International Journal of Solids and Structures 254–255 (2022) 111902

Contents lists available at ScienceDirect

International Journal of Solids and Structures

journal homepage: www.elsevier.com/locate/ijsolstr

Analysing wrinkle interaction behaviour with Z-fold crease pattern in

thin-film planar membrane reflector
Parth K. Kamaliya a, Ajitesh Shukla a, S.H. Upadhyay a, *, H.M.Y.C. Mallikarachchi b
a
Smart Materials and Structures Lab, Department of Mechanical and Industrial Engineering, Indian Institute of Technology Roorkee, Uttarakhand, India
b
Department of Civil Engineering, University of Moratuwa, Moratuwa, Sri Lanka

A R T I C L E I N F O A B S T R A C T

Keywords: Limited stowage volume of launch vehicles demands compact packaging of large deployable membrane struc­
Inflatable antenna structure tures. Employment of different folding approaches led to the unavoidable fold-lines(creases). Moreover, wrinkle
Planar membrane reflector reduction is essential in such thin-film solar sail and inflatable antenna structures. This article analyzes the
Wrinkling
combined effect of multiple creases and wrinkles on deployed thin Kapton’s surface accuracy. Authors have
Crease orientation
Crease-wrinkle interaction
introduced a novel approach for employing membrane length shortening because of multiple creases and fold
Surface accuracy characteristics. Real-world logarithmic behaviour of crease relaxation and fold-region properties help in
computing the crease curvature. A rhombus-shaped shell model with Z-folding is introduced and explored for the
effect of the number of creases and their orientation. Stress concentration and wrinkle contours for applied edge
tensile loading state that the fold lines affect the stress transfer path. Non-linear post-buckling analyses for the
Kapton® membrane express higher wrinkle evolution in the central region over the loading edge. Besides, a
discrepancy in wrinkle amplitude and wavelength reveals the effect of fold-line orientation on wrinkling.
Sighting actual conditions for antenna structure, creasing profile normal to the loading direction was found
preferable for maintaining reflector surface-effectiveness. Additionally, double creased membranes achieve
maximum surface correctness for any angular position, as the mirror effect fold shapes reduce wrinkles the most.
Results are validated by comparing with experimental outcomes.

1. Introduction coiling-wrapping, spiral, Z-folding, or Origami approaches. Though it

seems promising for folding, in-space deployment in one or simulta­
Space structures have limited stowage volume in the launch vehicle neously both directions by employing an external force in a relevant
to keep overall transportation costs as low as possible. However, just for means is fairly complex (An et al., 2020; Lee and Pellegrino, 2014;
weight reduction, an inflatable antenna and solar sail cannot be of Natori et al., 2015; Papa and Pellegrino, 2008; Pehrson and Banik, 2018;
smaller size at the expense of performance, which relies on structural Urbinati, 2020).
size (Arya, 2016; Fenci and Currie, 2017; Fu et al., 2016; Liu et al., Membrane reflectarray is attached to the torus, an inflatable frame of
2017b; Santiago-Prowald and Baier, 2013). The first attempt at an the planar antenna structure. This connection using multiple tie cables
inflatable structure was made with a launch of ECHO balloon in 60s. causes structural instability, which induces asymmetric loading during
Interest started growing after Inflatable Antenna Experiment (IAE) in in-space deployment and post the deployment. This stress imbalance is a
1996 (Pellegrino, 2001). The solar sail concept also came into the pic­ cause of out-of-plane membrane displacement, known as wrinkles
ture, omitting the additional weight to generate thrust. IKAROS project (Jenkins, 2006). Wrinkles, an elastic deformation of thin-film under
demonstrated successful solar sail deployment (Tsuda et al., 2013), compression (Cerda and Mahadevan, 2003; Roddeman et al., 1987; Zak,
which is helpful for current lightweight CubeSats. Polyimide films, 1982), alter the stress transfer path (Wong and Pellegrino, 2006) and
Kapton and Mylar, are found suitable under space environmental con­ disturb the surface topology (Chung et al., 2011). Thick and hard ma­
ditions observing ultra-lightweight, compact packaging ability and terials can replace the low rigidity thin films to reduce wrinkles. Yet,
thermal stability. Raised demand for folding techniques surfaced with lower stowage efficiency limits their use (Secheli et al., 2017). In this

* Corresponding author at: Department of Mechanical and Industrial Engineering, Indian Institute of Technology Roorkee, Roorkee 247667, Uttarakhand, India.
E-mail addresses: [email protected] (P.K. Kamaliya), [email protected] (A. Shukla), [email protected] (S.H. Upadhyay), yasithcm@uom.
lk (H.M.Y.C. Mallikarachchi).

https://doi.org/10.1016/j.ijsolstr.2022.111902
Received 20 July 2021; Received in revised form 26 June 2022; Accepted 22 July 2022
Available online 27 July 2022
0020-7683/© 2022 Elsevier Ltd. All rights reserved.
P.K. Kamaliya et al. International Journal of Solids and Structures 254–255 (2022) 111902

