Method for identifying fault type of power distribution network

Abstract

The invention discloses a method for identifying the fault type of a power distribution network. The method is characterized by including the steps: a fault waveform of the power distribution networkis obtained; compressed encoding is conducted on the fault waveform, wherein the compressed encoding includes the step that similarity operation and threshold encoding are conducted on the fault waveform to obtain characteristic compressed codes of the fault waveform; and the characteristic compressed codes are input into a classification model to obtain the type of the fault waveform of the powerdistribution network.

Description

technical field

[0001] The invention relates to the field of electric power technology, in particular to a classification method for distribution network fault types. Background technique

[0002] The distribution network is an important part of the power system. With the rapid development of the smart grid, a large number of uncertain connections of distributed power sources make the fault information of the distribution network more and more complicated, and the accurate and rapid analysis of the fault becomes more and more difficult. In order to ensure the highly intelligent operation of the distribution network, real-time monitoring of feeder operation data, timely warning of abnormal conditions, and rapid fault detection and processing are required. Among them, the identification of distribution network fault types is an important function of intelligent distribution network. The traditional distribution network fault type identification either uses the expert library t...

Examples

Embodiment Construction

[0065] In order to make the purpose, technical solutions and advantages of the present invention clearer, the present invention will be described in further detail below in conjunction with the accompanying drawings.

[0066] Figure 8 It is the original fault waveform diagram. After collecting 16 cycles of the original fault waveform, with 1312 sampling points, three-phase and zero-sequence 4-phase currents are collected, and the fault waveform matrix is ​​obtained as a 1312×4 matrix.

[0067] Intercept the fault waveform segment P with s=3, p=82 i , then get P 1 to P 411 There are 411 waveform segments in total, and the above 411 waveform segments are sequentially input into the template similarity model with 30 waveform templates, ie d=30. The parameters in the deep neural network in the template similarity model are obtained using the training method of the model of the present invention. When performing model training, the parameters in the loss function are set to γ=0...

technical field

[0001] The invention relates to the field of electric power technology, in particular to a classification method for distribution network fault types. Background technique

[0002] The distribution network is an important part of the power system. With the rapid development of the smart grid, a large number of uncertain connections of distributed power sources make the fault information of the distribution network more and more complicated, and the accurate and rapid analysis of the fault becomes more and more difficult. In order to ensure the highly intelligent operation of the distribution network, real-time monitoring of feeder operation data, timely warning of abnormal conditions, and rapid fault detection and processing are required. Among them, the identification of distribution network fault types is an important function of intelligent distribution network. The traditional distribution network fault type identification either uses the expert library t...

Examples

Embodiment Construction

[0065] In order to make the purpose, technical solutions and advantages of the present invention clearer, the present invention will be described in further detail below in conjunction with the accompanying drawings.

[0066] Figure 8 It is the original fault waveform diagram. After collecting 16 cycles of the original fault waveform, with 1312 sampling points, three-phase and zero-sequence 4-phase currents are collected, and the fault waveform matrix is ​​obtained as a 1312×4 matrix.

[0067] Intercept the fault waveform segment P with s=3, p=82 i , then get P 1 to P 411 There are 411 waveform segments in total, and the above 411 waveform segments are sequentially input into the template similarity model with 30 waveform templates, ie d=30. The parameters in the deep neural network in the template similarity model are obtained using the training method of the model of the present invention. When performing model training, the parameters in the loss function are set to γ=0...

Images