Analysis

When a hand calculation is enough, and when you need FEM

How to decide between a quick hand calculation and a finite element model for sheet-metal parts and brackets, with a worked bracket example, a decision table and the checks that keep an FEM result honest.

3 October 2026 · 8 min read

Finite element analysis is now available in most 3D CAD packages, and it is tempting to run it on everything. A coloured stress plot looks convincing. But for a large share of sheet-metal parts, a few lines of beam theory give an answer that is just as good, much faster, and far easier for someone else to check. Equally, there are parts where a hand calculation quietly misses the thing that will actually make them fail.

This article sets out where simple formulas are enough, where they run out, what FEM adds and how it can mislead, and how to use each method to check the other.

What a hand calculation does well

Most brackets, frames and supports behave like beams, plates or simple connections. If the load path is clear and the geometry is regular, classical formulas give stresses and deflections that are accurate enough for design, and the assumptions are visible on one page.

Beams and cantilevers: bending stress σ = M ÷ Z, shear, and deflection from standard cases.
Simple plates: tabulated solutions for rectangular and circular plates under uniform or point loads.
Connections: bolt shear and bearing, weld throat stress, net-section tension at a hole line.
Cold-formed sections: EN 1993-1-3 covers thin-gauge members such as channels, angles and lipped sections, including effective section properties that allow for local buckling of thin walls.

A hand calculation also forces you to understand the load path. You have to decide where the load goes, what holds the part, and which section is critical. That understanding is exactly what you need later to judge whether an FEM model is set up correctly.

A worked example: a plate cantilever bracket

Figure 1. A plate bracket bolted to a wall through a flange, carrying a load P at distance L. The bending moment is greatest at the root, and the root section is a rectangle t × h.

Take a bracket web in 5 mm S275 steel plate, 60 mm deep, with a design load of 1.5 kN applied 150 mm from the root. The load acts in the plane of the plate, so the web bends about its strong axis.

M = P × L = 1500 N × 150 mm = 225,000 N·mm
Z = t × h² ÷ 6 = 5 × 60² ÷ 6 = 3000 mm³
σ = M ÷ Z = 225,000 ÷ 3000 = 75 N/mm² (MPa)

Elastic section modulus of a solid rectangle. P is the factored design load.

S275 has a nominal yield strength of 275 MPa for plate up to 16 mm thick (EN 10025-2), so the utilisation in bending is 75 ÷ 275 ≈ 0.27, taking the partial factor γM0 as 1.0. The tip deflection is just as quick:

I = t × h³ ÷ 12 = 5 × 60³ ÷ 12 = 90,000 mm⁴
δ = P × L³ ÷ (3 × E × I) = 1500 × 150³ ÷ (3 × 210,000 × 90,000) ≈ 0.09 mm

E = 210,000 MPa for steel. This assumes a perfectly rigid root, which a bolted flange is not.

Now turn the same plate through 90° so the load acts across its thickness. Z becomes h × t² ÷ 6 = 60 × 5² ÷ 6 = 250 mm³ and the stress rises to 900 MPa, twelve times higher and well past yield. The ratio is simply h ÷ t. This is the kind of insight a hand calculation gives in seconds, and it is why sheet-metal brackets get their stiffness from flanges and folds rather than from thickness.

Notice what the calculation does not tell you: how much the flange and bolts let the root rotate, whether the bolt holes near the root raise the local stress, and whether the thin web could buckle sideways if it were much longer. For this short, stocky web those effects are small and a hand check is enough. For a long, thin web or a cut-out near the root, they may not be.

Where hand calculations run out

Beam theory gives the nominal stress. Real parts fail at the places where the stress is not nominal.

Stress concentrations: holes, slots, notches, bend reliefs and the inside of tight bends all raise local stress above M ÷ Z.
Local buckling and crippling: thin flanges and webs in compression, and webs under concentrated loads, can buckle long before the material yields. EN 1993-1-3 handles standard cold-formed shapes, but unusual folded geometry is harder to idealise.
Complex load paths: parts loaded in several directions at once, or where load shares between parallel members according to their stiffness.
Contact: parts that bear on each other, lift off, or clamp through bolts, where the contact area changes with load.
Fatigue: vibrating or cyclically loaded parts depend on the local stress range at welds, holes and edges, which needs either a detail category approach (EN 1993-1-9 for steel) or a reliable local stress.
Assemblies: the stiffness of welds, bolted joints and the supporting structure changes how load is shared.

