Mechanism and Machine Theory 49 (2012) 21–39

Contents lists available at ScienceDirect

Mechanism and Machine Theory

j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / m e c h m t

A compliant mechanism kit with flexible beams and connectors along with
analysis and optimal synthesis procedures☆
Padmanabh Limaye, G. Ramu, Sindhuja Pamulapati 1, G.K. Ananthasuresh ⁎
Mechanical Engineering, Indian Institute of Science, Bangalore, India

a r t i c l e i n f o a b s t r a c t

Article history: We present a compliant mechanism kit as a parallel to the kits available for rigid-body
Received 18 December 2010 mechanisms. The kit consists of flexible beams and connectors that can be easily hand-
Received in revised form 9 July 2011 assembled using snap fits. The mechanisms assembled using the kit accurately capture the
Accepted 14 July 2011
aspects of the topology, shape, and size of joint-free compliant mechanisms. Thus, the kit
Available online 25 August 2011
enables designers to conceive and design new, practicable, single-piece compliant mechanisms
that do not require assembly. The concept of the kit also resolves a discrepancy in the finite
Keywords:
element (FE) modeling of beam-based compliant mechanisms. The discrepancy arises when
Compliant mechanism kit
two or more beams are joined at one point and thus leading to increased stiffness. After
Beam finite element analysis
Topology optimization
resolving this discrepancy, this work extends the topology optimization to automatically
Building blocks generate designs that can be assembled with the kit for quick and easy validation instead of
Fillets time-consuming prototyping. Thus, the kit and the accompanying analysis and optimal
synthesis procedures comprise a self-contained educational as well as a research and practice
toolset for compliant mechanisms. The paper also illustrates how human creativity finds new
ways of using the kit beyond the original intended use and how it enables even a novice to
design compliant mechanisms.
© 2011 Elsevier Ltd. All rights reserved.

1. Introduction

Compliant mechanisms are joint-free mechanisms that transmit or transform motion or force through elastic deformation
rather than through hinges and sliders as in their rigid-linked counterparts. The primary advantages of compliant mechanisms are:
fewer parts, fewer assembly steps, absence of backlash, and reduced need for lubrication. The absence of hinges makes compliant
mechanisms attractive for many applications [1,2] including the emerging areas of micro and nano-scale systems [3]. The twofold
motivation for the work presented in this paper is explained next.
In spite of their many advantages, compliant mechanisms are not yet widely used nor are they taught widely in undergraduate
engineering courses. A reason for this slow, but gradually increasing, adoption of compliant mechanisms in practice and education
may be that designing them is slightly involved because one has to deal with elastic deformations—often geometrically nonlinear.
Designing or even analyzing compliant mechanisms usually requires access to finite element analysis (FEA) software. The pre- and

☆ Enhanced version of a paper presented at the 14th National Conference on Machines and Mechanisms (NaCoMM-09), NIT, Durgapur, India, December 17–18,
2009.
⁎ Corresponding author. Tel.: + 91 80 2293 2334; fax: + 91 80 2360 0648.
E-mail address: suresh@mecheng.iisc.ernet.in (G.K. Ananthasuresh).
1
Summer intern from the Indian Institute of Technology, Hyderabad, India.

0094-114X/$ – see front matter © 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.mechmachtheory.2011.07.008
22 P. Limaye et al. / Mechanism and Machine Theory 49 (2012) 21–39

