Application of different acoustic emission descriptors in damage assessment of fiber reinforced plastics: A comprehensive review

https://doi.org/10.1016/j.engfracmech.2020.107083Get rights and content

Abstract

The review emphasis on the innovative methods of utilizing the Acoustic Emission (AE) parameters in characterizing Fiber Reinforced Plastics (FRPs). This review is structured into different sections based on different AE descriptors used in characterizing FRPs. This work summarizes all the acoustic parameters, their advantages, limitation and applications in the damage characterization of FRPs. The first section focuses on the peak amplitude/frequency and their limitations in characterizing damage modes. Followed by the cumulative acoustic energy: the novel and innovative methods in utilizing cumulative acoustic energy for damage characterization and crack propagation in FRPs. Emphasis on acoustic counts and acoustic events: some of the novel methods developed in utilizing the acoustic counts, follows this section. The clustering of AE data and its advantages in effective damage characterization is discussed. Finally, the list of other popular and unpopular AE descriptors and the reason for their limited usage in the FRPs is discussed.

Introduction

Before the 1960 s, the general use of sound waves or acoustic emission signals to characterize any material, ranging from wooden structures, mine shafts, composite pressure vessels, architectural structures, to name a few, under loading has largely been given trivial importance and often neglected by the engineering and scientific community [1]. Within a period of 10 years, the whole scenario was reversed. Acoustic emission gained attention and grew exponentially after 1970s. Several names were used to refer the acoustic sources such as clicks, sounds, micro seismic activity, rock noise, elastic radiation, elastic shock, seismo acoustical activity and so on. These sound waves were then used to characterize the material. The reason for the notoriety of the acoustic emission technique can be explained by Liptai [2], who quoted, “Acoustic emission analysis is very sensitive to transient instabilities”. And as inferred by Gillis [3], acoustic emission is stress waves or pressure waves arising from different sources owing to different phenomena.

Despite of the great interest in those first years in implementing Acoustic Emission (AE) only a countable number of reports can be found on characterizing Fiber Reinforced Plastics (FRP) using Acoustic Emission (AE) prior to 1970s. Although it has been argued by Liptai [1] that the AE has greater applicability in composites, it has not been implemented efficiently until the beginning of 21st century.

Fiber Reinforced Plastics (FRPs) are composite materials, in which high strength and modulus fibers such as glass fibers, carbon fibers, aramid fibers, etc., are reinforced in thermo or thermosetting plastics. Owing to the superior stiffness and high strength to weight ratio, the FRPs are replacing metals in most of the aerospace, domestic and industrial applications. The main advantage of the FRP is that their properties can be tailored for specific applications depending on the stacking type, stacking sequence and orientation of the fibers in the matrix. More details about the fundamentals and applications of FRP can be found in numerous literatures. The recent book by Beaumont and Zweben summarizes the fundamentals and applications of FRP/Polymer Composites. [4], [5]

The reason for the limited usage of AE in FRP in the early years is partially due to the non-isotropic nature of the FRP. The conventional principles are not applicable for them. For instance, Kaiser Effect proposed in 1950s for isotropic materials, which implies that the AE signals are irreversible in nature and are only emitted in subsequent loadings only after the previous maximum load is exceeded. However, in composites/FRPs, acoustic activity is observed even before the previous maximum load is exceeded. This can be calculated by Felicity ratio, which is the ratio of load applied when the acoustic activities resume to the maximum previously applied load. [6]. However, due to the recent advancements in using AE descriptors, characterizing the composite materials, not to mention, FRP has become possible.

It is a well-known phenomenon that the acoustic events increase with the increase in stress of a composite material. Generally, when there are no large stress concentrations, acoustic events are generated uniformly throughout the FRP. This large number of acoustic events can start accumulating as early as 10% of the maximum the ultimate failure stress of a composite material. This is because the fiber oriented in a direction of the FRP laminate will induce microdamage locally along the transverse direction of the matrix. This microdamage is distributed along the FRP due to the weak and compliant nature of the matrix. This uniform damage state during the initial loading of a composite material is known as characteristic damage state [7]. By the Felicity effect, acoustic events are generated in the FRP even before the initial maximum load is reached during the subsequent loading cycles [6].

