Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
SEPARATION AND IDENTIFICATION METHOD FOR MULTI-SOURCE HYBRID ULTRA-HIGH FREQUENCY PARTIAL DISCHARGE DIAGRAM
Document Type and Number:
WIPO Patent Application WO/2023/213332
Kind Code:
A1
Abstract:
The present invention discloses a separation and identification method for multi-source hybrid ultra-high frequency partial discharge diagram, which comprises: obtaining a PRPD diagram for ultra-high frequency partial discharge signals, carrying out standardization processing, and constructing and labeling a partial discharge sample database; constructing a PRPD diagram dataset of partial discharge signals and noise signals; determining a deep learning network to serve as a detection model, and inputting detection model parameters; dividing up the PRPD diagram dataset, and inputting same into deep learning network detection models for training; adjusting the detection models and training parameters; retaining and evaluating the trained models, and selecting the model with the optimal effect as a PRPD diagram detection model; and using the selected detection model on a diagram to undergo identification to analyze and confirm detection targets one-by-one, and retaining a final detection result and a multi-source separation result. The present invention solves the problem where when a plurality of partial discharge signals are actually present when using an image recognition algorithm to perform discharge type identification on a GIS ultra-high frequency PRPD diagram, there is no way to individually detect each signal.

Inventors:
ZHAO HONGYI (CN)
SHEN DAOYI (CN)
HU YONG (CN)
QIAN DAZHAO (CN)
HE CHENGWEN (CN)
LI LIBIN (CN)
Application Number:
PCT/CN2023/101087
Publication Date:
November 09, 2023
Filing Date:
June 19, 2023
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
GLOBAL SCIENCE & TECH SHANGHAI CO LTD (CN)
International Classes:
G06T7/00; G06V10/764
Domestic Patent References:
WO2015070513A12015-05-21
Foreign References:
CN115187527A2022-10-14
CN112149549A2020-12-29
CN112198399A2021-01-08
CN114186589A2022-03-15
CN106556781A2017-04-05
CN114417926A2022-04-29
KR20200116425A2020-10-12
Download PDF: