How To Predict Pipe Failures

Pipe failures, and the resulting loss of non-revenue water, have long plagued water systems all over the world. It can be a tough thing to watch as treated effluent spills underground or onto city streets. While no shortage of new technologies have come forth promising to curb the problem, over 400 water mains still break every day, according to WaterMainBreakClock.com.

Such an extensive problem doubtlessly requires a robust solution. For five years, the Water Research Foundation (WRF) and several international research partners were conducting a million-dollar research project to find a solution, “Advanced Condition Assessment and Failure Prediction Technologies for Optimal Management of Critical Water Supply Pipes.”

“There was general acknowledgement that our tools and technologies to make intelligent investment in [buried potable water] infrastructure were lacking,” said Frank Blaha, WRF regional liaison, about the impetus for the project. “The age of buried infrastructure has typically been used as a measure of its degree of deterioration, but it has repeatedly been found that age is a poor predictor of the actual condition of those buried assets. This research program was launched to improve our understanding of pipe deterioration and advance our technologies to accurately measure the true physical state of the pipe, and to help prevent pipe failures.”

Blaha estimated that pipe failure costs utilities over $1 billion per year and that complete replacement happens only once every 200 years, on average.

“By having improved failure prediction techniques, failures can be prevented and less time and money will go into responding to those breaks,” he said. “Those saved resources can be refocused on rehabilitation and replacement activities.”

The project was divided into three parts: failure prediction, condition assessment, and corrosion modeling. By building a comprehensive picture of the status of pipes, the research team could calculate when they will break and allow systems to take preventative measures. It ultimately reached a time-dependent probability method for failures along pipelines.

“The method improves the confidence in condition assessment and failure prediction, and primarily focuses on in-between interpretation of pipelines with different condition assessment techniques and soil data, emerging technologies, and probability of failure in order to address which sections of the pipeline may need to be replaced as a matter of priority,” said Jian Zhang, a research manager with WRF.

The probability of failure calculation determined by the researchers was based on a prediction of remaining life that took into account potential input of loading, the material properties, and corrosion progression. While other prediction models have already been developed, Zhang believes his team has made an improvement.

“Previously known methods focused on different aspects of condition assessment for estimating pipe life,” he said, “while this project integrates … the concept of leak before burst, a better understanding of the most widely used technologies currently employed for the condition assessment of critical cast iron pipes, and an effective predictive model for pipe corrosion into one comprehensive method. Moreover, the method has been verified with actual pipe sections provided by participating utilities.”

The model relies heavily on corrosion rates as an indicator of potential main failures, a problem that costs U.S. water and wastewater systems north of $50 billion per year, according to The Heartland Institute.

“The predictive model is based on the premise that corrosion of cast iron in soils should follow the same basic ideas as those for corrosion of cast iron and other ferrous metals in other natural environments, such as fresh water, immersion in seawater, and in the tidal and atmospheric zones,” said Zhang.

The researchers determined that while the corrosion rate for pipes is not constant, it does reach a predictable rate after 10 to 15 years. Since most cast iron water pipes are over 50 years old, they have reached this steady state, they concluded. They exhumed over 30 pipes, grit blasted and laser scanned them, and took soil samples. They found that perforation from outside the pipes allowed freshwater from inside to be forced out under pressure, causing more general corrosion and leading to larger breaks.

“It explains the time between the first occurrence of a leak and the eventual bursting of a pipe under pressure,” said Zhang. “A mechanism is now available to explain the time between first occurrence of a leak and eventual pipe failure through bursting under internal pressure.”

WRF expects a final report on the project to be released by the end of the year. In the meantime, one of the research partners, Australia’s Sydney Water, has funded a new project to implement the discoveries and figure out how they can best be applied by water utilities around the world.

“The Water Research Foundation will work closely with Sydney Water and the research team in many ways to provide practical methods and knowledge to water utilities in the United States,” said Zhang.