Optimal sensor placement for leak location in water distribution networks: A feature selection method combined with...

Area of Science:

• Hydraulic Engineering
• Network Monitoring
• Data Science

Background:

• Real-time monitoring of smart water networks relies on Supervisory Control and Data Acquisition (SCADA) systems and strategically placed sensors.
• Current sensor placement methods often depend on well-calibrated hydraulic models, which can be computationally intensive and costly.
• Efficient sensor placement is crucial for accurate leak diagnosis and operational cost management in water distribution networks (WDN).

Purpose of the Study:

• To propose an efficient Optimal Sensor Placement (OSP) method for Water Distribution Networks (WDN) that minimizes reliance on hydraulic models.
• To develop a novel approach for node selection in WDN sensor placement using feature selection and graph signal processing.
• To evaluate the effectiveness of the proposed OSP method in terms of monitoring accuracy and node selection efficiency.

Main Methods:

• A heuristic optimization algorithm is employed for the sensor placement process.
• User nodes in the WDN are treated as features, utilizing a feature selection algorithm combined with graph signal processing.
• Graph signal reconstruction is applied to estimate pressure data across all nodes based on data from selected sensors.

Main Results:

• The proposed OSP method effectively analyzes node importance and considers redundancy during iterative sensor selection.
• Graph signal reconstruction demonstrates the relationship between the number of sensors and data reconstruction error.
• Testing on benchmark networks shows the OSP method's applicability and scalability to different WDNs.

Conclusions:

• The novel OSP method offers an efficient alternative to traditional model-dependent sensor placement techniques.
• The approach achieves superior monitoring outcomes and rapid node selection compared to existing methods.
• The proposed technique is adaptable and can be readily extended to various water distribution network configurations.