Stormnet: a connectionist model for dynamic management of wastewater treatment plants during storm events

Abstract

Models for solid transport in sewers during storm events are increasingly used. An important application of these models is the management of treatment plants during storm events so as to improve the quality of receiving waters. However, a major difficulty that prevents more general use of these tools is their calibration, which requires field data, accurate information about catchments and sewers, and a specific methodology. For that reason, a connectionist model called STORMNET has been designed to reproduce and replace usual conceptual and deterministic models. This model requires fewer data, can be automatically calibrated, and is comparatively simple. It is composed of two recurrent neural networks for the simulation of hydrographs and pollutographs of suspended solids, respectively. In this paper, we present an updated version of STORMNET designed for optimal management of wastewater treatment plants during storm events. This model has been validated using both model and real data. The results show the efficiency of STORMNET as a computational tool for simulating stormwater pollution.

Hacia un cambio en el paradigma del manejo y modelación de sistemas de drenaje urbano en países en desarrollo

“…During the last 30 years a number of sewer model-#30 revista de ingeniería dossier ling tools, including sediment modules, have been developed such as: Mosqito [68], Flupol [69], Mousetrap [70], Hypocras [72], Stormnet [73], Simpol [74], Sewsim [34,75,76], Horus [77,78], STSim [79], Remuli [80], and Cosmoss [81] among others. Despite the efforts that have been made to understand pollutant accumulation, erosion and transport dynamics, stormwater quality modelling still poses difficulties, and the generality and transferability of stormwater pollution models are still limited [82,83].…”

Including snowmelt in influent generation for cold climate WRRFs: comparison of data-driven and phenomenological approaches

“…[9][10][11] Therein, the artificial neural networks (ANN) are one of the most commonly used approaches, not only for forecasting the WRRF influent disturbance in view of improving wastewater treatment operation and real-time control, [12][13][14] but also for the simulation of rainfall-runoff and solids transport in sewers during storm events. 15 So far, research on IGs has not widely tackled WRRF operated under cold climate. Different research demonstrates that the snowmelt period at the end of winter causes an important increase in flowrate and a significant change in influent temperature and pollutant concentrations.…”

“…9–11 Therein, the artificial neural networks (ANN) are one of the most commonly used approaches, not only for forecasting the WRRF influent disturbance in view of improving wastewater treatment operation and real-time control, 12–14 but also for the simulation of rainfall-runoff and solids transport in sewers during storm events. 15…”