Vibration & Dynamics — Engineering Reference

1. At a glance

Vibration is the oscillatory response of a mass-elastic system to a disturbance — internal (rotating unbalance, combustion pulses, gear-mesh impulses) or external (base motion, fluid pressure, acoustic, seismic). Dynamics is the broader Newtonian study of motion under unbalanced forces; vibration is its most common engineering manifestation because every real structure has finite stiffness and finite mass and therefore every real structure has natural frequencies.

A vibration problem is characterised along five orthogonal axes:

Axis	Cases
Degrees of freedom	Single-DOF (SDOF) → multi-DOF (MDOF, lumped) → continuous (distributed mass + stiffness)
Excitation	Free (initial conditions only) vs forced (sustained input)
Damping	Undamped (idealised) vs damped (viscous / Coulomb / hysteretic)
Time response	Transient (decaying) vs steady-state (sustained)
Determinism	Deterministic (harmonic, periodic, swept) vs random (PSD-defined)

Every rotating machine — every motor, pump, fan, compressor, turbine, gearbox — operates in the regime where the running speed must be placed safely away from system natural frequencies. Every flexible structure — every aircraft wing, every robot arm, every PCB on a vibrating mount, every bridge in wind — has modes that determine whether it shakes itself apart, limits servo bandwidth, transmits noise, or fatigues. Resonance (excitation frequency at or near a natural frequency, with light damping) routinely produces response amplitudes 10–100× the static deflection and is the leading mechanical failure mode of well-built machinery.

Where it sits in the stack: statics → mechanics of materials → dynamics → vibration → control & signal-processing. The same mass/stiffness matrices used to find static deflections become eigenproblems that yield natural frequencies and mode shapes; the same FEA mesh used for stress yields a modal analysis with one solver-deck change.

2. First principles

Newton-Euler equations of motion

For a particle of mass m at position x, Newton’s second law is F = mẍ. For a rigid body, Euler’s equations add ΣM_G = I_G · α (about the centre of mass) plus the translation equation. For a multi-DOF lumped system with generalized coordinates {x}, free-body decomposition produces a coupled linear system:

$[M]\{\ddot{x}\}+[C]\{\dot{x}\}+[K]\{x\}=\{F(t)\}$

• [M] — mass matrix (symmetric, positive-definite)
• [C] — damping matrix (symmetric, positive-semidefinite)
• [K] — stiffness matrix (symmetric, positive-semidefinite; positive-definite if no rigid-body modes)
• {F(t)} — generalized force vector

This is the canonical vibration equation. SDOF is the 1×1 special case; continuous systems are its infinite-dimensional limit (PDE form).

Lagrangian formulation

For complex geometries — multi-body linkages, robot arms, gear trains — direct free-body work becomes impractical. Lagrange’s equations use scalar kinetic energy T and potential energy V:

$\frac{d}{dt}\left(\frac{\partial T}{\partial {\dot{q}}_i}\right)-\frac{\partial T}{\partial q_i}+\frac{\partial V}{\partial q_i}=Q_i$

where q_i are generalized coordinates and Q_i are generalized non-conservative forces. Lagrange’s method automatically eliminates internal constraint forces and yields the same [M], [K] matrices that Newton-Euler produces — but with less bookkeeping. It is the standard derivation route for any system with > 3 DOFs.

Energy view

In a conservative oscillator, kinetic and potential energy interchange: at the equilibrium point T is maximum and V is zero; at the turning point V is maximum and T is zero. The natural frequency falls out of equating ½m ω² X² = ½k X², giving ω² = k/m. This Rayleigh quotient generalizes to MDOF and continuous systems and provides upper-bound estimates of ω_1 from assumed mode shapes — useful before any solver is run.

3. SDOF — the canonical model

The single-DOF mass-spring-damper with mass m, viscous damper c, and spring k under force F(t):

$m\ddot{x}+c\dot{x}+kx=F(t)$

Free undamped response

ω_n = √(k/m) — the natural angular frequency (rad/s). f_n = ω_n / (2π) — the natural frequency in Hz. Period T = 1/f_n. Static deflection under self-weight δ_st = mg/k gives the convenient shortcut:

$f_n=\frac{1}{2\pi }\sqrt{\frac{g}{{\delta }_{\text{st}}}}$

For δ_st in mm: f_n ≈ 15.76 / √δ_st Hz. A bracket that sags 1 mm under its own weight has f_n ≈ 15.8 Hz.

