Fatigue simulation

[Gov45]

Hello all,

I have seen a case where while doing a fatigue calculation of an axle ball pin the load time history is divided into positive and negative signals. There is contact non linearity as well, since the ball pin will be in contact with an insert and will have a complex stress state.

ABAQUS was used to obtain the initial stress state (unit loads -Step-1=Pretension, Step-2=compression load, Step-3=Tension load) and FEsafe software was used to solve for fatigue.

My question is why the load time history is dived into positive and negative signals, does it have anything to do with contact non-linearity.

Thanks

 

[FEA way]

I'm not sure what you mean by load time history. Response from Abaqus ? For nonlinear analyses, you can't use the superposition (scale and combine) approach in fe-safe to multiply the frame with Abaqus results by a given signal. You have to use the dataset sequence loading where a series of frames (increments) from FEA defines the cyclic load.

 

[Gov45]

Load time history (positive and negative load signal) is the random signal that is given for fatigue calculation. This signal will be converted to different classes using rain flow cycle counting method . Each class will have a stress amplitude and mean stress which will be used for fatigue calculation.

So, for this contact non-linear calculation, this load time history was divided into positive and negative signals. This is where I am confused on why signal splitting is done.