Risk Based Asset Management Through Understanding Water Treatment Assets
Saskatchewan is facing a current infrastructure deficit, that is the current resources allocated to support its infrastructure are not in line with maintaining the current level of service into the future. This deficit has lead to a common assumption that assets are more likely to fail to provide their required service with age. While obsolescence and other factors may be related to age, the present work set out to investigate the relationship between probability of failure to provide service and expected life of assets, to allow asset managers to create more robust risk management strategies as part of their asset management plans. For the present work, failure is defined as any functional failure where the asset has not met its required level of service. By the functional failure definition, a failure can be any event from a short duration service interruption to catastrophic breakdown of an asset. The purpose of the present work is to look at probability of failure, and not consequence so assets will have two states in the study, fully meeting their required level of service (function) or not meeting their required level of service (functional failure). In the 1970s, two researchers with United Airlines, Nowlan and Heap, completed a study of when assets in aviation were likely to fail. The way failure is defined in their report is in line with the failure definition in the present work. Nowlan and Heap’s work lead to classifying failures into six different patterns. The largest group of failures, known as the F pattern, show that assets were more likely to fail early in their lives. The present work employed a similar study of its assets over a 13 month period to see if similar patterns could be found. To complete the study at the Saskatoon Water Treatment Plant, asset failure data was taken from the City’s Avantis Asset Management System. Failures were broken down by asset type and sub-type based on functionality to observe their age relative to expected life when they failed. The present work determined that the approach was sound and that there was sufficient data with some types of assets to fit to the patterns developed by Nowlan and Heap. The present work shows that likelihood of failure to provide service is not connected with age, and that additional study of the Saskatoon Water Treatment Plant asset’s is required in order to determine exact patterns of probability of failure for each of the assets types. After classifying the failures by each asset type, and accounting for errors in the data caused by the warranty period on assets that are maintained by the contractor, all types of assets were found to have higher occurrences of failure in the early portion of their lives than closer to the end of their expected life. Furthermore defining the shape of each asset type failures made it clear that age-based failure was not dominant. The study results lead to the conclusion that more research should be conducted not only at the Saskatoon WTP but with other infrastructure to better understand when assets are likely to fail to provide service so that probability can be integrated with risk management programs. Lastly this study has shown that it is not correct for asset managers to assume any pattern of failure for their assets without investigation.
DegreeMaster of Science (M.Sc.)
DepartmentCivil and Geological Engineering
SupervisorPutz, Gordon; Sparks, Gordon
CommitteeFonstad, Terry; Gustafson, Jan-Mark
Copyright DateDecember 2015