Optimization of structural solutions for hydrotechnical facilities using digital modeling and intelligent control systems

Abstract

The article is devoted to the optimization of structural solutions for hydrotechnical facilities under conditions of increasing uncertainty in the hydrological regime and long-term material degradation, proposing a transition from static calculation schemes to digital twins integrating information modeling, computational hydrodynamics, structural analysis, and intelligent control. It is shown that the traditional representation of a concrete mass as an isotropic medium leads to systematic errors in assessing stability and filtration reliability, since aging has a pronounced zonal character and is associated with microfracturing, leaching, and cyclic freezing-thawing. Based on the analysis of long-term observation series from more than 1500 sensors and the results of laboratory tests of cores, significant discrepancies between design and actual parameters have been established: in the zone of variable water level, the filtration coefficient increases by almost two orders of magnitude, and in the contact zone, a decrease in the modulus of elasticity by 17,56% is recorded with high variability, which requires the introduction of spatially heterogeneous parameters into calculation models. Numerical modeling of the stress-strain state with subsequent topological optimization of the gravity dam profile demonstrates the possibility of reducing the concrete volume by 11,59% while maintaining normative stability margins, but reveals an increase in tensile stresses in the crest zone, necessitating local structural reinforcement. For predictive control of the filtration regime, statistical and machine learning models are compared; the highest accuracy is provided by hybrid neural network approaches with physical constraints (RMSE 0.076 m, R² 0.991), allowing early detection of anomalies and reducing the integral accident risk by 24,5%. Barriers to implementation are noted, related to the lack of unified data exchange standards, algorithm validation, requirements for computational resources, and cybersecurity of critical infrastructure; the need for continuous updating of the digital model based on monitoring data to maintain the reliability of forecasts and adaptive management of spillway regimes is justified.