Damped response

Define the damping ratio:

$\zeta =\frac{c}{c_c}=\frac{c}{2\sqrt{km}}=\frac{c}{2m{\omega }_n}$

where c_c = 2√(km) is critical damping — the smallest c that prevents oscillation. The damped natural frequency:

${\omega }_d={\omega }_n\sqrt{1-{\zeta }^2}$

Three regimes:

• Underdamped (ζ < 1): oscillatory decay, x(t) = X·e^(−ζω_n t)·sin(ω_d t + φ). The common case for almost all structures (ζ typical 0.001 – 0.10).
• Critically damped (ζ = 1): fastest non-oscillatory return. Sometimes targeted in instrument design.
• Overdamped (ζ > 1): two real decaying exponentials, no oscillation. Rare in structures; common in fluid dashpots.

Logarithmic decrement

A clean way to measure ζ from a free-vibration decay trace:

$\delta =\ln \left(\frac{x_n}{x_{n+1}}\right)=\frac{2\pi \zeta }{\sqrt{1-{\zeta }^2}}\approx 2\pi \zeta (\zeta \ll 1)$

Hit a bracket with a modal hammer, record the response, measure two consecutive peaks: log decrement gives ζ in two lines of arithmetic. This is the foundation of impulse-response damping identification.

Harmonic excitation — the frequency-response function

For F(t) = F_0 sin(ωt), the steady-state response is x(t) = X sin(ωt − φ) with

$X=\frac{F_0/k}{\sqrt{(1-r^2{)}^2+(2\zeta r{)}^2}},\tan \phi =\frac{2\zeta r}{1-r^2}$

where r = ω/ω_n is the frequency ratio. The dimensionless amplitude X / (F_0/k) is the dynamic magnification factor. At resonance (r = 1) it equals 1/(2ζ) — the quality factor Q. For steel structures with ζ ≈ 0.01, Q ≈ 50; for rubber-isolated machinery with ζ ≈ 0.10, Q ≈ 5.

Transmissibility (base excitation / force isolation)

For a machine on isolators or a payload on a vibrating base:

$T_r=\frac{X}{Y}=\frac{F_T}{F_0}=\sqrt{\frac{1+(2\zeta r{)}^2}{(1-r^2{)}^2+(2\zeta r{)}^2}}$

Key features:

• r < √2 (i.e. ω_exc < √2 · ω_n): T_r > 1 — isolator amplifies the input.
• r = √2 exactly: T_r = 1 regardless of damping (crossover point).
• r > √2: T_r < 1 — isolation begins. Common design target: r ≥ 3, giving T_r ≈ 0.13 for ζ = 0.05.
• Higher damping reduces peak amplification at resonance but increases transmission at high r — the classic isolator trade-off.

Rotating unbalance

A rotor with eccentric mass m_e at radius e spinning at ω generates F(t) = m_e e ω² sin(ωt). The response X / (m_e e / m) = r² / √[(1−r²)² + (2ζr)²] — note the r² in the numerator: unbalance excitation grows with speed squared, so high-speed rotors are exquisitely sensitive to balance quality (ISO 21940 balance grades G0.4 – G4000).

4. Reference data — typical natural frequencies, damping ratios

Material / structural class	Damping ratio ζ
Steel (welded structure)	0.001 – 0.005
Steel (bolted, with friction)	0.005 – 0.02
Aluminum airframe	0.002 – 0.01
Reinforced concrete building (small amp)	0.02 – 0.05
Reinforced concrete (large amp / cracked)	0.05 – 0.10
Wood structure	0.05 – 0.10
Glass / ceramic	0.0001 – 0.001
Cast iron	0.005 – 0.02
Rubber elastomer (natural)	0.02 – 0.06
Rubber (butyl, high-damping)	0.10 – 0.25
Soil (typical for SSI)	0.05 – 0.30
Fluid-filled dashpot (viscous)	0.20 – 1.0
Magnetorheological damper (active)	0.10 – 0.80 (tunable)

Source: Inman Engineering Vibration 4th ed. Table 1.3.1; Genta Vibration Dynamics and Control §6.4; Rao Mechanical Vibrations 6th ed. Table 2.2; B&K Technical Review Damping of Materials (1982).

5p. MDOF / modal analysis

For an N-DOF system [M]{ẍ} + [K]{x} = {0}, assuming harmonic solutions {x} = {φ}e^(iωt) yields the generalized eigenvalue problem:

$([K]-{\omega }^2[M])\{\varphi \}=\{0\}$

Solutions are N pairs (ω_i², {φ_i}) — the natural frequencies and mode shapes. For typical structures the modes are real (undamped or proportionally damped); for general damping they become complex (real and imaginary parts, with a phase between DOFs).

