CN-122024926-A - Transformer oil chromatographic data anomaly identification method, device and equipment based on transducer characteristic collection of tunes of poems

CN122024926ACN 122024926 ACN122024926 ACN 122024926ACN-122024926-A

Abstract

The invention discloses a Transformer oil chromatographic data anomaly identification method based on a Transformer characteristic collection of tunes of poems, which comprises the steps of receiving input historical oil chromatographic time sequence data of a Transformer and preprocessing the historical oil chromatographic time sequence data through a preprocessing rule to obtain corresponding preprocessing data; the method comprises the steps of mapping the preprocessed data according to a preset collection of tunes of poems mapping rule to obtain corresponding features collection of tunes of poems, encoding the features collection of tunes of poems according to a preset feature collection of tunes of poems encoding model to generate corresponding fault collection of tunes of poems vectors, and carrying out abnormality recognition on the fault collection of tunes of poems vectors according to a preset abnormality recognition model to obtain corresponding abnormality recognition results. According to the Transformer oil chromatographic data anomaly identification method based on the Transformer characteristic collection of tunes of poems, the historical oil chromatographic time sequence data of the Transformer is mapped and encoded to obtain the fault collection of tunes of poems vector, and then the fault collection of tunes of poems vector is subjected to anomaly identification to obtain an accurate anomaly identification result, so that the accuracy of anomaly identification of the Transformer is greatly improved.

Inventors

FEI TAO
WU CHENGFU
DING YUEQIANG
DUAN KAIYUAN
SHI NING
LAI JUN
PENG CHENGUANG
CHENG XINGMIN
HAN CHUNLEI
JIANG WENDONG
ZHANG YONG
ZONG WANLI
JI RONGRONG
SUN LINTAO

Assignees

国网浙江省电力有限公司超高压分公司

Dates

Publication Date: 20260512
Application Date: 20260128

Claims (10)

