Modelling of Technological Parameters of Aluminium Melt Refining in the Ladle by Blowing of Inert Gas through the Rotating Impeller
Round 1
Reviewer 1 Report
Please correct the chapter numbering! In the article san, there are 2x chapters 3.3.
In the case of specific results, the evaluation should also be specific and not slip into the level of vague expressions such as "showed slight worse results" or another term "slight increase".
Please specify by how much, at least as a percentage, the results have improved or deteriorated.
In CFD simulations, water and aluminum melt were compared for different input values of temperature and viscosity. On the basis of what criteria is the conformity in the obtained results evaluated?
Measurements and visualization with water were performed at different speeds of rotary impellers. Based on what criteria was the suitability of the speed used also for the aluminum melt evaluated? Could you describe it mathematically and add it to the article? Based on what were the speeds chosen in Table 2?
How does the inert gas temperature affect the melt degassing process?
How does the energy balance affect the process of blowing the inert into the melt and how does the temperature of the melt change?
If we could compare the described method of degassing using an inert gas with the method of vacuuming the melt, which method is more efficient and less energy-demanding?
Author Response
Dear Reviewer,
We would like to thank you for your detailed feedback to improve our submission. We have considered each comment carefully and revised our manuscript to address the issues raised.
The individual responses to your comments are given in the attached PDF.
Thank you very much for your suggestions.
Authors
Author Response File: 
Author Response.pdf
Reviewer 2 Report
This paper presents experimental and CFD modelling of an aluminium refining in which inert gas is sparged into a stirred vessel. The effect of rotor speed, gas flowrate and rotor submergence are studied experimentally.
The experimental part of this paper appears to have been carried out competently, and the results are of interest. The CFD modelling part of the paper does not appear to have been carried out competently, and is not well described. For reasons I give below, I suggest the authors remove the CFD modelling work from the paper, and do further work on it. The experimental part is fine by itself, after some revision. My major concerns are given as follows:
Line 258: “The turbulent model 258 was complemented by a multiphase VOF model”. This is simply not possible. The VOF (Volume-of-Fluid) method seeks to resolve all gas-liquid interfaces, which is simply not possible in the case being studied here because of the small bubble size and the huge number of bubbles. See various references about the VOF method in general, and the Fluent VOF implementation in particular:
Sun et al (2012) Development of a vapor–liquid phase change model for volume-of-fluid method in FLUENT. International Communications in Heat and Mass Transfer, 39, Issue 8, October 2012, Pages 1101-1106.
I don’t know what option the authors have used in Fluent. If they have turned on VOF, I suspect that the code has simply not been able to operate correctly in VOF mode, but the results may be suspect. I hope the authors have used the multifluid model.
Fig 14: I find it hard to believe that the CFD model gives such good agreement with the experimental results without some optimisation of parameters. It’s not even clear what model the authors use for the interphase area. They claim they use the VOF model, but inspection of Fig 11 shows that the bubble size is very small, and I cannot believe that they have a sufficiently fine mesh to resolve all of the interfaces for all of the bubbles. Given that this would need to be transient, the computing requirements would be far beyond what is possible even in 2D, much less 3D. If the authors did not in fact resolve all the interfaces, then they need to assume a value for bubble size in order to determine interfacial area. Let me be clear: it might an acceptable procedure to tune a parameter (such as bubble size) to achieve agreement, but given only one result is shown, the procedure does not achieve anything, certainly not validation. I furthermore find it highly suspect that the CFD results for water and aluminium are almost the same. Given the large difference in surface tension, I am sure the bubble size will, in reality, be significantly larger in the aluminium case. I doubt if the authors are talking surface tension into account.
I recommend the authors remove the section on CFD modelling, since I believe more work needs to be done before it is publishable. The experimental work is itself suitable for publication without CFD.
The authors do not address the issue of similarity between the water and aluminium models. There is a difference of roughly a factor of two in density which will affect the volume of gas (assuming that the quoted flowrates are in normal volumes. But more importantly, there is a large difference in surface tension, which means that one would expect smaller bubbles in the water system. Furthermore, one would expect a difference in solubility of the gases. I suppose the results will still be valid in a RELATIVE sense, but the authors need to make clear that they do not expect similarity.
One would expect essentially exponential decrease in the gas concentration for a first order process, in which case the two times plotted in Figs 8 and 9 would not be independent, but would differ by a constant factor. The authors could check this. In any case, they should remark that a constant factor is expected, even if they choose to plot both times just in case there are deviations from first order behavior.
Minor comments, often related to written expression follow. In addition, an annotated version of the paper with suggestions for improvement of the English expression is attached.
Line 39: “their regular distribution and regular arrangement of bubbles in the whole volume of the refining ladle”. 'Regular ' is not the right word here. Perhaps 'distribution' is all that is needed, or perhaps 'wide-spread distribution". I’m not sure what you difference you intend to draw between ‘distribution’ and ‘arrangement’. The word 'arrangement' is not usually used in such a context, and its meaning is not clear.
