Abstract
This research article investigated the optimized process parameters for decreasing the hot cracking phenomenon and improving the microhardness of ultrasonic vibratory-assisted tungsten inert gas (U-TIG) welding of Inconel 625 alloy. The study employed two approaches: response surface methodology (RSM) and RSM coupled with a genetic algorithm (RSM-GA). The objective was to analyze the impact of welding process parameters, including welding current, gas flow rate, presence or absence of ultrasonic vibration, and filler material, on the crack length and microhardness of the welded joints. Experimental tests were conducted using RSM with a full factorial central composite design matrix, enabling comprehensive parameter space exploration. Parametric mathematical models were developed based on the obtained experimental data. These models were then utilized as fitness functions within the GA to determine the global optimal solution, aiming to minimize crack length and maximize microhardness. Additionally, artificial neural network (ANN) models were developed to predict the responses and optimize the welding process. The comparison between the experimental and predicted data demonstrated the reliability of the ANN model in accurately estimating the crack length and microhardness of U-TIG welded Inconel 625 alloy joints. The developed models achieved a prediction accuracy of less than 5 % error.