FEM software — NASTRAN, ABAQUS (concepts and use)

What FEM Software Does

The Core Workflow

Pre-processing: Build geometry, define materials, apply mesh (divide structure into elements)
Solver: Assemble stiffness matrices, apply boundary conditions, solve $\mathbf{K}\mathbf{u} = \mathbf{F}$
Post-processing: Visualize stress contours, deformation, safety margins

Why this structure? Because the physics (elasticity PDEs) → discretization (FEM) → linear algebra (solve) → engineering decisions (interpret) are separate concerns.

NASTRAN: The Aerospace Workhorse

NASTRAN (NASA Structural Analysis) was developed in the 1960s for the Apollo program. Philosophy: fast, reliable, linear analysis for large aerospace structures. Text-based input deck (.bdf Bulk Data File), proven algorithms, extensive validation heritage.

Why spacecraft engineers love it:

Launch vehicle qualification: proven for dynamic loads (random vibration, shock, acoustic)
Modal analysis: finds natural frequencies to avoid resonance
Linear = fast: analyze 1million DOF models in minutes

Key Capabilities

Analysis Type	What It Does	When You Use It
Linear Static	$\mathbf{K}\mathbf{u} = \mathbf{F}$ , find displacements/stresses	Steady loads (gravity, thermal)
Modal	Eigenvalue problem $(\mathbf{K} - \omega^2\mathbf{M})\boldsymbol{\phi} = 0$	Find vibration modes, avoid resonance
Frequency Response	Solve at each frequency $\omega$ : $(-\omega^2\mathbf{M} + i\omega\mathbf{C} + \mathbf{K})\mathbf{u} = \mathbf{F}(\omega)$	Harmonic forcing (rotating machinery)
Random Vibration (PSD)	Statistical response to random loads	Launch vibration environments

Formula The Static Equilibrium Equation

$\mathbf{K}\mathbf{u} = \mathbf{F}$ Why this form? Start from continuum mechanics: stress equilibrium $\nabla \cdot \boldsymbol{\sigma} + \mathbf{f} = 0$ . Apply virtual work principle: $\int_V \boldsymbol{\sigma} : \delta\boldsymbol{\epsilon} \, dV = \int_V \mathbf{f} \cdot \delta\mathbf{u} \, dV + \int_S \mathbf{t} \cdot \delta\mathbf{u} \, dS$ Discretize: $\mathbf{u} = \mathbf{N}\mathbf{u}_e$ (shape functions), $\boldsymbol{\epsilon} = \mathbf{B}\mathbf{u}_e$ (strain-displacement), $\boldsymbol{\sigma} = \mathbf{D}\boldsymbol{\epsilon}$ (Hooke's law). Substitute: $\int_V \mathbf{B}^T\mathbf{D}\mathbf{B} \, dV \cdot \mathbf{u}_e = \mathbf{f}_e$ The integral is the element stiffness matrix $\mathbf{k}_e = \int \mathbf{B}^T\mathbf{D}\mathbf{B} \, dV$ . Assemble all elements: $\mathbf{K} = \sum \mathbf{k}_e$ , $\mathbf{F} = \sum \mathbf{f}_e$ . Result: global system $\mathbf{K}\mathbf{u} = \mathbf{F}$ .

Problem: An L-bracket (aluminum, 100mm × 50mm × 3mm) is bolted at one end. Find first 3 natural frequencies.

NASTRAN Input Deck Structure:

$ Executive Control
SOL 103          $ Modal analysis
$ Case Control
METHOD = 1       $ Eigenvalue method
SPC = 10         $ Boundary conditions
$ Bulk Data
EIGRL,1,,,10     $ Lanczos, find 10 modes
GRID,1,0,0,0   $ Node 1 at origin
...
CQUAD4,1,1,2,3,4  $ Quad element
MAT1,1,70.E9,,2700  $ E=70GPa, rho=2700kg/m³
SPC1,10,123456,1    $ Fix all DOF at node 1

Why these cards?

