Artificial neural network prediction of aging effects on the wear behavior of IN706 superalloy

• Inconel 706 superalloy samples produced from elemental powders were aged for different periods were subjected to wear testing.
• The highest hardness value was measured for the samples aged for 12 h.
• The developed artificial neural network model successfully predicted weight loss.

In this study, the effect of aging parameters on wear behavior of PM Inconel 706 (IN 706) superalloy was experimentally investigated and an ANN model was developed to predict weight loss after wear tests. IN 706 superalloy powders were cold pressed (700 MPa) and sintered at 1270 °C for 90 min. The sintered components were gradually aged for 16 h at 730 °C and for 12–20 h at 620 °C. The samples of IN706 superalloy were subjected to wear test at a constant sliding speed of 1 m/s under three different loads (30 N, 45 N and 60 N) and for five different sliding distances (400–2000 m). The results clearly showed that δ, γ′ and γ″ phases were observed around grain boundaries of IN 706 superalloy aged for different periods. The highest hardness was measured for the samples aged for 12 h. Weight losses were found to increase as the sliding distance increased. Moreover, the ANN modeling of weight loss values for IN 706 superalloy gave effective results and can be successfully used to predict weight loss values in the parameters that were determined by the obtained high R2 value.

Figure optionsDownload as PowerPoint slide