Structural Health Monitoring of Aerospace Structures with Sol-Gel Spray Sensors

A. Ouahabi 1, a, M. Thomas 1, b, M. Kobayashi 2, c and C.-K. Jen 2,d 1Department of Mechanical Engineering, École de technologie supérieure,

1100 Notre-Dame Street West, Montréal, Québec, H3C 1K3, Canada 2Industrial Materials Institute, National Research Council Canada,

75 de Mortagne Blvd, Boucherville, Québec, J4B 6Y4, Canada aabdelhakim.ouahabi@etsmtl.ca, bmarc.thomas@etsmtl.ca,

cmakiko.kobayashi@cnrc-nrc.gc.ca, dcheng-kuei.jen@cnrc-nrc.gc.ca

Keywords: health monitoring, sensors, detection, localization

Abstract. A new approach is proposed for conducting structural health monitoring, based on newly

developed piezoceramic sensors. They are fabricated by a sol-gel spray technique. The potential

application of these sensors may be broad. These sensors have been evaluated for structural health

monitoring studies. The purpose of the present study aims the detection and the localization of

defects by the means of these new piezoceramic sensors. Nine sensors were integrated onto a

metallic plate with moving masses. The plate was excited by an impact at a specific location and the

vibratory signals from sensors were recorded simultaneously. The analysis of signals obtained from

nine locations was correlated with a numerical simulation in order to identify at each time the

location of the mass.

Introduction

Non-destructive testing (NDT) of materials are commonly performed to identify, characterize,

assess voids, defects and damage in metals, metal alloys, composites and other materials [1].

Furthermore, the increasing demand to improve the performance, reduce downtime, increase

reliability and extend the life of transportation vehicles, structures and engineering systems, requires

the use of systems that have integrated capabilities with built-in sensors that perceive and process

in-service information and take actions to accomplish desired operations and tasks [2]. Piezoelectric

ceramic sensors and actuators are commonly used as key candidates for smart materials and

structures. They have been used as structural vibration actuators, structural health monitoring

sensors, non-destructive evaluation probes for materials and structures, etc [3]. Thick (≥ 40 µm)

piezoelectric ceramic films can be made by the technologies of jet printing [4], dipping [5], tape

casting [6], hydrothermal method [7], etc.

Here, an alternative sol-gel spray technique is used. They take form of thick film layers with

substrates having different shapes. An important characteristic of these sensors is that they can be

coated directly onto desired sensing locations of planar or curved structures. These sensors present a

considerable advantage because their fabrication focuses on the use of hand held tools. It is possible

to obtain a thickness ranging between 25 and 150 µm and a diameter which can be less than

centimeter. Their operating temperature ranges from -100°C to 440°C [3]. For example, they can be

applied as sensors onto thin metallic and flexible membranes or around cylindrical surfaces. The

purpose of the present study aims the detection and the localization of defects by the means of new

piezoceramic sensors in the form of thin layers of piezoelectric materials coated directly onto the

structures.

Fabrication and Characterization

The fabrication process was first developed at Queen’s [8]. The piezoelectric particles are dispersed

in the sol-gel solution to produce a thick piezoelectric film [3, 9]. The spray can be carried out by an

air gun at room temperature and it is simple and inexpensive. The ball-milled sub-micron

Key Engineering Materials Vol. 347 (2007) pp. 505-510online at http://www.scientific.net© (2007) Trans Tech Publications, Switzerland

All rights reserved. No part of contents of this paper may be reproduced or transmitted in any form or by any means without thewritten permission of the publisher: Trans Tech Publications Ltd, Switzerland, www.ttp.net. (ID: 142.137.48.230-24/05/07,15:35:43)

piezoelectric lead-zirconate-titanate (PZT) or bismuth titanate (BIT) powders were dispersed into

PZT sol-gel solution. The PZT and BIT powders were chosen because of their high piezoelectric

constant and high Curie temperature (675°C), respectively. The final PZT/PZT or BIT/PZT mixture

