Unsteady Fluid Flows in the Slab Mold Using Anticlogging Nozzles

Round 1

Reviewer 1 Report

It is a well written contribution. Some remarks are as follows.

The authors claim (lines 95-96) that a SEN should operate without gas injection. This seems unrealistic. Gas injection delays clogging by inclusion attachment as well as avoids negative air leaking (due to venture effect caused by sliding gate partial opening). Thus gas seems unavoidable. The reported 4 hours until clogging looks like a short sequential.

Please check line 137: it should read mold instead of tundish

Please check table 1: the liquid flow rate should be 8.56 liters per second

Starting at line 291 the author report “spent time elapsed from dye dispersion”. This seems a rather crude assessment of velocities since the flow is highly turbulent. Dye dispersion will change according to the flow pattern observed at a given time. Thus it an average time resulting from how many shots? When the flow is double roll? Biased?

Why not PIV measurements? This would provide a much better comparison with numerical simulation. The authors should also make clear the software used in this simulation: home made? Fluent? CFX? Or else?

Please check line 344: it should read most instead of Most.

One misses some experimental details. As an example the level fluctuation graphs as detected by ultrasonic sensors. Their looks will be dependent on rate of capture and precision of the A/D board; also on the sonic sensor features such as precision. Please give details. The same applies to near surface velocities as given by the transducer.

The first part of the appendix, regarding wave height, might be useful in this analysis. It would be especially useful for comparison with the mold level actual fluctuations. The second portion of the appendix, regarding scale factor, could be deleted without loss of substance.

Author Response

Thank you for your observations and comments, we consider that they are a great contribution to the improvement of this article.

1. The authors claim (lines 95-96) that a SEN should operate without gas injection. This seems unrealistic. Gas injection delays clogging by inclusion attachment as well as avoids negative air leaking (due to venture effect caused by sliding gate partial opening). Thus gas seems unavoidable. The reported 4 hours until clogging looks like a short sequential.

 

In the mentioned lines it is not stated that the nozzles must operate without argon, the mention of operating without argon is for the 2 nozzles analyzed in this work only. Regarding the 4 heats, only 4 heats were made to make a comparison of the accumulated clogging in each of them, after this number of heats, nozzle A already had a considerable amount of clogging compared to nozzle B, nozzle B I could have endured one more heat.

 

2. Please check line 137: it should read mold instead of tundish.

 

The sentence was corrected and highlighted in the text so that it can be found faster.

 

 

3. Starting at line 291 the author report “spent time elapsed from dye dispersion”. This seems a rather crude assessment of velocities since the flow is highly turbulent. Dye dispersion will change according to the flow pattern observed at a given time. Thus it an average time resulting from how many shots? When the flow is double roll? Biased?

 

The sentence "spent time elapsed from dye dispersion" is a reference time to give an idea of the acceptable symmetry that exists in the flow using each of the nozzles, under the condition without and with cloggling, in addition to comparing the speed along the meniscus level, the spent time refers to the time that passes from when the tracing leaves the ports, until it touches the external wall of the nozzle, forming the upper roll flow, this measurement was reported only on the left side of the mold

 

4. Why not PIV measurements? This would provide a much better comparison with numerical simulation. The authors should also make clear the software used in this simulation: home made? Fluent? CFX? Or else?

 

The purpose of not doing it with PIV and doing it with the ultrasonic transducer is because we wanted to analyze the flow in the mold from a statistical point of view, obtaining a different analysis of the flow than those previously shown. The numerical simulation was carried out with ANSYS Fluent, in the Numerical Simulation section this information has already been added.

 

5. Please check line 344: it should read most instead of Most.

 

The sentence was corrected and highlighted in the text so that it can be found faster.

 

6. One misses some experimental details. As an example the level fluctuation graphs as detected by ultrasonic sensors. Their looks will be dependent on rate of capture and precision of the A/D board; also on the sonic sensor features such as precision. Please give details. The same applies to near surface velocities as given by the transducer.

 

The accuracy and capture speed of both the sensors and the ultrasonic transducer, records in real time the instantaneous level of the water and the instantaneous speeds, such captures are made second by second during the entire time in which the test is carried out without losing detail in no time. Such a description is given on lines 128 to 136.

 

7. The first part of the appendix, regarding wave height, might be useful in this analysis. It would be especially useful for comparison with the mold level actual fluctuations. The second portion of the appendix, regarding scale factor, could be deleted without loss of substance.

 

The authors consider the second part of the appendix, regardind scale factor, to be important, since it would be the continuation for a new research work on the matter.

Author Response File: Author Response.docx

Reviewer 2 Report

Dear authors, thank you for an interesting article, I hope my comments will help improve it!
1.Abstract
In the abstract, many terms are incomprehensible without reading the full article.
What is "neutralization of the biased flow"?
What is "balanced fluid"?
It is not clear how the loss function "Covariance Loss" is introduced.
I recommend rewriting the abstract so that the physical statement and the main result achieved are clear.
Formulation of the problem.

1. The authors assume the possibility of using the results in pouring metal, but in real conditions, heat transfer processes play a very important role and have a strong influence on hydrodynamics (free convection, temperature dependence of viscosity, phase transition), and there are no estimates for heat transfer factors in the article.

2. The clogging of the nozzle can be of any shape, so this calculation is a kind of special case, it is not clear how to make a generalization based on this work.

Calculation.
1. Equations are given in the paper, but the scheme of the numerical solution is not given. 

