Hindawi Publishing Corporation

EURASIP Journal on Embedded Systems

Volume 2010, Article ID 578591, 9 pages
doi:10.1155/2010/578591

Research Article
Model and Design of a Power Driver for
Piezoelectric Stack Actuators

M. Chiaberge, A. Tonoli, G. Botto, M. De Giuseppe, S. Carabelli, and F. Maddaleno

Mechatronics Laboratory, Politecnico di Torino, Corso Duca degli Abruzzi 24, 10129 Torino, Italy

Correspondence should be addressed to M. Chiaberge, [email protected]

Received 29 May 2009; Revised 10 November 2009; Accepted 31 December 2009

Academic Editor: Luca Fanucci

Copyright © 2010 M. Chiaberge et al. This is an open access article distributed under the Creative Commons Attribution License,
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

A power driver has been developed to control piezoelectric stack actuators used in automotive application. An FEM model of
the actuator has been implemented starting from experimental characterization of the stack and mechanical and piezoelectric
parameters. Experimental results are reported to show a correct piezoelectric actuator driving method and the possibility to obtain
a sensorless positioning control.

1. Introduction There are fundamentally two different driving methods

for piezoelectric actuators:
In linear piezoelectricity, the equations of linear elasticity are (1) charge driving,
coupled to the charge equation of electrostatics by means of (2) voltage driving.
the piezoelectric constants. However, the electric variables
are not purely static, but only quasistatic, because of the Piezoelectric stack displacement depends on the charge
coupling to the dynamic mechanical equations. stored on it; so using a charge feedback amplifier, it is possible
The aim is to obtain an analytical model in order to study to correctly drive the actuator. Considering the piezoelectric
the displacement of the free end of the stack and the system actuator as a 4-port element, for charge driving techniques,
dynamics. To improve the efficiency of the transconductance inputs are charge and external force (if present) and the
amplifier employed, a switching bidirectional converter has outputs are tip displacement and voltage drops on it (see
been developed to correctly control the tip displacement of Figure 3). Using instead voltage driving techniques, inputs
the stack and to recover the energy stored in the actuator are voltage and external force and outputs are charge and tip
during the discharge phase; this solution minimizes the displacement. For piezoelectric stack, we can create an FEM
required PCB area and reduces the input power. The control model that is useful to study the displacements of the free end
strategy is chosen in order to drive the piezoelectric actuator of the stack and the system dynamics.
in charge for accurate tip displacement estimation and This model does not consider damping in a first stage:
reduction of the intrinsic hysteresis of material. The complete this effect is introduced in a second stage taking into account
system can be described in Figure 1. the experimental response of the piezoelectric actuator in
resonance conditions. Including the mechanical dynamic
behavior, this equivalent model can be easily implemented
2. Piezoelectric Stack Actuator Model in electronics simulation software.
For a piezoelectric material, the equations we need are
Hysteresis is observable in open loop operation; it can be clearly standardized in the document ANSI/IEEE STD 176-
reduced by charge control (Figure 2) and virtually eliminated 1987, which expresses piezoelectric equations and physical
by closed loop operation. constants in SI units.
2 EURASIP Journal on Embedded Systems

Power Vin (t) Step-up

supply (V) converter

DC bus
(V)

Reference Eref (k) M Drive(t) Ipzt (t) qtip (t)

Control Bidirectional Piezo
profile (uint16) (V) converter (A) actuator (um)
Vpzt (t)
(V)
Vsense (t)
(V)
V Qpiezo (t) Sensors
Vpiezo (t) (V)
(V)

Figure 1: Schematic diagram of the proposed control architectures.

