Proceedings of the 4th Iranian International NDT Conference

Feb 26-27, 2017, Olympic Hotel, Tehran, Iran

IRNDT 2017

Finite Element Modeling of Ultrasonic Transducers

Mostafa Rafienezhad Masouleh1 and Farhang Honarvar2
1
M.Sc., K. N. Toosi University of Technology; [email protected]
2
Professor, K. N. Toosi University of Technology; [email protected]
More info about this article: http://www.ndt.net/?id=21268

Abstract
Ultrasonic transducers are the most important part of an ultrasonic testing system which significantly affect the
precision and accuracy of the measurements. Piezoelectric transducers are the most common ultrasonic transducers
which include three major components: the piezoelectric disk, backing material, and matching layer. The optimal
combination of these three components is traditionally done by trial and error. Analytical models used for this purpose
have their own pros and cons and are not capable of predicting the complete performance of the transducer. Finite
element modeling is an alternative method that can be used for designing ultrasonic transducers. This paper presents a
finite element model for a 4 MHz ultrasonic transducer excited by a spike signal in pulse-echo mode. The model
considers an ultrasonic transducer composed of a piezoelectric disk, a matching layer, and a backing material. The main
advantage of this model in comparison to earlier models is the incorporation of piezoelectric elements instead of
mimicking the piezoelectric effect by applying forces or displacements. In the developed model, the input is an
electrical signal and the output is an electrical voltage. This makes this model very similar to the way a real ultrasonic
transducer operates. The simulation results are compared with results obtained from experiments and very good
agreement is observed.

Keywords: Ultrasonic testing, Transducer, Finite element method, Piezoelectric

1- Introduction
An ultrasonic pulse-echo measurement process involves the generation of ultrasound by the transducer, propagation of
these ultrasonic waves into the medium and then reception of these waves by the transducer again. The pulser sends the
electrical pulse via a cable to an ultrasonic transducer. The transducer converts this electrical pulse into mechanical
vibrations which could then propagate as an ultrasonic pulse into the material. The transducer also converts the returned
ultrasonic pulse into an electrical pulse before transmitting it into the receiver. In ultrasonic measurement instruments,
the transducer is a crucial part, which can significantly affect the precision of the system.
In the past, the development of ultrasonic transducers was usually done by trial and error. This would take a lot of time
and cost a lot. It is difficult or even impossible to verify that the inadequate response of a transducer is due to design or
construction problems. However, if a numerical model is used for designing the transducer, the transducer response is
known in advance.
Researchers have already made efforts in modeling and simulating ultrasonic wave propagation to obtain further
understanding of this process [1]. The models commonly used to simulate the mechanical and electrical behavior of
piezoelectric transducers by incorporating simplifying assumptions. The geometries of practical transducers are often
two or three-dimensional. However, the most popular models, such as Mason's or KLM models are only one-
dimensional [2-5]. Comparison of a simulation with a related measurement proves that one-dimensional models can be
rather inadequate when applied to practical transducer geometries [6].
In designing and examination of ultrasonic piezoelectric ceramic transducers, accurate simulation tools such as the
finite element method (FEM) can be valuable tools. Application of the finite element method is one possible way to get
a realistic transducer simulation and to visualize the real acoustic wave propagation into the medium. FEM is an
efficient way to obtain the desired solution of a complex partial differential equations (PDE) which cannot be accurately
studied using analytic methods [7]. Limitations of one-dimensional analytical models such as geometric constraints for
accurate simulation and inability to simulate interaction with other objects can be avoided by using the finite element
method.
Finite element capabilities in simulation of piezoelectric structures were identified in late 1970s [8-10]. Since then,
much work has been done in improving the modeling of piezoelectric problems [11, 12]. Piezoelectric materials and
transducers are used in various industries and equipment such as microphones, speakers, pressure sensors, ultrasonic
motors and so on. Transducer structures vary according to their applications. Three-dimensional formulation of
ultrasonic transducers, analytical methods of piezoelectric materials, estimation of natural frequencies and the
corresponding eigenvalues, and limitations of various formulations were investigated by using the finite element
method. It was shown that the standard one-dimensional models are available only in certain cases [13-16]. Kocbach
reviewed the influence of material and geometry parameters on vibration, response functions, and radiated field of
 IRNDT2017

