PSI - Issue 5

Tomasz Tomaszewski et al. / Procedia Structural Integrity 5 (2017) 840–847 Tomasz Tomaszewski et al. / Structural Integrity Procedia 00 (2017) 000 – 000

846

7

summary is presented in Table 1. The K S coefficient may be deemed as the correction value used to shift the characteristics by the specified value. Such approach is justified in the area of the statistical size effect and is insufficient for the analyzed bending loads. As a result, fatigue characteristics σ a -N for the specified model of highly stressed volume was estimated (based on Eq. (2)) and the K S coefficient was taken into account. The results of implementation of the model and the experimental tests are presented on Fig. 7, and in Table 8. For a 95% probability (significance level α = 0.05) curves were determined that narrowed down the confidence interval.

Fig. 7. Fatigue characteristic determined based on experimental tests and analytical models

Table 8. Values of linear regression model parameters, error of the estimate against the empirical values.

Standard error of the estimate

Coefficient of residual variation

Linear regression line log σ a = a log N + b

n

1   i

2

(

)







S

Data

, a exp i

, a cal i

V

e

100% 



S



e



e

n

2



a exp

a

b

S e [MPa]

V e [%]

Experiment – minispecimen Experiment – standard specimen

-0.1104 -0.0689 -0.1104 -0.1104

3.232 2.941 3.163 3.134

9.2 9.3

2.0 2.0

Highly stressed volume

70.1 36.4

15.4

Highly stressed volume, coefficient K S

8.0

Fig. 8. Percent error for implemented models analytical models