post-processing involved in FEA, i.e., drawing or modifying the computer model and meshing it and then visualizing its
deformation, makes it difficult for the designer to exercise creativity and intuition. Prototyping a compliant mechanism is another
difficulty because one needs to machine it usually using a computer-numerically-controlled (CNC) machine because of their not-
so-simple geometry. Contrast this situation with rigid-body mechanisms whose motion is more easily visualized. Or, they can be
built easily using even cardboard and pins. Furthermore, several kits have been developed as practical toolsets for designing rigid-
body mechanisms wherein mechanisms can be realized by simple hand-assembly from the available parts. It is desirable to have
such a kit for compliant mechanisms too. Some compliant Lego building blocks are available to work in conjunction with the
Lego TM kit [4]. The concept of a stand-alone and general kit with computational support is the focus of this work.
There is a second motivation for developing a compliant mechanisms kit. A beam finite element is a popular choice for analysis
or topology optimization of compliant mechanisms [5], which have certain advantages over continuum finite elements. Beam
finite elements are attractive in analysis because most compliant mechanisms comprise slender beam-like segments. And they are
attractive in topology optimization because fewer elements are required and the potential to obtain distributed compliance is
better than what can be achieved with continuum element-based methods. However, there is an inconsistency in using beam
finite elements for modeling compliant mechanisms. It is due to the modeling of the connections where two or more beams of
different widths intersect. Consequently, a prototype of an analyzed or designed mechanism is found to be stiffer than its finite
element (FE) beam-model. This is illustrated with an example presented next.
Fig. 1(a) schematically shows a beam connection as it is assumed in the FE model while Fig. 1(b) shows the same beam
connection as it appears in the prototype. To accommodate the finite size of the cutting tool and/or to reduce stress
concentrations, a fillet is indeed unavoidable. But it adds extra material at the joint. It also reduces the actual deformable lengths
of the intersecting beams as opposed to the lengths assumed in the FE model. Hence, the joint becomes stiffer in the prototype
than its FE model [6].
When we use beam ground structures in topology optimization [7], there are many beam-to-beam intersections. Hence, the
aforementioned problem becomes even more prominent. An example is shown in Fig. 2(a–b). Fig. 2a shows the beam-model of a
displacement-amplifying compliant mechanism in its original and deformed configurations while Fig. 2b shows the same for an
actual 2D form with fillets. By nonlinear finite element analysis using two-node beam and nine-node 2D continnum elements, it
was found out that there was a discrepancy of as much as 109% in the strain energy between the two models. While beam model
had strain energy of 13.9E−3 Nm, the continuum model had only 6.65E−3 Nm. This huge discrepancy is not very surprising
because the effective lengths of beams decrease because of the junctions and fillets.
While developing the methods to resolve the issue of junction and fillet induced additional stiffness [8], it was realized that a
better way is to simply avoid the intersections of flexible beams by having a semi-rigid connector in analysis, design, and the real
prototype. This led to the idea of a compliant mechanisms kit that enables the design of compliant mechanisms with beams that
remain closer to reality. We note that the mechanisms assembled with the kit are only designs and not the final mechanisms to be
used in practice. This is in the same spirit as any other things designed using a kit.

L2

L1- Deformable length

Beam Overlapping Region

Fillet (extra material)
L2

Actual deformable length

Extra material L1

Fig. 1. (a) Ideal beam connection in FE model. (b) Actual beam connection in prototype.
 P. Limaye et al. / Mechanism and Machine Theory 49 (2012) 21–39 23

Fig. 2. (a) A displacement-amplifying compliant mechanism's beam model in original and deformed configurations, (b) its continuum model in the
manufacturable form with junctions and fillets.

It may also be noted that what can be done with the compliant mechanisms kit can also be done with a finite element analysis
software but it will not be as quick as checking a mechanism idea with the kit. This is because much time is required to create a
model in a computer, mesh it, and then analyze it. If it is large-displacement analysis, it will take a lot of time for computation. If
pre-loading or contact are present, the kit surely has further benefits over software-based trials of designers' ideas. This will
become clear after examining the examples presented later in the paper.
The remainder of the paper is organized as follows. We describe the novel concept of the kit and its building blocks in Section 2.
The use of the kit is also illustrated in this section. Finite element modeling of a compliant mechanism assembled using the kit is
explained in Section 3. Topology optimization, which can generate solutions that are realizable with the kit, is presented in
Section 4. Section 5 discusses the uses of the kit and illustrates how creative users can use it in ingenious ways even though they
may not be familiar with compliant mechanisms a priori. This point is underscored by the fact that a summer intern, the third
author, who after completing only the first year of mechanical engineering undergraduate programme, was able to design novel
compliant mechanisms effortlessly.

2. Compliant mechanisms kit

The proposed compliant mechanisms kit consists of two types of building blocks. One is a flexible beam and the second is a semi-
rigid connector. The connector undergoes very small to negligible deformation as compared with that of the beams. The connectors
are located at the junctions where different beams meet in different directions. Fig. 3(a) shows the geometric model of the
connector while Fig. 3(b) shows a sample assembly in which there are five connectors and eight beams. Four beams are longer
than the other four beams to make this possible. At this time, we have only two lengths for the beams but more could be added.
The point we illustrate in this paper is that just with two lengths, many interesting compliant mechanism designs can be
conceived. And, two lengths are adequate for a square grid with cross-beams that go diagonally. This is, by the way, the beam
ground structure used in topology optimization [8,9].
The geometry of the connector was designed (see Fig. 4) such that a beam can be easily snap-fitted into one of the eight slots.
Also seen in the connector are eight shorter crack-shaped notches. These help in inserting a beam by allowing slight deformation
24 P. Limaye et al. / Mechanism and Machine Theory 49 (2012) 21–39

Fig. 3. (a) Semi-rigid connector. (b) A simple assembly with five connectors and eight beams.

of the material on either side of the slot and in restoring the original configuration when the beam is positioned properly. This
helps in a tight fit of the beam inside the slot even though it does not take much force for insertion. This is the well-known
principle of snap-fits.

