Analysis of the Flow Field at the Tip of an Axial Flow Compressor during Rotating Stall Process Based on the POD Method
Round 1
Reviewer 1 Report
In this paper, the authors use the POD method (proper orthogonal decomposition) to decompose the flow field related to the rotating stall process of an unsteady numerical simulation of a transonic axial flow compressor. The mode characteristics of three stall stages have been analyzed.
First of all, the mathematical method used to setup the numerical method, namely the POD method, presents some errors. For example, each row of the matrix U is a snapshot at a given time, so there are M snapshots. The authors instead say "N moments" (not M). On the contrary, the number of elements in each snapshot is N, not M.
In the definition of the matrix C the factor 1/(M-1) is missing.
In equation (3) it is not clear if A is a matrix or a vector. Also subsequent equations are not clearly defined.
In the following of the paper there is no reference to this paragraph, maybe they use these concepts, but it is hidden from the reader.
The results are shown only in the form of graphics and there is no chance for the reader to verify what the author claim.
Author Response
Please see the attachment.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
In this report the authors presented chosen properties of the stall-related structures, not to be observed by the standard analysis ways but using the proper orthogonal decomposition (POD) method are understood. It is established that the stall cells are formed and developed, the flow conditions are gradually constituted by the low order modes, related to the properties of the stall cells and not obvious in the original flow field.
I have no fundamental objections to this interesting paper. It was written in accordance with the accepted rules of art.
After correcting minor errors, the paper will be acceptable.
Strong
Particularly good graphic layer. All drawings are clear and very carefully done.
Weak
Outdated literature is a significant drawback. Only 3 cited sources are from the last 2 years
Noticed errors/remarks
• All charts lack a space between the axis name and the unit, for example Figure 2. Is: Grid numer(million) ; Should be: Grid numer (million). This should be corrected throughout the paper.
• Links to equations should be standardized.
• The paper lacks conclusions for further research. Adding this would certainly increase the value of the work.
Small errors
• All variables (in the text and equation) must be italic.
• Line 35. Is: Yamada et.al.[5], should be: Yamada et.al. [5]; this should be corrected throughout the paper.
• Line 45. Is: „non-linearity [8] .Hah et.al.” ; Should be: „non-linearity [8]. Hah et.al.”
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 3 Report
The authors presented an analysis of the flow field at the tip of an axial flow compressor during rotating stall process based on the POD method.
The main findings are to be presented in the abstract.
The novelty of the work is to be clearly stated.
The used turbulence model is to be justified.
The boundary conditions are to be expressed mathematically.
More details about the numerical method are to be provided.
What is the convergence criterion?
What is the time needed for the convergence ?
What are the properties of the used computer.
In figure 4, how can you explain the differences between the numerical and experimental results?
The conclusion is to be shortened
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 4 Report
The paper explores an important and complex topic - separated flows in transonic flows for complex case (compressor). The paper provides a good introduction, a detailed description of the cfd methods and a good description of the data.
But the reviewer had difficulty understanding the method of calculating POD data. Firstly what variables are used in the calculation of POD? Only pressure? Or is a set of variables used (velocity, pressure, temperature)? Secondly, how was the time dependence of the POD mode amplitudes found (Figure 11b, Figure13b etc)? From the point of view of the reviewer, to obtain these data, one must use a singular value decomposition (SVD). In this case, in fact, the SPOD method is used. At the same time, based on the text of the chapter "discussion", the authors of the work did not use the SPOD method. The reviewer is confused about the method of decomposition of the calculated data, which is used in the paper. It can be seen that the authors of the paper have already tried to clarify this point (based on the selected sections of the text). But explanations on the POD method data processing are still insufficient and need to be expanded.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
I see no substantial improvement compared to the previous version, so I confirm my opinion.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 3 Report
Accept as it is.
Author Response
Point: Accept as it is.
Response: Thank you so much for accepting my paper!
Reviewer 4 Report
Thanks for the answers to my questions. I have no objection to publishing the paper.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Round 3
Reviewer 1 Report
Still don't see any progress
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
