Home
Search
Collections
Journals
About
Contact us
My IOPscience
A new active composite
This content has been downloaded from IOPscience. Please scroll down to see the full text. 2009 Smart Mater. Struct. 18 025020 (http://iopscience.iop.org/0964-1726/18/2/025020) View the table of contents for this issue, or go to the journal homepage for more
Download details: IP Address: 140.117.111.1 This content was downloaded on 22/08/2014 at 13:04
Please note that terms and conditions apply.
IOP PUBLISHING
SMART MATERIALS AND STRUCTURES
Smart Mater. Struct. 18 (2009) 025020 (7pp)
doi:10.1088/0964-1726/18/2/025020
A new active composite H Drobez1, G L’Hostis2,3 , K Buet Gautier2 , F Laurent1 and B Durand2 1
CETIM CERMAT, BP 2278, F-68068, Mulhouse Cedex, France Laboratoire de Physique et de M´ecanique Textiles, CNRS UMR 7189, 11 rue Alfred Werner, F-68093, Mulhouse Cedex, France
2
E-mail:
[email protected]
Received 3 October 2008 Published 22 January 2009 Online at stacks.iop.org/SMS/18/025020 Abstract CETIM CERMAT and the LPMT have been developing a new kind of active structure over the past few years: CBCM (controlled behavior composite material). The CBCM process consists in generating an internal source of heating within the composite structure and then in using the thermomechanical properties of the various components in order to deform the material. Carbon yarns are used as the internal heating source: being connected to a power supply, they are conductive and provide heating by the Joule effect. In this work, various aspects of the CBCM are studied by means of model plates. First, the plates and their constitutive layers are described. Second, the bending properties of the CBCM are presented. Third, it is shown how the structure can become active. Finally, in order to illustrate the capacities of this new active composite, a prototype of an aerodynamic flap integrating CBCM is developed. Experimental and numerical results are compared. (Some figures in this article are in colour only in the electronic version)
or adhesion. The CBCM (controlled behavior composite material) is one solution to these problems as the whole structure is active [9]. It works by using the anisotropic character of the thermomechanical behavior of the composite; then it just needs a variation in temperature. The anisotropic character should be well organized to be efficient. This principle has already been used: then the source of heating was either external or internal [10, 11]. In the CBCM, in order to prevent any problems of adhesion, only internal sources of heating are used by means of composites’ ‘natural’ materials. Carbon, which is a good electrical conductor and is traditionally used as fiber reinforcements, can play the role of internal source of heating. Connected to a generator (DC or AC), and thanks to their electrical properties, the carbon yarns are heated by the Joule effect. Then, the temperature rises either in a part of or the whole structure, and this leads to the deformation of the composite. As the amount of current will influence the internal temperature it will also allow control of the deformation of the composite structure. The aim of this paper is to make a general presentation of the CBCM. First, the CBCM principle is detailed, as well as the materials used to make it. Second, several results of bending tests are described to show the various performances of the structure. The third part of this work is the command
1. Introduction Smart materials can be defined as materials which are able to change their form, their mechanical, or any physical properties to adapt to their background. They are increasingly used for technical applications and particularly in the field of controlling structural deformations. In this case, preferential actuators are piezoelectric materials, shape memory alloys, or even electrostrictive/magnetostrictive materials. This list can be completed by thermal actuators: their principle is based on thermal elongation coupled to structure effects by using different materials or effects of structural instability. The use of one of the previous elements to make a composite become active can be done at several scales: at the scale of the fiber or the matrix [1–3], at the scale of the interface between the successive layers in a laminate [4–6], or at the scale of the linkage between the actuator and a passive composite structure [7, 8]. These elements or actuators are generally fixed on the surface of the structure or integrated within it. However, this may increase the weight and the volume of the structure, and can generate some problems of stress concentration 3 Author to whom any correspondence should be addressed.
0964-1726/09/025020+07$30.00
1
© 2009 IOP Publishing Ltd Printed in the UK
Smart Mater. Struct. 18 (2009) 025020
H Drobez et al
Figure 1. How the CBCM works.
Table 2. Properties of the plates.
Table 1. Characteristics of the layers. Reinforcement
Thickness (mm)
Estimated volume fiber fraction (%)
TVa MVb CAc Cd Me
0.25 0.25 0.1 0.1 5
41.7 23.6 23 22.8 —
Plate
Dimensions (mm) Weight (g) Resistance ()
Thin laminate 500 × 100 × 1.6 Sandwich 500 × 100 × 6.7
121 182
7.7 7.8
appears within the structure. The result is, as previously shown, the bending of the plate (figure 1).
