The Visual Measurement of Velocity Profile Distribution in Silt Carrying Flow by Using Ultrasound PIV and Iterative Multi-Grid Deformation Technique
Round 1
Reviewer 1 Report
Comments and notes can be found in the attached file.
Comments for author File: 
Comments.pdf
Author Response
Responses to reviewers’ comments:
Reviewer #1:
== General review
The presented article is a continuation of the authors’ research on the visual assessment of flow velocity. The proposed method is innovative and interesting, but the description of the research is unclear in several places. The text needs to be reviewed and edited, e.g.:
Answer: Thanks for your kind comments very much. Your comments are very valuable and helpful for the improvement of our manuscript, which is the visual assessment of flow velocity. I think this topic and method may be very interesting for the readers of this journal. We have followed your kind comments and revised our manuscript. Some related sentences are modified significantly and are marked in red or highlighted in the revision. Thanks a lot.
- Fig. 5;
- upper/lowercase initial letters in section 3.3;
- superscripts in units;
- in figure caption sometimes Figure, sometimes Fig.
An in-depth statistical analysis could significantly improve the presented results.
Answer: The detail of the answers is as follow:
-- I have added some description in Fig. 5. I think what you feel unclear is the meaning of the line arrow in the box. The line arrow presents the direction of the flow in the box. The box is divided from one to four and then to sixteen, which means that the direction of the flow becomes more precise and clearer. We have re-edited the figure and related descriptions in the revision.
-- The upper/lowercase initial letters in section 3.3 are revised accordingly.
-- The superscripts of word “cm2” are changed as “cm2” and rechecked the whole paper. After that, we changed accordingly in the following part.
-- The word “Fig.” is changed as “Figure” in the whole revision version. Sorry for that.
-- The statistical analysis part has been significantly improved accordingly, and some results are discussed in-depth. We hope you can understand it. Thanks!
== Detailed comments
[195-196] A concise comparison of many advanced image interpolators used for PIV image
deformation schemes can be found in Reference. Specific literature items should be indicated.
Answer: Specific literature items have been indicated in the revision “… in Reference [4,17,21,31,33]”. Thanks for your kind comments very much.
[203] However, the computational cost may be increased up to a more order of magnitude.
Unclear and incorrect sentence.
Answer: This sentence has changed as “However, the computational cost may be increased much if unsuitable approach is taken accordingly to Reference [35]”. The principle of the sentence you can see the Reference [Z. Ma, W. Hu, X. Zhao, W. Tao, A PTV method based on ultrasound imaging and feature tracking in a low-concentration sediment-laden flow, Measurement Science and Technology, 29 (2018) 025303.]
[248-249] (...) the sweep velocity becomes more approach to the next scan line.
What does it mean?
Answer: This sentence has changed as “Therefore, this sweep velocity can be corrected according to the frame rate and the lateral size of the view field according to the Doppler effect of flow velocity”. We measured velocity can be revised if it is affected by the Doppler effect. Thanks for your kind comments very much.
[270] the one-quarter rule.
Missing explanation what is one-quarter rule.
Answer: We have changed this sentence as “… the one-quarter rule which means than the box is divided from one to four.”. Sorry for making you confused.
[271-273] Step 3: Replace incorrect data of the velocity vector field using local mean and Lowpass filtering methods. A filter kernel with the same window size is enough to smooth spurious emission and then suppress fluctuations by using moving average filter.
How “incorrect data” are defined and determined? This is unclear from the description.
Why “Low-pass filtering” in uppercase?
Presented workflow is difficult to follow without some graphical aid (scheme?).
Answer: The “incorrect data” is the value which is out of the normal value because of incorrect calculation or without match, which is usually appeared if the flow field is changed in real time. The word “Low-pass filtering” has been changed as “low-pass filtering” because of wrong writing. The detail of the workflow is described step-by-step. I think the meaning is clear and easy to follow for you this time. Thanks for your kind comments very much.
[310-319]
There is no measurement sensitivity analysis - where did 3.5 MHz come from? Up to what depth were the measurements reliable?
Answer: The measurement of the ultrasound PIV measurement system is in real time. I don’t know how to go on the sensitivity analysis. Sorry for that. The frequency of 1 MHz, 2 MHz, 3.5 MHz, and 5 MHz ultrasonic wave is decided by the probe of the system, which cannot be changed. The reliable depth of the measurement is about from 5 cm to 35 cm. You can refer to Table 1. We hope you can understand that actual results can be affected by complex measuring conditions. We have added related description in the revision. Thank you.
