Projectech — Buy project kits, Learn, Earn, GrowOrder a Custom Project
Home/Guides/Mechanical/FEA Basics: Stress Analysis for Students

FEA Basics: Stress Analysis for Students

A rainbow stress plot means nothing without setup discipline. This guide covers the FEA workflow — defeaturing, materials, meshing, mesh convergence studies, fixtures, loads — how to read von Mises results honestly, handle singularities, and validate against a hand calculation, with a worked L-bracket example.

Written by Projectech7 min readPublished
For B.E./B.Tech Mechanical students running their first stress simulation for a design or final-year project Topics: FEA, Simulation, SolidWorks
Illustration of a finite element analysis: a meshed L-bracket with a von Mises stress rainbow plot, mesh convergence graph, and factor of safety callout.
Illustration generated for this guide.
In this guide

Your bracket design is finished in CAD, and your guide or faculty member says "run an FEA on it." You mesh it, apply some loads, and get a beautiful rainbow plot with a maximum stress of 412 MPa. Is the bracket safe? Over-designed? Or is the number meaningless? Without understanding what the software did, an FEA result is just colorful decoration — and examiners know exactly which questions expose that.

This guide explains FEA the way a student needs it: what the solver is actually doing, how to set up a trustworthy static stress analysis, how to read the results honestly, and the validation steps that separate engineering from picture-making.

What FEA actually computes

Finite Element Analysis breaks your continuous part into small elements (the mesh) connected at nodes. For linear static analysis, it assembles and solves one giant equation:

[K]{u} = {F}

  • [K] = global stiffness matrix (from geometry, material, and mesh)
  • {u} = nodal displacements (what the solver finds)
  • {F} = applied loads

From displacements, it derives strains, then stresses via Hooke's law. That's it — everything else (pretty plots, animations) is post-processing. Two consequences students must internalize:

  1. The answer is only as good as the mesh, the material model, and the boundary conditions. The solver never questions your inputs — garbage in, rainbow out.
  2. Linear static FEA assumes small deflections and linear-elastic material. If your part yields significantly, buckles, or has contact/gaps changing under load, you need nonlinear analysis — and you should say so in your report rather than silently running a linear study.

The workflow, in order

1. Simplify the geometry (defeature)

Remove what doesn't affect strength: tiny fillets, cosmetic chamfers, logos, small holes far from load paths. Each defeatured detail cuts mesh count and solve time. But keep fillets at stress concentrations you're actually evaluating — removing the fillet where the bracket bends removes the answer you're looking for.

2. Assign real material properties

You need Young's modulus (E), Poisson's ratio (ν), and yield strength at minimum. Use the actual grade: AISI 1020 (E ≈ 200GPa, yield ≈ 350MPa) is not "steel" generically — and the difference between 250MPa and 350MPa yield decides your factor of safety. For the report, cite the material data source.

3. Mesh: the most important step

Choice Guidance
Element type Second-order (parabolic) tetrahedral or hexahedral for stress; first-order tets are overly stiff — avoid for final results
Global size Start coarse; refine in the region of interest
Refinement Local mesh controls at fillets, holes, and section changes — 3–4 elements across a fillet radius minimum
Aspect ratio Keep elements chunky; long sliver elements corrupt results

Mesh convergence study (do this — it's the credibility of your whole analysis): run the same study at three mesh densities (coarse, medium, fine) and plot max stress vs element count. When the stress stops changing significantly (<5–10% between refinements), the mesh is converged. A single mesh density with no convergence check is not a result, it's a guess. Put the convergence plot in your report — examiners look for it.

4. Boundary conditions: fixtures

Fixtures must represent reality. A bracket bolted to a wall is not a fully fixed face unless the wall is infinitely stiff — but for student work, fixed faces on bolted interfaces are the standard simplification; state it as an assumption. Common errors:

  • Over-constraining (fixing faces that in reality can slide) → artificially low stress, artificially high stiffness.
  • Under-constraining → rigid-body motion, solver fails or gives nonsense.
  • Forgetting that a "fixed" bolt hole in reality allows micro-rotation — your peak stress at a perfectly-fixed hole edge is partly a modeling artifact (see singularities below).

5. Loads

Apply loads as they really act: pressure on a face, force at a bolt circle, torque about an axis, gravity for self-weight on large parts. Always do a hand calculation first — even a rough one. If FEA says 412MPa and your hand calc says ~80MPa, don't "trust the software"; find the discrepancy (wrong units on the load is the classic: newtons vs kilonewtons).

