A theoretical and experimental study on indentation creep with x-silica addition to an Al–8Zn composite alloy, based on Adaptive Neuro-Fuzzy Inference System (ANFIS)

Abstract

This research is based on the significance of using the adaptive neuro-fuzzy inference system (ANFIS) to advance scientific study and integrate the theoretical and experimental domains. (ANFIS) is one of the soft computing techniques that play a significant role in modeling. This model studies the effects of adding nanosilica particles and aging temperatures on the indentation creep behavior of Al–8Zn alloys. The optimal ANFIS model configuration will be chosen by comparing the RMSE values for different types and the number of membership functions (MFs) assigned to each ANFIS structure. The best results were found as root mean square error (RMSE) is $1.1\times {10}^{-8}$, R-squared (R²) is 0.9742, and mean absolute error (MAE) is $3.3\times {10}^{-5}$ by using the Gbellmf type with four MFs. Vickers hardness measurements were used to evaluate the indentation creep behavior of Al–8Zn-x nanoSiO₂, x= (0, 0.5, 1, 1.5, 2, 2.5, 3) wt.% under different aging temperature ranges from 398 to 478 K for 2 h, dwell times from 10 to 90 s, and a constant load of 200 gm. A scanning electron microscope (SEM) with an energy dispersive spectroscope (EDS) and X-ray diffraction (XRD) was used to study the change in microstructure. Increasing nanoSiO₂ concentration from 0.5 % to 1 % decreased minimum creep rates, but increasing nanoSiO₂ concentration from 1.5 % to 2 % increased minimum creep rates across all aging temperatures. Values of the stress exponent in the range of 5.9–8.82 decrease as the concentration of nanoSiO₂ rises to 1, and then they increase as the concentration of nanoSiO₂ rises, with activation energy ranging from 92.8 kJ/mol to 99.8 kJ/mol. ANFIS was utilized to confirm the experimental results further, and the method was also applied to predict the minimum creep rates at 2.5 % and 3 %. This study suggests that the ANFIS technique is a useful method for forecasting mechanical properties and deformation mechanisms of materials.