Acoustics and Vibrations (NVH) using FEA: An Overview
Acoustics and Vibrations (NVH): Modal and Frequency-Response Analyses for Resonance Prevention and Cabin Comfort
Why dynamic stiffness and frequency separation govern structural integrity and acoustic comfort
Unlike quasi-static structural evaluations—where external loads are applied slowly enough that inertial accelerations are negligible—real-world industrial machinery and transportation systems operate under continuous time-varying excitation. Whenever a cyclic forcing frequency coincides with an inherent structural natural frequency, energy transfers into kinetic oscillation with minimal dissipation, amplifying displacements and cyclic stresses far beyond static predictions. This phenomenon, structural resonance, is an intrinsic property of mass distribution, structural stiffness, and boundary conditions rather than a localized machine defect: rebalancing or realigning a rotating assembly cannot eliminate vibration if the supporting baseplate, foundation, or piping span resonates at running speed.
In automotive and mobility engineering, these structural dynamics directly govern Noise, Vibration, and Harshness (NVH)—the study of audible sound (noise), mechanical oscillations (vibration), and subjective ride quality (harshness). Ground reactions and powertrain forces propagate through suspension mounts and chassis rails into interior body panels, exciting cavity resonances as structure-borne noise. With the transition to electric vehicles (EVs), the absence of internal combustion engine (ICE) masking noise has elevated NVH simulation to a primary design gate: high-frequency electromagnetic tones from electric drive units (EDUs), road-load vibration, and wind noise become immediately audible unless controlled early through coupled finite element analysis (FEA) and acoustic modeling.
Modal vs. Forced Response
- Modal Analysis (Free Vibration): Identifies at what frequencies a structure will resonate and how it deforms (mode shapes). It does not yield absolute displacement, stress, or force magnitudes.
- Frequency-Response Analysis (Forced Vibration): Uses extracted modes and damping values to predict actual stress, velocity, and acoustic radiation magnitudes under defined cyclic or random loads.
- Model Debugging Role: Inspecting the first six zero-frequency rigid-body modes detects unconstrained bodies, missing contacts, or flying parts before running expensive dynamic sweeps.
Solving the eigenvalue problem and harmonic frequency response to prevent structural resonance

Eigenvalue extraction, mode shapes, and model connectivity checks
Modal analysis establishes the foundation for all linear dynamic workflows. Finite element software discretizes a structure's geometry, material elasticity, and boundary constraints into global mass (M) and stiffness (K) matrices. Under undamped free vibration, the equation of motion simplifies to a homogeneous system whose harmonic solution yields the generalized eigenvalue problem. Each resulting eigenvalue (λi) corresponds to the square of a circular natural frequency, while its paired eigenvector (φi) defines the spatial deformation pattern—the mode shape—at that frequency.
For standalone resonance screening, engineers typically extract 6 to 20 modes; when feeding modal frequency-response or modal transient simulations, the extraction upper bound is extended to at least 1.5× the highest loading or response frequency of interest. Solver selection scales with model size: the Lanczos method (EIGRL) computes exact eigenvalues and full eigenvectors efficiently for modest mode counts, whereas the Automated Multi-level Sub-structuring Eigenvalue Solution (AMSES / AMLS) computes approximate modal spaces up to 100× faster on multi-million-DOF NVH meshes when output is restricted to roughly 100 target degrees of freedom rather than every mesh node.
