Acoustic Emission Testing (AET) — Principles, Parameters & Applications in Welded Structures

Acoustic Emission Testing (AET) is one of the few non-destructive testing techniques that listens to a structure rather than interrogating it with an external energy source. When materials deform, crack, or corrode, they release energy in the form of elastic stress waves — transient mechanical disturbances that propagate through the material and can be detected by piezoelectric sensors attached to the surface. By capturing, locating, and characterising these signals in real time, AET gives engineers an unparalleled window into the dynamic behaviour of welds, pressure vessels, storage tanks, pipelines, and composite structures under actual or simulated service loading.

Unlike volumetric methods such as radiography or ultrasonic testing, which detect existing geometric discontinuities, AET detects only active mechanisms — those currently releasing energy. A fatigue crack propagating under cyclic load, stress corrosion cracking advancing under residual stress, or a delaminating composite joint will all produce distinct acoustic signatures. This intrinsic selectivity for active damage, combined with the ability to survey an entire structure from a sparse array of sensors, makes AET ideally suited for in-service monitoring, proof-load testing of pressure equipment, and structural health monitoring (SHM) programmes where continuous data are required without interrupting production.

This guide covers the complete technical foundation of AET: the physics of stress-wave generation, the critical signal parameters used in analysis, source-location algorithms, the Kaiser and Felicity Effects that underpin load-cycle interpretation, the principal industrial applications in welded and pressure-containing equipment, and the ASME/ASTM code framework that governs AE inspections. Whether you are a Level II or III AE practitioner, a welding or pressure-vessel engineer seeking to understand what AET can and cannot detect, or a student preparing for a certification examination, this article provides the depth and practical context you need.

The Physics of Acoustic Emission

Stress-Wave Generation Mechanisms

Acoustic emission arises whenever a rapid, localised energy release occurs within or at the surface of a solid material. The energy release perturbs the surrounding elastic continuum, launching elastic waves that propagate outward from the source. In metallic welded structures the most important source mechanisms are:

• Crack initiation and propagation — the most energetic and practically significant source; each incremental crack-advance event releases a burst of elastic energy proportional to the new crack surface area created.
• Dislocation motion and plastic deformation — individual dislocation glide and pile-up produce low-amplitude, high-frequency emission; collective plastic flow during yielding produces broad-band continuous emission.
• Inclusion fracture and debonding — brittle inclusion particles (oxides, sulphides) in the weld metal or HAZ can fracture under load, producing sharp, high-amplitude bursts.
• Martensitic transformation — the diffusionless, shear-type transformation from austenite to martensite (relevant to hardenable steel welds on cooling) produces characteristic burst emission.
• Stress corrosion cracking (SCC) — anodic dissolution, hydrogen embrittlement, and film rupture mechanisms each contribute to emission during SCC propagation.
• Leak-through defects — pressurised fluid escaping through a crack or pinhole produces sustained turbulence-driven emission in the frequency range 100 kHz to 1 MHz.

Wave Modes and Propagation

In solid structures, AE sources launch both body waves (longitudinal/compressional and transverse/shear) and, in plate-like geometries, guided Lamb waves. In the thin-shell structures typical of pressure vessels and pipelines, Lamb wave propagation dominates because the plate thickness is small relative to the acoustic wavelength. Lamb waves exist in two families — symmetric (extensional) and antisymmetric (flexural) modes — each with frequency-dependent velocities (dispersion). Understanding wave-mode content is critical for accurate source location because the velocity used in time-of-arrival calculations must match the mode actually being detected.

Attenuation and its Practical Consequences

As AE waves travel through the structure, their amplitude decreases due to geometric spreading (amplitude decreasing with distance), material absorption (conversion to heat by internal friction), and scattering at grain boundaries, welds, and structural features. The attenuation rate is material- and geometry-dependent and must be measured during the pre-test survey. Typical attenuation values for common materials are:

High attenuation in FRP and concrete structures necessitates dense sensor arrays and limits detection capability to sources within a relatively short range of each transducer.

