ANSYS Topology Optimization and Design Exploration: Parametric Studies and Weight Reduction

ANSYS Topology Optimization and Design Exploration: Parametric Studies and Weight Reduction

Topology optimization is one of my favorite features in ANSYS Workbench — there's something satisfying about watching the solver strip away material you don't need and leaving behind a structure that looks like it grew organically. I've used it on brackets, heat sinks, and aerospace ribs, and it consistently delivers 30-60% weight savings when done right. Let me show you how I set up both topology optimization and parametric studies.

Parametric Studies with DesignXplorer

Setting Up Parameters

1. Open Static Structural (or any analysis) in Workbench
2. In Mechanical, mark parameters:
   • Input parameters: Right-click a dimension > Parameterize
     • Geometry dimensions (thickness, radius, length)
     • Mesh element size
     • Load magnitude
     • Material properties
   • Output parameters: Right-click result > Parameterize
     • Maximum stress
     • Maximum deformation
     • Total mass
     • Safety factor
3. Parameters appear in Workbench Parameter Set

Design Points

1. Double-click Parameter Set in Workbench
2. Table of design points appears:
   • Each row = one design configuration
   • Input columns: Varied parameters
   • Output columns: Calculated results
3. Add design points:
   • Manually: Enter values for each input
   • Or use DOE (Design of Experiments)

Design of Experiments (DOE)

1. In Parameter Set > DOE
2. Select sampling method:
   • Central Composite Design (CCD): For response surface
   • Optimal Space-Filling: For broad exploration
   • Box-Behnken: For 3-4 parameters
   • Latin Hypercube Sampling: For many parameters
3. Set number of samples:
   • CCD: 2^k + 2k + 1 (k = number of parameters)
   • For 3 parameters: 15 samples
   • For 5 parameters: 43 samples
4. Workbench generates design points automatically

Response Surface

1. After running all design points:
2. Double-click Response Surface
3. Generate response surface:
   • Standard Response Surface: Polynomial fit
   • Kriging: Interpolation (more accurate for nonlinear)
   • Neural Network: For highly nonlinear responses
4. Visualize:
   • 3D surface: Two inputs vs. one output
   • 2D contour: Two inputs with output as color
   • Sensitivity: Which parameters have most influence
   • Local sensitivity: Bar chart of parameter influence

Optimization

1. Double-click Optimization in DesignXplorer
2. Set:
   • Objective: Minimize mass (or maximize safety factor)
   • Constraints: Maximum stress < yield, deformation < limit
   • Input parameter ranges: Min and max for each parameter
3. Select optimization method:
   • Screening: Simple, fast, approximate
   • MOGA (Multi-Objective Genetic Algorithm): For multiple objectives
   • NLPQL: For single objective, gradient-based
4. Run optimization
5. Results:
   • Candidate designs: Top 3 designs meeting objectives
   • Trade-off chart: Pareto front for multi-objective
6. Select best candidate and verify:
   • Insert as new design point
   • Run full analysis to confirm

Topology Optimization

Setup

1. Drag "Topology Optimization" to Project Schematic
2. Import geometry and define materials
3. Define boundary conditions and loads (same as static structural)
4. Define exclusion regions:
   • Preserved faces: Mounting surfaces, contact areas
   • Preserved bodies: Non-design regions (bolts, bearings)

Optimization Setup

1. In Topology Optimization:
2. Set objective:
   • Minimize mass: Reduce weight while meeting constraints
   • Minimize compliance: Maximize stiffness while meeting mass target
   • Maximize stiffness: For given mass fraction
3. Set constraints:
   • Mass retention: 30-60% of original mass (typical)
   • Maximum stress: Below yield (optional — may slow convergence)
   • Maximum displacement: Below allowable (optional)
4. Set manufacturing constraints:
   • Minimum member size: 3-5mm (avoid thin features)
   • Extrusion constraint: For extruded parts
   • Symmetry: Planar or cyclic symmetry
   • Demold direction: For cast parts (no undercuts)

Running Topology Optimization

1. Solve
2. ANSYS iterates:
   • Removes low-stress material
   • Retains high-stress material
   • Checks constraints
   • Converges to optimal material distribution
3. Monitor:
   • Mass fraction: Should converge to target
   • Compliance: Should decrease (stiffer structure)
   • Convergence: Should stabilize

Results

1. View topology:
   • Material distribution: Red = retained, blue = removed
   • Iso-surface: Smooth boundary at threshold (0.3-0.5)
2. Adjust display:
   • Threshold: Higher = more conservative (more material)
   • Clipping: Cut through to see internal structure
3. Export:
   • STL: For 3D printing
   • Geometry reconstruction: In SpaceClaim

Geometry Reconstruction

1. Export topology result to STL
2. Open SpaceClaim:
   • File > Open > STL file
3. Convert to solid:
   • Skin Surface: Wrap the STL mesh
   • Faceted: Keep as mesh body
   • Smooth: Smooth the rough topology surface
4. Clean up:
   • Remove small features
   • Add fillets at transitions
   • Re-add mounting features
5. Verify:
   • Run static structural on reconstructed geometry
   • Compare stress to original optimization

