Machine learning and swarm intelligence for dynamic control of collection systems

Recent technological advances in wireless sensor network technology and cloud computing has allowed utilities to optimize the operation of their collection systems. Low cost level and flow sensors powered by Internet of Things technology are being used to better understand the hydraulic performance of collection systems in real time. Also, machine learning and big data optimization techniques are being used to effectively merge real time sensor data and the hydraulic and hydrologic model into a single tool that can be used to dynamically control collection systems. This concept, called Real Time Decision Support Systems, or RT-DSS, effectively enables utilities to leverage both the hydraulic model and real time data to optimally operate and maintain the sewer system. This presentation will show how the EPA SWMM has been augmented with neural networks that utilize sensor data to model the hydraulics in a system and advanced control algorithms inspired on swarm intelligence to enable dynamic control of collection systems.