A Stator Slot Wedge Loosening Offline Detection System Based on an Intelligent Maintenance Robot of a Large Hydro Generator
Round 1
Reviewer 1 Report
This paper proposes a robot with a percussion device (instead of a hammer) in order to detect loose wedges in the stator of large hydro generators. Signal processing and artificial intelligence are used in order to provide the diagnostic.
The paper is quite well written and I have only some minor suggestions:
1- The title says "online detection" - but the abstract says "offline"(line 10).
2- Several acronyms are not defined. Please, define them the first time they appear in text and also create an "Abbreviations" section at the end of the document (before the references) - please, check the autors instructions.
Some of the acronyms I noticed appearing without definition:
-line 40: EDM;
-line 48: DSP;
-line 263: NTL.
3- First line of the conclusion (lines 458-459). I suppose that line was part of the template and the authors forgot to remove it.
Author Response
Response to Reviewer Comments
We would like to thank you for your careful reading, helpful comments, and constructive suggestions, which has significantly improved the presentation of our manuscript.Based on the instructions provided in your letter, we uploaded the file of the revised manuscript. Accordingly, we have uploaded a copy of the original manuscript with all the changes highlighted by using the track changes mode in MS Word. The comments are reproduced and our responses are given directly afterward in a different color (red). The following are our main modifications:
- The title and abstract have been changed, and the abstract explains the meaning of "offline"
- Add 150-161 lines to compare to other wedge condition monitoring devices
- Figure 4(e) , Figure 6 and lines 212-223 are added to explain how to set the tightness of the wedge
- Highlight the advances to other existing methods are added to the conclusion in the lines 500-514
Point 1: The title says "online detection" - but the abstract says "offline"(line 10).
Response 1: Thank you for pointing out this problem in manuscript. In the title of the article, "online" means that the generator is out of operation but does not need to move the stator and rotor. In the abstract, "offline" also means that there is no need to move the stator and rotor. However, this description is not clear enough and it is easy to be mistaken for the generator being offline but needs to move the stator and rotor or is operating. Therefore, I changed the "online" in the title to "offline" and explained the meaning of "offline" in detail in the abstract, I'm very sorry that I didn't explain it clearly in the article.
Point 2: Several acronyms are not defined. Please, define them the first time they appear in text and also create an "Abbreviations" section at the end of the document (before the references) - please, check the autors instructions.
Some of the acronyms I noticed appearing without definition:
-line 40: EDM;
-line 48: DSP;
-line 263: NTL.
Response 2: We thank the reviewer for pointing out this issue. We have rewritten each abbreviation according to the specification.
- EDM : electrical discharge machining.
- DSP: digital signal processor (DSP)
- Ntl: this is not an abbreviation of a word, but a mark that represents the length of the noise frame length.We have changed NTL to Ntl to avoid misunderstanding
Point 3: First line of the conclusion (lines 458-459). I suppose that line was part of the template and the authors forgot to remove it.
Response 3: Thank you very much for pointing out this problem in the manuscript. We have deleted this sentence. We are very sorry that it was retained in the original submitted last time due to our negligence.
Author Response File: 
Author Response.docx
Reviewer 2 Report
COMMENT TO THE AUTHORS
1.- Experimental setup
In your paper you present only laboratory tests, accomplished in a simulation bench.
It is not clear for me how the loose wedge and tight wedge have been completed?
In my opinion the experimental setup should be better described.
2.- Another question is about the degree of tightness of the wedges.
How have you considered that the wedge is “100%” tight? Or how have you set the limit to check that the wedge is correct?
3.- As far as I understand your intention is to use the robot in a hydro generator on-line (during operation).
Have you considered than in the hydro generator there are some parts of the stator winding that are under pressure (output of the cooling air), but there are some parts with positive pressure (cooling air inlet)?
4.- Sliding tapping device
What is the maximum applied force (radial direction) than the robot can withstand, without take off the stator?
5.- Another important issue is the different noise of the different machines. According to the generator’s speeds and characteristics, different noises are going to be present in the tests. In some cases, the generator noise is going to be stronger than the percussion noise.
Have you tested your algorithm with generator noise?
6.- Comparison to other wedge condition monitoring devices.
It would be convenient to compare the results of your proposed robot to other wedge monitoring devices.
7.- Highlight the advantages to other existing methods
It would be very interesting to highly the advantage of the proposed method to other existing wedge monitoring methods.
Other comments
8.- In order to clarify the figure 1, it should be included the stator and the rotor.
It should be included the two positions of the percussion device.
9.- Page 9 line 229. A letter is missing in the explanation.
10.- Page 19 lines 458-459
I think you should remove these lines.
Author Response
Response to Reviewer Comments
We would like to thank you for your careful reading, helpful comments, and constructive suggestions, which has significantly improved the presentation of our manuscript.Based on the instructions provided in your letter, we uploaded the file of the revised manuscript. Accordingly, we have uploaded a copy of the original manuscript with all the changes highlighted by using the track changes mode in MS Word. The comments are reproduced and our responses are given directly afterward in a different color (red). The following are our main modifications:
- The title and abstract have been changed, and the abstract explains the meaning of "offline"
- Add 150-161 lines to compare to other wedge condition monitoring devices
- Figure 4(e) , Figure 6 and lines 212-223 are added to explain how to set the tightness of the wedge
- Highlight the advances to other existing methods are added to the conclusion in the lines 500-514
Point 1: - Experimental setup
In your paper you present only laboratory tests, accomplished in a simulation bench.