Orthogonality

Modes are orthogonal with respect to both [M] and [K]:

{φ_i}^T [M] {φ_j} = 0 and {φ_i}^T [K] {φ_j} = 0 for i ≠ j

Normalizing such that {φ_i}^T [M] {φ_i} = 1 (mass-normalization) gives modal mass = 1, modal stiffness = ω_i². This is the standard FEA convention.

Modal coordinates and decoupling

Define the modal matrix [Φ] = [{φ_1} … {φ_N}] and modal coordinates {η} via {x} = [Φ]{η}. Substituting and pre-multiplying by [Φ]^T:

$[I]\{\ddot{\eta }\}+[\Lambda ]\{\eta \}=[\Phi {]}^T\{F(t)\}$

where [Λ] = diag(ω_i²). The MDOF problem decouples into N independent SDOF problems — the workhorse of structural dynamics. Each can be solved analytically or numerically and superposed.

Damping in modal coordinates

With Rayleigh proportional damping [C] = α[M] + β[K], the same transformation diagonalizes [C] and gives modal damping:

$2{\zeta }_i{\omega }_i=\alpha +\beta {\omega }_i^2⟹{\zeta }_i=\frac{\alpha }{2{\omega }_i}+\frac{\beta {\omega }_i}{2}$

Two measured ζ values at two frequencies pin α and β. Convenient but artificial — high-frequency modes get over-damped. For high-fidelity work, use modal damping ratios assigned per mode from test data.

Participation factor and effective modal mass

For base-excitation analysis the modal participation factor Γ_i = {φ_i}^T [M] {r} (where {r} is the rigid-body direction vector) and effective modal mass M_eff,i = Γ_i² describe how much of the system mass moves in mode i under a given input direction. ASCE 7 and many codes require enough modes to capture ≥ 90 % of total mass — anything less and the response is unconservative.

6p. Continuous systems — beams, bars, plates

When the wavelength of interest is comparable to the structure dimension, lumped models fail. Continuous systems are governed by PDEs:

• Axial bar: EA·∂²u/∂x² = ρA·∂²u/∂t² → c_bar = √(E/ρ)
• Torsional shaft: GJ·∂²θ/∂x² = ρI_p·∂²θ/∂t² → c_tor = √(G/ρ)
• String: T·∂²w/∂x² = ρA·∂²w/∂t² → c_string = √(T/(ρA))
• Euler-Bernoulli beam: EI·∂⁴w/∂x⁴ + ρA·∂²w/∂t² = 0
• Kirchhoff plate: D·∇⁴w + ρh·∂²w/∂t² = 0, where D = Eh³/[12(1−ν²)]

For a uniform Euler-Bernoulli beam:

${\omega }_n=({\beta }_{n}L{)}^2\sqrt{\frac{EI}{\rho A{L}^4}}$

The (β_n L) coefficients depend on boundary conditions:

BC	(β_1 L)	(β_2 L)	(β_3 L)	(β_4 L)	(β_5 L)
Cantilever (fixed-free)	1.8751	4.6941	7.8548	10.9955	14.1372
Simply supported (pin-pin)	π = 3.1416	2π = 6.2832	3π = 9.4248	4π = 12.5664	5π = 15.7080
Clamped-clamped	4.7300	7.8532	10.9956	14.1372	17.2788
Free-free	4.7300	7.8532	10.9956	14.1372	17.2788
Clamped-pinned	3.9266	7.0686	10.2102	13.3518	16.4934

(Free-free and clamped-clamped share eigenvalues but have different mode shapes; free-free additionally has two rigid-body modes at ω = 0.)

For SS-SS the result simplifies to f_n = (n²π/2) · √(EI/(ρAL⁴)). Timoshenko beam theory adds shear deformation and rotary inertia and is required when L/h < 10 or for higher modes; it lowers ω_n versus Euler-Bernoulli by typically 5–30 %.

