Model predictive control based on artificial intelligence and EPA-SWMM model to reduce CSOs impacts in sewer systems

Area of Science:

• Environmental Engineering
• Water Resource Management
• Urban Planning

Background:

• Urbanization and climate change increase precipitation intensity, overwhelming urban drainage systems (UDS) and causing combined sewer overflows (CSOs).
• CSOs lead to significant water pollution in urban environments globally.
• Existing UDS often lack the capacity to manage increased stormwater runoff effectively.

Purpose of the Study:

• To develop and evaluate a novel Model Predictive Control (MPC) system for mitigating CSOs.
• To optimize the operation of urban sewer systems for enhanced stormwater management.
• To reduce the environmental impact of CSOs through intelligent control strategies.

Main Methods:

• Implementation of a novel MPC framework integrating neural networks for flow prediction.
• Utilization of a Stormwater Management Model (SWMM) for simulating flow conveyance.
• Application of a genetic algorithm for optimizing sewer system operation and control strategies.
• Testing the model on the sewer system of Casablanca, Morocco.

Main Results:

• The developed MPC system demonstrated significant efficiency in reducing combined sewer overflows (CSOs).
• The proposed model achieved short optimization times, facilitated by parallel computing.
• Successful application and validation of the control strategy in a real-world urban sewer system.

Conclusions:

• The novel MPC approach offers an effective solution for managing CSOs in urban areas.
• Intelligent control strategies can significantly improve the performance of existing urban drainage systems.
• The integration of predictive modeling and optimization algorithms provides a robust framework for sustainable urban water management.