Breaking Through the Noise: Enhancing Acoustic Leak Detection With Analytics

Advanced analytics has further improved acoustic leak detection accuracy. By integrating analytic 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.

Defining Leak Detection Accuracy

The generally accepted rate of accuracy for is around 70%. But what exactly does that mean? Accuracy in leak detection takes several factors into consideration:

• Reliability. "Noise" in the water distribution system, such as valves turning on and off and traffic above the pipe, can interfere with detecting leaks and cause false positives. The more accurate the system, the better it can distinguish the acoustic profile of an actual leak from other sounds. This means more reliable identification of leaks, which reduces time and money wasted digging up pipes for nothing. In addition to being costly, misidentified leaks are also frustrating for utility managers and operators—and can damage their trust in the system itself.

• Repeatability. When two different sensors indicate the presence of the same leak, it increases confidence in the result and the system at large.

• Precision. Pinpointing the location of a leak is just as important as distinguishing it from noise. A system that can identify the presence of a leak but consistently misjudges its location by 50 to 100 yards will, like a false positive, lead to excess digging, wasted operator time and increased cost as crews dig up more and more pipe in search of the true source.

Enhancing Leak Detection Through Analytics

A typical leak detection system relies on interpreting noise profiles commonly associated with leaks. 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 analyzing the frequency and amplitude of sounds over time. This reduces false positives.

Pattern Recognition

Advanced algorithms learn to recognize 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.

Maximizing Impact: Best Practices for Integrating Analytics

Clearly, analytics are a powerful aid to acoustic leak detection, but analytics alone are not enough. Utilities should take these additional key steps to ensure the accuracy of the system:

1. 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. Any inaccurate data regarding pipe sizes, materials and lengths can affect the analysis.

2. Optimize Sensor Placement

   The deployment location of the sensors can drastically impact the quality of the measured acoustic sample data. For example, those placed too close to service lines or consumption points will experience more background noise than those located further away.

3. Focus on High-Consequence Areas

   While NRW in any area of the distribution network is costly, some parts of the system are more critical to monitor. Hospitals, service lines buried under highways or cross riverways, and mains that serve major metro areas that have high criticality should be the primary focus. These points also deliver the highest return on investment (ROI) for NRW initiatives.

4. Collaborate with Vendors

   Many vendors of analytics and acoustic leak detection are willing to partner with utilities to interpret analytic results. In addition, they are often willing to help with considerations such as sensor deployment planning, incorporating GIS data and other application parameters, and more. Maintaining a leak detection program does not have to rest solely on operators' shoulders; external support from a supportive vendor is readily available.

Final Thoughts

The integration of advanced analytics can propel acoustic leak detection systems to another level. By leveraging noise filtering, pattern recognition and comprehensive data analysis, utilities can improve their ability to identify and locate leaks, minimizing NRW and reducing operational costs in the process. With a focus on accurate data input, strategic sensor placement and collaboration with technology vendors, water utilities can harness these supercharged systems to better manage their distribution networks.