Level of Service-Based Asset Management Framework for Water Supply Systems

Ahmed Serag, Soliman Abu-Samra, Tarek Zayed

Research output: Journal article publicationJournal articleAcademic researchpeer-review

2 Citations (Scopus)

Abstract

Operators of North America's potable water systems are facing numerous challenges in meeting the current needs and future expectations. Even though utility management experts started building customer-driven asset management systems to prioritize the water mains' maintenance and replacement, the gap between the utility experts' and end users' perspectives still exists due to the lack of technical knowledge in terms of assessing the water quality. Therefore, this paper proposes a service-based asset management framework that evaluates the factors associated with the level of service (LOS) of the water supply networks and maps it to the physical condition. In this paper, the LOS for water supply networks is an indicator that measures the ability of a municipality to continuously supply the end users with adequate water quality to ensure fewer customer complaints and higher end-user satisfaction. The framework revolves around three phases: (1) data collection; (2) model implementation, which comprises LOS assessment and LOS and condition mapping models; and (3) results and analysis. To assess the LOS, a questionnaire was designed and analyzed using the best-worst method. Furthermore, an artificial neural network model was developed to map the relationship between the LOS and condition. Water quality, customer complaints, pressure, and continuity of water supply were used as mapping metrics between the LOS and condition. Toward the end, the framework was applied to the water distribution network of Montreal, Canada and it showed promising results in estimating the corresponding LOS from the condition. In addition, a cross-validation was carried out and the results displayed an 0.871 coefficient of determination (R2), which implies a strong existing relationship between the model inputs and outputs. This framework enables the utility experts to understand the customer perception of the service, optimize the budget allocation, and forecast the LOS based on the network condition.

Original languageEnglish
Article number04020026
JournalJournal of Pipeline Systems Engineering and Practice
Volume11
Issue number3
DOIs
Publication statusPublished - 1 Aug 2020

Keywords

  • Artificial neural network
  • Best-worst method
  • Level of service (LOS)
  • Multicriteria decision-making
  • Optimization
  • Pipelines
  • Water supply systems

ASJC Scopus subject areas

  • Civil and Structural Engineering
  • Mechanical Engineering

Cite this