System Identification Technique and Neural Networks for Material Lifetime Assessment Application - Complex System Modelling and Control Through Intelligent Soft Computations - page 780

Fig. 18 Multivariable and
multiaxial fatigue life
prediction of the NN-NARX
model for R and ʸ values of
the testing sets: R0.5 ʸ 0 ° , R-
1 ʸ 0 ° , R0.1 ʸ 15 ° , R-1 ʸ 30 ° ,
R10 ʸ 30 ° , R0.1 ʸ 45 ° ,
R0.5 ʸ 45 ° , R-1 ʸ 45 ° , R10 ʸ 45 ° ,

R-1 ʸ 60 ° , R10 ʸ 60 ° , R0.1 ʸ 75 ° ,

R0.1 ʸ 90 ° , R-1 ʸ 90 ° and

R10 ʸ 90 ° for E-glass/polyester

Fig. 19 Multivariable and
multiaxial fatigue life
prediction of the NN-NARX
model for R and ʸ values of
the testing sets: R0.5 ʸ 0 ° , R-
0.5 ʸ 0 ° , R-1 ʸ 0 ° , R-2 ʸ 0 ° ,
R0.1 ʸ 90 ° , R0.5 ʸ 90 ° , R-
0.5 ʸ 90 ° , R-1 ʸ 90 ° , R-2 ʸ 90 °
and R10 ʸ 90 ° for E-glass
fabrics/epoxy

R = − 2: ʸ =90 ° . The discrepancies will be examined further in the related S-N

curves.
Figures 20 and 21 show the S-N curves of E-glass/polyester obtained by the NN-
NARX model and the experimental data for stress ratios R =

− ʸ °

and
R = 10: ʸ =30 ° , respectively. Furthermore, Figs. 22 and 23 show the S-N curves of
E-glass fabrics/epoxy obtained by the NN-NARX model and the experimental data

1: =90