regard, Deng et al. (2019), Li et al. (2012), Ma et al. (2021), Martins antenna reflectors. Woo and Jenkins (2012) investigated the crease cusp
et al. (2020), and Wang et al. (2019; 2018) have studied wrinkling for effect on resisting global wrinkles’ growth. They found wrinkle-crease
different shaped gossamer structures. Jarasjarungkiat et al. (2008) interaction dependency on the applied load. Hence, a less stretched-
introduced a material modification approach to accurately predict out crease significantly impacts wrinkle formation. For lower crease
plastic behaviour in membrane. A reduced-order FE model based on the orientations, the crease can be replaced with a seam (Woo and Jenkins,
Fourier series method was introduced by Huang et al. (2015). Extended 2013) to avoid in-plane compliance (responsible for faster wrinkle for­
at later stage for profiles having non-uniform wrinkles (Khalil et al., mation). Thin-film dynamics for single and double-creased membranes
2020), which claim to define wrinkle evolution microscopically. Senda were discussed with a mode-shape frequency spectrum (Hossain et al.,
et al. (2015) examined how the geometry properties affect the wrinkling 2007). It was found that creased membranes behave like separate
behaviour by including a path-tracking method in typical FEM. Same structure and have different vibration characteristics. Woo (2010)
geometries are discussed for wrinkling strain relationship with out-of- showed the direct impact of sheet thickness and thus bending stiffness
plane displacements and the effect of wrinkle influencing factors on displacements. The latest study by Kamaliya et al. (2021) for metal-
(Poisson’s ratio, Young’s modulus, pre-stress) (Kumar et al., 2015; Wang polymer creased laminates recommend the layer thickness range to
et al., 2016). Liu et al. (2017a) introduced the concept of wrinkle-wave maintain the desired reflector surface accuracy. In a nutshell, most of the
model to precisely describe the wrinkling during FEA. Though all ana­ work with different folding approach targets solar-sail structures. None
lyses have conferred wrinkling, they skipped any irregularities/defects of them focus on planar antenna reflectors. Single crease models were
in the profile. Wang et al. (2010) proposed a tape connection in a analyzed for the effect of layer thickness, load variation, and crease
parabolic reflector, and Shinde et al. (2021) offered a material combi­ orientation in displacement of rectangular/square geometries. They
nation for a planar profile near edges to minimize the wrinkling effect haven’t considered multiple fold-lines or a specific folding pattern. Be­
(Kumar et al., 2020). Fleurent-Wilson et al. (2014) validated the use of sides this, all studies were for smaller/lab-scale membrane samples of up
micro-fiber composite patches to localize the wrinkles on square and to 500 mm in maximum measurements. Hence, there is a scope for
rectangular membrane profiles. A fuzzy-logic controller (Orszulik and research in analysing large-scale planar antenna reflector, which follows
Shan, 2012) was also proposed for membrane wrinkling under thermal a systematic fold pattern and undergoes asymmetric loading post the
loading. Sipos and Fehér (2016) implemented realistic material prop­ deployment.
erties to maintain the surface flatness by alleviating wrinkles. This paper proposes the favourable fold-line arrangement(s) for 1 m
On the other hand, crease(s) exist in the reflector surface as a sharp, × 1 m deployed thin Kapton membrane antenna reflector. Various crease
inelastic discontinuity. It may be permanent (owing to regular pack­ position and orientation combinations are simulated, targeting mem­
aging fold) or temporary and irregular (during handling). It affects the brane behaviour for crease-wrinkle interaction. An experimental pro­
membrane surface similarly to the wrinkle, but with additional bending cedure, in Section 2, quantify the crease relaxation (captures the neutral
elements due to its plasticity nature. High surface accuracy and flatness angle), effective modulus, and induced residual stresses. These material
are top priorities in antenna structures (Chandra et al., 2021). Hence, the parameters are essential for consideration in fundamental analysis. It is
distortions caused by creases and wrinkles need to be eliminated by followed by a newly derived crease profile formulation to characterize
overcoming the control issues (Jenkins and Kalanovic, 2000). However, the 1D Z-fold arrangement with membrane shortening & constant crease
fold lines are unavoidable in compact packing of large gossamer space curvature. Section 3 integrates these into the finite element model. A
structures. Hence, the interest starts growing for two primary concerns: non-linear post-buckling numerical simulation is conducted for a real­
the physical definition of a crease and its implementation in the large istically deployed creased membrane. A compatibly study in Section 4
antenna reflector/solar-sails. explains uniaxially tensioned non-creased & transversely creased thin-
A couple of methods have been proposed for crease characterization. films. The effect of crease anisotropy on wrinkle profile (wavelength,
One of them is a tape spring approach by Seffen et al. (2000) and amplitude, and out-of-plane displacement) and stress transfer path are
extended later for three-dimensional folds (Walker and Aglietti, 2007). thoroughly investigated. That is presented as a function of crease ori­
By declining the direct implementation of this approach as a hinge joint entations in Section 5. Exploring the surface topology aspect for varying
in the crease model, Abbott et al. (2014) and Dharmadasa et al. (2018) crease orientation & number of folds gives insight for reflector accu­
claimed realistic implementation of crease deformation. Parallel work rateness. By comparing root means square errors (RMSE), convincing
started by considering the nature of crease during folding and sequential remarks are made for an appropriate way to reduce wrinkles, thereby
deployment (Thiria and Adda-Bedia, 2011), targeting the solar-sail & maintaining the desired surface accuracy. Section 6 compares the
antenna reflector applications. Researchers have focused on crease experimental test results with numerical simulations using a non-contact
formulation. Based on deployment, they defined a fold as an elastic measurement technique for the scaled model. Section 7 concludes the
(Jules et al., 2019; Lechenault et al., 2014; Lee et al., 2019), viscoelastic, paper by recommending favourable crease arrangements in a square-
elasto-plastic (Dharmadasa and Jiménez, 2021; Li et al., 2020; Satou and shaped planar membrane reflector.
Furuya, 2018; Satou and Furuya, 2011; Torisaka et al., 2019; Xia et al.,
2018), or plastic (Dharmadasa et al., 2020; Jules et al., 2020) in nature. 2. Methodology of a Z-fold profile in planar membrane
All these were limited to the fold-lines only. Found no direct physical
implementation in the large-size membrane reflector, where wrinkling Introducing a crease before deep-diving into the wrinkle-crease
plays an important role. Though, for origami design, a creased-beam interaction is essential. Idealizing a crease profile in thin-film mem­
model (Papa and Pellegrino, 2008) was considered to predict the brane is a challenging task in itself. Additionally, selecting an appro­
load–displacement response and to check crease non-linearity. Cai et al. priate folding method (Liyanage et al., 2021) is important considering
(2017) considered crease as a fixed-width wide strip in a piece of Miura- the shape, size, and application. Measurement of neutral angle relaxa­
ori folding design and studied stress distribution by varying loading tion, crease curvature, and effective material properties plays a vital role
nodes. Later, Shen et al. (2021) presented an improved solution for stress in executing a non-linear analysis.
distribution by replacing crease with connector elements having elasto-
plastic behaviour. It provides a good improvement, but connector ele­ 2.1. Experimental characterization of the creased specimen
ments increase the design complexity and computational time (Piyumi
and Mallikarachchi, 2020). However, these possibilities are applicable The present study starts with crease formation and measuring crease
only when the surface accuracy is of no concern. relaxation in a rectangular piece of polyimide Kapton film. The dura­
Limited research ended with analyzing the effect of a crease in the bility and wide working temperature range of Kapton make it suitable
wrinkle behaviour, which directly affects the surface accuracy of for space applications. We cut the coupons of 25 mm in width and 125

2
P.K. Kamaliya et al. International Journal of Solids and Structures 254–255 (2022) 111902

Fig. 1. A step-wise procedure of measuring fold-line neutral angle started with the preparation of folded Kapton membrane: (a) Kapton membrane after removing
fold force is followed by (b) Neutral angle measurement and (c) Experimental setup for measuring creased membrane properties.

Fig. 2. Crease angle relaxation after removing the folding force for thin strips with two different thicknesses, 1 mil and 2 mil.

mm in height from defect-free Kapton sheets of 1 mil & 2 mils. We have

Table 1
taken utmost care to avoid any distortion due to unwanted cuts and
Properties of experimentally tested Kapton films.
scratch marks. Next, they were lightly bent and kept between two glass
plates such that a compressive force was applied uniformly. After 6 hrs, Property Value

the samples were allowed to unfold naturally by taking out the upper Young’s modulus (GPa) 3.605
plate. The viscoelastic nature of a material (Abbott et al., 2014) is the Effective modulus of a creased membrane (GPa) 3.139
Residual stress in a crease region (MPa) 0.169
cause of stress relaxation in natural unfolding of samples. Images of
Yield stress (MPa) 64.83
equilibrium fold angle were captured at timely intervals using a Nikon Material density (kg/m3) 1420
D3500 camera (24.3 megapixels, AF-P 18-55 lens). Later, they were Membrane thickness (µm) 50.5
post-processed in SOLIDWORKS® to extract the crease geometry. Poisson’s ratio 0.34
Drawing straight lines fit to both halves gives the angle between two
sides of a fold, known as a neutral angle. Figure 2 displays the relaxation
which is available from the manufacturer. Combining them with pro­
mechanism for all eight samples, four each (1 mil and 2 mil). The average
cured curvature profile and residual stresses, FE analysis delivers better
initial angle (measured just after load removal) is 1100 ± 50 and 1270 ±
results. Their consistency is confirmed from experiments at later stage of
50 respectively. After 10 hrs, the change in neutral angle observed to be
the paper.
370.
A crease introduces inelastic behavior in the membrane (near fold
portion). Measuring the mechanical properties of the creased sample is
2.2. One-dimensional crease model with multiple fold-lines
the most suitable way to detect it. The effective modulus method is
implemented to record this dissimilarity. Creased samples were fixed to
The model in the paper uses a one-dimensional approach for ideal­
the universal tensile testing apparatus (INSTRON® made) using a spe­
izing a crease in thin-film. From different folding techniques, Z-folding
cial jaw arrangement. It is an additional fixture attached to the existing
with parallel fold-lines (zig-zag) is the choice for analysis. The reason
setup (image c, Fig. 1). Instead of a conventional measuring method, we
being, it can be further combined with wrapping to induce a two-
used a digital instrumental cluster to extract material properties. Eight
dimensional approach. Figure 3(a) represents a part of this Z-fold
samples of the creased specimen were tested in addition to four non-
model, wherein a dotted line is the concave fold, and a full line (in blue)
creased pieces. An average value of Young’s modulus came out to be
is convex. In Fig. 3(b), the top image A shows that the membrane opened
3605 MPa (±234 MPa). An effective modulus of the creased membrane
at a crease angle θ just after removing the folding force. Applying a
(Eeff) was obtained as 3139 MPa (±183 MPa). These values are within
tensile force at two ends unfold the membrane with a corresponding
one standard deviation. Table 1 summarizes all material properties ob­
deployment angle θt (image B, Fig. 3(b)). Pulling it further, the mem­
tained experimentally (Kamaliya et al., 2021) except material density,
brane reaches the configuration image C, Fig. 3(b), a baseline profile for

