Research on Automatic Detection Method of DR Digital imaging for Brazing Seam Defects of Thrust Chamber Body

REN Wenjian; WANG Yonghong; LI Chunkai; SHI Yu; SUN Zhongcheng; QIN Jiangang

doi:10.7512/j.issn.1001-2303.2023.01.02

Important Issuse | Views : 0 下载量: 32 CSCD: 0

PDF
Export
Share
Collection
Album

Research on Automatic Detection Method of DR Digital imaging for Brazing Seam Defects of Thrust Chamber Body
Vol. 53, Issue 1, Pages: 9-14(2023)
DOI： 10.7512/j.issn.1001-2303.2023.01.02

扫描看全文

任文坚，王永红，李春凯，等.液体火箭发动机推力室熔焊缝数字化胶片图像缺陷识别方法研究［J］.电焊机，2023，53（1）：9-14.

REN Wenjian, WANG Yonghong, LI Chunkai, et al.Research on Automatic Detection Method of DR Digital imaging for Brazing Seam Defects of Thrust Chamber Body[J].Electric Welding Machine, 2023, 53(1): 9-14.
任文坚，王永红，李春凯，等.液体火箭发动机推力室熔焊缝数字化胶片图像缺陷识别方法研究［J］.电焊机，2023，53（1）：9-14. DOI： 10.7512/j.issn.1001-2303.2023.01.02.

REN Wenjian, WANG Yonghong, LI Chunkai, et al.Research on Automatic Detection Method of DR Digital imaging for Brazing Seam Defects of Thrust Chamber Body[J].Electric Welding Machine, 2023, 53(1): 9-14. DOI： 10.7512/j.issn.1001-2303.2023.01.02.

摘要

推力室的焊接质量可靠性与稳定性对于发动机服役性能以及火箭安全运行具有非常重要的意义，主要通过人工评定胶片图像来识别其焊缝缺陷，检测效率低、经验依赖性强。针对液体火箭发动机推力室胶片数字化图像的特点，通过对原始图像进行尺寸归一化、图像增强等图像预处理构建图像样本集，并采用YOLO V3算法建立了基于深度学习理论的熔焊缝缺陷自识别模型。结果表明，训练的YOLO神经网络模型能够准确识别胶片数字化图像中气孔、裂纹、未熔合、未焊透等4种典型熔焊缝缺陷，识别准确率达90%以上，有良好的工业应用前景。

Abstract

As one of the most important components of a liquid rocket engine, the reliability and stability of welding quality are very important for the service performance of the engine and the reliable and safe operation of the rocket.In actual production, the weld defects are mainly identified by manual evaluation of film images, which generally have shortcomings such as low detection efficiency and strong dependence on manual experience. According to the characteristics of the digitized image of the liquid rocket engine thrust chamber film, the image sample set was constructed by performing image preprocessing such as size normalization and image enhancement on the original image.The YOLO algorithm was used to establish a self-identification model of fusion weld defects based on deep learning theory.The research results show that the trained YOLO neural network model can accurately identify four typical fusion weld defects such as pores, cracks, lack of fusion, and lack of penetration in digital film images, and the recognition accuracy can reach more than 90%. which has good industrial application prospects.

关键词

无损检测钎焊缺陷X射线数字成像深度学习

Keywords

non-destructive testingbrazing defectsX-Ray digital radiographydeep learning

references

罗巍，张晗翌，矫慧，等. 液体火箭发动机推力室快速制造技术［J］. 航天制造技术， 2020（05）： 60-65.

LUO W， ZHANG H Y， JIAO H， et al. Rapid Fabrication Techniques of Liquid Rocket Engine Thrust Chamber［J］. Aerospace Manufacturing Technology， 2020（05）： 60-65.

任文坚，刘贞，张腾，等. 推力室钎焊身部X射线数字成像检测技术［J］. 火箭推进，2020， 46（03）： 90-95.

REN W J， LIU Z， ZHANG T， et al. X-ray digital imaging detection technology for brazed body of thrust chamber［J］. Journal of Rocket Propulsion， 2020， 46（03）： 90-95.

Lu Y， Zhang W. X-ray automatic detection image processing of welding defects［J］. Optical Instruments， 2018， 39（4）： 18-21.

Zhang J. Identification of weld defects based on principal component analysis［J］. Industrial Control Computer，2017（01）： 111-112．

迟大钊，马子奇，程怡，等. 不等厚板搭接焊缝缺陷数字X射线检测［J］. 焊接学报， 2019， 40（11）：45-48.

CHI D Z， MA Z Q， CHENG Y， et al. X-ray based defect testing method for a lap joint with unequal thickness steel plates［J］. Transactions of the China Welding Institution， 2019， 40（11）： 45-48.

姜洪权，贺帅，高建民，等. 一种改进卷积神经网络模型的焊缝缺陷识别方法［J］. 机械工程学报， 2020， 56（08）：235-242.

JIANG H Q， HE S， GAO J M， et al. An Improved Convolutional Neural Network for Weld Defect Recognition ［J］. Journal of Mechanical Engineering， 2020， 56（08）：235-242.

罗爱民，沈才洪，易彬，等.基于改进二叉树多分类SVM的焊缝缺陷分类方法［J］.焊接学报，2010（7）：51-54.

LUO A M， SHEN C H， YI B， et al. Method of multi-classification by improved binary tree based on SVM for welding defects recognition［J］. Transactions of the China Welding Institution， 2010（7）：51-54.

Zhang H， Chen Z， Zhang C. Weld Defect Detection Based on Deep Learning Method［C］. IEEE 15th International Conference on Automation Science and Engineering （CASE）， 2019， 1574-1579.

Liu J， Liu X， Qu F. A Defect Recognition Method for Low-Quality Weld Image Based on Consistent Multiscale Feature Mapping［J］. IEEE Transactions on Instrumentation and Measurement， 2022， 71： 1-11.

刘涵，郭润元. 基于X射线图像和卷积神经网络的石油钢管焊缝缺陷检测与识别［J］. 仪器仪表学报， 2018， 39（04）：247-256.

LIU H， GUO R Y. Detection and identification of SAWH pipe weld defects based on X-ray image and CNN［J］.Chinese Journal of Scientific Instrument， 2018， 39（04）： 247-256.

Alert me when the article has been cited

提交

暂无数据