Application of Machine Learning in Sanitary Sewer Condition Assessment for the City of Raytown, Missouri

The City of Raytown needed an asset management strategy for its sanitary sewer system. Raytown’s sanitary sewer system serves 30,000 residents. The system includes 165 miles of gravity sewer lines, 4,325 manholes, two miles of force main, and two equalization basins, each with six million gallon capacity. There are no combined sewers. The City engaged with NEER.ai to use a machine learning (ML) solution to identify the risk condition. NEER created and applied several ML techniques (clustering, classification, and regression) to populate missing values and score likelihood of failure (LoF), consequence of failure (CoF), and overall risk score for each sanitary sewer asset. NEER.ai produced more than 90% accuracy in predicting the LoF score for sanitary sewer assets. NEER also calculated CoF and the overall risk of the entire system. The City of Raytown saved several thousand dollars and significant staff hours by leveraging ML to implement its sewer asset management program.