PSI - Issue 75

13

Fabrice Deleau et al. / Procedia Structural Integrity 75 (2025) 392–418 Emmanuel Persent, Deleau Fabrice, Guillaume Coudouel, Guillaume Perrin/ Structural Integrity Procedia (2025)

404

A Basquin’s model (Manin et al., 2010) for the ( , ) a N  relationship is considered because it fits well the experimental results of the S-N data. It is written in equation (21), and the parameters log A and n are given in Table 2: ( )  − = a n N A or log( log )  = + a B C N (21)

Table 2: Fatigue parameters for steel forging in air and seawater with cathodic protection subjected to a high mean tensile stress (S-N curve “ DNV BM3” for Ra <3.2 µm)

Parameter

Air DNV BM3

Seawater with cathodic protection DNV BM3

log A

16.770

16.322

5

5

n

Stress range at 2×10 6 cycles [MPa]

248 (  36 ksi)

206 (  29.9 ksi)

The fatigue resistance of i-Clip / small sized prototype material can also be evaluated with Basquin calibrated model based on standard fatigue test results (Fig. 11). The high cycle fatigue tests were conducted according to the standard ASTM E466-21. It should be noted that no-failed specimens are not considered to estimate the fatigue resistance. These experimental values are done at 2×10 6 cycles for 95% survival probability and a confident level of 0.95. 3. Application of the improved methodology 3.1. Application of the improved methodology and estimated service life • Procedure Each of the three methodologies (API, DNV and IFPEN proposal) is applied to calculate the stress amplitude and the number of cycles to failure for the i-Clip connector, and more specifically for the small-sized prototype. The proposed procedure is a post-treatment performed after the cyclic computation treated by the Finite Element code, using an elastic-plastic material law for the connector. A material law is identified from tests conducted on the connector's material. The different stress amplitudes are calculated from the following equations : API a  from equation (3), DNV a  from equation (12), and a P  from equation (20). Then, the corresponding number of cycles to failure N for each of them is calculated using equation (21). • Results First, the formulation of each criterion being different in terms of range, mean stress, and correction, the fatigue hotspot (i.e the location providing the highest value of a  inside the lug grooves) is different for each methodology. An axial tension solicitation of 70 kN ± 50 is considered for a 1:5 small-sized prototype, equivalent to 7,000 kN ± 5,000 (~1,570 kips ± 1,120), for the i-Clip connector. 2 a    = at the hotspot, and the corresponding number of cycles, according to each methodology , and based on both “BM - 3” and “ i-Clip material ” S-N models. For our proposed methodology, the effect of preload was investigated. The preload considered is equal to the maximum capacity of the small sized prototype, which is 160 kN. It can be noted that the preload significantly reduces the mean stress as well as the stress range. The service life is thereby improved. The estimated number of cycles is multiplied by approximately 20. Table 3 presents the values of the corrected stress range