Peter Chalk Centre

University of Exeter

Stocker Road

Exeter

EX4 4QD

Tel: +44 (0)1392 263637

E-mail: CCWI2019@exeter.ac.uk 

17th International Computing & Control for the Water Industry Conference

1st - 4th September 2019
University of Exeter, UK
  • Facebook Social Icon
  • Twitter Social Icon
  • Instagram

2B Demand, leakage, energy

Please note the time shown on this page is automatically adjusted by the server according to the time zone set in your computer.

Green

Maria Cunha

Chair:

to

Leak Management in District Metered Areas with Internal Pressure Reducing Valves

G R Anjana

Presenter:

Authors:

G R Anjana, Avi Anthony and M S Mohan Kumar

Water utilities around the world are under tremendous stress to maintain the non-revenue water in their networks with in economic levels. Pressure management is identified as one of the most efficient way to reduce leaks in Water Distribution Systems (WDS), especially background leaks in the systems, which are difficult to locate. Most of the studies for pressure management deals with a single DMA and a boundary value corresponding to that DMA , which is located at the inlet. For the real-world case study under consideration, there are multiple internal Pressure Reducing Valves (PRVs) within the DMA , which makes direct application of the existing pressure management techniques difficult. Hence, in this study, the DMA is divided into multiple Pressure Management Areas (PMAs), based on the location and control area of each PRV in the DMA. A remote node-based modulation is formulated for obtaining the valve setting for each time period, as the network demand varies for 24 hours. The simulation optimization framework formulated for this problem is implemented using EPANET and MATLAB. The network was delineated into different PMAs, and it was found out that, in certain PMA, there are more than one PRV associated with it. For these PMAs, the dependent PRVs need to be throttled simultaneously, for achieving the target pressures.The results of the optimization show that, the pressures across the network was reduced , and the critical node pressure was maintained around a constant value for the duration of study. The value of the required minimum pressure is determined such that, the demand satisfaction is 100% across every customer node. For the water network under study, a 6% reduction in NRW was observed after implementation of this pressure management strategy.

to

Prediction of PaT energy production in water distribution systems under stochastic water demand

Stefania Fontanella

Presenter:

Authors:

Stefania Fontanella, Veronica Pepe, Luca Cozzolino and Renata Della Morte

The importance of clean energy production for the reduction of greenhouse gas emissions and global warming is becoming increasingly important, and this prompts attention towards of alternative energy uses. In this field, researchers have recently focused their attention on the exploitation of water distribution systems (WDSs) head losses for the production of electrical energy. Recently, the use of reverse Pumps that work as Turbines (PaTs) is becoming an attractive alternative, due to the lower price with respect to the most expensive conventional turbine. In order to maximize the energy production, it is important to know the actual values of the available head and flow rate, because they determine the operating point of a given PAT. Of course, the WDSs may be characterized by strong daily flow rate variability and oscillations of the pressures, and this requires a flexible management of the PAT. The modelling of random daily flow variability in the WDSs plays an important role in the evaluation of electrical energy production. In this research, it is presented a comparison between the actual daily energy production and the energy corresponding to the BEP, showing that the expression of the daily average energy production is dependent on the values that characterize daily demand.

to

Detection of Multiple Anomalies in Pipe Networks using a Paired-IRF Technique

Wei Zeng

Presenter:

Authors:

Wei Zeng, Jinzhe Gong, Aaron Zecchin, Angus Simpson, Benjamin Cazzolato and Martin Lambert

