Modeling the Frequency of Water Main Breaks in Water Distribution Systems: Random-Parameters Negative-Binomial Approach

AbstractWater main breaks can have significant adverse social, economic, and environmental impacts. As a result, water utilities seek to be proactive and implement asset management programs to reduce the frequency of water main breaks and mitigate their impacts. A key to the success of these asset management programs is the ability to quantify the effect that a variety of factors may have on the likelihood of water main breaks, and hence identify those pipes that need to be inspected frequently. Using water-main break data for a 21-year period from two U.S. cities in the Great Lakes region, this paper demonstrates a methodology to estimate the system-wide monthly frequency of water main breaks as a function of a number of explanatory variables. Using a random-parameters negative-binomial approach, the statistical estimations show that pipe diameters, average pipe age, distribution of pipe age, pipe material, time of year, and mean monthly temperature all have a significant impact on monthly water main break frequencies. The effect that some of these explanatory variables have on break frequencies also varies across months in several cases (as captured by random parameters). Finally, the results clearly show that the relationship between explanatory variables and monthly break frequencies is system specific, as reflected by the many differences in the estimation results between the two water systems considered in this study.