Warning: FEA tells you about the model, not the part. A converged, beautiful result with wrong boundary conditions is precisely wrong. The hand calculation is your reality check — never submit FEA without one, and never tune the model until it matches the hand calc (that's not validation, that's decoration).

Reading results honestly

Von Mises stress vs yield

For ductile metals, compare the von Mises equivalent stress against yield strength:

Factor of Safety = Yield strength / Max von Mises stress

Design targets: FoS ≥ 2 for static student-project structures with well-known loads; ≥ 3 where loads are uncertain, impact is possible, or failure hurts someone. FoS < 1 means predicted yielding — redesign, don't "accept" it.

Stress singularities: the spike that isn't real

Sharp internal corners, point loads, and perfectly-fixed constraint edges produce stresses that increase forever as you refine the mesh — they never converge. This is a mathematical artifact of linear elasticity, not a real stress. Real parts have some radius, some local yielding. How to handle:

  • Add the real fillet radius and mesh it properly — the stress will converge to a genuine concentration.
  • If the spike is at a fixture or load application point (not the region of interest), ignore the local peak and read stress a short distance away (Saint-Venant's principle).
  • Report what you did: "peak stress at the fixed edge is a singularity; reported stress is taken 5mm from the constraint."

An examiner who asks "why does your max stress keep rising with mesh refinement?" is testing exactly this.

Displacement and sanity checks

  • Deformed shape: does it bend the way you'd expect? A bracket that deforms sideways under a vertical load has a setup error.
  • Reaction forces: sum of reactions must equal applied loads (equilibrium). Most solvers report this — check it.
  • Displacement magnitude: 15mm deflection on a 100mm steel bracket under 500N is suspicious — recheck units and E.

Worked example: L-bracket

A 100×100×10mm steel L-bracket (AISI 1020, yield 350MPa), vertical leg bolted to a wall (fixed), 2000N downward load on the horizontal leg tip, 60mm from the bend.

Hand calc: bending moment at the bend M = 2000N × 0.06m = 120Nm. Section: b = 0.1m, h = 0.01m. I = bh³/12 = 0.1 × 1e-6/12 ≈ 8.33e-9 m⁴. σ = My/I = 120 × 0.005/8.33e-9 ≈ 72MPa. FoS ≈ 350/72 ≈ 4.9.

FEA: expect peak ~70–110MPa at the inner fillet (stress concentration factor ~1.5 on the nominal 72MPa, depending on fillet radius) — converging with mesh refinement. If your FEA reports 400MPa here, something is wrong with the setup (check: load units, fixture faces, material E). If it reports ~90MPa converged, the hand calc and FEA agree, and you have a defensible result.

What to put in the report

  • Objective and scope (what question the analysis answers)
  • Geometry simplifications and assumptions (fixtures, loads, material source)
  • Mesh details + convergence study (plot or table)
  • Hand calculation for comparison
  • Stress plot (von Mises), displacement plot (scaled, stated scale factor)
  • FoS with the design target and verdict
  • Limitations: linear static, no fatigue/buckling/thermal (unless analyzed), singularity treatment

Common mistakes

  1. No mesh convergence study — the single biggest credibility gap.
  2. No hand calculation — nothing to catch unit or setup errors.
  3. Reporting singularity spikes as the design stress.
  4. First-order tet mesh for final stress results (overly stiff, under-predicts).
  5. Over-constrained fixtures that don't match reality.
  6. Applying the load in wrong units (kN entered as N — factor-of-1000 errors).
  7. Forgetting gravity/self-weight on large assemblies.
  8. Claiming the design is "validated" from one linear static run — it's checked against static yielding, nothing more.

Where to go from here

More project guides

More in Mechanical

Heat Treatment Basics for Students: Annealing, Quenching and Tempering

Heat treatment turns soft steel into hard tools or tough parts by controlling how steel is heated and cooled. This guide explains annealing, normalising, quenching and tempering for student workshops: the temperatures, the quenching media, how to judge temper colours, how to check hardness, and the safety rules that keep the process under control.

Read guide

PID Control Tuning: A Practical Guide

Your robot oscillates, overshoots, or never quite reaches the target — that is a PID tuning problem. This guide explains what the P, I, and D terms actually do, gives a hands-on tuning procedure (P, then D, then I), and covers loop timing, noise, and windup.

Read guide

Vibration Measurement Basics

Machines talk through vibration — this guide teaches you to listen with numbers: displacement/velocity/acceleration, sensor selection and mounting, FFT fault signature table (1×, 2×, BPFO, gear mesh), ISO 10816 severity zones, and a seeded-fault student project setup.

Read guide