Predicting the Failure Performance of Individual Water Mains
This report was produced for the Urban Water Research Association of Australia, a now discontinued research program.
Report No UWRAA 114
Water-supply mains have been examined to determine the optimum timing for the replacement of a deteriorating asset. It has been found that, with suitable filtering of raw failure data for a pipe asset, it is possible to make reasonable predictions of the future failure behaviour of the asset. Drawing on work in data filtering by Minetti (1995) and by Constantine and Darroch (1995) for partial data sets, models have been developed which predict the time to the next failure and then suggest whether the main should be repaired or replaced in order to minimise life-cycle cost.
The report incorporates a set of guidelines for operators of water-supply mains. These can be applied by any Authority to a main with a history of four or more failures. Authorities are expected to refine these models according to their own operating environment and specific corporate goals. However, there is a strong recommendation that a standard format be adopted for the collection of performance data so that future research projects in this area can be assured of data consistency across the country (see Appendix B).
The findings of this project will enable more effective management of water-distribution networks by ensuring that a main is replaced at the end of its economic life.
Key words in this study are: asset management; water-supply pipe; failure; pipe replacement rules; data filtering; economic model.
Earlier the Urban Water Research Association of Australia (UWRAA) research has indicated the need to develop a reliable means to predict the failure performance of individual water mains. The project team has analysed the records of small-diameter asbestos-cement and cast-iron pipes and from the analysis developed failure and decision support models. The end product is a set of guidelines which operators can use to decide how to minimise the life-cycle cost of a main under their control.
A fundamental requirement of any water-main failure model is that it should represent the aging, or gradual deterioration, of the pipe in its operating environment. To develop a model on this basis it is necessary to remove data relating to externally imposed factors such as bad repair of a previous failure, faulty operation of the supply system, accidental damage during adjacent excavation work and intentional damage. To achieve this filtering, a process has been developed which screens the raw performance history data and leaves the failures due to inherent pipe characteristics. The latter are then employed in the failure prediction model to forecast the time to the next failure.
After extensive trials with data supplied by Melbourne Water and the Sydney Water Board, a model with an exponential form that predicts inter failure time was selected as the most appropriate. This failure model requires the full burst history of the main. In this regard, information on four or more failures are needed following the filtering process. If this is not available, the reader is referred to the companion UWRAA study presented in Research Report AM22 (Constantine and Darroch, 1995), which provides a failure prediction model for incomplete datasets.
The basic building block of predictive modelling is an accurate database of failure data on all mains.
There port presents results from a survey of water authorities that examined the data currently being collected in Australia. A major point coming from this survey was the obvious need for a set of standard definitions and fields for failure databases. It is strongly recommended that a standard be developed to provide consistency across the Australian water industry. A suggested format is provided in Appendix B.
The survey of water authorities also covered the currently used procedures for economic evaluation of water mains, and the replace/repair decision in particular. Greater emphasis is now being placed upon indirect costs associated with burst water mains and the total cost to the community. Using this information and applying the failure forecasting model developed in the of the project, a decision model was developed which employs the concept of Total Failure Cost to decide whether to repair or replace a failing main. Computer software has been written to assist authorities apply the model and this is available from the Infrastructure Research Unit, Department of Civil and Geological Engineering, RMIT.
The findings of this project are incorporated in a set of guidelines providing operators with a package of tools aimed at helping them minimise life-cycle costs. Advice on the collection and interpretation of performance data is followed by instructions on the application of filtering, the failure prediction process and the determination of an optimal replacement date.
It is suggested that authorities apply the guidelines included in this report and with local data, over a period of time, refine them to better manage their resources. The guidelines can be used immediately but they must be converted to your own environment using your own failure data.