TV = unbalanced 270 g m−2 glass fiber fabric (unbalance coefficient = 0.6). b MV = 300 g m−2 glass fiber mat. c CA = active layer of carbon yarns. d C = not connected carbon yarns. e M = PVC foam. a
2.2. The model plates To show the efficiency of both single and double effects, some model plates are made as described in the following. A thin laminate is used to demonstrate the capability of the CBCM concerning the single effect and a sandwich plate refers to the double effect. The characteristics of the various layers used are shown in table 1. In abbreviated form, the layouts are, for the thin laminate [TV2 /CA/MV4 ] (the number following the letter is the number of consecutive layers), and for the sandwich plate [MV2 /TV/C/M/CA/TV/MV2 ]. Except for the sandwich core, the plies are exactly the same for the two plates. The only difference is the order of the layout. Note, also, that for all the fabrics, the warp direction is along the length of the plate. The carbon yarns are used as an internal source of heating (the ‘active layer’). The reinforcements are impregnated with a polyester resin (Synolite 1717) whose glass transition temperature is 80 ◦ C. The plates are polymerized at room temperature for 24 h and then cured in an autoclave at 90 ◦ C for 4 h. The ‘active layer’ of carbon is connected to a DC power generator. The geometrical and electrical properties of the plates are given in table 2. These resistances can be compared to the 0.25 of Asanuma’s active composite [11], for a laminate (60 mm × 40 mm × 0.7 mm), made of one aluminum plate and two layers of carbon fiber reinforced plastics, and connected to a power supply.
description of the active structure. Finally, before concluding, an example of application is presented.
2. Materials and structure 2.1. How the CBCM works As it has been said before, a plate can warp or bend due to a rise in temperature under some conditions. In classical composite structures, this well-known effect is avoided as often as possible, particularly by using symmetric layouts. Two different manners to have a bending effect by heating the CBCM have been identified: one called the ‘single effect’ or ‘temperature effect’, and the other called the ‘double effect’ or ‘gradient effect’. The character ‘single’ or ‘double’ is relative to the possibility for CBCM plates to bend in one or two directions. The single effect is obtained thanks to a laminate made of layers with different coefficients of thermal elongation in a given direction. To work as indicated before, the laminate has to be asymmetric (see figure 1). The different coefficients can be obtained either by different materials or by different orientations of the reinforcement in the layers. The double effect can be obtained with any composite structure containing an insulating layer, for example a sandwich structure. If only one side of the composite is heated, the other side being insulated, a gradient of temperature
3. Performances of the CBCM A three-point bending test is chosen to characterize the behavior of the composite structure. The plates are supported 2
Smart Mater. Struct. 18 (2009) 025020
H Drobez et al
Table 3. Experimental results. Unit Thin laminate Sandwich Thickness (e) Free deflection ( f L ) Time to stabilize (ts ) Virtual load or equivalent weight ( Pv ) Mass (m ) Ratio equivalent weight/proper weight Mechanical work ( T ) Mechanical power ( T/ts ) Electrical power Power yield
mm mm s N
1.6 15 420 4.4
6.7 3.7 195 33.7
g
121 3.6
182 18.5
J W W %
0.06 1.42 × 10−4 57 2.91 × 10−4
0.12 6.4 × 10−4 53.4 1.12 × 10−3
Figure 2. Typical behavior of a thin laminate plate of CBCM.
on the lower side (the one which is insulated from the active layer). In order to compare their behavior with that of other active materials it is necessary to calculate the mechanical work and power of the structure [13]. So, the plates are first calibrated and their flexural rigidity ( K in N m−1 ) is determined by a three-point bending test under load but without connecting the carbon yarns to the generator. This value is used to calculate a ‘virtual load’ Pv (1) which can be associated to the maximum of free deflection f L . It is an equivalent weight which produces the same deflection as a 2.6 A current, when the plate is connected but free of load.
Pv = K × f L .
(1)
The mechanical work W is defined as Figure 3. Typical behavior of a sandwich plate of CBCM.
