A decision support model for the rehabilitation of “non-critical” sewers
Richard Fenner, L Sweeting
Abstract
Richard Fenner, L Sweeting
Abstract
20% of the UK drainage network is made up of “critical sewers” (those with the highest economic consequences of failure) and in the last 15 years these have been systematically rehabilitated. The remaining 80% of non-critical sewers are dealt with by reactive maintenance only. The paper describes how a flexible and rational decision support model for rehabilitating non-critical sewers has been developed by analysing existing sewer performance data and asset information. The method represents information contained in asset and event databases in a GIS to rank variable sized grid squares into priority zones for action. A second stage uses a Bayesian statistical analysis of each pipe length within those grid squares most at risk from sewer failure. The model has been validated on data from several water company regions and whilst it does not enable an absolute prediction of sewer condition, the procedures help to distinguish those parts of the system in greatest need of attention.
OpenAlex reports 34 citations for this work. Citation counts describe recorded attention and do not establish research quality.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
20% of the UK drainage network is made up of “critical sewers” (those with the highest economic consequences of failure) and in the last 15 years these have been systematically rehabilitated. The remaining 80% of non-critical sewers are dealt with by reactive maintenance only. The paper describes how a flexible and rational decision support model for rehabilitating non-critical sewers has been developed by analysing existing sewer performance data and asset information. The method represents information contained in asset and event databases in a GIS to rank variable sized grid squares into priority zones for action. A second stage uses a Bayesian statistical analysis of each pipe length within those grid squares most at risk from sewer failure. The model has been validated on data from several water company regions and whilst it does not enable an absolute prediction of sewer condition, the procedures help to distinguish those parts of the system in greatest need of attention.
Key concepts: Sanitary sewer, Rank (graph theory), Asset (computer security), Decision support system, Grid, Bayesian probability, Asset management, Engineering