CN-121980374-A - Deep learning-based power transmission wire ice edge corona discharge identification and prediction method, system, equipment and storage medium

CN121980374ACN 121980374 ACN121980374 ACN 121980374ACN-121980374-A

Abstract

The invention discloses a deep learning-based power transmission wire ice edge corona discharge identification and prediction method, a system, equipment and a storage medium, and relates to the technical field of power system safety monitoring; the method comprises the steps of training a mixed deep learning model comprising a corona discharge identification model and a time sequence prediction model, carrying out online incremental update on the mixed deep learning model by utilizing new data continuously collected in actual operation, generating identification probability and prediction trend based on the updated mixed deep learning model, calculating risk probability based on Bayesian theory and automatically generating corresponding risk level and early warning information.

Inventors

WU YU
ZHANG YAO
JIANG JIBIN
YIN FANGHUI
ZHOU WENXUAN
Nie Xianglun
HUANG JIE
LI YI
ZHENG XIAOHU
LIU QING
HOU YONGHONG

Assignees

贵州电网有限责任公司

Dates

Publication Date: 20260505
Application Date: 20251128

Claims (10)

1. The deep learning-based power transmission wire ice edge corona discharge identification and prediction method is characterized by comprising the following steps of: Collecting multi-mode data including acoustic, thermal imaging, meteorological and electrical signals, extracting the characteristics of each mode, and fusing to construct a multi-dimensional characteristic vector; Training a hybrid deep learning model comprising a corona discharge recognition model and a time sequence prediction model based on the multidimensional feature vector; performing online incremental update on the hybrid deep learning model by utilizing new data continuously collected in actual operation; Based on the updated mixed deep learning model, the recognition probability and the prediction trend are generated, the risk probability is calculated based on the Bayesian theory, and the corresponding risk level and the early warning information are automatically generated.
2. The deep learning-based power transmission line ice edge corona discharge identification and prediction method as set forth in claim 1, wherein the collecting multi-modal data including acoustic, thermal imaging, weather and electrical signals, extracting and fusing modal features, and constructing a multi-dimensional feature vector comprises: The method comprises the steps of filtering an acoustic signal, framing, calculating a power spectrum, filtering, calculating logarithmic energy, performing discrete cosine transform to obtain a characteristic coefficient, and performing characteristic extraction on a thermal imaging image through a convolutional neural network characteristic extraction network.
3. The deep learning-based power transmission line ice edge corona discharge identification and prediction method as set forth in claim 2, wherein the collecting multi-modal data including acoustic, thermal imaging, weather and electrical signals, extracting and fusing modal features, and constructing a multi-dimensional feature vector further comprises: and fusing the extracted modal features by using an attention mechanism, calculating corresponding attention weights according to the modal feature vectors, and carrying out weighted summation on the modal feature vectors based on the attention weights to obtain fused feature vectors, namely constructed multidimensional feature vectors.
4. A deep learning-based power transmission line ice edge corona discharge identification and prediction method as claimed in claim 3, wherein said online incremental updating of the hybrid deep learning model using new data continuously collected during actual operation comprises: And updating the mixed deep learning model according to a strategy of dynamically adjusting the learning rate along with time by adopting an online gradient descent self-adaptive learning algorithm, wherein the strategy enables the learning rate to be gradually attenuated along with the increase of the time step number.
5. The deep learning-based power transmission line ice edge corona discharge identification and prediction method as set forth in claim 4, wherein the generating the identification probability and the prediction trend based on the updated hybrid deep learning model, calculating the risk probability based on the bayesian theory, and automatically generating the corresponding risk level and the pre-warning information comprises: And generating the identification probability of corona discharge and the prediction trend of time sequence by using the updated mixed deep learning model.
6. The deep learning-based power transmission line ice edge corona discharge identification and prediction method as set forth in claim 5, wherein the generating the identification probability and the prediction trend based on the updated hybrid deep learning model, calculating the risk probability based on bayesian theory, and automatically generating the corresponding risk level and the pre-warning information further comprises: based on Bayesian theory, combining the observation feature vector and the risk event, calculating the probability of risk occurrence, namely calculating the probability of risk occurrence when a specific feature is observed through a Bayesian formula according to the known probability of occurrence of the observation feature when the risk event occurs, the probability of occurrence of the risk event and the probability of occurrence of the observation feature.
7. The deep learning-based power transmission line ice edge corona discharge identification and prediction method as set forth in claim 6, wherein the generating the identification probability and the prediction trend based on the updated hybrid deep learning model, calculating the risk probability based on bayesian theory, and automatically generating the corresponding risk level and the pre-warning information further comprises: and according to the calculated risk probability, classifying the risk into different grades according to a preset risk grade classification standard, and automatically generating corresponding early warning information according to the classified risk grade.
8. Deep learning-based transmission line ice edge corona discharge identification and prediction system, applying the method as claimed in any one of claims 1 to 7, comprising: The multi-mode data fusion characteristic construction module is used for collecting multi-mode data comprising acoustic, thermal imaging, meteorological and electric signals, extracting and fusing the mode characteristics, and constructing a multi-dimensional characteristic vector; the mixed deep learning model training module is used for training a mixed deep learning model comprising a corona discharge identification model and a time sequence prediction model based on the multidimensional feature vector; the online increment updating module is used for carrying out online increment updating on the mixed deep learning model by utilizing new data continuously collected in the actual operation; And the risk assessment and early warning generation module is used for generating recognition probability and prediction trend based on the updated mixed deep learning model, calculating risk probability based on Bayesian theory and automatically generating corresponding risk level and early warning information.
9. An electronic device, comprising: A memory and a processor; The memory is for storing computer executable instructions, the processor being for executing the computer executable instructions which when executed by the processor implement the steps of the method of any one of claims 1to 7.
10. A computer-readable storage medium, characterized in that it stores computer-executable instructions which, when executed by a processor, implement the steps of the method of any one of claims 1 to 7.

