[Method for optimal sensor placement in water distribution systems with nodal demand uncertainties]

Area of Science:

• Environmental Engineering
• Water Resource Management
• Systems Optimization

Background:

• Effective sensor placement is vital for monitoring water distribution systems.
• Nodal demand uncertainties pose challenges to accurate event detection and source identification.
• Existing methods may not fully account for dynamic system conditions.

Purpose of the Study:

• To introduce and evaluate 'identification fitness' for optimizing sensor placement in water distribution networks.
• To assess the impact of nodal demand uncertainties on sensor placement strategies.
• To determine the trade-offs between detection probability and identification accuracy under uncertainty.

Main Methods:

• Utilized Nondominated Sorting Genetic Algorithm II (NSGA-II) to identify the Pareto front.
• Optimized sensor placement based on minimizing detection time overlap and maximizing detection probability.
• Incorporated nodal demand uncertainties into the optimization process.

Main Results:

• Nodal demand uncertainties had a limited effect on the probability of detection and the number of potential sensor locations.
• Source identification accuracy significantly decreased as nodal demand uncertainties increased.
• The proposed identification fitness metric effectively balanced competing objectives.

Conclusions:

• Sensor placement optimization using identification fitness is robust to moderate nodal demand uncertainties regarding detection.
• Accurate source identification in water systems is highly sensitive to demand fluctuations.
• Further research should explore advanced algorithms to mitigate the impact of demand uncertainty on identification accuracy.