[320]
What is region 35x35 cm2?
Maybe region 35 cm x 35 cm will be easier to read and understand?
Answer: OK. We accepted your suggestion and changed accordingly. Thanks.
[327 vs. Table 1]
32x32 pixel/cm2 vs. 35*35 cm2
Answer: They are different contents. 32x32 pixel/cm2 is referred to spatial resolution of the ultrasonic image. 35*35 cm2 [35 cm * 35 cm] is referred to the monitored region of the ultrasound PIV measurement system. They are different concepts.
[328 vs. Table 1]
frames/sec vs. frame s-1
Units should be standardized and used consistently throughout the document.
Answer: Sorry for that. Units have been standardized and used consistently throughout the document. Thanks for your kind comments very much.
[332-342]
This description of used custom software is so enigmatic. No information about the license, website, computer language, requirements, operating systems, etc.
Answer: I don’t what the exactly meaning of you. Thanks for your kind comments very much, but we have revised this part again as “The processing software includes some PIV codes that can be referred to as the built-in soft-ware of the ultrasound imaging device and the program of the PIV method on the website. The software should include the following main features: (1) it can link a personal computer to the ultrasound imaging device directly by using a high-speed network card; (2) it can real-time display the ultrasound imaging results of silt carrying flow within sand particles in the personal computer; (3) it can real-time transform and process the image video by editing the interface program; (4) it can calculate the vector map and analyze flow field and depth-averaged velocity distribution by using iterative multi-grid deformation technique and related ultrasound PIV method. The program of our used ultrasound imaging measurement system is a special software for ourselves. The experiments about the re-al-time measurement of velocity profile distribution in silt carrying flow are carried out on the physical model system shown in Fig 7.”.
[364-365] The measured velocities of ADCP method are stable and accurate.
How stable and accurate? Such sentences should be based on references or data validation.
Answer: Thanks for your kind comments very much. We have changed this sentence as “The measured velocities of ADCP method can be taken as the standard velocity and then compared with the results from our ultrasound PIV method.”. Such sentences of the whole paper have been rechecked again. Thanks again.
[366-368] This is the best way to prove that the velocity profile vectors obtained by our ultrasound
PIV method are acceptable.
How can you prove this is the best method?
What is the “acceptance” level?
Answer: Sorry for that. This sentence is wrong expression and we have changed this sentence as “This is a suitable way to prove the velocity profile vectors obtained by our ultrasound PIV method.” We apology for this unprofessional statement. Sorry again.
[373] We took the follow measures to calculate the flow field ...Should be
We took the following measures to calculate the flow field ...
Answer: Thanks for your kind comments very much. We have changed this sentence as “We took the following measures to calculate the flow field….”
[373-379]
How this description is related to section 3.3?
Answer: Thanks for your kind comment. This is a short description and really related to section 3.3 which is the detail description. We have deleted this part in order to reduce repeated descriptions.
[395-398]
Why is this sentence written in a different mode than the others?
Answer: We have changed this part as “To estimate the accuracy of the measured profile velocities, firstly, a statistical value is gotten at each sampling position and taken as a true value. Secondly, the last obtained profile velocities are compared with the standard velocities, which are obtained by the ADCP method. Lastly, the velocity profile distribution...”. Thanks for your kind comments very much.
[405-408] is towards lower …. has a better agreement … has the worst performance
This is where the statistics should come in.
Answer: We have changed this part as “From Table 2 and Figure 10, we concluded that the velocity profile obtained by the WT64 is smaller than the actual velocity (ADCP). The velocity profile obtained by the WDI64-32 is nearly agreement with actual velocity. The stability of the velocity profile obtained by the WT16 is poor. In contrast, the velocity profile obtained by the WT16 has the worst among them.”. Thanks for your careful proofread very much.
Table 2. The profile velocities or The water velocities in profiles?
Answer: They are the measured profile velocities in Table 2.
Fig.10 Why are the axis labels inside the chart? This is very unusual. How to interpret fact that on average WT16 water velocity estimation is much better than WT64 for depths below 10 cm? This is not visible on Fig.11.
