The efficient long term management of large-scale public funded assets is an area of growing importance. Ageing infrastructure, growth and limited capital all result in the need for a more robust and rigorous methodology to prioritise rehabilitation and renewal decisions and, as importantly, to forecast future expenditure requirements. The overall objective of this research is to develop a Bayesian-based decision support system that will facilitate the identification of efficient asset management policies. The Bayesian approach enables us to formally incorporate, express and update our uncertainty when determining such policies. This is particularly relevant for water utilities that have incomplete or unreliable historical failure data sets and, as a consequence, rely heavily on past engineering experience.
An object oriented discrete event simulation has been developed to analyse existing maintenance policies, test the Bayesian methodology and to develop and identify improved maintenance policies. This paper focuses on the areas of research relating to the long term management of water distribution systems and, in particular, will present: (1) an overview of the Bayesian approach, (2) development and initial results for an object oriented discrete event simulation and (3) proposed future research and development.
Application of the Bayesian Approach to Incorporate Helium Isotope Ratios in Long-Term Probabilistic Volcanic Hazard Assessments in Tohoku, Japan