This study was conducted to develop the best model to estimate spatter rate for the short circuiting transfer mode of the gas metal arc (GMA) welding process. Utilizing an artificial neural network. The spatter rate generated during welding is a barometer of process stability for metal transfer, and it depends on the periodic waveforms of the welding current and arc voltage in the short circuiting mode. delve factors representing the characteristics of the waveforms as inputs and the spatter rate as an output were employed as variables for the neural network. Two neural network models were evaluated for estimating the spatter rate: one model did not consider arc extinction; the other model did. The input vector and the nodes of hidden layers for each model were optimized to provide an adequate fit. and estimated performance of each optimized model to the spatter rate was assessed and compared with the previously proposed model. It was, in addition, demonstrated in this study that the combined neural network model was more effective in predicting the spatter rate than other models through evaluation of the estimated performance of each optimized model.
[1]
Minoru Kobayashi,et al.
The Investigation of Arc Phenomena by means of A Computer
,
1984
.
[2]
M. Hermans,et al.
Process behavior and stability in short circuit gas metal arc welding
,
1999
.
[3]
Tsuneo Mita,et al.
Quantitative estimates of arc stability for CO2 gas shielded arc welding
,
1988
.
[4]
C. S. George Lee,et al.
Neural fuzzy systems: a neuro-fuzzy synergism to intelligent systems
,
1996
.
[5]
Marjan Suban,et al.
Determination of stability of MIG/MAG welding processes
,
2001
.
[6]
M J Kang,et al.
A study on the development of the arc stability index using multiple regression analysis in the short-circuit transfer region of gas metal arc welding
,
2001
.
[7]
C. Laymon.
A. study
,
2018,
Predication and Ontology.