NUMERICAL MODELING OF THE IMPACT OF COMPLEX LOADS ON BUILDING MATERIALS

Abstract

The article is devoted to the urgent problem of predicting the durability of building materials and structures exposed to complex multifactorial loads, including mechanical, thermal, seismic, and corrosive effects. Current trends in construction, especially in areas with high risks of military damage and natural disasters, require scientifically grounded methodologies for assessing the performance of materials under real operating conditions. The study presents a comprehensive analysis of numerical modeling methods, with a particular focus on the finite element method (FEM). This approach enables detailed reproduction of the stress–strain state and makes it possible to account for nonlinear interactions between different types of loads, which is essential for accurate predictions of material service life. Special attention is paid to algorithms that integrate mechanical, seismic, and thermal effects into a unified model, as well as the application of combined methods, including the boundary element method, the Monte Carlo method, and finite differences. The proposed numerical schemes were validated against experimental data, confirming high accuracy with deviations not exceeding a few percent. An additional innovative aspect of the research lies in the integration of classical numerical methods with machine learning technologies, particularly deep neural networks, which allow the consideration of complex nonlinear degradation patterns of materials over time. The study also emphasizes the integration of numerical models with monitoring systems based on IoT sensors. Such an approach ensures real-time dynamic control of the technical state of building structures and enables the timely identification of critical deviations. It has been demonstrated that the application of these algorithms not only improves the accuracy of residual life predictions but also significantly reduces costs by implementing resource-saving restoration technologies. The conclusions outline future research directions, including the extension of numerical modeling methodologies for novel high-performance materials, the advancement of machine learning techniques, and the creation of fully automated systems for monitoring and predicting the technical state of building structures.