Leak Prediction Model for Water Distribution Networks Created Using a Bayesian Network Learning Approach

2016 ◽  
Vol 30 (8) ◽  
pp. 2719-2733 ◽  
Author(s):  
Sou-Sen Leu ◽  
Quang-Nha Bui
Keyword(s):  
Prediction Model ◽  
Bayesian Network ◽  
Water Distribution ◽  
Distribution Networks ◽  
Water Distribution Networks ◽  
Learning Approach ◽  
Network Learning

scholarly journals Reliability based rehabilitation of water distribution networks by means of Bayesian networks

2017 ◽  
Vol 34 (1) ◽  
pp. 163-172 ◽  
Author(s):  
Abdelaziz Lakehal ◽  
Fares Laouacheria
Keyword(s):  
Decision Making ◽  
Bayesian Networks ◽  
Bayesian Network ◽  
Water Distribution ◽  
Distribution Network ◽  
Water Distribution Network ◽  
Distribution Networks ◽  
Dynamic Bayesian Networks ◽  
Water Distribution Networks ◽  
Network Rehabilitation

AbstractWater plays an essential role in the everyday lives of the people. To supply subscribers with good quality of water and to ensure continuity of service, the operators use water distribution networks (WDN). The main elements of water distribution network (WDN) are: pipes and valves. The work developed in this paper focuses on a water distribution network rehabilitation in the short and long term. Priorities for rehabilitation actions were defined and the information system consolidated, as well as decision-making. The reliability data were conjugated in decision making tools on water distribution network rehabilitation in a forecasting context. As the pipes are static elements and the valves are dynamic elements, a Bayesian network (static-dynamic) has been developed, which can help to predict the failure scenario regarding water distribution. A relationship between reliability and prioritization of rehabilitation actions has been investigated. Modelling based on a Static Bayesian Network (SBN) is implemented to analyse qualitatively and quantitatively the availability of water in the different segments of the network. Dynamic Bayesian networks (DBN) are then used to assess the valves reliability as function of time, which allows management of water distribution based on water availability assessment in different segments. Before finishing the paper by giving some conclusions, a case study of a network supplying a city was presented. The results show the importance and effectiveness of the proposed Bayesian approach in the anticipatory management and for prioritizing rehabilitation of water distribution networks.

scholarly journals Data-Driven Leak Localization in Urban Water Distribution Networks Using Big Data for Random Forest Classifier

Mathematics ◽  
2021 ◽  
Vol 9 (6) ◽  
pp. 672
Author(s):  
Ivana Lučin ◽  
Bože Lučin ◽  
Zoran Čarija ◽  
Ante Sikirica
Keyword(s):  
Random Forest ◽  
Prediction Model ◽  
Water Distribution ◽  
Prediction Models ◽  
Distribution Networks ◽  
Daily Basis ◽  
Random Forest Classifier ◽  
Water Distribution Networks ◽  
Model Accuracy ◽  
Demand Variation

In the present paper, a Random Forest classifier is used to detect leak locations on two different sized water distribution networks with sparse sensor placement. A great number of leak scenarios were simulated with Monte Carlo determined leak parameters (leak location and emitter coefficient). In order to account for demand variations that occur on a daily basis and to obtain a larger dataset, scenarios were simulated with random base demand increments or reductions for each network node. Classifier accuracy was assessed for different sensor layouts and numbers of sensors. Multiple prediction models were constructed for differently sized leakage and demand range variations in order to investigate model accuracy under various conditions. Results indicate that the prediction model provides the greatest accuracy for the largest leaks, with the smallest variation in base demand (62% accuracy for greater- and 82% for smaller-sized networks, for the largest considered leak size and a base demand variation of ±2.5%). However, even for small leaks and the greatest base demand variations, the prediction model provided considerable accuracy, especially when localizing the sources of leaks when the true leak node and neighbor nodes were considered (for a smaller-sized network and a base demand of variation ±20% the model accuracy increased from 44% to 89% when top five nodes with greatest probability were considered, and for a greater-sized network with a base demand variation of ±10% the accuracy increased from 36% to 77%).

scholarly journals Leak localization in water distribution networks using classifiers with cosenoidal features

2020 ◽  
Vol 53 (2) ◽  
pp. 16697-16702
Author(s):  
I. Santos-Ruiz ◽  
J. Blesa ◽  
V. Puig ◽  
F.R. López-Estrada
Keyword(s):  
Water Distribution ◽  
Distribution Networks ◽  
Water Distribution Networks ◽  
Leak Localization

scholarly journals Bottom-Up Generation of Peak Demand Scenarios in Water Distribution Networks

2020 ◽  
Vol 13 (1) ◽  
pp. 31
Author(s):  
Enrico Creaco ◽  
Giacomo Galuppini ◽  
Alberto Campisano ◽  
Marco Franchini
Keyword(s):  
Water Distribution ◽  
Distribution Networks ◽  
Water Distribution Networks ◽  
Peak Demand ◽  
Long Time ◽  
Cross Correlations ◽  
User Demand ◽  
One Year ◽  
The One ◽  
Best Fit

This paper presents a two-step methodology for the stochastic generation of snapshot peak demand scenarios in water distribution networks (WDNs), each of which is based on a single combination of demand values at WDN nodes. The methodology describes the hourly demand at both nodal and WDN scales through a beta probabilistic model, which is flexible enough to suit both small and large demand aggregations in terms of mean, standard deviation, and skewness. The first step of the methodology enables generating separately the peak demand samples at WDN nodes. Then, in the second step, the nodal demand samples are consistently reordered to build snapshot demand scenarios for the WDN, while respecting the rank cross-correlations at lag 0. The applications concerned the one-year long dataset of about 1000 user demand values from the district of Soccavo, Naples (Italy). Best-fit scaling equations were constructed to express the main statistics of peak demand as a function of the average demand value on a long-time horizon, i.e., one year. The results of applications to four case studies proved the methodology effective and robust for various numbers and sizes of users.

scholarly journals First Results in Leak Localization in Water Distribution Networks using Graph-Based Clustering and Deep Learning

2020 ◽  
Vol 53 (2) ◽  
pp. 16691-16696
Author(s):  
Luis Romero ◽  
Joaquim Blesa ◽  
Vicenç Puig ◽  
Gabriela Cembrano ◽  
Carlos Trapiello
Keyword(s):  
Deep Learning ◽  
Water Distribution ◽  
Distribution Networks ◽  
Water Distribution Networks ◽  
First Results ◽  
Leak Localization ◽  
Graph Based Clustering
Sign in / Sign up

Export Citation Format

Share Document