PSI - Issue 57

Grégoire Brot et al. / Procedia Structural Integrity 57 (2024) 53–60 G. Brot et al. / Structural Integrity Procedia 00 (2023) 000–000

54

2

et al. (2023), the present authors showed that fatigue testing using ultrasonic fatigue machines was sensitive to porosity and microstructure variations in Ti-6Al-4V-LPBF and is suited for process parameter optimization. It was also shown that the quality of printing batch can be assessed using few ultrasonic fatigue specimens. Fatigue limit assessment using self-heating (SH) testing is another possible fast characterization approach that might be used for process optimization. In this method presented in Luong (1995), the threshold stress amplitude between two self-heating regimes, σ SH , lim , is determined and used as an estimation of the fatigue limit of the material. To do so, temperature of the specimen is monitored while performing loading steps with a fixed number of cycle. During loading steps, stress amplitude σ a is constant and increased from one step to the following. When analysis SH results, di ff erent thermal indexes can be used to determine the threshold stress amplitude σ SH , lim . Most studies used the steady-state temperature increment ∆ T stab but it is also possible to use the initial heating speed S heat , i (Finis et al. (2017)) or directly the intrinsic dissipation d 1 . Krapez et al. (2000) proposed using lock-in infrared thermography to demodulate the temperature signal; i.e. to extract the first harmonic A f , synchronous with the imposed force signal and the second harmonic A 2 f synchronous with twice the loading frequency. A f is related to the thermoelastic coupling. A 2 f is an indicator of all sources of non-linearity in the material, notably microplasticity and is less sensitive to the variations of thermal conditions during testing compared to ∆ T stab . In the results of Krapez et al. (2000) on di ff erent steel grades, A 2 f presented two self-heating regimes with a threshold stress amplitude σ SH , lim close to the fatigue limit. However, it was not possible to confidently determine σ SH , lim when testing aluminum 7010. As compared to steels, mush fewer SH testing results were published on titanium alloys. Bustos et al. (2023) used SH testing on Ti-6Al-4V produced using Electron Beam Melting (EBM) process under two load ratios ( R σ = 0 . 1, R σ = − 1). They found out that temperature increase was not significant for specimens with an as-built microstructure for both as-built and polished surface and it was significant for Hot Isostatic Pressed (HIPed) and wrought specimens. Roue´ et al. (2018) performed SH testing on wrought Ti-6Al-4V at elevated temperature (200 and 400 ◦ C) and under di ff erent load ratios. Ti-6Al-4V presented two distinct SH regimes when tested at these temperatures. They found out that load ratios and testing temperature a ff ect the threshold stress amplitude but not the slope of the first and second SH regimes in the chart: log ( ∆ T stab ) = f ( log ( σ a )). The sensitivity of self-heating testing to microstructure and to porosity variations is not fully understood. Munier (2012) studied di ff erent steel grades using SH testing. He observed two self-heating regimes for all the studied grades with di ff erent dissipation intensities and di ff erent σ SH , lim . He also noticed a reduction of the dissipation after a stress relieving treatment even though this treatment did not a ff ect the threshold stress amplitude σ SH , lim . SH testing was also used on porous cast material (Ezanno et al. (2010)) or on porous additive manufactured materials (Bustos et al. (2023)) but without assessing the sensitivity of SH behavior to the porosity level. In this work, in order to study the e ff ect of microstructure and porosity on self-heating-based fatigue limit assess ment, SH tests are performed on di ff erent material grades of Ti-6Al-4V-LPBF. These grades have di ff erent porosity levels or microstructures. Moreover, SH tests were performed using two distinct load frequencies (15.67 and 800 Hz). Finally, SH behavior of virgin or previously deformed specimen ( ε pl = 4 %) were compared. SH fatigue tests are performed on five material grades of Ti-6Al-4V-LPBF. Three grades have the same microstruc ture but di ff erent porosity levels and three grade have the same porosity level but di ff erent microstructures. The di ff erent processing routes developed to obtained these grades and the characterization of their microstructure and porosity are detailed in Brot et al. (2023). Porosity levels P 1 ≈ 0 . 001%and P 2 ≈ 1 % are generated using di ff erent LPBF process parameters. Then, three distinct microstructures are produced using di ff erent thermal post-treatment. The third porosity level P 0 , almost fully dense, is obtained using a Hot Isostatic Pressing (HIP) treatment. The five studied grades are called according to their porosity level and the temperature of their thermal treatment: P 1 650 ◦ C, P 0 920 ◦ C, P 1 920 ◦ C, P 2 920 ◦ Cand P 1 1020 ◦ C. Microstructure 650 ◦ C is an ultrafine lamellar one in which width of α / α ’ lamellae is 0 . 4 ± 0 . 1 µ m. Microstructure 920 ◦ C is lamellar with a mean width of α lamellae of 2 . 2 ± 0 . 3 µ m. Super β -transus treatment at 1020 ◦ C generates an equiaxe microstructure in which colonies of α + β lamellae have the same orientation. These grains are 200 ± 24 µ m large. Following thermal post-treatment, specimens were machined to their final geometry presented on Figure 1 and then polished. For comparison purpose, wrought Ti-6Al-4V was also 2. Material and methods