3
P.K. Kamaliya et al. International Journal of Solids and Structures 254–255 (2022) 111902

this paper. This baseline configuration comprises of few assumptions: single crease is
First, the circular cross-section for crease; Second, crease radius Rc is √̅̅̅̅
same as crease depth dc (distance from crease tip to the tensioned E
Δlc = t⋅ ⋅f (σ t ), (1)
membrane surface); Third, there is no effect of gravity. The distance dc is σt
variable as the membrane opens up from a neutral angle state to the fully
where, t is membrane thickness, E is the elastic modulus and σ t is the
flatten configuration.
force during membrane deployment. Modulus is a function of plastic
The authors have derived a tape-spring model for the planar square
formation (near crease) and elastic one (in the remaining part of the
membrane in this section. The systematically arranged fold-lines are √̅̅̅̅̅̅̅̅̅̅
membrane) with yielding limit √2̅̅3 2Eσ t (MacNeal and Robbins, 1967).
subjected to transverse loading. Folding of the membrane introduces the
crease in it. A fold can be idealized as a cantilever beam (Fig. 3(c)). The effective modulus (Eeff ) of a thin film should be introduced in the
According to elastic beam theory, a bending moment is expressed as baseline configuration itself. It depends on two parameters, the length of
Mb = σ t × dc , where, σt is the applied force. This force overcomes the a membrane and shortening caused due to crease. Thus, changing the
residual stresses generated due to the formation of a crease. Incorpora­ number of fold-lines, as in the zig-zag folding, considerably affect the
tion of internal opening fold angle α is a must for crease characteriza­ effective modulus. For transversely creased membrane, it is represented
tion. Deformed coordinates depend on bending rigidity, membrane size by
(width and thickness), and neutral angle. Combining all such parame­ 1 1 1 ∂(Δlc1 + Δlc2 + … + Δlcn )
ters, a generalized function should be f(x) = sin(α/2) + 2EI σt
(d2c − R2c ). = − ⋅ . (2)
Eeff E l ∂σ t
dy 1
Integrating the differential equation dx = f(x)⋅√̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
̅ over a crease
1− {f(x)}2
So, in general, for equally spaced multiple creases in transverse direction
shape gives the deflection of a beam undergoing large deformation. That
of loading
is helpful for crease characterization and to get the relation of folding
force with deformation. Simplifying this, considering a flat membrane 1 1 1 ∑n
∂(Δlci )
= − ⋅ . (3)
with constant crease curvature for baseline configuration, the calcula­ Eeff E l i=1 ∂σ t
tion for this radius of curvature is expressed as R2C = σ2EI
r bt
(1 − cosα0 ).
Here σ r denotes residual stress, EI is the flexural rigidity, t and b are Substituting value of Δlc from Eq. (1) in to Eq. (3),
membrane thickness and width. Angle α0 is the interior opening angle ( √̅̅̅ ) √̅̅̅̅̅
just after removing folding force. 1 1
= +n
2 − 1 ⋅t E
√̅̅̅ ⋅ , (4)
This crease(s) reduces the overall size of a membrane to l − nΔlc from Eeff E 6⋅l σ 3t
original length l of the non-creased flat profile. The shortening due to a

Fig. 3. (a) A 1D Z-fold pattern showing the equally spaced fold-lines and folding sequence, (b) Crease model of the membrane under deployment: [image A] is the
membrane state just after released folding force, [image B] which undergoes unfolding with some external tensile load [image C] to make it flat baseline config­
uration, (c) is the profile description of a single crease from fully folded state to the unfolded state.

4
P.K. Kamaliya et al. International Journal of Solids and Structures 254–255 (2022) 111902

and Table 2
[ √̅̅̅̅ √̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅ ] Cases for numerical simulation.
1 1 n⋅t A 4 E − 16E − A
= + √̅̅̅̅̅ √̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅ + √̅̅̅ . (5) No. Membrane No. of Folding Crease angle (with respect to
Eeff E 4⋅l⋅ σ 3t 6E − A 3 size creases loading direction)

σ2
1 1m×1m – Pristine –
where, A =3⋅ σyt . Eqs. (4) and (5) are appropriate for observing effective model
modulus in the elastic and plastic region (creased portion) of the 2 1 1D Z-fold 300
3 2 pattern 450
membrane, respectively. For structures undergoing large deflection, a 4 3 600
geometrically non-linear model with such effectiveness describes its 900 (transverse crease)
behavior precisely.

3. Finite element model of creased membrane by pre-stressing the membrane before actual loading.
A crease is introduced in the model in the following fashion. First, the
A commercial software-package Abaqus/CAE is used to create non- cross-section is assumed to be a constant circular profile with a stress-
creased and creased membrane geometries. The FE model is prepared free initial membrane configuration. Crease radius accounts for the
for analyzing the wrinkle formation, evolution (with and without the unfolding computation from neutral angle relaxation. That will
presence of creases), and wrinkle-crease interaction. Wrinkle formation compensate for a nonlinearity convergence. Second, an effective
occurs when a critical buckling load is reached during applied modulus of the membrane with multiple folds from Eqs. (4) and (5) are
compressive stress. It can be induced by uneven loading, edge shearing, introduced in this baseline configuration. Effective moduli differentiate
or in-plane twist. A square-shaped Kapton film of 1 m × 1 m and 2 mil the creased and non-creased regions of the geometry. Furthermore, the
thick, in the elastic-plasticity model, is considered for wrinkling anal­ residual stresses account for the pre-stress in the geometry. In earlier
ysis. This planer membrane reflector undergoes tensile loading to cases, only transversely creased membrane was treated for tensile
distinctly catch the wrinkle formation and identify the direct effect of loading by introducing a single fold-line. They neglected either residual
crease on wrinkles. Applying a point load on a sharp corner of the thin- stress or effective modulus. The effectiveness of our technique can easily
film cannot provide enough room for out-of-plane displacement. Hence, be compared with geometry unfolding from the fully folded state. In
a truncated shape of a square membrane over a standard case is chosen. that, a connection between two halves was established using a connector
Two opposite ends were truncated with a width of 0.4 m while keeping element exhibiting a fixed hinge–deformable panel configuration.
the other two as fixed points to hold the membrane in a torus. By doing Next, selecting mesh elements plays a crucial role as they should
so, wrinkles evolve along the loading path, which directly intersects include transverse shear deformation effect, large rotations, and finite
with the crease, thus benefits in observing the wrinkle-crease interaction membrane deformation. Three-node linear element, S3 (a degenerated
behaviour more closely. form of S4), was the primary choice because of its high accuracy.
The finite element model of a meshed membrane with a zoomed-in However, its constant bending and membrane strain approximation
crease region is shown in Fig. 4(a). Side nodes A and B are con­ requires higher mesh refinement and thus high computational time for
strained in all DOFs except translation in Z-direction throughout the the solution. Hence, the reflector is modeled using S4R, a 4-node,
analysis. Top and bottom edges are applied forced displacement quadrilateral stress/displacement shell element with reduced integra­
(increased in every load step) in the direction shown in figure. All other tion and a large-strain formulation. After a mesh convergence study,
directional movements are restricted at those edges. With the same refined mesh size in a crease cusp region totaling 480 elements and 7680
boundary and loading conditions, total 13 models are prepared based on elements for the complete planar membrane.
one-dimensional Z-folding approach. The combination of number of The non-linear post-buckling analysis (flow chart, Fig. 5) method is
creases and its orientation totals 12 models, and one is the non-creased used to investigate crease-wrinkle interaction. In the base step itself, pre-
flat membrane. Table 2 shows all such combinations, and a couple of stress of magnitude 0.4 MPa is applied to overcome the residual stresses.
them are graphically presented in Fig. 4(b). Crease orientation is defined In the first step, displacement loading of magnitude 0.05 mm is applied
as an angle between the load direction and clockwise rotation of a fold- at the top and bottom edges. That induces out-of-plane displacement in
line (Fig. 4(b)). Crease radii are calculated and summarized in Table 3. small magnitude and generates wrinkles. Next, the wrinkling mode se­
Data also depicts respective residual stresses that need to be overcome lection from this eigenvalue buckling analysis should include all

Fig. 4. The said Z-fold crease arrangement fold-lines are transformed into (a) a square-shaped creased membrane simulation model of the size 1 m × 1 m along with
(b) other models of the same dimension, but with varying number of creases and its orientations.

