Paper Ref: S1801_P0239 3 rd International Conference on Integrity, Reliability and Failure, Porto/Portugal, 20-24 July 2009 SHM of CFRP-structures with impedance spectroscopy and Lamb waves Jürgen Pohl and Gerhard Mook Otto-von-Guericke-University /Institute of Materials and Joining Technology (IWF) Magdeburg, Germany E-mail: juergen.pohl@ovgu.de ABSTRACT Impedance spectroscopy and application of Lamb waves are attractive methods for permanent monitoring of integrity in Structural Health Monitoring (SHM). Investigations of CFRPstructures (Carbon Fibre Reinforced Polymers) with embedded or attached piezoceramic elements are presented for both methods. Examples regard impact damage detection as well as estimation of influencing factors (e.g. degradation of contacts or piezoceramic) and important characteristics (e.g. generation, velocity and attenuation of Lamb waves). INTRODUCTION Structural Health Monitoring (SHM) offers new possibilities for damage prevention of technical structures. SHM implies permanent monitoring of structural integrity as well as damage detection by inherent means [Giurgiutiu 2008, Park 2006]. Disadvantages of conventional maintenance concepts (time based with fixed inspection intervals) are overcome by introducing more and more condition based maintenance. Direct damage detection replaces assumptions for damage initiation and progress thus making damage tolerance concepts more effective. Monitoring and detection demand implementation of sensorial (and actuatorial) functions into the structure. For components made of CFRP (Carbon Fibre Reinforced Polymers) the use of embedded or attached piezoceramic elements is an attractive way for designing smart SHMstructures. Different approaches for SHM- purposes with piezoceramics in CFRP-structures are possible: Impedance spectroscopy is based on the coupling of the structures mechanical vibration behaviour with measurable electrical properties of the piezoceramics [Park 2006, Giurgiutiu 2008]. Another way is the evaluation of ultrasonic Lamb waves [Su 2006]. The piecoceramic acts in a passive variant as a receiver for Lamb wave signals, generated by damaging events such as impact damage in fibre reinforced materials. Actuatorial piezoceramics produce Lamb waves that are picked up by distributed sensors in an active alternative, so that distinct subject regions may be observed selectively. In both cases, detailed understanding of the interactions of Lamb waves with the structure and defects is necessary. In the following, examples of investigations for the use of impedance spectroscopy and Lamb waves as tools for SHM in CFRP-structures are presented. -1-
IMPEDANCE SPECTROSCOPY The combination of various elements (resistors, capacitors and inductors) in an electric circuit results in a characteristic corresponding frequency response of magnitude and phase of impedance. The measurement of impedance over a wide frequency range in impedance spectroscopy is a well-established method for the investigation of electrical properties in electrochemistry, characterising material and interface properties [Macdonald 1987]. In case of the regarded smart CFRP-structures, the piezoceramic is a lossy capacitor (capacitance C P ) with decreasing impedance for increasing frequency (ω). Including the other components of the structure the whole circuit may be described by an equivalent circuit with resistors built by the contact resistivities, the resistances of the in- and output conductors (summarized to R R ), and the resistance due to the lossy piezoceramic (determined by R P and C P ). According to Kirchhoffs law, the resulting impedance Z of the smart CFRP-structure is: Z = R R R + 2 ϖ C P 2 2 PRP 2 ϖ CPRP j 2 2 2 + 1 ϖ C R + 1 P P (1). Figure 1 shows the principal arrangement for electrical impedance spectroscopy in our investigations. Due to measuring constraints the admittance (reciprocal impedance) was recorded with a network analyser and is depicted as an absolute value Y=Z 0 /Z. The graph at the left side shows the general progress of the admittance with frequency that is controlled by the piezoceramic as a capacitance element. So at low frequencies it is dominated by the piezoceramic, but at higher ones by the frequency-independent resistors R R. So, general shape and position of the impedance curve reflect changes in electrical components directly. Examples for this are ageing effects of contacts, damage of the piezoceramic or loss of coupling. Fig. 1: Principal arrangement for impedance spectroscopy and spectrogram Additional resonance peaks become visible, explained by coupling of electrical and mechanical properties of the structure. An electric stimulation of the piezoceramic patch of a smart CFRP-structure causes mechanical deformations of the piezoceramic, resulting in mechanical oscillations of the whole structure. Depending on the strength of excitation, size of the structure and point of view the result may be defined as eigenvalues of structural vibrations or (standing) elasto-mechanical waves. Impedance has to be interpreted as coupled phenomena. -2-