1. A Transformer oil chromatographic data anomaly identification method based on a transducer characteristic collection of tunes of poems, the method comprising: Receiving input historical oil chromatographic time sequence data of the transformer and preprocessing the historical oil chromatographic time sequence data through a preprocessing rule to obtain corresponding preprocessed data; mapping the preprocessed data according to a preset collection of tunes of poems mapping rule to obtain corresponding features collection of tunes of poems, wherein the features collection of tunes of poems comprise oil chromatographic word vectors of each time node; The method comprises the steps of encoding the characteristic collection of tunes of poems according to a preset characteristic collection of tunes of poems encoding model to generate a corresponding fault collection of tunes of poems vector, wherein the fault collection of tunes of poems vector is used for representing the running state of a Transformer in a continuous period, and the characteristic collection of tunes of poems encoding model is a Transformer model; And carrying out anomaly identification on the fault collection of tunes of poems vector according to a preset anomaly identification model to obtain a corresponding anomaly identification result.
2. The Transformer oil chromatographic data anomaly identification method based on the Transformer characteristic collection of tunes of poems according to claim 1, wherein the receiving the input historical oil chromatographic time series data of the Transformer and preprocessing the historical oil chromatographic time series data by a preprocessing rule to obtain corresponding preprocessed data comprises the following steps: performing outlier elimination on the historical oil chromatographic time sequence data according to an outlier range in the preprocessing rule to obtain corresponding effective time sequence data; And carrying out missing value supplementation on the effective time sequence data according to a missing value supplementation function in the preprocessing rule to obtain corresponding preprocessing data.
3. The Transformer oil chromatographic data anomaly identification method based on the Transformer features collection of tunes of poems of claim 1, wherein the mapping the pre-processed data according to a preset collection of tunes of poems mapping rule to obtain the corresponding features collection of tunes of poems comprises: extracting corresponding multidimensional standard feature vectors from the preprocessing data; Mapping the multi-dimensional standard feature vector to an initial word vector space with a fixed dimension according to a linear embedding matrix in the collection of tunes of poems mapping rule to obtain an initial embedding vector; and superposing residual enhancement on the initial embedded vector according to a residual enhancement function in the collection of tunes of poems mapping rule to obtain a corresponding enhancement vector serving as the feature collection of tunes of poems.
4. The Transformer oil chromatographic data anomaly identification method based on the Transformer features collection of tunes of poems of claim 3, wherein the extracting the corresponding multidimensional feature vector from the preprocessed data comprises: acquiring concentration values corresponding to key characteristic gases of all time nodes; calculating typical ratio characteristics corresponding to concentration values of nodes at the same time; And combining the concentration value of each time node with the characteristic ratio to obtain a multidimensional characteristic vector.
5. The Transformer oil chromatographic data anomaly identification method based on the Transformer characteristic collection of tunes of poems according to claim 4, wherein the combining the concentration value of each time node with the typical ratio characteristic to obtain the multidimensional feature vector includes: Combining the concentration value of each time node with the characteristic ratio to obtain a corresponding combined characteristic vector; And carrying out normalization processing on the combined feature vector according to a preset normalization function to obtain a standard feature vector serving as a corresponding multidimensional feature vector.
6. The Transformer oil chromatographic data anomaly identification method based on the Transformer features collection of tunes of poems of claim 1, wherein the encoding the features collection of tunes of poems according to a preset feature collection of tunes of poems encoding model to generate corresponding fault collection of tunes of poems vectors comprises: intercepting a vector sequence corresponding to a continuous period of the sliding time window from the feature collection of tunes of poems according to the sliding time window set in the feature collection of tunes of poems coding model; adding position codes for the vector value of each time node in the vector sequence according to the code measurement in the characteristic collection of tunes of poems coding model to obtain a corresponding vector positioning sequence; Inputting the vector positioning sequence into an encoder of the characteristic collection of tunes of poems coding model, and performing oil chromatography time mode modeling on the coded vector sequence by using an attention mechanism of the encoder to obtain a corresponding coded vector sequence; Generating a fault collection of tunes of poems vector corresponding to the coded vector sequence through a hidden state layer of the feature collection of tunes of poems coding model.
7. The Transformer oil chromatographic data anomaly identification method based on the Transformer characteristic collection of tunes of poems according to claim 1, wherein the anomaly identification is performed on the fault collection of tunes of poems vector according to a preset anomaly identification model to obtain a corresponding anomaly identification result, including: Inputting the fault collection of tunes of poems vector by an input layer of the anomaly identification model; acquiring a node value which is obtained by the abnormal recognition model, carries out association analysis on the fault collection of tunes of poems vector and is correspondingly output by each output node; and acquiring the type corresponding to the output node with the maximum node value as a corresponding abnormal recognition result.
8. A Transformer oil chromatographic data anomaly identification device based on a transducer signature collection of tunes of poems, wherein the device is configured to perform the Transformer oil chromatographic data anomaly identification method based on a transducer signature collection of tunes of poems as defined in any one of claims 1 to 7, the device comprising: the preprocessing unit is used for receiving the input historical oil chromatographic time sequence data of the transformer and preprocessing the historical oil chromatographic time sequence data through a preprocessing rule to obtain corresponding preprocessed data; The mapping unit is used for mapping the preprocessing data according to a preset collection of tunes of poems mapping rule to obtain corresponding features collection of tunes of poems, wherein the features collection of tunes of poems comprise oil chromatographic word vectors of each time node; The encoding unit is used for encoding the characteristic collection of tunes of poems according to a preset characteristic collection of tunes of poems encoding model to generate a corresponding fault collection of tunes of poems vector, wherein the fault collection of tunes of poems vector is used for representing the running state of the transformer in a continuous period; and the abnormality recognition unit is used for carrying out abnormality recognition on the fault collection of tunes of poems vector according to a preset abnormality recognition model to obtain a corresponding abnormality recognition result.
9. A computer device, comprising a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other through the communication bus; a memory for storing a computer program; A processor for implementing the Transformer oil chromatographic data anomaly identification method based on the transducer features collection of tunes of poems of any one of claims 1-7 when executing a program stored on a memory.
10. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the Transformer oil chromatographic data anomaly identification method based on the Transformer feature collection of tunes of poems as defined in any one of claims 1 to 7.

