Research on Application of Convolutional Gated Recurrent Unit Combined with Attention Mechanism in Water Supply Pipeline Leakage Identification and Location Method
Hoi Yeung
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsI have reviewed the article entitled: "Research on Application of Convolutional Gated Recurrent Unit Combined with Attention Mechanism in Water Supply Pipeline Leakage Identification and Localisation Method". Its main contributions consist in the application of a new method for the localisation of leaks in water supply systems. This method is based on a novel convolutional gated recurrent unit method with attention mechanism to improve leakage identification in water supply pipelines. It combines CNNs for local spatio-temporal features, GRUs for long-term dependencies, and attention to highlight key features. The method is tested on an experimental platform and shows interesting good results.
I encourage the authors to consider or discuss the following issues:
Line 40: Given the difference between European billions and American billions, perhaps expressing the value "599.82 billion m3" in Hm3 might avoid confusion for readers.
Line 57: There are four digits after the comma. I think there is an error in the position of this comma. Nevertheless, I suggest again to express this quantity in Hm3.
Line 138: The authors say ".... produces a large negative pressure region,..." perhaps the authors mean "large negative pressure wave region" because the (relative) pressures at the leak point are not negative. In fact, if they were negative (sub-atmospheric pressure), the water would not flow out of but air would enter the pipe.
Line 142: Similarly as my last comment, I suggest to say something like "... manifested as a small negative pressure drop." or similar.
Line 172: In relation to the 1000-1200m/s wave velocity range, could the authors give some more information on the factors that could affect this propagation velocity? (degree of confinement, pipe material and thickness, etc.)
Section 3.1: Please specify the pipe diameter and thickness in the test platform description.
Finally, my main concern in relation to the work presented and the good results obtained on the experimental platform is based on the real application of the method presented. I understand that this application in a real water distribution system requires the installation of pressure sensors capable of detecting the pressure drops and negative wave fronts that will cause the leaks, and I wonder:
How far apart should these sensors be installed in a real system? I suspect they should be installed at distances greater than those used in the platform, so how would increasing this distance (compared to the results shown in the paper) affect the results?
Can the authors please comment on this?
Author Response
For research article
|
Response to Reviewer 1 Comments
|
||
|
1. Summary |
|
|
|
Thank you very much for taking the time to review this manuscript. Please find the detailed responses below and the corresponding changes in the re-submitted files. |
||
|
2. Questions for General Evaluation |
Reviewer’s Evaluation |
Response and Revisions |
|
Does the introduction provide sufficient background and include all relevant references? |
Yes/Can be improved/Must be improved/Not applicable |
|
|
Is the research design appropriate? |
Yes/Can be improved/Must be improved/Not applicable |
|
|
Are the methods adequately described? |
Yes/Can be improved/Must be improved/Not applicable |
|
|
Are the results clearly presented? |
Yes/Can be improved/Must be improved/Not applicable |
|
|
Are the conclusions supported by the results? |
Yes/Can be improved/Must be improved/Not applicable |
|
|
3. Point-by-point response to Comments and Suggestions for Authors |
||
|
Comments 1: Line 40: Given the difference between European billions and American billions, perhaps expressing the value "599.82 billion m3" in Hm3 might avoid confusion for readers. |
||
|
Response 1: [Agree] Thank you for pointing this out. To avoid ambiguity, based on your suggestion, we have changed the unit to Hm3, and the revised part is shown in Line 40.
Comments 2: Line 57: There are four digits after the comma. I think there is an error in the position of this comma. Nevertheless, I suggest again to express this quantity in Hm3. Response 2: [Agree] Thank you for pointing this out. We have reviewed the latest released data, revised the total water supply in China for 2022, and made modifications to the units. These modifications are reflected in Line 57 of the revised manuscript.