(paint) was then sprayed directly onto selected metallic substrates, such as stainless steel through an

airbrush. With this sol-gel spray technique, the films can be produced at desired locations using a

paper shadow mask. After spraying the coating, thermal treatments such as drying, firing and/or

annealing were normally carried out using a heat gun. In special cases in this study a furnace would

be used and mentioned specifically. Multiple coatings were made in order to reach desired film

thicknesses. The film thickness is between 40 and 200 µm. Piezoelectric films were then electrically

poled using a corona discharging technique. The corona poling method was chosen because it could

pole the piezoelectric film over a large area with complex geometries. Finally, silver paste, platinum

paste or silver paint spray method was used to form the top electrodes at room temperature. The

measured relative dielectric constant of the PZT/PZT film and BIT/PZT film was about 320 and 80,

respectively. The d33 measured by an optical interferometer was 30 10-12

m/V for PZT/PZT and 10

10-12

m/V for BIT/PZT. The thickness mode electromechanical coupling constant measured was 0.2

for PZT/PZT and that for BIT/PZT was lower. Fig. 1 presents the flow chart of the fabrication

process of the thick film described above.

Fig. 1: Flow chart of the fabrication process of piezoelectric thick film.

The numerical and experimental study of integration of this kind of sensors has been developed

for the defective and no defective plate in order to compare between them, and to extract a

conclusion concerning the effectiveness of these sensors for the detection and the localization of the

defects.

Experimental Modal Analysis

Nine of piezoceramic PZT/PZT sensors were deposited onto a steel plate (100 ×100 × 0.5 mm) as

shown in Fig. 2. The thickness of theses piezoceramics is 40 µm thick. The electrodes have been

mounted on each piezoceramic layer by using a silver-pen. The wires were glued with silver epoxy.

The plate is mounted in clamped configuration. A small moving mass of 3 grams is bonded to the

plate at different locations to simulate a defect in the plate. At each location of the mass, an

experimental modal analysis has been conducted by using the impact technique and compared with

the results obtained without mass. The vibratory response was simultaneously collected at each

Piezoelectric Powder with High Curie Temperature

Mixing with Solution of High Dielectric Constant

Air Spray

Drying, Firing and Annealing By Heat Gun or Torch

Desired Thickness

Yes

Corona Poling

Fabrication of Top Electrode

No

Damage Assessment of Structures VII506

location with the piezoceramic sensors and the impact was measured with force sensor. The first

five natural frequencies were analysed. Fig. 3 presents one of the frequency response functions

(FRF). As expected, the effect of added mass can be easily detected from the decrease of natural

frequencies. However, it can be noticed that the sensitivity of the decrease is not the same at each

natural frequency. The variation of each natural frequency accordingly with the position of the

added mass will be demonstrated in the latter section.

Fig. 2: Clamped plate under impact excitation (Hammer) and classification of sensors.

0.00 400.00 Hz

0.00

0.36

Amplitu

de

( g/N)

0.00

1.00

Amplitu

de

F FRF w ithout defectF FRF w ith defect

Fig. 3: FRF of defective (mass at location 1) and no defective plate.

We have noticed a hardening effect in time of the piezoceramic layers that have produced an

increase from a small amount (up to 10%), of the natural frequencies of the plate. Also, comparative

tests with an accelerometer show that this sol-gel spray sensor is able to be used as sensor because it

delivers the same type of signal without amplifier.

1

2

3

4

5

6

7

8

9

Key Engineering Materials Vol. 347 507

Numerical Modal Analysis

The cantilever plate has been discretized into 28×28 finite elements, as shown in Fig. 4, and the

effect of piezoceramic layers has been considered. The first five natural frequencies of the plate

were numerically computed without mass and with the mass located at various positions (i) (i =

1,…9). Figure 5 shows the corresponding mode shapes (ANSYS©). It was revealed that the mode

shapes are not affected by the position of the added mass.

Fig. 4: FEM modelization of plate with piezoceramic layers.

Detection and Localization of Defect

Accordingly to the location of the mass (0 means without defect), Fig. 5 shows clearly a decrease of

natural frequencies. That decrease is more sensitive to some specific frequencies accordingly with

the position of maximum amplitude of modes, as it is put in evidence by arrows). By investigating

simultaneously all the natural frequencies that are most affected when the added mass is located at a

specific position, and by analysing the location of maximum amplitude for each considered mode, it

is now possible to locate the possible positions of defect. Table 1 shows the process of

identification. By considering more modes, the process of identification will be more accurate. The

variation of frequencies when the mass was located close to the clamp was not enough sensitive to

conclude.

