Abstract

The ultrasonic backscattering method can be applied to determine the orientation of fibre composite materials whose achievable angle resolution depends on various parameters. This article describes the results of a study of those parameters. New models of the backscattering effect on fibre composite materials were developed and were compared with the measured results. The results showed that it is possible to detect the fibre orientation with better reliability. It was thus possible to use this method for an advanced defect analysis.
Table of contents
1. Introduction
2. The Measurement Principle
3. The Model
4. Comparison of theoretical and experimental results
5. Example of measurements on fibre composites
6. Summary

Introduction

Compared with standard materials the fibre composite materials provided higher stiffness together with low component weight. As a result, the new material is used in many fields of industry such as consumer, automotive and aeronautics. For safety reasons, it is always necessary to test composites using a nondestructive testing procedure [6].

When using fibre composites it is essential that the internal construction fulfill the specifications exactly. While standard NDT methods such as ultrasonic testing are good for detecting defects like voids, new methods are needed to determine the inner construction of these composites. Testing the orientation of the fibre is particularly important, since the E-module of the component depends on the angle between the tension and the fibre orientation [6].

The literature shows results which begin to determine the fibre orientation with the most important methods of Eddy Current [2], Microwave [3] and the ultrasonic backscattering method introduced here. The latter is highly interesting, since it can be applied to objects measuring a few millimeters in depth and no precondition of the material or structure is necessary. So it is not limited - as it is for eddy current - to application only to conductive materials. As another advantage the composite can be tested from only one side..

The applicability principle of the use of ultrasonic backscattering was proven in a previous work [4, 5]. This article introduced the results of a study in which the parameters of this method were investigated using models as well as the results of the applied measurements. The suitability of this method was confirmed by investigating many samples and the measured angle resolution matched well with the results of the model. Additionally, the ultrasonic backscattering method showed good results for detection of defects which were difficult to detect with other ultrasonic testing methods.

The Measurement Principle

The ultrasonic backscattering method is based on the impulse echo technique. However, in contrast to the impulse echo technique the signals from the object's back scattered sound wave are evaluated and not the direct reflected wave from the object. To prevent the direct reflection the sound beam is generated with an angle to the surface of the specimen. The determination of the fibre orientation is preceded by the intensity of the back scattered pulse which is plotted as a function of the probe's angle of rotation (Fig.1). Always when a fibre is oriented perpendicular to the probe, an intensity maximum of the back scattered signal is recorded. Based on this profile the condition of the fibre orientation can be determined. Averaging over a specific area was necessary since an intensity variation was also effected by defects ( it also reduced an effect of relatively small disorientation of some fibres within a fibre layer).

Fig 1: The measurement principle set-up

The Model

The investigation of the influence of specific parameters on the result is difficult since it is normally impossible to vary only one of them and hold the others constant. For instance, the sound beam is related to each probe and by changing the probe some parameters are changed at the same time (e.g., frequency, beam diameter). That's why an investigation that is based only on experimental work is not useful. In contrast, a model of scattering of ultrasound on fibre composites was developed, describing the sound field as single parameters and making possible the calculation of the back scattered relative signal amplitude. Since fibre composites consist of many coupled filaments, which are generators of back scattered signals, an exact description of the back scattering effect is very difficult. Thus, the following simplification was used for the model.