5
P.K. Kamaliya et al. International Journal of Solids and Structures 254–255 (2022) 111902

Table 3
Properties of crease profile for different orientations.
Residual stress from creasing (MPa) Crease radius (mm)

Crease orientation (in degrees)

Number of creases 30 45 60 90 30 45 60 90

1 0.195 0.239 0.338 0.169 3.345 4.219 5.967 2.983

2 0.256 0.313 0.443 0.221 3.011 3.687 5.214 2.607
3 0.300 0.367 0.519 0.260 2.781 3.406 4.817 2.408

quantitative characteristics for imperfection. All the displacement mode technique to solve a non-linear equation. It is an effective method where
shapes have magnitude in millimeters, but the imperfection usually re­ a computation starts from the last converged sub-step rather than the
lies on the first/second mode. Therefore, imperfection amplitude is first footstep. The subsequent section considers this perticular analysis
determined to be 1.0 × 10-6 m. It reproduces the same value when procedure and represents the detailed investigation of all the cases
membrane thickness is multiplied with a scaling factor of 2–10 %. In our (Table 2).
case, the second eigenvalue buckling mode shape is selected as a base for
the imperfection distribution region. 4. Simulating the membrane with transverse creases
The authors performed a non-linear post-wrinkling analysis in the
third and last step. The imperfection was introduced in the buckling Before interpreting wrinkles and crease-wrinkle interaction behav­
mode outcomes through subroutine. An imperative vertical displace­ iour (basically a post-deployment analysis), energy history should be
ment of 1 mm is applied to the tension edges in the post-buckling step to verified to find the model’s robustness. Typically, the system’s total
perform wrinkle-crease interaction. Few strategical implementations energy ETOTAL is equal to the energy introduced externally. This external
boost the convergence: First, the load function was replaced with energy is in any of the following forms: kinetic energy(ALLKE ); internal
displacement loading, say, tension displacement. This displacement work (ALLIW ): summation of artificial energy(ALLAE ), elastic & inelastic
control technique smoothens the wrinkling computation in the post- strain energy; viscous dissipation(ALLVD ); external work(ALLWK ). A
buckling phase. Second is the use of Newton-Raphson iteration combined form of mentioned terms creates an energy balance equation

Fig. 5. Flowchart reflecting the step-wise procedure of post-buckling analysis for crease-wrinkle interaction.

6
P.K. Kamaliya et al. International Journal of Solids and Structures 254–255 (2022) 111902

Fig. 6. Energy variation comparison for (a) non-creased membrane, along with (b) single creased, (c) double creased, and (d) triple creased Z-fold membrane. All
cases pertaining to the fold-lines modeled perpendicular to the applied force direction.

for a solar-sail/reflector antenna. crease vicinity, which is not the case with a non-creased profile. The area
near loading edges is almost flat and wrinkle-free with a single spike.
ETOTAL = ALLIW + ALLKE + ALLVD − ALLWK . (6)
Despite higher stress concentration in double and triple creased models,
Energy balance of the system indicates no discrepancy and adequate they keep the wrinkles concentrated in crease vicinity. It is a good sign
system convergence. Else, necessary actions are must to keep an analysis when we anticipate a wrinkle-free profile.
free from numerical instabilities. Energy plots in Fig. 6 for pristine and Figure 8 sights the surface configuration of the wrinkled membrane
creased models show zero value throughout the simulation for total after applied tensile loading. We have already seen wrinkling phenom­
energy, ensuring a conversed solution. For non-creased membrane, enon, which appears along the applied load direction and partially on
artificial energy rises with internal energy during the initial stage. This the fixed sides. Besides, the major principal stress direction agrees well
rise is limited to 4 % of the total internal energy. The same value is found with the wrinkle line. Vector diagrams (with red color) indicate the
to be 0.6 %, 0.2 %, and 0.3 % for single, double & triple creased direction of major principal stresses from wrinkling analysis adopting a
membranes respectively (see inset diagrams). Artificial force should be tension-field theory. Comparing both, the earlier stated remark on stress
applied to ensure no numerical distortion. Mesh elements used in the concentration stands true.
analysis favor in restraining artificial energy within a suggested limit of Following the wrinkle formation and evolution, the authors inves­
2 %. That increases total simulation run time and makes the simulation tigated the fold-line behaviour in these membranes. Out-of-plane
computationally expensive. Yet, it improves the overall accuracy of re­ wrinkle deformations for creased thin-films are plotted against
sults. To ensure minor/no effect of inertia, the kinetic energy should not normalized distance along positive X-direction at four crucial locations
exceed 5 % of internal energy at any time interval. The plots in Fig. 6 (Fig. 9(b-e)). Two are near the centre (at section XC-XC, 10 mm from mid-
indicate kinetic energy is a tiny percentage of internal energy. line) and near loading edges (at section XL-XL, 45 mm from the top edge.
The remaining two are at mid-sections (section XT-XT and XB–XB, 250
mm on either side of mid-line). Selection of this normalized distance over
4.1. Stress distribution and wrinkle evolution
actual distance is good for comparison fairness. Observation says that a
systematic Z-folding reduces wrinkle amplitude and increases the
After verifying energy history, surface profile contours of the non-
wavelength. Maximum wrinkle amplitude along XC-XC reduced to
creased & creased membranes are compared for the same boundary
0.4192 mm for double creased profile compared to 0.524 mm for non-
and loading conditions (Fig. 7). Primary results for pristine membrane
creased configuration. Furthermore, this value is close to that of a tri­
are compared to those of Wang et al. (2009) and found similar for a
ple creased profile (0.5914 mm). Conversely, a single crease arrange­
given 1 mm force–displacement on tensioned edges. According to Fig. 7
ment has shown an increment to 3.205 mm in maximum. Moreover, the
(a), wrinkle distribution in a non-creased membrane is uniform and
average wrinkle wavelength is 0.18 m and 0.22 m for double and triple-
streamlined along the loading direction. There are eight picks and
creased Kapton films compared to 0.1 m for non-creased geometry.
troughs in the wrinkle profile. These exact numbers are observed and
Minor principal stress directions are of concern for thoroughly
verified in Fig. 9(b-e). Similarly, results were found for creased mem­
investigating a wrinkle profile (i.e. wrinkle amplitude, wavelength, and
branes, but they witnessed the effect of fold-line on wrinkle distribution.
the number of wrinkles). Figure 10 compares the minor principal
An increasing number of folds creates higher stress concentration
stresses over the section XC-XC with those at loading edge XL-XL. These
and higher values for out-of-plane displacement. The wrinkling defor­
results are extracted from wrinkling analysis of non-creased as well as
mation is still symmetric on opposite sides of crease(s) (top and bottom
creased models using a non-linear method applicable for shell elements.
parts) with higher concentration in the central region (Fig. 7(c) and (d)).
Figure 10(a) shows lower stress values near the loading edge than at the
Also, the wrinkling waves are scattered over the entire reflector surface
central region. It conveys no excessive compressive stress generation,
compare to pristine and single crease models. The number of picks and
and the assumption of the in-extension theory stands appropriate for
valleys concurrently reduces to three/four. Results also depict how the
wrinkle edges. Results in Fig. 10(b) highlight the stress concentration
crease affects the stress transfer path. Stress gets concentrated in the

7
P.K. Kamaliya et al. International Journal of Solids and Structures 254–255 (2022) 111902

Fig. 7. Wrinkle configuration contours for (a) non-creased, (b) single, (c) double, and (d) triple creased membrane shows how the wrinkle distributes over the
membrane surface when the fold-lines are perpendicular to the direction of force.