Governing Equations · Free Vibration[M]{ü} + [K]{u} = {0}| [K] − λ[M] | = 0 → λi = ωi2fi =√λi2π
- [M], [K] Global mass and structural stiffness matrices
- λi, φi Eigenvalue (ω2) and mode shape eigenvector
- RESVEC Static residual vectors ortho-normalized with modes for external FRF
A full-assembly modal FEA of an industrial pump—encompassing shaft, impeller, motor, coupling, and support skid—verified natural frequencies against running speed. Confirming zero structural modes within ±10% of operating speed ensured API 610 compliance prior to commissioning.Field Diagnostics & VFD SweepsBaseplates, Piping, and Variable SpeedsBaseplates rocking at 1,780 CPM or decks bouncing at 900 CPM persist even after precision balancing when running speed aligns with a structural mode. Equipment on variable-frequency drives (VFDs) sweeps through speed bands on every ramp, requiring modal mapping across the entire drive envelope.Civil & Megastructure PrecedentsHistorical Resonance & DampingThe Tacoma Narrows Bridge (1940) collapsed from wind-induced aeroelastic flutter, and the 1985 Mexico City earthquake destroyed 6- to 15-story (especially 9-story) buildings matching lakebed frequencies. Modern towers counter this via tuned mass dampers (Taipei 101 pendulum) or aerodynamic profile shaping (Burj Khalifa).
Harmonic frequency-response analysis (FRF), damping, and multi-body coupling
When design constraints prevent shifting every structural mode completely outside a broad operating envelope, engineers perform harmonic (frequency-response) analysis to quantify the severity of forced vibration. Periodic excitations—such as rotor mass unbalance, electromagnetic stator pull, or blade-pass pressure pulses—are applied over a swept frequency band. Rather than solving millions of physical coordinates directly at every frequency step, modal superposition transforms the system into modal coordinates spanned by the extracted eigenvectors, reducing computation while capturing resonance peaks accurately.
At resonance, undamped theoretical amplitude approaches infinity; in physical hardware, structural and joint damping (ζ) caps peak magnification and introduces a phase shift across the natural frequency. Computed cyclic equivalent stresses are benchmarked against the material's endurance limit to verify high-cycle fatigue safety. To correlate virtual predictions with operating hardware, Operating Deflection Shape (ODS) testing measures forced motion under service loads; when an ODS pattern aligns in shape and frequency with a modal eigenvector, resonance is proven unambiguously. Furthermore, exporting reduced modal properties as Modal Neutral Files (MNF) via Component Mode Synthesis (Craig-Bampton) integrates flexible FEA subassemblies into Multi-Body Simulation (MBS) driveline and suspension models.
Harmonic & FRF Process
- 1. Boundary Constraints: Fix mounting feet, bearing stiffnesses, and contact interfaces to match operational clamping.
- 2. Periodic Load Sweep: Apply sinusoidal forces (unbalance, gear mesh, EMS) up to the target excitation ceiling.
- 3. Damping Specification: Assign modal or Rayleigh damping ratios (ζ) governing peak resonance amplification.
- 4. Endurance Evaluation: Compare peak von Mises cyclic stress against the S-N endurance threshold.
Frequency Response Function (FRF) & Damping Attenuation
Resonant Hit (120 Hz)2.
Safe Margin (75 Hz)3. Damped Mount (ζ=0.11)
Excitation Frequency (fexc)120 Hz
Modal Damping Ratio (ζ)0.025 (2.5%)
Current Damped FRF Response
Lightly Damped Baseline (ζ = 0.02)
Material Endurance Stress Limit (140 MPa)
ENDURANCE LIMIT · 140 MPaf₁ = 120 Hzf₂ = 295 Hz
CYCLIC STRESS (MPa)08016024032020 Hz120 Hz220 Hz320 Hz420 Hz
EXCITATION FREQUENCY (Hz)
Dynamic Amplification20.0×Peak Q ≈ 1 / (2ζ)
Cyclic Stress248 MPaEndurance limit: 140 MPaPhase Lag (θ)90°Crosses 90° at resonance
Fatigue Margin-77.3%Relative to endurance limit
ENDURANCE EXCEEDEDAt 120 Hz with ζ = 0.025, dynamic amplification (20.0×) drives cyclic stress to 248 MPa, exceeding the 140 MPa fatigue endurance limit.
Cyclic Fatigue Verification
Water Glider Assembly Harmonic Study
Periodic hydrodynamic and mechanical loads were evaluated at two distinct operating frequencies on a Water Glider assembly. Stress and deformation maps confirmed that induced cyclic stresses stayed safely below material endurance thresholds across both operational regimes.