AE Instrumentation and Signal Parameters

Transducer Design and Selection

AE sensors are piezoelectric devices that convert the mechanical surface displacement produced by an arriving stress wave into a voltage signal. Most industrial AE sensors use lead zirconate titanate (PZT) ceramic elements operating in resonance or broad-band mode. The key selection parameters are:

• Frequency response — resonant sensors optimised for a single frequency (typically 100 kHz, 150 kHz, or 300 kHz) offer maximum sensitivity in a narrow band; broad-band sensors (50 kHz to 1 MHz) provide waveform fidelity for modal analysis and source characterisation.
• Sensitivity — expressed as voltage output per unit velocity or displacement at the sensor face; higher sensitivity reduces the detection threshold but increases susceptibility to electromagnetic interference (EMI).
• Temperature rating — standard sensors operate to approximately 75 °C; high-temperature sensors with waveguide extensions are required for service above this limit.
• Housing and mount — magnetic hold-down, adhesive bond, or mechanical clamp; the mount must maintain consistent coupling pressure and couplant layer thickness throughout the test duration.

Key Signal Parameters in AET

Modern multi-channel AE data acquisition systems digitise the pre-amplified sensor output and extract a set of scalar descriptors from each detected hit. These parameters form the basis of hit-based analysis:

Threshold, Pre-amplification, and Dynamic Range

The detection threshold is set above the ambient noise floor to suppress electrical noise and mechanical background (e.g., pump vibration, fluid flow). A threshold of 35 to 45 dBAE is typical for well-shielded industrial systems in a quiet environment. Pre-amplifiers with gains of 40 or 60 dB are placed within 1 m of each sensor to boost the signal before transmission through instrument cables. The dynamic range of modern 18-bit systems spans approximately 100 dB, allowing simultaneous capture of microseismic sources and large burst events.

Source Location Techniques

Linear (1D) Source Location

Linear location uses two sensors and computes the source position along a linear structure (pipeline, beam) from the arrival-time difference. The formula is:

If Δt = 0 the source is equidistant from both sensors (midpoint). If Δt is at its maximum value (D/v), the source is at one of the sensors.

Planar (2D) Source Location

With three or more sensors, planar triangulation uses pairs of TDOA measurements to construct hyperbolic loci on the structure surface. The intersection of these hyperbolae provides the 2D source location. A minimum of three non-collinear sensors is required; four or more sensors over-determine the system and allow least-squares fitting that reduces the effect of timing errors.

Worked Example: 2D Location

The Kaiser Effect and Felicity Effect

Two phenomena central to the interpretation of load-cycle AE data are the Kaiser Effect and the Felicity Effect. Both were first described in metallic materials but have since been documented in composites, rock, and concrete.

Kaiser Effect

The Kaiser Effect (named after Josef Kaiser, 1950) states that a material subjected to repeated loading does not produce acoustic emission during reloading until the previously applied maximum stress level has been exceeded. In practical terms, if you load a structure to load P1, unload it, then reload it, AE activity is suppressed until the load again reaches P1. The Kaiser Effect is associated with materials in which the micro-damage mechanisms responsible for emission are not active at stress levels below the prior maximum. Its presence confirms that no new damage has accumulated between load cycles — a healthy sign during proof-load testing.

Felicity Effect and Felicity Ratio

The Felicity Effect describes the breakdown of the Kaiser Effect: significant AE is produced at a load below the previous maximum. This occurs when active, progressive damage mechanisms (growing cracks, active corrosion, delaminations) respond to any re-application of load. The Felicity Ratio (FR) quantifies this:

AE Signal Analysis Methods

Hit-Based (Parametric) Analysis

The most widely used industrial AE analysis method examines distributions of the extracted hit parameters — amplitude distribution, cumulative energy versus time, hit rate versus load, and source location maps. Amplitude distribution analysis (b-value analysis, analogous to seismology) characterises the relative proportion of small to large events. A decreasing b-value (population shifting toward high amplitudes) indicates progression from distributed microcracking to dominant macro-crack growth, an important warning sign in structural integrity assessment. Source location clusters that tighten or intensify with load increments indicate concentrations of damage that warrant follow-up inspection.

