Scholarly article on topic 'Low-frequency Inductive Power Transmission for Piezo-Wafer-Active-Sensors in the Structural Health Monitoring of Carbon-Fiber-Reinforced-Polymer'

Low-frequency Inductive Power Transmission for Piezo-Wafer-Active-Sensors in the Structural Health Monitoring of Carbon-Fiber-Reinforced-Polymer Academic research paper on "Materials engineering"

CC BY-NC-ND
0
0
Share paper
Academic journal
Procedia Technology
OECD Field of science
Keywords
{"Lamb waves" / PWAS / SHM / "wireless sensors"}

Abstract of research paper on Materials engineering, author of scientific article — Mariugenia Salas, Oliver Focke, Axel S. Herrmann, Walter Lang

Abstract Structural Health Monitoring (SHM) using surface applied or embedded Piezo-Wafer-Active-Sensors (PWAS) to generate and measure Lamb-waves on fiber-reinforced materials is the main target of this study. Wireless power transmission was achieved implementing the principle of inductive coils. The characteristics of the material where the coil system is located have a direct impact on the strength of the magnetic-field. Carbon Fiber Reinforced Polymer (CFRP) is known to behave as a conductor and depending on their frequency; radio waves can be shielded by such materials. Therefore, a new interest to evaluate the direct operation of PWAS via inductive coils at the low-frequency range, where the effect is lower, has been raised. A low-frequency wireless SHM system is achieved by combining the PWAS, in this case bonded to the surface, with an inductive coil in resonance. A CFRP plate with surface bonded PWAS was prepared in order to evaluate the power transmission; not only between antenna coils but also the energy necessary to stimulate the PWAS was measured. To enable this direct excitation of the PWAS, the same frequency which is provided on the primary side is used on the secondary side. Typical values of a PWAS are obtained form an electrical model based on its measured impedance.

Academic research paper on topic "Low-frequency Inductive Power Transmission for Piezo-Wafer-Active-Sensors in the Structural Health Monitoring of Carbon-Fiber-Reinforced-Polymer"

Available online at www.sciencedirect.com

ScienceDirect

Procedia Technology 15 (2014) 649 - 658

2nd International Conference on System-Integrated Intelligence: Challenges for Product and

Production Engineering

Low-frequency inductive power transmission for Piezo-Wafer-Active-Sensors in the Structural Health Monitoring of Carbon-

Fiber-Reinforced-Polymer

Mariugenia Salasa*, Oliver Fockeb*, Axel S. Herrmannc, Walter Langd

a Friedrich Wilhelm Bessel Institut Forschungsgesellschaft m. b. H., 28359 Bremen, Germany b Faserinstitut Bremen e.V., 28359 Bremen, Germany c University of Bremen (FB4), 28359 Bremen, Germany dInstitut für Mikrosensoren, -aktoren und -systeme (IMSAS),University of Bremen, 28359 Bremen, Germany

Abstract

Structural Health Monitoring (SHM) using surface applied or embedded Piezo-Wafer-Active-Sensors (PWAS) to generate and measure Lamb-waves on fiber-reinforced materials is the main target of this study. Wireless power transmission was achieved implementing the principle of inductive coils. The characteristics of the material where the coil system is located have a direct impact on the strength of the magnetic-field. Carbon Fiber Reinforced Polymer (CFRP) is known to behave as a conductor and depending on their frequency; radio waves can be shielded by such materials. Therefore, a new interest to evaluate the direct operation of PWAS via inductive coils at the low-frequency range, where the effect is lower, has been raised. A low-frequency wireless SHM system is achieved by combining the PWAS, in this case bonded to the surface, with an inductive coil in resonance. A CFRP plate with surface bonded PWAS was prepared in order to evaluate the power transmission; not only between antenna coils but also the energy necessary to stimulate the PWAS was measured. To enable this direct excitation of the PWAS, the same frequency which is provided on the primary side is used on the secondary side. Typical values of a PWAS are obtained form an electrical model based on its measured impedance.

© 2014 The Authors. Published by ElsevierLtd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.Org/licenses/by-nc-nd/3.0/).

Peer-review under responsibility of the Organizing Committee of SysInt 2014.