T = Pv × fL
(2)
and the mechanical power is defined as the ratio of mechanical work to the time when the displacement and temperature are stabilized. The electrical power is calculated from the current and the voltage. Table 3 summarizes the experimental results for the two plates. It is worth noting that the electrical powers are quite similar but the time before stabilization, the mechanical powers, and the yields are very different. Because the sandwich structure is more rigid, the deflection is smaller and it is able to lift an heavier load. Moreover, the thin laminate activates what we call the ‘temperature effect’ and the sandwich plate activates the ‘gradient effect’. Now, to compare to classical smart materials, the mass mechanical work is calculated. The various elements of comparison are summarized in table 4. If the energy used to activate the structure and the response time are considered, the characteristics of the CBCM are close to those of the shape memory alloy (SMA) Nitinol material. If the mass mechanical work and the way it operates are considered, they are similar to those of piezoelectric PZT materials. Although the mass mechanical work for the CBCM is the lowest, it is important to consider that the CBCM is a complete structure and can work independently, whereas the other materials have to be inserted into more or less heavy structures. In the same way, it is difficult to give some values for strain: the CBCM is
by two rigid cylinders placed 500 mm apart. Because of their asymmetry, the thin plates are not flat, but this state is defined as the initial state of the test. When the power is turned on, the current is circulating in the carbon yarns and the temperature rises in the plate, which is bending. The deflection from the initial state is measured thanks to an LVDT displacement sensor placed in the middle of the plate. Apart from its own weight there is no load applied to the plate [12]. The temperature of the room is 20 ◦ C. The temperature at the surfaces of the plate is measured thanks to thermocouples placed on each side. The typical behavior of CBCM plates is shown in figures 2 and 3, where the deflection and the temperature on each side are plotted versus time. These show that the thermal and mechanical phenomena stabilize after a while. To establish these curves the active layer was connected to the generator and the DC current was 2.6 A (0.28 A in each of the nine active yarns), the same for the two plates. The maximum of temperature and deflection (free deflection f L because there is no load) can be deduced from the curves. In the middle of the thin laminate the maximum deflection is 15 mm when the temperature reaches 57 ◦ C on each side. In the middle of the sandwich structure the maximum deflection is 3.7 mm when the temperature reaches 70 ◦ C on the upper side (the one near the active layer) and 42 ◦ C 3
Smart Mater. Struct. 18 (2009) 025020
H Drobez et al
Table 4. Comparison between CBCM and other smart materials [13].
Response time Energy used to activate Precision Mass mechanical work per cycle (J kg−1 ) Working mode Temperature range a
SMA Nitinol
Piezoelectric PZT
Thin laminate
Sandwich
Slow Temperature Low 247
Quick Electricity High 1.42
Medium-slow Temperature Good 0.495
Medium-slow Temperature Good 0.65
Linear-static 80 ◦ Ca
Linear-static 80 ◦ Ca
Binary-static up to 400 ◦ C
Linear-dynamic −20 to 250 ◦ C
Depends on the glass transition temperature of the resin. A security coefficient has to be applied.
Figure 5. Three-point bending test.
Figure 4. The CBCM plate equipped with sensors.
bending while the other smart materials are stretching, and what is important for the CBCM is the deflection. A limitation for the use of the CBCM is the glass transition temperature of the resin: 80 ◦ C here. In order not to damage the plates, a security coefficient was applied and the current was chosen not to exceed 70 ◦ C (to compare the two plates, the same current was used). Finally, all the characteristics are totally dependent on the constitution of the plates and, as an advantage, the latter can be modified to adapt the response of the composite. For example, by choosing different materials and different manners to lay them out, the mass mechanical work of some other plates we made reaches 20 J kg−1 [14].
Figure 6. Strain versus time during the experiment.
The aim of the experiment is to maintain a given strain, whatever the load. The carbon yarns are connected and the strain command is 300 μm m−1 . Note that because we speak about single effect the plate has to be mounted in such a manner that the deflection due to thermal effect appears at the opposite direction to the deflection due to the applied load. So when the 50 N load is applied (t = t0 ), the command sets on the regulation system and the temperature rises in the plate to cause a 300 μm m−1 strain. This value is reached after 7 min. If the load is now zero (t = t1 ), the strain becomes instantly higher than the command value, so the control system switches off the power. The temperature in the plate naturally decreases until the command strain is reached. Then the temperature is maintained to keep the right strain. This experiment can be done as often as required. The relative curve is shown in figure 6. The time necessary to reach the command value seems to be long (7 min) but it depends on many parameters that can easily be changed. On the one hand, there are the construction parameters such as the constitution of the plate; on the other hand, there are the working parameters such as the power supplied. Our current studies show that a higher current intensity (a working parameter) leads to a better time
4. Smart material As was presented before, the CBCM is an active material that can bend when heated. The ‘smart’ character of the structure comes from a control system made of sensors and a proportional–integral–derivative (PID) command. The PID regulation system is peculiar to the software Testpoint. The only possibility is to set the three regulation parameters ( P , I and D ). Those parameters were manually set. By adding temperature sensors and strain gauges to the CBCM (see figure 4), the system can detect variations in its background and then react to adapt to the new context. We can then speak of ‘smart structure’ considering the whole concept (CBCM with sensors and command). As said before, one of the interests of the CBCM is its being designed on demand to provide the desired properties. Here, to highlight the smart character of the structure, a thick laminated plate is made. This new plate is 7.45 mm thick and the equivalent load Pv , determined as previously, is 770 N. As shown in figure 5, a strain gauge is fixed in the middle of the lower side of the plate and a 50 N load is applied on the upper side. 4
Smart Mater. Struct. 18 (2009) 025020
H Drobez et al
Figure 7. Geometry and cinematic of existing flap.
