Investigation of welding parameters in the arc welding electrodes with intelligent systems

Arc sources, one of the most important tools of the manufacturing industry, are used at many points. Preparation stages and welding parameter selections are important before welding with these machines. The quality of the welds depends on the skill of the welder and the correct selection of the welding parameters. In the study, the yield strength (N / mm²) of the welding consumables was tried to be optimized by using different welding input parameters with two different algorithms. For this purpose, a hybrid model has been developed with the C # .NET programming language using the Artificial Immune System Clonal Selection Algorithm (CSA), which is one of the sub-methods of artificial intelligence, which is a very popular area in recent years, and the Fuzzy Logic method. YBA-BM hybrid algorithm method for the product with E7018-1 standard for the expected optimum yield strength based on the %Carbon, %Silicon, %Manganese inputs. It was verified with the MINITAB-17 statistical program that the regression model was successful with a rate of 99.99%. In addition, with the other study, the welding current, welding voltage and welding feed rate were determined as experimental input parameters and the yield strength (N / mm²) was estimated by artificial neural networks method. It has been observed that the artificial neural network (ANN) model trained according to the real data can predict the yield strength with the forward-feedback algorithm for the most ideal and standards-compliant values.