It has been established nearly 50 years ago by Krock and Broutman [8] and corroborated by Liptai [2] in his research work that the AE can be generated in three different modes by straining a FRP composite material. These modes of AE generation are plastic deformation of the matrix, fracture of the fiber and pullout at the interface between matrix and the fiber. Other than these classical damage modes, interlaminar crack growth, sliding of microcracked particles in the matrix, transverse cracks in cross-ply laminates can also produce both low-frequency symmetric and high-frequency nonsymmetric acoustic signals in different time–frequency domains. Over the years this has been experimentally proved by employing different types of in-situ damage monitoring techniques, which will be explained in detail in the subsequent sections.

The microcracks in the matrix is the major reason for this uniform stress concentration. Due to the viscoelastic nature of the polymer matrices, the acoustic events generated during this loading period has severe attenuation in high frequency components [9]. An experimental campaign by Awerbuch et al. [10] shows that the friction between the fractured laminates accumulates a significant number of acoustic events and in some cases, these accumulated acoustic events are much larger than the AE signals produced by the actual damage progression. Awerbuch and Ghaffari [11] followed these results by characterizing the AE signals produced by actual damage progression and the friction between the fracture laminates in double-edge notched unidirectional graphite/epoxy composites.

The other basic source of AE signal generation is majorly concentrated on the location close to the local stress concentration and accumulation of damage before the ultimate failure of the FRP [3], [7]. The location of these AE sources has a small region of origin due to the local stress concentration. However, the damage distribution can alter the elastic properties of the material and consequently, AE signals will be generated from the other locations than the local failure region. Nonetheless, the AE signals concentrated at the region of failure and the other regions vary significantly in terms of frequency, amplitude and duration [11].

Beyond these, the acoustic signals can be generated due to the rupture of the fibers, pullout of the fibers from the matrix, propagation of crack transversely through the interlaminar plies and so on. Each of the acoustic event associated with these damage propagations have unique features in terms of frequency, risetime among the other AE descriptors [2], [8]. This makes the AE technique a powerful tool in understanding and characterizing the damage progression and material properties of a composite material, particularly FRP. The AE descriptors associated with the different damage progressions has been often misinterpreted by the researchers. Thus, it is crucial to understand the major AE descriptors which has been commonly used to characterize the materials.

It has been eloquently said by Gillis [3] that the acoustic events recorded using a transducer is predominantly based on the settings selected by individual tester. In addition to that, there are numerous AE descriptors available to characterize the material properties, albeit, each having their inherent advantages and limitations.

Most of the researchers in the recent past have been using advanced data acquisition (DAQ) systems to record the piezoelectric transducer signal of the acoustic events. These signals are amplified by amplifiers and filtered through low and high bandpass filters before being converted into the AE descriptors such as peak amplitude, energy and also extracts features such as risetime, duration, counts and so on.

The most commonly used AE descriptor for discriminating the type of damage progression in FRPs is peak amplitude. The peak amplitude represents the largest voltage peak (Umax) in the recorded acoustic waveform with respect to the reference voltage (Uref) set at the pre-amplifier. The peak amplitude (A) can be expressed in dB as,A=20logUmaxUref

The other commonly used AE descriptor associated with the peak amplitude is the peak frequency. The peak frequency (Fp) is the maximum magnitude in the power spectrum of the recorded acoustic event. Most of the DAQs have an inbuilt system which transforms the recorded AE waveform by Fast Fourier Transform (FFT) and provides the maximum magnitude in the FFT spectrum. Traditionally, the FFT has been used to extract the peak frequency of the acoustic event.

The consideration of centroid frequency (Fc), however, provides more significance to the power spectrum than the peak frequency [12]. The frequency centroid is the weighted average of the frequency content and can be expressed as,Fc=fminfmaxUf.ffminfmaxUf

U(f) is the spectrum amplitude at each point of the frequency spectrum f, fmin and fmax are the minimum and maximum spectrum frequency, respectively.

Due to the resonant nature of the sensor, some researchers have used the weighted peak frequency (FWP) and integrated with the frequency centroid to characterize the damage progression in FRP [13], [14]. The weighted peak frequency can be expressed as:FWP=Fc.Fp

One of the most commonly used AE descriptors from the 1960 s until the present time and perhaps, beyond any stipulated time, is the counts. The count is an integer value representing the number of times the amplified sensor signal crosses the set trigger voltage of the counter. Larger acoustic events produce large number of counts for a single hit as amplitude of the acoustic wave requires more ring down time to go below the trigger voltage [15].

