness and accuracy of residual current waveform identification.
Need to check novelty before this filing date? Find Prior Art

Description

technical field

[0001] The invention relates to the technical field of pattern recognition of residual current signals in power systems, in particular to an automatic recognition method for residual current waveforms based on sparse representation. Background technique

[0002] With the advancement of science and technology and the continuous development of smart grids, the types of power electronic equipment in urban and rural low-voltage power grids are increasing. When a leakage fault occurs on a line, the residual current generated is no longer a single power frequency sinusoidal current, but has a pulsating DC component and high frequency AC. components, even complex waveforms such as smooth DC components. In this case, the traditional AC-type residual current-operated protector has been unable to effectively protect this complex AC-DC residual current, so A-type and B-type residual current-operated protectors have become a research hotspot in leakage protection. [00...

Examples

specific Embodiment approach

[0018] 1. Use the residual current transformer (RCD) to collect residual current data as a sample data set for constructing a dictionary. The specific implementation is as follows:

[0019] Use the residual current transformer (RCD) to collect the residual current signals of N different types of equipment loads when leakage faults occur, including sine AC, half-wave, 90° wave, 135° wave, half-wave superimposed 6mA smooth DC wave, half-wave Superimpose N types of 10mA smooth DC wave and high-frequency AC residual current, etc., filter out the noise signal through a filter, and then select the same residual current signal to discretize it to obtain a sub-sample data set:

[0020]

[0021] Among them, i=1,2,...,N represent different residual current signal types, M represents the number of sample points after discretization of each residual current, and the selected residual current samples contain data of multiple periods of the residual current signal as much as possible, t...