The stress-strain state of large-sized metal structures is investigated. The causes and consequences of the formation of residual stresses and strains are shown. Methods for predicting residual stresses and strains by the calculation method are presented. Destructive and non-destructive methods for determining the stress-strain state of large-sized metal structures are presented. The influence of local deformations and clearances during assembly on the value of residual stresses and deformations is shown on the example of a typical curved large-sized metal structure, characteristic for the design of antenna devices of radar stations and air traffic control systems. Conclusions are made about the importance of analyzing and evaluating the stress-strain state of large-sized metal structures. Radar stations and air traffic control systems during operation experience extreme multi-parameter loads and thermal effects. To ensure the high reliability of their work, a thorough and accurate analysis is required, followed by an assessment of the stress-strain state of the bearing large-sized component parts of metal structures already manufactured and only being designed at the stage of experimental design work, in order to be able to choose the correct technological, constructive and organizational sequence for their manufacture. In modern production, metalworking methods are used, based on a sharp increase in the energy concentration on the treated surfaces of the elements, which contributes to the uneven distribution of thermodynamic potentials over their volume. The critical state is stress concentration in the metal structure, which can lead to its destruction. In zones of stress concentration, a complex stress state always arises, volumetric or flat. The type of local stress state significantly affects the level of loads that the metal structure can withstand without destruction. The most dangerous is a comprehensive uneven stretching. The conditional characteristics of the mechanical properties of a material such as tensile strength or elongation, determined in accordance with current standards, are not enough to calculate the loads that the structure can withstand without breaking. Also, the stress-strain state of the metal structure affects the dimensional stability in the metal structure, which leads to the need to use special technological solutions to relieve and relax existing residual stresses and strains. A sufficiently accurate assessment of predicting the stress-strain state of large-sized metal structures can be a model model, which analyzes and evaluates residual stresses and strains in-situ, and the level of breaking load when testing a model model under appropriate temperature conditions is taken as a criterion for assessing the health of a material. However, this method for large-sized metal structures is not always technically feasible and often unprofitable due to the large size of structures, the duration and cost of testing, the difficulty of creating full-scale operating conditions, etc. The problem of determining the calculated stress-strain state of a metal structure can be solved by separate solution of thermomechanical and deformation subtasks according to empirical formulas using the finite element method or the extended finite element method. Moreover, for the reliability of determining the calculated stress-strain state, it is necessary in the mathematical model to take into account many factors affecting the magnitude of the residual stresses and strains. The indicated assumptions, as well as the complexity of the proposed calculations, do not allow accurate prediction of the subsequent stress-strain state of large-sized metal structures having complex geometric and spatially oriented shapes. It is possible to use non-destructive and destructive methods to determine the actual stress-strain state of metal structures. For a more accurate assessment of the stress-strain state of metal structures, we must cut the object and subject the interior to the measurement of residual stresses. For this, it is possible to use two main methods: the stress relaxation method and the method of intrinsic deformation. As practice shows, it is necessary to predict residual stresses during welding of various types of joints without performing complex calculations of thermal elastoplastic analysis. In these cases, the following two simpler methods can be used: the use of experimental databases and the use of effective internal deformation, which is a source of residual stress. As an example, deformations of welded large-sized metal structures, typical for antenna systems of radar stations and made of sheet metal, are predicted. Thus, we can conclude that a preliminary and actual assessment of the stress-strain state of welded metal structures, especially large ones, is a difficult task, but its importance can hardly be underestimated. In this regard, new methods and techniques are constantly appearing that allow this to be done with the greatest accuracy and less computational complexity.