Back to Projects

3D Materials AI / Topological Data Analysis

3D Topology-Aware Microstructure Screening Platform

Generate, upload, visualize, and rank 3D two-phase microstructures with morphology and topological descriptors before expensive FEM, CFD, or fabrication simulations.

Voxel Microstructure AnalysisVolumetric Field UploadConnectivity AnalysisDescriptor InterpretationCandidate Ranking

3D Microstructure Viewer

Rotatable solid-surface preview

Drag to rotate. Surface faces are shaded with a fixed overhead light.

Raw file dims:(only needed for .raw uploads)

54%

Solid Volume

phase occupancy

3

β₀ Components

solid connected regions

307

β₁ Tunnels

loops through solid phase

0

β₂ Voids

enclosed pore regions

-304

Euler χ

β₀ − β₁ + β₂

0.209

Surface Density

interface density estimate

Descriptor Interpretation

This candidate has Euler characteristic χ = -304 (β₀ = 3 solid components, β₁ = 307 tunnels, β₂ = 0 enclosed voids). It occupies 54% solid volume with 3 through-connected axises. For the transport objective, it is best described as multi-axis connectivity with balanced porosity.

Suggested next design move: Candidate is ready for downstream FEM/CFD-style simulation screening.

Topological Phase Diagram

Betti numbers and Euler characteristic computed across all thresholds. Each curve is normalized to its own maximum. The dashed line marks the applied threshold.

Fragmented
β₀ solid components (max 177)β₁ tunnels (max 310)β₂ enclosed voids (max 117)χ Euler (range -308→176)
t=0.480.000.250.500.751.00threshold →

Percolation

t = 0.43

Solid loses full connectivity above this threshold. Below it: bicontinuous. Above: fragmented islands.

Peak β₁

t = 0.47

Maximum tunnel density (310 loops). Best threshold for transport-optimized metamaterials — highest permeability proxy.

β₂ Onset

t = 0.43

Enclosed pore regions first appear. Below this threshold the solid traps void — relevant for closed-cell foams.

Spatial Correlation & Hyperuniformity

The two-point correlation S₂(r) measures the probability that two points separated by distance r are both in the solid phase. The structure factor S(k) is its Fourier transform; hyperuniform materials exhibit S(k→0)→0, suppressing large-scale density fluctuations—a key design target for phononic metamaterials.

Two-Point Correlation S₂(r) / φ̄²

φ̄²ξ≈1px1.01.50816r (voxels)

Correlation length ξ ≈ 1 voxels — distance at which correlations decay to halfway between peak and long-range limit.

Radial Structure Factor S(k) — log scale

k→011632k (wavenumber)
Non-hyperuniformS(k→0) ≈ 1.20e-1

Volume fraction φ̄

0.536

S₂(0) = φ̄ = 0.536

Correlation length ξ

1 vx

Characteristic microstructural length scale

Hyperuniformity

Not detected

S(k→0) suppressed vs bulk S(k)

Design To Additive Manufacturing Screening

Morphology descriptors and topology checks provide an early feasibility gate for microstructure designs before expensive simulation, fabrication, or inverse design refinement.

Feasible candidate

Design

Candidate 3D microstructure generated or uploaded as a thresholded two-phase volume.

Topology Gate

death threshold →birth →323 pairs
β₀ 3 / β₁ 307β₂ 0 / χ -304Objective: transportPaths 3/3

Additive Manufacturing

Ranked candidate prepared for downstream FEM/CFD simulation, printability checks, or inverse design refinement.

Screen Candidate Microstructures

Rank generated 3D candidates with descriptor-based pre-screening objectives. The highest-scoring structures are selected for downstream simulation pipelines.

Method Summary

The workflow combines 3D volume handling, microstructure descriptor extraction, topology-aware screening, spatial correlation analysis, and candidate interpretation before expensive simulation or additive manufacturing steps.

3D microstructure generation and volumetric field upload
Connectivity, surface, fragmentation, and transport descriptors
Ranked screening workflow for structure-property exploration