The AE energy (EAE) has been used widely in the recent years to characterize the mechanical properties of the FRPs. It was introduced by Harris and Bell [16], who experimentally proved its superior characterization power over number of acoustic counts and events. More details are provided in the subsequent section. The AE energy can be estimated by integrating the rectified transient voltage (Ui) of the recorded acoustic event over the period of time. The AE energy over time period t0 to ti can be expressed as,EAE=t0tiUi2tdt

Besides the AE energy, the partial power has been used by considerable number of researchers to characterize the damage progression, while others have used risetime, duration, number of acoustic hits and number of counts per event to name a few. Partial power is the percentage of power spectrum calculated from a specified range of frequency. It is the cumulative sum of the energy spectrum measured from the acoustic waveform in the user-specified frequency range. A pictorial representation of the acoustic waveform and the important descriptors are provided in Fig. 1.

In the subsequent sections, the characterization of damage progression and structural integrity of FRPs using different AE descriptors have been explained. The advantages, limitations and the best application in which each descriptor is more suited for has been discussed. It must be understood that each article which has been used for the review purpose has used AE in different ways. The purpose of this research work is to summarize the different perspectives and analyse the different methodologies proposed by various researchers.

Section snippets

Characterization of damage progression using peak amplitude and peak frequency

Peak amplitude of the acoustic events is rarely used a solitary parameter for damage characterization in composite material or in any materials in general. It is often commonly associated with the peak frequency. The peak amplitude is a function of peak voltage recorded during the acoustic event and the peak frequency simply represents the singular point of maximum frequency in the recorded signal. They are closely related to each other and it only makes sense to use them together to validate

Characterization of damage progression using acoustic energy

Cumulative acoustic energy (CEA) is one of the most commonly used AE descriptors in characterizing damage progression and material characterization in composite materials. Cumulative energy is relative and cannot be compared between different materials, nonetheless, they can be used effectively to map the damage progression. In simple words, cumulative energy curve of a tensile test results produced by a researcher cannot be compared with cumulative energy curve of the same test replicated by

Characterization of damage progression using acoustic counts and events

From the inception of acoustic emission in damage characterization, acoustic counts and events have played a crucial role. Over the years, many researchers have employed different ways of using acoustic emission counts/events as the prime function in assessing the material characteristics. Therefore, it is quite difficult to disregard any of their research work but in the meantime, even more difficult to summarize them all under a simple context. In that regard, a particular set of research

Clustering of acoustic data

Clustering of acoustic data in FRPs is popularly based on the fact that the failure modes in composites are in four different modes: matrix cracking, delamination in matrix/fiber interface, fiber pullout and fiber breakage. For this reason, most researchers would cluster their recorded acoustic parameter into four (sometimes three) classes. Using this, they can directly attribute each class of data to the failure mode.

Roundi et al. [58] has used an unsupervised pattern recognition algorithm,

Other popular and unpopular methods of using acoustic emission technique in characterizing FRPs

Besides the aforementioned parameters, the other popular AE descriptors rarely used by the researchers are partial power, rise time and duration. Based on the interest of the researchers while clustering the AE data and for other specific applications, some of these parameters are used. For instance, Kempf et al. [65] used three levels of partial power, 1: 0–250 kHz, 2: 250–450 kHz and 3: 450–800 kHz, to characterize the effect of tensile load on glass fiber reinforced polyurethane composites.

Conclusions

A review on predominantly used acoustic emission descriptors, their advantages, limitations and applications has been summarized in this paper. The important topics covered in this area are AE characterization using peak amplitude, peak frequency, energy, counts and events. Innovative ways of using each of these AE descriptors has been discussed.

  • a)

    The erroneous usage of peak amplitude and peak frequency has been discussed. The reason why the high amplitude/frequency signals being associated with

Declaration of Competing Interest

The authors declared that there is no conflict of interest.

References (85)

  • F. Meraghni et al.

    Micromechanical modelling of matrix degradation in randomly oriented discontinuous-fibre composites

    Compos Sci Technol

    (1995)
  • C. Barile et al.

    Experimental wavelet analysis of acoustic emission signal propagation in CFRP

    Engng Fract Mech

    (2019)
  • M.G.R. Sause et al.

    Quantification of failure mechanisms in mode-I loading of fiber reinforced plastics utilizing acoustic emission analysis

    Compos Sci Tech

    (2012)
  • C. Barile et al.

    Acoustic emission waveform analysis in CFRP under Mode I test

    Engng Fract Mech

    (2019)
  • D.J. Aggelis et al.

    Acoustic structural health monitoring of composite materials: damage identification and evaluation in cross ply laminates using acoustic emission and ultrasonics

    Compos Sci Tech

    (2012)
  • D.V. Soulioti et al.

