The Great Lakes Water Authority (GLWA) is a premier North American water utility that is responsible for managing the large and complex water transmission system that serves 112 communities across the southeast Michigan region, including the City of Detroit. These communities represent approximately 40% of the state’s population making GLWA the largest water utility in Michigan and one of the largest in the United States. To reliably deliver its quality water to regional customers, GLWA oversees over 816 miles of water transmission mains that delivers 1,720 million gallons of treated water daily. The GLWA transmission system was constructed over a 170-year period with large segments of the system installed during period of population growth in the region, especially in the first two decades of the twentieth century and immediately after World War II. Over that time, a wide variety of pipe types were used in the system design including prestressed concrete cylinder pipe (PCCP), cast iron (CI) pipe, steel (ST) pipe, reinforced concrete (RC) pipe, and ductile iron (DI) pipe. PCCP pipe is most common in the system representing 47% (389 miles) of linear assets with CI and ST pipe the second and third most at 20% (165 miles) and 15% (125 miles), respectively. Overall, the average age of pipe in the system is 70 years; with age comes deterioration. Hence, asset management of the system is essential to managing the risks of failure.

GLWA plans to develop a detailed plan to assess the condition of the transmission system and to develop a comprehensive risk management strategy to guide its decisions on inspection, maintenance and replacement. The need for a risk management strategy is evident given the age of the transmission system and a history of failures including the 2017 and 2021failures of PCCP assets along 14 Mile Road. In fact, 71% of the PCCP pipes in the GLWA system were fabricated in the 1960’s and 1970’s when the use of high-strength prestressing wires in PCCP resulted in more efficient pipe designs; however, time has proven these pipes to lack durability resulting in unusually high distress rates. GLWA Linear System Integrity Program (LSIP) will be responsible for developing and implementing a risk-based decision-making framework for its water transmission system linear assets. The LSIP program aims to implement a proactive strategy to maintaining the transmission system with data integral to the decision-making process. In addition to its evolving risk management strategies, the authority also has a 5-year capital improvement plan that aims to upgrade aging and high-risk segments of its water transmission system as a means of managing risk at a reasonable cost. LSIP aims to influence future capital improvement plans by setting investment prioritizations based on quantitative assessment of changes in component and system risk resulting from investment.

The state-of-practice in risk management of water transmission systems varies from utility to utility but can generally be described as being based on the use of population statistics of condition data to estimate damage accumulation as a function of pipe vintage and time; in turn, this allows the remaining asset life of aging pipe segments to be estimated. The data available to utilities to aid their decision making is based on inspection data and historical tabulations of failure. To collect data on the condition of the assets in a water transmission system, inspectors require access to the internal area of the pipes for visual and nondestructive evaluation (NDE) inspection with many methods that require dewatering of the pipe which can be inconvenient and costly. A variety of approaches are readily available with some specific to the pipe type. For example, visual inspection by an inspector or use of a video camera snaked into the system can offer qualitative evidence of pipe distress. Metallic pipe types can also have their wall thicknesses measured using ultrasonic NDE methods. PCCP pipes have benefited immensely from electromagnetic NDE inspection methods based on Eddy currents to assess the breakage of the prestress wires embedded in their walls. In some instances, more permanent monitoring solutions can also be adopted including acoustic fiber optics (AFO) that capture acoustic events from structural deterioration processes such as wire breakage in PCCP pipes. All of these methods have inherent costs associated with them in terms of acquisition of the data (e.g., equipment and labor costs) but also in the form of consumed life due to dewatering requirements for some of the methods. These costs demand utilities to perform a cost-benefit analysis to justify their adoption in terms of reducing uncertainty in their decision-making processes. In addition, the cost of these approaches makes their ubiquitous use infeasible; utilities must judiciously determine when and where to apply these methods to aid their decision making.

Once data is collected, there is an opportunity to aggregate data across a number of large utilities to begin to understand the performance of pipes based on their type and vintage. As previously mentioned, population statistical analysis has been extremely beneficial to highlighting problematic pipe types and to give utilities a more quantitative basis for prioritizing inspection, renewal and replacement. For example, the chronic poor performance of PCCP pipes segments manufactured in the 1970s has been well documented based on tabulation of PCCP conditions across the United States. While statistical analysis over a large number of utilities helps utilities make better decisions, there are some limitations to this approach. Specifically, statistical analyses tend to represent the “average” performance of pipe vintages and may not reflect the actual condition of the pipes in current use. In particular, it implicitly assumes the environmental and operational profile of one utility is, on average, the same as another which may be questionable in some instances.

Hence, there is an opportunity to extend current practices to render an asset management framework complementary to existing methods but one that derives greater precision based on expanded use of data and modeling. This proposal will explore a formal framework to enable GLWA to make better asset management decisions that lead to safer and more cost-efficient operation of the regional water transmission system. This proposal builds off of the team’s prior work with GLWA on developing a reliability assessment methodology to assess the probability of failure of PCCP pipe in the system using operational data such as steady state and transient pressure data and condition assessments done by contractors using electromagnetic methods. In the previous project, the team made a detailed analysis of load demand (D) in the transmission system including spatial statistical characterization of steady-state service and transient event pressures in the regional system. The team also developed a novel structural capacity (C) model approach using two- and three-dimensional elasticity models to predict structural response of pipes to the pressures observed in the system. These models can elegantly use input condition data coming from system inspections to accurately model the future behavior of pipe and predict remaining service life spans. A reliability methodology was previously proposed to calculate the probability of exceeding a defined limit state (which in essence defines “failure”) such as exceeding yield or cracking stresses in pipe materials. An attractive feature of the approach is that it calculates probability of failure in the form of a scalar reliability index (b) associated with a definable limit state function – this approach is consistent with load and resistance factor design (LRFD) codes used by the pipe industry (e.g., AWWA C301 & C304 pipe) thereby linking design and post-design asset management in the same decision framework. The team demonstrated the calculation of PCCP pipe reliability using steady-state and transient pressure data collected by GLWA from its transmission system and assumed corrosion levels of the prestressing wire wrap. A key innovation of the framework was the development of a distributed water hammer model to spatially map pressure transients measured by GLWA at specific pump stations to estimate the impact of those transients at each pipe location in the system.

This project builds on the team’s prior work in developing statistical spatiotemporal models of system pressures and creation of a robust reliability-based framework for assessing the residual capacity of system pipes.  The proposal will advance its predictive capabilities to help GLWA quantify how decisions made in inspecting and renewing pipe segments changes the risk profile of the transmission system.  The team specifically aims to work in direct collaboration with the GLWA LSIP team to contribute expertise in applying data science to model the operational pressure environment in the GLWA transmission system and using pressure data to more quantitatively assess pipe reliability as part of a comprehensive asset management framework. To further enhance the capabilities of the LSIP program, the proposal team will install a permanent monitoring system on a specific segment of the transmission system to assess in situ pipe behavior under pressure and to validate analytical models used in risk assessments.