No abstract

Most pipes in current urban water distribution systems have been laid down underground for several decades (AWWA, 2012;Che et al., 2021). These underground pipes are aging and fraught with leaks, which poses a tremendous challenge to water and energy savings as well as drinking water security (

No abstract

No abstract

No abstract

No abstract

Extended partial blockages in urban water supply systems (UWSS) are formed from complicated physical, chemical, and biological processes; thus, these blockages are commonly in random and nonuniform geometries. The transient-based blockage detection method (TBBDM) has been evidenced in many applications to be a promising way to diagnose these blockages. Despite the successful validation and application of the TBBDM in the literature, pipe blockages used in these studies were idealized and simplified to regular and uniform shape, which are however not common in practical UWSS, and thus invalidity and inaccuracy of this TBBDM has been widely observed in practical applications. This paper presents fundamental research on understanding the influence of more realistic and nonuniform blockages on transient wave behavior and the accuracy of current TBBDM. The blockage with a linearly varying diameter (termed as nonuniform blockage) is firstly investigated by the frequency domain analytical analysis for its impact on transient wave behavior, which is thereafter incorporated in the overall transfer matrix of transient frequency response for reservoir-pipeline-valve systems. The results indicate the nonuniform blockage may induce very different modification patterns on the frequency shift and amplitude change of transient waves from the uniform blockage situation.

In a recent paper (Alawadhi & Tartakovsky, 2020, https://doi.org/10.1029/2019WR025879), Alawadhi and Tartakovsky proposed an interesting, novel, efficient, and potentially powerful Bayesian-based approach to cope with the uncertainty in initial flow velocity when using transient waves to detect leaks in pressurized pipes. In this discussion, we illustrate that the uncertainty in initial flow velocity has negligible effect on the accuracy of transient-based leak detection methods. On the other hand, uncertainties in the wave speed have significant effects on the leak detection accuracy. Unfortunately, it appears that the proposed approach cannot be easily, if at all, extended to deal with uncertainties in wave speed and one may have to resort to classical Bayesian approaches which could be computationally intensive.