Changes in the mechanical properties (stiffness, damping) influence the transfer function of electrical to mechanical energy. So, damages influencing the structures oscillation behaviour, will become detectable by sensing electromechanical impedance. In summary, impedance spectroscopy offers potential for SHM-purposes as self-monitoring of the piezopatches [Park 2009] or detection of structural changes. Limitations for defect detection are given by the distance between defect and piezoceramic and the size of the examined structure. Impedance spectroscopy seems best suited for defect detection in relatively small dimensioned subjects, its size small at least in one direction. Figure 2 gives an example for defect detection, regarding a CFRP-strip with a stacking sequence of [0,-45, +45,90 ] 2 examined in a multiple impact test. Three circular piezoceramic patches with 10 mm diameter are surface-bonded (attached) in this case. The impedance curve exhibits a high number of resonance peaks due to different eigenmodes. After two impacts, causing delaminations respectively, the impedance curve shifted and position, magnitude and shape of resonance peaks changed, so indicating progression of damage. Figure 2: Impedance curves measured by piezoceramic 1 before and after two impacts (impact positions: x) Impedance spectroscopy offers opportunities for sensing inner degradation effects and environmental influences. The following shows consequences of outer load changes. A four point bending test revealed amplitude and shape variation of resonance peaks with changing load. The structure stiffens slightly under load and the impedance at the peak increases with a shift to higher frequencies, consequently. So, the influence of loading can be clearly distinguished from the effects of damaging, because degraded stiffness causes a shift toward minor frequencies. Figure 3: Influence of load on resonance-peak in four-point-bending test -3-
A continuous monitoring requests constant and stabile measuring conditions. Impedance spectroscopy can detect unintended changes of the system. Figure 4 gives an example for ageing effects of contacts. After a period of three months, degradation of contacts occurred, manifesting in a shift of the whole high-frequency part of the magnitude curve to lower values. Renewing of contacts showed similar effects. Such results have relevance for the planning of reference measurements, because harmless changes in electrical contacts shall not be confounded with structural ageing effects or damages. For this reason, reference measurements should be repeated in a scheduled manner or after contact renewals. Figure 4: Effect of ageing contacts: a) reference measurement b) measurement after three months Degradation of the piezoceramic is a worse situation for a health monitoring system. The next figure shows this for an impact of the piezoceramic causing fracture and delamination. These damages originate a drastic change in the impedance curve, shown in figure 5. A general change of the magnitude over a wide frequency range can be recognised, interpretable as degradation of the contact resistances with the impact by damaging of the electrical contacts. But in addition, remarkable changes in the resonance peaks become visible. These differences in resonance peaks mark the piezoceramics damage. Figure 5: Degradation of the piezoceramic - impedance curves before a) and after impact b) at piezoceramic LAMB WAVES Application of Lamb waves in an active or passive manner demands knowledge of the acting of the piezopatch as transmitter and receiver of the waves, velocity and attenuation behaviour of distributing Lamb modes and their interactions with defects [Su 2006, Diligent 2002, Giurgiutiu 2008, Köhler 2006, Cawley 2003]. Unfortunately, at least two modes exist at every frequency. Figure 6 displays the mode-shapes of the fundamental symmetric s 0 and antisymmetric mode a 0. -4-
Even in the case of quasi-isotropic CFRP-material anisotropy effects occur due to the layered design with strong directional dependence of properties in one layer. Additionally, Lamb modes are dispersive thus introducing extra attenuation effects for transient signals. The schematic dependence of mode velocities c from frequency and plate thickness is shown in the dispersion-diagram of figure 6. Figure 6: Symmetric and antisymmetric Lamb mode [Su 2006] and dispersion-diagram Excitation of Lamb waves strongly depends on geometry, position, size, composition and coupling of the pioezomaterial. The following results recorded with scanning laser vibrometer and ultrasonic testing devices show some effects of different form and position of the piezopatch on the generation of Lamb modes. Figure 7 gives an example of Lamb wave-signals, generated by an attached piezoceramic PZT-patch at a CFRP-stripe. The A-Scan (amplitude vs. time) demonstrates the complexity of these signals which consist of the two fundamental Lamb wave modes (symmetric mode s 0 and antisymmetric mode a 0 ). The doubled appearance of the signals becomes explained by reflection at the edge of the CFRP-structure. The B-scan (propagation coordinate vs. time) shows the spatial progress of the waves, making the different velocities of both modes clearly visible and revealing the edge of the piezoceramic as source. Fig. 7: Lamb waves generated by attached piezopatch: arrangement, A-scan, B-scan A different situation occurs for a central embedded (integrated) piezoceramic patch, presented in figure 8. In this case mainly the symmetric mode s 0 is launched and the weaker a 0 -signals are hidden in the mass of direct and reflected s 0 -signals. Even in the A-scan, the signals from the front (1) and trailing edge (2) of the piezoceramic become clearly resolvable. The B-scan admits the identification of reflected signals (Rs 0 ) from the edge of the CFRP-structure unambiguously. -5-