    Acoustic emission behaviour of steel fiber reinforced concrete under bending

    Constr Build Mater

    (2009)
  • R. Mohammadi et al.

    Correlation of acoustic emission with finite element predicted damages in open-hole tensile laminated composites

    Compos Part B-Eng

    (2017)
  • C. Baker et al.

    Transverse cracking in carbon fiber reinforced polymer composites: Modal acoustic emission and peak frequency analysis

    Compo Sci Tech

    (2015)
  • F. Cesari et al.

    Damage and residual strength of laminated graphite-epoxy composite circular plates loaded at the centre

    Compos Part A-Appl S

    (2007)
  • M. Saeedifar et al.

    Prediction of delamination growth in laminated composites using acoustic emission and Cohesive Zone Modeling techniques

    Compos Struct

    (2015)
  • F. Lissek et al.

    Acoustic emission for interlaminar toughness testing of CFRP: Evaluation of the crack growth due to burst analysis

    Compos Part B-Eng

    (2018)
  • W. Roundi et al.

    Acoustic emission monitoring of damage progression in Glass/Epoxy composites during static and fatigue tensile tests

    Appl Acoust

    (2018)
  • N. Godin et al.

    Clustering of acoustic emission signals collected during tensile tests on unidirectional glass/polyester composite using supervised and unsupervised classifiers

    NDT&E Int

    (2004)
  • J.P. McCrory et al.

    Damage classification in carbon fibre composites using acoustic emission: a comparison of three techniques

    Compos Part B-Eng

    (2015)
  • M.G. Baxter et al.

    Delta T source location for acoustic emission

    Mech Syst Signal Process

    (2007)
  • M. Kempf et al.

    Acoustic emission analysis for characterisation of damage mechanisms in fibre reinforced thermosetting polyurethane and epoxy

    Compos Part B-Eng

    (2014)
  • T.P. Philippidis et al.

    Using acoustic emission to assess shear strength degradation in FRP composites due to constant and variable amplitude fatigue loading

    Compos Sci Tech

    (2008)
  • M. Kharrat et al.

    A signal processing approach for enhanced Acoustic Emission data analysis in high activity systems: Application to organic matrix composites

    Mech Syst Signal Pr

    (2016)
  • Amir Mirmiran et al.

    Comparison of acoustic emission activity in steel-reinforced and FRP-reinforced concrete beams

    Constr Build Mater

    (2000)
  • A. Carpinteri et al.

    From criticality to final collapse: Evolution of the “b-value” from 1.5 to 1.0

    Chaos Soliton Fract

    (2009)
  • M. Shigeishi et al.

    Acoustic emission moment tensor analysis: development for crack identification in concrete materials

    Constr Build Mater

    (2001)
  • F. Yu et al.

    A novel method of identifying damage types in carbon fiber-reinforced plastic cross-ply laminates based on acoustic emission detection using a fiber-optic sensor

    Compos Sci Tech

    (2016)
  • C. Barile et al.

    Acousto-ultrasonic evaluation of interlaminar strength on CFRP laminates

    Compos Struct

    (2019)
  • P. Sedlak et al.

    Acoustic emission localization in thin multi-layer plates using first-arrival determination

    Mech Sys Signal Process

    (2013)
  • E. Maillet et al.

    Waveform-based selection of acoustic emission events generated by damage in composite materials

    Mech Sys Signal Process

    (2015)
  • R. Sleeman et al.

    Robust automatic P-phase picking: an on-line implementation in the analysis of broadband seismogram recordings

    Phys Earth Planet Inter

    (1999)
  • R.G. Liptai et al.

    An introduction to acoustic emission

  • R.G. Liptai

    Acoustic emission from composite materials. In. Corten H, Editors. Composite Materials: Testing and Design, ASTM STP 497. West Conshohocken, PA

    ASTM International

    (1972)
  • Gillis PP. Dislocation motions and acoustic emissions. In: Liptai R, Harris D, Tatro C, Editors. Acoustic Emission,...
  • Beaumont PWR, Zweben CH. Polymer Matrix Composites: Fundamentals. In: Beaumont PWR, Zweben CH. Comprehensive Composite...
  • Beaumont PWR, Zweben CH. Polymer Matrix Composites: Applications. In: Beaumont PWR, Zweben CH. Comprehensive Composite...
  • M.A. Hamstad

    Thirty years of advances and some remaining challenges in the application of acoustic emission to composite materials

  • Cited by (99)

    View all citing articles on Scopus
    View full text