Electric Powertrain NVHEDU Electromagnetic & MBS Integration
In electric drive units (EDUs), electromagnetic stator forces (EMS) are mapped onto coupled MBS + FEA flexible-body models via MNFs. This captures high-frequency gear mesh and motor whine orders that would otherwise be masked in conventional ICE powertrains.
Random vibration (PSD) criteria and vibro-acoustic coupling inside vehicle cabins

Road-load profiles, turbulent airflow, and launch acoustics vary unpredictably over time and cannot be represented by a single sinusoid. Power Spectral Density (PSD) analysis converts acquired time-domain vibration signals into the frequency domain via the Fast Fourier Transform (FFT), quantifying how vibrational energy distributes across frequencies. Because random responses are non-deterministic, stress evaluations follow a Gaussian probability distribution across 1σ (68.3%), 2σ (95.5%), and 3σ (99.7%) confidence thresholds—benchmarked against S-N fatigue curves and standards such as MIL-STD-810G (military), ISO 16750-3 (automotive), and DO-160 (aerospace).
Inside a vehicle cabin, structural vibrations couple directly with the interior air volume. Tire-road forces and powertrain excitations drive body panels—quantified via Panel Participation Factors—which radiate structure-borne noise (Sound Pressure Level, or SPL) into the passenger compartment. To localize acoustic hotspots early, engineers compute Equivalent Radiated Power (ERP) in Watts or decibels across frequency sweeps and plot 3D ERP surface density fields across doors, floorpans, and engine covers.
Statistical Stress Thresholds
- 1σ Level (68.3%): Majority of steady random response oscillations fall within ±1 standard deviation of the mean.
- 2σ Level (95.5%): Common industrial engineering design threshold covering 95.5% of response peaks.
- 3σ Level (99.7%): Extreme peak-stress envelope used for high-reliability aerospace and automotive fatigue life checks.
Step 01
Dynamic Load Calculation
- 1D system simulation
- Multi-Body Simulation (MBS) with MNF
Step 02
Excitation Spectra
- Random Road Load Data (RLD / PSD)
- Cyclic rotating orders (FRF)
Step 03
Structural Vibration
- Eigenvalues & mode shapes
- Panel participation & ERP density
Step 04
Cabin Noise Prediction
- Structure-fluid acoustic cavity
- Boundary Element Method (BEM)
Interactive Explorer 03 · Full Audible Spectrum (0–20,000 Hz)
Selecting the Right Acoustic Solver Across Frequency Regimes
Active Regime: Low Frequency (0–1,000 Hz)0 – 1,000 Hz
Finite Element Analysis (FEA)200 – 2,500 Hz
Boundary Element Method (BEM)1,000 – 20,000 Hz
Statistical Energy Analysis (SEA)2,000 – 20,000 Hz
Acoustic Ray Tracing
Deterministic Structural-Fluid Coupling
Low-Frequency Body & Cavity Modes (FEA)
Deterministic FEA governs low-frequency vibro-acoustics up to roughly 500–1,000 Hz, where wavelengths are large relative to panel dimensions. Models exceeding 50 million DOFs couple body-in-white meshes, suspension mounts, trims, and the interior air cavity to resolve cabin boom and dominant structural transfer paths.
Primary Engineering Deliverables- Target Phenomena: Body/chassis resonances, suspension vibration transmission, cabin acoustic cavity modes.
- Key Metric: Sound Pressure Level (SPL) at occupant ear positions and Panel Participation Factors.
- Hybrid Pairing: Combined with SEA above 1 kHz to span the full 0–20 kHz audible spectrum.
FPSS Space Assembly PSD Validation
Multi-axis PSD simulation on an FPSS assembly evaluated deformation and stress distributions under broadband spaceflight excitation. Peak 3σ stresses remained within material endurance limits, validating mechanical stability without physical retesting.
