PSI - Issue 24

Riccardo Panciroli et al. / Procedia Structural Integrity 24 (2019) 593–600 R. Panciroli and F. Nerilli / Structural Integrity Procedia 00 (2019) 000–000

598

6

1. Both plate and wires are modeled at first. 2. Shell elements are killed to apply the prestrain to wires only.

3. Prestrain is applied to the wires by imposing either the displacement or a prestraining force. Both approaches are valid, but in the latter case it is necessary to apply a zero-displacement condition in the out-of-plane direction because wires are o ff seted with respect to the reference nodes of the numerical model. The nodes are in fact located on the mid-layer of the panel, while the wires lay on the outer surface. The force, which is applied to the nodes standing on the reference surface, might thus introduce a moment on the wire if such boundary conditions are not correctly applied. 4. All boundary conditions are removed, and shell elements are set to alive. 5. The central node of the panel is locked to prevent rigid motions, and temperature is decreased to the panel only. This induces the panel to attain the curved shape in one of the two stable configurations. 6. Temperature on the wires is increased above A f . Its e ff ect is to shorten the wire as a result of the phase transforma tion. The wires get in tension, and a distributed bending moment generates to the panel. 7. We increase the number of wires until the distributed bending moment is large enough to morph the panel into to the second stable configuration. Results from the simulations obtained utilizing the proposed solution scheme reveals that indeed it is possible to morph the bistable panel through SMA actuation, but at a cost of a high number of wires. As title of example we report in the following the results for an antisymmetric square panel 200mm long and 1.3mm thick. We firstly introduced a 6% prestrain to the wires (which is the maximum admissible pseudo-plastic deformation found experimentally). We then activated the plate’s elements and cooled it down by 90 degrees, and finally thermally actuated the wires to induce the snap-back. At last, the wires have been cooled down below M f . Results revealed that it is necessary to apply a set of 100 wires equally spaced along the panel to induce the snap-back.

∆ T = 0 T SMA < A s

∆ T = 10 T SMA < A s

− 1 − 0 . 5 0

∆ T = 90 T SMA > A f

∆ T = 90 T SMA < A s

∆ T = 90 A s < T SMA < A f

0

− 10

0

− 15

− 5

− 5

5

− 10

0

10

20

Fig. 5. Numerical results about the out-of plane displacement of the panel during the morphing cycle.

Figure 5 shows the out-of plane deformation of the smart panel during the whole whole morphing cycle. The panel is initially flat, being the composite at the curing temperature and the prestrained wires below A s . As the temperature decreases, simulating the post-curing cooling phase, the panel assumes a stable configuration which is almost cylindrical. Please note that, di ff erently than the case with the sole composite plate, the SMA wires o ff er an increased bending sti ff ness which results in a shape which slightly di ff er from the cylindrical one. The e ff ect of the morphing cycle on the wire can be observed in Figure 6. The SMA wires yields (I) during the initial prestraining cycle (II); the wire then shortens due to the elastic spring-back as the prestrain force is removed