NDT of CFRP sandwich structures

Marek Listl, September 2015

Artikel auf Deutsch

The non-destructive testing of CFRP sandwich structures has particular value in aerospace as high demands on weight and stiffness conflict with safety regulations compliance. For this, there are several methods to choose from, however, there is a distinction between testing directly after production and the testing during application (“in field“).

Motivation

Since the 1930s, the increasing use of carbon fibre reinforced plastic (CFRP) has led to special requirements in non-destructive testing. Above all, sandwich structures are a particular challenge due to the material mix. However, in the field of aerospace safety regulations make testing during the production process and the operational phase essential.

Fundamentals of sandwich structures

The demands of light-weight construction require that panels and shell elements are as thin-walled as possible. As a result of their low local stiffness, they are however considered “stability-compromised”. They can crumple and bulges. This can be counteracted through the corresponding stiffenings (Fig. Stiffened fuselage of an Airbus A340. A further step involves the sandwich panel and sandwich shell. Here, two panel halves are separated by a core at an exact distance. ^[1] Common types of cores are foams and combs of aluminium, plastics, CFRP or paper. ^[2] Generally, such panels consist of: ^[2]

• Face sheets: Thin and stiff layers (in this case of carbon fibre reinforced plastics (CFRP)) for the transport of axial loads, flexure and thrust.
• Core: As light a core as possible to support the face sheets, for load transport between both sheets as well as for the load suspension of thrust perpendicular to the panel plane.
• Interface: Mostly an adhesive layer that ensures the connection and thus the power transmission between face sheets and core.

This structure has an especially good stiffness to weight ratio. In comparison to a panel of 2mm thickness, a sandwich shell with a face sheet thickness of 1 mm and a core height of 20 mm, has an areal moment of inertia about 150 times higher (P. 87). ^[1]

Failure mechanisms and types of failure

To determine the potential of the testing methods, the possible types of damage have to be known. However, it is important to distinguish between damage related only to the face sheets and that concerning the core i.e. the interaction between core and face sheet. For face sheet damage, please see the article about testing of CFRP components. Possible testing methods include thermography as well as Ultrasound and Eddy current test. The types of failure that can be detected with the testing methods introduced here are: ^[3][4]

• Bulges between the honeycomb struts
• Symmetrical and asymmetrical crumpling
• Crumpling of the honeycomb core
• Fracture of the honeycomb core
• Detachment of the face-sheet

In addition to these types of failure, different failures can occur during production: As with all the types of failures, these particular errors in the monolithic face sheets are neglected. The following failures can occur in the region of the core: ^[4]

• Fracture orthogonal to the surface
• Deformation
• Positional errors

Inhomogeneities of the connection between core and face sheet are: ^[4]

• Detachment
• Weak bond
• Excess adhesive
• Adhesive deficit
• Impurities

The actual influence of a defect is very dependent on the component. Thus, a general division and classification of the criticality is more difficult. It can however be said, that those defects that have one of the previously mentioned types of failures, directly or indirectly, are considered critical. ^[4]

Method principles

Generally, there are three types of waves available for the different testing methods. These are electromagnetic, thermal and elastic waves. It is crucial how strong the resistance to wave propagation the material is, how easily the wave can penetrate the component and how strongly the surfaces lead to reflection, absorption and transmission. Propagation resistance can be avoided to a certain extent by higher performance. It is therefore a little selective. Of greater importance is the behaviour at the interfaces of the three-material CFRP face sheets, adhesive and core as well as possible flaws such as air and impurities (e.g. Foil residues). Of decisive influence on the propagation is the electromagnetic impedance $//<![CDATA[ \begin{array}{l}\underline{Z}_W\end{array} //]]>$:

$//<![CDATA[ \begin{array}{l}\underline{Z}_W = \sqrt{\frac{j \omega \underline{\mu}}{\sigma + j \omega \underline{\epsilon}}}\end{array} //]]>$

• $//<![CDATA[ \begin{array}{l}\omega\end{array} //]]>$: angular frequency of the wave oscillation

• $//<![CDATA[ \begin{array}{l}\underline{\mu}\end{array} //]]>$: complex permeability

• $//<![CDATA[ \begin{array}{l}\sigma\end{array} //]]>$: electrical conductivity

• $//<![CDATA[ \begin{array}{l}\underline{\epsilon}\end{array} //]]>$: complex permittivity

acoustic impedance: $//<![CDATA[ \begin{array}{l}\qquad Z = \rho * c\end{array} //]]>$

• $//<![CDATA[ \begin{array}{l}\rho\end{array} //]]>$: density

• $//<![CDATA[ \begin{array}{l}c\end{array} //]]>$: sound velocity

as well as the thermal effusivity:  $//<![CDATA[ \begin{array}{l}\qquad e = \sqrt{\lambda \rho C}\end{array} //]]>$

• $//<![CDATA[ \begin{array}{l}\lambda\end{array} //]]>$: thermal conductivity

• $//<![CDATA[ \begin{array}{l}\rho\end{array} //]]>$: density

• $//<![CDATA[ \begin{array}{l}C\end{array} //]]>$: specific heat capacity

Large discrepancies of these parameters for sandwich components and flaws make the penetration of the wave through the component more difficult. A good detectability is the result. ^[4]

An additional classification of the measurement method results not only from the types of waves used but also from the one-sided or two-sided measurements as well as by the necessary connection. Two-sided measurements such as [[ultrasound transmission], involve additional effort due to the precise control of the opposing measurement heads. Furthermore, in the case of an “in-field” inspection, a testing from both sides can sometimes not be usefully applied due to the demounting effort. If coupling via air not possible, a coupling with water or another couplant has to be used. Complex infrastructures and extravagant preliminary work, such as sealing against water penetration, are the consequences. ^[4] An optimal test method is therefore one-sided and realisable with air coupling.