Fig. 8. Direction and magnitude of stress outcomes in the surface configuration of (a) non-creased and (b) double creased wrinkled membrane shows the stress
concentration in the vicinity of crease with higher magnitude. Wherein the non-creased profile shows equal magnitude along the force line.

8
P.K. Kamaliya et al. International Journal of Solids and Structures 254–255 (2022) 111902

Fig. 9. Measuring the out-of-plane displacement (a) at various locations on membrane to observe the wrinkle effect. Observations are made along a horizontal path
(b) near central region XC - XC, (c) near load edge XL - XL, and at section (d) XT–XT and (e) XB–XB mid-way towards the top and bottom edges from horizontal mid-
line AB.

9
P.K. Kamaliya et al. International Journal of Solids and Structures 254–255 (2022) 111902

Fig. 10. Minor-principal stress comparison on the center-line XC–XC and near loading edge of the wrinkled membrane XL–XL for (a) non-creased and (b) creased
profiles. With the increasing folds, a comparison of min. principal stress (c) over central section XC–XC and (d) for the whole membrane gives insight into the wrinkle
stress distribution.

effect near the crease. Estimated outputs gradually reach the maximum factors narrowly linked to the bending stiffness. These are the fold-line
amplitude as the number of fold-line increases (Fig. 10(c)). Unlike the (s), their style, numbers, location, and orientations. This paper targets
central section, a minimum stress value over the whole reflector surface the impact of these factors on wrinkling parameters (i.e., wrinkle
is of much smaller amplitude (Fig. 10(d)). That is due to discontinuous amplitude & wavelength, number of wrinkles & wrinkle angles, etc.).
wrinkle strain distribution in the spatial direction, as an effect of stress Numerically investigating a square membrane under tensile loading
attention near fold-lines. (Fig. 4) would offer a convincing recommendation for fold-line ar­
rangements. Based on that, the authors aimed to propose the possible
5. Analysis of crease-wrinkle interaction behaviour combination(s) that have controlled wrinkle-affected area and cause
minimum impact on surface accuracy.
Residual stress from creasing and different boundary conditions
affect the membrane’s stress state, thereby influencing the wrinkle 5.1. Wrinkle influencing factors
deformation. Factors that disturb the local buckling, Young’s modulus
and Poisson’s ratio, were analyzed extensively in the past. On the other The authors studied wrinkling parameters for two key locations: one
hand, Kapton’s local buckling under compression is influenced by the is near the crease, and the other is near the loading edge. Results for lines

Fig. 11. A sectional view of the wrinkled membrane. The profile on the left is a wrinkle geometry and related information while the right-side diagrams illustrate the
sectional view of a single wrinkle.

10
P.K. Kamaliya et al. International Journal of Solids and Structures 254–255 (2022) 111902

perpendicular to the minimum principal stress direction keep the trace

of wrinkle amplitude straight. Their interaction with the wrinkle profile
gives the maximum out-of-plane displacement and inclination to the
horizontal mid-line (Fig. 9(a)) defines the said wrinkle orientation.
Figure 11 is a surface outline of the creased planar membrane after in-
plane stretching. Here, a wrinkle profile indicates the major principal
stress direction. In addition, the appearance of wrinkling angle is a
consequence of the crease formation at a specific angle. Zoomed-in re­
gion, perpendicular to major principal stress, is a typical wrinkle profile.
This cross-sectional shape of wrinkle geometry gives a measurable view
for wavelength and amplitude.
Distance between two consecutive traces is a measure of wrinkle
wavelength (λn ), and the height of an individual trace delivers the
maximum amplitude (An ). Its average from a particular wavelength and Fig. 13. Trend of maximum deformation change over whole Kapton sheet, for
varied number of creases.
number of wrinkles (half waves) reflects wrinkling deformation. When
combined with average out-of-plane distortion, this deformation is
considered to be of a global level. Henceforth, the authors investigated and its orientation, which alter the distribution property of stress and
their changes under the effect of fold-line variations. The outcome affect the stress transfer path. That is discussed in the later part of this
benefits in preserving desired surface flatness by limiting the overall section.
surface RMS error. Effects of altered crease orientations are discussed here. The abscissa
in Fig. 13 is the crease orientation, whereas the ordinate represents
5.2. Effect of crease orientation maximum displacement. It is clear from the figure that the maximum
out-of-plane displacement decreases as the orientation angle becomes
This section discusses the effect of crease orientation on wrinkling normal to the applied tensile loading direction. In the case of multiple
parameters by comparing them with a non-creased wrinkled Kapton folds, double and triple crease arrangement shows a descending trend.
membrane model. The analysis was performed for four different crease Though it improves results, maximum displacements are much higher
orientations (300, 450, 600, and 900). Our focus remains on the locations for 300 and 450 orientation angles. They are observed to be 3.696 mm &
where wrinkles are prominently affected by crease and vice-a-versa. 2.732 mm than 0.527 mm for pristine geometry. Double creased mem­
brane observes minimum deviation in results due to the crease geometry
5.2.1. Out-of-plane displacement (1.323 mm, 1.254 mm, and 0.9054 mm). Z-folding arrangement formu­
The discussion begins with the results captured in central horizontal lates both creases in a totally reverse direction, thus contradicting each
section XC - XC. Figure 12(a) shows out-of-plane displacement curves other’s effect. This condition stands true for all cases with even number
with reference to the non-creased Kapton. Contour profiles of wrinkling, of fold-lines. The figure also shows that the examined results drastically
observed from the figure, exist just after stretching the membrane edges. reduce after 600 crease orientation for triple creased membrane. How­
As the crease orientation changes, out-of-plane displacement also varies ever, the drop is almost linear for single and double creased Kapton
from point to point. Each displacement curve reaches a maximum in the profiles when the inscribed alignment exceeds 500.
central area and remains relatively flat in the other portions. In parallel,
the number of wrinkles shrinks to two–three from seven-eight in the 5.2.2. Wrinkle-amplitude and wavelength
non-creased model. Figure 14 plots the mean wrinkle amplitude valued for defined
Subsequently, the results near the load edge (along XL - XL) of membrane models. Traces of amplitude are assumed to be straight and
wrinkled region are also discussed to examine the overall wrinkle-crease parallel for easier calculation. Width of individual peaks & valleys
interaction effect. Figure 12(b) compares displacements from original guides in extracting the mean wrinkle amplitude. As shown in the figure,
un-stretched conditions for all crease orientations and pristine models. each amplitude curve shows an upward trend for increased crease
Membrane with transverse crease shows a similar trend and attained a orientation. An overall mean amplitude is much higher than the pristine
maximum value at side corners. Results for other three cases are quite model due to fold line geometry.
different, constrained geometrically, because of crease orientations. In Fig. 15, the moving average of half-wave over wavelength-wise
Wherein the displacement is relatively high in magnitude and uneven in peaks is plotted against mean amplitude. The analysis is performed on
nature. Figure 12(a) observes a higher wrinkle profile spike than the single creased models with different crease orientations. The effect of
outputs near the load edge. The reason for this difference is the crease crease angle variation on wrinkle profile is the primary concern here. No

Fig. 12. Comparison of surface displacement for single creased membrane with four different orientations with pristine membrane (a) over central cross-section and
(b) near load edge for normalized distance in positive X direction.

