Using data from monitoring combined sewer overflows to assess, improve, and maintain combined sewer systems

Abstract

Using low-cost sensors, data can be collected on the occurrence and duration of overflows in each combined sewer overflow (CSO) structure in a combined sewer system (CSS). The collection and analysis of real data can be used to assess, improve, and maintain CSSs in order to reduce the number and impact of overflows. The objective of this study was to develop a methodology to evaluate the performance of CSSs using low-cost monitoring. This methodology includes (1) assessing the capacity of a CSS using overflow duration and rain volume data, (2) characterizing the performance of CSO structures with statistics, (3) evaluating the compliance of a CSS with government guidelines, and (4) generating decision tree models to provide support to managers for making decisions about system maintenance. The methodology is demonstrated with a case study of a CSS in La Garriga, Spain. The rain volume breaking point from which CSO structures started to overflow ranged from 0.6. mm to 2.8. mm. The structures with the best and worst performance in terms of overflow (overflow probability, order, duration and CSO ranking) were characterized. Most of the obtained decision trees to predict overflows from rain data had accuracies ranging from 70% to 83%. The results obtained from the proposed methodology can greatly support managers and engineers dealing with real-world problems, improvements, and maintenance of CSSsSupport for this project was provided by the ENDERUS Project (CTM-2009-13018) and the predoctoral grant FPI BES-2010-039247, founded by the Spanish Ministry of Science and Innovation. Lluís Corominas received the postdoctoral Juan de la Cierva grant (JCI-2009-05604) from the Government of Spain and the Career Integration Grant (PCIG9-GA-2011-293535) from the EU. The authors also acknowledge the support from the Economy and Knowledge Department of the Catalan Government through the Consolidated Research Group (2014 SGR 291)–Catalan Institute for Water Researc

Automated Model Construction for Combined Sewer Overflow Prediction Based on Efficient LASSO Algorithm

“…To get a predictive model with a size of one, the correlations between these candidate regressors and the CSO output are used to initialize the two vectors c 0) and b 1) ; consequently, the regressor leading to the largest absolute correlation is selected and added into the predictive model. Correspondingly, the variables ρ 1 , 1 , A 1) , b 1) , d 1) , k,θ 1,1 , and γ 1 are computed in sequence (where γ • 1 is assigned with zero in order to initiate the model learning process) to prepare the computing framework for locating the next LASSO solution.…”

“…The whole algorithm can be terminated by designating a specified number of model regressors first reached during the model learning process or using other criteria such as Akaike information criterion (AIC); thereby the selected model regressors and associated coefficients are retrieved. 1) , and k ← 1 in sequence. 3…”

“…from precipitation events and sewage/wastewater from domestic, industrial, commercial and municipal release together in the same sewer [1]- [3]. The wastewater gathered in the CSS is then directed to wastewater treatment plants (WWTPs), usually driven by gravity through paved inclined sewers together with a small number of lift pumping stations to assist water transportation between sewers.…”

A robust and accurate surrogate method for monitoring the frequency and duration of combined sewer overflows

“…The information is also useful to assess, improve and maintain combined sewer systems (e.g. Montserrat et al 2015 ) as well as calibrating hydraulic urban drainage models (Duchesne et al 2001 ; Montserrat et al 2017 ).…”

Pre-Conditioning Approach to Bayesian Decision Networks for Water Quality Sensors Positioning in Urban Drainage Systems

“…For this reason, models for the characterization of the wastewater have been widely studied for assessing the pollution load overflowed and/ or transferred to WWTP. Montserrat et al (2015) developed a methodology to evaluate the performance of combined sewer systems (CSS) using low-cost monitoring; in this case, the collection and analysis of real data is indispensable to assess, improve, and maintain CSSs in order to reduce the number and impact of overflows. Irvinea et al (2011) have implemented a project with municipalities in Western New York State to evaluate the low-cost options for illicit discharge trackdown.…”