Advanced Search  
JWJ > Volume 39(2); 2021 > Article
Journal of Welding and Joining 2021 April;39(2) :198-205.
Published online April 1, 2021.
doi:https://doi.org/10.5781/JWJ.2021.39.2.9
Spectrogram Based Detection Algorithm for Back-Bead in Gas Metal Arc Welding Process using Convolution Neural Network
Chengnan Jin*, Sangrin Park*, Sehun Rhee*
*School of Mechanical Convergence Engineering, Hanyang University, Seoul, 04763, Korea
Correspondence: Sehun Rhee  Email: srhee@hanyang.ac.kr
Received January 27, 2021  Revised February 4, 2021  Accepted February 23, 2021  Published online April 1, 2021
ABSTRACT
An automated welding system is essential to ensure a stable and good welding quality and improve productivity in the gas metal arc welding (GMAW) process. Therefore, various studies have been conducted on the establishment of smart factories and the demand for good weldability in the fields of production and manufacturing. In shipbuilding welding and pipe welding, the uniformly generated back-bead is an important criterion for judging the mechanical properties and weldability of the welded structure, and is also an important factor that enables the realization of an automated welding system. Therefore, in this study, the welding current signal measured in real-time in the GMAW process was pre-processed by a short time Fourier transform (STFT) to obtain a time-frequency domain feature image (spectrogram). Based on this, a back-bead generation detection algorithm was developed. To accelerate the training speed of the proposed convolution neural network (CNN) model, we used non-saturating neurons and a highly efficient GPU implementation of the convolution operation. As a result of applying the proposed detection model to actual welding process, the detection accuracies with and without the back-bead regions were 95.8% and 94.2%, respectively, which confirmed the excellent classification performance for back-bead generation.
Keywords: Gas metal arc welding | Convolution neural network (CNN) | Short time fourier transform (STFT) | Time-frequency domain analysis | Back-bead generation prediction
TOOLS
PDF Links  PDF Links
Full text via DOI  Full text via DOI
Download Citation  Download Citation
CrossRef TDM  CrossRef TDM
  E-Mail
Share:      
METRICS
0
Crossref
248
View
33
Download
Related articles
Evaluation of Gas Metal Arc Welding Characteristics according to Contact Tip Materials  2014 December;32(6)
Prediction of Heating Line for Plate Forming in Induction Heating Process Using Artificial Neural Network  2010 August;28(4)
Prediction of Heating-line Positions for Line Heating Process by Using a Neural Network  2003 August;21(4)
Determination of Optimal Conditions for a Gas Metal Arc Welding Process Using Genetic Algorithm   2000 October;18(5)
Development of an Effective Arc Sensing Algorithm for Seam-Tracking in Flux-Cored Arc Welding Process for Horizontal Fillet Joints .   1997 February;15(1)
Evaluation of Gas Metal Arc Welding Characteristics according to Contact Tip Materials  2014 December;32(6)
Prediction of Heating Line for Plate Forming in Induction Heating Process Using Artificial Neural Network  2010 August;28(4)
Prediction of Heating-line Positions for Line Heating Process by Using a Neural Network  2003 August;21(4)
Determination of Optimal Conditions for a Gas Metal Arc Welding Process Using Genetic Algorithm   2000 October;18(5)
Development of an Effective Arc Sensing Algorithm for Seam-Tracking in Flux-Cored Arc Welding Process for Horizontal Fillet Joints .   1997 February;15(1)
e-sciencecentral
Register for e-submission
Register here to access the e-submission system of Journal of Welding and Joining for authors and reviewers.
Manuscript Submission
To submit a manuscript, please visit the Journal of Welding and Joining e-submission management system at https://submit.e-jwj.org, read the Instructions for Authors, and log into the Journal of Welding and Joining e-submission system. For assistance with manuscript submission, please contact: koweld@kwjs.or.kr.
Free archive
Anyone may access any past or current articles without logging in.
Editorial Office
#304, San-Jeong Building,
23, Gukhoe-daero 66-gil, Yeongdeungpo-gu, Seoul 07237, Korea
TEL : +82-2-538-6511    FAX : +82-2-538-6510   E-mail: koweld@kwjs.or.kr
Browse Articles |  Current Issue |  For Authors and Reviewers |  About
Copyright© by The Korean Welding and Joining Society. All right reserved.           Developed in M2PI