Page 110 - MSAM-3-4
P. 110

Materials Science in Additive Manufacturing                           Bistable 3D-printed compliant structure



              As indicated above, noticeable differences could be   3.1.2. Partial bi-stability of structures in Group 1
            identified between the effective stress-strain curves for   (l’ = 30, g’ = 1)
            the 1  cycle and all the rest cycles. The area between the   Responses of the structures in Group 1 under quasi-static
                st
            loading and unloading curve indicates the dissipated   compression tests  were  captured  under  two  conditions,
            energy by the structure within one cycle.  It could   that is, only the bottom surface taped to the compression
                                                 41
            be observed from  Figure  5B,  D, and  F, for all three   plate and both top and bottom surfaces taped to the
            structures, the area from the 1  cycle is much larger than   compression plates.
                                     st
            those from the 15  and 30  cycles. It could be deduced
                                  th
                           th
            that the energy dissipation mostly happened within   Figure 7A depicts the force-displacement curves of the
            the 1   loading and unloading cycle. Furthermore, the   three structures in Group 1. The experiments with only the
                st
            structures dissipated a smaller but similar amount of   bottom surface glued to the compression plate were carried
            energy during later cycles.                        out first. The shaded areas suggested minor discrepancies
                                                               in experimental results among the three repeating tests for
              The critical buckling stress during the loading phase   each design. Some features of the snap-through among the
            and the residual strain after the unloading phase are   three different designs shared a similar phenomenon as in
            plotted in  Figure  6. The critical buckling stress for all   Group 2. To be specific, the structural stiffness was observed
            three structures drops drastically after the 1   cycle,   to increase with greater h’ values, leading to higher critical
                                                    st
            whereas slight differences are observed for cycles   force (P ). Besides, all three designs experienced the first
                                                                     cr
            afterward. Convergence of critical effective stress appears   snap-through at approximately the same displacement.
            after the 6  cycle for all three structures at the effective   No obvious second snap-through could be observed from
                    th
            stress of 2.5 kPa, 5.8 kPa, and 8.5 kPa, respectively. The   the design with h’ = 3. With the increase of h’, the second
            early convergence of critical effective stress indicates that   snap-through appeared to be more pronounced. During
            the plastic deformation mainly happened during the first   the experiment, there was a phase in which no force was
            few cycles. The residual strains for all three structures   recorded in the Instron machine for all three designs. Since
            are below 3%, suggesting their remarkable recoverability   the loading cell of this Instron machine is connected to  the
            and reusability. The precision of the Instron machine   top compression plate, the zero-force phase indicates the
            could cause the fluctuation of the residual strains.   detachment of the top surface of the structure and the
            Overall, structures in Group 2 are robust against repeated   top compression plate. After removing the compressive
            loadings.                                          load, one pair of the curved beams recovered to its initial
              The comparison among the three structures implies   configuration, and the other pair remained in its deformed
            that Design No. 6 (l’ = 60, h’ = 5) yields a slightly more   shape.
            resilient response toward loading and unloading cycles.   To investigate the causes of the occurrence of the zero-
            To be specific, the residual strain for Design No.  6 is   force phase, another group of experiments were performed
            smaller than Designs No.  4 (l’ = 60,  h’ = 3) and No.  5   with both the bottom and top surfaces of the specimens
            (l’ = 60,  h’ = 4). This observation suggests that Design   glued to the compression plates. The dished lines in
            No. 6 experienced the least amount of plastic deformation   Figure  7A are the force-displacement curves, which are
            among all. In terms of critical effective stress, all three   aligned with the previous results without the top surface
            structures witnessed an approximately 60% decrease   attached to the plate. In this case, negative forces were
            mainly from the 1  cycle.                          recorded during the zero-force phase in the previous tests,
                          st

                         A                                   B













            Figure 6. Results of cyclic loading and unloading experiment (A) Critical effective stress during each loading cycle. (B) Residual strain after each unloading
            process. The symbols δ and ε  represent critical effective stress and residual strain, respectively.
                          cr
                               r
            Volume 3 Issue 4 (2024)                         9                              doi: 10.36922/msam.4960
   105   106   107   108   109   110   111   112   113   114   115