Answer: The axis labels are moved outside the chart. The reason why the fact exists is because the disturbances of surface flow and wave. The average velocity estimation of WT64 is much better than WT16 for depths below 10 cm is because the WT64’ window templates is 64×64 and bigger than WT16. The average effect can be better if the window template is bigger. Thanks for your kind comments very much.
[420] actual velocity of depth y
Shouldn’t be actual velocity at depth y
Answer: Thanks for your careful proofread. We have changed related sentences.
[422] Use formulas (7) and (8) to obtain formulas (9) and (10).
This sentence means little. It is better to explain the meaning of the formulas.
Answer: Thanks for your kind comments. We have deleted this sentence and explained the meaning of the formulas. They have been changed as “The formulas (9) and (10) are obtained according to formulas (7) and (8), which can explain the effect of flow field.”. Thanks again.
[458-459] The flow velocity gradient across the window may exist some errors because not all the
particle images ... Something is wrong with this sentence.
Answer: This sentence has been changed as “There are some errors if the flow velocity gradient across the window because not all the particle images appear in the first interrogation window and then appear in the next interrogation window again”. Thanks for your careful proofread very much.
Table 3 and Fig. 11.
Please choose just one of them. What is the unit of this error? This is not obvious for the reader.
Answer: We deleted Table 3 accordingly. The unit of this error is %, which is shown in the axis label of Figure 11. We hope it is clear this time. Thanks a lot.
[489-490] As is shown in the curves of the WT64, the WT32, and the WT16 in Fig.12, the relative RMS error increases with the window size decrease.
In my opinion this isn’t true if you exclude depths 0-10 cm.
Answer: Thanks for your kind comments very much. Your opinion may be right when it excludes depths 0-10 cm for the measurement instantaneous velocity by using normal methods. However, the aim of this paper is the measurement of average velocity for the profile distribution along with the depths. The measuring accuracy becomes higher if the computing image is larger while the relative RMS error will decrease in theory for the average value. I think it can be accepted in some certain conditions and we still changed accordingly in order to reduce misunderstanding. Thanks a lot.
Two lines captions are missing in Fig.12 and this figure as the only one has black background.
Answer: The Figure 12 is re-edited and the missing two lines captions are added, and its background is changed as white. Thanks for your kind comments and careful proofread very much.
Reviewer 2 Report
See attachment
Comments for author File: Comments.pdf
Author Response
Responses to reviewers’ comments:
Reviewer #2:
The work "The visual measurement of velocity profile distribution in silt carrying flow by using ultrasound PIV and iterative multi-grid deformation technique" describes a velocity measuring method, based on the ultra-sound technique and on an iterative multi-grid deformation particle image velocimetry (PIV). The proposed method is employed to silt carrying flows and its accuracy is evaluated by comparing the measurements with the ones obtained by acoustic Doppler current profiler. The proposed approach is cost-effective and has certain advantages over other measuring techniques, when it comes to estimate the velocity field in non-transparent flows such as silt carrying flows. I signal that the manuscript is written in rather poor English and lacks some important details. I think that the work could be recommended for publication in “Applied Sciences”, provided that the manuscript is revised according to the following remarks.
Answer: Thanks for your recommendation very much. This manuscript has some advantages in the real-time measurement of velocity profile distribution in silt carrying flow and the measurement efficiency of field velocity vector. Your kind comments are very valuable and helpful for the improvement of our manuscript. We have followed your comments and revised the whole manuscript. Some related sentences are modified significantly and are remarked in red or highlighted in the revision. Lastly, the revision is uploaded to a professional software “Grammarly” for language rechecking. We think all the criticisms and suggestions are answered and revised carefully. We hope these can make your approval. Thank you again.
- Line 14 and elsewhere in the manuscript: could the Authors better explain in which sense the velocity measurements can be regarded as instantaneous? The ultrasound image is typically obtained over a finite time interval, due to the finite sweep velocity. Moreover in Sec. 5 the Authors made time averages of the velocity profiles. Please provide some evidence that the proposed measuring method is also able to reliable estimate the instantaneous velocities, or revise the text without claiming something that is unproven.
Answer: Thanks for your kind comments very much. We totally accept your suggestion and revise the text about the word “instantaneous velocity”, which may need more evidences. Sorry for our loose statements. We have deleted the word “instantaneous” through the manuscript and revised accordingly. Thanks a lot.