Water distribution systems (WDSs) typically consist of buried pipe networks that are often deteriorating with age. Many pipe sections are blocked, deteriorated or even cracked, which lead to a reduction of the water transmission efficiency, potential pipe bursts and potentially a large amount of water lost through leakage. Detection and localisation of these anomalies (leak, blockage and deterioration) in pipe networks is critical for targeted maintenance of WDSs. Controlled hydraulic transient pressure waves can be injected into a pipeline for the detection of anomalies. Theoretically, any physical discontinuities can induce wave reflections. Persistent hydraulic transient waves excited by a pseudo-random binary sequence (PRBS) signal have been found useful in extracting pipeline information such as a pipelines impulse response function (IRF), and the IRF can be determined for leak detection and localisation. However, the IRF response extracted from test data typically involves errors (i.e. resulting from background transients during testing or numerical artefacts generated during the IRF determination), and it is challenging to distinguish the anomaly-induced reflections that present as small spikes from these errors. The research reported here presents the application of a new paired-IRF method on a pipe network with multiple types of anomalies. The persistent PRBS signal and a dual-sensor configuration are used. Using the two measured transient pressure traces, a paired-IRF response can be determined through a deconvolution process. The anomaly-induced features manifest as paired spikes with specific patterns. The utilisation of these patterns can enhance the robustness of the detection process. Numerical simulations are conducted in a simple pipe network, and the paired-IRF response from leaks, blockages and are demonstrated.

to

Disaggregation of Household Water Use by Means of a Rule-Based, Automated Methodology

Filippo Mazzoni

Presenter:

Authors:

Filippo Mazzoni, Stefano Alvisi, Marco Franchini, Marco Ferraris and Zoran Kapelan

Strategies to improve the efficiency of water infrastructure and face the demand increase in urban areas are strictly related to the correct management of water resources. Concerning the residential sector, customers and utilities can be supported by smart-metering technology, which allows to understand when and where water is mostly used. Besides, if smart-metered data are coupled with a methodology for water end-use disaggregation, further information about how water is used within a household can be obtained. In this work, a new methodology for automated disaggregation of domestic water use is presented. The approach is rule-based and relies on volume information recorded with medium frequency (i.e. every minute) by a single smart meter at the household inlet point. The developed algorithm makes use of deterministic rules based on physical water use parameters (i.e. duration, flow rate, consumed volume) to classify water events into the corresponding end-use. The methodology for the automated disaggregation is calibrated and validated with smart-metered consumption data obtained from the intrusive monitoring of some households located in the Bologna district (Italy). Moreover, the method performance is compared against the one of a manual disaggregation approach and the accuracy of both methodologies is evaluated by comparing disaggregated data with the field-observed end-use events. Based on the obtained results, it emerges that the automated methodology is accurate in detecting water uses and disaggregating consumed volumes. In addition, unlike similar automated approaches, the disaggregation can be successfully achieved with medium frequency of readings, which characterizes most commercial smart demand meters.

to

Stochastic drinking water demand model parameterisation with smart meter data and data science algorithms

David Steffelbauer

Presenter:

Authors:

David Steffelbauer, Mirjam Blokker, Arno Knobbe and Edo Abraham

Ageing drinking water systems and rising water demands force water utilities to increase their systems efficiency. An increasing number of water utilities use hydraulic computer models to design and operate their systems in a more efficient way. Unfortunately, the accuracy of these models is limited due to a lack of measurements and a great number of unknown parameters.  Recently, smart water meters became available that promise unprecedented insights into water demand at household level in high temporal resolutions. However, these smart meters are not yet installed ubiquitously nor are high time resolutions commonly available from currently installed smart meters. This work shows (i) how data science algorithms can be used to generate knowledge out of existing low-resolution smart meter data and (ii) how to combine these novel insights with stochastic demand models aiming to infer information on both higher temporal as well as higher spatial scales. First, daily patterns are identified in the smart meter data using the k-Means clustering algorithm with Dynamic Time Warping (DTW) metric. This patterns are linked to inhabitants (a) with and (b) without jobs at home. Second, a method is presented for computing robust statistics of the occupants’ habits from smart meter data in terms of wake-up times and times of leaving the house as well as durations of being away and asleep. Both methods applied on simulated and measured smart meter data resulted in important insights on input parameters for the stochastic demand modelling software SIMDEUM. This research is part of the project “DASH of Water – DAta-driven Stochastic Hydraulic models of Water distribution systems”. Project DASH of Water aims to develop beyond state-of-the-art methods to simulate distribution systems in a more realistic and accurate way by utilising the potential of recently available smart meter technologies.

to