Waveform-Based Analysis

Modern systems digitise full waveforms at sampling rates of 1 to 40 MHz. Waveform analysis enables frequency-domain characterisation (FFT spectra), wavelet transform analysis to separate wave modes, and pattern recognition algorithms. Modal AE (MAE) analysis identifies the specific Lamb wave modes present in each hit and uses their known dispersion curves to deduce source depth and mechanism — distinguishing, for example, a surface crack from a mid-thickness inclusion fracture.

Noise Discrimination

Effective noise rejection is among the most challenging aspects of field AET. Established discrimination strategies include:

• Duration-amplitude filtering — mechanical noise sources (friction, impacts) typically produce high-count, long-duration signals relative to their amplitude. Plotting Duration vs Amplitude allows graphical separation of noise from genuine AE.
• Rise-time ratio — the RA value (Rise Time / Amplitude) is used in concrete and structural AET to separate tensile cracking (low RA) from shear or friction mechanisms (high RA).
• Guard sensors — sensors placed near known noise sources (valves, pumps) and used to veto hits arriving at nearby structural sensors within a defined time window suppress noise from those sources.
• Frequency filtering — many mechanical noise sources are richer in low-frequency content (<50 kHz) than crack-related AE; band-pass filtering (100–400 kHz) reduces their contribution to the data set.

AET in Welded Structure Inspection

Weld Defect Detection

Welded joints are the primary targets of AET in pressure equipment because welds concentrate residual stress, may contain embedded flaws (lack of fusion, porosity, hydrogen cracks), and are the most likely sites for in-service fatigue crack initiation. AET is applied to welded structures in two principal modes:

• Proof-load monitoring — the vessel or structure is loaded to a proof pressure (typically 1.25× MAWP for ASME vessels) while AE is continuously monitored. Active weld defects produce characteristic burst emission clusters that locate the flaw. The Kaiser Effect / Felicity Ratio criteria determine the severity classification.
• In-service surveillance — permanently installed sensor arrays monitor structures under normal operating loads. Fatigue crack growth during production cycles produces measurable AE well before the crack reaches critical size, enabling condition-based maintenance decisions.

For fracture mechanics-grade assessment, the AE hit rate from a growing crack can be correlated with crack velocity and stress intensity factor range using Paris law analogy. This is described in detail in our guide to mechanical testing methods, which covers fracture toughness test interpretation that underpins AE severity grading.

HAZ Monitoring During Post-Weld Heat Treatment

Some AE systems are applied during post-weld heat treatment (PWHT) of large fabrications. During heating through the stress-relief temperature range, residual stresses drive localised plastic flow and, in susceptible materials, reheat cracking. AE provides real-time detection of cracking events during PWHT, enabling the process to be halted and the weld area inspected before further damage accumulates. This is particularly valuable for heavy-section P91 and P22 chrome-moly welds where reheat cracking sensitivity is elevated — see our P91 welding requirements guide for the metallurgical background.

Limitations for Weld Inspection

AET is a passive, real-time technique and cannot image the volumetric geometry of a defect in the manner that phased array UT (PAUT) or TOFD can. It locates the source of emission but does not directly measure crack height, length, or through-wall extent. Accordingly, AET is most effective as a screening tool to identify weld zones requiring detailed volumetric follow-up, rather than as a standalone accept/reject method. Regulatory bodies and pressure-equipment codes treat AET as a monitoring technique, complementary to periodic volumetric methods.

Pressure Vessel and Storage Tank Applications

ASME Section V Article 12 — Metallic Pressure Vessels

ASME Section V Article 12 provides mandatory requirements for the AE examination of metallic pressure vessels. Key provisions include:

• All AE examinations must be performed by personnel certified to SNT-TC-1A Level II or III, or the equivalent ASNT certification scheme.
• The pre-test survey must include PLB (pencil-lead break) calibration at every sensor and a noise-rejection survey under pressurisation conditions.
• The vessel is pressurised in steps — typically 5 to 10% of test pressure per step — while AE data are continuously acquired. A hold period at each step allows transient noise from pressure-induced deformation to decay before evaluation.
• Evaluation criteria include: intensity classification (cluster significance), Felicity Ratio calculation at each re-pressurisation, and comparison against the Article T-1290 acceptance standard.
• Vessels failing the AET acceptance criteria are subject to additional examination by volumetric methods (RT or UT) to characterise the flagged regions.