Description

Transformer oil chromatographic data anomaly identification method, device and equipment based on transducer characteristic collection of tunes of poems Technical Field The invention relates to the technical field of power equipment state monitoring and fault diagnosis, in particular to a Transformer oil chromatographic data anomaly identification method, device and equipment based on a Transformer characteristic collection of tunes of poems. Background The transformer in the power equipment is key equipment for boosting or reducing conversion, a coil is arranged in the transformer, and the coil is soaked in transformer oil, so that insulation and heat dissipation of the coil are realized. In the prior art, whether the transformer is abnormal or fails is generally judged by observing the color change of the transformer oil, however, the judgment method needs to be judged according to human experience, and the judgment standards of different people are different, so that the accuracy of abnormality identification is poor. Therefore, the process of abnormal identification of transformer oil in the prior art method has the problem of insufficient accuracy. Disclosure of Invention The embodiment of the invention provides a Transformer oil chromatographic data anomaly identification method, device and equipment based on a Transformer characteristic collection of tunes of poems, and aims to solve the problem of insufficient accuracy in the process of anomaly identification of Transformer oil in the prior art. In a first aspect, an embodiment of the present invention provides a Transformer oil chromatographic data anomaly identification method based on a Transformer feature collection of tunes of poems, where the method includes: Receiving input historical oil chromatographic time sequence data of the transformer and preprocessing the historical oil chromatographic time sequence data through a preprocessing rule to obtain corresponding preprocessed data; mapping the preprocessed data according to a preset collection of tunes of poems mapping rule to obtain corresponding features collection of tunes of poems, wherein the features collection of tunes of poems comprise oil chromatographic word vectors of each time node; encoding the features collection of tunes of poems according to a preset feature collection of tunes of poems encoding model to generate a corresponding fault collection of tunes of poems vector, wherein the fault collection of tunes of poems vector is used for representing the running state of the transformer in a continuous period; And carrying out anomaly identification on the fault collection of tunes of poems vector according to a preset anomaly identification model to obtain a corresponding anomaly identification result. In a second aspect, an embodiment of the present invention further provides a Transformer oil chromatographic data anomaly identification device based on a Transformer feature collection of tunes of poems, where the device is configured to perform the Transformer oil chromatographic data anomaly identification method based on the Transformer feature collection of tunes of poems according to the first aspect, and the device includes: the preprocessing unit is used for receiving the input historical oil chromatographic time sequence data of the transformer and preprocessing the historical oil chromatographic time sequence data through a preprocessing rule to obtain corresponding preprocessed data; The mapping unit is used for mapping the preprocessing data according to a preset collection of tunes of poems mapping rule to obtain corresponding features collection of tunes of poems, wherein the features collection of tunes of poems comprise oil chromatographic word vectors of each time node; The encoding unit is used for encoding the characteristic collection of tunes of poems according to a preset characteristic collection of tunes of poems encoding model to generate a corresponding fault collection of tunes of poems vector, wherein the fault collection of tunes of poems vector is used for representing the running state of the transformer in a continuous period; and the abnormality recognition unit is used for carrying out abnormality recognition on the fault collection of tunes of poems vector according to a preset abnormality recognition model to obtain a corresponding abnormality recognition result. In a third aspect, an embodiment of the present invention further provides a computer device, where the device includes a processor, a communication interface, a memory, and a communication bus, where the processor, the communication interface, and the memory complete communication with each other through the communication bus; a memory for storing a computer program; and the processor is used for realizing the Transformer oil chromatographic data anomaly identification method based on the transducer characteristic collection of tunes of poems in the first aspect when executing the p