2.1. Physical building blocks of the kit

We chose spring steel (AISI 1040) for making the physical building blocks. It permits very thin beams that can be easily cut by
shearing from a sheet, or by using wire-cut Electro Discharge Machining (EDM) if precision is desired. The spring steel beams allow
reversible large elastic deformation without yielding. The connector, with its intricate and narrow cuts shown in Fig. 4, can also be
made using spring steel or aluminium using wire-cut EDM.
 P. Limaye et al. / Mechanism and Machine Theory 49 (2012) 21–39 25

Fig. 4. Top-view of the semi-rigid connector with eight slots for snapping the beams and eight crack-shaped notches for aiding the insertion.

Fig. 5(a) shows an assembled compliant mechanism using the physical building blocks. Its connectors at the four corners are held
fixed by gluing it to a base. When a rightward force is applied in the middle of the left edge as shown in Fig. 5(b), the midpoint of
the right edge moves to the left. This is a non-intuitive motion. Some amplification of the displacement is also noticeable in the
figure.
The horizontal and vertical spacing of the connectors is 50 mm. The beams are made of spring steel of thickness 0.2 mm and in
two standard lengths. The radius of the connector (see Fig. 5c) is 5.5 mm. In all the examples in this paper, the dimensions of
beams and connectors remain the same as above.
It is well known in the compliant mechanisms literature that the topology is of utmost importance. The connectivity of the
beams decides the way the mechanism deforms. As shown in Fig. 5(a–b), a particular arrangement of beams leads to non-intuitive
motion. This is what a topology optimization algorithm can do by removing the unnecessary beams in a ground structure. The kit
makes it easy for a human user to do the same and facilitates creativity. Selective removal/addition of the beams can also be done
interactively on a computer but a kit gives the feel and realistic geometrically nonlinear motion much more quickly than a
nonlinear finite element solution.
The topology designed by a human user can be readily used to make a real single-piece compliant mechanism made with
suitable material and appropriate manufacturing technique, for practical use. The mechanism assembled with the kit itself may not
be practical because the beams may get loosened and come out of the connectors under excessive deformations. It is the
conception of designs that the kit enables. Fig. 6 shows the mechanism of Fig. 5a in its continuum form with fillets and junctions
changed to facilitate CNC-milling using polypropylene. It may be noticed in Fig. 6 that the beam lengths and widths are used to
decide on the junction geometry based on the milling cutter's radius. The overlaid color plot of the simulated deformed
configuration was printed on a sheet of paper to show how the deformed mechanism looks like. This may be compared with Fig. 5b
to discern the resemblance. This example shows how a real compliant mechanism can be made using the design inspired by the
kit. Note that this asymmetric mechanism is able to make the output point move along an almost perfectly in line with the input
force direction over a large range of motion.
Next, we present some examples to illustrate the kinds of compliant mechanisms that are possible to design with the kit.

2.2. Examples

Consider three specifications for the design of compliant mechanisms as shown in Fig. 7(a–c). All three have some practical
relevance. The first one (Fig. 7(a)) requires that an input force at a point results in an output motion that approximates rotation
around a fixed point. This may be useful for disk-drive mechanisms so that they can be miniaturised with linear microactuators
rather than rotary micromotors. The problems of rotary micromotors such as stiction and wear are well known. It would be a good
exercise to imagine which beams are to be retained in Fig. 7(a) in order to get this input–output motion using the compliant
mechanism kit.
In Fig. 7(b,c) a different type of specification is shown. Shown in Fig. 7(b) is a fiber that needs to be grasped with two pairs
of points first and then pulled apart. This may be needed to test the strength of a fiber. Note that the specification is such that a
single force is enough to do it in order to minimize the need for control. Fig. 7(c) shows the same problem but there a rod
should be grasped first and then pushed in from both the ends. In both specifications, there is no restriction as to where the
compliant mechanism should be fixed. This is in contrast to what is usually done in topology optimization and design
methodologies developed for compliant mechanisms. Thus, this can be seen as an advantage over most existing computational
design methods.
26 P. Limaye et al. / Mechanism and Machine Theory 49 (2012) 21–39

Fig. 5. (a) A compliant mechanism assembled from the kit. (b) The deformed configuration. (c) The connector made of spring steel.

