Artificial neural network prediction of weld distortion rectification using a travelling induction coil

Barclay, Colin; Campbell, Stuart; Galloway, Alexander; McPherson, Norman

doi:10.1007/s00170-012-4713-z

Cited by 8 publications

(6 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The software used throughout this study was NeuroSolutions, which had previously been used by the authors for the successful prediction of weld geometry [1], distortion [22] and distortion rectification [24]. A total of 360 datasets were produced, 36 for each nozzle and shielding gas configuration, with an additional 20 for validation/production, thus allowing for a full range of results ranging from good quality welds performed using a low shielding gas flow rate and no cross draft, to poor quality welds with a high shielding gas flow rate and high cross draft velocity.…”

Section: Ann Model Developmentmentioning

confidence: 99%

See 1 more Smart Citation

Artificial neural network optimisation of shielding gas flow rate in gas metal arc welding subjected to cross drafts when using alternating shielding gases

Campbell

Ley

Galloway

et al. 2014

Proceedings of the Institution of Mechanical Engineers, Part B:

View full text Add to dashboard Cite

This study implemented an iterative experimental approach in order to determine the shielding gas flow required to produce high quality welds in the gas metal arc welding (GMAW) process with alternating shielding gases when subjected to varying velocities of cross drafts. Thus determining the transitional zone where the weld quality deteriorates as a function of cross draft velocity. An Artificial Neural Network (ANN) was developed using the experimental data that would predict the weld quality based primarily on shielding gas composition, alternating frequency and flowrate, and cross draft velocity, but also incorporated other important input parameters including voltage and current. A series of weld trials were conducted validate and test the robustness of the model generated. It was found that the alternating shielding gas process does not provide the same level of resistance to the adverse effects of cross drafts as a conventional argon/carbon dioxide mixture. The use of such a prediction tool is of benefit to industry in that it allows the adoption of a more efficient shielding gas flow rate, whilst removing the uncertainty of the resultant weld quality

show abstract

Section: Ann Model Developmentmentioning

confidence: 99%

“…ANN models have been applied extensively and relatively widespread to the welding process, predicting a number of weld aspects, including weld geometry, 1,17 mechanical properties, 18,19 weld quality, 20,21 weld-induced distortion 22,23 and weld distortion rectification. 24…”

Section: Introductionmentioning

confidence: 99%

Artificial neural network optimisation of shielding gas flow rate in gas metal arc welding subjected to cross drafts when using alternating shielding gases

Campbell

Ley

Galloway

et al. 2014

Proceedings of the Institution of Mechanical Engineers, Part B:

View full text Add to dashboard Cite

show abstract

“…They employed a set of FEM results obtained for various plate dimensions as the ANN inputs. Barclay et al [13] investigated the suitability of the induction heating process with the traveling induction coil to rectify the welding angular distortion. By using the experimental data, they considered an ANN to predict the distortion.…”

Section: Introductionmentioning

confidence: 99%

Using an optimized RBF neural network to predict the out-of-plane welding distortions based on the 3-2-1 locating scheme

Tafarroj

Kolahan

2018

Scientia Iranica

View full text Add to dashboard Cite

This study deals with the e ect of locator positioning in the 3-2-1 locating scheme to control the out-of-plane distortion in gas tungsten arc welding of sheet metals. To apply this locating scheme to the sheet metals, a suitable xture was designed. The distortion of the welded plates was predicted using the Radial Basis Function (RBF) neural network. To gather the experimental data employed in the RBF modeling process, a set of welding tests was performed on the sheet specimens by varying the positions of the three locators. The parameters of the network were optimally selected using the Simulated Annealing (SA) optimization algorithm. The average and maximum errors computed for the test dataset were respectively 2.43% and 5.30% while in some cases, the error fell below 1%. The results of the RBF network showed very good agreement with the experiments and it can be concluded that this modeling technique can be utilized successfully in predicting the welding distortions when the 3-2-1 locating scheme is used.

show abstract

“…As a result, residual stresses are introduced to the surrounding material, resulting in distortion in the final welded structures; a common problem associated with any welding process [24]. Although it has been shown that there is a critical limit to the angular distortion due to the welding process [25], the magnitude of the weld-induced distortion is primarily related to the specific thermal energy input of the welding process [26].…”

Section: Introductionmentioning

confidence: 99%

“…Computational modelling has become increasingly popular over the past few decades for the prediction of various aspects of the welding process. Extensive studies have been conducted in areas fundamental to the final structure, namely residual stresses in the weld region [29,30,31], weld induced distortion [25,26,30,31] and the solidified weld geometry [10,31].…”

Section: Introductionmentioning

confidence: 99%

Effect of shielding gas parameters on weld metal thermal properties in gas metal arc welding

Ley

Campbell

Galloway

et al. 2015

Int J Adv Manuf Technol

Self Cite

View full text Add to dashboard Cite

This study considered the effect of shielding gas parameters (composition, supply method and flow rate) on the post weld thermal properties (thermal expansion, specific heat capacity, thermal diffusivity and thermal conductivity) of the weld metal in gas metal arc welding. This is of importance as the thermal properties influences the temperature distribution and therefore the residual stresses and distortion present within the final structure. Due to the lack of accurate thermal data, computational modelling techniques (such as FEA, CFD, etc.) used for modelling the welding process generally make assumptions regarding the material thermal properties, and it is often the case that the parent material thermal properties are extended to the weld metal, introducing errors to the simulation. It was determined that the weld metal posses considerably different thermal properties to the DH36 grade steel parent material, and that the shielding gas parameters significantly alter key thermal properties of the solidified weld metal. A lower shielding gas flow rate exhibited beneficial properties including a lower thermal expansion, and higher specific heat capacity and thermal conductivity than a higher shielding gas flow rate

show abstract

Artificial neural network prediction of weld distortion rectification using a travelling induction coil

Cited by 8 publications

References 15 publications

Artificial neural network optimisation of shielding gas flow rate in gas metal arc welding subjected to cross drafts when using alternating shielding gases

Artificial neural network optimisation of shielding gas flow rate in gas metal arc welding subjected to cross drafts when using alternating shielding gases

Using an optimized RBF neural network to predict the out-of-plane welding distortions based on the 3-2-1 locating scheme

Effect of shielding gas parameters on weld metal thermal properties in gas metal arc welding

Contact Info

Product

Resources

About