Line 40: “and with long period of their effect in the melt.” I assume you are referring to residence time here. You probably cannot arrange for bubbles to have a ‘long’ residence time, but you can arrange to avoid very short residence time.
line 49: “In the case of physical modelling, a method is used, in which the real system is replaced by a tangible physical model, which is as close as possible to the behavior of the real system.” Wouldn’t it be clearer to just say “In the case of physical modelling, the real system is replaced by a tangible physical model that is as close as possible, in behavior, to the real system.”
Line 52: “In this method, both the prototype and the model have the same physical substance.” This is surely usually NOT the case. Your physical model is water, whereas the real system is molten aluminium!
Line 134: “This phenomenon is visible on the curves steepness increasing in comparison between the individual rotary speeds.” This can be expressed more clearly: “The dependence on gas rate can be seen in the different steepness of curves at the same rotary speed.” By the way, the two reddish colors are so similar to my eye, that they are almost indistinguishable. Cold they be made more different? There is also a problem when one line overlaps another, which happens a couple of times.
Line 260: “A moving reference frame with rotating speed 500 rpm was used to model the boundary between the fluid zone surrounding the impeller and the mesh interface.” This sentence is not logical. You mean: “A moving reference frame with rotating speed 500 rpm was used to model the fluid zone surrounding the impeller (and was patched to a stationary reference frame for the rest of the tank), and a simplified numerical scheme for simulating the presence of the rotating rotor.”
Line 261: “A wall rotating with zero angular speed relative to rotating reference frame was used to model the boundary conditions at impeller surface.” No, the wall rotates at zero angular speed relative to the stationary reference frame used for the outer part of the tank.
Comments for author File: Comments.pdf
Author Response
Dear Reviewer,
We would like to thank you for your detailed feedback to improve our submission. We have considered each comment carefully and revised our manuscript to address the issues raised.
The individual responses to your comments are given in the attached PDF.
Thank you very much for your suggestions.
Authors
Author Response File: Author Response.pdf
Round 2
Reviewer 2 Report
This paper presents experimental and CFD modelling of an aluminium refining in which inert gas is sparged into a stirred vessel. The effect of rotor speed, gas flowrate and rotor submergence are studied experimentally.
As I said in my review of the original version, the experimental part of this paper appears to have been carried out competently, and the results are of interest. However, I said in my review of the original version that the CFD modelling part of the paper does not appear to have been carried out competently, and is not well described. The authors have not engaged with my major concerns about the CFD work; perhaps they simply do not have the background that would enable them to discuss CFD meaningfully. Therefore, I repeat that the authors should remove the CFD modelling work from the paper.
Let me emphasize that my concerns are not just the way the modelling has been described, and cannot be fixed by slightly improved expression. The fundamental concern I have is that the model is described as a VOF model (which it cannot be), and several other major problems flow from that misunderstanding by the authors. The authors have not even tried to justify their statement that the model is a VOF model. I suspect that ‘VOF’ is just an option in Fluent, of which the authors have no understanding. I will try again to suggest a paper that the authors should read to gain an understanding of what the VOF model is: Zahedi et al (2014) Influence of fluid properties on bubble formation, detachment, rising and collapse; Investigation using volume of fluid method. Korean J. Chem. Eng., 31(8), 1349-1361 (2014) DOI: 10.1007/s11814-014-0063-x. This paper uses the Fluent VOF capability (there are many such papers, and I do not guarantee this particular work, but it appears to have been well carried out and described). You will see, when you read the paper, that they simulate a maximum of 3 or four bubbles in the computational domain. The experimental images in your paper indicate that the tank contains thousands of such small bubbles. It is not computationally possible to simulate so many bubbles using the VOF method – it would require far too much computer time and resources. Indeed, the plots you show of your CFD simulations show very smooth contours of pressure and velocity, which indicates that you have not used the VOF technique to resolve these thousands of bubbles. So the claims you make in the paper about a VOF model are untrue, and for that reason, the CFD modelling simply cannot be published, because of the ethical constraints that apply to publication.
Several other issues follow from the fact that you are not using a VOF model. Foremost is the fact that you cannot determine total interfacial (bubble) area from the CFD model. But the rate of H2 gas phase transfer to the bubbles depends completely on this area. So, you do not have any way to predict the degassing rate, unless you assume a bubble size. Again, I asked whether you had done that, but this question was not answered in your response. So it was not possible for you to obtain the good agreement you show or the degassing rate, unless it is purely by some strange chance event (perhaps the probably of this happening by chance would be one in a trilliontrillion …). So for a second time, the claims that you are making are untrue, and for that reason, the CFD modelling simply cannot be published, because of the ethical constraints that apply to publication.
I suggest that you collaborate with an expert in CFD of stirred tanks, of which there are several in chemical engineering departments in the Czech Republic. Multiphase CFD is actually rather complex, and requires some skill to carry out meaningfully.
Author Response
Dear Reviewer,
We would like to thank you for your feedback to improve our submission.
As per your recommendation, we have removed the section about numerical modelling from the article.
All the changes made are recorded in the article.
Thank you for your suggestions.
Authors