The reader should pause at this time and think of possible solutions to the above three problems. Usually, there are many
solutions to such problems. Here, through the compliant mechanism kit of two specific lengths of flexible beams, we are
restricting the possible solutions that can be realized. Even then many solutions can exist. The question is: which beams in the
grid of Fig. 7(a) ought to be retained to get the desired deformation? The same question is posed for the specifications in Fig. 7
(b–c). Solutions for the three problems are shown in Figs. 8–10. These were created partly with intuition and partly by
experimenting with the kit by the third author who had completed the first year of engineering program at the time of doing it.
 P. Limaye et al. / Mechanism and Machine Theory 49 (2012) 21–39 27

Fig. 6. Polypropylene prototype based on the design shown in Fig. 5a. The color sketch printed on the background paper shows the deformed configuration of the
mechanism upon actuation.

At that level, the concepts of kinematics and strength of materials are still unfamiliar and finite element analysis is unheard of.
Yet, the kit made it possible to experiment with the beams and connectors in various ways to develop ‘compliant mechanism
intuition’ so that all three problems (and more) could be systematically thought through to create at least one possible solution
for every problem.
The deformation, as can be seen in Figs. 8–10, is quite large. Developing an intuition for the large deformations is not easy. Finite
element solution, if done in a computer, does not give the result as quickly as the kit can provide. The force we feel at the input is
also an added advantage in understanding the kinestostatic behavior of the mechanism. This is useful in studying the mechanical
advantage aspects of compliant mechanisms.
Also worth noting here is the contact that takes place among different parts of the mechanisms. Such mechanisms are known as
contact-aided compliant mechanisms [6]. Intermittent contacts enrich the kinematic and stiffness adaptability of compliant

a Output Output
Fixed

Input force

b
Pull

Grasp Grasp

Input force
c

Push

Grasp Grasp

Input force
Fig. 7. Required input–output motion specifications for compliant mechanisms. (a) An input force giving an approximate rotation of two points about a fixed point,
(b) grasp and then pull a fibre, (c) grasp and then push on a rod.
28 P. Limaye et al. / Mechanism and Machine Theory 49 (2012) 21–39

Fig. 8. (a) Solution to problem shown in Fig. 5(a), (b) the deformed configuration wherein a large input force is applied. The desired approximate rotation of two
points about a fixed point is noticeable.

mechanisms. Such concepts can also be explored with the help of this kit without having to simulate the time-consuming contact
mechanics associated with it.

3. FE modeling of compliant mechanisms made with the kit

In compliant mechanisms made using the kit, the beams are not directly connected to each other; rather they are connected
through the semi-rigid connector. We call it semi-rigid here because it undergoes slight deformation during insertion of the
beams. However, once a mechanism is assembled, the connectors are essentially rigid. Therefore, connectors undergo a rigid-body
motion in the assembled mechanism. Consequently, there is a need to appropriately relate the displacements of nodes at which
beams are connected to those of the beams' degrees of freedom. This can be done using what are called multi-point constraints in
the finite element analysis (FEA) literature. These constraints are linear for small displacements and nonlinear for large
displacements.
In order to analyze these mechanisms, the equations obtained by regular FE assembly need to be solved along with the
constraint equations. It increases the computation. An alternate method is to use a super element in the place of each connector.
The stiffness matrix of this element should be large enough so that the element shows singular modes for translational
displacements as well as for small rigid rotations. Such a stiffness matrix for the super element can be obtained by the influence
coefficient method [10]. In this, a unit force is applied at each degree of freedom and the resulting displacements are stored as a
column in a matrix. By doing this for all the degrees of freedom, all the columns of the matrix are filled. This matrix is multiplied by
a large value to simulate the rigidity of the connector. Such connector stiffness matrices are assembled with beam elements and
the resulting FE equations are solved together.
We note that the problem of joint-stiffening in ordinary beam connections, noted in the introductory section, is avoided here.
The connector is rigid in both the model and the prototype. Hence, there is no discrepancy as long as the beams obey the
assumptions of the Euler–Bernoulli beam theory. In the kit, the beams are made sufficiently slender to ensure this; that is the
length is at least 10–15 times larger than the larger of the cross-section dimensions.
To solve the large deformation problem, we used co-rotational beam elements as implemented in [11,12]. The load is applied in
steps. At every step, the equilibrium equation is solved using the Newton–Raphson algorithm. It requires computing the tangent
stiffness matrix and the internal force at each step. It is necessary to separate rigid-body displacements from the total
displacements. Fig. 11 shows the large displacement of a connector which includes rigid rotation, α. The local coordinate system
 P. Limaye et al. / Mechanism and Machine Theory 49 (2012) 21–39 29

Fig. 9. (a) Solution to problem shown in Fig. 5(b), (b) the deformed configuration. Two pairs of points have come together to grasp and then they are pulled apart.

