Leakage discharge separation in multi-leaks pipe networks based on improved Independent Component Analysis with Reference (ICA-R) algorithm
Li, Yuanzhe and Gao, Jinliang and Wu, Wenyan (2018) Leakage discharge separation in multi-leaks pipe networks based on improved Independent Component Analysis with Reference (ICA-R) algorithm. In: International Joint Conference in Water Distribution Systems Analysis & Computing and Control in the Water Industry, 23-25 July 2018, Kingston Ontario Canada.
Preview |
Text
F107FinalPaperWDSACCWI Yuanzhe Li.pdf Download (1MB) |
Abstract
The existing leakage assessment methods are not accurate and timely, making it difficult to meet the needs of water companies. In this paper, a methodology based on Independent Component Analysis with Reference (ICA-R) algorithm was proposed to give an more accurate estimation of leakage discharge in multi-leaks water distribution network without considering the specific individuality of one single leak. The proposed algorithm has been improved is improved to prevent error convergence in multi-leak pipe networks. Then an example EPANET model and a physical experimental platform were built to simulate and evaluate the flow in multi-leak WDNs, and the leakage flow rate is calculated by improved ICA-R algorithm and FastICA algorithm. The simulation results are shown the improved ICA-R algorithm has better performance
Item Type: | Conference or Workshop Item (Paper) |
---|---|
Dates: | Date Event 30 June 2018 Accepted 23 July 2018 Published |
Uncontrolled Keywords: | physical leakage flow; blind source separation; Improved Independent Component Correlation algorithm with reference(ICA-R) |
Subjects: | CAH10 - engineering and technology > CAH10-01 - engineering > CAH10-01-01 - engineering (non-specific) |
Divisions: | Faculty of Computing, Engineering and the Built Environment > College of Engineering |
Depositing User: | Wenyan Wu |
Date Deposited: | 10 Oct 2018 15:31 |
Last Modified: | 20 Jun 2024 11:50 |
URI: | https://www.open-access.bcu.ac.uk/id/eprint/5947 |
Actions (login required)
View Item |