Enhancing Acoustic Leak Detection with Analytics

A typical leak detection system relies on interpreting noise profiles commonly associated with leaks. Advanced analytics has further improved acoustic leak detection accuracy. By integrating analytics with acoustic leak detection, water utilities can heighten the accuracy of these systems to quickly—and correctly—identify and locate leaks, providing valuable insight when it comes to the effective management and maintenance of the distribution network.

While routine inspections only offer a snapshot of pipeline conditions at a single point in time, receiving high-resolution and high-frequency data at regular intervals provides a more accurate picture of pipeline health.

Having an accurate system means a more reliable identification of leaks, which reduces operational costs and minimises non-revenue water. The generally accepted rate of accuracy is around 70%. Accuracy in leak detection takes several factors into consideration: reliability, repeatability and precision. Analytics takes this power to the next level in several ways:

Noise Filtering

Analytics can distinguish between true leak sounds and background noise (e.g., traffic, machinery, or water flow) by analysing the frequency and amplitude of sounds over time reducing false positives.

Pattern Recognition

Advanced algorithms learn to recognise the unique acoustic signatures of different types of leaks in different environments. By comparing measured acoustic data to known leak profiles, analytics help detect leaks more reliably.

Historical Data

Although there can be occasional inaccuracies when first starting out with leak detection analytics, the system learns over time. As mistakes are flagged, the system adapts, making increasingly accurate determinations as the algorithm matures.

Integration of Additional Parameters

In addition to sound profiles, analytics can incorporate other parameters like flow rates, transient events and network information, such as pipe material characteristics. By correlating these additional data points, the software can more confidently identify leaks and their locations.

Trend Monitoring

Not all leaks pop up overnight; most happen gradually. Analytics can track changes in acoustic patterns over time, catching slow-developing leaks that might otherwise go unnoticed, thus improving early detection.

How to Integrate Analytics Properly

Although powerful, analytics alone are not enough. To maximize impact and system accuracy, consider these best practices: 

• Include up-to-date information: When deploying analytics for any project, it is critical to frontload the software with the most up-to-date and accurate GIS information available about the network.
• Optimize sensor placement: Position sensors strategically across your network for optimal performance so that readings will be able to isolate service line disruption easily.
• Focus on high-consequence areas: Although NRW in any area of the distribution network is costly, some parts of the system are more critical to monitor. Service lines buried under motorways or mains that serve major metro areas that have high criticality should be the primary focus. It’s also worth noting, these points typically deliver the highest return on investment for NRW initiatives.
• Collaborate with vendors: Maintaining a leak detection program does not have to rest solely on your shoulders; external support from a supportive vender is readily available. They are often willing to help with sensor deployment planning, incorporating GIS data and other application parameters.

Integrating advanced analytics can propel acoustic leak detection systems to another level. By leveraging noise filtering, pattern recognition and comprehensive data analysis, you can improve your ability to identify and locate leaks all while minimizing NRW and reducing operational costs. Learn more about leak detection solutions and BlueEdge™ from Badger Meter.