LAUSR

ABSTRACT This article introduces methods used for the assessment and prediction of water distribution network (WDN) reliability. A failure rate expressed as the number of failures per operating day is considered to be the basic measure of the achieved WDN reliability level. The proposed method is based on the assumption that information about WDN failures is available in the form of long-term time series. The authors propose a mathematical model using specific forms of state space models which takes into account a seasonal effect and an intervention variable. The proposed model enables us to identify and assess both long-term trends in WDN reliability evolution and short-term reliability changes related to seasonal effects and also to predict the future development of WDN reliability. The practical applicability of the proposed method is demonstrated by assessing the information set based on long-term observations of the selected WDN failures.