FEA 101
Finite Element Analysis
(FEA) —The Fundamentals
Understanding the Physics Behind the
Software
From Stiffness Matrix to Stress
Interpretation
ANSHUL KUMAR
FEA 101
WHAT IS FINITE ELEMENT ANALYSIS
(FEA)?
Finite Element Analysis (FEA) is a numerical method used to approximate
solutions to complex engineering problems governed by differential
equations.
In structural mechanics, it solves the equilibrium equation:
∇·σ + b = 0
Where:
σ = stress tensor
b = body forces
and it predicts:
Displacement
Strain
Stress
Reaction forces
Instead of solving the governing equations for an entire complex
structure directly, FEA:
Divides the structure into smaller elements (finite number of
elements), hence the name finite element analysis.
Approximates displacement within each element using interpolation
(shape) functions.
Assembles a global system of equations
It transforms a continuous structure into a solvable algebraic system.
ANSHUL KUMAR
FEA 101
THE MATHEMATICAL FOUNDATION
The governing equation of linear static FEA:
[K]{u} = {F}
Where:
[K] = Global stiffness matrix
{u} = Nodal displacement vector
{F} = Applied load vector
The stiffness matrix is derived from:
Material properties (E, ν)
Geometry
Element shape functions
After assembling all element matrices into a global system
and applying boundary conditions, the system is solved for
displacements.
Strain = B{u}
Stress = D·Strain
FEA is fundamentally matrix algebra applied to mechanics.
ANSHUL KUMAR
FEA 101
DISCRETIZATION & ELEMENT
FORMULATION
The accuracy of FEA depends heavily on discretization.
Discretization (Meshing)
The structure is divided into:
Element Types
1D:
Truss (axial only)
Beam (axial + bending + shear)
2D:
Plane stress
Plane strain
Shell elements
3D:
Tetrahedral
Hexahedral
Each element contains:
Nodes
Degrees of freedom (DOFs)
Shape functions to interpolate displacement
Poor element selection can introduce:
Shear locking
Volumetric locking
Artificial stiffness
Mesh refinement improves resolution, but cannot correct modeling errors.
ANSHUL KUMAR
FEA 101
MATERIAL BEHAVIOR & MODELING
ASSUMPTIONS
FEA results are only as valid as the assumptions made.
Common assumptions in basic FEA:
Material Behavior
Linear Elastic:
σ = Dε
Valid below yield
Assumes proportional stress–strain relationship
Nonlinear Material:
Plasticity (yield surface, hardening rules)
Hyperelastic models
Creep and viscoplasticity
Geometric Assumptions
Small Deformation:
Linear strain-displacement
Constant stiffness matrix
Large Deformation:
Nonlinear strain
Stiffness becomes displacement-dependent
Material Directionality
Isotropic
Orthotropic (common in composites)
Fully anisotropic
Incorrect material modeling produces misleading stress predictions.
ANSHUL KUMAR
FEA 101
LOADS, BOUNDARY CONDITIONS &
LOAD PATHS
Loads and constraints define structural behavior.
Types of Loads:
Concentrated forces
Distributed pressures
Thermal loads
Inertial loads
Boundary Conditions:
Fixed supports
Symmetry
Contact interfaces
Multi-point constraints
Improper constraints can:
Artificially stiffen the structure
Prevent natural deformation
Distort load paths
Understanding load transfer through stiffness variations
is critical for correct stress interpretation.
ANSHUL KUMAR
FEA 101
STRESS TENSOR & FAILURE
INTERPRETATION
FEA outputs the full stress tensor:
σ= [ σx τxy τxz ]
[ τyx σy τyz ]
[ τzx τzy σz ]
From this tensor, we derive:
Principal stresses
Von Mises equivalent stress
Maximum shear stress
Failure depends on:
Yield criteria (Von Mises, Tresca)
Fracture mechanics
Fatigue crack growth
A high local stress may be acceptable if:
It is localized
It does not exceed allowable limits
It does not drive fatigue or crack propagation
Stress contours require engineering interpretation.
ANSHUL KUMAR
FEA 101
LIMITATIONS, VALIDATION &
ENGINEERING RESPONSIBILITY
FEA is an approximation method with inherent limitations:
Mesh dependency
Sensitivity to boundary conditions
Dependence on modeling assumptions
Linearization of complex behavior
Best practices include:
Mesh convergence studies
Sensitivity analysis
Correlation with hand calculations
Physical validation with test data
True simulation maturity means:
Understanding when the model is insufficient.
Recognizing when nonlinear, fatigue, or fracture analysis is
required.
Knowing that convergence does not imply correctness.
Software solves equations.
Engineers validate structural reality.
ANSHUL KUMAR