Structural Mechanics of Engineering Constructions and Buildings

Thermal stress during the hardening of mass concrete structures is a significant risk factor for early cracking, which directly impacts the durability and load-bearing capacity of buildings and structures. Simplified calculation methods based on hypotheses about the pattern of temperature and stress distributions often demonstrate low accuracy, necessitating the search for more advanced approaches to stress state prediction. This paper proposes a method for predicting the thermal stress in mass concrete foundation slabs based on artificial neural networks (ANNs) using real-time temperature monitoring data. Three ANN architectures were investigated: recurrent, feedforward, and cascade. A comprehensive dataset, including 499,800 records obtained from parametric finite element calculations, was compiled for training. The models demonstrated high prediction accuracy, with the feedforward neural network achieving the best result, with a mean-square error of 0.025 MPa². Verification using experimental data confirmed the practical applicability of the approach, including the ability to predict the timing of crack formation. The developed method enables efficient and less computationally expensive analysis of temperature monitoring data in real time compared to traditional modeling, thereby improving the reliability of building structures.

The object of the study was a composite spatial rod structure of cylindrical shape with a coating in the form of a truncated cone. The cylindrical part of this structure was a regularly ordered rod lattice formed of flat regular hexagons. Blocks including linear and rotary springs were used to connect the rods of the polygons. The conical cover was a conventional frame-truss structure. The purpose of the study was to numerically analyze the stress-strain state of a spatial core structure under dynamic influence, taking into account the malleability of nodal connections. The finite element method in the form of a displacement method is used as a mathematical apparatus. Computational experiments have made it possible to design the design of the block for the nodal connection of the lattice rods. Using a full-scale model that included three sections of the grid, the kinematic parameters of the designed block of the nodal connection of the rods were studied.

Roofs in the form of plates and shells of complex curvilinear plan are common structural solutions in architecture. Such structures have a number of advantages. The mid-surface of shells and plates of curvilinear trapezoidal plan is constructed using parametric and vector equations and has a number of special aspects to consider when calculating their stress-strain state. For structures of this shape, no exact analytical solution has been obtained, but it is possible to obtain a numerical solution, for example, by the finite element method and the variational-difference method. In such a situation, for verification of calculations, comparing the results obtained using different numerical procedures is useful and relevant. A comparative analysis of the results of calculating the stress-strain state of a plate curvilinear in plan, obtained by the methods mentioned above, was conducted. In the literature, the topic of calculating plates and shells of curvilinear trapezoidal plan is insufficiently developed. The aim of the study is to obtain data on the calculation of the stress-strain state of a plate of curvilinear trapezoidal plan, as well as to assess the applicability and specifics of the two methods in calculating such structures. To accomplish the tasks, the following software was used: ANSYS APDL software for calculation by the finite element method, and the author-developed SHELLVRM program for calculation by the variational-difference method. The parameters of the stress-strain state of a plate of curvilinear trapezoidal plan have been obtained and analyzed, verification of the obtained results has been carried out, recommendations for implementing both calculation methods in the practice of structural analysis have been given, and computational difficulties and special aspects of both methods have been identified.

This research aims to improve methods for testing steel fibers, concrete, and fiber-reinforced concrete elements in tension, bending, and under the combined action of tensile, shear, and bending forces. The development of new and the improvement of existing experimental methods of preparing initial data for the design of fiber-reinforced concrete structures has become a relevant issue today, as the rate at which new types of steel fibers enter the market outpaces the research into the mechanical properties of fiber-reinforced concrete created using these types of fibers, which is necessary for creating a comprehensive calculation model for the behavior of fiber-reinforced concrete structures under service conditions. This article discusses the author’s proposed testing method and the device for studying the mechanical properties of fibers, concrete, and fiber-reinforced concrete elements, describing the operating principle of the developed experimental device and the features of the recording equipment. A detailed description of practical evaluation of the developed experimental setup is provided, along with references to the results obtained using the test method described in the article. A series of studies conducted using the developed experimental setup allowed to obtain new data and refine existing data on the diagrams characterizing the behavior of steel fibers in concrete. Such diagrams were used to construct a mathematical model for the stress analysis of normal cross-sections of flexural fiber-reinforced concrete elements and to assess the strength and reliability of fiber-reinforced concrete structures in bending.