No more holes…

With advances in ultrasound, non-destructive testing of composite structure is becoming increasingly accessible. Stefano Beltrando of QI Composites in Turin explains the progress his company have made pioneering this field.
Since ‘fibreglass’, as it was originally known first appeared being able to nondestructively diagnose anomalies and defects in composite structures has been the dream of every boatbuilder — and composite engineer. Come 1998 and it was exactly this challenge that pushed me to undertake the research necessary to develop a satisfactory method tu meet what remained a compelling need.
The method I was seeking had to fit a broad variety of situations, from boats under construction to finished ones, in the water as well as in a yard. It had to work on every kind of component, from masts to rudders to hulls. And it had to overcome problems arising from the thickness of the gel coat of the fillers and of the most complex lamination geometry.
My own relatively short experience in the aerospace industry played a fundamental role in leading me to the potential of ultrasonic techniques. In the ‘pulse echo’ method, as used for ultrasound tests on humans, a transducer placed un the sample surface emits ultrasound waves, which are reflected at the interfaces between different materials and layers. The echoes are in turn registered by the transducer and the resulting pattern can be interpreted to unravel the material condition.ULTRASOUND
Definition: A sound generated above the human hearing range (about 20KHz) is defined as ultrasonic. The frequency range normally used in ultrasonic non-destructive testing is 100KHz to 50MHz. Although ultrasound behaves in a similar manner to audible sound, it has a shorter wavelength. To have reflection and refraction phenomena, beam wavelength dimension has to be similar to the dimensions of the obstacle (ie discontinuities, defects, etc). This means ultrasound can be reflected off very small surfaces such as defects inside materials. This property makes ultrasound ideal for non-destructive testing of materials.
As with any wave-like perturbation, As with any wave-like perturbation, ultrasounds are characterised by wavelength, frequency, period and propagation velocity. They are a mechanical wave, so they need an elastic medium in which to be propagated (solid, Liquid or gas).Wave types
Non-destructive methods use both longitudinal and transverse waves. Ultrasound runs over particles Laid on planes perpendicular to the propagatiun direction. Ultrasound propagation causes compression and rarefaction areas within the medium.· Propagation velocity (V)
This represenrs uLtrasound wave displacement occurring within a time unit — it is measured in m/s.
The velocity of ultrasound in a perfectly elastic material, at fixed temperature and pressure, is constant; it depends on material and wave type.

· Acoustic impedance (Z)
Acoustic impedance characterises ultrasound wave propagation and depends upon material type. It is calculated as the product of material sound velocity and material density p (kg/(m2* s)) such that:
z =p v

This represents resistance offered by the medium to sound propagation. Among the consequences of this phenomenon are thedifferent wave propagation velocities in different materials and varying signal attenuation as ultrasound progresses through a medium. The higher the acoustic impedance, the less the ultrasound wave is hindered and attenuated.

· Reflection and transmission

When an ultrasound beam strikes perpendicularly on an interface between two mediums, A and B, with different acoustic impedance, Z, part of the beam is reflected and the other part is transmitted.
A relative measure of reflected and transmitted beams is given by two coefficients, dependent upon the impedance Za and Zb of the two mediums.
Table A shows the reference ultrasound speed in different materials. Remember that composite materials, whether of sandwich or monolithic structures, are composed of different materials characterised by different ultrasound velocities. Thus ultrasound velocity in a composite structure depends not only on material types, but also on their volume fraction.

Table A

MATERIAL	LONGITUDINAL ULTRASOUND WAVE SPEED (M/S)
Epoxy Resin	2,200
Epoxy+glass monolithic	2,200-2,600
Epoxy+carbon monolithic	2,600-2,900
Sandwich with Nomex core	2,100-2,300
Sandwich with balsa core	2,900-3,000

Unfortunately, I discovered early in m work that specific issues of the nautical world hampered the straightforward application of the ‘conventional’ pulse echo method as it is normallv used in the aerospace industry. The main problem arises from the fact that the overwhelming majority of the laminates used are handmade and hand-impregnated. This results in an inhomogeneous resin distribution as well as in a great variation of the filler, gel coat and bond thickness. Further, dealing with hybrid materials (eg glass-Kevlar and glass-carbon) was a key difficulty from the start of our work since we were primarily concerned with raceboat structures.

