User login


You are here

Discussion of fracture paper #24 - The sound of crack growth

ESIS's picture

Carbon fibre reinforced polymers combines desired features from different worlds. The fibres are stiff and hard, while the polymers are the opposite, weak, soft and with irrelevant fracture toughness. Irrelevant considering the small in-plane deformation that the fibres can handle before they break. It is not totally surprising that one can make composites that display the best properties from each material. Perhaps less obvious or even surprising is that materials and composition can be designed to make the composite properties go beyond what the constituent materials are even near. A well-known example is the ordinary aluminium foil for household use that is laminated with a polymer film with similar thickness. The laminate gets a toughness that is several times that of the aluminium foil even though the over all strains are so small that the polymer hardly can carry any significant load. 

In search of something recent on laminate composites, I came across a very interesting paper on material and fracture mechanical testing of carbon fibre laminates::

"Innovative mechanical characterization of CFRP using acoustic emission technology" by Claudia Barile published in Engineering Fracture Mechanics Vol. 210 (2019) pp. 414–421

What caught my eye first was that the paper got citations already during the in press period. It was not less interesting when I found that the paper describes how acoustic emissions can detect damage and initiation of crack growth. The author, Barile, cleverly uses the wavelet transform to analyse the response to acoustic emission. In a couple of likewise recent publications she has examined the ability of the method. There Barile et al. simulate the testing for varying material parameters and analyse the simulated acoustic response using wavelet transformation. This allow them to explore the dependencies of the properties of the involved materials. 

They convincingly show that it is possible to both detect damage and damage mechanisms. In addition, a feature of the wavelet transform as opposed to its Fourier counterpart is the advantages at analyses of transients. By using the transform they were able to single out the initiation of crack growth. Very useful indeed. I get the feeling that their method may show even more benefits.

A detail that is unclear to me, if I should be fussy, is that there are more unstable phenomena than just crack growth that can appear as the load increases. Also regions of damage and in particular, fracture process regions may grow. When the stress intensity factor K alone is sufficient there is no need to consider neither size nor growth of the fracture process region. The need arises when K, J, or any other one-parameter description is insufficient, e.g. in situations when the physical size of the process region becomes important. Typical examples are when cracks cross bi-material interfaces or when they are small relative to the size of the process region. When the size seems to be the second most important feature, then the primary parameter may be complemented with a finite size model of the process region to get things right. There is a special twist of this in connection with process region size and rapid growth. In the mid 1980's cohesive zones came in use to model fracture process regions in FEM analyses of elastic and elastic-plastic materials. Generally, during increasing load, cohesive zones appear at crack tips and develop until the crack begins to grow. One thing that at first glance was surprising, at least to some of us, was that for small cracks the process region first grows stably and shifts to be fast and uncontrollable, while the crack tip remains stationary. Later, of course the criterion for crack growth becomes fulfilled and crack growth follows.

Is it possible to differentiate between the signals from a suddenly fast growing damage region or fracture process region vis à vis a fast growing crack?

It would be interesting to hear from the authors or anyone else who would like to discuss or provide a comment or a thought, regarding the paper, the method, or anything related.

Per Ståhle



Dear Prof. Ståhle,

Thank you for your appreciation and interest in my paper. The analysis of the signal-based data connecting to the acoustic events arising during the damage progression seems to be a powerful way in assessing the integrity of the material.

Your concern regarding the importance of stress intensity factor and/or energy release rate is actual and plays a crucial role in the damage progression. To this scope, me and my colleagues Casavola and Pappalettera, had done subsequent research works. In one of these studies the Cohesive Zone Modelling (CZM) was analysed in the Finite Element approach to study and describe the behaviour of the CFRPs subjected to the same Mode I load (

In the meantime, the Acoustic Emission data were deeply investigated even further in understanding the damage progression in CFRPs subjected to more than one damage mode. At this purpose, the Wavelet Packet Transform (WPT) proved to be a powerful even if delicate tool in distinguishing the different damage modes. Different frequency bands of different damage modes are accounted in a CFRP using this approach (

At this moment, we are also attempting to explore other ways of identifying the damage modes in a more accurate way, by analysing the acoustic emission data according to different approaches.

Once again, I thank you for your interest in our research works.

Claudia Barile

ESIS's picture

Dear Claudia,

Your follow up experiments and simulations are very interesting. When you say that you use cohesive zone models to simulate the fracture processes it strikes me that how smart that must be. I guess that you by that can figure out the length of the crack while the short crack are expected to develop long cohesive zones vis à vis cohesive zones at long cracks. Also the major part of the cohesive zone would host fracture processes that endure high stresses while the fracture processes at crack tips of longer cracks are exposed stresses that are more evenly distributed between zero close to the crack tip and a maximum stress at the tip of the cohesive zone. 

I guess that if the acoustic spectrum of the high stress initial fracture processes and the low stress fracture processes are different you would be inform of many things. Perhaps state of the process region, crack length and length of the process region. At least theoretically.

I don't know what the capabilities are but at least to me it seems like the sky is the limit.

Per Ståhle

PS Sorry about the delay with this reply. DS

Subscribe to Comments for "Discussion of fracture paper #24 - The sound of crack growth"

Recent comments

More comments


Subscribe to Syndicate