Anna, an Example of AI Application in Leak Detection, Localization, and Sizing in Water Distribution Networks

2024-04-16

Innovative Pilot to Maximize the Efficiency of Water Leak Detection in Potable Water Distribution Networks


Aqlara, a company managing potable water supply services and wastewater systems, in collaboration with Adasa Sistemas, an engineering firm specializing in developing and implementing digital solutions for the complete water cycle, have carried out a leak detection project in the water distribution network of the municipality of Anna using artificial intelligence (AI) based on FIDO Tech technology.

This initiative has allowed Aqlara to test this disruptive technology for detecting, locating, and sizing leaks, where the AI component plays a pivotal role in enhancing traditional technology that uses acoustic sensors.

FIDO Tech utilizes machine learning algorithms, resulting in continuous improvement as it:

Collects and verifies information files.
Filters out noises caused by other urban sources like traffic, valves, pumps, etc., which can generate false alarms.
The project involved analyzing the urban core of Anna, covering three sectors with a total network length of 16.5 km.

Two Phases

During the first phase, a total of 220 acoustic sensors known as “bugs” were deployed over a period of one week. The data collected by these sensors enabled the creation of a comprehensive map pinpointing leak-free zones and critical areas.

Acoustic Sensor or "Bug"

The AI has optimized the analysis of areas likely to contain potential leaks by clustering detected points, which will be further studied in the project's second phase.

 

Global Map Indicating the Absence and Possible Presence of Leaks After the First Deployment Stage

The second stage focused on sites identified with potential leaks using correlation techniques and field marking. It also categorized the size of water losses as small, medium, or large.

Conclusions

The methodology implemented in this project has enabled:

The elimination of potential problem areas in the distribution network after confirming them as leak-free zones.
The localization of several detected leaks that were hidden.
The identification of existing issues in different network components, such as valves and fire hydrants, which were not in proper working condition.

The value and differentiating contributions of this service are based on:

  1. Non-intrusive technology: installation on the surface of the material, avoiding damage to existing structures.
  2. Applicability on any pipe material, equipment, or accessory: versatility and adaptability in each situation.
  3. High precision and reliability: reducing human errors and false alarms and ensuring accurate leak detection.
  4. Leak size classification: prioritizing management according to their magnitude and optimizing resources.

With the launch of this service in the Spanish market aimed at maximizing water efficiency, Adasa Sistemas:

  1. Provides customers with an end-to-end leak detection, localization, and sizing service in the water distribution network, or segments of it, using disruptive acoustic technology backed by AI power.
  2. Reaffirms our commitment to continuous improvement in preserving a resource as scarce as water.
  3. Contributes to the advancement of digitalization in the water sector through innovative technological solutions.