Comments 3: Line 138: The authors say ".... produces a large negative pressure region,..." perhaps the authors mean "large negative pressure wave region" because the (relative) pressures at the leak point are not negative. In fact, if they were negative (sub-atmospheric pressure), the water would not flow out of but air would enter the pipe. Response 3: [Agree] Thank you for pointing this out. Our previous description was not rigorous enough. Thank you for pointing out this deficiency. We have made modifications to the negative pressure wave area. This change can be found in the revised manuscript in Line 138.
Comments 4: Line 142: Similarly as my last comment, I suggest to say something like "... manifested as a small negative pressure drop." or similar. Response 4: [Agree] Thank you for pointing this out. Based on your suggestion, we have adjusted and modified the expression. The change can be found in lines 142 to 143 of the revised manuscript.
Comments 5: Line 172: In relation to the 1000-1200m/s wave velocity range, could the authors give some more information on the factors that could affect this propagation velocity? (degree of confinement, pipe material and thickness, etc.) Response 5: [Agree] Thank you for pointing this out. I agree with this comment. We have added some factors that affect the propagation speed of negative pressure waves, such as pipeline materials, fluid density, etc. These changes can be found in lines 172-190 of the revised manuscript.
Comments 6: Section 3.1: Please specify the pipe diameter and thickness in the test platform description. Response 6: [Agree] Thank you for pointing this out. We supplemented the values of pipeline diameter and wall thickness in lines 345-346.
Comments 7: How far apart should these sensors be installed in a real system? I suspect they should be installed at distances greater than those used in the platform, so how would increasing this distance (compared to the results shown in the paper) affect the results? |
||
|
Response 7: [Agree] Thank you for pointing this out. The question you raised is very valuable. Actually, we have installed sensors on a real pipeline in Shanghai at present. Through experiments, we found that sensors located about 5 kilometers away from the leak point can also capture negative pressure drop signals well. However, due to limitations in experimental conditions, we have not yet calculated and compared the effective distance of sensor installation. But as research progresses, we will also focus on this important issue.
|
||
|
4. Response to Comments on the Quality of English Language |
||
|
Point 1: The English is fine and does not require any improvement. |
||
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for AuthorsAfter overcame to 'shock' of the paper was wrongly titled ('Localisation' should be 'Location'), the reviewer did enjoy the paper. I have to admit that I am not an expert of different neural network NN software but my previous students did use NN to identify flow regimes in oil and gas transfer pipelines. Thus I did not go into the detailed descriptions in the paper. I also instigated many engineering projects to reduce leakage from water distribution network in the UK many years ago and lectured on pressure transients. The authors should find out the state of technologies in the UK.
For detailed comments/suggestions, please refer to the attached paper.
A number of points that the authors need to take note
- it is not clear from the paper where the flow and pressures were measured. Improvement to the schematic of the test set up and operation are needed. Is there a flow source? How the pressure in the main is maintained.
- Figure 10 needs more explanation. Why plot pressure vs flow? Same as Figure 2
- Why 1000l/s was used instead of 2000l/s. What implications on results?
Comments for author File: Comments.pdf
Do check the use of English!
Author Response
For research article
|
Response to Reviewer 2 Comments
|
||
|
1. Summary |
|
|
|
Thank you very much for taking the time to review this manuscript. Please find the detailed responses below and the corresponding changes in the re-submitted files. |
||
|
2. Questions for General Evaluation |
Reviewer’s Evaluation |
Response and Revisions |
|
Does the introduction provide sufficient background and include all relevant references? |
Yes/Can be improved/Must be improved/Not applicable |
|
|
Is the research design appropriate? |
Yes/Can be improved/Must be improved/Not applicable |
|
|
Are the methods adequately described? |
Yes/Can be improved/Must be improved/Not applicable |
|
|
Are the results clearly presented? |
Yes/Can be improved/Must be improved/Not applicable |
|
|
Are the conclusions supported by the results? |
Yes/Can be improved/Must be improved/Not applicable |
|
|
3. Point-by-point response to Comments and Suggestions for Authors |
||
|
Comments 1: After overcame to 'shock' of the paper was wrongly titled ('Localisation' should be 'Location'), the reviewer did enjoy the paper. I have to admit that I am not an expert of different neural network NN software but my previous students did use NN to identify flow regimes in oil and gas transfer pipelines. Thus I did not go into the detailed descriptions in the paper. I also instigated many engineering projects to reduce leakage from water distribution network in the UK many years ago and lectured on pressure transients. The authors should find out the state of technologies in the UK. |
||
|
Response 1: [Agree] Thank you for pointing this out. Sorry for the inconvenience caused due to our incorrect use of English words. We have made modifications to the article title. We learned that you have instigated many engineering projects to reduce leakage in the UK water distribution network and lectured on pressure transients. We are honored to receive your review comments. We also admire the work you and your research team have done. Next, we will further understand and learn about the development of leakage technology in water distribution networks in the UK.
