XVIII Congresso Brasileiro de Automtica / 12 a 16-setembro-2010, Bonito-MS

ACTIVE MODAL DAMPING OF TRUSS STRUCTURES USING INTEGRAL CONTROL STRATEGY GUSTAVO LUIZ C. M. ABREU, VICENTE LOPES JR. GMSINT, Depto. de Engenharia Mecnica, Faculdade de Engenharia de Ilha Solteira-FEIS Av Brasil, 56, 15385-000 Ilha Solteira, SP, Brasil E-mails: gustavo@dem.feis.unesp.br, vicente@dem.feis.unesp.br
Abstract This paper presents an active modal damping control of a truss structure using one pair of piezoelectric stack actuators collinear with force transducers and integral control strategy. A finite element model of the structure is constructed using the three-dimensional frame elements considering electro-mechanical coupling between the host structure and piezoelectric stack actuator. In this paper, the active member placement is determined by using the fraction of modal strain energy as an optimal index. The integral controller, obtained by using the root locus technique, is designed to maximize modal damping of the truss structure. The numerical simulation results illustrate that the active members of the structure and the integral controller can effectively reduce truss vibrations. Keywords Active Damping, Truss Structure, Integral Control, Piezoceramic Stack Actuators, Finite Element Method.

Introduction In this case, the control system consists of independent SISO loops, whose stability can be readily established from the root locus technique. The choice of the actuator/sensor location is another important issue in the design of actively controlled structures. The actuators/sensors should be placed at locations so that the desired modes are excited most effectively (Lammering et al., 1994). A wide variety of optimization algorithms were proposed to this end in the literature. Two popular examples are the Simulated Annealing method (Chen et al., 1991) and the Genetic Algorithm method (Rao et al., 1991). See also Padula and Kincaid (1999) for review of the different placement strategies for sensor and actuator. Although these methods are effective, they fail to give a clear physical justification for the choice of the actuator/sensor placement. In this paper, it was chosen a more physical method that has been used by Preumont et al. (1992). It merely consists in placing the transducer in the truss structure with maximal fraction of modal strain energy which is a measure that represents the ability of a vibration mode to concentrate the vibrational energy into the actuator and can be interpreted as a compound indicator of controllability and observability of the specific mode by the actuator. Active damping of truss structures with integral control was introduced at the beginning of the 90s (Preumont et al., 1992) and has since been thoroughly studied both theoretically and experimentally (Preumont, 2002). This paper investigates numerically an integral force feedback controller for suppressing the undesired structural vibrations in a truss structure containing piezoelectric stack actuators and collocated force sensors forming a so-called smart/intelligent truss structure. It is shown that the control system consists of independent SISO loops, i.e. a decentralized active damping with local controllers connecting each actuator to its collocated force

A truss structure is one of the most commonly used structures in aerospace and civil engineering (Yan and Yam, 2002). Because it is desirable to use the minimum amount of material for construction, the trusses are becoming lighter and more flexible which means they are more susceptible to vibration. A convenient way of controlling a truss structure is to incorporate a piezoelectric stack actuator into one of the truss members (Anthony and Elliot, 2005). Research on damping of truss structures began in the late 80s. Fanson et al. (1989), Chen et al. (1989) and Anderson et al. (1990) developed active members made of piezoelectric transducers. Preumont et al. (1992) used a local control strategy to suppress the low frequency vibrations of a truss structure using piezoelectric actuators. Their strategy involved the application of integrated force feedback using two force gauges each collocated with the piezoelectric actuators, which were fitted into different beam elements in the structure. Carvalhal et al. (2007) used an efficient modal control strategy for the active vibration control of a truss structure. In that approach, a feedback force is applied to each node to be controlled according to a weighting factor that is determined by assessing how much each mode is excited by the primary source. An important feature of the truss structures control is the collocation between the actuator and the sensor. An actuator/sensor pair is said to be collocated if it is physically located at the same place and energetically conjugated, such as force and displacement or velocity, or torque and angle. The properties of collocated systems are remarkable; in particular, the stability of the control loop is guaranteed when certain simple, specific controllers are used (Preumont, 2002). It requires that the control architecture be decentralized, i.e. that the feedback path include only one actuator/sensor pair, and be thus independent of others sensors or actuators possibly placed on the structure.