To overcome these problems, an ultrasound characterisation of all the aforementioned composite materials, as well as the whole range of their combinations, was undertaken. This preliminary step was followed by the characterisation of all the defects usually affecting these material compositions.
As a result of this detailed research we discovered that every unaltered interface between different materials is characterised bv a tvpical ultrasound spectrum. When the signature typical of a certain interface is unreadable or altered, the presence of an anomaly in the structure can be inferred.
The step that led to the complete understanding of the ultrasound spectrum has been critical, as it now allows us to extract a wealth of information from virtually every sample.
The control or reference samples that we analysed to characterise the defects were created by us for that specific purpose or collected from a variety of production boatyards or taken directly from existing structures. After the ultrasound analysis the samples suffered destruction or chemical attack, to increase the extent of the detected defects until they became visible, thus providing a check on the predicted weaknesses. This experimental work provided us with an absolutely original set of observations which were used to unravel the mechanisms behind defect nucleation and growth. lnterestingly, some of our results clearly contradicted commonly held views on composite behaviour based solely upon a superficial approach to the problem without any understanding of the real dynamics involved within it.
After this experimental part our efforts were drawn towards two main aspects:
1. How to name defects
2. How to evaluate the relative importance of a defect on the composite material life

1. The need for an unambiguous nomenclature arose from the confirmation that different kinds of delamination result from different causes and have in turn different effects on material properties. Up to that point, disregarding our observations, no distinction was made among the different delamination types. Our efforts allowed us to distinguish two end-members among the great variety of typical delaminations, which we named ‘glossy’ and ‘ragged’. In a glossy delamination the two adjacent layers are never in contact; the resin is distinctively polymerised on the two layers and as a result it is as glossy as paint (it has a glossy look). A ragged delamination results from the separation of two originally adjacent layers, causing the interposed resin to fracture. As a result of this process the surface looks ragged.
It can be observed that a glossy delamination is a ‘primary’ defect, meaning that it results from a construction mistake, while the ragged delamination is a ‘secondary’ defect, induced by use. Composites can also be affected by a whole range of defects whose characteristics range between the two aforementioned end-members. We named this intermediate condition as ‘diffuse and concentrated porosity’ (since it resembles the bonding between a layer of sandpaper and a perfectly polished and flat surface). A simplified picture of this defect is provided by rare glued spots surrounded by numerous voids. Only a limited effort is required to delaminate the two layers. As a result of the delamination, a composite pattern made of glossy and ragged spots is observed, thus testifying to its intermediate nature between glossy and ragged delamination.

2. The evaluation of the importance of a defect is the main focus of current research.
When I say to a skipper who is supposed to sail solo across the Atlantic Ocean at 3Okt, ‘There is a 30 cm2 delamination along the waterline’, does it mean that he is doomed or not? Put another way, can the limited presence of delamination in the hull be regarded as a ‘normal’ and trivial feature or does it represent a first step in the degenerative process of the structure?

The answer in this crucial question requires a net set of investigations, aiming in find an answer in the following questions:

1. Where is the delamination located and what is its extent?
2. What kind of delamination are we dealing with?
3. Does it result from a construction mistake or did it nucleate afterwards

4. Does it affect vital parts of the structure?
5. Was it generated by sailing, transporting or assembly-related strain?
6. Is it linked to the construction technique and to the materials used? (if ‘yes’, then the whole structure is likely to be biased)
7. Can it be repaired quicklv without causing even more damage (especiallv in the long term)?
8. How much will the repair cost and how long will it take?
9. Have we dealt with similar problems in the past?
10. What kind of navigation is the boat being prepared for?
Once an answer to these 10 questions is found we know whether we (and the solo skipper) have to be worried or not (or to what extent our, and the skipper’s, concerns are justified). Most of the questions can be answered with the help of the skipper and boatbuilder, but for some of the answers a check with the designer or the structural engineer is required. As a result of this process the knowhow of a team composed of ourselves, the designer, the skipper and the constructor steadily grows.

QI Composites were created in 1999 to provide research and consultancy to boatyards, filling a clearly defined knowledge gap. Thanks to the development of a reliable non-destructive method of analysis based on ultrasound we can now perform careful evaluations of production quality and methodologies. The formal standardisation of this procedure, the current focus of our work, will ultimately lead to the construction of ‘provisional’ models which will allow us to predict the evolution of composite structures on the basis of defects observed over their life.
Among its raceboat clients QI Composites enjoyed a major collaboration with Team Prada. Current commercial clients include Nautor, Baltic and Wally Yachts, and also the Ferretti Group.