11
P.K. Kamaliya et al. International Journal of Solids and Structures 254–255 (2022) 111902

significant variation was observed in the average half-wavelength,

though the angle changes. Maximum values (19.903 mm and 20.249
mm) attained for non-creased membranes are at two corners. The profile
is almost flat in the central region with a relatively very small amount of
1 to 2 mm. Completely reverse trends are observed for 300, 450 & 900
crease orientations with maximum amplitudes of 13.946 mm, 11.479
mm & 10.718 mm respectively. Comparing with results found over the
central cross-section in Fig. 12(a), both agree well. A concluding remark
is that the lower the wavelength, the higher the number of wrinkles and
vice-a-versa. This alteration in wrinkle amplitude and wavelength re­
veals the effect of fold-lines on wrinkling. Principally, a crease concen­
trates the stress in its locality and strengthens the membrane bending
stiffness.
Fig. 14. Variation of mean wrinkle amplitudes at the centre of the wrinkled
region versus the non-creased membrane (a cross-mark) and crease orienta­
tions, for varied number of fold-lines. 5.3. Effect of number of fold-lines

Numbers of fold-line were molded considering shape, size, and

different folding arrangements for compact membrane structures. So, it
is essential to analyze the effect of these multiple fold-lines on stress
concentration, wrinkling parameters, and surface accuracy before
putting the system into actual practice. In this section, the authors have
investigated three different cases with single, double, and triple crease in
addition to those discussed in Section 4. Boundary conditions and other
parameters are also carried forwarded.

5.3.1. Wrinkle profile

Figure 16 shows stress distribution, at maximum strain energy points
(Fig. 6), for all combinations of creases and their orientation angles. The
detailed results are discussed and compared subsequently in this section.
Fig. 15. Average half-wavelength variation over different crease orientations Wrinkle direction is mainly affected when a flat membrane undergoes
for single creased membrane and comparison of the same with behaviour in shear or twisting. In our case, that is due to the fold-line orientation and
pristine membrane geometry. that too for planar membranes under in-plane tension. Wrinkle direction
changes in the same fashion as the fold-line orients. When we change the
crease number from single to two & three, the number of wrinkles,

Fig. 16. Wrinkle contours and its orientations when fold lines are positioned at 30 degrees, 45 degrees, and 60 degrees with respect to loading direction: for varying
number of creases from one to three.

12
P.K. Kamaliya et al. International Journal of Solids and Structures 254–255 (2022) 111902

Table 4 amplitude, and wrinkle affected area reduces. However, at the same
Comparison of wrinkle affected region for different crease orientation cases. time, wrinkle wavelength increases. Here, the crease(s) alters the
Wrinkle affected region (% of total surface area) wrinkle path and spreads it to a wide-span over a thin-film surface. That
results in a reduced wrinkle-affected area and improved surface accu­
Number of Crease profile 30 45 60 90
creases angle (degrees) racy. Statistics in Table 4 are somewhat different because of the change
in wrinkle distribution property. Tension-field theory is applied to
Pristine 0.55735 – – – –
1 – 0.37657 0.07917 0.16644 0.39739
differentiate between taut and wrinkled regions. In Fig. 17, the black
2 – 0.24173 0.07918 0.10223 0.13335 area identifies the wrinkled region, while the white represents the taut
3 – 0.15031 0.63622 0.52759 0.08442 state. Comparison convinces the expansion of the wrinkled region’s
width over the entire thin-film, when the number of fold-lines is
increased.
Results of crease number variation are analyzed in detail, and a
comparison is shown in Fig. 18. As seen from Fig. 18(a), the mean
wrinkle wavelength deviation is negligible for an increased number of
folds. Though it shows a little increasing trend, overall half-wavelength
ranges near that of the non-creased membrane. 300 and 450 crease
orientation profiles show deviation for an odd number of fold-lines. As
even numbers counter the effect of each other, it causes stress diversion
from its actual trend. Comparing half wavelength with graphs in Fig. 18
(b), mean wrinkle amplitude falls and wavelength rises. That is a good
sign as far as the target of a flat reflector surface is concerned. The lowest
mean wrinkle amplitudes are 0.35237 mm, 0.45089 mm, and 0.09645
mm for altered crease; 450, 900, and 300 orientations. For a non-creased
membrane, this value is 0.25587 mm.

5.3.2. Surface accuracy

These values for mean wrinkle amplitude are under 3 mm, as seen in
the previous subsection. Hence, expecting an overall surface roughness
to be within the desired limit of 1 mm. Displacements from flat, un-
stretched thin-film are calculated for all analyzed cases and summa­
rized in Table 5. The surface precision RMS value is measured using Eq.
Fig. 17. Wrinkled region is given by the wrinkling analysis using tension-field
(7).
theory. The black portion indicates the wrinkled area, and the white one is the √̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
1∑
taut region. RMS = (x − x)2 , (7)
n

where, n represents the number of nodes, x is the mean value of

amplitude, while (x − x) represents the distance of a node’s current po­
sition from initial ideal location of a membrane plane.
From all angle variations of the crease, a double creased membrane
observed the lowest surface error value. The second crease will oppose
the out-of-plane wrinkling deviation caused by the first crease, thus,
reducing the overall surface error. Root Mean Square Error (RMSE) is
found to be minimum for non-creased membrane (≈ 0.2 mm) for obvious
reasons. However, crease(s) in the membrane is unavoidable as far as
larger reflectors are for use. For this practical aspect, triple-crease at an
angle of 90 degrees helps limit RMSE value to 0.377 mm, a minimum of
all creased membrane reflectors considered for analysis. That happened
because wrinkling could not favor as creases are in the transverse

Table 5
Surface error (RMS) for the different cases of crease considered.
Angle of crease orientation Number of Surface error (RMS)
(degrees) creases (mm)

Nil Non-creased 0.194

90 1 1.400
2 0. 377
3 0.626
60 1 0.816
2 0.615
3 1.000
45 1 0.985
2 0.554
3 1.400
30 1 1.200
Fig. 18. (a) Mean half wavelength and (b) mean wrinkle amplitude for
2 0.582
different crease orientations at horizontal mid-line for three cases, showing the
3 0.724
effect of the number of fold-lines.

13
P.K. Kamaliya et al. International Journal of Solids and Structures 254–255 (2022) 111902

2004), is used to obtain crease-wrinkle interaction details. The tech­

nique uses the triangulation principle and analyzes two-dimensional
photographs to reconstruct a three-dimensional profile. Camera cali­
bration is the first step, followed by target marking & camera arrange­
ments, capturing photos from different angles, and image-processing.
Experiments in this paper use the printed dot targets to obtain test re­
sults of small out-of-plane displacements. The calculated ideal target
diameter is 1.03 mm, which relies on sample dimensions and image
pixels. The distance between two consecutive dots is 7.5 mm, totaling
325 such dots over the whole membrane surface. Pictures were taken
from six different angles (four - in the same plane - at 450 angular in­
tervals, two - from top and bottom) and analyzed using image-
processing software. Figure 20 shows the creased specimen’s arrange­
ment. It is connected to the load cells from top and bottom, while the
side corners are kept fixed. Initial values of load cells are tare to remove
the effect of self-weight. The tension force on the membrane is gradually
increased to 21.48 N (equivalent to 1 mm displacement) with digital load
cells. Captured photographs are post-processed and reconstructed as a
three-dimensional profile.
Fig. 19. Surface accuracy radar for combination of creases and its orientation
helps in getting the surface roughness measure and selection of suitable com­
bination among the analyzed cases. 6.2. Comparison of experimental and numerical results

