PSI - Issue 59

Oleh Yasniy et al. / Procedia Structural Integrity 59 (2024) 17–23 Oleh Yasniy et al. / Structural Integrity Procedia 00 (2019) 000 – 000

21

5

a

b

0 100 200 300 400 500 600 700 800

0 100 200 300 400 500 600 700 800

Npred, cycle

Npred, cycle

0

200

400

600

800

0

200

400

600

800

N, cycle

N, cycle

Fig. 4. The predicted and experimental dependencies of the number of loading cycles obtained by the method of random forest (a) and neural network (b) a b

4,5

4,5

Exp Pred

Exp Pred

4

4

3,5

3,5

Δε , %

Δε , %

3

3

2,5

2,5

0

200

400

600

800

0

200

400

600

800

N, cycles

N, cycles

Fig. 5. The predicted and experimental dependencies of the strain range of NiTi alloy on the number of loading cycles obtained by the method of random forests (a) and neural networks (b) a b

800

0 100 200 300 400 500 600 700 800

600

400

200

Npred, cycle

Npred, cycle

0

0

200 400 600 800

N, cycle

0

200

400

600

800

N, cycle

Fig. 6. The predicted and experimental dependencies of the number of loading cycles, obtained using the method of random forests (a) and neural networks (b) The error of the random forest method was 7% for the test set and 7.5% for the neural network method. Fig. 7