Application Development for Objective-Based Asset Management of Wastewater Distribution Networks

The last major update to the United States wastewater infrastructure was implemented with the Clean Water Act of 1972 (CWA). Being 50 years from this update, most assets in current wastewater distribution networks (WWDN) are reaching the end of their service life. The industry itself is chronically hampered by a lack of adequate funding to rehabilitate assets needing attention, and inspections are often too expensive or not physically accessible in some parts of the WWDN. Consequently, the American wastewater infrastructure has received poor ratings from the American Society of Civil Engineers (ASCE), currently a “D+”. To raise the quality of the WWDN, assets must be regularly inspected and systematically replaced when required. However, these inspections and replacements are often limited by budget restrictions, forcing utility companies into a reactionary approach. Thus, actions are taken only when an asset failure occurs, resulting in large amounts of capital spending for the replacement and cleanup by the utility company. Therefore, there is a need for a more efficient way to conduct asset management.

To this end, we propose a framework to minimize the inefficiencies in current wastewater asset management plans by optimizing inspection schedules and predicting asset failures. We have developed an objective-based asset management approach as well as a software tool utilizing machine learning (ML) to assess the condition of WWDN infrastructure as well as prediction of the list of corresponding corrective actions. Using the Python programming platform, we employed the Random Forest Classification (RFC) to evaluate correlations in the data among several degradation and impact factors. Utilizing the data supplied by the Atlantic County Utilities Authority in NJ our tool can accurately evaluate and predict potential pipeline failures with high precision and provide a ranked list of contributing factors.