2467

XVIII Congresso Brasileiro de Automtica / 12 a 16-setembro-2010, Bonito-MS

sensor. It is also demonstrated that this control problem can be formulated with the root locus approach. 2 The Truss Structure The truss structure used in this work is depicted in Fig. 1. It consists of 12 bays of 140 mm each, made of steel bars of 4 mm diameter connected with plastic joints (mass block of 70g) and clamped at the bottom. It is equipped with active member as indicated in the Fig. 1. It consists of a piezoelectric linear actuator collinear with a force transducer.
f Detail of the active member

rial stacked together. The force exerted by the active member is defined by Preumont (2002)

f a = K a ( na d33V )

(2)

where Ka is the stiffness of the actuator, d33 is the piezoelectric coefficient, V is the voltage applied to the piezo that produces an free expansion , and is the projection of the displacements at the end nodes of the active member i.e., is the sum of the free piezoelectric expansion () and the elastic displacement (fa/Ka). The elongation is linked to the structural displacement by
T = ba x

(3)

The equation governing the structure containing the active member can be found by substituting Eqs. (2) and (3) with Eq. (1); the new equation is
T & + Cx & + K + K ababa M& x x = bf + ba K a na d33V

Force transducer

fa

(4)

= nd 33V

where K is the stiffness matrix of the structure excluding the axial stiffness of the actuator. The equation (4) can be transformed into modal coordinates according to where x = ,

Piezoelectric linear actuator

= [1 2 K n ] is the matrix of the mode shapes, solutions of the eigenvalue problem: T & + K + K a ba ba M& x x = 0 . Assuming normal modes

Active Member

normalized according to T M = I and introducing & ]T , the transformed the modal state vector xn = [ equation of motion (4) becomes

&n = An xn + Bn1 f + Bn 2V x
Figure 1. Truss structure with an active member

(5)

where

2.1 Governing Equations Consider the linear structure of Fig. 1 equipped with a discrete, massless piezoelectric stack transducer. The equation governing the motion of the structure excited by a force f and controlled by the piezoelectric actuator (fa) is

0 An = K

I 0 , Bn1 = T , C b

Bn 2

and K = diag i2 , C = diag (2 i i ) , i is the i-th natural frequency of the truss and i is the associated modal damping. Similarly to Eq. (2), the output signal of the force sensor, proportional to the elastic extension of the truss, is defined by

( )

0 = T b K n d a a a 33

(6)

& + Cx & + Kx = bf + ba f a M& x

(1)

where K and M are the stiffness and mass matrices of the structure, obtained by means of the finite element model using the three-dimensional frame elements (Kwon and Bang, 1997) (each node has six degrees of freedom), C is the damping matrix; b and ba are, respectively, the influence vectors showing the locations of the external forces (f) and the active member in the global coordinates of the truss (the non-zero components of ba are the direction cosines of the active bar in the structure), and fa is the force exerted by the active member. Consider the piezoelectric linear transducer of Fig. 1 is made of na identical slices of piezoceramic mate-

y = Cn 2 xn + Dn 22V
where
T Cn 2 = K a ba 0 , Dn 22 = K a na d33

(7)

(8)

2468

XVIII Congresso Brasileiro de Automtica / 12 a 16-setembro-2010, Bonito-MS

3 Actuator Placement More than any specific control law, the location of the active member is the most important factor affecting the performance of the control system. Good control performance requires the proper location of the actuator to achieve good controllability. The active member should be placed where its authority over the modes it is intended to control is largest. It can be achieved if one locates the transducer in order to maximize the mechanical energy stored in it. The ability of a vibration mode to concentrate the vibrational energy into the transducer is measured by the fraction of modal strain energy vi defined by (Preumont, 2002)

Considering the main characteristics of both transducers as: Ka = 65 N/m and nad33 = 6.31510-7 m/Volts, the fractions of modal strain energy vi, computed from Eq. (9), are shown in Table 1 for the first six structural elements (see Fig. 2a).
Table 1. Fraction of modal strain energy in the selected finite elements.

