Active vibration isolation of electronic components by piezocomposite clamped–clamped beam

Abstract

The sensitive electronic components used in military and aerospace applications endure some intense vibrations. These vibrations have some disturbing effects on the stability and on the service life of these devices. So, protecting these elements becomes a major economic and strategic stake. Vibration isolation can be applied to different levels of the on-board systems. Indeed, it is advisable to isolate electronic components either at the rack level or at the board level or at the component level. In this paper, the last solution is chosen because of low moving masses which imply low control energies.

An active suspension system is located between the host board and the sensitive element to be isolated. This designed control system uses a simple Integral Force Feedback strategy. This vibration isolation control is stable for its collocated version and does not need a numerical model of the system to be controlled. Robustness of the system is asymptotically guaranteed. The proposed isolation device, made of alumina for passive structure and made of PZT and PVDF for transducing layers, is experimentally tested. Experimental performances are compared with theoretical performances.

Introduction

Vibration isolation is necessary in two broad classes of problems:

• A vibrating element is fixed on a structure. Mechanical waves propagate through this whole structure. Thus, they can damage the different sensitive elements of the structure or reduce their service life.

• A sensitive element is fixed on a vibrating structure. So, vibrations can modify operating points of this element but also strongly damage it.

The passive solution is the simplest way to achieve vibration isolation. Several passive techniques are studied in the literature by using elastomer materials [1], by using shape memory alloys [2], [3] or by modifying mechanical impedances [4].

The passive suspension, sketched in Fig. 1, is considered.

The transmissibility of the system, i.e. the relationship between the acceleration of the mass (${\ddot{W}}_s$) and the acceleration imposed to the support (${\ddot{W}}_u$) is written in Laplace's variables:

$$T_{{\ddot{W}}_s,{\ddot{W}}_u}=\frac{(\ddot{W}{)}_s}{(\ddot{W}{)}_u}=\frac{s·C+K}{s^2·M+s·C+K}$$

where K is the stiffness of the suspension (N m⁻¹), M the mass of the structure (Kg), C the damping coefficient induced by the suspension (Kg s⁻¹) and s the Laplace's variable.

The objective of any suspension is obviously to limit the acceleration of the system to be isolated in the excitation frequency bandwidth. The behavior of the system is observed in Fig. 2, where different transfer functions are plotted for different damping ratio values $\xi =(C/2)\sqrt{(1/\mathit{MK})}$.

As shown in Fig. 2, this system presents a modal resonance peak that is to say an increasing of the acceleration of the system to be isolated in a narrow frequency bandwidth. Obviously, the modal peaks decrease when the damping ratio of the system increases. Consequently, the resonance amplification decreases and the sensitive mass stability increases in the narrow frequency bandwidth around the considered eigenfrequencies. However, this effect also results in the reduction of a high frequency filtering decay rate. Indeed, this high frequency decay rate evolves from −40 dB/decade, when no structural damping is considered in the suspension, to −20 dB/decade limit for high structural damping ratios, $\xi$. Thus, it results a bad isolation in high frequency range. Then, it becomes necessary to lower the cut-off frequency to increase the high frequency isolation capability of the suspension. This modification is performed by limiting the stiffness of the connection what involves a loss of stability in low frequencies. Ultimately, the traditional mechanical compromise is in the ratio between the stiffness and the damping ratio of the connection [5].

A best mechanical compromise can be obtained by using an active control process [5]. Sky-hook control is the simplest and the most common control strategy [6]. This isolation method can be stable and robust in its collocated version. The idea is to actively introduce, in the mechanical connection, a viscous damper rigidly fixed to a Galilean coordinate system. The design of this controller is based on an absolute sensing signal such as the acceleration, the transmitted force, the absolute velocity or the absolute displacement. Studies of various sky-hook-type strategies are broadly proposed in the literature [7].

The constant miniaturization of electronic components essential with electronic boards such as the frequency generators, the vibrating gyroscopes and certain accelerometers generates two major difficulties. On the one hand, the size of the soldered connection points is strongly reduced and so the yield strength of these soldered connection points is reached. On the other hand, the dimension reduction of these elements fatally involves the reduction of their useful mass. Then, these electronic components become very sensitive to the external noise and become inefficient and inaccurate. Consequently, the measurement accuracy is lost in signal noise level.

There are various methods to avoid these limitations. Indeed, it is possible to stabilize the sensitive components at level of the racks containing carrying electronic cards. Generally, this solution is obtained by using passive isolation strategies. A second solution involves stabilizing at level of the boards. It is possible to apply some active isolation strategies by using piezoelectric patches [8], [9], [10]. The last solution is to control the electronic components themselves. This paper examines this method. The principal idea is to create “small isolation islands” to individually isolate each sensitive component [11]. The control electronics already presents onto the electronic boards is exploited. The isolation structure will be integrated into the design phase of the electronic circuits. The used control law is founded on a classical collocated “sky-hook” strategy with an internal sensor.

The objective of this study is to design, manufacture and test a isolation meso-structure, so as to prove an active vibration isolation feasibility for a vibration isolation device. The paper is organized as follows. Section 2 describes the operation of the studied active suspension, the experimental device and the way to manufacture it. In Section 3, modeling of the active suspension is provided and an experimental characterization is achieved. This section is a base for the numerical development of the control architecture in Section 4. In the following section, experiments are achieved to clearly set the validity of the control strategy and, thus, of our designed active suspension. These experimental results are compared with the numerical ones. Finally, concluding remarks are discussed.