Lattice Structure Generation

Creating Lattice Structures

1. In SpaceClaim:
   • Insert > Lattice
2. Select cell type:
   • Gyroid: Smooth, isotropic (good for 3D printing)
   • Diamond: Similar to gyroid
   • Star: Stiff in multiple directions
   • Octet: Stiff and lightweight
   • Kelvin: Good thermal properties
3. Set parameters:
   • Cell size: 2-10mm (typical)
   • Wall thickness: 0.3-1.0mm
   • Density: 20-50% (volume fraction)
4. Apply to region:
   • Select body or face
   • Lattice fills the volume

Lattice Optimization

1. Combine topology optimization with lattice:
   • Outer skin: Solid (for mounting and loads)
   • Interior: Lattice (for weight reduction)
2. Set lattice density based on stress:
   • High-stress regions: Dense lattice (or solid)
   • Low-stress regions: Sparse lattice
3. Export for 3D printing:
   • STL with lattice structure
   • Direct to metal 3D printer (DMLS, SLM)

Practical Applications

Bracket Weight Reduction

1. Original: Steel bracket, 2.5 kg
2. Setup:
   • Preserve mounting holes and load application face
   • Objective: Minimize mass
   • Constraint: Max stress < 150 MPa (< yield 250 MPa)
   • Mass retention: 40%
3. Result: Optimized bracket, 1.0 kg (60% reduction)
4. Reconstruction: Smooth in SpaceClaim, add fillets
5. Verification: Max stress = 145 MPa (within limit)
6. Manufacturing: 3D print in aluminum or investment cast

Heat Sink Optimization

1. Original: Pin-fin heat sink, 200g
2. Setup:
   • Thermal analysis with heat load (50W)
   • Objective: Minimize mass
   • Constraint: Max temperature < 80°C
   • Design variable: Pin diameter, height, spacing
3. DOE: 15 design points (CCD with 3 parameters)
4. Response surface: Kriging
5. Optimization: MOGA (minimize mass, minimize temperature)
6. Result: Optimized heat sink, 120g (40% reduction), T = 78°C
7. Pareto front: Shows mass vs. temperature trade-off

Aerospace Rib Optimization

1. Original: Aluminum rib, 1.8 kg
2. Setup:
   • Preserve mounting edges and fuel passage
   • Objective: Minimize compliance (maximize stiffness)
   • Constraint: Mass < 1.0 kg (55% reduction)
   • Manufacturing: Milling (demold constraint = vertical)
   • Symmetry: Planar (about center plane)
3. Result: Optimized rib, 0.9 kg, 50% reduction
4. Reconstruction: Convert to machinable geometry
5. Verification: Stiffness within 5% of original, stress < yield

Verification of Optimized Designs

Static Structural Verification

1. Run full static structural on optimized geometry
2. Check:
   • Maximum stress: Must be below yield
   • Safety factor: Must be > 1.5
   • Deformation: Must be within allowable
3. If stress exceeds limit:
   • Increase mass retention in topology optimization
   • Add material at high-stress regions
   • Use higher strength material

Modal Verification

1. Run modal analysis on optimized geometry
2. Check:
   • Natural frequencies: Should not match excitation frequencies
   • Stiffness: Should be similar to original (if objective was stiffness)
3. If frequencies shift significantly:
   • Check if shift is beneficial (moved away from excitation)
   • Or detrimental (moved toward excitation)

Fatigue Verification

1. Run fatigue analysis on optimized geometry
2. Check:
   • Fatigue life: Must meet design life (e.g., 10⁶ cycles)
   • Stress amplitude: Must be below fatigue limit
3. Optimized designs may have new stress concentrations:
   • Add fillets at all transitions
   • Smooth surfaces to reduce notch effects

Verification Checklist

• [ ] Input parameters cover all design variables
• [ ] DOE samples cover the design space adequately
• [ ] Response surface accuracy is acceptable (R² > 0.95)
• [ ] Topology optimization preserved all critical interfaces
• [ ] Manufacturing constraints are applied (member size, demold, symmetry)
• [ ] Optimized geometry is reconstructed as a clean solid
• [ ] Static structural verification passes (stress < yield, SF > 1.5)
• [ ] Modal verification shows no new resonance issues
• [ ] Fatigue verification meets design life
• [ ] Weight reduction target is achieved

Wrapping Up

The biggest mistake I see people make with topology optimization is skipping the verification step. The optimized shape looks cool, but you need to run a full analysis on the reconstructed geometry — I've had cases where the stress was fine in the optimization but jumped 40% after I cleaned up the geometry and added fillets. Always verify. And don't forget manufacturing constraints — a beautiful organic shape that you can't actually machine or cast isn't much use. Set your member size limits, add your symmetry, and you'll get results you can actually manufacture.