Weld quality prediction in linear friction welding of AA6082-T6 through an integrated numerical tool

Risultato della ricerca: Articlepeer review

22 Citazioni (Scopus)

Abstract

A numerical and an experimental campaign were carried out with varying oscillation frequency and interface pressure. The local values of the main field variables at the contact interface between the specimens were predicted by a Lagrangian, implicit, thermo-mechanical FEM model and used as input of a dedicated Neural Network (NN). The NN, integrated in the FEM environment, was designed in order to calculate both a Boolean output, indicating the occurrence of welding, and a continuous output, indicating the quality of the obtained solid state weld. The analysis of the obtained results allowed three different levels of bonding quality, i.e., no weld, sound weld and excess of heat, to be correctly identified and predicted.
Lingua originaleEnglish
pagine (da-a)389-396
Numero di pagine8
RivistaJournal of Materials Processing Technology
Volume231
Stato di pubblicazionePublished - 2016

All Science Journal Classification (ASJC) codes

  • ???subjectarea.asjc.2500.2503???
  • ???subjectarea.asjc.1700.1706???
  • ???subjectarea.asjc.2500.2506???
  • ???subjectarea.asjc.2200.2209???

Fingerprint

Entra nei temi di ricerca di 'Weld quality prediction in linear friction welding of AA6082-T6 through an integrated numerical tool'. Insieme formano una fingerprint unica.

Cita questo