1. Introduction
A piezoelectric actuator is a device used to convert electric energy into a displacement of mechanical structure or into a force acting of selected structure. The basic part of such an actuator is made from piezoelectric material in which electric energy is converted into strain of this material. Mohith et al. [
1] enumerated five groups of traditional structure of piezoelectric actuators: unimorph, bimorph, tube, multi-layer and amplified. The actuators from the first two groups are most often based on beam which is only fixed at one end. The actuator, which the mechanical structure is based on, has been called a piezoelectric cantilever actuator for several dozen years, e.g., in [
2].
The piezoelectric cantilever actuator consists of a clamped-free beam, a voltage amplifier and a control system. The cantilever beams are most often constructed from one layer of non-piezoelectric substrate and one or two layers of piezoelectric material. Non-piezoelectric and piezoelectric materials are attached to each other by gluing. Initially, PZT ceramics [
3] and PVDF polymers [
4] were applied in the construction of this type of actuator. Since the beginning of the twentieth century, the rapid development of research on the control of cantilever actuators, built on the basis of Macro Fiber Composite (MFC), can be noticed. Some of the first research results were presented by Azzouz et al. [
5] in this area. In the years that followed, many researchers pointed out advantages of piezoelectric actuators based on MFC in comparison to actuators based on piezoceramics, e.g., Nguyen et al. [
6] noticed that the MFC have a better dynamic actuation than the bulk PZT type for range of high frequency.
Macro Fiber Composite (MFC), invented by NASA and commercialized by Smart Material Corp, is a material that consists of piezoceramic fibers with rectangular cross-section, non-piezoelectric epoxy filling the gaps between the fibers, copper electrodes adjacent to both sides of the fibers and polyamide kapton holding the structure together. Dimensions of fibers, epoxy layers, electrodes and polyamide kapton are known. P1 and P2 are two basic types of MFC: the first is dedicated to actuators and the second is used in the energy harvesting process. Structures of these types are varied in terms of the arrangement of the electrodes [
7]. MFC P1 is equipped with interdigitated electrodes which gather both positive and negative charges on each side of piezoelectric fibers. A direction of polarization and a direction of stress in the beam are the same thanks to the application of such an electrodes arrangement [
8]. The MFC P1 composite was modeled by many scientists. Williams et al. [
9] presented for MFC patch the nonlinear tensile and shear stress–strain behavior and Poisson effects using various plastic deformation models. Deraemaeker et al. [
10] proposed mixing rules for the determination of longitudinal and transverse piezoelectric coefficients of MFC patches. Zhang et al. [
7] presented the FE model which is based on the Reissner–Mindlin hypothesis using linear piezoelectric constitutive equations. Steiger et al. [
11] present a FE model, which was used for the effective computation of material constants of MFC. Emad et al. [
12] proposed a model of MFC actuator based on the replacing of MFC actuator with an equivalent simple monolithic piezoceramic actuator using two electrodes only. Research is also conducted in the field of optimization of the localization of MFC in the actuator structure, e.g., Padoin et al. [
13] proposed an optimum localization of the MFC in beam structure in order to maximize a controllability index. The cantilever actuator structures, containing both MFC and substrate, are most often modeled by the use of a linear motion equation, in which a tip displacement is only included, e.g., [
14], or by the use of a linear matrix motion equation, e.g., [
15]. The linear equations are expanded by descriptions of nonlinear phenomenon: hysteresis and creep, which appear in a control system of the actuator containing MFC. Hysteresis is modeled by the use of models applied for the actuators based on PZT, e.g., Yang et al. [
14] proposed the Bouc-Wen approach and Xu et al. [
16] applied the Prandtl-Ishlinskii model. The creep is also modeled by the use of approaches applied to describe piezoelectric ceramics, e.g., Schrock et al. [
17] presented the application of the Prandtl-Ishlinskii model for the creep in actuator based on MFC.
Deflection control systems, which contain the cantilever beam actuator based on MFC, A/D board, a control algorithm implemented in a dedicated computer program, and a voltage amplifier, are also intensely researched. A selection of the A/D boards is not dependent on the kind of used piezoelectric material, so A/D boards are the same for the actuators based on PZT and actuators based on MFC. Control algorithms used for actuators based on PZT are also used for the actuators based on MFC: algorithms based on a linear function, e.g., PID [
18], algorithms based on a state space, e.g., LQR [
15], algorithms based on fuzzy functions, e.g., [
19] and algorithms based on genetic methods, e.g., [
20].
