Vibration Analysis

Vibration analysis is the systematic study of the oscillatory motion of structures and mechanical components. In the context of structural acoustics, it provides the bridge between mechanical vibrations and the acoustic fields they generate. Mastery of the terminology is essential because each term carries a precise definition that influences how data are interpreted, how models are built, and how mitigation strategies are devised.

Natural frequency – the frequency at which a structure tends to vibrate when it is disturbed and then allowed to vibrate freely. It is a property of the mass and stiffness distribution of the system. For a simple cantilever beam the first natural frequency might be around 5 Hz, while a steel plate may have several natural frequencies ranging from 50 Hz to several kilohertz. Knowing natural frequencies is critical because if an external excitation contains energy near one of these values, resonant amplification will occur.

Mode shape – the spatial pattern of deformation that a structure exhibits at a particular natural frequency. Mode shapes are orthogonal, meaning that the deformation pattern of one mode does not interfere with that of another. In practice, engineers often visualise mode shapes using coloured contour plots or animated deformation sketches. A common challenge is that real structures rarely vibrate in a single pure mode; instead, multiple modes can be excited simultaneously, leading to complex superpositions.

Damping ratio – a dimensionless measure of how quickly vibrational energy dissipates. It is defined as the ratio of actual damping to critical damping. Typical values for steel structures are on the order of 0.01 To 0.03, Whereas polymeric components may exhibit ratios of 0.05 Or higher. Low damping ratios mean that resonant peaks are sharp and high; high damping smooths the frequency response but may also reduce the effectiveness of certain isolation strategies.

Resonant frequency – often used interchangeably with natural frequency, but in practice it refers to the frequency at which the system’s response is maximised under a particular type of excitation. For example, a harmonic force applied at a frequency that matches a natural frequency will produce a resonant response, whereas a broadband random excitation may cause resonance at a slightly shifted frequency due to damping effects.

Amplitude – the magnitude of the vibration, which can be expressed in terms of displacement (meters or micrometres), velocity (meters per second), or acceleration (g‑units). The choice of unit depends on the sensor used and the type of analysis. In many acoustic applications, acceleration is preferred because it correlates directly with sound pressure levels generated by vibrating panels.

Phase – the relative timing of a sinusoidal vibration compared with a reference. Phase information is crucial when combining multiple vibration sources, as constructive or destructive interference depends on the phase relationship. In modal testing, the phase angle between force and response at each frequency is used to compute the complex frequency response function.

Frequency response function (FRF) – the ratio of output to input as a function of frequency, typically expressed in complex form (magnitude and phase). The FRF encapsulates the dynamic behaviour of a structure and is the cornerstone of experimental modal analysis. For a given force input, the FRF predicts the resulting acceleration at a measurement point across the frequency range of interest.

Transfer function – a synonym for FRF, often used in control engineering contexts. The transfer function can be represented analytically as a rational function of the Laplace variable s, where the poles correspond to natural frequencies and the residues reflect modal participation.

Modal analysis – the process of identifying the natural frequencies, mode shapes, and damping ratios of a structure. It can be performed analytically (e.G., Using finite‑element models) or experimentally (e.G., Using impact‑hammer or shaker testing). The result is a set of modal parameters that enable prediction of the response to arbitrary excitations through superposition.

Operational deflection shape (ODS) – the deformation pattern of a structure when it is subjected to its actual operating loads, not just a single mode. ODSs are obtained by measuring the response at many points while the structure is running under realistic conditions. They help engineers understand how multiple modes interact in service.

Spectral density – a function that describes how the power of a random vibration signal is distributed over frequency. The term “power spectral density” (PSD) is commonly used for acceleration signals and has units of (g²/Hz). Engineers use PSDs to assess fatigue damage caused by broadband noise, such as that generated by wind turbines or vehicle engines.

Coherence – a statistical measure ranging from 0 to 1 that indicates how well two signals are linearly related at each frequency. In vibration testing, a high coherence (above 0.9) Between force and response signals confirms that the measured FRF is reliable. Low coherence can arise from sensor noise, non‑linear behaviour, or insufficient excitation.

Fast Fourier Transform (FFT) – an algorithm that converts a time‑domain signal into its frequency‑domain representation. FFTs are the workhorse of vibration analysis because they enable rapid calculation of spectra from long time records. However, the FFT assumes that the signal is stationary and periodic within the analysis window; violations of these assumptions lead to spectral leakage.

Windowing – the application of a weighting function (such as Hann, Hamming, or Blackman) to a time‑domain record before performing an FFT. Windowing reduces leakage but also widens the main lobe, affecting frequency resolution. Selecting an appropriate window is a trade‑off that depends on the bandwidth of interest and the desired resolution.

Aliasing – an artefact that occurs when a signal is sampled at a rate insufficient to capture its highest frequency content. According to the Nyquist theorem, the sampling frequency must be at least twice the highest frequency present. If this condition is not met, high‑frequency components fold back into lower frequencies, corrupting the spectrum.

Nyquist frequency – half of the sampling rate. It defines the highest frequency that can be unambiguously represented in a sampled signal. For a 10 kHz sampling rate, the Nyquist frequency is 5 kHz. Engineers often design anti‑alias filters to attenuate frequencies above the Nyquist limit before digitisation.

Time domain – the representation of a vibration signal as a function of time. Time‑domain analysis is useful for transient events, such as impact loads or shock pulses. It allows direct observation of peak values, rise times, and decay rates, which are essential for assessing structural integrity after a sudden event.

Frequency domain – the representation of a signal as a function of frequency. Frequency‑domain analysis excels at identifying resonant peaks, quantifying broadband noise, and designing filters. Many vibration‑related standards, such as ISO 10816, prescribe limits in terms of RMS velocity measured over a specific frequency band.

Transient analysis – the study of non‑steady‑state responses that evolve over time, typically following an impulsive or step input. Transient analysis is performed using time‑domain integration methods (e.G., Newmark‑β) or by applying the inverse Laplace transform to the transfer function. In structural acoustics, transient analysis predicts how a panel reacts to a sudden impact and the resulting sound emission.

Steady‑state analysis – the study of responses that have settled into a constant periodic pattern, usually under harmonic excitation. In this regime, the system’s behaviour can be described entirely by its FRF. Steady‑state analysis is the basis for vibration‑based sound power calculations, where the acoustic power radiated by a panel is proportional to the squared velocity amplitude.

Harmonic excitation – a sinusoidal force applied at a single frequency. Harmonic testing is the simplest way to probe a structure’s dynamic behaviour, as each frequency component can be examined independently. However, many real‑world sources (e.G., Engines) generate a spectrum of harmonics, requiring multi‑frequency or broadband testing.

Random vibration – excitation characterised by a stochastic distribution of frequencies and amplitudes, typically described by a PSD. Random vibration testing is used to simulate environments such as aircraft turbulence, road roughness, or seismic ground motion. The response is also stochastic, and statistical methods are employed to predict fatigue life.

Broadband noise – vibration energy that spans a wide frequency range without distinct peaks. Broadband noise is common in HVAC systems, fans, and compressors. In acoustic terms, broadband noise leads to a smooth sound pressure level spectrum, often described using A‑weighted decibels.

Vibration isolation – the practice of reducing the transmission of vibratory energy from a source to a receiver. Isolation is achieved by inserting compliant elements (springs, elastomers, or air mounts) between the two. The isolation effectiveness depends on the mass of the isolated equipment, the natural frequency of the isolation system, and the frequency content of the source.

Tuned mass damper (TMD) – a passive device consisting of a mass, spring, and damper tuned to a specific target frequency. When the primary structure vibrates near that frequency, the TMD absorbs energy and reduces the amplitude of the primary system. TMDs are widely used in skyscrapers, bridges, and precision instruments.

Viscoelastic damping – energy dissipation arising from the time‑dependent deformation of polymeric or rubber‑like materials. Viscoelastic layers are often incorporated into sandwich panels to increase damping without adding excessive mass. The damping loss factor of viscoelastic materials is frequency‑dependent, which must be accounted for in accurate acoustic predictions.

Stiffness – the resistance of a component to deformation under load. In linear systems, stiffness k is defined as force divided by displacement (N/m). Stiffness directly influences natural frequencies; increasing stiffness raises the frequencies, while reducing stiffness lowers them. In acoustic panels, high stiffness is desirable for radiating sound efficiently, whereas low stiffness may be preferred for vibration isolation.

Mass – the quantity of matter that resists acceleration. In vibration analysis, mass contributes to inertia and determines the kinetic energy stored in a vibrating system. Adding mass to a structure generally lowers its natural frequencies, which can be beneficial for moving resonances away from harmful excitation bands.

Compliance – the inverse of stiffness, expressed as displacement per unit force (m/N). Compliance is often used when describing the flexibility of mounting elements. High compliance mounts provide better isolation at low frequencies but may permit excessive motion at higher frequencies if not properly damped.

Modal mass – the effective mass associated with a particular mode shape. It is obtained by integrating the product of the mode shape and the mass distribution over the structure. Modal mass is a key parameter in the superposition method, where the contribution of each mode to the overall response is weighted by its modal mass.

Modal stiffness – the effective stiffness of a mode, defined as the ratio of modal force to modal displacement. Together with modal mass, it determines the natural frequency of the mode (ω = √(k / m)). Modal stiffness can be extracted from finite‑element eigenvalue analyses or from experimental FRFs.

Participation factor – a scalar that quantifies how much a particular mode contributes to the global response under a given load distribution. It is calculated by projecting the load vector onto the mode shape and normalising by the modal mass. Modes with high participation factors dominate the response and are the primary targets for mitigation.

Eigenvalue – in vibration analysis, the square of the natural angular frequency (ω²) that satisfies the homogeneous equation of motion. Solving the eigenvalue problem yields both eigenvalues (frequencies) and eigenvectors (mode shapes). Numerical methods such as the Lanczos algorithm are used for large finite‑element models.

Eigenvector – the mode shape associated with a particular eigenvalue. Eigenvectors are normalised for convenience, often so that the maximum displacement is unity or the modal mass is one. The eigenvector defines the pattern of deformation that the structure will assume when vibrating at its corresponding natural frequency.

Finite‑element model (FEM) – a discretised representation of a structure using elements (beams, plates, shells, solids) that approximate the governing differential equations. FEM enables prediction of natural frequencies, mode shapes, and stress distributions. Accuracy depends on mesh quality, material property definition, and appropriate boundary conditions.

Lumped‑parameter model – a simplified representation where mass, stiffness, and damping are concentrated at discrete points. Lumped models are useful for quick hand calculations, such as estimating the resonance of a simple mass‑spring‑damper system. While less accurate than FEM, they provide insight and are valuable for initial design iterations.

Continuous system – a model that treats mass, stiffness, and damping as distributed along a spatial dimension, such as a beam or plate. Analytical solutions for continuous systems often involve solving differential equations using methods like separation of variables or the Rayleigh‑Ritz technique. Continuous models are essential for understanding wave propagation phenomena in structural acoustics.

Beam theory – the collection of analytical models (Euler‑Bernoulli, Timoshenko) that describe the bending behaviour of slender members. Beam theory predicts natural frequencies, mode shapes, and deflection under loads. In vibration analysis, the Timoshenko beam model is preferred when shear deformation and rotary inertia are non‑negligible, especially at higher frequencies.

Plate theory – models the flexural behaviour of flat, thin elements. Classical plate theory (Kirchhoff) assumes that normals to the mid‑surface remain perpendicular after deformation, neglecting shear deformation. For thicker plates, the Mindlin‑Reissner theory incorporates shear effects. Plate natural frequencies are crucial for predicting sound radiation from panels.

Shell theory – extends plate theory to curved surfaces, accounting for membrane and bending stresses. Shells are common in aerospace and automotive structures. Their vibration characteristics are more complex because curvature couples bending and stretching, leading to a richer set of natural frequencies.

Acoustic coupling – the transfer of vibrational energy from a structure into the surrounding fluid (air or water). Coupling is quantified by the coupling loss factor, which relates the vibrational power input to the acoustic power radiated. Efficient coupling occurs when the structure’s surface velocity is high and the acoustic impedance of the medium matches the structural impedance.

Sound power – the rate at which acoustic energy leaves a source, measured in watts. In vibration‑driven sound, the radiated sound power is proportional to the square of the normal surface velocity integrated over the radiating area. Sound power is a fundamental metric in noise‑control engineering because it is independent of distance and environmental conditions.

Sound pressure level (SPL) – the logarithmic measure of acoustic pressure relative to a reference value (20 µPa in air). SPL is expressed in decibels (dB). When relating vibration to SPL, the conversion involves the radiation efficiency of the panel and the frequency‑dependent relationship between velocity and pressure.

Decibel (dB) – a dimensionless unit that expresses ratios on a logarithmic scale. In vibration, it is common to use dB for velocity (dB (v)) or acceleration (dB (a)). For example, a vibration velocity of 1 mm/s corresponds to 0 dB (v). Decibel scales compress large dynamic ranges into manageable numbers, but care must be taken to retain the correct reference.

A‑weighting – a frequency‑dependent filter that approximates the human ear’s sensitivity, emphasising frequencies around 2–4 kHz and attenuating very low and very high frequencies. A‑weighted SPL (dB (A)) is standard for environmental noise assessments. In vibration, A‑weighting can be applied to the velocity spectrum before converting to SPL for a more realistic perception‑based metric.

Vibration fatigue – the progressive and cumulative damage that occurs when a structure is subjected to cyclic stresses. Fatigue life is often estimated using S‑N curves (stress versus number of cycles) or Miner’s rule, which sums the damage fractions of each stress range. In structural acoustics, vibration fatigue is a concern for panels that experience continuous low‑amplitude oscillations, such as aircraft fuselage skins.

Failure criteria – quantitative limits that define when a component has reached an unacceptable condition. Common criteria include a maximum allowable stress, a maximum displacement, or a fatigue damage limit of 1.0 (According to Miner’s rule). Selecting appropriate failure criteria is essential for ensuring safety while avoiding over‑conservative designs.

Damage tolerance – the ability of a structure to sustain a flaw (e.G., A crack) without catastrophic failure. Damage‑tolerant design incorporates inspection intervals and crack‑growth predictions based on fracture mechanics. Vibration monitoring can be used to detect changes in modal parameters that indicate the presence of damage.

Non‑destructive testing (NDT) – techniques used to assess the integrity of a component without causing harm. In vibration analysis, modal testing itself is a form of NDT because changes in natural frequencies or damping ratios can reveal hidden defects. Complementary NDT methods include ultrasonic testing, radiography, and thermography.

Modal testing – the experimental procedure for extracting modal parameters. It typically involves exciting the structure with an impact hammer or electrodynamic shaker, measuring the response with accelerometers, and processing the data using techniques such as peak‑picking, circle‑fit, or polyreference methods. Accuracy depends on proper sensor placement, adequate excitation, and high coherence.

Impact hammer testing – a simple, low‑cost method of exciting a structure with a sudden force applied by a calibrated hammer. The hammer’s force transducer records the input, while accelerometers record the response. Impact testing is suitable for low‑frequency modes and for structures that cannot be attached to a shaker.

Shaker testing – the use of an electrodynamic or hydraulic shaker to apply controlled, repeatable forces over a wide frequency range. Shaker testing can provide higher excitation levels than impact hammers, making it appropriate for measuring high‑frequency modes and for investigating linearity by varying the input amplitude.

Accelerometer – a sensor that measures acceleration, typically using a piezoelectric ceramic element. Accelerometers are the most common transducers in vibration analysis because they have a wide frequency range and high sensitivity. Selecting the correct sensitivity (e.G., 10 MV/g versus 100 mV/g) is important for matching the expected amplitude and avoiding saturation.

Velocity sensor – devices such as laser vibrometers or electromagnetic velocity pickups that directly measure surface velocity. Velocity is often preferred for acoustic applications because the radiated sound power is proportional to velocity squared. Laser vibrometers provide non‑contact measurement, eliminating mass loading effects.

Displacement sensor – instruments that measure the absolute or relative motion of a point, such as capacitive probes or LVDTs (linear variable differential transformers). Displacement sensors are useful for low‑frequency, large‑amplitude motions, but they typically have lower bandwidth than accelerometers.

Laser vibrometer – a non‑contact optical instrument that measures surface velocity or displacement by detecting the Doppler shift of a reflected laser beam. It offers high spatial resolution, making it ideal for mapping mode shapes without attaching sensors. Challenges include surface reflectivity requirements and susceptibility to ambient light interference.

Data acquisition (DAQ) – the hardware and software system that digitises analogue sensor signals for analysis. Key DAQ specifications include sampling rate, resolution (bits), input range, and number of channels. For high‑frequency vibration work, a sampling rate of at least 5 kHz and a 16‑bit resolution are common minimums.

Sampling rate – the number of samples taken per second, expressed in hertz (Hz). The sampling rate must satisfy the Nyquist criterion for the highest frequency of interest. In practice, engineers often sample at 10‑20 times the highest frequency to allow for anti‑alias filtering and to improve signal‑to‑noise ratio.

Quantisation – the discretisation of an analogue signal into digital levels. Quantisation introduces noise that appears as a flat floor in the spectrum. Using higher‑resolution ADCs (e.G., 24 Bits) reduces quantisation noise, which is especially important when measuring low‑amplitude vibrations.

Signal‑to‑noise ratio (SNR) – the ratio of the desired signal power to the background noise power, usually expressed in decibels. High SNR is essential for accurate modal parameter extraction. Techniques to improve SNR include averaging, using low‑noise amplifiers, and proper grounding.

Filter – a signal‑processing element that attenuates specific frequency ranges. Filters can be low‑pass, high‑pass, band‑pass, or notch. In vibration analysis, filters are applied to remove unwanted noise (e.G., Mains hum at 50 Hz) or to isolate a frequency band of interest before performing an FFT.

Low‑pass filter – allows frequencies below a cutoff to pass while attenuating higher frequencies. It is used to eliminate high‑frequency noise that could alias into the frequency band of interest. Care must be taken to select a cutoff well below the Nyquist frequency.

High‑pass filter – permits frequencies above a cutoff to pass while suppressing low‑frequency drift. High‑pass filtering is useful when the sensor output contains large DC offsets or low‑frequency environmental vibrations that are not relevant to the analysis.

Band‑pass filter – combines low‑ and high‑pass characteristics to isolate a specific frequency band. Band‑pass filters are employed when focusing on a narrow resonant peak amidst broadband noise. The filter’s quality factor (Q) determines the sharpness of the passband.

Notch filter – a narrow band‑stop filter that removes a specific frequency, often used to suppress mains interference (50 Hz or 60 Hz) without affecting surrounding frequencies. Notch filters must be designed carefully to avoid removing energy from nearby structural modes.

Anti‑alias filter – a low‑pass filter placed before the A/D converter to attenuate frequencies above the Nyquist limit. Proper design of the anti‑alias filter ensures that out‑of‑band energy does not fold back into the usable spectrum, preserving the integrity of the measured data.

Window function – a mathematical function applied to a time‑domain signal to reduce spectral leakage. Common windows include Hann, Hamming, and Blackman, each offering a different trade‑off between main‑lobe width and side‑lobe suppression. The choice of window influences frequency resolution and amplitude accuracy.

Hann window – a cosine‑based window that provides good side‑lobe attenuation while moderately widening the main‑lobe. It is frequently used in vibration analysis because it balances resolution and leakage for most practical cases.

Hamming window – similar to the Hann window but with slightly higher side‑lobe levels and a narrower main‑lobe. It is useful when a modest improvement in amplitude accuracy is required.

Blackman window – offers the highest side‑lobe suppression among the common windows, at the cost of a broader main‑lobe. It is chosen when extremely low leakage is needed, for example when analyzing weak resonances adjacent to strong ones.

Averaging – the process of combining multiple spectra to reduce random noise. Averaging improves SNR proportionally to the square root of the number of averages. In practice, engineers may average 8, 16, or 32 spectra depending on the required confidence.

Ensemble averaging – averaging over multiple repetitions of an experiment under identical conditions. This technique isolates the deterministic part of the response and eliminates random variations. Ensemble averaging is essential when measuring FRFs with low excitation levels.

Statistical analysis – the application of probability theory to quantify uncertainty, confidence intervals, and repeatability of measured parameters. Techniques such as Monte Carlo simulation, bootstrapping, and Bayesian inference are increasingly used to assess the reliability of vibration‑based predictions.

Confidence interval – a range around an estimated parameter (e.G., Natural frequency) within which the true value is expected to lie with a given probability (commonly 95 %). Confidence intervals are derived from the variance of repeated measurements and are vital for specifying design margins.

Uncertainty – the combined effect of measurement errors, model approximations, and environmental variations. Uncertainty analysis follows guidelines such as the Guide to the Expression of Uncertainty in Measurement (GUM) and helps engineers allocate safety factors rationally.

Calibration – the process of verifying and adjusting the output of a sensor or measurement system against a known standard. Regular calibration of accelerometers, force transducers, and DAQ channels ensures traceability and reduces systematic errors.

Sensor placement – the strategic positioning of transducers to capture the most informative response. Placement guidelines include locating sensors at antinodes of expected mode shapes, avoiding nodal lines, and ensuring that the sensor’s mass does not significantly alter the dynamics.

Mounting – the method by which a sensor is attached to the structure. Common mounting techniques include adhesive bonding, magnetic mounts, and stud‑type fixtures. Improper mounting can introduce additional stiffness or damping, skewing the measured FRF.

Wiring – the routing and shielding of sensor cables. Poor wiring can lead to ground loops, electromagnetic interference, and signal attenuation. Twisted‑pair cables with proper shielding are standard practice to maintain signal integrity.

Grounding – establishing a common electrical reference for all measurement equipment. Correct grounding eliminates noise caused by potential differences and reduces the risk of damaging sensitive electronics. In vibration labs, a single‑point grounding scheme is often employed.

Environmental effects – factors such as temperature, humidity, and pressure that influence material properties and sensor performance. For example, the stiffness of an elastomeric mount can change dramatically with temperature, shifting the isolation natural frequency. Accounting for these effects is crucial when designing for outdoor or aerospace applications.

Temperature drift – the gradual change in sensor output caused by temperature variations. Many piezoelectric accelerometers exhibit a temperature coefficient of sensitivity, requiring temperature compensation or periodic recalibration.

Humidity – can affect the adhesive bond of sensors and the damping characteristics of polymeric materials. In high‑humidity environments, moisture absorption may increase material loss factors, altering the vibration response.

Structural damping – the inherent energy dissipation within the material and joints of a structure. Unlike viscoelastic damping, structural damping is often modelled as a constant loss factor across frequency. Accurate estimation of structural damping is essential for predicting the amplitude of resonant peaks.

Structural acoustics – the field that studies the interaction between structural vibrations and the surrounding acoustic field. It encompasses the prediction of sound radiation from vibrating panels, the design of quiet structures, and the assessment of vibration‑induced noise.

Coupling loss factor – a parameter that quantifies the efficiency of energy transfer from a vibrating structure into the acoustic medium. It is defined as the ratio of radiated acoustic power to the vibrational power input. Designers aim to minimise the coupling loss factor for noise‑control applications while maximising it for loudspeaker design.

Transmission loss – the reduction in sound power transmitted through a partition, expressed in decibels. Transmission loss is related to the vibrational response of the partition; a stiff, heavily damped panel typically yields high transmission loss.

Vibration transmission – the propagation of vibrational energy through a structural path. Transmission paths can be direct (through solid connections) or indirect (through fluid coupling). Mapping transmission paths helps identify critical points where isolation or stiffening will be most effective.

Isolation design – the process of selecting and sizing isolation elements to achieve a target reduction in transmitted vibration. Design steps include calculating the required natural frequency of the isolation system, choosing appropriate stiffness and damping, and verifying performance through FRF measurements.

Isolation criteria – quantitative guidelines that dictate acceptable vibration levels. Standards such as ISO 10816 provide vibration severity zones based on RMS velocity. Isolation criteria may also be defined by specific mission requirements, such as limiting cabin noise in aircraft to a certain dB(A) level.

ISO 10816 – an international standard that classifies vibration severity for rotating machinery based on measured velocity. It defines four zones, ranging from “good” to “dangerous,” and provides limits for different machine categories. Although primarily for machinery, the same limits are often applied to structural components when assessing vibration health.

ISO 16750 – a standard that addresses environmental test methods for road vehicles, including vibration testing. It specifies test spectra, test durations, and acceptance criteria for components destined for automotive applications. Familiarity with ISO 16750 is essential for engineers involved in vehicle‑level vibration validation.

ASTM standards – a collection of methods developed by the American Society for Testing and Materials. Relevant ASTM standards for vibration include ASTM E756 (Vibration Testing of Structures), ASTM E2275 (Vibration Testing of Aircraft Structures), and ASTM E292 (Procedures for Modal Testing). Compliance with ASTM methods ensures repeatability and comparability of results across laboratories.

Measurement uncertainty – the quantified doubt about the result of a measurement. It incorporates contributions from sensor accuracy, calibration, repeatability, and environmental influences. Reporting measurement uncertainty alongside vibration data is required for rigorous engineering documentation and for meeting certification requirements.

Repeatability – the variation observed when the same operator repeats a measurement under identical conditions. High repeatability indicates that the measurement system is stable and that random errors are low.

Reproducibility – the variation observed when different operators or different equipment perform the same measurement. Reproducibility assesses the robustness of the measurement protocol across different labs or test rigs.

