Wyniki 1-1 spośród 1 dla zapytania: authorDesc:"Michał Zielaskowski"

Influence of material heat treatment on low cycle fatigue of austenitic steel at working conditions of gas turbine exhaust structures DOI:10.15199/28.2019.5.2

  1. PROBABILISTIC CHARACTER OF THE FATIGUE PHENOMENON Despite the continuous development of knowledge about the phenomenon of fatigue, predictive models describing this phenomenon often differ significantly from the cases observed in reality. This discrepancy results from the probabilistic nature of the phenomenon of fatigue, the relationship between the initiation and development of the crack from randomly distributed defects of the metal structure, which are additionally superimposed by uncertainties resulting from geometry, load, boundary conditions, material properties, measurement errors and finally simplifications adopted in the analytical models themselves. Equally important is the fact that fatigue tests are expensive and time-consuming, which means that researchers are forced to work on a small amount of data, which significantly extends the confidence range of the estimated models. Such a wide range of confidence means that the influence of some parameters that may possibly affect life resistance of a material remains undetectable. Understanding and considering the above-mentioned features of the fatigue phenomenon is crucial for the correct interpretation of test data and, as a result, for the safe and responsible construction of machine elements. This fact is confirmed by the growing number of scientific papers analysing the phenomenon of fatigue with mathematical statistics tools that allow to assess the reliability of mathematical models, probability of damage and design an experiment so that the obtained data give the most reliable results [1, 2]. Figure 1 shows example of data from fatigue tests for the same material, tested under the same conditions and in the same laboratory [3, 4], with the confidence interval of these data for the 2μ confidence level. It can be seen that although the linear regression curve well describes the measurement points, the confidence range of the estimation is very wide and m[...]

 Strona 1