AcoustoSolver
Live Explorer

Real-time acoustic metamaterial simulator with physics-informed neural surrogates. Compute band structures via Plane Wave Expansion on square and hexagonal lattices, analyze Lamb wave plate dispersion, explore SSH topological edge states, and estimate transmission loss — all computed client-side in your browser.

Crystal Parameters

0.55.0 mm20
0.100.300.48
110 mm50
Fast60 k-ptsPrecise
Ready. Select materials and click Compute.

Bandgap Properties

Bandgap Width
Center Frequency
Lower Edge
Upper Edge
Impedance Contrast
Lattice Type
Max Group Velocity
Compute Time

SSH Topological Model

0.11.02.0
0.10.52.0
41030
Adjust t₁/t₂ to explore

Band Structure — Dispersion Relation

Unit Cell & Brillouin Zone

Group Velocity & Density of States

Lamb Wave Plate Dispersion — Rayleigh-Lamb Equation

Exact solution of the Rayleigh-Lamb frequency equation for the selected host material plate. Shows symmetric (S0, S1) and antisymmetric (A0, A1) Lamb modes — essential for plate-based metamaterial design.

SSH Topological Dispersion & Finite-Chain Edge States

Transmission Loss Estimate (N Unit Cells)

1520

Bandgap Map (Filling Fraction Sweep)

Sweeps filling fraction from 0.10 to 0.48 — may take 5-15 seconds

AcoustoSolver Architecture

Material Database 12+ acoustic materials
Unit Cell Square / Hexagonal lattice
PWE / FDTD / SSH Physics solvers
Neural Surrogate ResidualMLP + Physics Loss
Band Structure Bandgaps + Topology

Physics-Informed Loss

L = LMSE + λsym·Lsymmetry + λord·Lordering + λsm·Lsmoothness

Multi-component loss enforces band symmetry ω(k) ≈ ω(-k), monotonic ordering ωn+1 ≥ ωn, and smoothness between adjacent k-points.

Residual Backbone

Input(8) → Embed(256) → 4×ResBlock(256) → Output(n_k × n_bands)

Skip connections and layer normalization enable training on the non-convex physics-informed loss landscape.

SSH Topological Model

E(k) = ±√(t&sub1;² + t&sub2;² + 2t&sub1;t&sub2;cos(ka))

Zak phase γ = π when t&sub2;/t&sub1; > 1 (topological phase), enabling robust edge states protected by chiral symmetry.

← All ML/AI Projects View Source on GitHub