Smart Water Infrastructures Laboratory: Reconfigurable Test-Beds for Research in Water Infrastructures Management

Round 1

Reviewer 1 Report

The paper deals with the facility laboratory research of Albalborg University.

The aim of the Laboratory research is to describe and test of desired networks.

The paper focus interest on validating each designed solution before the deployment in a real network.

The paper is well constructed and both the introduction and the other parts of the work are well proposed.

The authors should clarify only few aspects:

-    How they can define the effective value of input data in a real case? We suggest to carry on, in the future, other experiments by using an approach that can take into account the calibration aspects in assuming the parameters and to estimate technical value for the model.

-    How it is possible to define the number of measures and to choose the positon of loggers to compare the behaviour of the model and the real data acquired?

- a brief analysis of the costs of the proposed approach.

Very good work

I propose to consider these paper to afford the problem to define the appropriate value of input data and to guarantee a best accord between real data and obtained results. The problem, in my opinion is to know the real network. The Laboratory studies can give good and reliable performances only if we use reliable data.

Wu Z.Y. & Sage P., 2007, “Pressure Dependent Demand Optimisation for Leakage Detection in Water Distribution Systems”, in Water Management Challenges in Global Change, (Ulanicki et al. editors), Taylor & Francis Group, London, ISBN 978-0-415-45415-5, pp.353-361.

Identification of Leakages by calibration of WDS Model. DOI:10.1016/j.proeng.2014.02.072. pp.660-667. In PROCEDIA ENGINEERING - ISSN:1877-7058 vol. 70. Fiorini Morosini, Attilio; Costanzo, F; Veltri, Paolo; Savic, D. A.

Identification of Measurement Points for Calibration of Water Distribution Network Models. DOI:10.1016/j.proeng.2014.11.496. pp.693-701. In PROCEDIA ENGINEERING - ISSN:1877-7058 vol. 89 Fiorini Morosini, Attilio; Costanzo, F; Veltri, Paolo; Savic, D.

Stephens M., Misiunas D., Lambert M., Simpson A., Vitkovsky J. & Nixon J., 2005, “Field Verification of a Continuous Transient Monitoring System for Burst Detection in Water Distribution Systems.” CCWI 2005 International Conference, Exeter University Press, Exeter, Vol.2, pp. 257-262.

Kapelan, Z., Savic, D. A., Walters, G. A., Optimal sampling design methodologies for water distribution model calibration, Journal of Hydraulic Engineering 131 (3), pp. 190–200, 2005

Author Response

We would like to thank the reviewer for spending time to help us improve our manuscript. The significant corrections are highlighted in the manuscript in blue, the reviewers’ comments are written in italic and ours are written in plain text. We have given point-to-point responses to the reviewer comments where we have indicated below how we have addressed them:

1. How they can define the effective value of input data in a real case? We suggest to carry on, in the future, other experiments by using an approach that can take into account the calibration aspects in assuming the parameters and to estimate technical value for the model.
   1. The test-bed model describes the qualitative properties of the network. The calibration with respect to the sizing of a real network is not addressed in this paper. After your remark, we decided to work with calibration methods in the future.
2. How it is possible to define the number of measures and to choose the positon of loggers to compare the behaviour of the model and the real data acquired?
   1. In the laboratory there is redundancy of measurements (loggers). With a vast number of sensors it is possible to choose a subset that matches the configuration of a real system. We have reformulated part of the result section to emphasize the sensor selection. (See lines 373-377).
3. a brief analysis of the costs of the proposed approach.
   1. We have added a sentence discussing the resources needed to operate the laboratory. (See lines 504-508).
4. Recommended literature:
   1. We believe that some of the studies in the recommended literature are related with our manuscript’s domain, this kind of studies could benefit from the laboratory use. We have included some of these references for leakage detection in the introduction. 

Reviewer 2 Report

The implementation of a reconfigurable laboratory to represent water distribution networks and wastewater collection systems and to test different control strategies is interesting. It could be a valuable tool to be exploited to improve the operation of these systems. However, the paper needs to be reorganized and rewritten, it is necessary to define clear research objectives.

Revise text between lines 21-28, sentence’s structure is confusing.

Research objectives should be clearly defined in the introduction section. Disperse ideas that justify the implementation of the test modules and the possibility of evaluate different control strategies are presented but, a clear statement of the objectives of the paper is necessary.

