Integrating physically based simulators with Event Detection Systems: Multi-site detection approach

Area of Science:

• Control theory
• Water resource engineering
• Environmental monitoring

Background:

• Fault detection (FD) is crucial for system monitoring.
• Existing event detection systems (EDSs) often use signal processing, analyzing monitoring stations independently.
• Model-based FD, using system simulations, offers an alternative approach.

Purpose of the Study:

• To compare the performance of model-based EDS with signal processing-based EDS for water distribution systems.
• To demonstrate the advantages of a Multi-Site EDS (MSEDS) for simultaneous network-wide analysis.
• To highlight the potential of physically based simulation models in contamination event detection.

Main Methods:

• Development and application of a model-based EDS utilizing water quality and hydraulics simulation models.
• Implementation of a Multi-Site EDS (MSEDS) for integrated analysis of data from multiple monitoring stations.
• Comparative analysis of detection accuracy and time between model-based and signal processing-based approaches.

Main Results:

• Model-based EDS demonstrates superior performance compared to signal processing-based EDS.
• MSEDS achieves higher detection accuracy, including more true positive alarms and fewer false alarms.
• MSEDS significantly reduces the detection time for contamination events.

Conclusions:

• Physically based, model-based EDS are more effective for detecting contamination events in water distribution systems.
• Integrated, multi-site analysis within an MSEDS framework offers substantial improvements in event detection capabilities.
• The study advocates for the adoption of model-based and multi-site approaches for enhanced water safety and security.