The constitutive equations of a piezoelectric material at a element. If we consider charge driving, the equations of the
microscopically level are the following: stack are

S = sE T + dE, M q̈stack (t) + K Q qstack (t) − ΓQpiezo (t) = F ext (t),

(1)
D = dT + εT E, 1 (3)
ΓT qstack (t) + Qpiezo (t) = Vpzt (t).
Cpzt
or in an equivalent form:

T = cD S − hD, While using voltage driving, the equations are

(2)
E = −hS + βD. M q̈stack (t) + K V qstack (t) − ΘVpzt (t) = F ext (t),
(4)
The meaning of each symbol conventionally used in previous ΘT qstack (t) + Cpzt Vpzt (t) = Qpiezo (t).
expressions is the following:
(i) S : strain of the material, Mechanical and electrical dissipations are not considered
during this step. The meaning of each symbol conventionally
(ii) sE : Young’s modulus (m2 /N), used in previous expressions is the following:
(iii) T : mechanical stress (N/m2 ),
(iv) d : charge constant (C/N), (i) M : is the mass matrix,
(v) E: electric field applied (V/m), (ii) K V : is the stiffness matrix (short circuit Vpzt = 0),
(vi) D : electric displacement (C/m2 ), (iii) K Q : is the stiffness matrix (open circuit Qpiezo = 0),
(vii) εT : permittivity (F/m), (iv) Θ : is the electromechanical coupling matrix,
(viii) cD : elastic stiffness constant (N/m2 ),
(v) qstack : is the nodal displacement,
(ix) h : piezoelectric constant (V/m) = (N/C),
(vi) Fext : is the external force on stack tip,
(x) β : impermittivity constant (m/F).
(vii) Vpzt : is the voltage drops on piezo stack,
The relations that regulate the behavior of the entire
piezoelectric actuator are derived from a classic FEM (viii) Cpzt : is the equivalent piezo capacitance,
approach based on the standard defined equations.
(ix) Γ = Θ /Cpzt ,
The core of the actuator is a stack of piezoelectric layers.
Electrically speaking, the layers are in parallel, while from (x) Qpiezo : is the charge stored on piezo stack.
a mechanical point of view they are in series. Applying
a voltage to the electrical leads of the stack will cause a It is possible to transform the nodal equations into modal
displacement of the mechanical sides and a force on the faces equations in order to consider each mode independent from
that are respectively the sum of displacements and forces of the others. Considering only the resonances in the frequency
each layer. range between 10 Hz and 100 kHz, the model has been
Once defined the potential and kinetic energy associated reduced to the first five modes. The goal is to obtain electrical
to the element, it is possible to obtain the two electromechan- impedance in order to know the load to be applied to the
ical equations that regulate the behavior of the piezoelectric power driver.
EURASIP Journal on Embedded Systems 3

Performing a low-frequency analysis (s → 0), the inertial

properties of each mechanical parallel branches can be
neglected, while elastic contributes are summed to the
electrical capacitance:
30
 ϑ2
qpzt (um)

i
C∗ = C + . (6)
i
mi ωi2
20
Introducing the residue term ϑ2i /mi = hi , the static
capacitance (at s → 0) can be written as
10
 hi
C∗ = C + . (7)
i
ωi2
0 20 40 60 80
Vpzt (V)
So, the quantity sC ∗ corresponds to the equivalent low-
frequency admittance. Doing then a first-mode analysis, we
(a) should detract the first-mode quantity from equivalent low-
frequency capacitance:
 
sh h
Y (s) = 2 1 2 + s C ∗ − 12
s + ω1 ω1 s→0
30 ⎡  ⎤
h1  
qpzt (um)

⎢ C ∗ − 2 s2 + ω12 ⎥ (8)
sh1 ⎢ ω ⎥
= 2 + s⎢
⎢
1 ⎥
⎥ .
20 s + ω12 ⎣ s2 + ω2
1 ⎦
s→0
10 Impedance poles are admittance zeros; so in order to find
impedance natural frequencies, admittance’s numerator is
imposed equal to zero.
0 200 400 600 800 Keeping apart the s parameter in the gradient of the
Qpiezo (uC) curve, it is possible to impose equal to zero just on the part
in brackets:
(b)
   
h h
Figure 2: Piezoelectric hysteresis behaviour with voltage and charge h1 + s2
C − 12
∗
+ ω12 C − 12 ∗
= 0. (9)
driving strategies. ω1 ω1
First zero’s pulsation is
Fpzt1 Fpzt2 h1 + ω12 C ∗ − h1 ω12
s2z1 = − 2 =− , (10)
C − h1 /ω1
∗ 1 − h1 /ω12 C ∗
qpzt1 qpzt2
where ωi are pulsations of admittance antiresonances, which
are in correspondence to the impedance resonances. From
Ipzt experimental tests on the piezoelectric stack or from FEM
analysis, it is possible to obtain the value of the first
Vpzt
antiresonance pulsation (which also is the first admittance
Figure 3: Schematic representation of a piezoelectric element. resonance value) sz1 . Substituting this value in the last
expression, the first residue h1 is
 
