Introduction
Addressing the issue of water leaks and gas fugitive emissions using an innovative solution that combines Artificial Intelligence (AI) and IoT technology to minimize waste and optimize maintenance activities. This is the mission of Onyax, an innovative company active in the field of telecommunications and remote industrial monitoring, capable of providing an effective and advanced digitalization response to the needs of water and gas distribution network managers. Thus, the EVALD project was born, leveraging the great AI potential of Onyax’s cloud platform ACE, already used by several utilities nationwide to remotely monitor their plants, and the support of an IoT sensor network with electro-vibro-acoustic technology, designed with an integrated hydrophone, autonomous battery, compact size, and instant installation. A comprehensive project aimed at sustainable and effective innovation in managing two fundamental resources for our territory.
Challenges
The challenges Onyax faces in Leak and Anomaly Detection involve issues related to operational simplification and data acquisition and processing technology. Onyax’s first challenge began by considering the diversity and vastness of distribution infrastructures where highly scalable and adaptable solutions must be applied to different technical contexts. This is combined with the need for the most sensitive possible acoustic and vibrational analysis with noise-free signals, thanks to a comprehensive experimental process based on IoT logic, and minimizing “false positives/negatives” through the development of reliable machine learning algorithms applicable on Onyax’s ACE platform. Additionally, long-term sustainability focused on low-power is of great importance, supported by Onyax’s continuous research operations aimed at raising awareness of minimizing water waste and reducing gas emissions in city and industrial pipeline systems.
The solutions
The EVALD solution, an acronym for Electro-Vibro-Acoustic-Leakage-Detect, is designed to integrate data from IoT devices, deep learning algorithms, and numerical modeling techniques into a single digital ecosystem, analyzing acoustic, pressure, electrical, and vibrational parameters. The idea behind the EVALD project was not to limit itself exclusively to the identification of acoustic anomalies through a passive approach, as in most traditional leak detection systems currently on the market, but to consider an “active” acoustic analysis supported by various electrical, pressure, and vibrational sources.
The “active” listening approach used by Onyax and refined through a valuable technological collaboration with the Department of Electronics, Information and Bioengineering at the Politecnico di Milano, consists of deliberately injecting a sound into the water/gas network and, through the IoT device TUBE-T3 with an integrated hydrophone, listening to it at different points to derive the average flow, thus analyzing the frequency deviation of the received signal while excluding possible environmental noise. This Audio Analysis technique, combined with the quantities related to the network nodes (pressure, temperature, and vibration) already acquired in the distribution networks through the installation of Onyax IoT devices, allows for the creation and subsequent “progressive training” of the digital twin of the network, built using artificial intelligence models, to identify, with higher precision, the areas with the highest probability of leakage or the so-called “false positives.” Finally, all collected data is sent to the cloud platform, where it is analyzed and visualized through intuitive interfaces, enabling managers to intervene promptly.
The advantages
The advantages of this new solution that combines IoT and artificial intelligence are primarily noticeable in the ability to reduce the error rate in leak detection, allowing managers to carry out targeted measurement campaigns even before deploying entire teams dedicated to acoustic field analysis. Additionally, the following benefits are provided:
- Energy Efficiency: Installation of devices with long-lasting batteries and low power consumption with minimal intervention actions by operators and a consequent sustainable approach.
- Technological Improvement: Data constantly received from the installed devices allows for increasingly precise analysis and, in fact, a more accurate realization of the digital twin of the physical network.
- Continuous Monitoring: The use of advanced IoT sensors allows continuous monitoring of the network, ensuring a constant view of the infrastructure’s state and enabling targeted interventions in case of anomalies.
- Scalability: The EVALD solution is scalable and can be easily adapted to distribution networks of different sizes and complexities.
- Predictivity: With networks made of steel and protected with cathodic protection techniques, in addition to evaluating the actual protection of water and gas pipelines, for example, through the BLACKBOX-CAT device, predictive techniques can be applied to both the lifespan of sacrificial anodes and understanding if a pipeline is starting to undergo a process of reduction, coating removal, or corrosion, all indicators of future leaks.
