Random Bagging Classifier and Shuffled Frog Leaping Based Optimal Sensor Placement for Leakage Detection in WDS

Water Distribution Systems (WDS) are the large scale systems that demand design of enhanced leak detection and isolation techniques to prevent from water waste. Leakages lead to imperative loss of water in water distribution networks. Many works are published on leak detection of WDS. However, the e...

Full description

Saved in:
Bibliographic Details
Published in:Water resources management 2019-07, Vol.33 (9), p.3111-3125
Main Authors: Rayaroth, Rejeesh, G, Sivaradje
Format: Article
Language:English
Subjects:
Accuracy
Amphibians
Atmospheric Sciences
Bagging
Civil Engineering
Classification
Classifiers
Data
Decision trees
Detection
Distribution
Distribution management
Earth and Environmental Science
Earth Sciences
Environment
Error detection
Fitness
Geotechnical Engineering & Applied Earth Sciences
Hydrogeology
Hydrology/Water Resources
Leak detection
Leakage
Nodes
Optimization
Placement
Pressure
Sensors
Training
Water
Water distribution
Water distribution systems
Water engineering
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Water Distribution Systems (WDS) are the large scale systems that demand design of enhanced leak detection and isolation techniques to prevent from water waste. Leakages lead to imperative loss of water in water distribution networks. Many works are published on leak detection of WDS. However, the existing methods failed to improve the leakage detection accuracy and reduce the time. In this paper, Random Decision Tree Bagging Classifier based Shuffled Frog Leaping Optimization (RDTBC-SFLO) Technique is introduced. In RDTBC-SFLO, the collected pressure data are taken as training data. Random ID3 decision forest classifier process constructs the ID3 decision tree and produces the classification results of pressure data. After classification, number of nodes at abnormal pressure data is randomly generated as initial population in shuffled frog leaping optimization process. The fitness value of every node is calculated and the optimal nodes are chosen for the sensor placement in WDS with higher accuracy and minimal error rate. The simulation results show that RDTBC-SFLO technique increases the performance of water leakage detection with minimum classification time when compared to state-of-the-art works.
ISSN:0920-4741
1573-1650
DOI:10.1007/s11269-019-02296-7