MODELING OF MECHANICAL SYSTEMS TO IMPROVE THE DURABILITY OF STRUCTURES

Abstract

The relevance of this study lies in the need to enhance the reliability and extend the service life of modern structures operating under complex conditions of dynamic loads and environmental influences. It has been established that traditional calculation methods do not always adequately account for nonlinear effects, the stochastic nature of loads, and the multifactorial processes of material degradation. As a result, premature failures occur, and the efficiency of technical systems decreases. This underscores the necessity of developing new approaches to modeling mechanical systems that can replicate real operating conditions and ensure the reliability of durability forecasts. The purpose of this article is to develop scientifically grounded approaches to modeling mechanical systems that take into account actual operating regimes, thereby increasing their durability and functional reliability. The research methodology is based on a systematic analysis of operating conditions and dynamic loads, the generalization of modern numerical methods, including the finite element method and the finite difference method, as well as the use of machine learning algorithms and the concept of digital twins. Methods of comparative analysis, stress–strain state modeling, and degradation process assessment were applied, ensuring the comprehensiveness of the results obtained. The findings of the study include the identification of key operational factors that determine the durability of structures, the systematization of modeling methods, and the substantiation of their effectiveness in predicting service life. The impact of environmental factors such as corrosion, thermal cycling, wear, and radiation exposure on the gradual degradation of materials was investigated. The study revealed critical challenges limiting the accuracy and practical implementation of models, including insufficient quality of input data, the complexity of describing multiphysics processes, and the lack of unified standards. The conclusions confirm that the development of multilevel models integrated with full-scale testing and sensorbased monitoring improves the accuracy of service life prediction and helps prevent premature structural failures. It is recommended to account for manufacturing and assembly defects, apply non-destructive testing, and establish unified methodological standards. Future research should focus on improving adaptive multiphysics models designed to operate with incomplete or stochastic data, developing artificial intelligence algorithms for analyzing large volumes of operational information, and creating digital platforms for unified durability assessment across different engineering domains.