* Corresponding author. Tel.: +49-421-218 -62609; fax: +49-421-4088-9-6262; E-mail address: salas@fwbi-bremen.de

* Corresponding author. Tel.: +49-421-218 -58655; fax: +49-421-58710; E-mail address: focke@faserinstitut.de

2212-0173 © 2014 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/).

Peer-review under responsibility of the Organizing Committee of SysInt 2014. doi:10.1016/j.protcy.2014.09.026

Keywords: Lamb waves; PWAS; SHM; wireless sensors

1. Introduction and Background

There is a growing interest towards wireless technologies that would enable Structural Health Monitoring (SHM) of carbon fiber reinforced polymers (CFRP) with applications in different industries, including but not limited to aeronautics and wind energy. These structures are interesting due to their complex mechanical and electrical anisotropic properties. Embedded sensors in CFRP have been demonstrated showing that the major challenge is the integration without affecting the low weight of the structure or generating an impact [1, 2]; hence the interest in the use of small and very thin wireless powered systems. The goal is to create a wireless sensor network for use in the structural health monitoring systems by implementing the approach of low frequency (LF) inductively coupled coils together with Piezo-Wafer-Active-Sensors (PWAS).

Nomenclature

D coil diameter

r coil radius

h height of the solenoid coil

a planar coil mean radius

c winding thickness for planar coil

N number of windings

L coil inductance

k coupling factor between coils

M mutual inductance

C capacitance

z distance between coils

x distance between PWAS-actuator and PWAS-sensor

1.1. Structural Health Monitoring using Piezo- Wafer-Active-Sensors

Lamb waves, or guided ultrasonic plate waves, propagate in a solid medium between two parallel surfaces over large areas; therefore they are suitable for sparse sensor arrays. More than one wave modes exist for each frequency. The modes at low frequencies are S0, A0 and SH, the latter one, denoting the shear horizontal wave. The symmetrical modes are called, S0, Si S2...and the anti- symmetric ones A0 Aj, A2..., [Fig. 1] starting with the mode that has the lowest frequency for a given wave number [3]. These waves are easily excited and measured, for example by integrated or surface-applied piezoelectric elements. They have successfully been used to identify and localize both various discrete, localized damage, such as holes, notches, cracks, delamination or weak bonds [4] as well as distributed, non-localized damage such as fatigue in composites [5]. Examples for other investigations are characterization of the influence of moisture absorption on Lamb wave propagation [6] or the inverse measurement of mechanical properties [7]. Especially the detection of non-visible and barely visible damages such as delamination, these are typical and hard to detect from outside the structure when using classical visual inspection methods.

Fig. 1. Theoretical calculated curves for the lowest Lamb wave modes S0, A0 and SH in an UD-CFRP plate in reinforcement direction [8, 9]: group velocity.

Lamb wave based SHM methods can be distinguished in a passive [8] or an active system. The first one uses the waves; which are created by damaging events, such as crack initiation and growth or an impact event (hail, bird strike) on the other side an actuator is used to create the Lamb waves in a controlled fashion. For SHM purposes, where the sensing ability is integrated in or onto the structure, they are most commonly excited and measured using piezoelectric elements. Fig. 2 shows a simple example of a SHM-System with piezoelectric elements bonded to the plate's surface. An excitation signal is applied to a piezoelectric element which acts as an actuator. These generated waves depend therefore on the interaction between actuator, adhesive and local material properties. While traveling dispersive through the structure, these waves are under laying damping due to viscoelastic material behavior and contact to surrounding fluids or materials. Reaching the sensor, the strains are transmitted through the bonding layer to the piezoelectric element, based on the inverse piezoelectric effect an electric output signal is created which then can be displayed and subsequently can be analyzed. In this study piezoelectric PIC255 from Piceramic was used [9].

Fig. 2. SHM with Piezo Elements using Pitch Catch Method [8].

