Search

Submerged arc welding (SAW) is a widely used technique in various industries for welding thick plates. The quality of welded joints produced by this process depends on the selection of appropriate parameters that yield weldments with desirable mechanical properties. Among these parameters, weld bead penetration is a crucial indicator of weld quality. In this study, the effect of arc voltage, welding current, distance between the contact tip and the workpiece, and arc travel speed in the presence of Chromium oxide (Cr2O3) nanoparticles on penetration has been investigated. In addition, a new hybrid optimization algorithm has been developed by combining the differential evolution (DE) and wingsuit flying search (WFS) algorithms. The hybrid algorithm is evaluated by using standard benchmark functions and combined with the adaptive network-based fuzzy inference system (ANFIS) to create a new model. The model predicts weld bead penetration in the SAW process based on input parameters. The proposed hybrid algorithm improved the effectiveness of the main ANFIS in predicting weld bead penetration. The results showed that the addition of Cr2O3nanoparticles to the weld pool increased its penetration.

ABSTRACTOne of the mechanical properties of weldments influenced by the heat input generated by the Submerged Arc Welding (SAW) process is the Hardness of Weld Metal (HWM). This research was carried out to develop a non-linear regression model using the Response Surface Methodology (RSM) to predict the HWM in the SAW of low-carbon steel plates (St37) in the presence of Cr2O3nanoparticles. For developing the regression model, current, voltage, stick-out, welding speed, and Cr2O3nanoparticles were considered as the input variables and the HWM as the response. Analysis of variance was carried out to check the validity of the regression model and decide on the significant variables affecting the HWM. SEM and EDX analysis were performed on the weld to study the microstructure and elemental variables. The main and interactive effects of the input variables on the HWM were presented in graphical forms and analysed. The results showed that increasing the thickness of Cr2O3nanoparticles increased the HWM due to grain refinement in the weld metal. Furthermore, optimisation of the model was developed for finding the optimum values of the HWM. Finally, confirmation tests were carried out and they indicated that the error of the empirical model was 4.36%.