ω2
h1 = ω12 1 + 21 C ∗ . (11)
Reducing the mechanical model, it is possible to convert sz1
it into a frequency domain as shown in (5) as equivalent
admittance. In this way, the admittance associated to the first natural
frequency is
 
sϑ2 sϑ2 sϑ2 sh
Vpzt  1 2 +  2 2 +· · ·+  n  + sC Y1 (s) =  2 1 2  . (12)
2 2
m1 s + ω1 m2 s + ω2 mn s2 + ωn2 s + ω1
= Ipzt . Keeping then in consideration all the successive modes, it is
(5) possible to extend the matter in order to determinate all the
4 EURASIP Journal on Embedded Systems

successive residues. For the first n modes, impedance can be

written as
⎡ ⎤ L1 L2 L3 Ln
n
 n

shi h i
Y (s) = 2 + ω2
+ s⎣C −
∗ ⎦ . (13)
i=1
s i i=1
ωi2
s→0 C C1 C2 C3 Cn

Residues hi can be written as

hi Rs R1 R2 R3 Rn

  i    i  
i−1
1/s2zi C ∗ 2
j =1 ω j ω2j − s2zi + j =1 h j
2
k=2 ωk ωk2 − s2zi
=− i−1   .
2
j =1 ω j ω2j − s2zi
Figure 4: Electrical equivalent circuit for piezoelectric actuator
(14)
model.

3. Introduction of Mechanical and

Electrical Losses
A term-by-term comparison with the first parallel branch
In order to obtain a model closer to the reality, it is parametric admittance leads to
necessary to introduce a modal damping element into each 1 s
resonance term of the transfer function. It is important Y1 (s) = . (19)
L1 (s2 + 1/L1 C1 )
to keep in consideration a resistance in series with the
piezo capacitance, which models the electrical losses due The following relations can be used in order to define lumped
to electrical connections. Including damping, the resulting parameters values for the electrical equivalent circuit:
admittance is
⎛⎡ ⎤ ⎞ 1 1
n n h1 = , L1 = ,
 shi  h L1 h
Y (s) = + ⎝s⎣C −
∗ i ⎦ ⎠// 1 . 1 −→ 1
1
(20)
s2 + 2ζω s + ω2 ω2 RS ω12 = , C1 = .
i i i=1 i
i=1 s→0 L 1 C1 L1 ω12
(15)
All the relations before can be used to evaluate the parame-
Damping can be considered just near natural frequencies ters relative to the first branch and can be extended to each
relative to the branch, while at each other frequency can be successive branch of the electrical equivalent circuit:
neglected.
1 1
hi = , Li =,
Li hi
4. Definition of the Equivalent 1 −→ 1 (21)
Electrical Circuit Parameters ωi2 = , Ci = .
L i Ci Li ωi2
From (15), it is clear that the electric equivalent circuit of the Residues value can be used to characterize the inductive
mechanical properties is made up by n parallel branches to and the capacitive elements of the electrical equivalent
piezoelectric equivalent capacitor C. Each branch is made up implementation, while resistance value is determined by
by a capacitor and an inductor (in first approximation losses comparison with experimental data obtained on piezo stack
are neglected), so the physical admittance of each parallel sample.
branch of the electrical equivalent circuit is
1 1 s 5. Electrical Equivalent
Yi (s) = = . (16)
sLi + 1/sCi Li (s2 + 1/Li Ci ) The admittance can be represented using the electrical equiv-
Starting from the admittance obtained from modal analysis, alent circuit in Figure 4, where parallel branches correspond
it is possible to rewrite the last expression, in order to do a to the first 5 modes of the stack (electrical equivalent of
term-by-term comparison and determinate the parameters mechanical modes).
for the electrical equivalent circuit. A comparison between the experimental impedance and
⎡ ⎤ the result of electrical equivalent model is showed in Figure 5.
n
  hi n
shi The electrical admittance is a good approximation of
Y (s) = ⎣ ∗ ⎦
2 +s C − . (17) the piezo stack electrical behavior, so it will be used as a
i=1
2
s + ωi ω2
i=1 i s→0 “reference load” for power driver design.
Our analysis is limited to the first five most important modes,
starting from the admittance related to the first mode: 6. Power Driver Design
sh1 Injector controller is aimed at the control of the fuel quantity
Y1 (s) =  . (18)
s2 + ω12 injected in the combustion chamber of the cylinders. This
EURASIP Journal on Embedded Systems 5

