Criticality Analysis of a Water Distribution System Considering Both Economic Consequences and Hydraulic Loss Using Modern Portfolio Theory

Round 1

Reviewer 1 Report

The study is focused on the strategies for the management of water distribution systems. In particular, a new methodology is proposed for assessing the priorities in the replacement of sigle pipes, in order to reduce the economical impact of system failures. The mehodology is based on the Modern Portfolio Theory (MPT) used in optimal investment strategies, and applied in the wate rmanagement contest in order to combine the resiliences coming from hydraulic losses and econimic consequences on the users, minimizing the assets risk.

Basically, the use of MPT makes less questionable the common practice of combinining funcionality factors of different nature in the optimization process.

Unfortunately, it is quite common, when a theory from a different scientific field is applied in water management and civil engineering, to give a great emphasis to the methodology but losing in the scientific soundness of the application.

In my experience, the economic consequence of pipe leakage could come from two kind of event:

the direct and indirect cost of the water volumes lost in the soil

the cost for the end users, and, consequently, for the water utilities coing from system failures.

The first event is common in the water distribution network and it is not connected only to a single pipe failure, but to pipe age. The strategy to face this kind of diffuse network failure is not based only on pipe replacement, but in pressure management, districtualization, energy recovery, and so on. Pipe rehabilitation is scheduled, and it is not limited to a single branch but to a sector of the network.

The second event, represented by a pipe breakage leading to a service interruption is observed mainly in water transmission. For the anular shape of the distribution network, water can be redirected on a different pathway in case of pipe breakage with very short recovery times. Considering the reduction in pressure due to pressure managment, by operating on the valves and reducing dissipation the same water volumes can be supplied by the new emergency pathway.

As a consequence, the case studies of the paper, based on an artificial network, NAICS users concentrated in the node, reduced flexibility of the network, pipe replacement as unique leakage reduction strategy, etc., make the analysis unrealistic.

Without any knowledge of the network more than Fig. 2, a pipe replacement strategy based on the priorities of Figure 4, and supported by the graph of figure 6 is in my opinion meaningless.

In my opinion, the paper has to be resubmitted including:

- a better discussion in the introduction of the problems of pipe leakages, available strategies and problems of pipe failures leading to high economic risks.

- a clear statement that in the paper the author is making an excercise, but that additional factors and combined rehabilitation strategies have to be included to male the method a ready to use approach

- a warning that in this simplified view the use of resilience is proposed even if these factors are used in literature for high rank risk, like flooding and heathquakes, where no emergency operation on the network is possible

- a better discussion of ECLIPS, and an explanation of the meaning of lines 125-127 concerning the limitations of ref 20 assumptions in the test case analysed in the paper

- a complete description of the network with pipe and node numbers and pressure distribution

- a better description of the EPANET model and of the results. For example, I don't understand why does the hydraulic resilience increase as a consequence of a pipe replacement with no breakage. Are you giving a Q(h) law in the nodes? In wich nodes are you already in a flow deficit condition before pipe replacement?

Finally, you have 432 nodes, but the users node ID are between 4 and 23. Are all users considered in the study concentrated in a small zone. In this case, why? Or this node number is a simple identification number?

Author Response

We appreciate the helpful comments from Reviewer #1. Based on the comments, we substantially revised the original manuscript. Detailed responses explaining how we addressed each comment are given in a separate file.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comment 1: What is the logic of equation 2? Alpha appears both numerator and denominator? How come same weightage has been retained both design demand and actual supply? Is it viable? Under failure conditions, nodal behavior will be changing.

Comment 2: Why is it two curves for ρ = 0.5? Kindly explain. Or is it a typographical mistake?

Comment 3: Quantification of economic losses is not clear in the paper. Authors should demonstrate with some numerical values how it can be calculated.  

Comment 4: Pipe Breakage location should be shown in the network layout. At least major locations may be highlighted.

Comment 5: Include the following papers in the literature discussion.

1.         T.R. Neelakantan, C.R. Suribabu and Srinivasa Lingireddy (2008) “Pipe-sizing optimization including break-repair and replacement economics” Water SA Journal 34(2), 217-224.

2.         C.R.Suribabu and T.R.Neelakantan (2012). “Sizing of water distribution pipes based on performance measure and break-repair-replacement economics. ISH Journal of Hydraulic Engineering, Taylor and Francis Publishing, 18(3), 241-251.

3.         T.R.Neelakantan and C.R.Suribabu (2017) Discussion on “New Resilience Index for Urban Water Distribution Networks” by G. P. Cimellaro,  A. Tinebra, C. Renschler, and M. Fragiadakis. DOI:10.1061/(ASCE)ST.1943-541X.0001433, Journal of Structural Engineering,143(8),0701001-1.

Author Response

We appreciate the helpful comments from Reviewer #2. Detailed responses explaining how we addressed each comment in the revised manuscript are given in a separate document.

Author Response File: Author Response.pdf