Skip to main content

Fast determination of fatigue properties of materials beyond one billion cycles

Periodic Reporting for period 2 - FastMat (Fast determination of fatigue properties of materials beyond one billion cycles)

Reporting period: 2019-01-01 to 2020-06-30

If a cyclic loading is applied to a material, it can fail for stress amplitude lower than the Ultimate tensile strength. This phenomenon is called fatigue. Fatigue of materials has been studying for a long time but it remains nowadays a crucial step in the mechanical design. For example, in the transport and energy production industries it is estimated than 80% of fracture are due to fatigue.

Moreover, the increase of the lifespan of many structures leads to an increase of the number of cycles applied to this structure. Nowadays, it is rather common to find mechanical system like rotating machine components which can fail for number of cycle larger than ten million and sometimes beyond one billion.
To assure the safety of the structures it is necessary to characterize the fatigue properties of the material for this very high number of cycles.

The current fatigue design methods are based on standards which recommend to plot the evolution of the stress amplitude according to the number of cycles to fracture. Each point of this curve corresponds to a fatigue test carried out until fracture generally at low frequency of typically 10 hertz. To draw a SN curve until 107 cycles, it can take more than one month. And if one wants to explore the VHCF domain, the duration of only one test until one billion cycles is about 3 years. Therefore, for testing time reduction reasons fatigue characterization is limited to ten million cycles. And the standards suppose that the SN curve can be extrapolated in the very high cycle fatigue domain by a horizontal asymptote called the fatigue limit. Many results in the literature show that this fatigue limit does not always exist.

The objective of the FastMat project is to provide answers to these two technological challenges: the reduction of the testing time and the exploration of the VHCF domain. We thus suggest to develop a completely new method for fatigue characterization based on the analysis of short interrupted fatigue tests with complementary measurement of self-heating. The final objective of the project is thus to be able to estimate the fatigue behavior of material in the VHCF domain from these self-heating data. This new method will have the advantage to reduce drastically the testing time.
To applied this new method, we suggest is to establish the physical relationship between the self-heating, the fatigue damage and finally the number of cycles to fracture. In the case of metallic materials, the self-heating is related mainly to intrinsic dissipation. At the microscopic scale this phenomenon can be explained by the cyclic movement of dislocations in the crystal.
On the other side, the fatigue damage is due to the irreversibility of this dislocations movement. This irreversibility cumulates according to cycles and generates an evolution of the microstructure which can be quantified by the stored energy in the material. This explanation shows that the stored energy characterizes thus very well the level of the fatigue damage. Finally, an extrapolation of the evolution of the stored energy over cycles will give information on the number of cycles to fracture.

To solve the problematic of this project, we propose a methodology based first on an experimental study where both the intrinsic dissipation and the stored energy were quantified from measurement of the temperature, the stress and the strain during ultrasonic tests and second on a modelling at the dislocation scale to establish the relationship between the fatigue mechanisms and their thermal signature. This modelling will be carried out with a Discrete Dislocation Dynamics which simulate the movement and the interaction of a great number of dislocations during an cyclic Loading. The results of this simulation and it will help us to interpret the experimental results.

The main experimental challenge is the stress measurement during one cycle, 50µs. To do that time resolved X ray diffraction technique will be used with intense sources from a synchrotron beamline. In the VHCF domain the stress and the strain are small and the mechanical work depends directly on these two quantities and their time shift which is about few ten of nanosecond. Thus it is necessary to improve the time resolution and to do that we propose two methods. The second method which use the synchrotron in a pulsed mode, allows to obtain time resolution lower than one tenth of nanosecond. The first period of the FastMat project was mainly devoted to the developement of this two experimental techniques.
The scientific objective of the FastMat project is to understand the relationship between the self-heating of materials during a cyclic loading and the fatigue damage level in order to develop very fast method to characterize the fatigue behavior beyond one billion cycle. The ultimate objective is to be able to write a standard which can be used by the engineers to design structures submitted to cyclic loading and thus to assure their security.
Experimental device : ultrasonic fatigue machine on the diffAbs beamline of Soleil synchrotron
Classical stress amplitude versus number of cycle curve