Fig 2: Spherical model of ultrasonic backscattering on a fibre

First it was decided that for determination of the angle resolution of the back scattered signal there is no need for an absolute intensity value. So just the relative profile of the intensity maximum is necessary (perpendicular position of probe to the fibre). Previous investigations showed that in many cases a model with its description of the fibre composite structure can be based on a single fibre. It is essential that on a single fibre a coordinate transformation on a system of coordinates can be done by a two dimensional description of the back scatter effect. The first axis of this system of coordinates is the beam axis. The second axis stands perpendicular to this and lies on the level of beam axis and fibre. The third axis stands perpendicular to the other axis. Since the third axis stands perpendicular to the fibre as well as to the beam axis, the back scattered intensity does not depend on this direction. The only variable is the angle between the beam axis and the fibre (precondition: a fixed angle of incidence ), that is called angle . Although this model is based on a strong simplification (considering the scattering conditions) an analytical solution for that problem can't be found. The reason is the sound beam, which has still not been given detailed consideration. The scattering theory is based on sound fields of plane waves of one frequency which have the same amplitude at each point. Such sound fields can be applied only when the reflector is smaller than the wavelength. Since this is not the situation of one fibre direction of the investigated case, we need to use the real sound field for the model. The solution was found by theoretically cutting the fibre into small elements on which the sound amplitude is approximately constant. The acoustic reaction of the fibre is calculated by the sum (considering the true phase) of each back scattered wave from the elements. Considering the fact that an element far from the beam axis is less excited than that one on the beam axis, all parts of the back scattered signal were calculated in relation to the applied sound beam profile. Each part of the elements were approximated as a spherical shape with the properties of the fibre (sound velocity and density), allowing a calculation by use of partial waves (Fig 2).

An additional method for calculation of all parts is the use of a long thin element as an approximation of the real fibre. Because both methods showed a similar relative intensity, it allowed the conclusion that the exact material property determines the absolute intensity profile of the back scattered signal. However, the relative intensity profile, especially near the maximum, is less affected by the exact value of the material property.

Comparison of theoretical and experimental results

By use of the model it was possible to calculate an exact intensity profile of the back scattered signal. The example of measurement results in Fig 3 show this clearly.

Fig 3: Comparison of theoretical and experimental results ( = 0° as standard) 10K

The measurements were obtained with ultrasonic equipment consisting of five axes. The probe could be controlled in x,y,z - axis as well as in two rotary axes of the probe. For each position of the probe's angle, as determined from generated C-scan pictures, the sound pressure (as a physical result) was calculated and averaged. Each averaged sound pressure related intensity value - hereafter called 'mean intensity' - (after it was set as standard = 0°) could be compared with the corresponding relative intensity value.

A simultaneous record of all existing fibre orientations with the component was maintained by using a wide time gate. In contrast, using a small time gate, it was possible to achieve the result of the time position for in-depth analysis of the internal structure of the component (Fig 7).

The model showed that the following significant parameters have a major effect on the intensity profile near the maximum.

• Sound beam characteristics at the scattering element means the width of the sound field of focused probes at the focus point.
• The probe's center frequency

It was already mentioned that a separation of both parameters by experiments is not possible since probes are manufactured with fixed specifications. Since a good verification of the scattering model was proven by measurements (Fig 3), the investigation proceeded using theoretical methods. The result of variation of the focus radius (Fig 4) and of the centre frequency (Fig 5) was calculated.

Fig 4: Variation of the nominal frequency (focus radius = 0.3 mm is approx. the focus radius of the 5 MHz probe (7K)

Fig 5: Variation of the sound beam diameter (nominal frequency = 2.25 MHz) (6K)

Figures 4 and 5 show clearly that an improvement of the angle resolution can be achieved either by a higher frequency or a wider sound beam at the fibre position.

Example of measurements on fibre composites

The applicability of the ultrasonic backscattering method was shown on different multilayer composites. A bi-directional (0°/90°) manufactured Prepreg is introduced as an example (Fig 6). The low thickness of approx. 2.5 mm made the use of a high frequency and non-focused 10MHz probe possible.

Fig 6: Measurement of the fibre orientation of a Prepreg (23K)

Figure 6 shows that the fibre orientations (0°/90°) could be determined with an accuracy of approx. 1° by the position of the intensity maximum.

Evaluation of deeper sections within the specimen needs a lower test frequency and has because of that the disadvantage of a lower angle resolution, since high attenuation would occur if a high frequency were used. This can be generally necessary for all other high absorption materials. The same method was tested on other Prepreg specimens as well on a carbon fibre texture (CFK) and on a thermoplastic object (Table 1).