direction, and creases block the wrinkling path. Figure 19 and Table 5 Targeting wrinkle characteristics, the same 1/5th scale model with 1
also illustrate mostly good agreement over surface accuracy (aka. RMSE) mm displacement loading is simulated and compared for both test cases.
for double-creased membrane reflectors, as wrinkles are contained only The experimental mean wrinkle amplitude is 6.201 × 10-4 m for case A:
in the region between creases. triple creased (arranged transversely, α = 900 ) and 3.388 × 10-4 m for
case B: double creased (fold-line orientation, α = 450 ). On the other end,
6. Experimental verification numerical analysis outcomes are 5.957 × 10-4 m and 2.995 × 10-4 m,
respectively. Displacements on the central cross-section are compared in
6.1. Experimental set-up and procedure for loading analysis Fig. 21 for experimental results and numerical simulations.
Tables 6 and 7 compare the displacement induced at the central lo­
Wrinkling experiments are performed on 1/5th scale model (200 mm cations. Point IDs are from 3D point cloud data. The percentage error in
× 200 mm) to verify the crease position effects on the wrinkled mem­ the experimental and simulation results is less than 11 %.
brane. A fully non-contact method, photogrammetry (Black and Pappa, Consistency in wrinkle distribution results and overall wrinkling

Fig. 20. Experimental procedure of crease-wrinkle interaction analysis in tensioned membrane.

14
P.K. Kamaliya et al. International Journal of Solids and Structures 254–255 (2022) 111902

Table 7
Comparison of displacements for double creased membrane case B.
Point Numerical outcomes (×10-4 Experimental values (×10-4 Error (%)
ID m) m)

411 0.1907 0.2311 17.49

415 0.1925 0.2312 16.75
419 0.4758 0.2298 − 107.07
423 1.3315 1.2970 − 2.66
427 1.2414 1.5930 22.07
430 1.5343 1.7180 10.70
434 3.4869 2.8570 –22.05
437 0.2253 0.2884 21.89
440 4.9226 3.9090 − 25.93
443 10.1732 8.9370 − 13.83
446 9.7364 10.9690 11.24
449 2.3333 5.9760 60.96
452 10.2958 9.0330 − 13.98
455 9.7110 10.3040 5.75
458 5.3374 7.0670 24.47
461 1.3513 3.1160 56.64
464 2.6564 3.1520 15.72
467 2.7545 3.1630 12.92
470 0.2642 1.5030 82.42
473 1.8624 2.2450 17.04
476 1.1325 0.9030 − 25.41
479 0.2924 1.1340 74.22
482 0.1907 0.5730 66.72
485 0.1902 0.8840 78.49

The difference in RMS error for the whole membrane reflector surface is
3.1 % and 4.5 % in both cases, respectively, which is reasonable and
acceptable.
Fig. 21. Comparison of numerical outcomes and experimental results for (a)
case A: triple creased membrane (fold-lines arranged transversely, α = 900 )
7. Conclusion
and (b) case B: double creased profile (fold-line orientation, α = 450 ).
This paper analyses the wrinkle-crease interaction for deployed
Table 6 planar membrane reflector. The crease-lines are in a systematic 1D Z-
Comparison of displacements for triple creased membrane case A. fold arrangement. Based on a one-dimensional membrane model, the
study derives analytical relations for multiple folds considering crease
Point Numerical outcomes (×10-4 Experimental values (×10-4 Error
ID m) m) (%)
plasticity properties. The 50.5 μm thin-film geometries are modelled
using shell elements. We attained out-of-plane wrinkle deformation by
294 3.2043 3.540 9.48
extracting buckling mode. Moreover, geometric imperfections are
232 4.8497 4.570 − 6.12
233 6.2771 6.690 6.17 introduced to obtain convergence by force stabilization. The authors
234 6.7918 7.230 6.06 investigated the effect of various crease orientations and number of
235 6.1480 5.950 − 3.33 creases on wrinkle profiles. An experimental comparison with the lab-
236 4.1579 4.160 0.05
scale model justifies these analyses’ important conclusions.
146 0.7436 1.610 53.82
238 4.5401 5.890 22.92
The introduction of fold lines increases stress concentration. It ben­
284 8.7659 10.050 12.78 efits in reducing the wrinkle-affected regions, thus maintaining the
241 10.7429 9.940 − 8.08 overall surface accuracy of the planar membrane reflector. Crease
242 10.6119 9.950 − 6.65 changes the wrinkling direction and influences the stress transfer path as
243 10.8084 10.490 − 3.04
well. Compared with a non-creased membrane, both the wrinkle
244 8.8484 10.490 15.65
245 5.2523 6.470 18.82 amplitude & wavelength increases for creased models. The only
246 2.4878 3.230 22.98 exception is a double creased membrane. Surface configuration and
247 3.0526 2.150 − 41.98 vector diagrams convey that, implementing a systematic folding pattern
248 6.0799 5.380 − 13.01 counters the effect of mirror profiles. Hence, the surface accuracy is
249 6.7460 7.190 6.18
250 6.2497 5.770 − 8.31
found best in double creased membranes. That makes it the most suit­
251 4.8370 5.210 7.16 able choice among all. For crease orientation, the maximum surface
252 3.9029 4.250 8.17 accuracy (average of three cases) is in the case where creases are
transverse to the applied external load. However, a double creased
membrane is the most suitable geometry. Hence, fold-lines should be
perpendicular to the applied load direction to resist the wrinkle
prediction reveals that our computational analysis can simulate the
evolution.
experiments accurately. The slight difference is the influence of the
surrounding environment, accuracy with which photographs were
taken, or assumed constant crease curvature in finite element analysis.

