Image processing technique for measuring the weld dilution and heat-affected zone to model the gas metal arc welding process by a neuro-fuzzy system

Subject of study. Measuring the weld dilution and heat-affected zone to model gas metal arc welding is proposed. Purpose of the work. Gas metal arc welding is a critical process for a high-quality permanent joining of metal workpieces. The voltage, wire feed rate, shielding gas flow, nozzle angle, and distance from the workpiece are essential to control parameters for gas metal arc welding. Typically, the quality of a weldment is explained in terms of the geometry of the weld bead, dilution, and heat-affected zone. The degree of dilution plays a vital role in preventing undesirable phases from forming and distributing elements during welding. Additionally, the heat-affected zone affects the mechanical properties and microstructure alterations in the welded metal. Method. The effect of input parameters such as welding velocity, voltage, and wire feeding speed on dilution and the heat-affected zone is investigated in this paper. This is accomplished using an image processing technique. Then, the correlation coefficients of the dilution and heat-affected zone were predicted using neuro-fuzzy modeling. Main results. Measuring the weld bead geometries was done using the image processing technique with a mean error percentage of 1.66%. Correlation coefficients for the dilution and heat-affected zone were predicted to be 93 and 99, respectively, based on experimental data. Practical significance. The results indicate that the image processing technique is highly capable of measuring the geometry of the weld bead and that the resulting adaptive neuro-fuzzy inference system model is suitable for predicting gas metal arc welding dilution and the heat-affected zone. (c) 2024 Optica Publishing Goud.