The dissertation aims to develop a digital twin tool leveraging sensor data for predicting structural deterioration. Computational aspects will involve numerical methods like Finite Element Analysis (FEA) to solve equations describing stress and strain states of solids. The mathematical framework for dynamic system evolution, encompassing stochastic parameters, will be developed to adapt models for representing and predicting the mechanical behavior under study, using experimental data. This adaptation will be achieved through a combination of critical parameter descriptions with physics-based digital twin approaches and machine learning methods (data-driven digital twin). The goal is to create a tool capable of real-time adjustment of mechanical stress predictions in systems, enabling proactive maintenance with immediate economic and environmental benefits. The tool will accurately estimate system lifespans and facilitate extension through continuous algorithm feedback developed within the dissertation. It will establish a connection between the current functional state of the component and critical strength and functionality parameters in both current and future timeframes.