We analyze models of summation of fatigue damages under irregular variable loading on the basis of linear and nonlinear interpretation of the damage function. We show the necessity of taking into account the main physicomechanical processes in the metal under cyclic loading, which can improve the accuracy of computations.Keywords: models of summation of fatigue damages, physicomechanical processes, cyclic loading.The demands of engineering practice in designing new and perfecting the existing technical articles that work under regular and irregular variable loading, as well as the introduction of new promising materials, require the development of models and computational methods that would enable one to better predict the durability of an engineering object under in-service conditions for guaranteeing its high reliability and competitiveness [1,2]. In doing so, it is important both to carry out a comparative analysis of existing models and to develop new models (especially those available for engineering practice) with regard for the actual mechanisms of damage accumulation, improving the accuracy of prediction.The solution of such a complex problem is connected with searching for the most general laws describing the kinetics of damage accumulation in metals. Note that the intention to take into account the phenomena occurring in the structure of metals more accurately leads to the complication of the existing relations. One of the approaches in the theory of summation of fatigue damages is the concept of crack initiation and growth, which has found its reflection in the models of summation where fatigue consists of two stages.The first of them, i.e., the stage of crack initiation, or the development of dispersed damages, is connected with the kinetics of accumulation of microplastic strains, which lead to the development of damages in the form of short dispersed cracks. When the critical point is achieved, the process of damage passes to the second stage, namely, the propagation and growth of a main crack, which, upon reaching its critical length, leads to the fracture of the material or of the structural element. According to another approach, a defect in the form of a crack can always be present in the structure of a metal, especially in the case of stress concentration, and, therefore, one should consider durability as the survivability of a cracked structure, whose lifetime is determined by the kinetics of crack propagation up to reaching its critical size. Such a concept is appropriate for structural elements in operation, especially for preventive measures and for the determination of their lifetime resource. Depending on operational, technological, and design factors, this stage can constitute 10 to 30% of the lifetime of the structure [3].In the present work, we consider the problem of evaluating the durability of materials and structural elements in the stage of dispersed damages (the stage that determines the main lifetime of the structure under consideration up to its failure). In engineering practic...