2. The situation with modeling the shape of the free surface is not clear. The authors write “no-slipping boundary condition applies to all walls”, but what is the boundary condition on the top surface? If there is also no-slip, then this requires justification.
3. The article mentions the meniscus many times, in such modeling it does not give the shape of the surface, it is required to use, for example VOF model.

Experiment
1. Several ultrasonic sensors are used in the work. How are the specified sensor installation points explained, i.e. why are they located in these places?2. Parameters and model of ADC, characteristics and models of sensors are not specified.

Results
1. The results of comparison of calculation and experiment are not presented. There are only pictures with flow fields.
2. The hydrodynamic pattern of water flow in the calculation hardly corresponds to the pattern of metal flow, since the geometry is quite complex. Nevertheless, estimates of the Reynolds number can be given. Do they match the Reynolds number in a real installation?

Article formatting.

I recommend using numbers instead of words where possible "200" instead of "two hundred" (line 120), "1e7" instead of "ten million", etc...
Use abbreviations: “Hz” instead of “Hertz”
(Line 133-134).
Use times symbol \times (multiplication sign) instead of "x"
See text in tables, for example: Standard Deviation: 8.713?10...

Comments for author File: Comments.docx

Author Response

Thank you for your observations and comments, we consider that they are a great contribution to the improvement of this article.

1.Abstract
In the abstract, many terms are incomprehensible without reading the full article.
What is "neutralization of the biased flow"?

The neutralization of the biased flow refers to the fact that thanks to the internal deflectors of the nozzle, the flow that is loaded only towards one side of the nozzle, the side where the opening of the sliding gate valve is located, when passing through the deflectors so it manages to center the flow inside the nozzle, which is why it is called flow neutralization.

What is "balanced fluid"?

A balanced fluid refers to the fact that without the internal deflectors of the nozzle, there is a biased flow, which means that the flow presents an increase and decrease in speed along the nozzle. With the internal deflectors it is possible to avoid this effect inside the nozzle, achieving a balance in the flow through the ports, as shown in figure 14.

1. The authors assume the possibility of using the results in pouring metal, but in real conditions, heat transfer processes play a very important role and have a strong influence on hydrodynamics (free convection, temperature dependence of viscosity, phase transition), and there are no estimates for heat transfer factors in the article.

Of course, there are still other phenomena to take into account, it is true, but we believe that the results shown are a good principle to take into account in the pouring metal, keeping in mind that in the present work the flux delivered by each of the metals is analyzed. the nozzles in the mold and how this affects the variations of the meniscus and its speed, and considering that it is an area of vital importance in the mold.

2. The clogging of the nozzle can be of any shape, so this calculation is a kind of special case, it is not clear how to make a generalization based on this work.

Yes, of course, the form of clogging can be given in different ways and the form that was analyzed in this work is very specific to the clogging that was formed in the 2 holes analyzed, however, if it can give us a general idea of how the clogging affects the flow patterns in the mold and the effect it has on the meniscus region, which is the area of the mold that was analyzed in most detail.

Calculation.
1. Equations are given in the paper, but the scheme of the numerical solution is not given.

The schematic of the model used for the numerical simulation can be found in the section on numerical simulation. 

2. The situation with modeling the shape of the free surface is not clear. The authors write “no-slipping boundary condition applies to all walls”, but what is the boundary condition on the top surface? If there is also no-slip, then this requires justification.
   3. The article mentions the meniscus many times, in such modeling it does not give the shape of the surface, it is required to use, for example VOF model.

To account for the presence of the upper layer phase, the shear boundary condition is applied at the bath surface. This part has been added in the numerical simulation section.

Experiment
1. Several ultrasonic sensors are used in the work. How are the specified sensor installation points explained, i.e. why are they located in these places?2. Parameters and model of ADC, characteristics and models of sensors are not specified.

The position of the sensors was defined taking into account the covariances of the flow of the two sides of the mold, wanting to have a position near the narrow wall of the mold, another position near the external wall of the nozzle and a position more between. the 2 positions mentioned above, and looking for their symmetrical positions on the other side of the mold. The purpose of these positions is precisely to analyze how the flow behaves, to analyze if when on one side of the mold its height with respect to the lower surface, in the symmetrical position of the other side of the mold how the flow is behaving, and to have then an understanding of how the flow on the 2 sides of the mold is related.

Results
1. The results of comparison of calculation and experiment are not presented. There are only pictures with flow fields.

Figures 7, 8, 17 and 18 certainly show images of the flow field using each of the nozzles with and without clogging, comparing and evidencing that the flow obtained both experimentally and numerically, have a good match, however, in figures 13b and d and figures 22 b and d show the comparison of the speed calculations obtained both experimentally and numerically.

2. 
   The hydrodynamic pattern of water flow in the calculation hardly corresponds to the pattern of metal flow, since the geometry is quite complex. Nevertheless, estimates of the Reynolds number can be given. Do they match the Reynolds number in a real installation?

It could be said that they coincide relatively.

Article formatting.

I recommend using numbers instead of words where possible "200" instead of "two hundred" (line 120), "1e7" instead of "ten million", etc...
Use abbreviations: “Hz” instead of “Hertz”
(Line 133-134).
Use times symbol \times (multiplication sign) instead of "x"
See text in tables, for example: Standard Deviation: 8.713?10...

The observations was corrected and highlighted in the text so that it can be found faster.