13th International Conference on Fracture June 16–21, 2013, Beijing, China —5— Spec. No. Specimen diameter /mm ΔS/S /% Damage coefficient λ Measured ΔAD/AD /% △AD /mm2 Calculated crack length /mm Constant d0/mm Real crack length /mm Corrected crack length /mm Relative error /% 1 10.07 5.33 0.33 16.32 12.82 1.0 2.20 2.10 4.39 2 10.08 9.35 0.47 19.85 15.59 1.5 2.54 2.60 -2.46 3 10.07 14.00 0.53 26.56 20.86 2.0 3.15 3.10 1.61 4 10.10 19.17 0.57 33.545 26.35 2.5 1.10 3.74 3.60 3.72 Table 1 shows that parameter d0 really is a constant. Although there is some difference between average of real equivalent crack length and that from prediction model, relative error is not more than 5.0%, which manifests that the prediction model is reliable and applicable. 5. Predetermination crack initiation life by means of the prediction model Generally, for low cycle fatigue test using smooth specimen, specimen will fail quickly once crack initiates, due to larger stress and strain acted on the specimen. For most of metallic materials, crack initiation life will dominantly take up 80% or so of the entire life. It is very difficult to directly measure crack initiation life through normal low cycle fatigue tests using smooth specimen. In order to investigate applicability of the prediction model, relation of crack initiation life to entire life was studied below. Simple-specimen-method was employed in order to avoid data distribution. Amount of cycle, corresponding to percentage of stress attenuation which relates to some equivalent crack length, is crack initiation life of the equivalent crack length. Different literatures defined different crack initiation, and it is difficult to determine one proper equivalent value for steel such as 10CrNiMo structural steel, which will rely on concrete test. In this paper, three of specimens were performed, with series of number 5 to 7, and percentage of stress attenuation was supposed to 0.65%, 2.0%, 5.33%, 9.35%, 14.0%, 19.17%, 20%, 30% and 50% from a low-level to a high-level, recording relevant repeated cycles. See figure 4 for final test results. Curve of △S/S-N in figure 4 indicates that with cycle increasing, percentage of stress attenuation is improved, and when the cycle is close to 800 to 1000 cycles, specimens fail quickly with a high speed. Based on above analysis, when repeated cycle attains to 800 to 1000 cycles, it is suitable taking corresponding equivalent crack length as crack initiation for 10CrNiMo steel. In the range from 800 to 1000 cycles, △S/S of the three specimens are all 5.33% and corresponding equivalent length of crack initiation is 1.0mm. Considering of correctness of the prediction model, real crack initiation should be as 2.1mm. See table 2 for data analysis of three specimens.
RkJQdWJsaXNoZXIy MjM0NDE=