7p. Damping models

• Viscous (linear): F_d = c·ẋ. Mathematically convenient — yields linear ODEs and the [C] matrix above. Approximates fluid dashpots and some elastomers at low strain. The model behind almost all linear vibration analysis.
• Coulomb (dry friction): F_d = μN·sgn(ẋ). Discontinuous; produces linear envelope decay (not exponential). Dominant in bolted joints, brake systems, leaf springs, foundation-soil contact. Energy dissipated per cycle ΔE = 4μN·X is independent of frequency, unlike viscous (ΔE = πcωX²).
• Hysteretic / structural: F_d = (η·k/ω) · ẋ, energy per cycle ΔE = πη·k·X² — independent of frequency. Matches measured behavior of most metals far better than viscous. Loss factor η ≈ 2ζ for light damping. Typical η values: steel 0.0001–0.01, aluminum 0.0001–0.005, cast iron 0.001–0.02, rubber 0.05–0.5, viscoelastic damping tape (3M ISD-112) 0.5–1.5 at design temperature.
• Rayleigh proportional: [C] = α[M] + β[K] — analytically convenient but not physical.
• Modal: assign ζ_i per mode directly from test. Highest fidelity for linear analysis.

Logarithmic decrement (revisited, MDOF context)

For a single dominant mode in a free-decay record:

$\zeta \approx \frac{1}{2\pi N}\ln \left(\frac{x_0}{x_N}\right)$

over N cycles. Use ≥ 5 cycles for a stable estimate; fewer and noise dominates.

8p. Random vibration

Most real excitation — road inputs to a vehicle, jet noise on an airframe, turbulent pressure on a heat exchanger, seismic ground motion — is broadband and stochastic. Described by the Power Spectral Density S_xx(f) [units²/Hz]: the mean-square contribution per Hz of bandwidth.

${\sigma }^2={\int }_0^{\infty }{S}_{xx}(f)df$

so the RMS is σ = √(area under PSD).

Miles equation (the workhorse)

For an SDOF system at f_n with damping ζ subjected to flat input PSD S_0 [g²/Hz] across a broad band that fully contains f_n:

${\sigma }_{\text{response}}=\sqrt{\frac{\pi }{2}{f}_{n}Q{S}_0}=\sqrt{\frac{\pi f_n{S}_0}{4\zeta }}$

with Q = 1/(2ζ). Used everywhere in spacecraft, electronics, and aerospace random-vibration screening (MIL-STD-810H method 514.8, NASA-STD-7001B, GEVS).

Level crossing (Rice formula)

For a narrowband Gaussian response with RMS σ and centre frequency f_n, the expected rate of upward crossings of level a per second is:

$N^{+}(a)=f_n\cdot \exp \left(-\frac{a^2}{2{\sigma }^2}\right)$

This feeds rainflow counting and Miner’s rule (linear damage accumulation, D = Σ n_i/N_i ≤ 1) for fatigue under random load — see mechanics-of-materials for S-N and Goodman diagrams.

9p. Practical math — three worked examples

Example A — Automotive engine mount design

A 1500 kg engine block must be isolated from a 30 Hz excitation (typical 6-cylinder firing harmonic at 600 rpm). Design the mount stiffness.

Design rule for isolation: f_exc / f_n > √2; practical target f_exc / f_n ≥ 3 → 4. Pick f_n = 6 Hz (so ratio = 5).

ω_n = 2π · 6 = 37.70 rad/s

Required stiffness (total, all 4 mounts combined):

k = m · ω_n² = 1500 kg · (37.70 rad/s)² = 1500 · 1421.2 = 2.132 × 10⁶ N/m = 2132 N/mm

Per mount (assume 4 mounts equal): k_mount = 533 N/mm. A Lord Corporation LM-1116 / LM-1119 flexible engine mount (rubber-in-shear, durometer 50 IRH) is rated ≈ 450–600 N/mm radial — direct fit.

Static deflection check: δ_st = mg/k = (1500·9.81) / (2.132·10⁶) = 6.90 × 10⁻³ m = 6.9 mm. Cross-check with the shortcut: f_n = 15.76/√6.9 = 6.0 Hz ✓.

Transmissibility at 30 Hz, assuming ζ = 0.08 (typical rubber):

r = 30/6 = 5.0 T_r = √[(1 + (2·0.08·5)²) / ((1 − 25)² + (2·0.08·5)²)] = √[(1 + 0.64) / (576 + 0.64)] = √(1.64 / 576.64) = √0.002844 = 0.0533

So 94.7 % of the 30 Hz vibration is blocked. In US units: m ≈ 3307 lb, k_total ≈ 12180 lbf/in, δ_st ≈ 0.272 in.

Check resonance crossing. On start-up, the engine sweeps through 6 Hz at idle approach. With Q = 1/(2·0.08) = 6.25, the engine will transiently amplify 6× while sweeping past — usually acceptable for short transit. If not, add a snubber or auxiliary damping (Lord active mount with controlled fluid damping, e.g. LM-2400 series).

