First things first clean up the mess.
There are basically two directions in addressing combined overflow problem: First is to expand the sewer system’s capacity to deal with more water by adding new green/grey infrastructures. Second is to optimize the use of existing facility like installing the smart real-time flow control control system. One not so sexy but truly effective aspect of optimizing existing sewer capacity is cleaning it.
In one case study of city of Camden, it is reported that the modeling of the camden sewer system predicts that 90 percent of the flooding problem in a normal year of rainfall could be eliminated with proper cleaning and maintenance of sewer system. It is not easy to inspect and maintain hundreds of miles of sewer system. Two interesting numerical methods have the potential to help to reduce the cleaning task to minimum cost.
One method uses graph theory and existing sewer system structure data to identify the most critical parts of the sewer networks. It therefore can prioritize the most important elements of sewer networks for inspection. Another method uses neural networks to accurately classify the pipe defects in image data collected during inspection. This way sewer condition could be assessed more objectively and and quickly. Since these two methods are based on abstract math algorithm, they can be easily applied to any sewer system that has the proper data they need.
References
[1] Andrew,Kricun (March 2017)Impact of Proper Maintenance of Combined Sewer Overflow System on Flooding in the City of Camden, retrieved on Septemember 30, 2019, from https://www.jerseywaterworks.org/wp-content/uploads/2017/11/CCMUA-case-study-Fact-Sheet.pdf
[2] Dulcy M. Abraham, UTILIZING NEURAL NETWORKS FOR CONDITION ASEESSMENT OF SANITARY SEWER INFRASTRUCTURE, retrieved on Septemember 30, 2019, from https://pdfs.semanticscholar.org/5700/ffa634b7c4d31abd48148c3ed96201ff4203.pdf
[3]Didrik Meijer(January, 2018) , Identifying Critical Elements in Sewer Networks Using Graph-Theory, retrieved on Septemember 30, 2019, from https://pdfs.semanticscholar.org/c2ee/ec3c10dd3956dd9a2b05988964e88776d9d0.pdf