Figure 8. Cinematic of active composite flap.
opening : release of the flap
flap under stress and locked
closing : heat
heat and flap locked
closing : stress
Figure 9. Modified opening/closing cinematic of the flap.
of response. Another parameter, the density of active yarn per unit area, which is a construction parameter, is under study.
The aim of this study is to suppress all the link elements and the jack by the use of a single active composite structure (figure 8). For this application, the response time of the CBCM, and the influence of external parameters such as convection phenomena, can constitute obstacles for the use of a CBCM structure. So, the closing/opening cinematic of the flap has been modified (figure 9): the activation of the composite flap does not open it but close it. Moreover, the flap is molded in its deformed shape; thus, when the flap is closed, it is under stress. Using the specifications of the former flap, the active composite flap is predesigned with simplified calculus tools in order to define the kind and the thickness of reinforcements
5. Example of active structure The example we choose to illustrate this work is a retractable aerodynamic flap. It is used on a car to stabilize the vehicle at high speed. The existing flap is made with a thermoplastic matrix and carbon fibers, and is moved by means of a jack (figure 7). There are three rotational joints combined to the jack (its associated hydraulic system is not represented) to realize the opening/closing cinematic of the flap. The flap is 280 mm long, 360 mm wide, and the running opening of the flap is 40 mm. 5
Smart Mater. Struct. 18 (2009) 025020
H Drobez et al 45
140
T (°C)
d (mm) 40
120
Tac
Tnac Tac Tac fe Tnac fe d exp d fe
100 80
35 30 25 20
60
.
A
Tnac
15 40 10
dexp
20
5
time (s)
0 0
500
1000
1500
2000
2500
0 3000
Figure 11. Experimental results: temperatures and displacement versus time.
Figure 10. Flap prototype.
to be used, and to quantify the working parameters such as electrical input power. The layout of the calculated laminate is the following: 1 aramid fabric (thickness: 0.4 mm, 230 g m−2 ), 1 glass unidirectional fabric (UD) at 0◦ , along the longitudinal direction of the flap (thickness: = 1.1 mm, 650 g m−2 ), 1 carbon UD at 90◦ (thickness: = 0.25 mm, 160 g m−2 ) and 2 glass UD at 90◦ (thickness: 1.1 mm, 650 g m−2 ). The active layer is between the aramid layer and the first UD layer (thickness = 0.05 mm). The total thickness of the flap is 4 mm. A prototype of the flap was realized (figure 10); it is composed of an active area (dark) and a non-active area (white). The resin is an epoxy resin, Epolam 2025 from Axson (the glass transition temperature is Tg = 135 ◦ C). The structure has been cured according to the manufacturer recommendations. Figure 11 shows, during a cycle of opening/closing (950 s of heating and 1500 s of cooling), the temperature Tac of the active area (with a maximum of 125 ◦ C), the temperature of the non-active area Tnac (with a maximum of 31 ◦ C), and the deflection at the end of the flap dexp (with a maximum of 39 mm). This result is obtained for a supplied power of 180 W. An unsteady thermomechanical calculation using finite elements has been performed on the final prototype (Abaqus software). The part was modeled with its initial curvature. The properties of each layer was obtained by homogenization. The Joule effect in the active layer is taken into account via an internal heating source φ (the hetval subroutine was used). Because φ = Pe /(Sac eac ) ( Pe is the electrical power supplied, Sac and eac are, respectively, the area and the thickness of the active layer) then φ = 0.06 W mm−3 . The finite elements are the S8RT elements of Abaqus: thermally coupled quadrilateral general thick shell, with biquadratic displacements and bilinear temperature in the shell surface. The anisotropic behavior of the layers is defined in local anisotropic coordinate systems that follow the structure deformation. The convection coefficients on and under the flap are taken to be equal to h = 12 W(m2 ◦ C)−1 . In figure 10, good accordance can been seen between the experimental and the numerical results concerning the values of displacement and temperature at the end of the test. There
is a difference between the results in terms of kinetics. This difference can come from the homogenized characteristics of the layers, and the thermomechanical properties of each constituents, particularly those of the reinforcements, being taken from the literature and not from measurement. This example highlights the interest of the CBCM compared to other actuators; this means its capability to generate important efforts and to deform thick composite structures. The virtual load Pv to be applied at point A to have a maximal deflection of 38.7 mm is 37 N.
