Translation. Region: Russian Federation –
Source: Novosibirsk State University –
An important disclaimer is at the bottom of this article.
Employees Center for Artificial Intelligence of Novosibirsk State University (CAI NSU) Received a state registration certificate for a mathematical model of a heating network designed to analyze and optimize the operation of heating systems. The new tool will enable more accurate predictions of emergency situations in infrastructure operation, reduce energy losses, and improve the efficiency of utility system management.
The development is a so-called model heating network—a digital object that can be used to test various calculation and control algorithms without risking real-world assets. This model is not tied to a specific city, but reproduces the key characteristics of real-world heating systems.
"This is a kind of virtual testing ground where we test various approaches to calculating and analyzing heating networks. It allows us to study processes that are difficult or expensive to investigate in real-world conditions, as well as conduct initial testing of the solutions our developers are developing," explained Alexander Dekterev, head of the Krasnoyarsk branch of the Institute of Thermophysics SB RAS and senior researcher at the NSU Center for Innovative Research.
Furthermore, using this model network, scientists create extensive databases reflecting the system's behavior under various scenarios—from changing weather conditions or network loads to various emergency situations. This data is then used to train neural networks, which can then quickly predict the state of the heating network and offer optimal solutions to operators.
"As we know, before using a neural network on real utility infrastructure facilities, where it can quickly assess how the system is changing and the consequences of such changes for both the heat supplier and the consumer, it needs to be trained. Our model network is well suited to this task, as it has sufficient databases of both normal operating conditions and emergency scenarios," Dekterev noted.
We previously reported that the NSU Artificial Intelligence Center, in collaboration with the Federal State Unitary Enterprise "Department of Energy and Water Supply," is implementing a project to digitalize the heating system in the Sovetsky District of Novosibirsk. As part of this project, a digital model of a specific network section is being created (Figure 2), which will enable real-time monitoring of operating parameters, leak detection, and prediction of potential accidents.
Such solutions are particularly in demand in the housing and utilities sector, where deteriorating infrastructure is coupled with the need for uninterrupted supply to consumers. The use of digital models and artificial intelligence algorithms allows not only for faster response to emergency situations but also for proactive prevention, reducing repair and maintenance costs.
The patent awarded marks an important milestone in the NSU Center for Information Technologies' work in this area: the creation of a universal model for developing methods for analyzing and managing utility infrastructure. The next step will be the development of digital twins for specific heating networks, taking into account their individual characteristics.
According to the developers, in addition to the joint project with FSUE UEV, a similar project is expected to launch this year with the participation of the Novosibirsk mayor's office. The technology could potentially be scaled up to other cities. This will improve the reliability of heat supply, reduce energy losses, and make urban infrastructure management more efficient.
Please note: This information is raw content obtained directly from the source. It represents an accurate account of the source's assertions and does not necessarily reflect the position of MIL-OSI or its clients.