Testing procedure

The table below shows specific measurement methods for testing CFRP sandwich structures and their physical principles. When describing individual methods it is important to note that the possibilities differ from component to component. Most of the measurements described here were carried out on a plane sample and under ideal conditions. In reality, the result can differ in detail.

Computer tomography

Computer tomography (CT) makes high resolution 2D or 3D analysis of samples possible. A significant difference to medical CT is the technical system of a flexible test set-up and it can be operated with significantly higher radiation dosages and exposure times due to not having to consider patient safety. ^[5] An additional advantage is the potential conversion of measurement data into CAD data. This allows for a FEM analysis of a real component with defects. ^[6] CT is one of the most reliable and precise testing methods. Disadvantages are limitations on component size as well as an obligatory two-sided measurement. In addition, the required systems are very cost-intensive to purchase. An unsolved problem is the calibration of CT systems to determine the degradation of the detector.

Thermography

The detailed article Infrarot-Thermographie an Kohlefaserverbundwerkstoffen describes the different variants of thermography. Advantages are the efficient and cost-effective application of measurement. Specialized programmes can automatically evaluate the results. Furthermore, a one-sided measurement is also feasible. However, attention has to be paid to the thermal influence of the environment as well as air circulation and reflection.

In the case of testing sandwich structures, delaminations, weak bonds and honeycomb faults are easy detectable. Difficulties arise from impurities and cuts perpendicular to the component’s surface. With a one-sided measurement, reverse side damage is only detectable with low reliability. Stimulation with ultrasound does not make sense. ^[4]

Ultrasound

Various articles describe the mode of operation in detail:

• Ultrasound
• Air-coupled ultrasound: operating principle and applications
• Ultraschall-Durchschallung

The well-established and advanced pulse-echo method provides good detectability in face-sheets and the front-sided core connection. If a paper honeycomb is used as a core, failures are almost undetectable. Reverse-side defects remain mostly undiscovered. ^[4]

In the case of transmission via air-coupled ultrasound, no differences between front and back can be noted. Both sides can be well tested. Also, defects in the core are well detectable, as long as they are not perpendicular to the surface. These only appear as advantages however, as measurements from both sides are very complicated and expensive. The coupling with air is made possible by very low stimulation frequencies. ^[4]

Narrowband ultrasonic spectroscopy

Narrowband ultrasonic spectroscopy (NBUS) measures mechanical impedances of the specimen. A well-known example of a test device is the: Bondmaster from Olympus. A coupling via a couplant can also be useful here. Yet the results of an examination are very extensive. Only foil impurities on the reverse face sheet as well as vertical cracks in paper honeycombs are difficult to detect. ^{[4] [7]}

Lamb waves analysis

Lamb waves are a combination of pressure and shear waves. They are very attractive for NDT due to their high propagation ability and the associated large test area. Should they be tested in a wide frequency range, a coupling with a couplant is also necessary here. ^[8] In the case of narrow specimens or nearby walls, good wave propagation leads to interfering reflections. Testing is typically conducted one-sided with a piezo oscillator. Defects can be detected well on both sides of the sandwich. Only weak bonds, foil impurities and perpendicular cuts in a honeycomb core cannot be detected. ^{[4] [8]}

Shearography

In shearography, deformations based on thermal or mechanical stimulations are measured. Of particular hindrance for the measurement is the necessity for high surface quality. The detectability in this testing method is very limited. Flaws can only be detected in the front face sheet and in the interconnecting layer. ^[4]

Linked articles

• Infrarot-Thermographie an Kohlefaserverbundwerkstoffen
• Ultrasound transmission technique
• Shearography

Literature

1. Skriptum zur Vorlesung "Leichtbau" vom SS 2014. Lehrstuhl für Leichtbau. 2012.
2. Skriptum zur Vorlesung "Luft- und Raumfahrtstrukturen" vom SS 2015. Lehrstuhl für Leichtbau. 2012.
3. Handbuch Struktur Berechnung 55111-01. LTH.
4. Mosch, M.: Untersuchung und Potentialanalyse verschiedener ZfP-Prüfmethoden für Eignung zur einseitigen Prüfung von CFK-Wabenstrukturen. 2012.
5. FAQ - Häufige Fragen zur 3D Computertomographie. 3D Padelt GmbH. Strausberg, 2015.
6. Industrielle Computertomographie. Wikipedia. 25.8.2015.
7. Narrowband Ultrasonic Spectroscopy for Inspecting Multilayered Aerospace Structures. NDT.net. 2006.
8. Hillger, W.: Lamb-Wellen zur Schadensanzeige in faserverstärkten Kunststoffen . DLR Institut für Faserverbundleichtbau und Adaptronik. DGZfP-Jahrestagung 2005. Braunschweig, 2005.