Neuro-Fuzzy Technique for Micro-hardness Evaluation of Explosive Welded Joints

In this research work, aluminium with low-carbon steel was joined through the explosive welding process at different loading ratios. The high impact pressure caused by explosive energy at the weld interface results in increase in hardness value. This hardness property plays an important role, as it affects the mechanical properties of welded plates. A model based on an intelligent technique named adaptive network-based fuzzy inference system (ANFIS) has been developed to predict the various micro-hardness values across the weld interface of explosive welded plates. The obtained experimental data were utilized for training and testing of the ANFIS model to predict micro-hardness values. The developed model was used on both sides of the weld interface, i.e. aluminium and steel. The model performance evaluations were carried out using different statistical criteria such as cross-correlation, mean absolute percentage error (MAPE) and root-mean-square error (RMSE). In comparison with aluminium, the steel side showed good results with a value of adj. R-square (0.95955) when compared to that in aluminium (0.85343). This observation was also supported by MAPE and RMSE data. The experimentally obtained micro-hardness values were found to be in good agreement with predicted ones through the ANFIS model.