Example B — Cantilever beam first natural frequency

AISI 1018 steel cantilever: L = 200 mm, cross-section 20 mm × 3 mm (rectangular). E = 200 GPa, ρ = 7870 kg/m³.

Section properties (bending about strong axis, h = 20 mm wide × b = 3 mm thick → for bending about the 3 mm height, I = (20·10⁻³)·(3·10⁻³)³/12):

A = 20·3 = 60 mm² = 6.0 × 10⁻⁵ m² I = b·h³/12 = (0.020)·(0.003)³ / 12 = (0.020 · 2.7×10⁻⁸) / 12 = 4.5 × 10⁻¹¹ m⁴

(Beam is thin-direction bending; bandwidth-limiting mode.)

Mass per length: ρA = 7870 · 6.0×10⁻⁵ = 0.472 kg/m

EI: 200×10⁹ · 4.5×10⁻¹¹ = 9.00 N·m²

First mode (β_1 L = 1.8751):

ω_1 = (1.8751)² · √(EI / (ρA·L⁴)) = 3.5160 · √(9.00 / (0.472 · 0.2⁴)) = 3.5160 · √(9.00 / (0.472 · 1.6×10⁻³)) = 3.5160 · √(9.00 / 7.552×10⁻⁴) = 3.5160 · √(11919) = 3.5160 · 109.18 = 383.9 rad/s

f_1 = ω_1 / (2π) = 61.1 Hz

Tip deflection under 1 N static load: δ = PL³/(3EI) = 1 · (0.2)³ / (3·9.00) = 0.008 / 27 = 2.96 × 10⁻⁴ m = 0.296 mm

Cross-check via static-deflection shortcut: the equivalent SDOF mass for a cantilever (Rayleigh, distributed mass) is 0.2427 m_total = 0.2427·(0.472·0.2) = 0.0229 kg. Equivalent stiffness k_eq = 3EI/L³ = 27/0.008 = 3375 N/m. ω_n = √(3375/0.0229) = √147380 = 384 rad/s → 61.1 Hz ✓ — exact agreement.

Servo bandwidth implication: for a servo controlling a payload on this beam, classical rule of thumb is closed-loop bandwidth ≤ ω_1 / (5–10). So a 61 Hz first mode caps useful servo bandwidth at 6–12 Hz unless active vibration suppression is included — see classical-control (planned).

Example C — Random vibration RMS stress on a bracket

A mounting bracket carries an electronic enclosure (effective lumped mass m = 0.50 kg at the bracket tip). Bracket modelled as SDOF with f_n = 100 Hz, ζ = 0.05. The base is excited with a flat PSD S_0 = 1.0 g²/Hz from 20 Hz to 2000 Hz (typical military qualification spectrum, MIL-STD-810H method 514.8 cat 24).

Miles equation, response RMS acceleration:

σ_a = √[(π/2) · f_n · Q · S_0] = √[(π/2) · 100 · 10 · 1.0] = √(1570.8) = 39.6 g_rms

(Q = 1/(2·0.05) = 10.)

Convert to RMS displacement (SDOF tip displacement under steady-state harmonic at f_n equals acceleration/(2πf_n)², so RMS:

σ_x = σ_a · g / (2π f_n)² = 39.6 · 9.81 / (628.3)² = 388.5 / 3.948×10⁵ = 9.84 × 10⁻⁴ m = 0.984 mm RMS

Propagate to RMS stress. If the bracket’s static-deflection-to-stress relation is σ_stress / δ = k_σ (units Pa/m, determined by geometry — e.g., from a static FE run, suppose k_σ = 5.0 × 10⁸ Pa/m), then:

σ_stress,RMS = k_σ · σ_x = 5.0×10⁸ · 9.84×10⁻⁴ = 4.92 × 10⁵ Pa = 0.49 MPa RMS

3-sigma peak (Gaussian assumption): 3 · 0.49 = 1.47 MPa

Compare against the AISI 1018 fatigue endurance limit (~ 220 MPa for unnotched, ~ 80 MPa for typical machined-and-notched bracket per Shigley Table 6-2 with K_f = 2.5) — bracket is far below endurance limit, infinite life. If the bracket had f_n = 30 Hz instead, σ_a would scale as √(f_n) → σ_a ≈ 21.7 g — still safe, but the displacement response (which scales as σ_a/(2πf_n)² ) blows up: 21.7 · 9.81 / (188.5)² ≈ 6.0 mm RMS — large clearance issue.