It is not clear for me how the loose wedge and tight wedge have been completed?
In my opinion the experimental setup should be better described.
Response 1: Thank you for the comment. We are sorry we didn't explain clearly before. We added a description according to your opinion. We explained how to set loose wedges and tight wedges in lines 222-226，”As shown in Figure 76, after the simulation test bench is completed, the shape varia-ble of the corrugated plate in stator slot wedge is adjusted by replacing the gaskets of dif-ferent thicknesses, simulating the corresponding three different states of the slot wedge.”, and Figure 6 and lines 209-220 are added to explain how to set the tightness of the wedge.
Point 2: Another question is about the degree of tightness of the wedges.
How have you considered that the wedge is “100%” tight? Or how have you set the limit to check that the wedge is correct?
Response 2: We are very sorry for our negligence of the explanation. We define the maximum thickness gasket as 100% tight, so as long as we use the maximum thickness gasket, it is 100% compact. And we added Figure 6 and lines 209-220 to explain how to set the tightness of the wedge.
Point 3: As far as I understand your intention is to use the robot in a hydro generator on-line (during operation).
Have you considered than in the hydro generator there are some parts of the stator winding that are under pressure (output of the cooling air), but there are some parts with positive pressure (cooling air inlet)?
Response 3 : We are very sorry for not explaining clearly the meaning of “online” and ”offline”. We changed the title to “offline” and add an explanation that we detect the generator without moving the stator and rotor when generator stops running in the abstract.
Point 4: Sliding tapping device
What is the maximum applied force (radial direction) than the robot can withstand, without take off the stator?
Response 4: As shown in the figure below, the angle between the connecting rod 1 below the arrow and the horizontal direction is 30 degrees, the maximum force in the horizontal direction is 35N, and the radial force is equal to 35N multiplied by the cosine 30 degrees, which is about 30N . Therefore, the maximum radial force borne by the robot is about 30N.
Point 5: Another important issue is the different noise of the different machines. According to the generator’s speeds and characteristics, different noises are going to be present in the tests. In some cases, the generator noise is going to be stronger than the percussion noise.
Have you tested your algorithm with generator noise?
Response 5: In line 228, we mentioned that each group tested 100 sets of tests. The 100 sets of tests in each group we use include the excessive noise of the generator when it is operating, so the sound containing generator noise has been tested in the later algorithm
Point 6: Comparison to other wedge condition monitoring devices.
It would be convenient to compare the results of your proposed robot to other wedge monitoring devices.
Response 6: Compared with other wedge-shaped condition monitoring devices, our advantage is that it is written in lines 150-161. The contents are as follows : ” Compared with other wedge-shaped condition monitoring equipment, our ad-vantages are as follows:
- Compared with permanent magnet adsorption type and crawler type robots, the generator maintenance robot based on the track mechanism can detect not only the stator but also the rotor, and does not need to contact the surface of the stator and rotor, which will not affect the detection results of knocking sound.
- Compared with the crawling detection robot, the sliding tapping device moves faster, and multiple sliding tapping devices can be installed on the guide rail at one time, and the position of the knocking device can be adjusted according to different generator models. The surface of multiple stator slot wedges can be in-spected at one time, and the efficiency of detecting the generator can be greatly improved.
Point 7: Highlight the advantages to other existing methods
It would be very interesting to highly the advantage of the proposed method to other existing wedge monitoring methods.
Response 7: We highlight the advantages to other existing methods in lines 500-514. The content is as follows:
- The sliding tapping device is installed on the guide rail and moves faster. The sliding tapping device does not contact the surface of the stator and rotor, and will not affect the detection results of the knocking sound. It can detect not only the stator under offline state, but also the rotor under offline state, with wider applicability.
- Multiple sliding tapping devices can be installed on the guide rail at one time to detect the stator slot wedge faster.
- Compared with manual knocking on the stator slot wedge, the knocking force of our automatic knocking device is very stable, which ensures the stability of the inspection slot wedge state.
- Compared with the detection results of traditional manual identification and detection robot slot wedge, this paper combines BP neural network to classify and identify the tightness of generator slot wedge, which identifies the tightness of slot wedge faster.
Point 8: In order to clarify the figure 1, it should be included the stator and the rotor.
It should be included the two positions of the percussion device.
Response 8: Thank you for this detail. We didn't mark it clearly. We modified Figure 1 and marked the positions of stator and rotor, 2-stator and 3-rotor.
Point 9: Page 9 line 229. A letter is missing in the explanation.
Response 9: We are very sorry that due to our negligence and font display problems, the letters of the formula in this line are not displayed, we have added all the letters in this line.
Point 10: Page 19 lines 458-459
I think you should remove these lines.
Response 10: Thank you very much for pointing out this problem in the manuscript. We have deleted this sentence. We are very sorry that it was retained in the original submitted last time due to our negligence.
Author Response File: Author Response.docx
