Learning Forum

[Lena Westerburg Burr]

I have to analyse a structure for fatigue failure using ansys. The structure has both a constant load applied to it, as well as two identical oscillating loads acting at different locations (so the loading condition would be of the type R > 0). How do I set up these oscillating forces and the fatigue study such that I can see how many loading cycles (i.e. oscillations) the structure can withstand?

How does the fatigue tool know what to count as one cycle, should I input this? At the moment my 'Life' solution just shows the whole model as being able to withstand 10^8 cycles which I think might be unrealistic. 

I have already added S-N curve data to the engineering data.

I am new to Ansys and any help would be much appreciated.

[peteroznewman]

Hello Lena,

Do the oscillating loads at different locations have not only the same magnitude, but also the same frequency?  In other words, is the phase angle between those loads fixed or variable?  If the frequencies of the oscillating loads is identical, what is the phase between them, is it zero or some other fixed value?

It would be helpful to see an image of the structure and the location of the two oscillating loads and the constant load.

Regards,
Peter

[Lena Westerburg Burr]

Here is a picture of the structure. It is a tandem bicycle frame that has a constant downward force on each of the seat joints and then an oscillating force for each of the pedals. To simplify I think it is appropriate to assume that the two oscillating forces do have the same frequency and that the phase is 0. (This allows us to model the worst case scenario where the two riders exert the maximum downward force on the pedal on same side and at the same time)

[peteroznewman]

Hello Lena,

Thanks for the image and detailed explanation. It would be great if you also showed a coordinate system, so I will assume that X is horizontal, Y is vertical and Z is out of the page.

I assume the pedal forces are offset from the plane of the frame to represent the center of each pedal so they create some twisting in the frame.

What constraints are on the Remote Displacement? The only constraint should be Z = 0 and all others free.  You could move the coordinates of the Remote Displacement down by the radius of the wheel since it is the tire on the ground that is the constraint location, but that might not make a difference to the stress in the frame caused by the pedal loads.

I would delete the Fixed Support and put a Frictionless Support on the hole in the front tube. This would be a bit closer to how the frame is supported. 

The seat forces imply a 100 kg (220 lb) rider is sitting on each seat. Have you considered that when a rider wants to put maximum force into a pedal, they lift up from the seat?  One simple scenario for fatigue is that 981 N oscillates from the left pedal to the right pedal with no load on the seat, the opposite pedal or handle bar. If you do that, then a simple fatigue calculation would be two zero-based loads: left pedal and right pedal. You could define a fatigue cycle for this problem as one revolution of the pedal crank arms.

Or did you mean that two 171 kg (377 lb) riders are each putting 100 kg of their weight on their seat while also pushing down on the pedal? That is a scary thought!

Regards,
Peter

[Lena Westerburg Burr]

Thanks Peter, I will make those changes to the boundary conditions. Your other points are also helpful but the setup has mainly been defined by my university project brief so I'm aware that it might not be quite accurate as a model but it is a simplification for the purpose of the project. So yes, assume riders are each putting 100kg of their weight on the seat while also pushing down on the pedal.

My question would be how do you define a fatigue cycle in the fatigue tool? And how do you do it such that two forces are oscillating and two are static?

[peteroznewman]

Use the Fatigue Tool and set the Loading Type to Non-proportional Loading. 

The Non-proportional Loading option is a non-proportional constant amplitude loading type for models that alternate between two different stress states (for example, between bending and torsional loading). Problems such as an alternating stress imposed on a static stress can be modeled with this feature. Non-proportional loading is only supported for Fatigue Tool objects used with Solution Combination where you select exactly two stress states (only two Base Cases present in the Solution Combination Worksheet).