Location of

mass

Affected

frequencies

nodes of maximum amplitude of

modes

Identification of mass

position

1 1st

2nd

1, 2 and 3

1 and 3

1 or 3

2 1st

4th

1, 2 and 3

1, 2, 3 and 5

1, 2 or 3

3 1st

2nd

1, 2 and 3

1 and 3

1 or 3

4 3rd

5th

2, 4, 5 and 6

1, 3, 4, and 6

4 or 6

5 3rd

4th

2, 4, 5 and 6

1, 2, 3 and 5

2 or 5

6 3rd

5th

2, 4, 5 and 6

1, 3, 4, and 6

4 or 6

Table 1: Identification of possible locations of defect.

Damage Assessment of Structures VII508

Mode 1

Mode 2

Mode 3

Mode 4

Variation de la 4th frequency

305

310

315

320

325

330

335

0 1 2 3 4 5 6 7 8 9 10

Position of the added mass

Fréquency [Hz]

Variation of the 1st frequency

38

39

40

41

42

43

0 1 2 3 4 5 6 7 8 9 10

Position of the added mass

Frequency [Hz]

Variation of the 2nd frequency

94

96

98

100

102

104

106

0 1 2 3 4 5 6 7 8 9 10

Position of the added mass

Frequency [Hz]

Variation of the 3rd frequency

235

240

245

250

255

260

265

0 1 2 3 4 5 6 7 8 9 10

Position of the added mass

Fréquency [Hz]

Key Engineering Materials Vol. 347 509

Mode 5

Fig. 5: Variation of natural frequencies with added mass.

Conclusions

New integrated thick piezoceramic sensors were used for the detection and the localization of

defects. They seem effective for potential structural health monitoring. Their advantage lies in the

fabrication that focused on the use of the handheld and readily accessible equipment to perform sol-

gel spray technique. The merits of these sensors are mainly in their miniature and lightweight (film

thickness ≥ 40 µm), flexible, non-destructive test and smart sensing site. An application of these

sensors in order to detect a light added mass was successful, by analysing the decrease of frequency

produced by the mass. Furthermore, an analysis of the more affected frequencies and their

corresponding modes allowed for the identification of possible locations of the mass.

Acknowledgements

The authors are grateful to the support of the Industrial Materials Institute (IMI), National Research

Council of Canada (NRC), and the Consortium for Research and Innovation in Aerospace in

Quebec (CRIAQ).

References

[1] A.S. Birks, R.E. Green, Jr and P. McIntire in: Nondestructive Testing Handbook, 2nd

Ed.,

volume 7 of Ultrasonic Testing, ASNT (1991).

[2] M T. Kundu: Ultrasonic Nondestructive Evaluation: Engineering and Biological Material

Characterization (CRC Press, New York, 2004).

[3] M. Kobayashi and C. K. Jen: Smart Materials and Structures Vol. 13 (2004), p.951

[4] H. Adachi, Y. Kuroda, T. Imahashi and K. Yanagisawa: J. J. Appl. Phys. Vol. 36 (1997),

p.1159

[5] K.L. Gentry, J.M. Zara, S.-D. Bu, C.-B. Eom and S.W. Smith: Proc. IEEE Ultrason. Symp.

Vol. 2 (2000), p. 977

[6] C. Galassi, E. Roncari, C. Capiani and P. Pinasco : J. Eur. Ceram. Soc. Vol. 17 (1997), p. 367

[7] M. Shimomura, T. Tsurumi, Y. Ohba and M. Daimon: J. J. Appl. Phys. Vol. 30 (1991), p. 2174

[8] A. Barrow, T.E. Petroff, R.P. Tandon and M. Sayer: J. Appl. Phys. Vol. 81 (1997), p. 876

[9] M. Kobayashi, T. R. Olding, M. Sayer and C. K. Jen: Ultrasonics Vol. 39 (2002), p. 675

Variation of the 5th frequency

355

360

365

370

375

380

0 1 2 3 4 5 6 7 8 9 10

Position of the added mass

Fréquency [Hz]

Damage Assessment of Structures VII510