Artificial intelligence supporting sustainable district heating
Artificial intelligence supporting sustainable district heating
Leading Italian multi-utility optimizes its networks with Exprivia: a tool for anomaly detection and predictive analysis
Exprivia has recently launched a collaboration with one of Italy’s leading multi-utility companies, active in the energy, environment, and water service sectors, with a strong presence across the national territory. Among its most strategic assets are district heating networks, which in several cities provide heat to homes while reducing the use of domestic boilers and promoting a more efficient and sustainable energy model.
However, the complexity of this type of infrastructure, vast and largely underground, exposes it to the risk of leaks along the pipelines, resulting in energy losses, high costs, and maintenance challenges.
The need
The industrial player needs to implement an advanced system capable not only of detecting operational anomalies and leaks within one of its district heating networks, but also of predicting potential future issues. This can be achieved by analyzing thermal images acquired annually through aerial surveys and integrating them with other features/data sources within an anomaly detection and predictive analysis tool. The goal is to understand the evolution of the network and its anomalies over time, in order to intelligently plan maintenance activities and reduce waste.
Exprivia's solution
Exprivia is supporting its partner in the industrialization of a digital platform based on artificial intelligence (deep learning) models for anomaly detection.
Through the collection and mosaicking of thermal images from annual aerial surveys, the system builds a progressive dataset in which anomalies are compared with verified field leak cases. The neural networks thus learn to distinguish between physiological conditions or background noise and truly significant anomalies.
The result is an intelligent map capable of ranking and prioritizing anomalies according to severity, extent, and potential impact — a valuable strategic decision-making tool for the multi-utility. Moreover, diachronic monitoring makes it possible to observe how anomalies evolve over time, transforming a traditionally reactive approach into a predictive and proactive model.
Benefits
The platform will make it possible to proactively identify critical points in the network that are potentially exposed to anomalies or failures, and to optimize maintenance planning, with a positive impact both economically and environmentally. In this way, the multi-utility can reduce energy losses and improve the overall efficiency of the network, strengthening its role as a key player in the energy transition of the areas it serves.
