When you receive a finite element analysis (FEA) report from a third-party engineering firm, how can you be confident that it’s reliable?
Most project managers and clients aren’t trained in simulation. Yet they are often responsible for acting on FEA results—results that may affect structural integrity, cost, and safety. That’s why we developed a practical, 14-point checklist that anyone can use to perform a high-level review of an FEA report, which may help you to feel more confident in the FEA results.
This guide outlines what to look for, what questions to ask, and what red flags to raise.
Why High-Level FEA Checks Matter
An FEA report may look impressive—with colourful stress plots and dense terminology—but that doesn’t mean it’s correct. We’ve encountered cases where:
-
Boundary conditions didn’t match reality.
-
Loads were incorrectly applied or omitted.
-
The mesh was too coarse in critical areas.
-
Stress plots were interpreted without understanding yield limits.
In such cases, the final design can be either unconservative or over-conservative. A simple, structured review can help to limit this.
14 Key Things to Check in an FEA Report
1. Does the model match the real structure?
Check that the FEA model dimensions and thicknesses match the engineering drawings. Pay special attention to areas of concern. Ensure that tolerances (like corrosion or mill allowance) are accounted for.
2. Was the correct element type used?
-
Use shell elements when t/L < 0.1 (thin plates, tanks, sheet metal).
-
Use beam elements when members are long and slender (length > 10–20× cross-section).
-
Use solid elements when local 3D stress states, welds, or fatigue are of concern.
3. Are the boundary conditions realistic?
Ensure constraints don’t prevent natural movements (e.g., using fully fixed supports when simple supports are expected). Also, boundary conditions should not be too close to the critical area under evaluation.
4. Was the correct material used, with the correct units?
Verify material properties—especially yield strength, density, and elastic modulus—and ensure they match what’s specified in the design or datasheet.
5. Does the model mass match reality?
The model’s total mass (usually reported in the simulation summary) should be within ~5–10% of the actual expected mass. Watch out for missing components not included in the model.
6. Were all relevant loads applied?
This includes self-weight, pressure, wind, seismic, nozzle loads, and any live loads. If a load wasn’t applied, the report should state why (e.g., negligible effect). Check load magnitudes, directions, and locations.
7. Were reaction forces shown and explained?
Reaction forces should balance the applied loads. For example, if a pressure is applied on one side, the support reactions should match the resulting force. A mismatch suggests a modeling issue.
8. Was the mesh refined in critical areas?
Good practice suggests at least 2 elements through thickness in bending regions, and smaller elements in areas with holes, welds, or fillets. Check for jagged or irregular stress plots—this often indicates a coarse or poorly shaped mesh.
9. Are stress plots interpreted correctly?
Localized high stresses may not always mean failure. Understand whether the stress is due to self-limiting effects (like local bending) or from continuous loading (like pressure).
10. Were iso-clipped plots included?
These help visualize where stresses exceed a critical threshold. If stress spans the entire thickness, the average stress might exceed the allowable, which is critical in static analysis.
11. Was the deformed shape shown and explained?
The deformed shape should make intuitive sense. If the structure bends in the “wrong” direction, or displaces more than expected, check for incorrect loads or boundary conditions.
12. Is the magnitude of displacement realistic?
Even if the shape looks right, the magnitude might not. A 5 mm deflection is realistic for a bracket—500 mm is not.
13. Do the results stay within linear elastic limits?
If you’re using linear static analysis, check that stresses stay below yield and displacements are within the small-deflection assumption. Otherwise, the analysis type may be invalid.
14. Was anything verified with hand calculations?
Simple formulas like:
-
Hoop stress: σ=pr/t
-
Axial stress: σ=F/A
-
Bending stress: σ=My/I
What Else Is in our Checklist?
In addition to the 14 main points, our checklist includes:
-
Specific mesh quality metrics (e.g., Jacobian, aspect ratio)
-
How to interpret elemental vs. nodal stress plots
-
Guidance on contact modeling and symmetry
-
A reminder to check report quality and assumptions
Here is a full Check List for Static FEA’s: