Issue 62

D. Milone et alii, Frattura ed Integrità Strutturale, 62 (2022) 505-515; DOI: 10.3221/IGF-ESIS.62.34

Tab. 1 reports the expected (i.e. assessed by the operator) and the predicted value of the limit stress by the neural network. Each material has been identified according to its use. If it has been used as a training set its identifier is "TRAIN", if it has been a part of the test set its identifier is "TEST". Finally, if it has been used for validation, its identifier is "VALIDATION".

(a) (b) Figure 11: a) PA12 Expected vs. Predicted limit stress (validate); b) PA66GF35 Expected vs. Predicted limit stress (validate).

Expected Value

Predicted Value

RMSE

Material

TEST TYPE

t[s]

T[K] σ lim [MPa] t[s] T[K] σ lim [MPa] t[s]

T[K] σ lim [MPa]

TRAINING TRAINING

8.61 6.31 6.65 7.20 8.49

-0.21 -0.28 -0.34 -0.21 -0.25 -0.30 -0.06 -0.14 -0.18 -0.84 -0.78 -0.81 -0.70 -0.45 -0.29 -0.93

38.41 32.38 31.50 37.66 38.45

7.88 -0.21 7.17 -0.28 7.13 -0.34 7.88 -0.21 7.62 -0.26

38.95 32.92 32.04 38.20 38.21

0.520 0.0001 0.600 0.0001 0.330 0.0001 0.480 0.0007 0.620 0.0007

0.5172 0.6021 0.3345 0.4785 0.6199 0.0554 0.0176 0.6218 0.0029

PA66GF35 (composites)

TEST

VALIDATION VALIDATION

TRAINING

93.27 87.57 98.09 31.94 53.87 62.17 36.26 40.19 31.15 15.08 32.99

279.96 262.45 294.49 213.32 169.32 192.96

93.35 -0.30 280.02 0.050 0.0011 87.60 -0.07 262.47 0.020 0.0036 97.21 -0.14 295.11 0.620 0.0001 31.93 -0.18 213.32 0.002 0.0001 53.82 -0.84 169.35 0.030 0.0005 62.12 -0.78 193.00 0.040 0.0005

AISI316L (steel)

TEST

VALIDATION

C45 (steel)

TRAINING

TEST TEST

0.033

S355 (steel)

0.0406

TRAINING

29.82 30.66 27.56 13.75 13.84

36.26 -0.81 40.19 -0.70 32.98 -0.40 15.07 -0.29 34.48 -0.86

29.82 30.66 28.85 13.75 14.65

0.005 0.00003 0.0005 0.001 0.00001 0.0017

PA12 (plastic)

TEST

VALIDATION

1.294 0.0353

0.9121

TRAINING

0.005 0.00004 0.0045

PE100 (plastic)

VALIDATION 1.053 0.04949 0.5727 Table 1: Comparison of expected vs. predicted values of the limit stress for several kind of material.

The trained network is able to predict the transition temperature at which the first damage occurs. As can be seen from Tab. 1, regarding the predictions made by the neural network, a maximum of 1.2 s of difference between the obtained value and the predicted value can be attested on the entire data set (PA12 tensile tests). As far as the temperature is concerned, a value of 0.049 K (PE100) is attested as the difference between the value obtained and the predicted value. Finally, with

513