SOL 103: Solution sequence for modal analysis
EIGRL: Eigenvalue solver (Lanczos method for large sparse matrices)
CQUAD4: 4-node shell element (good for thin structures)
MAT1: Isotropic material (Young's modulus, density)
SPC1: Single-point constraint (boundary condition)

Interpretation: Output shows $f_1 \approx 850$ Hz (first bending mode). If launch vibration has energy at 850 Hz, add stiffener or damper.

ABAQUS: The Nonlinear Specialist

When Linear Isn't Enough

ABAQUS (from Dassault Systèmes) handles nonlinearities:

Material nonlinearity: Plasticity, hyperelasticity (rubber seals), composite damage
Geometric nonlinearity: Large deformations (deployable boms), buckling
Contact: Bolted joints, friction, separation

Why spacecraft need this:

Composite layup analysis: delamination, fiber failure modes
Deployable mechanisms: hinges, cables, large rotations
Crush simulations: landing gear impact

Formula Nonlinear Equilibrium (Newton-Raphson)

Linear assumes $\mathbf{K}$ is constant. Nonlinear: stiffness depends on displacement. $\mathbf{R}(\mathbf{u}) = \mathbf{F}_{\text{ext}} - \mathbf{F}_{\text{int}}(\mathbf{u}) =0$ Why iterate? Solve by Newton-Raphson:

Start with guess $\mathbf{u}^{(0)}$
Compute residual: $\mathbf{R}^{(i)} = \mathbf{F}_{\text{ext}} - \mathbf{F}_{\text{int}}(\mathbf{u}^{(i)})$
Compute tangent stiffness: $\mathbf{K}_T = \frac{\partial \mathbf{F}_{\text{int}}}{\partial \mathbf{u}}\bigg|_{\mathbf{u}^{(i)}}$
Solve correction: $\mathbf{K}_T \Delta\mathbf{u} = \mathbf{R}^{(i)}$
Update: $\mathbf{u}^{(i+1)} = \mathbf{u}^{(i)} + \Delta\mathbf{u}$
Repeat until $\|\mathbf{R}\| < \text{tol}$ Why this works: Taylor expansion: $\mathbf{R}(\mathbf{u} + \Delta\mathbf{u}) \approx \mathbf{R}(\mathbf{u}) + \mathbf{K}_T\Delta\mathbf{u} =0$ , so $\mathbf{K}_T\Delta\mathbf{u} = -\mathbf{R}$ drives residual to zero.

Example: Composite Panel Progressive Failure

Problem: A carbon fiber panel (8 plies, quasi-isotropic layup) under compression. When does it fail?

ABAQUS Approach:

Material model: Hashin damage criteria (fiber tension/compression, matrix cracking)
Elements: Continuum shell elements with layup definition
Analysis: Static Riks (arc-length method for post-buckling)
Output: Load-displacement curve, damage initiation locations

Why this step-by-step?

Step 1: Hashin checks stress in fiber/matrix separately: $\left(\frac{\sigma_{11}}{X_T}\right)^2 + \alpha\left(\frac{\tau_{12}}{S}\right)^2 \ge 1$ (fiber tension failure)
Step 2: Each ply is a separate layer with orientation (0°, 45°, -45°, 90°)
Step 3: Riks method follows equilibrium path even after peak load (snap-through)
Step 4: Color contour shows "failure index" = 1.0 where damage starts

Result: Panel fails at 12.3 kN (matrix cracking at ply interfaces), then catastrophic fiber breakage at 14.1 kN. Design margin insufficient—add plies or change layup.