Comments 2: it is not clear from the paper where the flow and pressures were measured. Improvement to the schematic of the test set up and operation are needed. Is there a flow source? How the pressure in the main is maintained. Response 2: [Agree] Thank you for pointing this out. The experimental pipeline is supplied with water from an external water pipe, with a pressure of 0.25MPa~0.35MPa. Due to interference from other water equipment in the entire laboratory, the water pressure inside the test pipeline fluctuates greatly, with a maximum fluctuation amplitude of 0.15 MPa. Therefore, we stabilized the pressure in the experimental pipeline by installing a pressure reducing valve. The pressure reducing valve can stabilize the pressure within the range of 0.15MPa~0.25MPa, avoiding the impact of water supply pressure transients on the testing equipment. Based on your comments and suggestions, in order to make our experimental introduction more complete, we have added Figure 9 (a) to illustrate the water source and pressure adjustment and measurement methods of the experiment. The corresponding textual description can be found on lines 349-359 of the revised manuscript.
Comments 3: Figure 10 needs more explanation. Why plot pressure vs flow? Same as Figure 2. Response 3: [Agree] Thank you for pointing this out. When a pipeline leaks, a sudden decrease in liquid flow near the leak point will create a large negative pressure wave area, where liquid molecules transfer energy through collisions and vibrations, forming a wavefront area of NPW, manifested as a sharp drop in pressure. In the area far away from the leakage point, liquid molecules begin to fill the space where the flow decreases, forming the back wave region of NPW, manifested as a small negative pressure drop. This pressure drop will propagate at a relatively stable speed along both ends of the pipeline like waves over time. According to the principle of negative pressure waves, there are two key factors that affect the accuracy of leak point positioning: time difference and wave velocity. This paper focuses on the time difference of negative pressure waves collected by two sensors, without considering changes in flow. Your suggestions are very important, and we will explore the impact of flow in future research. In order to make the presentation of the paper clearer, we have supplemented the descriptions of Figure 2 and Figure 10. Please refer to lines 142 to 149 and lines 398 to 407 of the revised manuscript for details.
Comments 4: Why 1000l/s was used instead of 2000l/s. What implications on results? Response 4: [Agree] Thank you for pointing this out. There are two ways to obtain the value of negative pressure wave velocity. The first is to calculate it using analytical expressions based on pipeline conditions and other factors. Additionally, experience values have always been used. In general, the NPW wave velocity range of urban water supply pipelines is 1000~1200m/s. According to the principle of negative pressure waves, the influence of wave velocity on the accuracy of leak point positioning is also crucial, which is also an important aspect of our research work in the next step. For the negative pressure wave velocity, this manuscript takes an empirical value of 1000 m/s. We have added some factors that affect the propagation speed of negative pressure waves, such as pipeline materials, fluid density, etc. These changes can be found in the revised manuscript in Line 172-190.
|
||
|
|
||
|
4. Response to Comments on the Quality of English Language |
||
|
Point 1: Do check the use of English! Response: We have carefully checked and revised the English version of the entire text. |
||
Author Response File: Author Response.pdf
