A Two-Stage Data-Driven Spatiotemporal Analysis to Predict Failure Risk of Urban Sewer Systems Leveraging Machine Learning Algorithms

Risk-informed asset management is key to maintaining optimal performance and efficiency of urban sewer systems. Although sewer system failures are spatiotemporal in nature, previous studies analyzed failure risk from a unidimensional aspect (either spatial or temporal), not accounting for bidimensional spatiotemporal complexities. This is owing to the insufficiency of good-quality data, which ultimately leads to under-/overestimation of failure risk. Here, we propose a generalized methodology/framework to facilitate a robust spatiotemporal analysis of urban sewer system failure risk, overcoming the intrinsic challenges of data imperfections-e.g., missing data, outliers, and imbalanced information. The framework includes a two-stage data-driven modeling technique that efficiently models the highly right-skewed sewer system failure data to predict the failure risk, leveraging a bidimensional space-time approach. We implemented our analysis for Bogotá, the capital city of Colombia. We train, test, and validate a battery of machine learning algorithms-logistic regression, decision trees, random forests, and XGBoost-and select the best model in terms of goodness-of-fit and predictive accuracy. Finally, we illustrate the applicability of the framework in planning/scheduling sewer system maintenance operations using state-of-the-art optimization techniques. Our proposed framework can help stakeholders to analyze the failure-risk models' performance under different discrimination thresholds, and provide managerial insights on the model's adequate spatial resolution and appropriateness of decentralized management for sewer system maintenance.

[Substrate Flow by Different Biochemical Activities in the Urban Sewage Network]

In order to investigate the consumption mechanism of organic contaminants in different biological metabolism pathways within an urban sewage network, a pilot-scale system using two kinds of sediments (urban sewage sediment and artificial sediment) was constructed. The pilot system was operated to study the migration and transformation characteristics of COD, methane, nitrate, and sulfate between sediment and sewage. Results showed that the variation of COD in sewage was 170.58 mg·L^-1, with change of COD by deposition and bioreaction of 101.53 mg·L^-1 and 69.05 mg·L^-1, respectively. Due to biological metabolism, the generation of methane in sewers was of 7.39 mg·L^-1; the decrease of nitrate and sulfate in sewage was 0.33 mg·L^-1 and 21.35 mg·L^-1, respectively. Based on our calculations, the consuming concentration of COD was 32.51 mg·L^-1for methane generation, 8.04 mg·L^-1 for denitrification, and 6.41 mg·L^-1for sulfate degradation by sulfate reducing bacteria. The decrease in COD by deposition was responsible for 65.38% of total variation of COD in sewage. The decrease in COD by bioreaction was responsible for 34.62% of total variation in COD. Meanwhile, total variation values of COD for methane generation, denitrification, and sulfate degradation were responsible for 68.01% of COD variation by bioreaction. Deposition was therefore the main pathway for removal of organic contaminants from the sewer system; biochemical activities of methane generation, denitrification, and sulfate degradation also played important roles.

[Distribution Characteristics of Methanogens in Urban Sewer System]

A study was conducted through a 1200 m-PVC-pipe-reactor, which was used to simulate the urban sewer system. Gas chromatography, liquid chromatography and 454 high-throughput sequencing were utilized to study the variation of substrates during the methanogenic process and the distribution characteristics of methanogens in the sewer system. The results showed that the concentration of methane increased along the sewer system, which illustrated that methanogens existed in the sewer network. The methanogens mainly contained Methanosarcina, Euryarchaeota_unclassified and Methanobacteriaceae_unclassified. The distinct succession which Euryarchaeota_unclassified replaced Methanosarcina to be the first dominant microbial genus between 800-1000 m of the sewer system. Formic acid, methanol, methylamine, acetic acid and hydrogen were available substrates for methanogens. Among these substrates, acetic acid was the primary substrate for methanogen. The variation trends of these substrates were first increasing and then decreasing along the length of the sewer system, which led to the succession phenomenon of methanogens in the sewer system.