PSI - Issue 34

8

Tim Koenis et al. / Procedia Structural Integrity 34 (2021) 235–246 Tim Koenis et al. / Structural Integrity Procedia 00 (2019) 000 – 000

242

To obtain the cyclic stress field as employed in the fatigue analysis, the loading is simulated without the residual stress and strain from the process simulations. In this analysis, the model is clamped along the edge of the baseplate, and a load of 5 kN is applied at the hole in the deposited bracket. The stress field obtained from this analysis is imported in the fatigue life analysis using scaling from 1 to -1 of the stress values to obtain a cyclic stress field. Multiple fatigue life analyses have been performed adopting different means stress fields. The first analysis is without any residual stress, and thus a mean stress of 0. To observe the effect of the LMD process chain on the fatigue life, the residual stress field obtained after wire EDM is adopted in a second fatigue life analysis. Finally the residual stress field is obtained when heat treatment is not included in the virtual process chain. This residual stress field is adopted in a third fatigue life analysis to observe the effects of heat treatment on the fatigue life of a LMD manufactured part. 3. Results and discussion 3.1. LMD process simulation calibration and validation Fig. 7 a displays the measured interlayer temperature over the normalized manufacturing time of the T-shaped part deposited with strategy 1, which is used to calibrate the thermal model. To calibrate the thermal model, the laser absorptivity, emissivity, convection coefficient and conductive heat transfer coefficient are varied to match the trend and temperature values obtained experimentally. The calibrated thermal parameters are displayed in Table 2 and relate well with values observed in literature such as by Chiumenti, et al. (2017) and Ye et al. (2018). These parameters have been employed to model the use of strategy 2 and strategy 3 to verify the thermal modelling method. Fig. 7 b shows the comparison between the experimental and numerical interlayer temperatures over the normalized manufacturing time for both strategies. As can be observed in this figure, the experimental results are less consistent compared to the calibration experiment, as multiple breaks were required to correctly manufacture the parts. However, these breaks are captured in the machine log data and thus can be implemented in the numerical simulations. The simulations generally show good correspondence with the experimental data, which is sufficient to implement the temperature

(a)

(b)

results in the mechanical simulation.

Fig. 7. (a) interlayer temperature profile for model calibration; (b) interlayer temperature profiles for model validation. Fig. 8 (a) illustrates the extraction of the plate deformation to use in the mechanical calibration from the 3D scans of the deformed baseplate and deposition. The coordinates of the nodes located near the edge indicated by the red arrow and line in Fig. 8 (a) have been extracted for all three manufacturing strategies. These coordinates are displayed in 2D in Fig. 8 (b) to illustrate the deformation profile along this edge. From the numerical simulation of the T-shape manufactured with strategy 1, the same deformation profile has been extracted to compare with the experimental results. By varying the initial temperature at which elements are activated, the mechanical model has been calibrated