Physical Correlation & UpdatingEMA, Laser Vibrometry, and MAC
Full-vehicle NVH digital twins are validated with up to 110 accelerometers and microphones, impact hammers, electrodynamic shakers, and laser Doppler vibrometers. Curve-fitting estimators (LSCE for lightly damped time data; LSCF for heavily damped frequency data) extract test modes correlated via the Modal Assurance Criterion (MAC).
04 · Methodological TaxonomyClassification of key NVH, vibration, and acoustic analysis methods

Selecting an appropriate simulation or experimental method depends on excitation determinism, target frequency range, and whether structural nonlinearity or acoustic radiation dominates the engineering question. Filter the classification below by discipline to inspect each method's governing domain, outputs, and role in an integrated NVH workflow.
Normal Modes (Eigenvalue) Analysis
Core
Solves |[K] − λ[M]| = 0 using Lanczos, AMSES, or AMLS to extract natural frequencies, mode shapes, and rigid-body connectivity checks.
- ExcitationFree undamped
- Key Outputfn (Hz), φi, residual vectors
- Primary UseResonance avoidance & FRF basis
Harmonic Frequency-Response (FRF)
Core
Evaluates steady-state structural response to periodic cyclic forces (unbalance, gear mesh, blade-pass) using modal superposition and damping.
- ExcitationSinusoidal / cyclic sweep
- Key OutputAmplitude, phase, cyclic stress
- Primary UseEndurance limit & EDU whine
Random Vibration (PSD) Analysis
Core
Propagates broadband stochastic energy (road-load data, acoustic launch loads) via FFT spectra to compute probabilistic Gaussian fatigue stress.
- ExcitationStochastic PSD (g²/Hz)
- Key Output1σ / 2σ / 3σ RMS stress
- Primary UseISO 16750-3, MIL-STD-810G, DO-160
Linear & Nonlinear Transient Analysis
Core
Integrates equations of motion directly in time for non-periodic impulses, pothole impacts, wind gusts, or large-displacement contact events.
- ExcitationTime-history f(t)
- Key OutputTransient u(t), shock stress
- Primary UseImpact, shock, nonlinear contact
Equivalent Radiated Power (ERP)
Acoustic
Estimates acoustic power radiated by vibrating structural surfaces directly from surface normal velocity without meshing the surrounding air volume.
- FrequencyLow–Mid (≤ 1.5 kHz)
- Key OutputERP (Watts, dB) & 3D density map
- Primary UseDoor, floorpan & cover ranking
Boundary Element Method (BEM) & Fluid Cavity
Acoustic
Models interior cabin air cavities via coupled structure-fluid FEA or exterior open-space radiation using surface-only BEM meshes.
- Frequency20 – 2,500 Hz
- Key OutputCabin SPL, transfer intensity
- Primary UseInterior boom & pass-by noise
SEA, Ray Tracing & CMS Reduction (MNF)
Acoustic
Tracks high-frequency energy balance (>1–2 kHz) through trims via SEA/ray tracing, and reduces flexible subassemblies into MNFs via Craig-Bampton CMS.
- Frequency1 – 20 kHz (SEA) / System MBS
- Key OutputEnergy flow & MNF super-elements
- Primary UseFull-spectrum NVH & drivelines
Experimental & Operational Modal (EMA / OMA)
Validation
Measures physical FRFs using impact hammers, shakers, accelerometers, and laser vibrometers (EMA), or ambient loads in service (OMA), fitted via LSCE/LSCF.
- MetricFRF poles, damping ζ, MAC
- Key OutputMeasured modes & model update
- Primary UseFEA correlation & large civil assets
Operating Deflection Shape (ODS)
Validation
Captures forced spatial deformation of machinery or vehicles running under real service loads to verify whether operating vibration matches a natural mode.
- ExcitationIn-service machine loads
- Key OutputForced deflection pattern
- Primary UseRoot-cause resonance proof