Worked example Bolted Joint Contact Analysis

Setup: Two aluminum plates (5mm thick) joined by M6 bolt, torqued to 10 N·m. Find contact pressure distribution and slip. ABAQUS Input:

*SURFACE INTERACTION, NAME=CONTACT: Define friction coefficient μ = 0.2
*CONTACT PAIR: Specify master/slave surfaces (bolt head/plate)
*STEP, NLGEOM=YES: Nonlinear geometry (bolt preload changes stiffness)
*CLOAD, OP=NEW: Apply bolt preload as equivalent force $F = \frac{T}{0.2d} \approx 8333$ N Why each setting?
Friction matters: without it, plates slide freely (unphysical)
NLGEOM=YES: bolt tension stiffens the joint (geometric nonlinearity)
Preload formula: $T = kFd$ , where $k \approx 0.2$ for lubricated threads, $d = 6$ mm Output: Contact pressure peaks at 180 MPa (within bearing strength), no slip under 5 kN shear load. Joint is safe.

Practical Workflow Differences

| Aspect | NASTRAN | ABAQUS | |-----|------| | Input | Text deck (.bdf) | Text deck (.inp) or CAE GUI | | Speed | Fast for linear (optimized solvers) | Slower (iterative nonlinear) | | Typical Model Size | 1-10M DOF | 10k-1M DOF (nonlinear overhead) | | Best For | Spacecraft primary structure, launch loads | Composites, deployables, detailed joints | | Validation |50+ years of flight heritage | Industry standard for crash/impact |

Why text decks? Version control, automation (parametric studies), reproducibility. GUI is great for learning, but production analysis uses scripts.

Software Selection Decision Tree

When to use NASTRAN:

Linear analysis (most spacecraft structures are linear elastic)
Large models (millions of DOF)
Frequency domain (random vibration, acoustic)
Heritage/certification requirements (NASA, ESA standards)

When to use ABAQUS:

Material nonlinearity (composites with damage, plasticity)
Geometric nonlinearity (post-buckling, large rotations)
Contact problems (bolted joints, seals, deployables)
Explicit dynamics (impact, shock, crash)

When to use both:

Global model in NASTRAN (linear, fast) → extract forces → local detailed model in ABAQUS (nonlinear joint)

Recall Explain Like I'm 12

Imagine you're building a trehouse. You want to know: Will the floor hold my weight? Will the wind blow it down?

You could build it and see what breaks—but that wastes wood and is dangerous. Instead, you build a computer model: draw the treehouse, tell computer it's made of wood (which bends a little under force), and say "I weigh 50 kg and the wind pushes at 60 km/h."

The computer chops the treehouse into tiny pieces (like LEGO bricks), figures out how each piece bends and pushes on its neighbors, and calculates the total. It tells you: "The floor bends 2 cm (safe!) but the rope on the left is at90% of breaking strength (add a second rope!)."

NASTRAN is like a super-fast calculator that's perfect for simple materials (wood, metal) and small bends. ABAQUS is like a scientist that can handle weird stuff: rubber bands that stretch a lot, pieces that crash together, or materials that crack. Spacecraft engineers use both: NASTRAN for the big structure, ABAQUS for the tricky parts (hinges, seals, composites).

Verification and Validation

Standard checks:

Equilibrium: $\sum F = 0$ , $\sum M = 0$ (reactions balance applied loads)
Energy: Strain energy $U = \frac{1}{2}\mathbf{u}^T\mathbf{K}\mathbf{u}$ should equal work $W = \mathbf{F}^T\mathbf{u}$
Symmetry: If geometry/load are symmetric, deformation should be too
Limiting cases: Slender beam → Euler-Bernoulli formula, thin plate → Kirchhoff theory

Connections to Spacecraft Design

Structural Analysis Methods: FEM is one method; compare to analytical solutions, testing
Vibration and Modal Analysis: NASTRAN modal = find $\omega_n$ to avoid resonance with Launch Vehicle Loads
Composite Materials: ABAQUS layup analysis for Carbon Fiber Structures
Thermal-Structural Coupling: Export NASTRAN temps → thermal stresses (or coupled analysis)
Stress Analysis and Margins: FEM outputs stresses → apply Factors of Safety → design margins
Model Correlation: Match FEM predictions to Ground Test Procedures data

