In the welding processes, parametric optimization is crucial, and intelligent prediction makes use of data availability to cut the cost of experimental operations. This article proposes adopting the adaptive neuro-fuzzy inference system (ANFIS) model for predicting ultimate tensile strength in TIG-MIG hybrid welding. Experiments are designed and optimized according to Taguchi's principles. Proposed neural network models are developed using experimental data. Three input process parameters ( MIG voltage, TIG current and gas flow rate) were designed in an L9 orthogonal array at three levels each. The maximum tensile obtained was 868.3 MPa. The signal-to-noise ratio shows that the optimum parameter setting that maximizes the tensile strength corresponds to MIG Voltage (V) = 25, TIG Current (A) =180, and Gas flow rate =19 L/mm. The analysis of variance shows that the gas flow rate had the most influence on the ultimate tensile strength with a 42.35% contribution, followed by the MIG voltage with 31.67%, and TIG current with 18.13% contribution. The developed ANFIS model is 99.9 % accurate at the training (MAPEtraining=0.1670) and 96.3% accurate at the testing (MAPEtestung=0.1670) for predicting the ultimate tensile strength. The R2-values of the models at training and testing were closer to unity depicts a good fit between the experimental and predicted values of the response. The lower RMSE values (RMSEtraining=1.8963, RMSEtesting =4.8194) indicates the lower deviation of the experiment values of ultimate tensile strength from the predicted values. These results imply that ANFIS models can reduce experimental costs and hurdles associated with the trial and error approach to get the appropriate welding parameters. Therefore experimental designs for other plate thicknesses and similar processes could be built and predicted without actual experimentation.
|Number of pages
|E3S Web of Conferences
|Published - 6 Oct 2023
|15th International Conference on Materials Processing and Characterization, ICMPC 2023 - Newcastle, United Kingdom
Duration: 5 Sept 2023 → 8 Sept 2023