SHM using Lamb waves is based on the fact that damage represents a discrete, local change of the waveguide which interacts with the propagating Lamb wave or, in case of a passive system, that the damaging event generates a Lamb wave of certain properties. The signal analysis has to take into account that these waves are reflected at edges or transmitted through interfaces, and therefore undergoing also mode changes. Lamb waves should be excited at frequencies in the range of 0 kHz to 500 kHz, here only the lowest Lamb wave modes are generated (Fig. 1). And these waves travel with different group velocities. Depending on the distance between actuator and sensor the S0 and A0 signal arrive at different arrival times. Therefore the modes can be identified. The Lamb wave frequencies depends on the excitation time of the actuator, the output voltage on the sensor materials Young's modulus, its piezoelectric constant, the sum of extensional surface strains, the sensor thickness and the sensor area [4]

1.2. Inductive Wireless Power Transmission: Low-Frequency Selection

By means of inductive coupled coils we eliminate the need for electrical connectors or wiring the PWAS. In consequence, it possible to have a material integrated passive system, meaning no external source of power will be necessary. Moreover the lifetime of such system will be larger than those active systems depending on a battery. In [10] wafer-type piezoelectric transducers combined with inductive coils were successfully used to generated Lamb-waves, however the reading range between coils was for gaps of a few millimeters between the probe coils.

On a previous research [11] different tests to validate the influence of CFRP on the transmission were carried. First, a unidirectional laminate was cut into small square pieces and placed on top and bottom of the receiver, sealing it in order to observe the effect on the coupling at different distances between coils. At HF the efficiency was

reduced to one third of its original value. In the LF range, the efficiency decreased to half the original value. On a second test a one layer unidirectional laminate was placed directly under the receiver in order to observe the effect on the reflection coefficient of the antenna coil. The quality of the antenna coil in the LF range decreased, but the effect of CFRP is stronger in the HF range, where the resonance frequency shifted.

Receiver S11 at 13.56 MHz

Fig. 3. Microwave CST Simulation for HF coils placed near CFRP and Aluminium.

In order to confirm the results, the high-frequency behaviour of the coefficient S11 was simulated by using CST Microwave Studio, placing a very large unidirectional plate with thickness 0.5 mm at a distance of 1 cm from the receiver coil, as shown in Fig. 3. The anisotropic behavior of CFRP material was defined using a model, where conductivity and permittivity are determined by the ratio of carbon fibre to resin content [12, 13]. The results were similar to those found in [14] and therefore this technology has been set aside in the meantime. A low-frequency wireless SHM system is achieved by combining the PWAS with an inductive coil in resonance similar to a Radio-Frequency-Identification System [15]. In this case the voltagetransferred can beobtained according to Eq. 1, where C2 is the capacitance in resonance with the receiver coil at a given frequency foEq. 2.

jMk^L1L2 i1

l + (juL2 + R2) + ;wC2)

here Rl is the load on the secondary side and fo is the resonance frequency with ra = 2rcfo. In this case we substitute the load in the receiver for the piezoelectric. A PWAS is connected in parallel with the coil, while the resonance frequency is adjusted by means of a capacitor, as seen in Fig. 4. Additional components may be necessary to keep the PWAS power supply stable and to obtain the pulse from the primary side.

f= 250 to 500 KHz

Primary Side: Transmitter Secondary Side: Receiver + Load

Fig. 4. Wireless Power Transfer to a PWAS.

The desired z,.ead distance between coils goes ideally up to 5 cm with maximum at 2.5 cm. Using Eq. 3, D can be found.

Zread = 2vF (3)

In the following sections the system design will be explained. Exploring the constrictions set by the selected piezoelectric sensors. Finally, some initial results are presented from a system (Actuator- Sensor) bonded to a CFRP plate. The here mentioned coils were evaluated away from the CFRP plate. The magnetic flux flows through the primary coil and depends on its area. Consequently the geometry of both inductors determines the energy. High energy content is linked to high inductance. In this study we take into consideration K as the dimensionless coupling factor given by Eq. 4 which describes the mutual inductance M between the two coils and is decisive for the energy content of a wireless coil system [15].

k = ^rUN^rlrj = M " ijL^r^tf+z)3 ()

2. Low Frequency Coil Optimization

Transmitter and receiver coils were made with Litz wire of 25/0.071mm. The inductance was calculated according to Wheeler's formula for a solenoid inductance [16]:

L = h + 0.45 D (4)

Indicated in Table 1 are the coils used on the first coupling test at 125 kHz resonance frequency. The system has a maximum read at 2.8 cm according to Eq. 3. The power sent by the primary coil is 250 mW; the maximum voltage peak-to-peak obtained at the secondary coil was 33.4 V as seen in Fig. 5, where the received voltage has been plot versus the distance z between coils. The PWAS was the load (approximately 150 Q). This system has a power sent to power received ratio of efficiency 88.2%. The calculated coupling is 0.5 according to Eq. 4.

