So you’ve run a finite element analysis (an FEA) of a mild steel “thing” and have a lovely looking stress plot, but wait! Some stresses are 800 MPa and glowing red!
What should you do?
It’s extremely disappointing to know that some engineers – so-called experts in their field – would increase member and plate sizes to reduce these high stresses. Yet, in most cases, they represent singularities in the FEA model that can be dismissed.
Think about it. Mild steel yields around 250 MPa, and has an ultimate strength of 410 MPa. 800 MPa is a fairytale stress – it CAN’T happen.
But you do need to understand what the computer is telling you. You can’t just ignore them per se.
A stress singularity is a point of the mesh where the stress does not converge towards a specific value. If we keep refining the mesh, the stress at this point keeps increasing, and increasing, and increasing… Theoretically, the stress at the singularity is infinite.
The most common examples of this occurrence is when:
- applying a point load;
- creating abrupt load changes;
- constraining a model at a point;
- creating abrupt transition between materials; and
- sharp re-entrant corners.
If you know what issue you’ve modelled to cause the singularity, you can be confident in dismissing it.
When loaded statically (ie not repeated fatigue loading), ductile materials, like steel, yield before failure. Where small portions of the structure see stresses beyond the yield stress they will in actual fact locally yield, and in the process, redistribute this load to adjacent structure elements.
As the “bits” experiencing the high stress hit yield, the stiffness of the material changes and load is picked up by the structure around it (which is still operating below yield, with a higher stiffness than the yielded structure).
If you’re not happy to make a judgement call, there’s a feature in most good FEA software that actually models this behaviour, and loads up adjacent elements rather than stressing elements beyond their yield limit.
In short, if the model runs successfully, the structure is sound under-strength limit state loads (i.e. much higher than expected).
If the model doesn’t run, well, the loads are too high or the section too weak, and now it’s time to review the member sizes and plate thicknesses.
Fatigue is a bit of a different story. We’re now talking expected loads being repeatedly loaded and unloaded, not those factored ones above which we accept failure.
If these repeated loads produce yield in some elements, that’s not so good. Because when the load is removed the yielded elements get compressed by the stiffer elements of steel around them, and if repeated long enough will lead to fatigue failure.
We want to ensure our designs are fatigue-friendly, and we want to model those designs properly if the software to predict expected stress ranges.
Cracked Steel Beam Subjected to Non-Stationary Fatigue Cycle
The moral of this story? Just because you buy the licence to use FEA software, doesn’t imply you know how to use it!
Used properly, Finite Element Analysis (FEA) an amazing method that proves complex structures are structurally sound in cases when there’s no equivalent simple code formula.
We regularly use Strand7, Staad.Pro and SolidWorks Premium Simulation to solve problems using FEA. Each of these applications have their merits and the choice of use is dependent on the structures or objects being analysed and the results you wish to determine.
Have a Project that You Want to Discuss?
Book a meeting with us and we will assist you with the structural design and analysis of your structure to help you produce a structure that is fit-for-purpose and capable of resisting all applied loads without failure during its intended life.
Like this article?
Subscribe to our newsletter to receive a weekly email from Yenem’s director, Dave Meney.