6. Conclusions and prospects In this first study, a CBCM has demonstrated that it could be of great interest in the field of active control of materials. Considering its time response and its activation mode, it can be compared to an SMA. Unless they can develop high loads, SMAs have small precision and their behavior is binary. Those limitations do not exist with the CBCM. Moreover, it has an actual load that could be in the same order or even higher than those of piezoelectric actuators. Finally, in many cases, it could be profitable to substitute a given smart material by a CBCM, provided that the reaction speed is not a problem. The CBCM becomes smart when provided with sensors and a control system. The last section of this study shows an example of application of a CBCM. Nevertheless, this first study on such a structure has also highlighted some problems which still have to be solved, and many aspects remain to be studied.
• The first generation of connections between the plates and the generator was of bad quality and led to damage into the plates. The solution is to insert small trips of copper within the composite and to join the carbon yarns to the trips. • The rigidity of the plates had been measured at room temperature, but it should be different when the temperature is higher. Some experiments are in progress to characterize the evolution of the rigidity with the temperature and to measure a blocking load. 6
Smart Mater. Struct. 18 (2009) 025020
H Drobez et al
• The level of the current intensity sets the maximum value of deflection and temperature. The only limit in temperature is the glass transition temperature of the resin. By changing the resin (epoxy instead of polyester, for example) it is possible to have better performances. The influence of the current level on the behavior of the CBCM is a study in progress. • Some problems of local overheating have appeared during the tests. These are linked to an inefficient thermal conduction of the resin. At the moment, two solutions are under study in order to ameliorate this point. It is possible to optimize the disposition of the active yarns in the composite to have a better distribution of the heat. It is also possible to add some elements that play a role in the heat transfer but do not have a part in the electric conduction: for example, some carbon nanotubes can be used.
[4] Zhang C S and Ni Q Q 2007 Bending behaviour of shape memory polymer based laminates Compos. Struct. 78 153–61 [5] Barrett R and Gross R S 1996 Super-active shape-memory alloy composites Smart Mater. Struct. 5 255–60 [6] Elzey D M, Sofla A Y N and Wadley H N G 2005 A shape memory-based multifunctional structural actuator panel Int. J. Solids Struct. 42 1943–55 [7] Yoon K J, Park K H, Park H C and Perreux D 2003 Thermal deformation analysis of curved actuator LIPCA with a piezoelectric ceramic layer and fiber composite layers Compos. Sci. Technol. 63 501–6 [8] Schultz M R, Hyer M W, Williams R B, Wilkie W K and Inmann D J 2006 Snap-through of unsymmetric laminates using piezocomposite actuators Compos. Sci. Technol. 66 2442–8 [9] Drobez H, L’Hostis G, Laurent F, Durand B and Meyer G 2006 Controlled behavior of composite material (CBCM) with a conductive active layer JEC Mag. 22 66–9 [10] Wetherhold R C and Wang J 1996 Controlling thermal deformation by using laminated plates Composites B 27 51–7 [11] Asanuma H, Haga O, Ohira J I, Hakoda G and Kimura K 2002 Proposal of an active composite with embedded sensor Sci. Technol. Adv. Mater. 3 209–16 [12] Drobez H 2006 Mat´eriaux composite a` comportement contrˆol´e PhD Thesis Universit´e de Haute Alsace [13] Monner H P 2005 Smart material for active noise and vibration reduction NOVEM: Proc. Noise and Vibration Emerging Methods (Saint Raphael, April 2005) [14] Drobez H, L’Hostis G, Laurent F, Durand B and Meyer G 2006 Thermal ageing of an active composite material with an internal source of heat Proc. 12th European Conf. on Composite Materials (Biarritz, Aug.–Sep. 2006)
References [1] Chung D D 1998 Self monitoring structural materials Mater. Sci. Eng. R 22 57–78 [2] Psarra G C, Parthenios J and Galiotis C 2001 Adaptative composites incorporating shape memory alloy wires J. Mater. Sci. 36 535–46 [3] Donadon M V, Almeida S F M and De Faria A R 2002 Stiffening effects on the natural frequencies of laminated plates with piezoelectric actuators Composites B 33 335–42
7