Element 1 2 3 4 5 6

v1 (%) 0.92 4.27 0.23 1.46 3.36 0.21

v2 (%) 2.60 0.92 5.44 3.23 0.50 4.05

vi =

T iT K ababa i

T iT K + K ababa i

T K a ba i

( )
i2

(9)

The Eq. (9) is readily interpreted as the ratio between the strain energy in the actuator and the total strain energy when the structure vibrates according to i-th mode. Physically, vi can be interpreted as a compound indicator of controllability and observability of mode i by the transducer. The best location consists in placing the transducer in the truss structure with maximal fraction of modal strain energy vi, where i is the mode to be controlled. In this work, the control objective is to damp the first two modes of the structure by using two active elements. The search for candidate locations where these active members can be placed is greatly assisted by the examination of the first two structural mode shapes as can be seen in Fig. 2.

From the Fig. 2 and the Tab. 1 one can see that active member number 2 has a large influence on mode 1 and almost no influence on mode 2, and that the opposite occurs for active member number 3 i.e., substituting the active member for the bar 2 provides a strong control on mode 1 (v1 = 4.27 %), but offers a reasonable control on mode 2 (v2 = 0.92 %). By contrast, an active member substituted for the bar 3 offers almost no control on mode 1 (v1 = 0.23 %) and good controllability on mode 2 (v2 = 5.44 %). This result motivated the positions of the transducers in the actual truss. 4 Integral Controller Design Consider the truss structure with the active members of section 3. Each active member consists of a piezoelectric linear actuator collocated with a force transducer. In this section, it is considered a decentralized active damping with local controllers connecting each actuator to its collocated force sensor. Hence, assuming one pair of actuator/sensor, the response of the truss (Eq. 4) to a control input voltage (V) can be written alternatively, in Laplace form, by

[Ms + (K + K b b )] x = b K n d
2 T a a a a
13 14 15

a a 33V ( s )

(10)

10

12 11 9 8 6 5 3

where the damping matrix C and the external forces f have been omitted for simplicity. According to the integral control technique (Preumont, 2002), the collocated force sensor (y) is integrated and fed back to the control input voltage (V)
V (s ) = g y (s ) K a na d 33 s

(11)

(a)

(b)

(c)

Figure 2. a) disposition of the active elements; b) first mode shape and c) second mode shape

where g is the gain of the controller and the constant K a n a d 33 at the denominator is for normalization purpose. Note that the negative sign in Eq. (11) is combined with the negative sign of the force output (Eq. 7 or 2) to produce a positive feedback. The integral term 1/s introduces a 90 phase shift in the feedback path and
2469

XVIII Congresso Brasileiro de Automtica / 12 a 16-setembro-2010, Bonito-MS

thus damping in the system (Chen et al, 1989). It also introduces a -20 dB/dec slope in the open-loop frequency response, and thus reduces the risks of spillover instability (Preumont, 2002). Combining Eqs. (11) and (7) or (2), yields
V (s ) =
T g ba x na d33 s + g

2 g 2 T T Kababa = 0 Is + s+g

(16)

(12)

After using the definition (9) of the fraction of modal strain energy, Eq. (16) is reduced to a set of decoupled equations

Substituting (12) into (10), one gets the closed-loop characteristic equation

s 2 + i2

gi vii2 = 0 s + gi

(17)

2 g T T Kababa Ms + K + Kababa s + g
From this equation, for g 0

x = 0

(13)

For a small gain gi, one can assume a solution of (17) of the form s i i j i (Preumont, 2002), where

[Ms + (K + K b b )] x = 0
2 T a a a

i is the closed-loop structural damping ratio of mode i, this results approximately in (neglecting the terms of the second order in i )

(14)

which corresponds to the situation when the electrodes of the piezoelectric transducer (see Fig. 1) are short-circuited. The solutions of the eigenvalue problem (14) are the open-loop poles ( i - resonance frequencies). On the other hand, for g

gi vi 2i

(18)

which provides a formal justification for the placement of the active members in the places where vi is maximum, as suggested in section 3. Finally, it can be demonstrated (Preumont, 2002) that the maximum modal damping is achieved for

[Ms

+K x=0

(15)

gi = i

which corresponds to the situation where the axial contribution of the active member has been removed (corresponding to the stiffness matrix K). The solutions of Eq. (15) are the open-loop zeros ( z i - antiresonance frequencies). Figure 3 shows the evolution of the closed-loop poles when g increases from 0 to + (only one half of the locus is shown). The pole trajectories go from the open-loop poles ( i ) to the open-loop zeros ( z i ).
Im(s)

i zi

(19)

