Multi-Phase Field Method for Solidification Microstructure Evolution for a Ni-Based Alloy in Wire Arc Additive Manufacturing
Round 1
Reviewer 1 Report
This paper reports multi-Phase field method for solidification microstructure evolution for an Ni-based alloy in wire arc additive manufacturing. Some comments are listed below:
1. In Fig. 5, the text size is not large enough for readers.
2. Fig. 6 shows microstructure but it is impossible to see PDAS. Images with high resolution are needed.
3. In Fig. 10, there is no error bar.
Author Response
Dear Reviewer
We wish to express our appreciation to the Reviewer for their insightful comments, which have helped us to improve the paper significantly.
Comment 1: In Fig. 5, the text size is not large enough for readers.
Response: This figure was redrawn with enlarged letter and thick line.
Comment 2: Fig. 6 shows microstructure but it is impossible to see PDAS. Images with high resolution are needed.
Response: These figures were replaced with originally measured high resolution pictures.
Comment 3: In Fig. 10, there is no error bar.
Response: Experimental error range values in Table 6 were laid on Figure 10. MPFM calculation errors were estimated from Figure 9. These values were put on Table 6 and laid on Figure 10.
We wish to thank the Reviewer again for their valuable comments.
Reviewer 2 Report
Microstructure evolution has become increasingly attractive while metal additive manufacturing moves rapidly in both academics and industry. The authors simulated a Nickel-based alloy single wall by WAAM process using a multi-phase field method coupled with CALPHAD calculation. As far as I’m aware, it’s novel, and the method and the findings are helpful to other researchers and engineers in the field. Particularly, the predicted micro-segregation of solute elements looked accurate and meaningful even in such complex microstructures, and the diffusion-controlled primary dendrite arm spacing, even though the latter wasn’t well supported by their experimental measurements. As the authors said, the PDAS from simulation and experiments qualitatively agreed with each other, but there were quantitative differences. It was pointed out the gaps could be linked to the absence of 3D simulation capability due to the limitations of the TQ-interface function. It may be, or may not be the case, but certainly the “quantitative differences” are worth further studies to investigate and confirm.
There are two minor issues which could be improved (one is a typo, I believe):
1) The microstructures images in Figure 6 are not good enough to show the primary dendrites and their “arms”. Better quality micrographs are needed.
2) The “x-y cross section” in Figure description should be “x-z cross section” instead?
Summarily, I think it’s a good work, and recommend to accept after minor corrections and improvement are made.
Author Response
Dear Reviewer
We wish to express our appreciation to the Reviewer for their insightful comments, which have helped us to improve the paper significantly.
Comment 1: The microstructures images in Figure 6 are not good enough to show the primary dendrites and their “arms”. Better quality micrographs are needed
Response: These figures were replaced with originally measured high resolution pictures.
Comment 2: The “x-y cross section” in Figure description should be “x-z cross section” instead?
Response: I really appreciate your indication. In the caption of Figure 1, “x-y cross section” was wrong, it was replaced with “x-y cross section”.
We wish to thank the Reviewer again for their valuable comments.
Reviewer 3 Report
1. Can the authors explain whey they chose an arc power efficiency of 0.2?
2. In Figure 2, the leading edge of the added (vertical) layer is in "red color" in (b) and (d), but then at a lower temperature in (c), which is at a distance between (b) and (d); why is the temperature at the leading edge of the added layer experiencing this temperature dip? The substrate seems to be at a uniform temperature.
3. Since the cooling rates are much lower than in LPBF, does dendrite coarsening play a role? Secondary dendrite arm coarsening?
4. What segregation model was used for the calculations?
5. In the conclusion the authors mention a constant solidification velocity. But this velocity likely changes during the melt-pool solidification. Can the authors please comment on this solidification velocity.
Author Response
Dear Reviewer
We wish to express our appreciation to the reviewer for their insightful comments, which have helped us to improve the paper significantly.
Comment 1: Can the authors explain why they chose an arc power efficiency of 0.2?
Response: This value was chosen by comparing melt poor size in the first layer between FEM analysis and experimental measurement to be approximately same each other. This explanation is inserted in lines 92-94 of the revised paper by blue letter.
Comment 2: In Figure 2, the leading edge of the added (vertical) layer is in "red color" in (b) and (d), but then at a lower temperature in (c), which is at a distance between (b) and (d); why is the temperature at the leading edge of the added layer experiencing this temperature dip? The substrate seems to be at a uniform temperature.
Response: In Birth and Death model of FEM in this study, the deposition was discretized by the hexahedral element of constant size, 0.5 mm×0.5 mm×0.5 mm. Temperature in the leading edge fluctuates because added area is periodically heated periodically by the discretized heat element. This temperature fluctuation can be seen over 1350℃ in Figure 3. The fluctuation disappears as heat diffusing with time under 1350℃. Figure 2 is replaced with that temperatures at the leading edges are selected as same level between (a), (b), (c) and (d), for readers not confused.
Temperature in the substrate region was also solved by FEM, in which boundary condition was set as insulated. Temperature far form fusion area in the substrate is considered to be not so increased in the time scale of this thermal analysis.
Comment 3: Since the cooling rates are much lower than in LPBF, does dendrite coarsening play a role? Secondary dendrite arm coarsening?
Response: Phase field method flamework includes prediction of coarsening. Secondary dendrite arm length tends to be short in Figures 7, 8 and 9. Typical coarsening can not be seen in the narrow liquid space between the primary dendrites. From experimentally measured microstructures in Figure 6, which are replaced by higher resolution pictures, typical coarsen of secondary dendrite arm also can not be seen.
Comment 4: What segregation model was used for the calculations?
Response: Phase field method coupled with CALPHAD database provides solute partitioning in interface by quasi-equilibrium condition. Thus, this method includes segregation calculation mechanism with solving solute diffusion equation.
Comment 5: In the conclusion the authors mention a constant solidification velocity. But this velocity likely changes during the melt-pool solidification. Can the authors please comment on this solidification velocity.
Response:. Your feedback is highly appreciated. I found a typo in table 3. Temperature gradient value in 1p is corrected as 9000K/m, not 90000K/m.
Explanation of steady solidification velocity definition is inserted in lines 269-270 by blue letter. Explanation of inverse relation between 1-layler and 5&10-layers is replaced in line 273 by blue letter.
In the present MPFM calculations, Figs.7-9, unsteady microstructure evolutions were performed in narrow space of the melt pool even if using constant cooling rates and temperature gradients. In order to obtain solidification microstructure of constant solidification velocity estimated by cooling rate and temperature gradient, it is considered that so long calculation time and vertical region size will be necessary. This point is inserted in lines 178-180 of the revised paper by blue letter.
We wish to thank the reviewer again for their valuable comments.