Bode magnitude diagram-piezo with case-comparison Ipzt

104

103

102 Transconductance Vpzt

Vsupply
amplifier
Impedance (Ohm)

101

100

10−1

10−2 PWM
Signal
10−3
Figure 6: Schematic architecture of the proposed driving stage.
10−4
102 103 104 105
Frequency (Hz)
in a complementary way. To obtain the complete discharge of
Experimental the piezoelectric equivalent capacitance, a dissipative section
FEM model (braking transistor) has been added.
Figure 5: FEM equivalent model and experimental measurements The most important physical quantity for a piezoelectric
comparison. actuator is the charge stored in the equivalent capacitance
which depends on temperature. To cover all the possible
driving techniques, the current flow to/from the actuator and
can be done by performing a position control on the injector the related voltage on it must be carefully measured and con-
needle. trolled. The basic converter scheme is proposed in Figure 7.
Unfortunately, no velocity, position, or force measure can
be performed inside the injector, neither on the piezoelectric 7. Power Driver Control
stack nor on the leverage.
So, in order to estimate the injector needle position, it Power driver control is based on an inner current loop and
is important to measure a related quantity. The constitutive an outer energy loop that controls the equivalent energy of
equations of a piezoelectric stack showed that the needle the actuator. The inner loop is a hysteretic current mode
position is strictly related to the electrical charge loaded controller with quasi constant frequency that solves the
in the equivalent capacitance of the injector itself. A direct problem of sub harmonic instability, guarantees a quasi
charge measurement is also possible but it is an intrusive constant switching frequency, and reduces the inductance
technique. Nevertheless, loaded charge is proportional to current ripple (Figure 8).
the current transferred to the piezo stack by the power The hysteretic “window” is adjusted by varying the valley
driver. current reference value starting from the difference between
The implemented control is based on two different reference switching periods and measured ones.
feedback loops: Starting from a hysteretic width initial value, the control
system measures the switching period, and using a frequency
(1) an inner loop which controls the injector current,
feedback loop, the integral regulator determines the hys-
(2) an outer loop which controls the charge stored in the teresis variation in order to maintain the frequency quasi
injector. constant [2].
The outer charge/energy loop (Figure 9) is a classic PI
Hence from the charge stored in the piezo equivalent
control law with anti-windup that follows the reference
capacitance, it is possible to estimate the needle position.
profile.
From the models above, it is possible to consider the
piezo stack as a capacitive load. A good method to drive
a capacitive load is to use a current generator driven by 8. Implementation
an external signal: the simplest way to implement this type
of current generator is to use a transconductance amplifier The proposed system is implemented using a DSP/FPGA
driven by a PWM signal (see Figure 6). based prototyping control platform developed at Mechatron-
A good solution to obtain a high-efficiency amplifier is to ics Lab. In Figure 10, it is possible to see the power board
use a class D amplifier with current mode control in order to with the bidirectional converter (on the left) and the FPGA
have a quasi-ideal current generator [1]. controller board (on the right) between dummy loads (top)
The chosen topology is based on a synchronous bidirec- and the piezoelectric injector (bottom).
tional Buck Converter operating in a Continuous Conduc- The power board is also provided with a boost DC/DC
tion Mode (CCM) in order to reduce current stresses on converter [3] also controlled from the same FPGA device
electronic components, where the two transistors are driven in order to make the system compliant with automotive
6 EURASIP Journal on Embedded Systems

M1 Rdis
+
Vsupply Cboost
M2
M3
Rsense

Bidirectional Braking Current Voltage

converter circuit sense sensor
Power driver
V Ipiezo
Vpiezo

Figure 7: Schematic circuit of the bidirectional converter.