- Could the Authors better define a “silt carrying flow” in terms of the solid volume fraction? What is the range of the solid volume fraction, for which the proposed measuring method is expected to work well?
Answer: The silt carrying flow refers to a sediment-laden flow with a suspended sediment concentration from 0.0001 % to 0.5 % in terms of the solid volume fraction during the physical model test [12, 14]. The proposed measuring method can both work well in the range of the solid volume fraction. More information you can refer to the References [12, 14]. Thanks for your kind comments very much.
- About the image deformation (section 3.1 - point 1): a first image deformation is necessary to convert the arc-annular shaped image into a standard rectangular image. A further image deformation is related to the image-shifting technique implemented in the PIV code. So as I understand, there are two different image deformations procedures. They should be explained/differentiated more clearly in the manuscript to avoid ambiguities.
Answer: Thanks for your understanding. There are two different images obtained from the system, but just one kind of image (obtained from the convex array probe shown in Figure 3) needs image deformation. Therefore, there is only one image procedure in practice, that is, the image deformation for the image obtained from the convex array probe in section 3.2. Thanks a lot.
- Section 3.3 is tedious. It should be shortened without loss of detail.
Answer: Sorry for that. The concern may be right. We have revised this part according to the comments of both the Reviewer 1 and Reviewer 2. Your suggestions should have been both taken. Thanks a lot.
- Lines 414-440: the discussion reported in Sec. 5, which uses the law of the wall, is unclear. Moreover, a logarithmic law of the wall holds for the turbulent regime, not for a laminar flow. Please better explain or, since this short discussion seems not very useful for the purpose of the paper, it might be deleted in my opinion.
Answer: Firstly, thanks for your careful proofread very much. Your suggestion can be acceptable. We also think this part more than one time, but we still want to remain it because this part can explain the reason of the measurement error in theory, I think, and can improve the discussion part even though this part is small. Additionally, this part has been revised according to. I think the quality of this part has been improved and can be accepted as publication. Thanks for your kind comments very much.
- Section 4: the Authors stated that the employed PIV code is a “Special UPIV for River Model Test”. Would it be possible to cite some reference about this PIV software? Would this PIV code be available to other researchers, if requested to the Authors? Moreover, would it be possible to provide some information about the economical cost of the employed ultrasound device, SIUI APOGEE 1200?
Answer: Thanks for your kind comments very much. Your concern can be understood, I think. We have changed this part in the revision and will provide some codes or some information on website for the Authors who contact us in private. You can contact me through the Corresponding E-mail and I will give some codes for help if you need them. We hope you can understand that and thanks a lot.
- What is the expected accuracy of the reference measurements by ADPC?
Answer: The expected accuracy of the reference measurements by ADPC is about 0.3% accordingly many published articles on website, and there are some different between different ADCP devices for different measuring flow conditions. Thank you.
- Fig. 12 should be revised according with the same style of the other figures. Moreover, the legend does not report the meaning of some color lines reported in the plot.
Answer: Thanks for your careful proofread very much. The Figure 12 has been re-edited accordingly. The background and style are changed as the same style as the other figures. The missing axis labels are added and explained. Thanks a lot.
- In the Conclusion (line 525), the sentence "The velocity vectors obtained by the iterative multi-grid deformation algorithm have a good agreement with standard values" should be revised in order to include some quantitative details (in particular it could be useful to shortly recall the RMS errors already discussed in the manuscript).
Answer: Thanks for your kind comments very much. This sentence has been changed as “The velocity vectors obtained by using the WDI64-32 algorithm is the most agreement with standard values (ADCP algorithm) when compared with the WT64 algorithm, the WT32 algorithm, and the WT16 algorithm. Their RMS errors are 0.148, 0.4214, 0.8796, and 1.4148 at the certain condition, respectively”.
- The English should be substantially improved, including in the abstract. I suggest a professional translation service.
Answer: I am sorry to make you feel confuse because of my poor English, and thanks for your big patience for proof-reading this paper this time. We have revised the abstract and asked for help from others. Lastly, the whole revised manuscript has been proof-read and uploaded to a professional software “Grammarly” for language rechecking. We think all the criticisms and suggestions are answered and revised carefully. Thank you again.
Author Response File: 
Author Response.docx