transducers in transmission mode without considering the backing material [17]. Medina carried out simulations on
ultrasonic transducers in transmission mode and observed that the simulated signals are noisier than those obtained from
experiments [18]. Aanes studied the benefits and limitations of finite element modeling of transducers [19]. Only one
piezoelectric crystal was modeled and electrical conductivity was compared with experimental measurements. Nygren
introduced the material damping into the transducer but was not able to obtain the desired results [20]. Imperiale
developed a code for the finite element model of an ultrasonic transducer by using domain decomposition technique
[21]. Bilgunde studied the impact of changes in time period and the mesh size on his FEM model. He compared his
simulation results in terms of received echo time with experimental results. Echo time was selected as the output
parameter because transducer components information were taken from the literature and compared with experiment
[22]. Over the years, much work has been done by using simplifying assumptions where not all three components of the
transducer are modelled or just considering the transducer as a transmitting element [23-25].
The aim of this paper is to develop a finite element model that could be used as a designing tool for optimizing the
performance of custom made ultrasonic transducers. We use the complete transducer model including the piezoelectric
disk, backing layer, and matching layer in both transmit and receive modes. To validate the model, experimental
measurements are conducted on an in-home constructed piezoelectric transducer.

2- Transducer General Description

An ultrasonic transducer is a layered structure as shown in Figure 1. The basic components of the transducer are the
piezoelectric disk, backing material, and matching layer.

2-1 The Piezoelectric Disk

The piezoelectric disk is a crystal which converts the electrical energy into mechanical (ultrasonic) energy and vice
versa. The crystal is a piezoelectric ceramic with two electrodes coated on its two opposite faces. Applying an electric
impulse makes the ceramic vibrate at its resonance frequency [26]. The piezoelectric ceramic thickness determines its
central frequency. The constitutive matrix equations relating the mechanical and electrical quantities that are the basis
for the derivation of the finite element model are:

(1)

(2)

where T is the tensor of mechanical stress ( ), S is the tensor of mechanical strain, E is the electric field vector ( ), D
is the vector of dielectric displacement ( ), CE is the stiffness matrix for constant electric field ( ), εS is the
permittivity matrix for constant mechanical strain ( ), and e is the piezoelectric matrix ( ) [27].

Figure 1. Schematic design of a typical ultrasonic transducer.

2-2 Backing Material

In order to support the piezoelectric disk, a backing material is used at the back of the piezoelectric disk. The backing
material absorbs the back-transmitted energy and controls the vibration of the disk. To minimize the internal reflections,
the acoustic impedance of the backing material matches that of the piezoelectric crystal. A typical backing material
consists of a mixture of tungsten powder and epoxy [28]. To have an appropriate backing layer, several design
considerations need to be considered such as appropriate attenuation coefficient and impedance of the backing material
that would result in the desired bandwidth of the transducer [29].

2-3 Matching Layer

One or more front layers can be used to improve the power transmission between the piezoelectric disk and the
propagating fluid. These layers also acts as wear protection plates for a piezoelectric element. The addition of more than
 IRNDT2017

one matching layer increases the complexity to manufacture the transducer. However, this also increases the efficiency
of the transducer and results in the improvement of both bandwidth and sensitivity.
Collin [30], based on transmission line theory, mentions that the optimal acoustic impedances for one matching layer
(Zm) is between impedance of the piezoelectric ceramic (Zpiezo) and impedance of medium (Zmed):

(3)

For the design of a wide-band transducer, Desilets et al. in [5] modify the choice of front layer using the KLM theory:

(4)

3- Transducer Modeling
The models are done with finite elements simulating a transducer radiating waves in a medium. The models are
developed in ABAQUS v6.12 software. In the case of circular transducer, the model has been considered an axial
symmetric geometry in order to reduce the computational time of simulations, Figure 2. In the model X axis is parallel
to the transducer face, and Y axis is in the direction of wave propagation. The total time of propagation was 25 µs with
0.01 µs for the time sampling.

Figure 2. Axisymmetric model.

As the model is axisymmetric, all components of the transducer have been modeled with axisymmetric plane elements.
The models were shaped with square 2D elements of 4 and 8 nodes. For wave propagation problems, the spatial
discretization must be defined such that it can resolve the shortest wavelength of interest. It is recommended to use 10
element per wavelength to get an adequate solution [15].
For the finite element modeling, the piezoelectric material is modeled as orthotropic (transverse isotropic) material. The
ceramic used in the construction of ultrasonic transducers has symmetry in the XY plane and is polarized on Z axis. The
PZT-5A ceramic used in this work, with 10 mm diameter, thickness of 0.54 mm, and central frequency of 4 MHz.
The piezoelectric crystal has not constraints or load, except for the electrical point of view. To create the wave, it is
enough to excite the piezoelectric crystal and creating the potential difference between two faces above and below it. A
spike excitation is applied on the upper surface of the crystal, while the bottom one is grounded. Excitation pulse is
defined by using the following model [31]:

(5)

where and the three parameters t, t0, and V0 are time (µs), maximum pulse time (µs), and
maximum voltage (V). Figure 3 shows the pulse excitation obtained from the above equation.
 IRNDT2017

Figure 3. Spike excitation pulse.

