Asset owning companies struggle to develop reliable estimates of asset remaining useful life of individual assets. This situation is exacerbated by the way in which reliability engineers traditionally, and still do, make end-of-life decisions for mechanical and electrical assets. Current practices focus on predicting the mean life of assets and do not coherently account for uncertainties. This results in overconservative decisions for the renewal or disposal of assets. The effect of these conservative decisions is the generation of work from over-frequent planned renewals plus work generated from unplanned premature failures. Both contribute to the large maintenance backlogs.
The specific aims of this project are to look at systems where there are multiple failures on the same (repairable) component and consider the case that the repair is not as-good-as-new. The aim is to develop new approach for the Water Corporation for this situation and compare this with the results of the way data is currently processed.
The project will look at a specific class of assets and the first task for the successful applicant is to consider different asset classes and make a reasoned selection. Examples include pipe repairs, waste water blockages, clear water pipe and valve repairs. The project will involve the use of Bayesian statistics so strong coding in R or Python and solid statistical background is required.