Table 1. First System - Receiver and Transmitter Characteristics.

_Number of Windings Diameter [mm]_L[|iH] Calculated_L[|iH] Measured

Receiver 45 80 273 272.2

Transmitter 15 80 36 34.4

The Spice simulations using a transformer model are very similar to the values measured, however the theoretical approximation does not account for the negative coupling which causes a shift on the resonance frequency. The calculated values on the graphic from Fig. 5 have been shifted to match the simulation, observing that at 2 cm the negative coupling ends.

Voltage at Secondary Coil

^ 30,00 &

M 20,00 at

£ 10,00

erv 0,00

1,00 2,00 3,00 4,00 5,00 Distance Between Coils [cm]

* Average Measured ■ LTSpice Model x Calculated

Fig. 5. Coupling Results between Coils at 125 kHz Resonance Frequency with 150 Ohm load.

The goal is to surface apply the coils to the CFRP with the possibility of future use of other materials like GFRP to protect from external damages. One way would be by glue or binders and another approach is the use of sewing techniques (tailoring). However it might be applied it should be consider that the quality of the receiver antenna will be damped to about half the value here presented and depending on the frequency use we might observe a shift in resonance.

In order to produce thinner systems, receiver coils were made with magnetic bifilar wire AWG 32 (MWS Wire Industries) with Polyimide high temperature up to 200°C isolation. The inductance was calculated according to Wheeler's formula for planar coil inductance [16] where N is the number of turns, a is the coil mean radius, and c is the thickness of the winding.

a2 (5)

L = 31.33 ^0N2--—

8a + 11c

The transmitter solenoid coil was built with the same diameter as the receiver in order to obtain a better understanding. Similarly to the previous system its inductance was calculated using Eq. 5. By means of these two coils a maximum of 30 Vpp was obtained at a frequency of 500 kHz using the PWAS- actuator as the receiver load. The receiver was tuned by adding a capacitor as explained in section 3.

Table 2. Receiver and Transmitter Coil Characteristics. (Maximum Read 0.7cm).

_Number of Windings_Diameter [mm]_L[[H] Calculated_L[[H] Measured

Receiver 30 40 34.46 32.4

Transmitter 18 40 20.54 21.7

An extra coil with an inductance of 9.7 |iH available in the market (from Vishay) was selected due to its high quality factor given by the use of iron particles on its core. This coil was used as transmitter in some cases, where indicated.

3. PWAS Characterization

The PWAS are surface bonded to a Quasi-Isotropic plate; these are the load for the inductive wireless coil system. Impedance measurements were made using HP 4194A Impedance Analyzer. Based on five different PWAS an average was found and used as the departing point for an easy model (another form of the Van Dyke Model) approximation for a PZT that uses a parallel RLC tank circuit as seen on Fig. 6e. This model is explained on [17]. Results are shown in Fig. 6 where (a) is the resistance of the PWAS. (b) Is the phase in radians, (c) is the magnitude of the impedance and (d) is the reactance.

Frequency [KHz]

^ -2.6 C

.2 -2.7 "D

ÛÉ -2.8

m -2.9

Frequency [KHz]

200 250 300 350

Frequency [KHz]

100 150 200 250 300 350

Frequency [KHz]

Fig. 6. PWAS Electronic Model based on Impedance Measurements

In Fig. 6 the samples were plot together with the approximation given by the easy-model. The electrical model is a resonance circuit based on the phase and resistance of the PWAS as seen in Fig. 6 (d). From the reactance curve and the magnitude curve one can observe that the PWAS has a series resonance at 250 kHz and a parallel resonance at 275 kHz. It can also be noted from the resistance curve that the base resistance R is about 37.5 Q. The calculation of the electrical component values was done and used for further in the modelling process. Using the approximation model Co is about 3.4nF, having therefore an influence on the resonance system when connected in parallel to the receiver coil. Consequently the resonant capacity must be adjusted.

