Learning Forum

[Conor Leighton]

Hello all,
I’m a mechanical engineer learning Ansys – my first contact with Multiphysics in 15 years so be gentle!
I am working on a problem at the moment, and I am unsure as to the correct workflow in workbench.

We have two components, A & B. Component A has an interference fit with component B. This interference fit increases stress in some of Component A’s more slender regions. I have managed to assess this – happy days.
The next stage is in understanding the fatigue life on component A, for both random vibration and harmonic. Again, I can use modal analysis and harmonic/random to determine fatigue life.

However, how can I assess the impact of the press fit principle stresses on the fatigue life? In essence – how can I use the press fit analysis results as a mean stress input to both Random and Harmonic fatigue?

A few notes:
I’m happy to consider the impact of the press fit is ONLY to increase the mean stress ‘elsewhere’ on component A – I am not that interested in the press fit itself.
I do not think the prestress has any significant impact on resonance
I am happy to consider the contacts used from the press fit analysis as bonded for subsequent fatgiue analysis to enable linear assessment.

Any and all help greatly appreciated,
CL

[Chandra Sekaran]

The in-built Fatigue tool in Mechanical does not take into account mean stress from a static analysis. However you can use the ANSYS nCode Designlife add-on. This does allow adding a static analysis result to account for mean stress effects when doing PSD fatigue calculations. However note that ANSYS nCode Designlife does not directly use the PSD results created by ANSYS solver. Instead you provide a harmonic analysis results with unit load and the PSD curves. Additionally you also point to the static results (for mean stress effects).

[Daniel Shaw]

I concur with Chandra' reply.  Ansys nCode DesignLife is our best tool for directly combining the effect mean stress from a static load with PSD loading.  You just need to associate a static stress result set with a PSD loading in the DesignLife Load Mapper.  If you do not have an Ansys nCode DesignLife license and need to use the Mechanical Fatigue Tool (FT), there is a workaround that might be useful.  If the mean stress is relatively uniform in the critical region, you could extract it and assess its effect using a mean stress correction theory (probably either Goodmand or Gerber).  You could then include that mean stress effect by adjusting the material's S-N curve to create a S-N curve that includes the mean stress effect.  If the mean stress is not relatively uniform, it would require more user effect to find the mean stress at multiple locations and create multiple S-N curves.  A conservative approach is to use the maximum mean stress in the entire structure to adjust the S-N curve.