Title:
HETEROGENEOUS FIELD SIGNAL-BASED INTELLIGENT INVERSION METHOD FOR PIPELINE DEFECTS
Document Type and Number:
WIPO Patent Application WO/2022/088226
Kind Code:
A1
Abstract:
A heterogeneous field signal-based intelligent inversion method for pipeline defects, relating to the technical field of non-destructive detection of pipeline defects. Said method comprises: first, acquiring heterogeneous field signals, performing anomaly determination, then performing basic value correction on the heterogeneous field signals by using an improved mean median method, and performing denoising processing by using a wavelet analysis method; then performing a padding operation on the heterogeneous field signals corresponding to pipeline defects after the denoising processing, unifying the heterogeneous field signals of different sizes into the same size, and performing non-linear transformation on the amplitude values of the signals; designing a sparse autoencoder having an axisymmetric structure, to obtain primary features of the heterogeneous field signals corresponding to the pipeline defects; classifying the length, width and depth of the pipeline defects, to obtain classification labels of various pipeline defects; designing a multi-classification neural network, classifying the heterogeneous field signals corresponding to the pipeline defects, and extracting deep features containing defect size information; and constructing a random forest regression model, so as to achieve intelligent inversion of the sizes of the pipeline defects.
Inventors:
ZHANG HUAGUANG (CN)
LIU JINHAI (CN)
WANG LEI (CN)
FENG JIAN (CN)
WANG GANG (CN)
MA DAZHONG (CN)
LU SENXIANG (CN)
LIU JINHAI (CN)
WANG LEI (CN)
FENG JIAN (CN)
WANG GANG (CN)
MA DAZHONG (CN)
LU SENXIANG (CN)
Application Number:
PCT/CN2020/126885
Publication Date:
May 05, 2022
Filing Date:
November 06, 2020
Export Citation:
Assignee:
UNIV NORTHEASTERN (CN)
International Classes:
G01N27/83; G06K9/00
Foreign References:
CN109783906A | 2019-05-21 | |||
CN106018545A | 2016-10-12 | |||
CN104990977A | 2015-10-21 | |||
CN110220966A | 2019-09-10 | |||
US20130124109A1 | 2013-05-16 | |||
CN109632942A | 2019-04-16 |
Other References:
LIU JINHAI, MINGRUI FU, JIANHUA TANG: "MFL inner detection based defect recognition method", CHINESE JOURNAL OF SCIENTIFIC INSTRUMENT NOV., vol. 37, no. 11, 30 November 2016 (2016-11-30), pages 2572 - 2581, XP055925952, DOI: 10.19650/j.cnki.cjsi.2016.11.021
Attorney, Agent or Firm:
SHENYANG DONGDA INTELLECTUAL PROPERTY AGENCY CO., LTD (CN)
Download PDF: