Analysis of the combined effects of preheating and welding wire feed rates on the FCAW bead geometric characteristics of 1020 steel using fuzzy logic-based prediction models

Abstract Single weld beads were deposited on a steel plate using three different welding wire feed rates (slow, medium, and fast). The samples were preheated before welding at three different temperatures (100, 150, and 200°C). Fuzzy logic models were developed and integrated into the analysis for predicting weld bead geometries. The experimental results demonstrated that preheating and wire feed rates had significant impact on the geometric shape characteristics of 1020 weld beads. Higher preheating temperatures and optimal wire feed rates led to improved weld bead geometry. The integrated fuzzy logic model predicted the weld bead geometry with optimal input variables of 23 V, 150 A, 3 mm/s welding speed, and 540 J/mm heat input, with an optimal bead width, bead height, depth of penetration, heat affected zone (HAZ) width, and height (9.72, 2.02, 1.62, 12.54 and 2.73 mm). The accuracy of the fuzzy models were examined via regression plots, which yielded R 2 values of 0.9146, 0.9909, 0.9467, 0.9805, and 0.8239, for the bead width, bead height, depth of penetration as well as HAZ width and height. This implies that the fuzzy models were effective in predicting the bead height, justifying from its very high degree R 2 value of 0.9909. This showcased the viability of fuzzy logic for predicting weld bead geometry.