Leak detection in water distribution network using steady-state modelling approach | |
| Author | Olorundare, Ranti Matthias |
| Call Number | AIT Thesis no.UWEM-22-01 |
| Subject(s) | Leak detectors Water--Distribution--Management Water leakage--Management |
| Note | A thesis submitted in partial fulfilment of the requirements for the degree of Master of Science in Urban Water Engineering and Management Jointly offered by Asian Institute of Technology, Thailand and IHE Delft Institute for Water Education, the Netherlands |
| Publisher | Asian Institute of Technology |
| Abstract | Leakage detection is a pertinent issue in water distribution networks (WDNs). Poor leakage management and control result in huge losses of valuable resources such as drinking water quality and can also lead to its contamination, thereby creating a public health risk. To prevent these losses, water utilities have been using hardware devices for leakage detection and localization, but these devices all have their drawbacks. It is expensive to apply especially when detecting leaks for the whole city, it is time-consuming and requires lots of human force. However, the idea of the software leaks detection approach emerged among researchers, which has been classified into two, namely, steady-state and unsteady-state approaches for leakage detection. The steady-state approach has been tested by some water utilities on an experimental site and it has been proven to be effective in determining the size of leak but not adequate in pinpointing leaks. Although, the two approaches have not yet been implemented in a real water network, it is still at research stage. For this reason, this research therefore digs deep into steady state analysis for leakage detection using optimisation tool. A synthetic network was created on EPANET with an artificial leak, similar network was also built on Jupyter notebook, a python integrated development environment with the help of WNTR python package. This network was optimised with a python optimisation package called PyGAD. The strategy used in this study was to try several times to get the optimizer to converge to the assumed leak size added in the artificial network. For particular timesteps, the network is first simulated using an artificial leak (with defined leak location and size). After that, the results are then sent into the optimization algorithms as function inputs. Genetic algorithms are given a challenge of identifying the leak location and size comparable to predetermined values based on these function inputs. This allowed the user to see if the optimization algorithms had converged on the best option for the desired outcomes. |
| Year | 2022 |
| Type | Thesis |
| School | School of Environment, Resources, and Development (SERD) + School of Engineering and Technology (SET) |
| Department | Department of Civil and Infrastucture Engineering (DCIE) |
| Academic Program/FoS | Urban Water Engineering and Management (UWEM) |
| Chairperson(s) | Babel, Mukand Singh;Ferras, David (Co-Chairperson); |
| Examination Committee(s) | Xue, Wenchao;Shanmugam, Mohana Sundaram; |
| Scholarship Donor(s) | Orange Knowledge Programme (OKP);Ministry of Foreign Affairs, The Netherlands/NUFFIC/IHE Delf;Asian Institute of Technology; |
| Degree | Thesis (M. Sc.) - Asian Institute of Technology - UNESCO-IHE, 2022 |