#flashcards/physics

What is the fundamental equation solved in linear FEM? :: $\mathbf{K}\mathbf{u} = \mathbf{F}$ , where $\mathbf{K}$ is the global stiffness matrix (assembled from element stiffness matrices $\mathbf{k}_e = \int \mathbf{B}^T\mathbf{D}\mathbf{B} \, dV$ ), $\mathbf{u}$ is the displacement vector, and $\mathbf{F}$ is the applied load vector. It represents discretized equilibrium.

Why is NASTRAN preferred for spacecraft primary structures?

Fast linear analysis with validated solvers, excellent for large DOF models (1-10M), proven heritage for launch loads and frequency-domain analysis (modal, random vibration), industry-standard for certification (NASA, ESA).

What does ABAQUS's Newton-Raphson solver do differently than linear FEM?

Iteratively solves

\mathbf{R}(\mathbf{u}) = \mathbf{F}_{\text{ext}} - \mathbf{F}_{\text{int}}(\mathbf{u}) = 0

by computing tangent stiffness

\mathbf{K}_T

, solving

\mathbf{K}_T\Delta\mathbf{u} = \mathbf{R}

, and updating

\mathbf{u}

until convergence. Handles nonlinear material/geometry/contact where stiffness depends on displacement.

What is a mesh convergence study and why is it critical?

Rerun analysis with progressively finer mesh (2×, 4× elements) until key outputs (peak stress, max displacement) change less than 5%. Critical because FEM is an approximation—too coarse a mesh mises stress concentrations, underestimates deformations. Verification step, not validation.

Name three nonlinearities that require ABAQUS over NASTRAN :: (1) Material nonlinearity: plasticity, composite damage, hyperelasticity; (2) Geometric nonlinearity: large deformations, post-buckling, snap-through; (3) Contact: friction, separation, bolted joints. NASTRAN is linear elastic only.

How do you derive the element stiffness matrix from first principles?

Start with virtual work:

\int_V \boldsymbol{\sigma}:\delta\boldsymbol{\epsilon} \, dV = \delta W

. Substitute FEM discretization:

\mathbf{u} = \mathbf{N}\mathbf{u}_e

(shape functions),

\boldsymbol{\epsilon} = \mathbf{B}\mathbf{u}_e

(strain-displacement),

\boldsymbol{\sigma} = \mathbf{D}\boldsymbol{\epsilon}

(Hooke's). Result:

\mathbf{k}_e = \int_V \mathbf{B}^T\mathbf{D}\mathbf{B} \, dV

What is the modal analysis eigenvalue problem?

(\mathbf{K} - \omega^2\mathbf{M})\boldsymbol{\phi} = 0

, where

\mathbf{K}

is stiffness,

\mathbf{M}

is mass matrix,

\omega

are natural frequencies (rad/s), and

\boldsymbol{\phi}

are mode shapes. Finds frequencies to avoid resonance with launch vibration environments.

Why do text input decks remain standard despite GUI availability?

Version control (track changes in git), automation (parametric sweps via scripts), reproducibility (exact rerun), collaboration (code review), integration with optimization tools. GUIs are good for learning, decks for production.

Concept Map

Hinglish (regional understanding)

Intuition Hinglish mein samjho

Spacecraft structures ko design karte waqt, aap physically prototype bana kar test nahi kar sakte har baar—bohot expensive hai aur time bhi lagta hai. Isliye FEM software use karte hain jo virtual testing allow karta hai. NASTRAN aur ABAQUS aerospace industry ke do main workhorses hain. NASTRAN NASA ne 1960s mein Apollo program ke liye develop kiya tha—yeh linear analysis mein expert hai, matlabagar apka material (aluminum, steel) thoda