The preceding result has been established for a single active member; if there are several active members operating with the same control law and the same gain g, this result can be generalized under similar assumptions. 5 Simulations and Numerical Results Numerical simulations are presented to illustrate the efficacy of the integral controller applied to the truss structure. The structure considered is the 12-bay of 140 mm each with 111 members and 41 nodes, and the nodes at the bottom are clamped (see Fig. 1). The passive members are made of steel with a diameter of 4 mm, and the damping is assumed to be proportional to the stiffness and mass matrices so that C = 10-1M + 10-7K. At each node there is a centralized mass block of 70g and has six degrees of freedom (dof), translations and rotations in x, y and z directions, so the truss structure has 228 active dofs, and the state-space model consequently has an order of 456. The strategy is to control the first two modes (17.80 Hz and 20.61 Hz) by using two active members positioned in the finite elements shown in Fig. 2a, and two decentralized integral controllers (Eq. 12) connecting each actuator (considering Ka = 65 N/m and nad33 = 6.31510-7 m/Volts) to its collocated force sensor.

2 z2 1 z1

Structure

Integrator Re(s)

Figure 3. Root locus of the structure (for two modes) with integral control

It is readily established from the root locus (Fig. 3) that the system is unconditionally stable for every value of the gain g (Preumont, 2002). To evaluate the i-th modal damping ( i ), Eq. (13) must be transformed in modal coordinates with the change of variables x = . Assuming that the mode shapes have been normalized according to T M = I and taking into account that T T K + K a ba b a = diag i2 = 2 , yields

( )

2470

XVIII Congresso Brasileiro de Automtica / 12 a 16-setembro-2010, Bonito-MS

5.1 Controller Design In practice, it is not advisable to implement plain integral control (11), because it would lead to saturation. A forgetting factor () can be introduced by slightly moving the pole of the compensator from the origin to the negative real axis, leading to (Preumont, 2002)
gi y (s ) Vi (s ) = K a na d33 s +

4 3 2 1 0 -1 -2 -3 -4

Force Transducer 2
Open-loop Closed-loop

(20)

with lower than the first natural frequency of the structure. In this paper, is assumed be equal to 1 / 2 . Table 2 presents the open-loop poles ( i - solutions of 14) and open-loop zeros ( z i - solutions of 15), in rad/s, when either actuator 1 or actuator 2 is active.
Table 2. Open-loop poles and open-loop zeros of the structure.

Amplitude (N)

5 Time (s)

10

Figure 5. Uncontrolled and controlled responses at the force transducer 2 with impulsive disturbance forces

Actuator 1 2

Mode 1 2

i 111.76 127.92

zi 70.24 107.48

This type of force is used as it will excite many modes of vibration and hence is a difficult test for the control system. From the results it can be observed that the sensor responses are reduced greatly. Figure 6 presents the corresponding control voltages.
0.15
Actuator 1 Actuator 2

0.1

0.05

Note that zi is lower than the corresponding pole i as shown in Fig. 3. When the distance between zi and i increases, consequently the gain g and the modal damping i increase as shown in Eqs. (19) and (18). Substituting the corresponding zi and i in Eq. (19), the following results are obtained: g1 = 140.97 for the first mode and g 2 = 139.56 for the second mode. 5.2 Simulation Results To verify the controller performance numerically, open loop and closed loop simulations were conducted and the results are presented and discussed. Impulsive forces are applied in all directions on each node at the top of the structure (see Fig. 1). The uncontrolled and controlled responses of the force transducers 1 and 2 in time domain are shown in Figs. 4 and 5.
8 Force Transducer 1
Open-loop Closed-loop

Voltage (V)

-0.05

-0.1

-0.15

-0.2

5 Time (s)

10

Figure 6. Feedback control voltages applied by the piezoelectric actuators with impulsive disturbance forces.