What are the aspects of the test-beds implementation that are addressed in the paper: How has been the system designed and implemented? The implementation of control strategies? Possibility of reconfiguration to adapt to different systems? Impact of system behavior that it is difficult to represent in simulation environments?  All these aspects had been mentioned, but none of them is appropriately addressed in the paper.

In materials and methods section model, components description and implementation issues are mixed. It is confusing for readers. How demand profiles recreated by real data are implemented in experimental system (lines 123-125)?

 It should be separated experimental objectives as the indicated between lines 115 and 119, and system representation and configuration. WDN? WWC? Expl98

What happen with the accumulation term in equations 2, 3?

Figure 6 and figure 8, describe how modules are integrated to represent the systems presented in figure 2? An explanation is necessary to understand how the systems presented in figure 2 are represented by test beds modules.

In the results section, the selection of the case studies and system representation should be explained and justified. Which characteristics and advantages of the experimental system are demonstrated considering these cases studies?  How are test beds prepared to represent each case study?

For each case study, relevant variables, objectives of experiment and control problem should be described. 

Discussion and conclusions are vague. A clear statement of paper objectives is necessary to define the relevant aspects to be analyzed.

Author Response

We would like to thank the reviewer for spending time, very careful reading and providing us with the comments. We appreciate very much that the reviewer kindly suggested new formulations. We have found all the comments very valuable in improving the manuscript. The major corrections are highlighted in the manuscript in blue, the reviewers’ comments are written in italic and our replies are written in plain text. We have given point-by-point responses to the reviewer comments below where we have indicated how we have addressed them:

1. Revise text between lines 21-28, sentence’s structure is confusing.
   1. Indeed, this part has not been expressed clearly. The introduction has been rewritten to clarify the motivation.
2. Research objectives should be clearly defined in the introduction section. Disperse ideas that justify the implementation of the test modules and the possibility of evaluate different control strategies are presented but, a clear statement of the objectives of the paper is necessary.
   1. We agree with the reviewer, a clear statement of the paper objectives is necessary. We have restructured the introduction section such that the paper objectives are listed explicitly. The objectives are now divided into laboratory objectives and research objectives. (See lines 67-141).
3. What are the aspects of the test-beds implementation that are addressed in the paper: How has been the system designed and implemented? The implementation of control strategies? Possibility of reconfiguration to adapt to different systems? Impact of system behavior that it is difficult to represent in simulation environments?  All these aspects had been mentioned, but none of them is appropriately addressed in the paper.
   1. We have restructured the “materials and methods” section in three parts. In the new version, the design of the modules and the design of the networks and implementation are addressed separately. (See correction lines 155-205 and 206-312 respectively). The third part contains the hardware design.
   2. This paper proposes a test facility to validate control strategies, the paper gives evidence of the usefulness of the system by giving examples of four control strategies implemented at the laboratory; we have included a clarification for emphasizing the contribution of this paper (See lines 138-141)
   3. The possibility of reconfiguration is addressed in the materials and methods. This section explains why we choose to design both systems (hydraulic circuits and communication network) with a modular architecture. The original submission needed a better definition of the objectives a new structure, the new version aligns the paper objectives with the laboratory development and findings.
   4. We have removed the part where we compare the laboratory with simulation environments. Although we find it relevant for some specific applications, we believe that it might confuse the reader from the main scope of the paper.
4. In materials and methods section model, components description and implementation issues are mixed. It is confusing for readers. How demand profiles recreated by real data are implemented in experimental system (lines 123-125)?
   1. These sections have been separated into “module design”, where the laboratory modules are designed based on mathematical models, and “network design”, where the design of a pipe network with laboratory modules is addressed. The first activity addresses the fixed design of the modules, whereas the latter is the part that will be changed between experiments making the lab flexible. (See correction in lines 155-205 and 206-312 respectively). Additionally, we have added a new paragraph on the “results” section where the configuration needed to transform a real network problem to a laboratory experiment is described. (See lines 365-381)
5. It should be separated experimental objectives as the indicated between lines 115 and 119, and system representation and configuration. WDN? WWC? Expl98
   1. We have restructured the introduction to include a list of the experimental objectives. Later, in the results section we provide four examples of the experimental objectives achieved for optimal management and fault tolerant control. (see lines 106 -141)
6. What happen with the accumulation term in equations 2, 3?
   1. We have changed the notation of the accumulation term in equation (2). Equation (3) is an algebraic equation relating nodal pressure and tank level, it does not have accumulation term. (See correction in lines 173-174)
7. Figure 6 and figure 8, describe how modules are integrated to represent the systems presented in figure 2? An explanation is necessary to understand how the systems presented in figure 2 are represented by test beds modules.
   1. We have modified the figures 6 and 8 to illustrate the link between the structure of a real network and the laboratory modules. A subsection is included to clarify the design process from the real network to test-bed (See lines 365-381).
8. In the results section, the selection of the case studies and system representation should be explained and justified. Which characteristics and advantages of the experimental system are demonstrated considering these cases studies?  How are test beds prepared to represent each case study?
   1. In general, the main advantage of performing laboratory experiments is the fast implementation in a safe environment, we have added a brief description of the contribution of the experiment for each study case.
   2. We have given an extensive description of how the characteristics of the study cases are emulated in the result section for each study case.
9. For each case study, relevant variables, objectives of experiment and control problem should be described.
   1. We have given additional information of each experiment in the result section; this means that we have edited some labels in the figures to match the legends of the plots. We believe that this change will help the reader the interpretation of the results.
10. Discussion and conclusions are vague. A clear statement of paper objectives is necessary to define the relevant aspects to be analyzed.
    1. In the new version, we have reformulated the discussion to summarize the findings of the paper and align them with the paper objectives. The contribution of the laboratory in the validation of control solutions is highlighted.  The conclusion is edited to contain a better description of the findings.

