This article presents the development of an experimental methodology based on acoustic emission wave detection for determining delamination onset and propagation in carbon fibre composite materials under quasi-static and fatigue loading. Delamination was investigated in quasi-static interlaminar fracture testing over a wide range of mixed-mode ratios (G II /G T = 0, 0.3, 0.5 and 1) for unidirectional and woven samples. An acoustic emission wave detection method was developed to detect delamination onset, and the corresponding fracture toughness was computed. Interlaminar fracture toughness was also calculated by beam theory and from finite element analysis with the virtual crack closure technique. The mechanical testing results, acoustic emission monitoring and numerical model's interlaminar fracture toughness were used to define delamination initiation criteria by drawing two-dimensional envelopes corresponding to G C = f(G II /G T ). The acoustic emission wave detection method showed damage accumulation before observable crack propagation, and its failure envelope corresponded to lower fracture energies than the standard test and modelling methods. Mode I fatigue testing with acoustic emission monitoring was performed on the woven samples for different energy release rate ratios (G IMAX /G IC = 0.3-0.8). A first series of samples were tested to construct an onset delamination fatigue curve DG = f(N). A second series of samples were used to study the cumulative acoustic emission energy distribution during delamination growth. An unsupervised pattern recognition methodology is presented for crack opening and closing testing, in order to discriminate between fatigue signal noise and acoustic emission signals emitted from crack initiation and crack growth. Correlations were observed between the acoustic emission energy distribution, the load range, the delamination length and the crack growth rate.

In this paper, a structural health monitoring strategy for detecting interlaminar delamination in a carbon fiber reinforced polymer structure using Lamb waves is proposed. The delamination is simulated by inserting a Teflon tape between two transverse plies and the Lamb wave generation and measurement is enabled by using piezoceramic elements. The Lamb wave theoretical propagation and through thickness strain distribution are studied, in order to determine the optimal configuration of the final system in terms of mode and frequency selection, and piezoceramic sizing and spacing, for detection of cross-sectional delamination. Pitch and catch measurements are performed by comparing wave propagations for different frequencies and along damaged and undamaged paths of the structure, and the analysis of results is performed using the reassigned short time Fourier transform. It appears that in the low frequency range (below 300 kHz), the A0 mode is sensitive to the damage, while in the high frequency range, S1 and A1 modes are both very sensitive to the damage while the propagation of the S0 mode is not affected very much.

SUMMARY Damage monitoring, failure prognostics, and remaining service life prediction represent great technological challenges for reliable maintenance of aeronautical structures. Consequently, there is a whole variety of non‐destructive evaluation techniques that are available to ensure the quality of the metallic structures. The majority of these techniques are able to detect defects such as discontinuities of surface or volume and the variations of section and are applied to discrete intervals. The application of these techniques proves too long and generates high maintenance costs. This paper proposes experimental methodologies to monitor fatigue damage growth in real time and also use a physics‐based model for fatigue life prediction. The aluminum alloy samples, with inserted pre‐cracks in the fastener holes, was tested mechanically in fatigue tension–tension cyclic loading with follow‐up of two complementary health monitoring techniques such as acoustic emission (AE) and acousto‐ultrasonic (AU). The approach uses AE to detect fatigue crack initiation and crack growth in aluminum samples and also applies AU measurements in order to assess global health conditions and damage accumulation during tension–tension cyclic loading. Nasgro analytical fracture mechanic model was used to predict crack growth and to determine the number of the load cycles Nf required to grow the initial crack to final crack size ac. The results indicate that exploiting health monitoring data such as AE signals coupled with analytical physics‐based models provides a convenient methodology to determine the fatigue life and to estimate safety factor on life of the materials tested. Furthermore, this paper demonstrates that AU flexural Lamb wave (A0) offers high potential to track damage before the occurrence of the first crack, such as fatigue damage caused by plastic deformation, and quantitatively to assess fatigue damage stages of the material. Copyright © 2011 John Wiley & Sons, Ltd.

The global mechanical properties of composite structures in service depend on fiber/matrix interface and interlaminar strength. The paper proposes to use Acoustic Emission (AE) and advanced signal processing to evaluate the interlamnar performance of polymeric composites. A delaminating process simulated with a Double Cantilever Beam (DCB) in opening mode (Mode I) coupled with a n Acoustic Emission (AE) technique has been employed. Different samples were analyzed to observe the damage evolution and to evaluate the interlaminar decohesion processes. The resistance to delamination growth is expressed in terms of the interlaminar DCB mode (mode I) fracture toughness, measured by strain energy release rate, GI, dissipated per unit area of delamination growth in composite. Three categories of samples were used: two unidirectional carbon fiber/epoxy resins with one degraded by heat and one with a commercially used resin. It was found that sample that was exposed at a temperature greater than the glass transition temperature Tg of the epoxy had a higher cumulative energy release rate than the two other samples types. The original type having the lowest release rate. Acoustic emission parameters have been found to be powerful indicators of the intensity of the damage. Multivariate analysis of up to 49 parameters was performed in order to group classes of AE signals with matching characteristics. A correlation was established between the energy release rate and the acoustic emission energy.

No abstract