Aboveground Storage Tank Floor Testing

Scanning tank floors by conventional NDT requires out-of-service periods, product removal, and entry into confined spaces. AET offers an in-service alternative: sensors are attached to the tank shell near the floor-to-shell junction and the floor is flooded with product or water; active corrosion, active weld cracking, and through-floor pinholes all produce characteristic emission. The technique is governed by ASTM E1930 (Standard Practice for Examination of Liquid-Filled Atmospheric and Low-Pressure Metal Storage Tanks) and is widely used in refinery and petrochemical facility integrity management programmes.

Source location on a flat tank floor uses a planar sensor array on the shell perimeter. Active areas are ranked by emission intensity and location, directing the follow-up inspection to specific quadrants of the floor rather than requiring full manual scanning of the entire surface area.

FRP Pressure Vessels and Pipelines

Acoustic emission is the primary volumetric inspection method for fibre-reinforced plastic (FRP) pressure equipment because conventional RT and UT are poorly adapted to composite laminates. During proof-test pressurisation, fibre breakage, matrix cracking, and delamination each produce characteristic AE signatures distinguishable by amplitude, frequency, and waveform shape. ASTM E1186 and E1932 provide the acceptance criteria framework, and the Felicity Ratio is the principal integrity indicator, with acceptance thresholds typically set in the range 0.85 to 0.90 depending on vessel age and service history.

AET in Structural Health Monitoring

Beyond discrete inspection events, AET is increasingly deployed as a continuous structural health monitoring (SHM) technology, integrating with plant automation systems and digital twin platforms. Permanently installed sensor arrays on bridges, offshore structures, crane booms, pressure vessels, and nuclear containment structures provide continuous, condition-based monitoring. Data management platforms compute hit-rate trends, location cluster evolution, and energy accumulation rates, triggering maintenance alerts when pre-defined thresholds are exceeded.

The integration of AE with other SHM modalities — vibration analysis, strain gauging, and corrosion sensors — enables sensor fusion approaches that improve both detection reliability and false-alarm discrimination. Distributed sensing over large structures using optical fibre Bragg grating (FBG) sensors in place of piezoelectric elements is an active area of development, offering inherent immunity to EMI and the ability to interrogate sensor arrays over multi-kilometre runs from a single instrument.

AET for Corrosion and Leak Detection

Active Corrosion Detection

Corrosion reactions — particularly those involving hydrogen evolution (acid attack, cathodic protection breakdown) and stress corrosion cracking — generate detectable AE. The hydrogen micro-bubbles produced at corroding surfaces collapse and produce characteristic high-frequency emission that can be distinguished from mechanical background noise by its statistical distribution and frequency content. In pipelines carrying hydrocarbon streams with H2S content (sour service), AET monitoring provides early warning of active SCC before crack sizes reach detectable dimensions by conventional UT. Our sour service materials guide covers the metallurgical context for this application.

Pressurised Leak Detection

Pressurised fluid escaping through a tight crack or pinhole produces intense turbulent flow that excites AE in the frequency range 100 kHz to 1 MHz. This signal is typically continuous rather than burst-type and decays in amplitude with distance from the leak point according to the attenuation characteristics of the structure. AE leak detection is used on gas pipelines, buried utilities, valve seats, heat exchanger tube-to-tubesheet joints, and pressure-relief valve seats. The technique can detect leaks with flow rates as low as 1 ml/min in favourable acoustic conditions and is significantly more sensitive than conventional snoop or tracer gas methods for tight, active leaks.

Applicable Codes and Standards Summary

Comparison of AET with Other NDT Methods

In practice, AET and volumetric UT/PAUT are complementary. AET efficiently identifies the weld zones with active emission; PAUT then provides accurate sizing of the flagged flaws to support fitness-for-service assessments under fracture mechanics frameworks.

Recommended Books on Acoustic Emission Testing

Disclosure: WeldFabWorld participates in the Amazon Associates programme (StoreID: neha0fe8-21). If you purchase through these links, we may earn a small commission at no extra cost to you. This helps support free technical content on this site.

Frequently Asked Questions

Related Technical Resources