To model the backing layer, epoxy material was considered by attenuation of 8.13 dB per meter. In order to eliminate
interference echoes reflected from surface of backing material, the minimum thickness can be achieved by using
following equation [32]:

(6)

where L, α, and are propagating distance (m), attenuation (dB/m), and ratio between transmitted and received wave
amplitude.
For our case, acoustic impedance of piezoelectric disk is 33 MRayl, whereas, acoustic impedance of fluid (propagating
medium) is 1.5 MRayl. So, based on Desilets theory, matching layer with the acoustic impedance 4 MRayl is required
for the wideband transducer. In the existing transducer, Plexiglas is used as the front layer and has an acoustic
impedance of 3 MRayl. This value is very near to the required impedance for the wideband transducer. Therefore, the
existing selection is a good (not ideal) selection as the matching layer. The thickness of the front layer is set equal to
quarter wavelength (λ/4) to improve the transmitted wave into the propagating medium [33]. Table 1, Summarizes all
material parameters used in the finite element model.

Table 1. Summary of material parameters used in the models.

Material Parameter Value Material Parameter Value
Density (Kg/m3) 3674 d31 -175
Epoxy Poison Ratio 0.33 Piezoelectric Constants (10-12 m/v) d33 400
Elasticity Module (GPa) 6.7 d15 590
C11 121
Density (Kg/m3) 1158 C12 75.4
Plexiglas Poison Ratio 0.33 PZT-5A Elastic Constants (GPa) C13 75.2
Elasticity Module (GPa) 5.4 C33 111
C44 211
Density (Kg/m3) 1000 ε33 830
Dielectric Constants
Water Bulk Module (GPa) 22 ε11 916
Sound Velocity (m/s) 1500 Density (Kg/m3) 7650

Contact between transducer components and contact between transducer and propagating medium were defined in such
a way to allow the waves to be transmitted. Therefore, the nodes are tied to each other. In this case, the acoustic
impedance of material is the only influential factor in the reflecting and transmitting of the waves. In the experiment,
there is a sheet on the tank floor which is simulated with zero displacement of low end of fluid in the model. This
boundary condition is used rather than modeling sheet because simulating the thin sheet needs to use finer mesh which
increases the solution time.
 IRNDT2017

4- Numerical Results
We have performed three simulations in order to study the effect of components and to complete our model: one with a
piezoelectric disk, the other with a piezoelectric disk and backing material, and finally with all three main components
(piezoelectric disk, backing material, and matching layer).

4-1 Piezoelectric Disk

In the first part, the FEM model of the piezoelectric disk propagating in a water medium was performed. In this form,
only the piezoelectric disk was simulated. Figure 4 shows the output signal of the piezoelectric crystal. The first echo
shows vibration of piezoelectric crystal after excitation. The second echo is due to the wave that is transmitted in the
water and been received by piezoelectric disk after reflection from the rear surface of water.

Figure 4. Received signal by piezoelectric disk.

4-2 Piezoelectric Disk and Backing Material

This time, we present another configuration to illustrate the effect of backing material on the response of transducer.
Backing material is used to minimize internal reflections, controlling piezoelectric vibration and limit the pulse
duration. We consider a backing material directly in contact with piezoelectric disk. Backing material with thickness of
15 mm and impedance of 6 MRayl was added to simulate the effect of backing material on transducer response. As
expected (Figure 5.), signal amplitude has fallen due to absorption of wave energy by the backing material. Also, due to
energy absorption, pulse length decreased which is highly desirable in the designing of ultrasonic transducer because it
will result in increasing transducer resolution. The use of backing material produces a short response by reducing the
ringing following the initial excitation of the piezoelectric disk.

Figure 5. Received signal by piezoelectric disk with backing material.

 IRNDT2017

4-3 Completed model

In the last stage of modeling, matching layer with 3 MRayl impedance and thickness of a quarter wavelength was added
to the model. The thickness of a quarter wavelength increases the amplitude. The selected Impedance, which is very
close to proposed impedance for Desilets theory, shortening the pulse and thereby increase the bandwidth, Figure 6.

Figure 6. Received signal by completed model.