Cross-link to mechanics-of-materials for the Goodman/Soderberg fatigue criterion and S-N curve methodology.

10p. Edge cases & assumptions

• Modal truncation error. Mode-superposition keeps only the first N modes (typically 90 % effective mass). For shock loads with broad spectra and for stress in members far from the load point, neglected high-frequency modes can carry significant deflection — required correction is the residual mode method or mode acceleration method. Industry guidance: keep modes up to ≥ 2 × highest excitation frequency, plus a static residual.
• Mass loading from instruments. A typical PCB Piezotronics 352C03 accelerometer weighs 5.8 g. Mounted on a 50 g lightweight composite panel, it lowers the measured f_n by √(50/55.8) − 1 ≈ −5 %. For accurate modal test on light structures use miniature accels (PCB 352A24, 0.8 g) or non-contact LDV (Polytec OFV-505).
• Nonlinear stiffness (Duffing). Springs with hardening or softening nonlinearity (rubber, large-deflection beams, magnetic levitation) lead to amplitude-dependent natural frequency and jump phenomenon — the frequency-response curve folds, allowing two stable response amplitudes for one excitation frequency. Identified by hysteresis in swept-sine response.
• Closely-spaced modes. Two modes within ~1 % of each other in frequency (common in symmetric structures, e.g. a square plate’s first two bending modes, or aircraft wing flap-bending pairs) cause mode coupling and mode swapping — operating deflection shapes look like superpositions, and small perturbations swap which mode is “first.” Modal complexity factors > 0.1 are a warning sign.
• Coupled axial-torsional-bending in rotors. A rotor in bearings exhibits forward and backward whirl, gyroscopic stiffening, and bearing-coupled modes that depend on speed. The Campbell diagram (natural-frequency vs speed, overlaid with engine orders 1×, 2×, 3×) identifies operating-speed resonance crossings — every turbomachine design includes one.
• Self-excited vibration. Flutter (aerodynamic-elastic coupling), chatter (machine-tool regenerative cutting), stick-slip (friction-driven), shaft whip — these have no external harmonic input but extract energy from a steady source (flow, cutting force, sliding contact). The instability is in the system itself; linear damping models predict ζ < 0 (negative apparent damping). Mitigation: detune, add damping, or break the regenerative path.
• Parametric resonance. Periodically time-varying stiffness k(t) = k_0(1 + ε cos Ωt) — e.g. a swing pumped at 2× the natural frequency, or a shaft with anisotropic stiffness — produces unbounded growth even with damping when Ω ≈ 2ω_n/n. Mathieu equation. Common in cable-stayed bridges and parametrically excited MEMS.
• Floor-borne vs airborne paths. Vibration in buildings transmits through both structural (floor → wall → equipment) and acoustic (air-pressure → sensitive surface) paths. ISO 10137 separates the two. A vibration-sensitive instrument (electron microscope, optical interferometer) needs both pneumatic isolation (floor-borne) and acoustic enclosure (airborne).
• Temperature dependence. Damping in elastomers and viscoelastics is sharply temperature-dependent. A rubber mount that gives ζ = 0.08 at 25 °C may give ζ = 0.03 at 60 °C and ζ = 0.20 at 0 °C. Always specify operating-temperature range when sizing isolators. Viscoelastic damping tapes (3M ISD-112, ISD-110) have peak loss-factor temperatures published.

11p. Tools & software

FEA (modal, frequency-response, transient, random)

• ANSYS Mechanical — MODAL, HARMIC, TRANS, SPECTRUM, PSD solution types; the broadest commercial coverage.
• Abaqus — *FREQUENCY, *STEADY STATE DYNAMICS, *RANDOM RESPONSE; modal-based and direct-integration variants.
• MSC Nastran / Simcenter Nastran — SOL 103 (modal), 108 (direct freq response), 111 (modal freq response), 112 (modal transient), 200 (optimization with frequency response). The aerospace incumbent.
• Altair OptiStruct — strong for topology optimization with frequency constraints.

Multibody dynamics (rigid + flexible)

• MSC Adams — automotive ride/handling, suspension, drivetrain.
• FunctionBay RecurDyn — strong on contact and large-deformation flexible bodies.
• Altair MotionSolve — closely integrated with HyperWorks.
• MapleSim / Modelica / OpenModelica — equation-based modeling.

Lumped-parameter / system simulation

• MathWorks Simulink Simscape (Multibody, Driveline) — common in control system co-simulation.
• ESI SimulationX, Gamma Technologies GT-SUITE — driveline, hydraulic, thermal-fluid systems.
• AVL Excite — engine and powertrain NVH.