(LCS) makes an angle α with the global coordinate system (GCS). For simplicity, only two nodes are considered, which have
coordinates (X1, Y1) and(X2, Y2), with respect to origin O. UL and Ug are displacements in LCS and GCS, respectively.

h iT
UL = Ux1 Uy1 Uθ1 Ux2 Uy2 Uθ2
L
h iT ð1Þ
Ug = Ux1 1
Uy
1
Uθ
2
Ux
2
Uy
2
Uθ
g

The mapping between LCS and GCS is given by the transformation:

δUL = λδUg
ð2Þ
F intL = λF int g

where transformation matrix λ is given by

2 3
cosðα Þ sinðα Þ 0 0 0 0
6 −sinðα Þ cosðα Þ 0 0 0 07
6 7
6 0 0 1 0 0 07
λ=6
6
7 ð3Þ
6 0 0 0 cosðα Þ sinðα Þ 077
4 0 0 0 −sinðα Þ cosðα Þ 05
0 0 0 0 0 1

UL is obtained from Ug by subtracting displacements due to rigid rotation of α and mapping them to LCS by transformation
matrix λ.

 
UL = λ Ug −Γ ð4Þ
30 P. Limaye et al. / Mechanism and Machine Theory 49 (2012) 21–39

Fig. 10. (a) Solution to problem shown in Fig. 5(c), (b) an intermediate deformed configuration when two pairs of points have come together to grasp; (c) further
application of the input force results in the pushing-in of the grasped points.

2 32 3
cosðα Þ−1 −sinðα Þ 0 0 0 0 X1
6 sinðα Þ cosðα Þ−1 07 6 7
6 0 0 0 76 Y1 7
6 α 07 6 7
Γ=6
0 0 0 0 76 1 7 ð5Þ
6 cosðα Þ−1 −sinðα Þ 07 6 7
6 0 0 0 76 X2 7
4 0 0 0 sinðα Þ cosðα Þ−1 0 54 Y2 5
0 0 0 0 0 α 1
int
FL = KUL ð6Þ

From Eqs. (2), (4), and (6), we can write

 
int T
Fg = λ Kλ Ug −Γ ð7Þ
 P. Limaye et al. / Mechanism and Machine Theory 49 (2012) 21–39 31

Fig. 11. Solid lines show the original connector and dashed lines show displaced configuration. Mapping is obtained between LCS and GCS.

Hence, the tangent stiffness matrix is obtained as

tgt ∂F int g
Kg = ð8Þ
∂Ug

tgt T
Kg = λ Kλ ð9Þ

A test problem is analyzed as shown in Fig. 12, where two beam elements are connected together with a rigid connector. The
stiffness of the joint is kept very high so as to make it perfectly rigid. We notice that the joint undergoes a large rotation just as its
physical counterpart made with the kit would do.

4. Optimal synthesis by topology optimization

Using the FE model discussed in the previous section, we now proceed to beam ground structure-based topology optimization.
The algorithm in topology optimization can do what the human user can do with the kit: add/remove connectors and beams as
needed to satisfy the specified functionality. The optimization problem is stated as follows.

Maximize MSE
xi
SE
subject to KU = F ð10Þ
KUd = Fd
Vb <V ⁎
xi = w or 0

Fig. 12. Test problem of the rigid connector for large deformation problem.
32 P. Limaye et al. / Mechanism and Machine Theory 49 (2012) 21–39

F Fd

Fig. 13. (a) Inverter mechanism shown with deformed configuration for F = 100 N, (b) Distribution of widths plotted. Displacement amplification is 1.96. p = 1.