Description

Deep learning-based power transmission wire ice edge corona discharge identification and prediction method, system, equipment and storage medium Technical Field The invention relates to the technical field of safety monitoring of power systems, in particular to a deep learning-based power transmission wire ice edge corona discharge identification and prediction method, a deep learning-based power transmission wire ice edge corona discharge identification and prediction system, deep learning-based power transmission wire ice edge corona discharge identification and prediction equipment and a storage medium. Background Icing of the transmission line is an important factor affecting safe and stable operation of the power system. When the surface of the lead forms the ice ridge, the electric field intensity of the tip of the ice ridge increases sharply, and a corona discharge phenomenon is easy to generate. The corona discharge not only can cause electric energy loss, but also can cause flashover accidents, thereby seriously threatening the safety of the power grid. The existing infrared thermal imaging detection, ultraviolet imaging detection, acoustic detection and other technologies have the problems that single-mode detection is easy to generate false alarm or missing alarm, effective prediction and early warning capability is lacked, environmental adaptability is poor, intelligent degree is low, and the problems that different lines and meteorological conditions cannot be adapted, and the like. Disclosure of Invention In view of the above problems, the present invention provides a method, a system, a device and a storage medium for identifying and predicting ice edge corona discharge of a power transmission wire based on deep learning. Therefore, the invention solves the technical problems of low accuracy and insufficient prediction capability in the prior art by integrating various detection means and timing prediction technologies to realize high-precision identification and early warning. In order to solve the technical problems, the invention provides the following technical scheme: In a first aspect, the invention provides a deep learning-based power transmission wire ice edge corona discharge identification and prediction method, which comprises the following steps: Collecting multi-mode data including acoustic, thermal imaging, meteorological and electrical signals, extracting the characteristics of each mode, and fusing to construct a multi-dimensional characteristic vector; Training a hybrid deep learning model comprising a corona discharge recognition model and a time sequence prediction model based on the multidimensional feature vector; performing online incremental update on the hybrid deep learning model by utilizing new data continuously collected in actual operation; Based on the updated mixed deep learning model, the recognition probability and the prediction trend are generated, the risk probability is calculated based on the Bayesian theory, and the corresponding risk level and the early warning information are automatically generated. As a preferable scheme of the deep learning-based power transmission wire ice edge corona discharge identification and prediction method, the method comprises the following steps: The acquisition comprises multi-mode data of acoustic, thermal imaging, meteorological and electric signals, the extraction of each mode characteristic and the fusion are carried out, and the construction of the multi-dimensional characteristic vector comprises the following steps: The method comprises the steps of filtering an acoustic signal, framing, calculating a power spectrum, filtering, calculating logarithmic energy, performing discrete cosine transform to obtain a characteristic coefficient, and performing characteristic extraction on a thermal imaging image through a convolutional neural network characteristic extraction network. As a preferable scheme of the deep learning-based power transmission wire ice edge corona discharge identification and prediction method, the method comprises the following steps: the acquisition comprises multi-mode data of acoustic, thermal imaging, meteorological and electric signals, the extraction of each mode characteristic and the fusion are carried out, and the construction of the multi-dimensional characteristic vector further comprises: and fusing the extracted modal features by using an attention mechanism, calculating corresponding attention weights according to the modal feature vectors, and carrying out weighted summation on the modal feature vectors based on the attention weights to obtain fused feature vectors, namely constructed multidimensional feature vectors. As a preferable scheme of the deep learning-based power transmission wire ice edge corona discharge identification and prediction method, the method comprises the following steps: the online incremental updating of the hybrid deep learning model by utilizing the new dat