Using a sine signal the coil system in resonance can be tested together with the PWAS Actuator- Sensor as shown in Fig. 7. The signal from the PWAS Sensor is an average of 64 samples with about 60 mVpp. For this test the transmitter coil was used as the transmitter and the receiver is from Table 2. The sine is given by a signal generator at a frequency of 500 kHz.

System at 500 KHz, L1=9.7 pH L2=32 pH

JË 10 ■§ 0

Signal from PWAS Sensor

Time[s]

Timefsl

Fig. 7. (a) Sine Excitation of PWAS actuator, (b) PWAS sensor response.

4. System Integration and Measurements

Lamb wave generation occurs by excitation of a PWAS actuator via a Hann-windowed sine signal. In order to try to mimic this behavior the primary coil is connected to microcontroller and amplification circuitry that will provide the necessary signal and power for the secondary system. Fig. 8 is a representation of the test here carried.

Fig. 8. Measurement system set-up.

The entire system can be divided into three sections. A primary side responsible for the signal generation and amplification connected to a primary coil that oscillates at the necessary frequency. The secondary side is formed by the receiver coil together with the PWAS actuator and tuning capacitors necessary to bring the system to resonate at the selected frequency. Finally, there is a third section including a PWAS sensor and the signal detection via oscilloscope. The simulations were carried using a spice model depicted on Fig. 9(a) where L1 and C1 correspond to those from Fig. 8. On the right side [Fig. 9(b)] the signals indicate the results for the simulations and the measurements with a five peak signal at a frequency of 250 kHz. The signal seen at the sensor indicates that the wave travels in one direction without seen any modes. This is given by the nature of the excitation signal being too long. In this test the transmitter and receiver were those from Eq. 4.

Fig. 9. Spice Simulation of the behaviour from a receiver coil with PWAS Actuator. Where coupling k=2. (a) Schematic representation, (b) top: Simulated results and measurements, bottom: Signal observed at the PWAS sensor. Nerveless the excitation time was not short enough to generate Lamb waves.

A pulse of 2 |is width with a waiting time of 248 |is is provided via microcontroller to the primary coil which is set at resonance as indicated on the system set-up. Such resonating system is capable of giving an oscillation at such desired frequency as see in Fig. 10. The purpose of this signal is to quickly excite the PWAS system and wait to see a response. Such response will be composed by symmetric and asymmetric modes. One of the biggest challenges in the future is to find out at what frequency it is possible to eliminate the interference from the symmetric and the asymmetric modes as explained in [18]. It can be seen in Fig. 10 that the signal from the sensor is delayed due to the distance to the actuator, but also the number of oscillations increased while keeping the centre-frequency. This indicated that the lamb-wave is traveling along and transversally on the plate. This indicates that an Ao and a S0 Lamb wave is generated, this has to be proved by reference measuring with a wired system on the same plate at the

same frequency and a modified Hann-windowed sine signal.

System at 500 KHz, L1=9.7 mH L2=32 mH

-Signal sent va Receiver Coil

Timers]

Signal from PWAS Sensor

Timers]

Fig. 10. (a) Excitation PWAS-Actuator and (b) Received Signal at PWAS-Sensor.

5. Conclusions and Future Work

Even though it is not yet possible to distinguish the lamb-wave modes received from the PWAS-Sensor it has been shown that a short pulse together with resonating coil system can excite a PWAS-Actuator according to the selected frequency. The power transmitted will depend on the coil-system coupling and efficiency, which in turn is affected by the conductivity of CFRP, with worse results at higher frequencies. The next step in this research is the embedding of the coils in FRP; an example is shown in Fig. 11a. here a coil is realized by Tailored-Fiber-Placement with a modified embroidery machine (Tajima TMHL- G 108) [21]. Also the PWAS and the necessary electronics should be placed in the same way on an embroidery ground. The signal integrity should be evaluated and compared to that from a non-wireless system. Fig. 11b. is an example of a receiver coil bonded to carbon fiber before the addition of epoxy in order to form the composite.

Fig. 11. (a) sewed coils into textile, (b) surface bonded coil to carbon fibres.

