About preliminary statistical processing of information on the study of fatigue strength of machine parts

Abstract

A generalized approach to statistical processing of the results is proposed fatigue parameters of static strength using methods of mathematical statistics and conventional statistics using the following probabilistic parameters: the mean square deviation, the initial moment, the central moment of dispersion and the coefficient of variation of the distribution series of physical and mechanical characteristics of the part material. Further, the analysis of fatigue strength research at the initial stage is carried out according to statistical information processing, graphical design of a number of stress distributions, statistical stress analysis, a graphical approach is applied in the form of a histogram of a number of distributions, a polygon of frequencies, a polygon of accumulated frequencies, the selection of theoretical stress distribution laws with their empirical confirmation is compared with a more rigorous assessment of the conformity of the distribution laws, which is performed using special consent criteria, for example, the Pearson criterion, a parallel verification of the correctness of the chosen approach using classical dependences of the resistance of materials is proposed. the total error is estimated in the form of a methodological error and a direct discrepancy between the theoretical and experimental values of stresses and temperatures of the physico-mechanical process according to experimental and theoretical normal and tangential stresses arising during the operation of the part as a result of the application of external force factors, a temperature-time superposition is used, in the form of a function of the predicted durability of the fluctuating kinetic theory of strength, in which the temperature, as a linear function, it can be replaced by any energy or force criterion, in particular: specific energy, relative deformation, normal or tangential stresses. The proposed approach requires substantial experimental study on a basic batch of the same type of samples and coordination according to schematized diagrams of the limiting amplitudes of Goodman, Sorensen - Kinasoshvili, Kogaev, subject to the conditions of safe operation of parts in the field of low-cycle fatigue. The proposed probabilistic model of statistical processing of fatigue strength can be recommended for solving applied problems of the theories of mechanics of materials, elasticity, plasticity and creep, resistance of materials, structural mechanics