| This study combined a national survey with field visits and laboratory tests to evaluate the performances of small diameter water pipelines. It focused on the historical performances of various types of pipelines over the years, pipeline breaks per day, causes of pipeline breaks, and methods to repair the leaks. Based on a national survey, representing over a population of 11 million, pipeline breaks varied from 0.33 to 12 breaks per day. Data from over 22,000 pipe failures in the City of Houston were collected and analyzed. Investigation showed that performances of the pipes were affected by ages, types of material, pipe lengths, pipe sizes, soil conditions, and climate factors. The field study indicated that there were correlations between types of soil and buried depths and water pipeline failures. Dominant failures were also in 6-inch and 8-inch diameter pipes.;Based on historic performance data, two statistical models were developed to predict the performances of different types of pipes in the system: the time-dependent exponential model and the non-homogeneous statistical model. The time-dependent exponential model includes time-dependent factors such as soil moisture content and temperature changes with time. The non-homogeneous statistical model includes static factors such as pipe lengths, pipe sizes, and buried depths (a surrogate parameter for the climates). The non-homogeneous model represents the relative risks of these static factors. Both models predict the behaviors of different types of pipes in the system very well in time scale. Since the water pipeline failures are mainly due to corrosion and cracking, the risk-based corrosion-fracture model was developed to represent the failures. Corrosion results in the reduction of effective wall thickness in the water pipes, hence corrosion is modeled in terms of crack development with time. The corrosion-fracture model predicts the failure trends observed in various water pipeline networks. |
