PSI - Issue 22

Yang Ai et al. / Procedia Structural Integrity 22 (2019) 70–77 Y. Ai et al. / Structural Integrity Procedia 00 (2019) 000 – 000 As aforementioned, once , , ( ) and / 0 were determined, using the already known 0 ( , ) and ( , ) under the maximum local stress , as inputs, the resulting output is the predicted number of loading cycles ( , ) corresponding to ( ) and , . Therefore, under the assumption of equivalent highly stressed volume, the proposed probability model is applied for describing both effects of size and notch. 3. Model validation and comparison 3.1 Determination of highly stressed volume In this section, the experimental data of titanium alloys and aluminum alloys are utilized for model validation. Each geometry of the tested specimens is indicated in [22, 23], where TC11 alloy [22] are round bar specimens, Al 7075-T6 alloy [23] with sheet specimens. All presented specimens are tested under R= -1. By the highly stressed node method, the highly stressed volume subjected to a stress level between 0.8 , [13] and , are shown in Table 1 . Table 1 Calculation results of the highly stressed volume TC11 specimens The highly stressed volume ℎ ( 3 ) Al 7075-T6 specimens The highly stressed volume ℎ ( 3 ) = 1.6 (A) 4.257 = 1 5162.44007 = 1.6 (B) 31.018 = 2 22.9656 = 2.06 1.363 = 4 0.595856 3.2 Size effect analysis This section presents predicted results obtained using proposed probability model under the experimental data of reference specimens. To verify the applicability of the established procedure on describing the statistical and geometrical size effects, the tested specimens of each alloy are respectively defined as three kind of specimens i.e. the reference specimen, the specimen for model calibration and the predicted specimen. Fatigue life distribution of predicted specimens can be established under the reference specimen and the specimen for calibration. Especially, for each of materials, six predictive combinations can be defined from the specimens of three geometric shapes. Furthermore, for same corresponding volume relationship, it's worth noting that the interchange of reference specimen and the specimen of calibration can obtain the same model coefficient ( , ) . Hence, under developed procedure, three sets of ( , ) and predicted results for each material are respectively showed as presented in Table 2-3 and Fig. 1-2 . 73 4 Table 2 Parameters of model coefficient ( , ) for TC11 specimens Table 3 Parameters of model coefficient ( , ) for Al 7075-T6 specimens Reference specimen Specimen for calibration Predicted specimen Parameters of model coefficient = 1.6 (A) = 1.6 (B) = 2.06 −2.215 × 10 −7 9.388 × 10 −4 −6.22 × 10 −1 = 1.6 (A) = 2.06 = 1.6 (B) −3.653 × 10 −7 1.609 × 10 −4 −1.003 = 1.6 (B) = 2.06 = 1.6 (A) −3. 63 × 10 −7 1.252 × 10 −3 −7.98 × 10 −1 Parameters of model coefficient = 1 = 2 = 4 5.015 × 10 −8 −6.464 × 10 −5 2.794 × 10 −1 = 1 = 4 = 2 −1.775 × 10 −7 2.887 × 10 −4 2.383 × 10 −1 = 2 = 4 = 1 −5.277 × 10 −7 8.215 × 10 −4 1.759 × 10 −1 Parameters of model coefficient Parameters of model coefficient Reference specimen Specimen for calibration Predicted specimen Parameters of model coefficient Parameters of model coefficient