Run a linear-elastic FEA on a pressure-containing component and it’s common to get a result that says the design fails — calculated stress above yield at a local hot spot — on a part that, in reality, has years of safe service behind it or would clearly survive a hydrotest. That’s not a wrong answer; it’s the wrong question. Linear-elastic analysis is often too conservative to qualify a real design, and elastic-plastic and limit-load methods exist specifically to answer the question linear-elastic analysis can’t: does the part actually collapse, or does a local region simply yield and redistribute load to its neighbors the way real steel does?
What linear-elastic analysis assumes — and where that breaks down
A linear-elastic FEA assumes stress and strain stay proportional no matter how high the calculated stress goes, and it reports whatever peak stress the mesh produces, even at a stress concentration where real material would yield locally and stop climbing. At a fillet, a thread root, or a bolt hole, that peak can sit well above yield in the model while the component as a whole is nowhere near collapse. Design codes handle this with stress categorization — splitting the result into primary, secondary, and peak stress and applying different allowables to each — but categorizing stress correctly in a complex geometry is itself a skilled, sometimes ambiguous exercise.
What elastic-plastic and limit-load analysis actually checks
Elastic-plastic analysis models the material’s real, non-linear stress-strain behavior — once a local region reaches yield, it stops carrying additional load proportionally and sheds it to surrounding material, exactly as real steel does. Limit-load analysis takes this further, incrementally increasing the applied load in the model until the structure as a whole loses the ability to carry any more — true plastic collapse, not a local stress reading. ASME BPVC Section VIII Division 2 Part 5 defines both routes explicitly: Protection Against Plastic Collapse (limit-load or elastic-plastic) and Protection Against Local Failure, each with its own acceptance criteria, as an alternative to the more conservative elastic stress-categorization route.
When you actually need it
- A linear-elastic result fails at a local stress concentration that engineering judgment says isn’t a real collapse risk — before redesigning the part, an elastic-plastic check may show it already has adequate margin.
- Geometry with unavoidable local stress risers — complex wellhead bodies with multiple bores, threaded connections, sharp internal transitions — where elastic stress categorization becomes genuinely ambiguous to apply correctly.
- Re-rating an existing, proven design to a higher pressure class, where the goal is finding the real margin that’s actually there, not the conservative margin a simpler method reports.
- Design optimization — reducing material or simplifying geometry where linear-elastic analysis would otherwise force an oversized, over-conservative design.
The trade-off
Elastic-plastic and limit-load analysis aren’t automatically the right call for every job. They take longer to run, need accurate non-linear material data, and produce a result that’s genuinely harder to review than a stress plot against a single allowable — which is exactly why the assumptions, boundary conditions, and load-incrementing approach need to be documented as carefully as the result itself. For a straightforward design with generous margin, a linear-elastic check is faster and just as defensible. The judgment call is knowing which situation you’re actually in before committing to either.
We run both linear-elastic and elastic-plastic/limit-load analysis for pressure-containing equipment, matched to what the design actually needs rather than defaulting to the more conservative method by habit. See our FEA & Simulation services, or talk to an engineer about a design that’s failing a linear-elastic check.