PSI - Issue 57

Martin Matušů et al. / Procedia Structural Integrity 57 (2024) 327 – 334 Author name / Structural Integrity Procedia 00 (2019) 000 – 000

333

7

was thus set to 7000 cycles. With higher load amplitudes, the transition to Phase 2 occurs earlier. The testing blocks with amplitudes higher than 100 MPa were tested with the block length of 5000 cycles. Fig. 5 illustrates that the most noticeable differences are observed in the upper section of the step-test, where the stabilized temperature varies significantly amongthe series. The highest and the lowest recorded temperatures were found in T300 and T200 series, respectively, with T240 and NoHT series exhibiting intermediate values. A significant temperature increase is observed above 90 MPa, well above the observed fatigue limit values from Fig. 3. A closer examination of the trends in Fig. 5, right reveals that the position of the first knee point could be speculated to be around 65 MPa, as depicted in the close-up graph in Fig. 5. Such value would be much closer to the fatigue limits found from the S-N curve results shown in Fig. 3.

Fig. 5. The Self-Heating step tests result in four different cases representing each of the four different heat treatments. A close-up of the region between 30-90 MPa was taken to observe the transition from the region under the fatigue limit (from K&V regression at 10 7 cycles) with points connected via line function, that does not represent any regression, but to emphasize the bilinear trend in this region. Some researchers [7; 8] claim that there should be one knee point in a bilinear evolution and that it should be related to the position of the fatigue limit. When observing the trends in Fig. 5, and not knowing the S-N curves in Fig. 3, one would be tempted to estimate the fatigue limits of all cases to about 90-100 MPa. The first transition at lower stress amplitude values is less visible, which can be partly caused by using uncooled thermal cameras with lower resolution. However, if the first knee point should be focused and found, a decision at which stress amplitudes values to clip the records would have to be taken – otherwise the right-most part of the trends will hide it. 3. Conclusion This paper evaluates the influence of 4 different heat treatments (HT) on static and fatigue behavior of AM-built specimens from AlSi10Mg. From the perspective of the high-cycle fatigue performance, the best HT concerns the T300 variant (300 °C for 2 hours and water quenched). It yields the highest fatigue strengths in HCF region, despite the silicon network being decomposed, which results in lower tensile properties. If the low-cycle fatigue response is to be optimized however, the T200 series with artificial age-hardening at 200°C for 2 hours / furnace and air cooled performs the best. Overall, any heat treatment seems to improve the fatigue response compared to untreated NoHT series in this case. From the perspective of heat dissipation during fatigue testing, the step self-heating test with increasing stress amplitudes was used to assess the regions of stabilized temperature for each load amplitude. The results show that though some hints of a bilinear dependency between stress amplitude and the observed temperatures can be detected and the knee point might relate to the fatigue limit position, a third region starting at higher stress amplitudes masks it very well. A consistent use of this solution for fatigue limit detection would require defining a criterion for clipping the stress-temperature dependency at a specific stress level. Such criterion is missing today.