PSI - Issue 37

Mohammed Algarni et al. / Procedia Structural Integrity 37 (2022) 676–683 Mohammed Algarni/ Structural Integrity Procedia 00 (2019) 000 – 000

681

6

MATLAB was used to calibrate two material parameters in the KV fatigue model: α = 600 and β = -0.21. The KV model predicted the fatigue life of the large notch accurately yet failed to predict that of the medium and sharp notched specimens with the help of K f , as shown in Figure 3(a). Therefore, the fatigue notch factor was replaced with the factor of stress triaxiality in Equ. (3) to include the effects of notches. The parameters C η = 4.5 and η o = 0.33) are material constants calibrated by MATLAB, and a and r are the geometrical parameters presented in Figure 1 and Table 1. This factor is explained in detail by (Algarni, 2019). The combination of the stress triaxiality factor from Equ. (4) with the KV model in Figure 3(b) showed excellent fatigue life predictions for all notch sizes. = 1 + ( − 1) where = 1+ 1 w/ (1) = ( + ) (2) (  ) = [1 −  (  −  )] (3)  = 3√Ω 1+2Ω 2 + Ω+1 , where Ω = 2 (1 + 2 ) (4)

Figure 3: KV fatigue life model (at R= -1) failure to predict the influence of notches with fatigue notch factor (a) and accurately predicted fatigue life with the stress triaxiality factor (b).

4.2. Effect of mean stress on fatigue life model

The mean stress effect on the fatigue life is more significant in the HCF than in the LCF regime (S. Lu et al., 2018). Many fatigue models consider the mean stress influence on fatigue life, the most common being is Gerber model (1874), Goodman model (1919), Soderberg model (1930), Morrow (1968), SWT model (1970), and Walker model (1970). Based on the literature review, the Walker model is recommended for the mean stress influence on thermoplastics materials (Z. Lu et al., 2018; Mellott & Fatemi, 2014; Mortazavian & Fatemi, 2016). This model is presented in Equ. (5), where the equivalent stress amplitude at zero mean stress is S ao , A is the stress at N f = 1, γ is a material constant, and b is the fitted fatigue life slope. The model prediction with A = 44.2, b = -0.08, and γ = 0.23 predicted the experimental results accurately, as shown in Figure 4. Note that γ is relatively low, which indicates that polymer materials are highly sensitive to mean stress. = ( ) ( 1− 2 ) 1− (5)