Development of fuzzy logic-based expert systems for optimizing TIG welding parameters of low carbon steel plates

The complex experimental procedure involved in TIG welding sequence, selection of ideal input parameters and further challenges during tensile testing to obtain the ultimate tensile strength (UTS) value of a given material may lead to inaccurate output depending on the experimental and test procedures. Moreover, while traditional optimization techniques which is sometimes based on trial and error, or prolonged experimental process may be time-consuming and costly, fuzzy logic-based expert systems offer a more effective solution to optimizing welding input variables for optimal output response. In this study, TIG welding input parameters derived from design of experiment for 20 weld runs were employed in welding AISI 1015 low carbon steel plate, and the welded samples were subjected to tensile testing to determine the UTS values. UTS of the control sample was obtained as 385 MPa, while UTS values of 409.670, 412.265, 410.256, 414.624, 415.563, 413.314, 411.432 and 408.306 MPa were obtained from the tensile test with corresponding welding input parameters (voltage, current, welding speed and gas flow rate, from weld runs No. 4, 5, 7, 8, 13, 14, 15 and 18. However, optimal UTS values of 407.923, 411.186, 408.874, 414.123, 415.116, 411.874, 409.702 and 406.827 MPa were obtained from the fuzzy logic optimization at the same No. of weld runs. Regression plot of the experimental and fuzzy logic predicted output response revealed R ² value of 0.9892 which is very close to 1, indicating the accuracy of the optimization sequence. This was further confirmed from a correlation plot which clearly indicated a strong positive correlation between the experimental and fuzzy optimized UTS values. The robust results of the study demonstrate the effectiveness and reliability of the fuzzy logic-based expert system in optimizing TIG welding parameters for low carbon steel plates.