Vpzt (t)
V Iref (t) eI (t) M1 drive(t) Ipzt (t) (um)
(V) (V) (bit) (A)
Power Piezo
Current driver stack
control
qpzt (t)
(um)
Kh Tsw (s)
(uint16)
Integral eT (s) Tmis (s)
Time
control + counter
law −
Tref (s) Frequency
feedback
control
Current
sense

M1 drive(t)
V Ipeakref (t) (V) (bit)
V Ivalleyref (t) (V) Comparators
V Ipiezo (t) (V)

Hysteretic
Kh
(uint16)

Figure 8: Power driver control architecture with current and frequency control loops.

applications where the standard 12 V power source is not signals are used to schedule the algorithms. The profile
useful to drive the piezoelectric actuators. generator, for instance, uses a clock enable with a period of
The entire FPGA project takes up to 12 k logic elements 1 µs, the same used to trigger the ADC sampling.
(on a maximum of about 33 k available on the ALTERA Using a single main clock for most of the synchronous
Cyclone II device used in the application) divided into 18 logic allows to reduce problems related to different clock
main entities; the implemented code uses about 150 kbits of domains; only the ADC and DAC IP cores are fed by
the memory embedded on the FPGA and 4 DSP elements clock signals at lower frequency for the SPI communication
for the fast fixed point multiplications. The main clock has a between the FPGA and the converter devices; the ADCs clock
frequency of 100 MHz, generated by an internal PLL fed by has a frequency of 20 MHz which allows 1 µs of sampling
an external 50 MHz oscillator. period while the current loop DACs clock has a frequency of
All the state machines and processes are synchronized by 50 MHz resulting in 500 ns of maximum update period (the
the main clock, and some internally generated clock enable real frequency is controlled by the inner control loop).
EURASIP Journal on Embedded Systems 7

Brake and
EOI
comparator

Vpzt (t)
M1 drive(t) Ipzt (t)
Epiezo ref (t) eE (t) V Iref (t) eI (t) (um)
(bit) (A)
(V) (V) PI (V) (V) Current
Saturation Power Piezo
control control driver stack
qpzt (t)
V Ipiezo (t)
Anti (um)
Eelt (t) (V) Current Qpiezo (t)
wind-up
(V) sense (uC)
Qpiezo mis (t)
(V) Analog
integrator

Vpzt (t)
(um)

Figure 9: Power driver control architecture with charge/energy loop.

Table 1: Piezoelectric actuator tip estimated displacement versus experimental measurements.

Measured Estimated
Charge storage Piezo Voltage Tip displacement Piezo Voltage Estimated position Position Error
Qpiezo [µC] V pzt [V] qpzt [um]  pzt [V]
V qpzt [µm] Err [%]
275 45 16 45 16.56 3.38
384 65 24 65 23.88 0.48
545 83 34 83 34.7 2.03
604 90 40 90 38.67 3.44
714 110 48 110 46.06 4.21
878 120 56 120 57.08 1.89
933 130 60 130 60.78 1.28
1098 145 68 145 71.87 5.38
1164 155 80 155 76.3 4.85

