Investigation of feed-forward back propagation ANN using voltage signals for the early prediction of the welding defect

Dinu Thomas Thekkuden, Abdel Hamid Ismail Mourad

Research output: Contribution to journalArticlepeer-review

41 Citations (Scopus)


The research paper investigates the prediction capability of the artificial neural network for weld quality assessment from the captured voltage signals in a gas metal arc welding process. The bead-on-plate welds and v-groove welds were made on SA 516 grade 70 material by altering different parameters such as stickout distance, gas flow rate and travel speed. The voltage signals of each weld were captured using a data acquisition system having 8000 Hz data acquisition rate. The descriptive statistics of the voltage data such as mean, standard error, median, mode, standard deviation, sample variance, kurtosis, skewness, minimum and maximum corresponding to bead-on-plate welds and v-groove welds were used for training and testing the neural network respectively. The quality of the weld was assessed by the visual inspection, and from control charts plotted using voltage data. Overall classification accuracy of 94.7% was achieved in the training process. The feed-forward back propagation neural network predicted the quality of test v-groove welds accurately with a 90.9% prediction rate. The results proved that the developed method is promising for the immediate and early prediction of the weld quality.

Original languageEnglish
Article number1615
JournalSN Applied Sciences
Issue number12
Publication statusPublished - Dec 2019


  • Artificial neural network
  • Fast-forward back propagation
  • Weld quality
  • Welding voltage

ASJC Scopus subject areas

  • General Chemical Engineering
  • General Materials Science
  • General Environmental Science
  • General Engineering
  • General Physics and Astronomy
  • General Earth and Planetary Sciences


Dive into the research topics of 'Investigation of feed-forward back propagation ANN using voltage signals for the early prediction of the welding defect'. Together they form a unique fingerprint.

Cite this