Font size

A
A

Line spacing

A
A

Color

A
A
3 августа 2026 г.

SFedU has taught a neural network to select the “recipe” for the strongest concrete from waste

3 августа 2026 г.

SFedU has created a neural network that selects the optimal composition of concrete for construction 3D printing from waste. The algorithm finds a compromise between environmental friendliness and strength, making it possible to reduce the consumption of natural sand and speed up the development of new materials.

The main goal of the study was to create an algorithm that ensures a compromise between the environmental friendliness and strength of concrete – as was emphasized at the Southern Federal University. Additional requirements for concrete with waste were imposed because it was intended specifically for 3D printing of buildings, rather than for traditional construction methods.

“The mixture for construction 3D printing must be easy to pump and flow evenly through the nozzle, but immediately after laying, it must retain its shape and support the weight of subsequent layers.” “Therefore, when selecting the composition, it is important to take into account not only the final strength of the concrete, but also its suitability for continuous and stable printing,” explained one of the authors of the new program, Ekaterina Kravchenko, head of the Center for Circular Biomaterials Competencies and leading researcher at Southern Federal University. The specialist emphasized that when assessing the strength of 3D‑printed concrete, it is necessary to take into account its layered structure, the quality of adhesion between the layers, and the dependence of the mechanical characteristics on the direction of load application.

Scientists from Southern Federal University have developed a program that allows for the use of fewer primary natural resources for construction 3D printing – that is, it partially replaces some components, the extraction of which requires separate work, with construction waste. This algorithm will significantly speed up the creation of new formulations for construction 3D printing, reduce raw material consumption, the number of laboratory experiments, and the time required to select the mixture. Using such materials, it is possible to print interior partitions, facade elements, small architectural forms (for example, flower beds or benches), and individual non‑structural components of low‑rise buildings.

“On the one hand, the composition should contain the maximum possible amount of secondary raw materials obtained from construction waste in order to reduce the consumption of natural sand and return the waste to the production cycle. On the other hand, an increase in the proportion of recycled components should not lead to a significant loss of concrete strength. As a result, machine learning helps to select such a compromise composition,” explained one of the authors of the work, Professor Tatyana Minkina, head of the Department of Soil Science and Land Resource Assessment at Southern Federal University.

In the future, scientists plan to continue developing eco‑friendly building materials for 3D printing for urban infrastructure, including for vertical landscaping projects.

“We aim to make cities ‘greener’ in two ways at once: by increasing the area of vegetation “upward,” and by using eco‑friendly materials and low‑carbon technologies in the creation of the structures themselves,” - Minkina added.

The work was carried out with the support of the federal “Priority‑2030” program (the “Youth and Children” national project) at Southern Federal University. The results are presented in the Journal of Materials in Civil Engineering, ASCE.

 

Short link to this page sfedu.ru/news/81111

Additional materials on the topic