Neural Network Modeling of the Strength of Normal Sections of Prefabricated Reinforced Concrete Ribbed Slabs

Introduction. Precast reinforced concrete ribbed slabs are broadly used as floors and coverings for industrial, residential and public buildings. Their use in this capacity is due to the high technological efficiency of manufacturing, efficient use of concrete and the possibility of automating factory production. One of the critical tasks in designing such structures is to calculate the bearing capacity of normal cross sections. Traditional calculation methods are reliable, but they are outdated. Machine learning methods are increasingly being employed in engineering, where researchers are opting for artificial neural networks (ANNs). The use of traditional methods in processing structured data such as tables and databases has its limitations. Neural networks are capable of analyzing unstructured data such as text, images, and videos, which opens up new prospects for analyzing and comprehending information. The article sets forth an approach to neural network modeling of the bearing capacity of normal sections of prefabricated reinforced concrete ribbed slabs.

Materials and Methods. A structured and processed data array (dataset) includes 20 samples for which a computational model based on a multilayer perceptron has been developed and verified. The input parameters are the geometric as well as physical and mechanical characteristics of the slabs and the applied load, the output parameter is the limiting bending moment calculated using the limit state method.

Research Results. Training on a limited sample did not lead to retraining of the model due to the correct division of data into test, training and control batches and the use of the quasi-Newton optimization method. The model has displayed a high level accuracy and reliability. Artificial neural networks are capable of identifying nonlinear dependencies between the parameters with no a priori assumptions.

Discussion and Conclusion. The suggested model is not a substitute for the existing calculations, but it serves as an efficient digital tool for quick verification of design solutions, optimization of reinforcement and improvement of structural reliability. Its implementation into BIM systems and digital construction platforms is in compliance with the requirements of Industry 4.0 and creates new opportunities for designing prefabricated reinforced concrete structures.

1. Rimshin VI, Anpilov SM, Usanov SV. Application of Cognitive Technologies to Predict the Strength of Thin Walls of I-Beams. Expert: Theory and Practice. 2024;1(24):42-52. https://doi.org/10.51608/26867818_2024_1_42 (In Russ.).

2. Rimshin VI, Solovyov AK, Suleymanova LA, Amelin PA. Neural Network Forecasting of Physical and Mechanical Characteristics of Composite Materials Used to Strengthen Building Structures. Expert: Theory and Practice. 2023;4(23):101–107. (In Russ.) https://doi.org/10.51608/26867818_2023_4_101

3. Tarasov IV. INDUSTRY 4.0: CONCEPT & DEVELOPMENT. Business Strategies. 2018;6(50):43–49. (In Russ.) https://doi.org/10.17747/2311-7184-2018-5-43-49

4. Rimshin VI, Usanov SV. Neural Network Modeling of the Strength of Normal Cross-Sections of Beams with Rigid Composite Reinforcement. News of Higher Educational Institutions. Construction. 2025;7(799):5–15. (In Russ.) https://doi.org/10.32683/0536-1052-2025-799-7-5-15

5. Rimshin VI, Usanov SV, Vorobyov AE, Savelyev ES. Restoration of the Bearing Capacity of Reinforced Concrete Floor Slabs on the Example of a Multi-Storey Civil Building. BST: Bulletin of the Construction Equipment. 2025;6(1090):46–48. (In Russ.) https://elibrary.ru/item.asp?id=82379614 (accessed: 21.10.2025)

6. Rimshin VI, Usanov SV, Vorobev AE, Savelev ES. Numerical Simulation of the Stress-Strain State of Beams with Rigid Composite Reinforcement. Vestnik of Volga Tech. Series: Materials. Constructions. Technologies. 2025;1(33):63–73. (In Russ.).https://doi.org/10.25686/2542-114X.2024.4.63

7. Rimshin VI, Anpilov SM, Usanov SV. Practical Approaches to Eliminating Collisions in Information Models of Buildings. Expert: Theory and Practice. 2024;1(24):42–52. (In Russ.) https://doi.org/10.51608/26867818_2024_3_87

8. Kurbatov VL, Rimshin VI, Shubin IL, Volkova SV. Information Modeling and Artificial Intelligence in Modern Construction and Housing and Communal Services (2nd edition, revised) Moscow: Publishing House ASV; 2025. (In Russ.) https://elibrary.ru/item.asp?id=82433619. (accessed: 21.08.2025)

9. Perera R, Barchin M, Arteaga A, De Diego A. Prediction of the Ultimate Strength of Reinforced Concrete Beams FRP-Strengthened in Shear Using Neural Networks. Composites Part B: Engineering. 2010;41:287–298. http://dx.doi.org/10.1016/j.compositesb.2010.03.003

10. Peng F, Xue W, Xue W. Database Evaluation of Shear Strength of Slender Fiber-Reinforced Polymer-Reinforced Concrete Members. ACI Structural Journal. 2020;117(3):273–282.http://dx.doi.org/10.14359/51723504

11. Estep DD. Bending and Shear Behavior of Pultruded Glass Fiber Rein-forced Polymer Composite Beams with Closed and Open Sections. West Virginia University; 2014. 545 p. https://doi.org/10.33915/etd.545

12. Lagaros ND. Artificial Neural Networks Applied in Civil Engineering. Applied Sciences. 2023;13(2):1131. https://doi.org/10.3390/app13021131

13. Afrifa RO, Adom-Asamoah M, Owusu-Ansah E. Artificial Neural Network Model for Low Strength RC Beam Shear Capacity. Journal of Science and Technology (Ghana). 2012;32(2):119–132. http://dx.doi.org/10.4314/just.v32i2.13

14. Ahmad A, Cotsovos DM, Lagaros ND. Framework for the Development of Artificial Neural Networks for Predicting the Load Carrying Capacity of RC Members. SN Applied Sciences. 2020. https://doi.org/10.1007/s42452-020-2353-8

15. Mansour MY, Dicleli M, Lee JY, Zhang J. Predicting the Shear Strength of Reinforced Concrete Beams Using Artificial Neural Networks. Engineering Structures. 2004;426:781–799. http://dx.doi.org/10.1016/j.engstruct.2004.01.011

16. Imam A, Anifowose F, Azad AK. Residual Strength of Corroded Rein-forced Concrete Beams Using an Adaptive Model Based on ANN. International Journal of Concrete Structures and Materials. 2015;9(2):159–172. http://dx.doi.org/10.1007/s40069-015-0097-4