A Two-Stage Model for Data-Driven Leakage Detection and Localization in Water Distribution Networks
Round 1
Reviewer 1 Report
The work has a research character. concerns a very important and current issue of water losses in water distribution systems (WDS). The issue also appears in the amendment to the Water Directive. Water companies face the challenge of reducing water losses. The paper proposes a two-stage model based on data analysis to predict leakage incidents and their specific locations in water distribution networks. The adopted methodology is correct. Pressure and flow data collected from multiple points in the network were used. Statistical measures were then applied to these error distributions are used to classify the occurrence and location of leaks. Please indicate the novelty of the proposed approach. The limitations of the model should also be indicated. There is no pattern numbering on page 10. Figures 6-9 are hard to read. I suggest separating diagrams from graphs.
Author Response
Please see the attachment
Author Response File: 
Author Response.pdf
Reviewer 2 Report
The authors present an interesting methodology for leakage detection and localization in water distribution networks. The structure of the paper is correct. Nevertheless, some aspects of the paper should be revised and improved:
- Leakage detection and localization in water distribution networks is a widely studied topic. A research paper must be on the frontier of knowledge. What is the scientific novelty of the work carried out? The introduction should be clear about the novelty of the work presented compared to other works already published. This should be perfectly clear.
- Lines 193-196: “The flow readings for the pair can only partially explain the variation in the difference between pressure heads for the two nodes, as we cannot fully explain the outcomes of the high-dimensional model using a low-dimensional dataset”. What implications does this have?
- Lines 272-273: “The demand standard deviation was assumed to be in between 16% - 20% for all the nodes”. Why? The authors should better explain this assumption.
- Lines 291-292: “The choice of these sensor locations is based on factors such as criticality and distance from the sources”. The authors need to better explain this statement.
- Are the pressure sensor and the flow sensor always together? Why?
- Many figures are too small.
- The methodology has been applied to three "controlled" WDNs. Have the authors applied the methodology to any real WDN? Would it be possible to add the results obtained in a real WDN?
- The proposed methodology is interesting, but the practical application to real WDNs is not clear. The authors should add something about it in the paper.
To conclude, in my opinion, the paper needs some modifications to be published in a prestigious scientific journal. The authors must clarify the scientific novelty of the work carried out. On the other hand, the authors must complete and improve some sections.
Author Response
Please see the attachment
Author Response File: Author Response.pdf
Reviewer 3 Report
Please, see attached file.
Comments for author File: Comments.pdf
Author Response
Please see the attachment
Author Response File: Author Response.pdf
Round 2
Reviewer 2 Report
The authors have adequately responded to all my comments, but modifications in the manuscript are necessary. The authors should add responses to comments in the text of the paper.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
