Recognition method of power grid monitoring alarm events based on convolution and long short-term memory network

A long-term and short-term memory and alarm event technology, applied in neural learning methods, biological neural network models, electrical and digital data processing, etc. Low efficiency, improve work efficiency, and reduce the effect of monitoring screen pressure

Active Publication Date: 2021-11-12
HOHAI UNIV
View PDF5 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] Purpose of the invention: Aiming at the problem that the manual judgment in the existing power grid monitoring and alarming is easy to miss and misjudgment, and the recognition efficiency is not high, the present invention provides a power grid monitoring and alarming event recognition method based on convolution and long-term short-term memory network

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Recognition method of power grid monitoring alarm events based on convolution and long short-term memory network
  • Recognition method of power grid monitoring alarm events based on convolution and long short-term memory network
  • Recognition method of power grid monitoring alarm events based on convolution and long short-term memory network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0063] Refer figure 1 , figure 2 with image 3 The method of the present invention according to the following steps:

[0064] The first step, collecting all the substations and lines in the history of the name of the grid monitoring system to monitor alarms and alarm information for each time scale, alarm information contained in grid monitoring alarm events constitute recognition model required training data set;

[0065] A second step of monitoring alarm information history data preprocessing, unsupervised training to monitor alarm information word2vec model, generating information vector signal including characteristics of the specific process:

[0066] (1) word go and stop words

[0067] Update power thesaurus, thesaurus and access to the collection and export of electricity from historical monitoring alarm information line and substation name as the name of the import thesaurus dictionary word power through the use of data. Using the exact mode Jieba segmentation tool is monit...

Embodiment 2

[0109] Taking a city grid company 2016 and 2017, a total of more than 14 million historical monitoring alarm information is the original language library, extracts 9 types of alarm event samples to train and test the identification model. 90% of each type of alarm event sample as a training set, 10% as a test set, alarm event type, and the number of samples per class, as shown in Table 1.

[0110] Table 1 Alarm event sample quantity

[0111]

[0112] In the classification task of the event, the classification result of the identification model is generally expressed in the confusion matrix, and the meaning of the second classification confusion matrix is ​​shown in Table 2.

[0113] Confusion matrix in Table 2 Event Identification

[0114]

[0115] The confusion matrix will measure the identification effect of the model according to its actual home and identification belonging to four classes, define accuracy (Accuracy), precise, and precise, recall, and F1 value, four indicat...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention discloses a power grid monitoring alarm event recognition method based on convolution and long-term short-term memory network. The method generates information vector through historical monitoring alarm information and time stamp in the power grid monitoring system, and extracts events from the collected historical monitoring alarm information. samples to build a sample library of alarm events; secondly, establish a deep learning recognition model based on the combination of long-term short-term memory network and convolutional neural network, and use the alarm event samples to train the model; finally, use the trained deep learning model to monitor the alarm information. Recognition, take the event category with the highest probability as the output of the recognition result. The invention combines the excellent performance of the long-short-term memory network in dealing with timing problems and the convolutional neural network in mining the local features of short texts to establish a combined model, which can realize the rapid identification of power grid alarm events and effectively reduce the monitoring pressure of the monitoring business personnel. Improve the efficiency of daily monitoring and abnormal handling of accidents.

Description

Technical field [0001] The present invention belongs to the power system intelligent alarm control technology, and specific relates to a grid monitoring alarm event identification method based on convolution and long short-term memory network. Background technique [0002] As the grid size is expanding, the regulator is quickly responding to the power grid equipment failure, and timely restoring the power network operation method, the intelligent level of the operating monitoring of the grid equipment is improved, and the independent recognition of the power grid alarm event is enhanced. The work efficiency of daily monitoring and abnormal accidents is of great significance. [0003] The power grid monitoring alarm information is a Chinese text data, which is an important data foundation for regulators to monitor the operating status of the grid. With the scale expansion of the grid equipment and the increase in intelligent monitoring, the output of power data has emerged, the nu...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Patents(China)
IPC IPC(8): G06F16/35G06F40/284G06N3/04G06N3/08
CPCG06F16/35G06N3/088G06N3/044G06N3/045
Inventor 臧海祥白子瑜程礼临孙国强卫志农
Owner HOHAI UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap