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

4B 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

Shuming Liu

Chair:

to

Field Survey of Oversized Water Meters and Service Lines

Steven Buchberger

Presenter:

Authors:

Christopher Douglas, Steven Buchberger and Peter Mayer

Water meters and service lines are critical pieces of a building’s plumbing system. The water meter is designed to track water use so utilities and their customers understand how much water is consumed. Consumption data are vital for appropriate billing by the utility and for managing water as a resource. The service line is critical because it connects the rest of the plumbing system to the supply line. A small service line could create head loss such that the rest of the plumbing system cannot deliver water at adequate pressure. Damage to the piping might even occur due to water hammer. On the other hand, a large service line will add cost and increase residence times. Longer residence times give pathogens more opportunity to proliferate within the pipe. A comprehensive study of indoor water use at sites in the western United States was conducted to evaluate water meter sizing practices for buildings with water meters no smaller than 1 1/2” in diameter. Water meters and service lines were routinely oversized across 43 different buildings. Of the sites analysed, 86% were found to have oversized meters and 67% were found to have oversized service lines. The Water Demand Calculator (WDC) was found to be a useful tool for accurately estimating peak indoor water demand in residential buildings. Moreover, the WDC was recently adopted into the Uniform Plumbing Code. More research is required at non-residential buildings to calibrate the WDC and apply it other building types.

to

Optimal location and setting of time modulated PRVs for water loss reduction with leakage modelling by pressure driven analysis

Ricardo Gomes

Presenter:

Authors:

Joaquim Sousa, João Muranho, Alfeu Sa Marques and Ricardo Gomes

Water losses have several negative impacts and so water companies are always willing to reduce them [1][2]. Unreported leaks and background leakage usually represent a major component of water losses and pressure management is an effective, easy, economic and quick solution to reduce it. Adjusting the pressure to the minimum required reduces the leakage flows and so water losses. But pressure management projects must be preceded by specialized studies (identify the location and setting of the PRVs to install) and cost benefit analysis (assessment of economic viability). Following a previous work [3][4][5], this paper presents a methodology to help in those tasks, by identifying the optimal location and setting of time modulated Pressure Reductions Valves (PRVs) to reduce water losses in water distribution networks (WDNs) and maximize the Net Present Value (NPV) of pressure management projects. This methodology was embedded in WaterNetGen [6], an extension of EPANET. Leakage assessment is performed using pressure driven analysis and the optimal location and settings of the PRVs are identified by a Simulated Annealing algorithm. The results from a hypothetical case study are used to illustrate the potential of this methodology and to take conclusions.

to

Designing a Common Evaluation Framework for Leakage Detection in Water Distribution Networks

Magdy Abdullah Eissa

Presenter:

Authors:

Magdy Abdullah Eissa, Stelios Vrachimis, Demetrios Eliades and Marios Polycarpou

Various research groups are developing algorithms considering the water leakage detection problem, using different methodologies. These methods make different assumptions, in relation to the network topology, the type of sensors used and the measurement frequency, the uncertainty in the measurements and the model. Typically, these algorithms are trained and tested on different closed datasets, sometimes provided by end users, with very specific characteristics. In terms of accuracy, various metrics have been used to evaluate and compare each methodology, and different works may consider different metrics. To be able to evaluate different leakage detection algorithms, a common evaluation framework is required to allow leakage diagnosis researchers and practitioners to compare the different methodologies under various topologies and leakage types. This work proposes the upgrade of a previously proposed benchmark dataset for water leakage detection, the Leakage Diagnosis Benchmark (LeakDB), and the demonstration of a common analytical framework for assessing and evaluating different leakage detection methodologies published in the literature. In specific, the main contributions of this work are summarized in the following points 1) Enhance LeakDB by adding new leakage scenarios and realistic benchmark distribution networks; 2) Implement published leakage detection algorithms that use different solution methodologies (model-based, model-free, data-driven and signal-processing based approaches); 3) Design of a common framework for evaluating and analyzing all detection techniques using different metrics. The results from the multi-perspective assessment are presented and the performance of the algorithms is compared on different illustrated networks and scenarios provided.

to

Influence of Disinfectant Corrosion on Loss of Longitudinal crack on Different Pipe Materials in Water Distribution Network

Shiyuan Hu

Presenter:

Authors:

Jinliang Gao, Shiyuan Hu, Wenyan Wu, Yuanzhe Li, Shuhe Zou, Zhihui Liu and Chenhao Ou

The leakage model of water supply system can be used to guide the prediction of the leakage flow. The corrosion of disinfectant on the leakage of the pipe in water supply system increases the leakage flow and even produces potential safety hazards. Galvanized steel pipe, ductile iron pipe, UPVC pipe and PE pipe that are commonly used in water supply network are selected in this study, and the longitudinal crack of the pipe is selected as the study object. The corrosive effect of chlorine disinfectant on the leaked pipe is studied by performance test of corrosion pipe, dynamic corrosion experiment, experiment of pressure influence on leakage, and numerical simulation. On the basis of the test data, the improved FAVAD model of the longitudinal crack pipel is established by using ANSYS and 1stOpt software platform. The results show that the corrosive action of chlorine disinfectant has a certain destructive effect on the material properties such as the yield strength and elastic modulus of the leaked pipe. The performance of the pipe material has the greatest influence on the leakage flow, followed by the concentration of chlorine disinfectant and the influence of the longitudinal crack length. The improved FAVAD model can be used to guide the selection and replacement of water supply pipes, the dosage of chlorine disinfectants in water plants.

to

Comparison of seasonal domestic water use and the impact of household characteristics on per capita water consumption in Sirte, Libya

Iman Alharsha

Presenter:

Authors:

Iman Alharsha, Fayyaz Memon and Raziyeh Farmani

The water scarcity problem requires effective and efficient usage of the available fresh water resources. For this, knowledge of the seasonal domestic water consumption variability and daily water end-uses trends is vital. This knowledge is lacking for Sirte city in Libya. Basing on questionnaire survey data, this paper uses 25 IBM statistics models to investigate the seasonal per capita water consumption and the impact of household characteristics on per capita water consumption in Sirte city for summer season. Then regression techniques will be used to predict the future water demand for the study area in summer season. The statistical analysis results showed that the average water consumption was 255l/p/d and 380 l/p/d in winter and summer, respectively. Moreover, statistical analysis of the results showed that all water end uses, except hand wash basin, have significant differences between seasons. The per capita water consumption increases with the built-up area, number of rooms and number of flows, but it is not affected by the household income.

to

Short-term Drinking Water Demand Patterns and Forecasting

Rebecca M. Page

Presenter:

Authors:

Rebecca M. Page, Thomas Kunimünch, Michael Blessing and Nikolai Fink

The aim of this study was to forecast drinking water demand for a small community in north-western Switzerland for the next 24 hours. The approach was based on a multivariate dataset including demand time-series and weather information, namely hourly air temperature, atmospheric pressure and humidity data. The forecasting model were based on a Recurrent Neural Network (RNN) with Long Short-Term Memory Cells (LSTM). The in- and output time-series used to forecast the next day’s demand had a one-hour time step and were aggregated in a cloud-based system linked to the community SCADA system. The results of the forecast model are discussed in context with changes in the process and water distribution system, as well as with model-specific parameters. The acceptance of such data-driven management support systems greatly depends on the implementation and ease-of-use for the daily operational tasks. The study presented lays emphasis on the integration of the field-based measurements with contextual information with the aim to generate the basis for decision-making processes, such as minimizing the cost associated with energy consumption for pumping water during high-priced daytime hours, as oppose to lower cost night-time energy.

to