Title: Water distribution networks leak detection using graph theory and optimised LSTM

Authors: R. Sengothai; R. Sivaraman

Addresses: Pie Mathematics Association, Chennai, 600094, India ' Dwaraka Doss Goverdhan Doss Vaishnav College, Chennai, 600106, India

Abstract: Water supply issues are made worse by leakage in the water distribution network (WDN). Because there is a higher chance of contamination from leaky openings, leakage lowers the quality of water. It also results in pressure drops and power losses. These drawbacks are worse in undeveloped nations because of the substandard design quality and advanced age of distribution systems. Identification of the WDN leak is crucial due to the limited supply of water resources. In this research work, a graph theory and artificial network-based leak detection method are proposed to address the problem of leakage in WDNs. The proposed method is designed to detect leaks in WDNs with high accuracy. The technique uses graph theory to split the WDN into two sections, and then pressure loggers and the optimised long short-term memory (O-LSTM) algorithm are used for determining which section leaks.

Keywords: water distribution networks; WDNs; long short-term memory algorithm; long short-term memory; LSTM; leak detection; African vulture optimisation algorithm; AVOA; graph theory.

DOI: 10.1504/IJMOR.2024.140072

International Journal of Mathematics in Operational Research, 2024 Vol.28 No.3, pp.347 - 373

Received: 16 Mar 2023
Accepted: 26 Mar 2023
Published online: 19 Jul 2024 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article