Early Failure Detection Model for Water Mains Due to Seasonal Climatic Impacts

Abstract

Breakage patterns of homogeneous groups of water mains are influenced by non-time-dependent (static) factors, such as material, diameter and vintage and timedependent (dynamic) impact factors, such as soil temperature, soil moisture and changing operating pressure. A steady increase in deterioration is responsible for increasing breakage rates and year-to-year variations of this breakage rates. Some of the year-to-year breakage rates variations among individual mains are a result of irreducible random natural variation (aleatory uncertainty), while some of the variation (epistemic uncertainty) can be explained by the existence of different factors. Partially the breakage rate variations are attributed to climatic impact factors. Hence these seasonal climatic impact factors were examined. The investigations show that yearly breakage rates with above-average failure frequencies are triggered by severe frost periods and are followed by years with reduced failure frequencies as the weak spots were purged due to climatic extremes. Further with an increasing number of previous observed breakages (NOPF) the probability of a pipe to break in freezing periods increases as well. In this paper an early failure detection model due to climatic impact “freezing periods” was derived. The aim of such a model is to identify individual water mains for early failure detection within leakage monitoring programs and further for whole of life cost analyses.