Measurement instrumentation

• Accelerometers (IEPE / ICP): PCB Piezotronics 352C03 (general purpose, 100 mV/g, 5.8 g, 0.5 Hz – 10 kHz), 352A24 (miniature, 100 mV/g, 0.8 g), 356A16 (triaxial, 100 mV/g, 7.4 g); Brüel & Kjær 4534-B triaxial (10 mV/g, 9 g); Dytran 3263A2 (high temperature, 500 °C); Endevco 27 series (charge-mode for high temperature / high shock).
• Modal impact hammers: PCB 086C03 (0.32 kg, sensitivity 2.25 mV/N, force range 2200 N — general-purpose); 086C04 (mini, for small structures); 086D50 (large, sledge, for civil structures up to 22 kN).
• Laser Doppler Vibrometers (non-contact): Polytec OFV-505 / OFV-5000 (single-point), PSV-500 (scanning, full-field mode shape mapping), MSA-600 (micro-system analyzer for MEMS).
• IEPE signal conditioners / DAQ: PCB 482C series, Kistler 5134, Dewesoft SIRIUS-HD-STG (24-bit, 200 kS/s/ch).
• Shaker tables / electrodynamic shakers: Brüel & Kjær LDS V8 (89 N peak, lab use), LDS V455 (489 N, lab/test), MB Dynamics Modal 50A (modal exciter, 222 N); Unholtz-Dickie T1000 for environmental qualification (up to 26.7 kN). Hydraulic shakers (MTS) for seismic / large-displacement.
• Analyzers / DAQ + software: Siemens Simcenter Testlab (LMS Test.Lab), Brüel & Kjær PULSE / Connect, HBM SoMat eDAQ (field), Dewesoft DewesoftX, NI DIAdem / Sound and Vibration Toolkit.

Open-source / Python ecosystem

• scipy.signal — windowing, PSD (Welch), filtering, transfer function fitting.
• OpenSeismoMatlab — earthquake response spectra and structural dynamics in MATLAB/Octave.
• PyFEM, FEniCS, CalculiX, Code_Aster — open FEA with modal solvers.
• PyOMA / pyEMA — operational and experimental modal analysis libraries.
• OpenModal — Python-based modal testing toolkit (LMS-style FRF measurement).

ISO 10816 / 20816 — machine vibration severity (broadband velocity RMS, mm/s)

Compact alarm/trip table for a Class III machine (large machines, rigid foundation, e.g. 75 kW – 300 kW motor):

Zone	RMS velocity (mm/s)	Interpretation
A	≤ 1.8	New-machine quality
B	1.8 – 4.5	Long-term continuous operation acceptable
C	4.5 – 7.1	Limited duration; restricted operation
D	> 7.1	Severe — damage imminent; trip

Class definitions and zone boundaries vary with machine size (Class I – IV) and foundation type. The ISO 20816-1:2016 series supersedes ISO 10816 and includes specific parts for steam turbines, hydraulic machines, wind turbines, etc.

12. Cross-references

• beam-theory — Euler-Bernoulli and Timoshenko PDEs; mode-shape derivation for the table in §6p.
• mechanics-of-materials — stress/strain prerequisite; fatigue (Goodman, S-N) for random-vibration life prediction in Example C.
• statics-fundamentals — equilibrium foundation that becomes [K] in modal analysis.
• electric-motors — rotor dynamics, bearing-induced vibration, Campbell-diagram analysis of motor-driven trains.
• bearings — bearing vibration signature (BPFO, BPFI, BSF, FTF frequencies), envelope-detection diagnostics, ISO 15243 failure modes.
• gears-power-transmission — gear-mesh frequency (GMF = n_teeth · rpm/60) and sideband spectra are dominant rotating-machinery excitation sources.
• fasteners-bolts — bolted-joint damping and frictional contributions to structural ζ.
• classical-control (planned) — closed-loop bandwidth vs structural ω_n trade-off; notch filters for active vibration suppression.
• structural-analysis (planned) — multi-storey building dynamics; response-spectrum and time-history seismic analysis.
• kinematics-dh (planned) — joint compliance and link flexibility as MDOF systems on the robot dynamics tree.
• pid-control (planned) — vibration suppression and input-shaping for high-acceleration manipulators.
• construction-bim — BDF / INP / OP2 file formats for modal data exchange.