5- Experimental Validation
The transducer was developed in laboratory with the same characteristics as the FEM model. PZT-5A piezoelectric
crystal used in transducer with 4 MHz nominal frequency. Matching layer and backing material attached directly to the
front and rear of crystal. Figure 7 shows the final transducer image. The experimental setup was performed in water
tank in pulse-echo mode, Figure 8. A Pulser/Receiver Panametrics 5072PR was used to generate excitation signal.

Figure 7. In-house manufactured ultrasonic transducer. Figure 8. Experimental test setup.

In this section, the simulation results will be compared with test results. Figure 9 shows a comparison of echoes
obtained by the experiment and simulation. Two echoes are very similar but there are little differences in the ringing.
For the best understand of the graphics all the curves are normalized.
The frequency response curve is extracted. Figure 10 shows the frequency response curve obtained by simulation and
experiment. The center frequency and frequency bandwidth of 3.5 MHz and 28.57 percent, respectively, is obtained
from the simulation, while the amount of 3.7 MHz and 24.32 percent, respectively, shown in experimental test. Good
agreement can be seen between the simulation and experimental result. In simulation many parameters are assumed to
be ideal and some parameters are simplified which may affect system performance. Therefore it could be expected that
difference exist between simulation results and experimental measurement. The difference in the received signal and
frequency response can be caused by the following factors:
- Errors in the manufacturing process.
- Ideal assumption of component connections, frequency and thickness of crystal, and … in simulation.
- Piece surface finishing, no parallel surfaces, and … are factors that may cause errors in experimental tests.
 IRNDT2017

Figure 9 - Comparison between an experiment and a simulated signal.

Figure 10. Comparison between experiment and simulation frequency response.

6- Conclusions
Many designers employ one dimensional analytic models to predict the behavior of ultrasonic transducers. However,
due to certain limitations these models are not adequate for understanding the complete performance of an ultrasonic
transducer. In recent years, finite element modeling has been considered for designing the ultrasonic transducers. The
objective of modeling and simulation of transducers is to optimize the design parameters without going through time-
consuming tests. It also makes it possible to easily evaluate new materials and study the output signal of the system.
In this paper, ultrasonic transducer simulation was carried out by using the finite element method. A 4 MHz ultrasonic
transducer consisting of a piezoelectric disk, a backing material, and a matching layer was modelled. The modelled
transducer was excited by a spike signal in pulse-echo mode. The FEM results were compared with experiments and
good agreement was observed. This kind of simulation can be used as a powerful tool for predicting the behavior of
ultrasonic transducers.

References
1. Chang, J., Ni, Q., and Iwamoto, M., 2004. "Simulation of ultrasonic wave propagation in model composite
materials". Japanese Journal of Applied Physics, 43, p. 2926.
2. Mason, W. P., 1948. Electromechanical transducers and wave filters. 3rd ed., D. Van Nostrand Co.
3. Krimholtz, R., Leedom, D. A., and Matthaei, G. L., 1970. "New equivalent circuits for elementary piezoelectric
transducers". Electronics Letters, 6, pp. 398-399.
4. Kossoff, G., 1966. "The effects of backing and matching on the performance of piezoelectric ceramic transducers".
IEEE Transactions on Sonics and Ultrasonics, 13, pp. 20-30.
5. Desilets, C. S., Fraser, J. D., and Kino, G. S., 1978. "The design of efficient broad-band piezoelectric transducers".
IEEE Transactions on Sonics and Ultrasonics, 25, pp. 115-125.
 IRNDT2017

