Enhancing inflow and infiltration detection in urban sewer networks with a new deterministic sensor placement method

Area of Science:

• Environmental Engineering
• Water Resource Management
• Network Optimization

Background:

• Inflow and infiltration (I&I) significantly impacts urban sewer network (USN) functioning.
• Optimal sensor placement is crucial for reliable I&I event detection but remains challenging.
• Existing heuristic search-based methods (HSBMs) yield variable sensor placement strategies (SPSs).

Purpose of the Study:

• To develop a novel deterministic two-stage clustering method for optimizing SPS based on information entropy.
• To improve the detection reliability of I&I events in USNs.
• To provide a consistent and optimal SPS for engineering applications.

Main Methods:

• A two-stage clustering approach utilizing Spectral Clustering based on joint entropy.
• Incorporation of topology structure properties for enhanced clustering and reliability.
• Optimization of SPS by maximizing joint entropy for sensor assignment within selected clusters.

Main Results:

• A strong correlation (R > 0.95) was found between joint entropy and SPS detection reliability.
• The proposed method achieved 7-15% higher detection reliability than two HSBMs.
• The developed method deterministically provides optimal SPS, outperforming HSBMs.

Conclusions:

• The novel deterministic two-stage clustering method offers superior performance for SPS optimization in USNs.
• Information entropy is a key factor in achieving high I&I detection reliability.
• This approach provides a reliable and consistent solution for practical engineering applications.