Reviewer 3 Report

Article by Val Ladesma et al. is presenting Smart Water Infrastructures Laboratory (SWIL) hosted at Aalborg University. It is a very well written and easy to read article. Highly recommended for undergraduate students in Environmental Engineering.  

However, this is not scientific nor applied research article. This is presentation of very good laboratory-scale university facility. Article seems to based on materials and methods described for students' laboratory works manual.

Authors did not properly addressed scale effect, basing only on selected relatively small-scale real cases and have drawn general conclusions based on that. Comparing real WDN systems to physical laboratory-scale models based on DN25/DN15 is very ambitious and should be thoroughly supported by dimensional analysis (at least by comparing flow regimes in terms of Re, Fr) and proper analyses of scale-effect.

I strongly recommend this article to be sent for publication in popular magazines which are targeting appropriate audience (IWA Source).

Author Response

We would like to thank the reviewer for his valuable time and advice. The major corrections are highlighted in the manuscript in blue, the reviewer’s comments are written in italic and ours are written in plain text. We have given point-by-point responses to the reviewer comments where we have indicated below how we have addressed them:

1. Article by Val Ladesma et al. is presenting Smart Water Infrastructures Laboratory (SWIL) hosted at Aalborg University. It is a very well written and easy to read article. Highly recommended for undergraduate students in Environmental Engineering.
   1. We have reformulated the introduction section to highlight the main objective of the paper. The objective of this paper is to inform about the laboratory validation methods. This laboratory focuses on the discovery of monitoring and control solutions. However, other disciplines can also benefit from the data generated with the laboratory experiments. (see explanation in lines 114-118)
2. However, this is not scientific nor applied research article. This is presentation of very good laboratory-scale university facility. Article seems to based on materials and methods described for students' laboratory works manual.
   1. This is a legitimate observation; this paper does not propose a collection of control methods, it presents a tool and methods to verify the control solutions. To accommodate this comment, we have modified the structure of this paper to clearly state the link with the content in “materials and methods” with the findings presented in the section “Results”.
   2. This paper addresses the development of a laboratory and proposes four verification methods for the study of control solutions. We consider that the validation method that we propose with the SWIL contribute to facilitate the research in this field. We have reformulated the paper objectives to emphasize the contribution. (See the research objectives section line 106)
3. Authors did not properly addressed scale effect, basing only on selected relatively small-scale real cases and have drawn general conclusions based on that. Comparing real WDN systems to physical laboratory-scale models based on DN25/DN15 is very ambitious and should be thoroughly supported by dimensional analysis (at least by comparing flow regimes in terms of Re, Fr) and proper analyses of scale-effect.
   1. We agree with the reviewer, we have included a new subsection that addresses the scale-effect and describes considerations taken for reducing a real scale network to a small-scale test bed. (See correction in lines 218-251 )
4. I strongly recommend this article to be sent for publication in popular magazines which are targeting appropriate audience (IWA Source).
   1. We appreciate this comment, however we consider that the audience of this special issue, that is devoted to findings within monitoring and control solutions, could specially benefit from reading this article.