Fig. 8: Lamb waves generated by an integrated piezopatch: arrangement, A-scan, B-scan The geometry and the form of the eigenmodes of the piezoceramic patch play an essential role for the design of the resulting sound field. A circular source with symmetric eigenmodes obviously generates a circular field but a rectangular one transmits a directed pattern, governed by the distribution of the primary vibrations of the piezoplate. Figure 9 illustrates this behaviour showing the preferred emission perpendicular to the edges with strongly decreased amplitudes in the diagonal directions for the rectangular source. Fig. 9: Sound field of a circular and a rectangular piezopatch Knowledge of directional distribution of velocity and attenuation of Lamb waves is an essential feature for generating a sensor-actuator network for SHM-purposes. Material composition, stacking sequence, geometry and manufacturing of the inspected structure influence these properties. Figure 10 shows an example of a CFRP-structure, anisotropic in material (made of 0 /90 carbon fibre fabric) and geometry (backside stiffening by stringers). The measurement was performed with an attached circular transmitter (T) in a receiver distance of 150 mm. Strong differences for s 0 -mode-velocity in 0 /90 and 45 -directions occur as expected. A detailed analysis showed a 25%-increase of the velocity in the stringer regions, explained by the material properties of the stringers. -6-
Fig. 10: Velocity-distribution of s 0 -mode in an anisotropic CFRP-structure In the case of a more homogenous material, as shown in figure 11 by the example of an multidirectional quasi-isotropic structure with [0, +45, -45, 90 ] s stacking sequence, the velocity fluctuations grow less. This variation is explained by the distribution of the Lamb wave energy over depth with a preference of the layers with energy concentration. So, the slight higher value in the 90 -direction becomes apparent. Fig. 11: Velocity-distribution of s 0 -mode in a quasi-isotropic CFRP-structure The spatial distribution of the attenuation of Lamb waves plays an important role for practical applications, too. Attenuation effects determine the maximal possible distances between actuators and sensors and by this the area that is monitored. The attenuation of a structure exhibits a similar behaviour as velocity, but as figure 12 presents, is much more influenced by material and geometry. The anisotropic structure possesses beside conventional attenuation (damping) mode conversion effects (leakage) at the stringers that result in additional attenuation. -7-
Fig. 12: Attenuation of s 0 -mode in an anisotropic (left) and a quasi-isotropic (right) CFRP-structure For SHM-purposes interactions of Lamb waves with defects as reflexion, transmission and mode conversion offer potential for defect detection and evaluation. Figure 13 gives an example for detection of impact damage in a quasi-isotropic strip specimen with a transmission arrangement. The impact resulted in a small delamination with a diameter of about 5 mm displayed by a slight amplitude degradation over the transmission path of 200mm between attached transmitter (T) and receiver (R). The difference signal before and after impact more clearly expresses these changes. Beside the direct transmission path (denoted by I) a more pronounced signal (II) becomes visible which is explained by a preferential way along the +/- 45 -fibre-directions of the structure with multiple reflexions at the edges. This signal gives an additional indicator of defect presence. Fig. 13: Detection of impact damage: arrangement and transmission pathes, signals before (1) and after impact (2) (left), difference signal (right) Figure 14 demonstrates the interaction of the wave field with a small impact-induced delamination of 6 mm size. The embedded piezoceramic produces the symmetric s 0 -mode, which is reflected at the delamination with mode conversion to the antisymmetric mode a 0. At the marked position of the delamination the nearly circular radiation of the a 0 -mode starts. -8-
Fig. 14: Detection of impact damage by mode conversion CONCLUSIONS Impedance spectroscopy is attractive for self monitoring of piezoceramic sensors or actuators to detect changes in coupling, connections or environmental factors. A detection of defects in the surrounding areas is possible, the action range strongly depending on geometry and material properties. Lamb waves offer a high potential for monitoring large areas, but further investigations regarding interaction with defects and solving the inverse problem of damage prediction from Lamb wave data are necessary. REFERENCES Giurgiutiu, V.: Structural Health Monitoring with Piezoelectric Wafer Active Sensors, Academic Press, (2008) ISBN 9780120887606 Park, G., Sohn, H., Farrar, C.R. and Inman, D.J.: Overview of piezoelectric impedance based health monitoring and path forward. Shock and Vibration Digest, 35(6) (2003). 451 463 Su, Z.; Ye, L.; Lu, Y.: Guided Lamb waves for identification of damage in composite structures: A review. J. of Sound and Vibration 295 (2006), 753-780 Macdonald, J. R.: Impedance spectroscopy. New York: John Wiley & Sons; 1987 Park, S.; Park, G.; Yun, C.-B.; Farrar, C. R.: Sensor Self-diagnosis Using a Modified Impedance Model for Active Sensing-based Structural Health Monitoring. Structural Health Monitoring 2009; 8; 71-82 Diligent, O.; Grahn, T.; Boström, A.; Cawley, P.; Lowe, M. J. S.: The low-frequency reflection and scattering of the S0 Lamb mode from a circular through-thickness hole in a plate: Finite Element, analytical and experimental studies. J. Acoust. Soc. Am. 112 (6), 2002, 2589-2601 Köhler, B.: Dispersion Relations in Plate Structures Studied with a Scanning Laser Vibrometer. 9 th European NDT Conference, ECNDT 2006, Berlin, paper Mo.2.1.4 Cawley, P.; Lowe, M.J.S.; Alleyne, D.N.; Pavlakovic, B.; Wilcox, P.: Practical long range guided wave inspection - applications to pipes and rail, Materials Evaluation, Vol 61, (2003) 66-74 -9-