The significant difference in operation between the control system for actuator based on MFC and the control system for actuator based on PZT ceramics is apparent in the required control voltage values for these actuators. The control system for actuator based on MFC requires significantly higher control voltage in comparison to control system for actuator based on PZT ceramics. The initial depolarizing field for typical PZT ceramic (PZT-5A) is around 500 kV/m [
21], which causes the maximum control voltage to be ±150 V for the thickness of the PZT layer equal to the thickness of the MFC layer (0.3 mm). A reducing the thickness of the PZT layer reduces the value of the maximum voltage, e.g., for 0.267 mm it will be ±134 V [
22]. The maximum values of control voltage are significantly larger for MFC patch: maximum operational positive voltage is equal to +1500 V and maximum operational negative voltage is equal to −500 V [
23]. Higher control voltage generally leads to higher control cost and lower economic efficiency [
24]. In addition, it may lower the stability and reliability of the control system in practice because the piezoelectric actuator has the risk of being destroyed by too high control voltage [
24]. The higher control cost mainly results from the necessity to use amplifiers that generate higher voltages. Nowadays, the amplifier generating maximum voltage of 1500 V is offered only by Smart Materials Corp. and is significantly more expensive than some of amplifiers of maximum voltage about ±200 V, which are on offer by other companies.
A deflection increase of the piezoelectric actuator (for a given carrying substrate) can only be achieved by an increase of control voltage [
24]. Hence, the biggest deflection of cantilever actuator based on MFC may be obtained for +1500 V for each carrying substrate. The actuator containing two MFC patches can achieve a maximum deflection, whose value is the same as for control voltage equal to +1500V, for the use of lower control voltage than +1500 V. However, the application of two MFC patches causes a change of control system because the actuator with two MFC patches has two control inputs. Kumar et al. [
25] presented an algorithm which generated one control signal for two PZT patches: for one patch this control signal was positive and for the second it was negative. A similar approach is presented by Jain et al. [
26]. Kaci et al. [
27] proposed another approach in which two control signals were generated, but one of these signals was a real part of the external force and the second was an imaginary part of this force. Control algorithms, presented in [
25,
26,
27], enable effective control assuming that the cantilever beam is symmetrical so it can be treated as a SISO control object. A prismatic cantilever beam with two MFC patches is symmetrical if one MFC layer is glued on exactly the same as the second MFC layer and if both glued connections have the same properties, e.g., a contact area. In practice, meeting these conditions is difficult.
The novelty of this article is a control algorithm which generates two control signals which are independent of each other. In contrast to the literature approaches presented, the cantilever piezoelectric actuator containing two MFC patches was treated as a MISO (multi input and simple output) control object which allows for effective control regardless of whether both MFC layers are symmetrically glued on both sides of the carrying substrate layer. Obtaining the required control quality was realized by the use of two integrating action, which can be created independently for each control signal. A comparison of the control quality of piezoelectric actuator with one input and the control quality of piezoelectric actuator with two inputs was not presented before in the literature.
6. Conclusions
The control system of beam actuator, based on Macro Fiber Composite, was tested in laboratory and numerical research. The tested beam actuator contained PCB FR-4 substrate and two MFC patches.
An application of two control signals in the beam actuator led to a reduction of voltage, which controls this actuator, compared to the control voltage of actuator with one control signal. However, the application of two control signals does not lead to a two-fold reduction of the control voltage. The decrease in the maximum voltage was from 30.22% to 39.39% in conducted laboratory experiments.
An application of a control system with two control signals causes a reduction of the overshoot value compared to the system with one control signal. The decrease in overshoot occurred in eight out of nine experiments. The settling time was shortened in two experiments for two-input actuator in comparison to one-input actuator; in the remaining seven experiments, this time did not change significantly.
An application of limit of maximum control voltage leads to a greater decrease of the control quality for one-input actuator compared to two-input actuator. In the laboratory experiments, it was most influenced by the increase of overshoot, which was much greater for one-input actuator compared to two-input actuator.
The advantage of the presented control system is the possibility of generating two control signals, which are independent of each other. Thanks to this, it is possible to generate two control signals that are different from each other, which may be useful in the case of lack of symmetry in the structure of the actuator. The lack of symmetry may be caused by, e.g., unequal gluing of both MFC patches on a carrying substrate. The generation of two control signals that are different from each other will be the subject of future research.