6. Lerch, R., 1988. "Finite element analysis of piezoelectric transducers". IEEE Ultrasonics Symposium, pp. 643-654.
7. Qi, W., and Cao, W., 2000. "Finite element study on 1-D array transducer design", IEEE Transactions on
Ultrasonics, Ferroelectrics, and Frequency Control, 47, pp. 949-955.
8. Tiersten, H. F., 1967. "Hamilton's principle for linear piezoelectric media". Proceedings of the IEEE, 55, pp. 1523-
1524.
9. EerNisse, E. P., 1967. "Variational method for electroelastic vibration analysis". IEEE Transactions on Sonics and
Ultrasonics, 14, pp. 153-159.
10. Allik, H., and Hughes, T. J. R., 1970. "Finite element method for piezoelectric vibration". International Journal for
Numerical Methods in Engineering, 2, pp. 151-157.
11. Allik, H., Webman, K. M., and Hunt, J. T., 1974. "Vibrational response of sonar transducers using piezoelectric
finite elements". The Journal of the Acoustical Society of America, 56, pp. 1782-1791.
12. Naillon, M., Coursant, R. H., and Besnier, F., 1983. "Analysis of piezoelectric structures by a finite-element
method". Acta Electronica, 25, pp. 341-362.
13. Ostergaard, D. F., and Pawlak, T. P., 1986. "Three-dimensional finite elements for analyzing piezoelectric
structures". Ultrasonics Symposium, pp. 639-644.
14. Lerch, R., 1990. "Simulation of piezoelectric devices by two-and three-dimensional finite elements". IEEE
Transaction on Ultrasonics, Ferroelectrics and Frequency Control, 37, pp. 233-247.
15. Abboud, N. N., Wojcik, G. L., Vaughan, D. K., Mould Jr, J., Powell, D. J., and Nikodym, L., 1998. "Finite element
modeling for ultrasonic transducers". Ultrasonic Transducer Engineering Conference, pp. 19-42.
16. Wang, J. S., and Ostergaard, D. F., 1999. "A finite element-electric circuit coupled simulation method for
piezoelectric transducer". IEEE Ultrasonics Symposium, pp. 1105-1108.
17. Kocbach, J., 2000. "Finite element modeling of ultrasonic piezoelectric transducers". PhD Thesis, Department of
Physics, University of Bergen, Bergen.
18. Medina, M., Buiochi, F., and Adamowski, J. C., 2006. "Numerical modeling of a circular piezoelectric ultrasonic
transducer radiating water". ABCM Symposium Series in Mechatronics, pp. 458-464.
19. Aanes, M., Storheim, E., and Vestrheim, M., 2010. "Finite element analysis and measurement of ultrasonic
piezoceramic transducers in air and water". BNAM, Norway.
20. Nygren, M., 2011. "Finite element modeling of piezoelectric ultrasonic transducers". Ms Thesis, Department of
Electronics and telecommunications, Norwegian University of Science and Technology, Norway.
21. Imperiale, S., Marmorat, S., Leymarie, N., and Chatillon, S., 2012. "A complete FE simulation tools for NDT
inspections with piezoelectric transducers". Acoustics Nantes, France, pp. 2651-2656.
22. Bilgunde, P., and Bond, L., 2015. "A 2D finite element simulation of liquid coupled ultrasonic NDT system". AIP
Conference Proceedings, pp. 1543-1552.
23. De Paolis, S., Lionetto, F., and Maffezzoli, A., 2009. "Finite Element Modeling of Ultrasonic Transducers for
Polymer Characterization". COMSOL Conference, Milan, p. 4.
24. Han, G., Zhang, J., Li, B., Lu, J., and Dou, W., 2012. "A Novel FEM Method of Modeling and Visualization".
Proceedings of the Second International Conference on Instrumentation, Measurement, Computer, Communication
and Control, pp. 1476-1479.
25. Imperiale, S., and Joly, P., 2012. "Mathematical and numerical modelling of piezoelectric sensors". ESAIM:
Mathematical Modelling and Numerical Analysis, 46, pp. 875-909.
26. Cobbold, R. S. C., 2007. "Foundations of biomedical ultrasound", Oxford University Press, USA.
27. Meitzler, A., Tiersten, H. F., Warner, A. W., Berlincourt, D., Couqin, G. A., and Welsh Iii, F. S., 1987. IEEE
standard on piezoelectricity, Standards Committee of the IEEE Ultrasonics, Ferroelectrics, and Frequency Control
Society, p. 66.
28. Sayers, C. M., and Tait, C. E., 1984. "Ultrasonic properties of transducer backings". Ultrasonics, 22, pp. 57-60.
29. Brown, L. F., 2000. "The effects of material selection for backing and wear protection/quarter-wave matching of
piezoelectric polymer ultrasound transducers". IEEE Ultrasonics Symposium, pp. 1029-1032.
30. Collin, R., 1955. "Theory and design of wide-band multi section quarter-wave transformers". Proceedings of the
IRE, 43, pp. 179-185.
31. Schmerr, L., and Song, J.-S., 2007. Ultrasonic nondestructive evaluation systems: models and measurements,
Springer.
32. Bray, D. E., and Stanley, R. K, 1996. Nondestructive evaluation: a tool in design, manufacturing and service. CRC
press.
33. Wang, H., Ritter, T. A., Cao, W., and Shung, K. K., 2001. "High frequency properties of passive materials for
ultrasonic transducers". IEEE Transactions on Ultrasonics, Ferroelectrics and Frequency Control, 48, pp. 78-84.