Acknowledgements

This research was conducted as part of the IGF- project Nr. 17649N from the German Research Association for Measurement, Controls and Systems Engineering supported by the AiF in the funding structure for industrial cooperative research (IGF) by the Federal Ministry of Economics and Technology based on a decision from the

German Bundestag. References

[1] Lang W, Boll D, Schotzko T. Function Scale Integration- Embedding Sensors in Materiasl for Structural Health Monitoring. 6th European Workshop on Structural Health Monitoring, 2012.

[2] Lang W, Jakobs F, Tolstosheeva E, Sturm H, Ibragimov A, Kesel A, Lehmhus D, Dicke U. From embedded sensors to sensorial materials— The road to function scale. Sensors and Actuators A: Physical 2011;171/1:3-11.

[3] Herrmann AS, Calomfirescu M. Theoretical and experimental studies of Lamb wave propagation in attenuative composites. Sensors and Smart Structures Technologies for Civil, Mechanical, and Aerospace Systems 2007;6529.

[4] Giurgiutiu V. Structural Health Monitoring With Piezoelectric Wafer Active Sensors. Academic Press; 2008.

[5] Rheinfurth M, Kosmann N, Sauer D, Busse G. Lamb waves for non-contact fatigue state evaluation of composites under various mechanical loading conditions. Composites: Part A 2012;43:1203-1211.

[6] Herrmann AS, Schubert KJ. On the influence of moisture absorption on Lamb wave propagation and measurements in viscoelastic CFRP using surface applied piezoelectric sensors. Composite Structures 2012;94:3635-3643.

[7] Castaings M, Hosten B, Kundu T. Inversion of ultrasonic, plane-wave transmission data in composite plates to infer viscoelastic material properties. NDT & E International 2000;33/6:377-392.

[8] Calomfirescu M. Lamb Waves for Structural Health Monitoring in Viscoelastic Composite Materials. PhD, University of Bremen: Logos Verlag; 2008.

[9] Konstantin ASH, Schubert J. On attenuation and measurement of Lamb waves in viscoelastic composites. Composite Structures 2011 ;94:177-185.

[10] Herrmann AS, Schubert KJ. A Compensation Method for Environmental Influences on Passive Lamb Wave Based Impact Evaluation for CFRP. Key Engineering Materials 2013:1265-1272.

[11] Piceramic. http://www.piceramic.de/piezokeramische_werkstoffe_4.php.

[12] Greve DW, Sohn H, Yue P, Oppenheim IJ. An Inductively Coupled Lamb Wave Transducer. IEEE Senors Journal 2007;7/2:395-301.

[13] Salas M, Focke O, Lang W, Herrmann A. Wireless Power Transmission for Structural Health Monitoring of Fiber-Reinforced-Composite Materials. IEEE Sensors Journal Special Issue on Material-integrated Sensing 2014.

[14] Galehadar A, Nicholson K, Rowe KST, Ghorbani K. The Conductivity of Unidirectional and Quasi Isotropic Carbon Fiber Composites. Preceedings of the 40th European Microwave Conference 2010.

[15] Lee SE, Oh KS, Kim CG. Electromagnetic Characteristics of Frequency Slective Fabric Composites. IEEE Electronic Letters 2006;42/8:439-441.

[16] Mayordomo I, Dräger T, Bernhard J. Technical Challenges for the Integration of passive HF RFID. IEEE International Conference on RFID-Technologies and Applications 2011.

[17] Finkenzeller K, Müller D. RFID Handbook, 3 Hrsg., Chichester: Wiley; 2010.

[18] Wheeler HA. Simple Inductance Formulas for Radio Coils. Proceedings of the Institute of Radio Engineers 1928;16/10.

[19] Kim J, Grisso BL, Kim JK, Ha DS, Inman DJ. Electrical Modeling of Piezoelectric Ceramics for Analysis and Evaluation of Sensory Systems. IEEE Sensors Applications Symposium 2008.

[20] Su Z, Ye L. Identification of Damage Using Lamb Waves- From Fundamentals to Applications. Lecture Notes in Applied and Computational Mechanics. Proceedings of the Institution of Mechanical Engineers, Part L: Journal of Materials Design and Applications. Berlin: Springer; 2004.

[21] Schiebel P, Herrmann AS. Analyse des Verformungsverhaltens von Kohlenstoff-Rovings im textilen Fertigungsprozess für belastungsgerechte CFK-Strukturen, 13th Chemnitzer Textiltechnik Tagung 2012.