where MSE = U TKUd and SE = 0.5 U TKU. Here, the ratio of mutual strain energy (MSE) and strain energy (SE) is considered as the
objective function to be maximized with respect to widths of the beams in the ground structure grid as the design variables, xi. A
spring is provided at the output port where the displacement is to be maximized. Fd denotes a unit dummy load applied at the
output node in the direction in which the displacement is to be maximized. The volume constraint has V* as the upper limit on the
volume V of the beam elements used.
As the designs are intended to be made using the kit, the design variables xis are constrained to take a value of either 0 or w,
where w is the standard width of the beams in the kit. Thus, it is a binary optimization problem. In order to solve it efficiently using
gradient-based optimization, we consider it as a continuous optimization problem wherein intermediate values of xi are penalized.
A penalization factor p is used for this purpose. Instead of xi, now Xi is used as the beam width for computing the element stiffness
matrices.

x p
i
Xi = w ð11Þ
w

We solved few benchmark problems using this formulation. The Method of Moving Asymptotes (MMA) algorithm [13] is used
for optimization. In all following examples, w is 0.2 mm, the same as that in the kit. Lower bound on xi is kept at 1E−3 to avoid
numerical singularities and yet is small enough to simulate the absence of a removed beam.
 P. Limaye et al. / Mechanism and Machine Theory 49 (2012) 21–39 33

F F

Fig. 14. (a) Inverter mechanism shown with deformed configuration for F = 100 N, (b) Distribution of widths plotted. Displacement amplification is 2.19. p = 2.

Figs. 13(a) and 14(a) show the results of optimization for p=1 and p=2 respectively for the inverter mechanism wherein the direction of
the input force is reversed in the output displacement. A volume fraction of 0.4 is used. The spring at the output has a value equal to 1E−10
times E, the Young's modulus of 210 GPa for spring steel was used. Random asymmetric initial guesses were used to get these designs.
Symmetric solutions obtained with symmetric initial guesses for these problems were stiffer than the asymmetric designs.
The distribution of beam widths is shown in Figs. 13(b) and 14(b). We notice that the widths of almost all the members are
pushed to either 0 or w, not only in the case where p = 2 but also for p = 1. The reason is attributed to the fact that the bending
stiffness of beams is proportional to the cube of their widths. This acts as inherent penalization for beams that are in between 0 and
w and are not preferred by the optimization algorithm. Although some of the beams have attained intermediate widths, it is
negligible and the mechanism behaves similarly, when those particular beams are deleted. It can be noticed that the number of
such intermediate elements has decreased with p = 2.
The mechanism solution assembled using the kit as per Fig. 14(a) is shown in Fig. 15. It is important to note that a human
designer might come up with a different solution as the solution is not unique. Also important to note is that the optimization
algorithm based on linear analysis takes, in general, less time on a desktop computer than a human designer can take to
conceive a mechanism with the kit. It may therefore appear that topology optimization has an edge over human creative
endeavor. But it is not true when we implement the nonlinear analysis into the topology optimization program [14]. Those
results are not presented in this paper. But it suffices to note that human user with the compliant mechanisms kit can then
compete with the computer in coming up with a solution for a desired input–output motion specification.
34 P. Limaye et al. / Mechanism and Machine Theory 49 (2012) 21–39

Fig. 15. Compliant mechanism assembled using the kit as per the topology solution shown in Fig. 12(a): (a) without force, (b) with force.

5. Discussion and other uses of the kit

We showed in Section 2 how the kit can be used to conceive new compliant mechanisms. The one shown in Fig. 10(a) is
difficult among the three examples. It is not an easy one to be solved by topology optimization algorithm as it involves two steps
(grasp first and then push) and very large deformations as can be seen in Fig. 10(b–c). In order to see how this was designed,
consider the actual hand-sketched ideation process shown in Fig. 16.
The mechanism in Fig. 10(a–c) was designed in steps following a simple idea of grasping. The first set of sketches in Fig. 16
show this. The next set of sketches shows how the grasped points can be pulled apart when a force is applied at the bottom. The
push-part is added in a few steps. This was achieved by beginning to assemble the kit and trying out different designs and by fixing
different connectors. Once the basic functionality was achieved, it was made sufficiently stiff by experimenting with the kit further.
The verbatim description of the designer (the third author) is rather long to include here. But is it not unreasonable to assume that
many others can also do this to create novel compliant mechanisms.
Human creativity enhances the utility of the kit. Some observations were made when this kit was used by different people.
Consider the mechanism shown in Fig. 17(a–b) with its undeformed and deformed configurations. At first sight, it is also a grasp
and pull mechanism albeit with two inputs that act on the left and right hand sides. But there is more to it than can be understood
at first sight. Observe how the almost rectangular outer profile of the mechanism changes to groundnut (peanut) shape when it is
deformed. This is an example of a profile-morphing design. A number of researchers are working on this problem to conceive
designs where an aircraft wing can smoothly morph from one shape to another [15–17]. We find that the kit can be of help in such
problems as well. Next, we briefly describe our preliminary attempt at this.
 P. Limaye et al. / Mechanism and Machine Theory 49 (2012) 21–39 35

Fig. 16. Hand-sketched thought-process that led to the mechanism assembled with the kit and shown in Fig. 8(a). The thought process is too long to include here
but may be partly understood from the sketch.