Modal damping – the damping associated with a specific mode, often expressed as a damping ratio or loss factor. Modal damping can differ from the overall structural damping because it reflects the combined effect of material damping, joint friction, and energy radiation for that particular mode.

Modal participation factor – a scalar that indicates the contribution of a mode to the overall response under a given excitation. It is derived from the modal shape and the load distribution. Modes with high participation factors dominate the response and are the primary focus for mitigation.

Modal superposition – a mathematical technique that combines the contributions of individual modes to predict the total response to an arbitrary load. The method assumes linearity and requires knowledge of modal masses, stiffnesses, and damping ratios. It is widely used in finite‑element post‑processing to generate frequency‑response predictions.

Rayleigh damping – a damping model that expresses the damping matrix as a linear combination of the mass and stiffness matrices (C = αM + βK). Rayleigh damping provides a convenient way to assign frequency‑dependent damping ratios in FEM analyses. Selecting α and β to match desired damping ratios at two reference frequencies is a common practice.

Proportional damping – a special case of Rayleigh damping where the damping matrix is proportional to either the mass or stiffness matrix alone. Proportional damping leads to uncoupled modal equations, simplifying the modal analysis. However, it may not accurately capture the true damping behaviour of complex structures.

Non‑proportional damping – occurs when damping cannot be represented as a simple combination of mass and stiffness. In such cases, modes become coupled, and the standard modal superposition approach must be modified. Non‑proportional damping is often encountered in composite structures with viscoelastic layers.

Complex eigenvalue – an eigenvalue that includes an imaginary component, representing both frequency and damping. Complex eigenvalues arise naturally when solving damped vibration problems and provide direct access to the damped natural frequency and modal damping ratio.

Finite‑element eigenvalue extraction – the computational process of solving the generalized eigenvalue problem [K − ω²M]Φ = 0, where K is the stiffness matrix, M the mass matrix, ω the natural frequency, and Φ the eigenvector. Iterative solvers such as the subspace iteration method are used for large models.

Mass‑stiffness‑damping matrix – the three fundamental matrices that define the dynamic behaviour of a linear system. Accurate assembly of these matrices is vital; errors in mass distribution or stiffness definition can lead to significant discrepancies between predicted and measured natural frequencies.

Boundary conditions – constraints applied to the model that represent physical supports, connections, or symmetry conditions. Correctly modelling boundary conditions is essential because they strongly influence natural frequencies and mode shapes. Common boundary conditions include fixed, simply supported, and free edges.

Symmetry exploitation – the reduction of computational effort by modelling only a symmetric portion of a structure and applying appropriate symmetry boundary conditions. This technique halves or quarters the model size while preserving accuracy for symmetric mode shapes.

Mesh convergence – the practice of refining the finite‑element mesh until the predicted natural frequencies change by less than a prescribed tolerance (often 1 %). Convergence studies ensure that numerical discretisation does not dominate the results.

Element type selection – choosing between beam, shell, or solid elements based on the geometry and the dominant deformation mode. For thin panels, shell elements capture both bending and membrane actions efficiently; for thick components, solid elements may be necessary.

Mass scaling – a numerical technique that artificially increases mass in a model to allow larger time steps in transient simulations. While mass scaling can accelerate analyses, it must be applied judiciously because it alters the natural frequencies and can lead to inaccurate predictions.

Model validation – the process of comparing simulation results with experimental data to assess the fidelity of the model. Validation metrics include the percentage error in natural frequencies, mode shape correlation coefficients, and damping ratio discrepancies. Successful validation builds confidence for using the model in design optimisation.

Model updating – adjusting model parameters (e.G., Stiffness, damping, boundary conditions) to improve agreement with measured data. Techniques such as sensitivity analysis, gradient‑based optimisation, and Bayesian updating are employed to refine the model iteratively.

Sensitivity analysis – the study of how variations in model parameters affect the output (e.G., Natural frequencies). Sensitivity information guides designers on which parameters most influence vibration performance, allowing targeted design changes.

Design optimisation – the systematic adjustment of design variables to achieve objectives such as minimising weight while maintaining acceptable vibration levels. Multi‑objective optimisation may balance structural stiffness, acoustic radiation, and cost. Gradient‑based or evolutionary algorithms are commonly used.

Acoustic radiation efficiency – the ratio of radiated acoustic power to the vibrational power input.