Genetic Algorithm Leakage Detection in Hydraulic Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current leakage detection methods in water distribution systems are inefficient and costly, particularly in highly looped networks with many valves and pumps, as they struggle to accurately identify leakage nodes due to damping of transient waves and difficulty in distinguishing leak reflections from demand changes, and existing techniques are not effectively applicable to large-scale systems.
Innovation Solution
The implementation of model-based pressure-dependent demand optimization techniques using a genetic algorithm to optimize emitter coefficients and locate leakage nodes in hydraulic simulation models, allowing for the representation of leakages as pressure-dependent emitter flows, which minimizes the difference between model-predicted and field-observed data for pressure and flow.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If inverse transient analysis is used for leakage detection, then leakage identification capability is improved, but applicability to large-scale looped networks deteriorates due to damping of transient waves
Solution Approach 1:
The patent segments the water distribution network into multiple zones or sub-networks, each analyzed independently using inverse transient analysis. This allows the method to be applied to large-scale systems by breaking them into manageable sections where transient waves remain detectable, thus resolving the contradiction between maintaining measurement precision and adapting to large-scale networks.
Solution Approach 2:
The patent transitions from time-domain transient analysis to frequency-domain spectral analysis by applying Fourier transforms to pressure signals. This dimensional change from temporal to spectral domain allows leakage detection to work effectively in looped networks where transient waves are damped, as spectral analysis can identify leakage signatures across a range of frequencies rather than relying on single transient events.
2Measurement precision
If step-testing is used for leakage detection, then leakage detection accuracy is improved, but operational convenience deteriorates due to supply interruptions
Solution Approach 1:
The patent replaces the mechanical step-testing procedure (which physically isolates and pressurizes sections of the network) with a computational method using spectral analysis of continuous pressure measurements. This substitution eliminates the need for operational disruptions while maintaining leakage detection accuracy, as the mathematical analysis can identify leaks from normal operating pressure fluctuations.
3Measurement precision
If acoustic loggers are deployed extensively for leakage detection, then leakage detection coverage is improved, but cost increases due to equipment quantity
Solution Approach 1:
The patent makes existing pressure sensors and flow meters multi-functional by using them not only for their primary monitoring purposes but also as sources of data for leakage detection through spectral analysis. This universality eliminates the need for separate acoustic logger equipment, achieving comprehensive leakage detection coverage without increasing equipment quantity.
Solution Approach 2:
The patent creates a virtual model or digital twin of the water distribution network that replicates its hydraulic behavior. By analyzing pressure and flow data through this virtual model using spectral methods, the system achieves extensive leakage detection coverage without deploying physical sensors throughout the entire network, thus reducing equipment quantity while maintaining detection capability.
Data Source
AI summary
In one embodiment, a hydraulic simulation model corresponding to a real-world hydraulic network is loaded in a hydraulic modeling and simulation application executing on a computer system. The hydraulic simulation model represents leakages as pressure dependent emitter flow at selected nodes (leakage nodes). Optimization criteria include a specified maximum of possible leakage nodes. A genetic algorithm (GA) generates trial solutions for an optimization, each trial solution representing locations for leakage nodes and corresponding emitter coefficients. A hydraulic analysis is performed for the trial solutions to generated model-simulated results. The model-simulated results are compared to field-observed data for the real-world hydraulic network to generate goodness-of-fit values. The process is repeated until a particular goodness-of-fit value is achieved or a maximum number of iterations is reached. Predicted locations for leakage nodes from a particular selected trial solution are then displayed to a user.