Fig. 17. A mechanism designed for grasping and pulling on an object using two inputs. (a) undeformed, (b) deformed. Notice also how the almost rectangular outer
profile of the undeformed mechanism changes to groundnut (peanut) shape in the deformed mechanism. This is a clue to designing morphing topologies using the kit.
36 P. Limaye et al. / Mechanism and Machine Theory 49 (2012) 21–39

Fig. 18. The compliant mechanism kit used for designing morphing wings. (a) An approximate aerofoil shape assembled using the same building blocks—beams of
two different lengths and a connector. (b) The changed profile with a single actuation at the middle connector on the left side.

In morphing wing design, the intent is to conceive an internal arrangement of the beams inside an aerofoil so that its shape can
be changed with just one internal actuation. For this, a user assembled an approximate aerofoil shape with the existing kit. See
Fig. 18(a–b). By fixing the upper and lower connectors on the left side and moving the middle connector, the shape can be

Fig. 19. Two stable positions of a bistable single actuator-based trailing edge design of a morphing airfoil.
 P. Limaye et al. / Mechanism and Machine Theory 49 (2012) 21–39 37

changed so that the trailing edge can be substantially deflected while the leading edge keeps the shape in comparison. Thus, not
unexpectedly, users think of novel ways of using the kit. Using bent beams (see Fig. 10(a–c)) to prevent the buckling or
excessive and undesirable deformation of an important beam is another such example.
Another use of the kit arises in the context of multi-stable mechanisms [18,19]. The nonlinearity inherent in the compliant
mechanisms assembled using the kit enables multi-stable compliant mechanisms. The freedom in pre-loading of some
elements can also be exercised using the kit. The designer who has enough experience with the kit can create bistable and
multi-stable mechanism within a short span of time. The time needed by designers, on the average, is not more than an hour just
as it is with Lego TM-like building blocks or rigid-link mechanism kits. Fig. 19 (a–b) show the bistable airfoil whose trailing edge
can be smoothly morphed between two states. A similar behavior for the leading edge is shown in Fig. 20(a–b). Such designs are
advantageous because actuation is needed only to move the airfoil from one shape to another but not for retaining it in either
shape.
Fig. 21 shows a 16-state multi-stable compliant mechanism created using the kit in all its 16 stable configurations. Each of
its four segments were made to be bistable by way of pre-loading. It is an extreme example that may not have an immediate
practical use. However, the imaginative avenues that the kit provides may be put to practical use by inventive designers. The
bistable morphing airfoils shown in Figs. 19 and 20 are examples of such utility arising out of the kit. Furthermore, the
process of realizing a compliant mechanism made using the kit is straightforward: the flexible beams can be permanently
fastened or a unitized mechanism can be made out of a single block of material, as it is usually done for compliant
mechanisms.

6. Conclusions

In this paper, we proposed a kit for intuitively conceiving designs of compliant mechanisms. The kit comprises a semi-rigid
connector and beams of two lengths, all made using spring steel. It is possible to design compliant mechanisms quickly by using
the kit by mere hand-assembly and by exercising one's intuition. It also helps in comprehending the results of systematic synthesis
process.
It is also worth noting that the kit helped circumvent a modeling discrepancy encountered in beam finite element based simulation
of compliant mechanisms. Additionally, a finite element analysis technique for the mechanisms made using the kit was described. Also
presented in the paper is a topology optimization method that gives a solution to arbitrary specifications. The topology-optimized
solutions can be made using the kit because the constraints of the kit are incorporated into the topology optimization program. This
program uses small-displacement linear modeling and hence is fast. When nonlinear modeling (which is not presented here) is used

Fig. 20. Two stable positions of a bistable single actuator-based leading edge design of a morphing airfoil.
38 P. Limaye et al. / Mechanism and Machine Theory 49 (2012) 21–39

Fig. 21. A 16-state multi-stable compliant mechanism created using the kit in all its stable configurations.
 P. Limaye et al. / Mechanism and Machine Theory 49 (2012) 21–39 39

to account for the large displacements, the program takes more time. In such a case, human designers can design faster than the
computer program.
The kit's parts along with the optimal design method prove to be a pragmatic toolset for design of single-piece compliant
mechanisms that do not require assembly. Such things as contact among the members, buckling, and unspecified fixed locations,
etc., can be easily attempted with the kit but not with topology optimization. By presenting the examples involving non-intuitive
motions, shape-morphing, and multi-stable configurations, the creative ways in which the kit can be used are illustrated in the
paper.

