Abstract:
One of the most significant concerns of humankind has been the provision of adequate
quantities and quality of water. Unfortunately, modern water distribution networks are complex
and challenging to manage due to increased levels of urbanisation, varying consumer demands,
and limited resources. Therefore, authorities must make critical decisions to solve problems
associated with different phases and time frames of water distribution networks. Most of the
management issues are challenging to solve using conventional problem-solving techniques.
One of the versatile approaches for solving these management problems is to use customer
complaints to develop optimisation techniques. National Water Supply & Drainage Board
(NWSDB) receives various complaints about their services. This thesis presents a case study
on the Biyagama water supply scheme, focusing on analysing leakage occurrence in the
pipeline network. The study utilised a dataset collected by the National Water Supply and
Drainage Board (NWSDB) for three years. Aim is to investigate the factors influencing leak
occurrence in the Biyagama water supply scheme and develop a comprehensive understanding
of pipe network vulnerability, with the objective of informing targeted maintenance strategies
and improving the overall reliability and efficiency of the water supply system. The primary
objective was to investigate the factors influencing leakages in the network and develop a
comprehensive understanding of pipe network vulnerability. The research employed the Kmeans
algorithm, implemented using the Python programming language, to conduct a cluster
analysis of the dataset. The analysis incorporated factors, including age, inner diameter, and
hydraulic parameters, such as pressure and velocity, to identify patterns and correlations
associated with leak occurrence. The cluster analysis results revealed that age physical
characteristics and hydraulic parameters alone were insufficient to explain the variations in
leak frequency among the clusters. To account for this, additional factors were considered,
including pipe material, installation quality, maintenance practices, temperature, soil condition,
traffic load and environmental conditions. Incorporating hydraulic parameters, such as pressure
and velocity, provided valuable insights into the relationship between fluid dynamics and leak
vulnerability. High fluid velocities also correlated with leak occurrence. The findings
emphasise the complexity of pipe network vulnerability and highlight the importance of
considering multiple factors in assessing leakage occurrence. By gaining a comprehensive
understanding of the factors influencing leaks, stakeholders can develop targeted maintenance
strategies and improve the overall reliability and efficiency of the water supply system. This
research contributes to the field of water supply management by providing insights into the
vulnerabilities of pipe networks and the factors influencing leak occurrence. The findings can
inform decision-making processes for infrastructure maintenance and support the development
of sustainable water management strategies.
