Mathematic Model by Response Surface Methodology and Artificial Neural
Network for Predict Result of Tensile Shear and Nugget Size of Zinc Coated
Steel JIS G3313 Welded by Resistance Spot Welding
Type Of :
Academic Manuscript
Type :
Machanical Engineering
This research described to the determine an optimization mathematic model using response surface
methodology in central composite design method and artificial neural network (ANN) for predicting the of tensile
shear and nugget size in the zinc coated steel JIS G3313. The following resistance spot welding (RSW) parameters
were studied: the welding current, welding time, and electrode force. The resulting materials were examined using
tensile shear tests which were observed nugget size and microstructure with scanning electron microscopy (SEM).
The microstructure phenomenon could be explained by the welding optimum condition that fine pearlite and
intensity in heat affected zone. The research results reveal that an optimum RSW parameters were welding current
of 12 kilo amperes, welding time of 9 cycle and 1.5 kilo newton electrode force. The fine acicular ferrite occurred in
the nugget size, which results in increased welding material high mechanical property. The ANN model with the
proposed mathematical model, which tensile shear represents 3 neurons for the input 10 neurons for 1 hidden layer
and 1 output neurons (3-10-1). The ANN model was developed to establish of the nugget predict represents 3
neurons for the input 5 neurons for 1 hidden layer and 1 output neurons (3-5-1). The mean square error (MSE) and
coefficient of determination (R2) for tensile shear predict was showed that of 0.0026 and 0.956 respectively, which
nugget size predicted MSE of 0.0004 and R2 of 0.958. This research, the related manufacturing sector can use
research data and mathematical models was used to predict and quality control of the RSW processes to obtain
tensile shear and the nugget size according to the acceptance criteria.
Date22/12/2019
705

Keyword :