Acknowledgments

The authors thank Mr. Pakeeruraju Podugu for his inputs on the aerofoil example. Synergistic efforts among the members of the
Multi-disciplinary and Multi-scale Device and Design (M2D2) laboratory in the Mechanical Engineering department of the Indian
Institute of Science, Bangalore, also helped greatly in developing the compliant mechanisms kit idea into a potential viable toolset
in practice. Useful comments of an anonymous referee helped improve the clarity of presentation.

References

[1] L.L. Howell, Compliant Mechanisms, John Wiley, New York, 2001.
[2] N. Libontiu, E. Garcia, Mechanics of microelectromechanical systems, Kluwer Academic Publishers, New York, 2005.
[3] G.K. Ananthasuresh, L.L. Howell, Mechanical design of compliant microsystems: a perspective and prospects, Journal of Mechanical Design 127 (4) (2005)
736–738.
[4] http://research.et.byu.edu/llhwww/clegos.htm.
[5] M.I. Frecker, Recent advances in optimization of smart structures and actuators, Journal of Intelligent Material Systems and Structures 14 (4–5) (2003)
207–216.
[6] N.D. Mankame, G.K. Ananthasuresh, Synthesis of contact-aided compliant mechanisms for non-smooth path generation, International Journal of Numerical
Methods in Engineering 69 (12) (2007) 2564–2605.
[7] A. Saxena, G.K. Ananthasuresh, On an optimal property of compliant topologies, Structural and Multidisciplinary Optimization 19 (2000) 36–49.
[8] P. Limaye, “Investigating the Accuracy of Beam-based Finite Element Modeling in the Analysis of Compliant Mechanisms”, Master of Engineering (ME)
Project Report, Indian Institute of Science, Bangalore, 2009.
[9] C.B.W. Pedersen, Topology Optimization Design of Crushed 2D-frames for Desired Energy Absorption History, Structural and Multidisciplinary Optimization
25 (5–6) (2003) 368–382.
[10] W.T. Thompson, Theory of Vibrations, Nelson Thornes Ltd., New York, 1996.
[11] P. Sivanagendra, Geometrically nonlinear elastic analysis of frames with application to vision-based force-sensing and mechanics of plant stems, ME Project
Report, Indian Institute of Science, Bangalore, 2006.
[12] P. Sivanagendra, G.K. Ananthasuresh, Size-optimization of a cantilever beam under the deformation dependent load with application to wheat plants,
Structural and Multidisciplinary Optimization 39 (2009) 327–336.
[13] K. Svanberg, The method of moving asymptotes - a new method for structural optimization, International Journal for Numerical Methods in Engineering 24
(1987) 359–373.
[14] A. Saxena, G.K. Ananthasuresh, Topology synthesis of compliant mechanisms for nonlinear force-deflection and curved path specifications, Journal of
Mechanical Design 123 (2001) 33–42.
[15] M. Abdulrahim, H. Garcia, R. Lind, Flight characteristics of shaping the membrane wing of a micro air vehicle, Journal of Aircraft 42 (2005) 131–137.
[16] L. Saggere, S. Kota, Static shape control of smart structures using compliant mechanisms, AIAA Journal 37 (5) (1999) 572–578.
[17] A.Y.N. Sofla, S.A. Meguid, K.T. Tan, W.K. Yeo, Shape-morphing aircraft wing: status and challenges, Materials and Design 31 (2010) 1284–1292.
[18] Y.Se. Oh, S. Kota, Synthesis of multistable equilibrium compliant mechanisms using combinations of bistable mechanisms, Journal of Mechanical Design 131
(Issue 2) (2009) 021002-1-11.
[19] G. Chen, D.L. Wilcox, L.L. Howell, Fully compliant double tensural tristable micromechanisms, Journal of Micromechanics and Microengineering 19 (No. 2)
(2009) 025011-1-8.