13. Citations

1. Rao, S.S. Mechanical Vibrations, 6th ed., Pearson, 2017. ISBN 978-0-13-436131-1. The standard graduate text — comprehensive, mathematically thorough.
2. Inman, D.J. Engineering Vibration, 4th ed., Pearson, 2013. ISBN 978-0-13-287169-3. The clearest undergraduate text; strongest on physical interpretation.
3. Meirovitch, L. Fundamentals of Vibrations, McGraw-Hill, 2001 (reprinted Waveland 2010). ISBN 978-1-57766-691-2. Mathematical rigor; the reference for advanced modal analysis.
4. Meirovitch, L. Methods of Analytical Dynamics, McGraw-Hill, 1970 (reprinted Dover 2003). ISBN 978-0-486-43239-3. The deep theoretical companion for Lagrangian / Hamiltonian dynamics.
5. Genta, G. Vibration Dynamics and Control, Springer, 2009. ISBN 978-0-387-79579-9. Strong on rotor dynamics, nonlinear effects, and active control.
6. Ewins, D.J. Modal Testing: Theory, Practice and Application, 2nd ed., Research Studies Press, 2000. ISBN 978-0-86380-218-8. The bible of experimental modal analysis.
7. Newland, D.E. An Introduction to Random Vibrations, Spectral and Wavelet Analysis, 3rd ed., Dover, 2005. ISBN 978-0-486-44274-3. The random-vibration reference.
8. Wirsching, P.H., Paez, T.L., Ortiz, K. Random Vibrations: Theory and Practice, Dover, 2006. ISBN 978-0-486-45015-1. Engineering-practice companion to Newland.
9. Den Hartog, J.P. Mechanical Vibrations, 4th ed., McGraw-Hill, 1956 (reprinted Dover 1985). ISBN 978-0-486-64785-2. The historical classic; still the clearest treatment of tuned mass dampers.
10. Harris, C.M., Piersol, A.G., editors. Harris’ Shock and Vibration Handbook, 6th ed., McGraw-Hill, 2010. ISBN 978-0-07-150819-6. The encyclopedic practitioner reference.
11. Bendat, J.S., Piersol, A.G. Random Data: Analysis and Measurement Procedures, 4th ed., Wiley, 2010. ISBN 978-0-470-24877-5. Spectral analysis fundamentals for measurement.
12. ISO 7626 series — Mechanical vibration and shock — Experimental determination of mechanical mobility (parts 1–5, 1989–2017). The reference for FRF measurement methodology.
13. ISO 10816 / ISO 20816 series — Mechanical vibration — Evaluation of machine vibration by measurements on non-rotating parts. ISO 20816-1:2016 (general), 20816-2 (steam turbines), 20816-3 (industrial machines 15 kW – 1000 kW), 20816-4 (gas turbines), 20816-5 (hydraulic machines), 20816-21 (wind turbines).
14. ISO 2041:2018 — Mechanical vibration, shock and condition monitoring — Vocabulary. Authoritative definitions of technical terms.
15. ISO 21940 series — Mechanical vibration — Rotor balancing. Replaced the older ISO 1940. Defines balance quality grades G0.4 – G4000.
16. ANSI S2.61:1989 (R2006) — Guide to the Mechanical Mounting of Accelerometers. ASA / Acoustical Society of America.
17. MIL-STD-810H Method 514.8 — Vibration. U.S. DoD environmental engineering test method; the de facto random-vibration qualification standard for defense and aerospace electronics.
18. NASA-STD-7001B — Payload Vibroacoustic Test Criteria. The spacecraft-test standard.
19. GEVS, GSFC-STD-7000B — General Environmental Verification Standard. NASA Goddard launch-environment test spec.
20. PCB Piezotronics Technical Information — Introduction to Piezoelectric Accelerometers and Modal Testing Handbook. https://www.pcb.com/resources. Useful vendor-authored tutorials.
21. Brüel & Kjær Technical Reviews — Damping of Materials and Members in Structures (Lazan reprint, 1982); Modal Analysis primer series (Avitabile reprints). https://www.bksv.com.
22. Polytec Application Notes — Laser Doppler Vibrometry. https://www.polytec.com/us/vibrometry. Especially AN-VIB-G-1 (Principles).
23. Avitabile, P. Modal Testing: A Practitioner’s Guide, Wiley, 2017. ISBN 978-1-119-22289-3. The modern practical companion to Ewins.
24. Adhikari, S. Structural Dynamic Analysis with Generalized Damping Models, Wiley, 2014. ISBN 978-1-84821-521-2. Authoritative on non-proportional damping.