Learning Forum

[waltbeach]

Hello everyone,

I'm performing a fatigue analysis with the following load:

I have defined Infinite Life as 10^6 cycles and 1 block as 1 cycle, as shown below:

However, when I get results, the maximun life obtained (infinite life) is not corresponding with 10^6 cycles as I defined previously, inifinite life is shown with a value of 1171 cycles (I don´t know why 1171). I mean, once I obtain 1171 cycles in a zone of my component, even decreasing loading, the result of life keeps constant, instead of keep growing until 10^6 (which I defined infinite life).

Someone has any idea what's happening?

Thanks in advance.

[peteroznewman]

Read the ANSYS Help to understand what defining an Infinite Life means in this analysis.

Infinite Life

Stress Life Analysis

This option appears only if Type is set to History Data (non-constant amplitude loading) and defines what life will be used if the stress amplitude is lower than the lowest stress on the SN curve.

In your model, the stress amplitude is not lower than the lowest stress on the SN curve, so it calculates Life from the SN curve. 

When you plot Life, it shows that the minimum value on the legend is 1171 and shows you where on the part that minimum value can be found. This means that the part will have a 50% probability of failure at that location after it has been exposed to 1171 repetitions of that loading history.

If you were to replace that loading history with one where the amplitudes were 1000 times smaller, then you might find that the Life plot will show 1e6 cycles as the minimum life.

[waltbeach]

I found and read the fatigue analysis manual from Ansys, but still, I didn´t find the solution.

I tried exactly what you have recommended me, perform the analysis with very low amplitude load, but the results doesn´t change.  I let several pictures showing results with amplitude loads 1000 times lower:

(Non-constant amplitude loading. The resolution doesn´t allow to see the scale, but the maximum value is 0.999)

(Fatigue tool configuration)

(Load included in the analysis) 

(Life results)

(SN curve of my material. Y axis (x10^3))

Honestly, I don´t see where the error is, because maximum 1N amplitude loads don't generate amplitude stresses bigger than 390 MPa  (stress at 10^6 cycles).

[peteroznewman]

It looks like the last point on the SN curve is 370 MPa, so if the 1 N amplitude load generates a stress of 390 MPa, that is larger than the last point on the SN curve and the Infinite life will not be used.  Try an even larger scale reduction in the time history amplitude of 10,000 instead of 1,000. Or just delete the two points off the SN curve.  All this is just in service of trying to see if the software will report the Infinite life value you entered, but it seems like that won't apply when using your proper data.

[waltbeach]

Sorry if I explained wrong myself before, 1N amplitude load don't generate 390 MPa, not even 1 Pa barely.

Moreover, I have tried to change the load history data, and tested with SineWave.dat, which Ansys contains by default and...Surprisingly, now I get the value of infinite life which I defined previously (1e6 cycles) . However, when I try back with other load history data like SAETransmission, the error reoccurs, and I get a value lower than I defined as Infinite life.

I has no idea why happen this, because a load history shouldn't have influence in those parameters, when stresses generated are too much lower than stress at the number of cicles defined as Infinite life (1Pa vs 390 MPa), in both cases.

[peteroznewman]

If you provide the source files to reproduce this error, then it might be entered as a defect that the developers can remedy in the next release.