PSI - Issue 36

Iryna Didych et al. / Procedia Structural Integrity 36 (2022) 166–170 Iryna Didych et al. / Structural Integrity Procedia 00 (2021) 000 – 000

167

2

class of dispersions

i f

frequency y prediction predicted element of sample y true

true value of the sample element

sample size

n

1. Introduction Machine learning is an area that has developed from artificial intelligence. At present it affects many industries and science. Machine learning is often used in autonomous robotics, computational biology, etc. One of the reasons for the machine learning widespread was the scaling of data flow in the network and the growing degree of automation by Konovalenko and Maruschak (2021). In particular, the amount of accumulated data is being increased. Therefore, in the presence of experimental data, it is advisable to apply machine learning methods for data modeling by Hastie et al. (2009). It is known that by using the machine learning methods, in particular, neural networks, it is possible to predict the fatigue crack growth rate with high accuracy, as demonstrated by Pidaparti et al. (1995), Mohanty et al. (2009), Yasnii et al. (2018, 2020). Moreover, methods of machine learning are powerful tools to model the stress – strain diagrams of AMg6 aluminum alloy, shown by Yasniy et al. (2020). The aim of this study was to predict the magnitude of the jump during the creep by method of machine learning, in particular, neural networks. 2. Material and methods 2.1. Analysis of the jump-like creep of AMg6 aluminum alloy AMg6 alloy successfully competes with steels, titanium and its alloys due to low specific weight and high resistance to fracture under the static, cyclic and dynamic loads, good corrosion resistance and manufacturability by Fedak (2003). It is widely used in aircraft and shipbuilding for the production of load-supporting structures. Therefore, the study of the mechanical properties of this material is very important. It is noted in the previous investigations by Yasniy et al. (1998) that the plastic deformation of AMg6 alloy is accompanied by jump-like increments. The total deformation of the jumps exceeds all other deformation in 5 – 10 times. It is known that under the conditions of uniaxial tension test of a composite materials the stress – strain diagram is observed at the soft type of loading by Strizhalo et al. (1997), (1993), (1999), Fedak (2003). In particular, by Yasnii et al. (2002) the jump-like increments of strain were revealed in both creep and dynamic creep experiments. Experimental studies showed that the creep of AMg6 aluminum alloy is accompanied by jump-like increments of strain, the values of which are similar to the "steps" of the tensile strain by Fedak (2003). One distinguishes four typical regions on the creep curve of AMg6 aluminum alloy (Fig. 1).

р

 p

I

II

I І

І V

р ст

III

t