The final goal of this FPGA implementation is to demon- In Figure 12, piezo actuator current (yellow) and piezo
strate that all the injector control strategies and algorithms voltage (magenta) are shown. The blue line in the oscil-
can easily fit in a final ASIC device for the automotive market loscope picture is the STROBE signal used to start charge
using standard available BCD silicon technologies. (fall edge) and discharge (rise edge) sequence during piezo-
electric actuator test. To avoid resonance oscillation, current
reference profile has a rise time of about 200 µs in order
9. Experimental Results
to guarantee a soft start charge phase while at the end
Experimental results were obtained using the test bench in of discharge phase the braking MOSFET turns on for the
Figure 11 in order to validate the relation between charge and complete discharge of the load (a small voltage/current peak
displacement as shown in Table 1, where in the last column is visible almost at the end of the discharge phase).
errors between indirect measure and experimental one are Figure 13 shows the dynamic behavior of the system
highlighted. applying a reference profile with multiple steps. Waveforms
As previously mentioned, tip displacement is strictly in the above figure are piezoelectric actuator current(yellow),
related to the charge stored in the actuator, so using an voltage (red), tip displacement (green), and stored charge
inverse kinematics it is possible to indirectly obtain the tip (blue), respectively. From the above pictures, it is possible
displacement measuring the charge. to notice a direct relationship between piezoelectric actuator
Typical voltage and current waveforms obtained with this stored charge and measured tip displacement while the
type of actuator driving stage are shown in Figures 12 and 13. voltage suffers of the intrinsic hysteresis phenomena.
8 EURASIP Journal on Embedded Systems

Ipzt (t) (A)

2
0
−2
−4
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
Time (s) ×10−3

(a)

Vpzt (t) (V)

200
100
0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
Time (s) ×10−3

(b)

qpzt (t) (um)

Figure 10: The experimental set-up of the proposed FPGA 60
controlled power driver with dummy loads and a piezoelectric 20
actuator connected. −20
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
Time (s) ×10−3

(c)
Sensor signal XYZ Linear stages
conditioning block
Qpzt (t) (uC)

Sensor controller 1500

1000
500
0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
Time (s) ×10−3
Thermocouple
(d)
Injector

Figure 13: Piezoelectric actuator current, voltage, measured tip

Injector heater
displacement, and charge during a multilevel driving test.

Laser sensor
Temperature PID controller Injector support 10. Conclusion
In this paper, a complete approach to piezoelectric actuators
Figure 11: The experimental test bench for piezo injector charac- modeling and control strategy design has been presented.
terization. The results reported in the present paper show that the
model obtained is a good approximation of the real dynamic
response of the piezoelectric actuator and the tip displace-
ment can be estimated from the charge measurements with
an error less than 5% without a direct measure on the
C3
actuator (not possible in many applications).
The designed power module correctly drives the piezo-
electric actuator and allows controlling the tip displacement
C1
using the suggested indirect method.
Moreover, the presented bidirectional power driver
allows energy recovery from the actuator during the dis-
charge phase with an overall efficiency above 80%.
With respect to other implementations [4], the proposed
system is able to guarantee good performances with a single
C2 DCBUS rail voltage. Moreover, two nested control loops
(current and charge/energy) allow a very precise estimation
of needle position strictly related with fuel injection in the
cylinder combustion chamber. Those top performances lead
Figure 12: Piezoelectric actuator voltage and current during to high efficiency in engine combustion and less exhaust gas
charge/discharge phases. released in the atmosphere.
EURASIP Journal on Embedded Systems 9

References
[1] L. H. Dixon, “Average current mode control of switching power
supply,” in Proceedings of the Unitrode Power Supply Design
Seminar, pp. 5.1–5.14, 1988.
[2] X. Yang and Z. A. Wang, “A novel quasi-constant frequency
hysteretic current mode control approach,” in Proceedings of the
IEEE Annual Power Electronics Specialists Conference (PESC ’03),
vol. 3, pp. 1147–1150, June 2003.
[3] Q. Zhao and F. C. Lee, “High-efficiency, high step-up DC-DC
converters,” IEEE Transactions on Power Electronics, vol. 18, no.
1, pp. 65–73, 2003.
[4] G. Gnad and R. Kasper, “Power drive circuits for piezo-electric
actuators in automotive applications,” in Proceedings of the IEEE
International Conference on Industrial Technology (ICIT ’06), pp.
1597–1600, 2006.