Table 1: Result of the fibre orientation with ultrasonic backscattering
Sample	Material	Fibre orientation	Experimental result
1 (s. o.)	CFK-Prepreg	0°; 90° (-1 ± 1)°	(89 ± 1)°
2	CFK-Prepreg	0°; ± 45° (-1 ± 1)°	(46 ± 1)° ; (134 ± 1)°
3	CFK-texture	0°; 90° (-1 ± 3)°	(91 ± 3)°
4	CFK-Thermoplastic	unknown	(0 ± 1)° ;(90 ± 2)°

While the method is used to detect all fibre orientations, in-depth information is not possible. However depth characteristics using specific time gates were also investigated. (Fig 7).

Fig 7: In-depth determination of fibre orientation (11K)

The results of the Prepreg specimen show three different fibre orientations which were detected by using two time gates. The first gate records mainly a 0° orientation close to the surface, while the second gate is positioned in a deeper section with an orientation of 45° detect and shows a result of 135°. Since a 4 mm thick specimen was investigated, a focused 7.5 MHz was applied.

Besides the determination of the fibre orientation the method can be also used for detection of other defects in fibre composites. Especially for detection of small defects the backscattering method shows better results than the impulse echo technique.

Fig 8: Detection of air inclusions with ultrasonic backscattering (49k)

The first example shows the result for an artificially applied crack detected by use of the impulse echo and the backscattering method. It's almost impossible to detect this crack with the impulse echo technique (Fig. 8 left), while the backscattering method shows a clear result (Fig. 8 right). For achieving a minimum echo amplitude of the fibre, angle position was adjusted so that the back scattered signal showed a minimum amplitude. With the same method it is also possible to detect small air inclusions (Fig 9), which could not be detected with traditional testing methods.

Fig 9: Detection of a fibre crack with ultrasonic backscattering (41k)

Summary

This article introduced which parameters have an essential influence on the detection of the fibre orientation of composites by use of the ultrasonic backscattering method. It was shown that the sound beam characteristic (frequency and diameter) have a major influence on the determination of intensity profile, by means of the achievable angle resolution. It is possible to achieve a good angle resolution by using a high frequency and a large beam diameter. By use of specific parameters a detection of the fibre orientation at different depths is possible for different kind of materials. Furthermore, the method of angle testing can find defects in fibre composites which can not be found so easily using other ultrasonic test methods.

References

1. Krautkrämer, J.; Krautkrämer, H.: Werkstoffprüfung in Ultraschall. 5. Aufl. Springer Verlag, Berlin 1986.
2. Maaß, M.; Crostak, H.-A.: Einsatz von Drehfeld-Wirbelstromsensoren zur Bestimmung von Materialvorzugsrichtungen in Verbundwerkstoffen und von Texturen in Blechen. Vortrag, DGZfP-Jahrestagung 1995, Aachen. In: DGZfP Berichtsband 47.1, Berlin 1995, S. 127-142.
3. Diener, L.: Abbildung von Orientierungen in Polymerwerkstoffen mit Mikrowellen. Vortrag, 13. Stuttgarter Kunststoff-Kolloquium. In: Eyerer, F.; Fritz, H.-G.: Tagungsband 13. Stuttgarter Kunststoff-Kolloquium. Hütig-Verlag, Heidelberg, 1993, S. 199-206.
4. Bar-Cohen, Y.; Crane, R.L.: Nondestructive Evaluation of Fibre Reinforced Composites with Acoustic Backscattering Measurements. Materials Evaluation 40 (1982) 8, S. 970-975.
5. Pfeifer, T.; von Wachter, F.K.; Schuster, J.: Backscattering - ein neues Ultraschallprüfverfahren zur Analyse des inneren Aufbaus von Faserverbundbauteilen. Ingenieur-Werkstoffe 2 (1990) 9, S. 46-49.
6. Pfeifer, T.; Rümenapp, S.; Feldhoff, J.: Ultraschall zur Bestimmung von Faserorientierungen in Verbundkunststoffen. Kunststoffberater 7/8 (1996) S. 42 - 46
7. W. Michaeli, W.; Wegener, M. : Einführung in die Technologie der Faserverbundwerkstoffe. Carl Hanser Verlag München Wien (1990)

Authors

* Laboratory for machine tools and production technology (WZL).
Chair of Metrology and Quality Management.
Homepage: http://www-mtq.wzl.rwth-aachen.de (in German)