PSI - Issue 40

V.S. Kanakin et al. / Procedia Structural Integrity 40 (2022) 194–200

195

2

Kanakin V.S. et al. / Structural Integrity Procedia 00 (2019) 000 – 000

1. Introduction Metal matrix composite materials are becoming increasingly widespread due to their higher physical and mechanical properties in comparison with the alloys on which they are based (Chawla and Chawla, 2013; Samal et al., 2021; Smirnov et al., 2018). In turn, this leads to the appearance of a large number of new materials, whose properties and behavior under various external influences have yet to be studied. As a consequence, it is necessary to study the properties of new synthesized composites and to construct their mathematical models for designing technological processes of material processing in order to optimize the thermomechanical conditions of workpiece loading for producing a part with high performance properties. For this purpose, technological processes are simulated, and this requires information on the rheological behavior of the material under conditions of assumed thermomechanical loading. To solve this problem, a large number of mathematical models of flow stress have been developed. They can be divided into several types: phenomenological, structural-phenomenological, physically based, and neural networks (Kondratev and Trusov, 2016; Konovalov and Smirnov, 2008; Lin and Chen, 2011; Lin et al., 2008; Smirnov et al., 2020). The advantage of neural networks over all the other types of models seems to be that they are well suited for approximating experimental data without the need for deep immersion in understanding the physical processes occurring in the material. As a result, it becomes possible in principle to predict the behavior of the material under conditions for which no experimental data have been obtained and to study the mechanisms proceeding on their basis. The purpose of this work is to build a neural network for describing and predicting the flow stress of the AlMg6/10% SiC metal matrix composite at deformation temperatures ranging from 300 to 500  С. 2. Materials and research methodology The AlMg6/10% SiC metal matrix composite was manufactured by powder metallurgy technologies (Smirnov et al., 2018). The size of the used SiC particles with an average diameter of 2.0±0.4 μm corresponded to the F1500 standard. The diameter of the AlMg6 alloy particles ranged between 1 and 60 μm. The AlMg6 alloy (the 1560 aluminum alloy according to GOST 4784-97) had the following chemical composition, wt%: Al-92.4, Mg-6.56, Mn-0.5, Fe-0.27, Si-0.16, Cu-0.013, Ti-0.04, Zn-0.02, and Be-0.0012. The chemical composition was determined by means of a Spectra Maxx LMF04 analyzer. The microstructure of the specimens before deformation was studied by the electron backscatter diffraction (EBSD) technique using a Vega II Tescan scanning electron microscope with an EBSD Oxford HKL NordlysF+ analysis accessory. The size of the scanned step was 0.3 μm for the analysis of specimens by the EBSD method. The grains were considered to have a misorientation greater than 15°, with a subgrain misorientation of 2  to 15  . Cylindrical specimens were tested for compression on a plastometric installation designed by the Institute of Engineering Science, Ural Branch of the Russian Academy of Sciences. The compression specimen had the diameter d 6 ±0.05  0 mm and the height h 9 ±0.05  0 mm. In compression tests, a graphite-based lubricant was

used to reduce friction between the punch and the specimen. 3. The rheological behavior of the composite and its modeling

The rheological behavior of the AlMg6/10% SiC metal matrix composite and its modeling were studied for deformation temperatures ranging from 300 to 500  С and strain rates ranging from 0.1 to 5 s − 1 . The specimens were annealed at 520 °C for 15 hours before deformation. The microstructure of the speci mens is shown in Fig. 1 (Smirnov et al., 2018). As it can be seen from Fig. 1b, the particles of the SiC reinforcer are located along the boundaries of the sintered particles of the AlMg6 alloy matrix. The AlMg6/10% SIC MMC matrix had a granular structure (Fig. 1a) with an average grain diameter of 5 μm.