A time-varying chirp forces f (amplitude of 10 N) from 1 Hz to 20 Hz with a target time of 3 seconds are applied in all directions on each node at the top of the structure (see Fig. 1). The uncontrolled and controlled responses of the force transducers 1 and 2 in time domain are shown in Figs. 7 and 8.
8000 6000 4000 2000 0 -2000 -4000 Force Transducer 1
Open-loop Closed-loop

Amplitude (N)

-2 -6000 -4 -8000 -6 0 1 2 3 4 5 Time (s) 6 7 8 9 10 0 1 2 3 4 5 Time (s) 6 7 8 9 10

Figure 4. Uncontrolled and controlled responses at the force transducer 1 with impulsive disturbance forces

Figure 7. Uncontrolled and controlled responses at the force transducer 1 with time-varying disturbance forces

2471

Amplitude (N)

XVIII Congresso Brasileiro de Automtica / 12 a 16-setembro-2010, Bonito-MS

3000

Force Transducer 2
Open-loop Closed-loop

References Anderson, E. H.; Moore, D. M. and Fanson, J. L. (1990). Development of an Active Truss Element for Control of Precision Structures. Optical Engineering, Vol. 29, N 11, pp. 13331341. Anthony, D. K. and Elliot, S. J. (2005). On Reducing Vibration Transmission in a Two-dimensional Cantilever Truss Structure using Geometric Optimization and Active Vibration Control Techniques. Journal of the Acoustical Society of America, Vol. 110, pp. 1191- 1194. Carvalhal, R.; Lopes Jr., V. and Brennan, M. J. (2007). An Efficient Modal Control Strategy for the Active Vibration Control of a Truss Structure. Shock and Vibration, Vol. 14, pp. 393-406. Chen, G. S.; Lurie, B. J. and Wada, B. K. (1989). Experimental Studies of Adaptive Structures for Precision Performance. SDM Conference, pp. 1462-1472. Chen, G. S.; Bruno, R. J. and Salama, M., (1991). Optimal Placement of Active/Passive Members in Truss Structures using Simulated Annealing. AIAA Journal, Vol. 29, N 8, pp. 1327-1334. Fang, J. Q.; Li, Q. S. and Jeary, A. P. (2003). Modified Independent Modal Space Control of m.d.of. System. Journal of Sound and Vibration, Vol. 261, 421-441. Fanson, J. L.; Blackwood, G. H. and Chu, C. C. (1989). Active Member Control of Precision Structures. SDM Conference, pp. 1480-1494. Kwon, Y. and Bang, H. (1997). The Finite Element Method using Matlab. CRC Press. Lammering, R.; Jia, J. and Rogers, C. A. (1994). Optimal Placement of Piezoelectric Actuators in Adaptive Truss Structure. Journal of Sound and Vibration, Vol. 171, pp. 67-85. Padula, S. L. and Kincaid, R. K. (1999). Optimization Strategies for Sensor and Actuator Placement. Technical report TM-1999-209126, NASA Langley Research Center. Preumont, A.; Dufour, J. P. and Malekian, C. (1992). Active Damping by a Local Force Feedback with Piezoelectric Actuators. Journal of Guidance, Control and Dyamics, Vol. 15, pp. 390-395. Preumont, A. (2002). Vibration Control of Active Structures: An Introduction, Kluwer. Rao, S. S.; Pan, T. S. and Venkayya, V. B. (1991). Optimal Placement of Actuators in Actively Controlled Structures using Genetic Algorithms. AIAA Journal, Vol 29, N 6, pp. 942-943. Yan, Y. J. and Yam, L. H. (2002). A Synthetic Analysis of Optimum Control for an Optimized Intelligent Structure. Journal of Sound and Vibration, Vol. 249, pp. 775-784.

2000

1000 Amplitude (N)

-1000

-2000

-3000

5 Time (s)

10

Figure 8. Uncontrolled and controlled responses at the force transducer 2 with time-varying disturbance forces

Figure 9 presents the corresponding control voltages.

150

100

Voltage (V)

50

-50

-100

-150

5 Time (s)

10

Figure 9. Feedback control voltage applied by the piezoelectric actuators with time-varying disturbance forces.

6 Conclusions An integral controller was designed and numerically implemented on a truss structure containing a pair of piezoelectric linear actuators collinear with force transducers. The procedure used for placing actuators along the structure was proven to be effective. Besides, the optimal index has a strong intuitive appeal. The integral controller, obtained by using the root locus technique, was designed to maximize modal damping of the truss structure. The control system consisted of a decentralized active damping with local controllers connecting each actuator to its collocated force sensor. A set of numerical simulations was performed, which illustrated the effectiveness of the developed controller in reducing the vibrations of a truss structure. The experimental tests however, are relegated to the future. Acknowledgments The first author would like to thank the FAPESP (N 2008/05129-3) for the financial support of the reported research.

2472