Figure 2. A small round hole in a wide plate under tension. Across the section through the hole, the elastic stress rises to about three times the nominal stress at the hole edge.

Some of these can still be done by hand. Stress concentration factors for standard features are tabulated in reference books, and a factor of about 3 for a small round hole in a wide plate is a classic result. But when several features interact, or the geometry is not in any table, a numerical model becomes the sensible tool.

What FEM adds, and how it misleads

FEM solves the stiffness of the actual geometry. It shows where load really goes, finds local peaks around features, predicts buckling modes, and handles contact and assemblies. Used carelessly, it produces confident numbers that are wrong.

Singularities

At a perfectly sharp re-entrant corner, at a point load, or at a single fixed node, the theoretical elastic stress is infinite. FEM cannot show infinity, so it shows whatever the mesh allows: refine the mesh and the peak rises, refine again and it rises further. A stress that never settles is a modelling artefact, not a result.

Figure 3. Peak stress against element size, illustrative values. At a sharp corner the peak keeps climbing as the mesh is refined. With a real fillet radius modelled, it converges to a finite value above the nominal hand-calculated stress.

Other common pitfalls

Mesh: too coarse a mesh under-predicts peaks at holes and fillets. Refine locally and check that the result stops changing.
Boundary conditions: fixing a face completely makes it infinitely stiff, which is rarely true of a bolted flange or a wall. Over-stiff supports move stress around and hide flexibility.
Loads: applying a load to a single node or a tiny face creates its own singularity. Spread it over a realistic area.
Shell versus solid: sheet metal is usually best modelled with shell elements at mid-thickness. Solid elements need several elements through the thickness to capture bending, or they become far too stiff.
Linear versus nonlinear: a linear analysis that shows stress above yield, large deflection or parts in contact is outside its own assumptions.

Choosing the method

A guide, not a rule. The deciding factor is whether the critical stress can be found reliably by hand.
Situation	Hand calculation	FEM
Simple bracket or beam, clear load path, static load	Usually enough	Optional, for checking
Standard cold-formed section (channel, angle, lipped C)	EN 1993-1-3 methods	Rarely needed
Holes, slots or cut-outs in highly stressed areas	Nominal stress plus tabulated factors	Useful where features interact
Thin folded part in compression, unusual shape	Rough check only	Buckling analysis recommended
Several load cases in different directions	Possible but laborious	Efficient
Bolted or clamped parts with contact	Simplified checks	Recommended
Vibration or cyclic loading	Detail category approach	Often needed for local stress or natural frequencies
Weight or material optimisation	Good for first sizing	Useful for the final iteration

Using one to check the other

The best practice is not to choose one method but to use both. A hand calculation should come first, because it tells you what answer to expect. When the FEM result arrives, check it against that expectation before trusting any coloured plot.

Sum the reaction forces and confirm they equal the applied load in every direction.
Compare the deflection with the beam formula. It should be in the same range, usually a little larger because real supports are not rigid.
Read the stress away from holes and corners, where it should be close to M ÷ Z.
Check that peak stresses converge as the mesh is refined, and ignore peaks at sharp corners, point loads and fixed nodes.
Ask whether the deformed shape looks physically sensible, using an exaggerated scale.

If the FEM and the hand calculation disagree by more than you can explain, one of them is wrong, and it is worth finding out which before anything is released for manufacture.

Summary checklist

Start with a free body diagram and a hand calculation of the critical section.
If the critical stress sits in plain material with a clear load path, the hand calculation may be all you need.
Use FEM where holes, folds, buckling, contact, fatigue or assemblies control the design.
Model real fillet radii, realistic supports and spread loads, and check mesh convergence.
Reconcile the FEM result with the hand calculation and record both.

If you would like a second pair of eyes on a bracket calculation or an FEM model that does not quite add up, send us your drawings.

All insights