Application of statistical approaches within the scope of fatigue life prediction methods for notched specimens using the example of unalloyed steels

• Applicant: Professor Dr.-Ing. Peter Starke
• Subject Area: Mechanical Properties of Metallic Materials and their Microstructural Origins
• Term: since 2024
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 542796320

The design of dynamically loaded metallic materials and components is usually based on the relationship between stress amplitude and number of cycles to failure shown in a S-N curve. Statistical statements regarding the probability of fracture or failure lead to a significantly improved fatigue life prognosis, but increase both the costs and the experimental effort. The aim of the proposed research project is to develop a fatigue life prediction method (LPV), which enables the most accurate possible prognosis with regard to cyclic deformation behaviour and (residual) fatigue life despite a reduced specimen effort. The process-oriented consideration of the fatigue process with the help of non-destructive testing methods (e.g. thermography, DIC, resistance measurements, magnetic methods) makes the highest possible information output accessible. An essential aspect is the application of parametric and non-parametric mathematical models, which enable a statistical validation of the results obtained. The statistical evaluation is based on the data recorded from several load increase and constant amplitude tests, whereby individual S-N curves result from the input variables derived in each case. These differ in their position and slope due to the individual microstructure. The resulting material-dependent scatter will be included in the calculations as a scatter range. The results of the proposed research project will thus make it possible to demonstrate the potential of combining fatigue life prognosis methods and the statistical evaluation of fatigue data. Furthermore, the LPVs developed so far have mostly been used and validated on unnotched specimens. For a first step towards the provision of component S-N curves by means of LPV, a further goal is to make the measurement techniques and methods also suitable for notched specimens in order to detect processes occurring in the notch base during the fatigue process and to provide input variables for the LPV to be newly developed. If this is successful, the range of applications for LPV can be significantly expanded, resulting in a clear added value for the current state of the art.

• DFG Programme: Research Grants