Abstract—The weld bead geometry influences the mechanical properties of the weld. The reinforcement height and bead width are important factors of the bead geometry. This study is concerned with the multi-response optimization of Shielded Metal Arc Welding (SMAW) process for an optimal parametric combination to yield favorable weld bead width and reinforcement height of welded joints produced on 5 mm thick mild steel plates using the Artificial Neural Networks (ANN). To disperse molten droplets of the electrode, a longitudinal external magnetic field was created by a bar magnet which was mounted on the tailstock side of a lathe machine with the help of a wooden structure. Speed of welding was made constant with the help of the cross slide of the lathe machine. Eighteen experimental data sets were used to train the ANN. Seven experiments were conducted to get other seven sets of data to compare the results obtained with the corresponding prediction made by ANN. It was found that the predictions made by the ANN were very close to the experimental values.

Index Terms—Artificial Neural Networks (ANN), Back propagation, Bead geometry, Input process parameters, Reinforcement height.

Cite: Rudra Pratap Singh, Ramesh Chandra Gupta, and Subhash Chandra Sarkar, "Analysis of Bead Width and Reinforcement Height during Shielded Metal Arc Welding under Magnetic Field Using Artificial Neural Networks," International Journal of Mechanical Engineering and Robotics Research, Vol. 2, No. 2, pp. 59-68, April 2013.