15
P.K. Kamaliya et al. International Journal of Solids and Structures 254–255 (2022) 111902

Declaration of Competing Interest Jules, T., Lechenault, F., Adda-Bedia, M., 2020. Plasticity and aging of folded elastic
sheets. Phys. Rev. E 102, 1–8. https://doi.org/10.1103/PhysRevE.102.033005.
Kamaliya, P.K., Upadhyay, S.H., Mallikarachchi, H.M.Y.C., 2021. Investigation of
The authors declare that they have no known competing financial wrinkling behaviour in the creased thin-film laminates. Int. J. Mech. Mater. Des. 17,
interests or personal relationships that could have appeared to influence 899–913. https://doi.org/10.1007/s10999-021-09559-5.
the work reported in this paper. Khalil, S., Belaasilia, Y., Hamdaoui, A., Braikat, B., Damil, N., Potier-Ferry, M., 2020.
A reduced-order modeling based on multi-scale method for wrinkles with variable
orientations. Int. J. Solids Struct. 207, 89–103. https://doi.org/10.1016/j.
Acknowledgments ijsolstr.2020.10.002.
Kumar, S., Upadhyay, S.H., Mathur, A.C., 2015. Wrinkling Simulation of Membrane
Structures under Tensile and Shear Loading. Columbia Int. Publ. J. Vib. Anal. 3,
This work was supported by the Department of Science and Tech­ 17–33. https://doi.org/10.7726/jvamc.2015.1002.
nology, Govt. of India under grant [No. DST/INT/SL/P-27/2016] and Kumar, S., Upadhyay, S.H., Koryanov, V.V., 2020. Research and modeling of wrinkles
Ministry of Science Technology and Research, Sri Lanka, under grant and control of rectangular membrane structures with high-class modeling in on-orbit
conditions. IOP Conf. Ser. Mater. Sci. Eng. 812 (1), 012011. https://doi.org/
[No. MTR/TDR/AGR/03/02/09]. The authors are thankful to Professor 10.1088/1757-899X/812/1/012011.
Marian Wiercigroch & Professor S. Kyriakides for their constructive and Lechenault, F., Thiria, B., Adda-Bedia, M., 2014. Mechanical response of a creased sheet.
valuable suggestions for improving the manuscript. Phys. Rev. Lett. 112, 1–5. https://doi.org/10.1103/PhysRevLett.112.244301.
Lee, N., Pellegrino, S., 2014. Multi-layered Membrane Structures with Curved Creases for
Smooth Packaging and Deployment, in: Spacecraft Structures Conference, AIAA
References SciTech. American Institute of Aeronautics and Astronautics, Maryland, USA, pp.
1–20. https://doi.org/10.2514/6.2014-1037.
Abbott, A.C., Buskohl, P.R., Joo, J.J., Reich, G.W., Vaia, R.A., 2014. Characterization of Lee, T.U., Yang, X., Ma, J., Chen, Y., Gattas, J.M., 2019. Elastic buckling shape control of
Creases in Polymers for Adaptive Origami Structures, in: ASME (Ed.), Smart thin-walled cylinder using pre-embedded curved-crease origami patterns. Int. J.
Materials, Adaptive Structures and Intelligent Systems. Newport, Rhode Island, USA, Mech. Sci. 151, 322–330. https://doi.org/10.1016/j.ijmecsci.2018.11.005.
pp. 7480:1–7. https://doi.org/10.1115/smasis2014-7480. Li, M., Li, X., Che, L.X., Li, F.W., Kang, Z., 2020. Non-uniform global-buckling and local-
An, N., Li, M., Zhou, J., 2020. Modeling SMA-enabled soft deployable structures for folding in thin film of stretchable electronics. Int. J. Mech. Sci. 175, 105537 https://
kirigami/origami reflectors. Int. J. Mech. Sci. 180, 105753. doi.org/10.1016/j.ijmecsci.2020.105537.
Arya, M., 2016. Packaging and Deployment of Large Planar Spacecraft Structures. Li, Y.L., Lu, M.Y., Tan, H.F., Tan, Y.Q., 2012. A study on wrinkling characteristics and
California Institute of Technology. dynamic mechanical behavior of membrane. Acta Mech. Sin. Xuebao 28, 201–210.
Black, J.T., Pappa, R.S., 2004. Photogrammetry and videogrammetry methods for solar https://doi.org/10.1007/s10409-011-0512-2.
sails and other gossamer structures. In: In: 45th AIAA/ASME/ASCE/AHS/ASC Liu, M., Huang, J., Wang, Y., 2017a. Analysis of wrinkled membrane structures based on
Structures, Structural Dynamics and Materials Conference. American Institute of a wrinkle-wave model. AIP Adv. 7 (1), 015301.
Aeronautics and Astronautic, Palm Springs, California, pp. 1671–1681. https://doi. Liu, Z.-Q., Qiu, H., Li, X., Yang, S.-L., 2017b. Review of large spacecraft deployable
org/10.2514/6.2004-1662. membrane antenna structures. Chinese J. Mech. Eng. 30 (6), 1447–1459.
Cai, J., Ren, Z., Ding, Y., Deng, X., Xu, Y., Feng, J., 2017. Deployment simulation of Liyanage, P.M., Gangasudan, N., Mallikarachchi, H.M.Y.C., 2021. Modified spiral folding
foldable origami membrane structures. Aerosp. Sci. Technol. 67, 343–353. https:// pattern for deployable membranes. Aerosp. Sci. Technol. 117, 106926 https://doi.
doi.org/10.1016/j.ast.2017.04.002. org/10.1016/j.ast.2021.106926.
Cerda, E., Mahadevan, L., 2003. Geometry and Physics of Wrinkling. Phys. Rev. Lett. 90, Ma, L., Tang, Z., Bian, Z., Zhu, J., Wiercigroch, M., 2021. Analytical solution for circular
4. https://doi.org/10.1103/PhysRevLett.90.074302. inhomogeneous inclusion problems with non-uniform axisymmetric eigenstrain
Chandra, M., Kumar, S., Chattopadhyaya, S., Chatterjee, S., Kumar, P., 2021. A review on distribution. Int. J. Mech. Sci. 194, 106213 https://doi.org/10.1016/j.
developments of deployable membrane-based reflector antennas. Adv. Sp. Res. 68, ijmecsci.2020.106213.
3749–3764. https://doi.org/10.1016/j.asr.2021.06.051. MacNeal, R.H., Robbins, W., 1967. Tensile Properties of A Tape With A Transverse
Chung, J.Y., Nolte, A.J., Stafford, C.M., 2011. Surface wrinkling: a versatile platform for Crease.
measuring thin-film properties. Adv. Mater. 23, 349–368. https://doi.org/10.1002/ Martins, A.D., Silvestre, N., Bebiano, R., 2020. A new modal theory for wrinkling analysis
adma.201001759. of stretched membranes. Int. J. Mech. Sci. 175, 105519 https://doi.org/10.1016/j.
Deng, X., Xu, Y., Clarke, C., 2019. Wrinkling modelling of space membranes subject to ijmecsci.2020.105519.
solar radiation pressure. Compos. Part B Eng. 157, 266–275. https://doi.org/ Natori, M.C., Sakamoto, H., Katsumata, N., Yamakawa, H., Kishimoto, N., 2015.
10.1016/j.compositesb.2018.08.088. Conceptual model study using origami for membrane space structures - a perspective
Dharmadasa, B.Y., Jiménez, F.L., 2021. Modeling the effects of creases in an unfolding of origami-based engineering. Mech. Eng. Rev. 2, 14-00368-14–00368. https://doi.
membrane. AIAA Scitech 2021 Forum 1 PartF, 1–15. https://doi.org/10.2514/ org/10.1299/mer.14-00368.
6.2021-1258. Orszulik, R.R., Shan, J., 2012. Fuzzy logic active flatness control of a space membrane
Dharmadasa, B.Y., Mallikarachchi, H.M.Y.C., Lopez Jimenez, F., 2018. Characterizing structure. Acta Astronaut. 77, 68–76. https://doi.org/10.1016/j.
the Mechanics of Fold-lines in Thin Kapton Membranes. In: AIAA Spacecraft actaastro.2012.03.008.
Structures Conference. American Institute of Aeronautics and Astronautics Inc, Papa, A., Pellegrino, S., 2008. Systematically creased thin-film membrane structures.
Kissimmee, Florida. https://doi.org/10.2514/6.2018-0450. J. Spacecr. Rockets 45, 10–18. https://doi.org/10.2514/1.18285.
Dharmadasa, B.Y., McCallum, M.W., Mierunalan, S., Dassanayake, S.P., Pehrson, N.A., Banik, J.A., 2018. Folding approaches for tensioned precision planar shell
Mallikarachchi, C.H.M.Y., Jiḿ enez, F.L., 2020. Formation of plastic creases in thin structures. AIAA Spacecr. Struct. Conf. 2018 https://doi.org/10.2514/6.2018-1439.
polyimide film. J. Appl. Mech. Trans. ASME. https://doi.org/10.1115/1.4026559. Pellegrino, S., 2001. Deployable Structures in Engineering, in: Pellegrino, S. (Ed.),
Fenci, G.E., Currie, N.G.R., 2017. Deployable structures classification: a review. Int. J. Deployable Structures. Springer-Verlag Wein, pp. 1–35. https://doi.org/10.1007/
Sp. Struct. 32, 112–130. https://doi.org/10.1177/0266351117711290. 978-3-7091-2584-7.
Fleurent-Wilson, E., Pollock, T.E., Su, W.J., Warrier, D., Salehian, A., 2014. Wrinkle Piyumi, C., Mallikarachchi, C., 2020. Influence of Intersected Fold-Lines on Thin-Folded
localization in membrane structures patched with macro-fiber composite actuators: Membranes. MERCon 2020 - 6th Int. Multidiscip. Moratuwa Eng. Res. Conf. Proc.
inflatable space antenna applications. J. Intell. Mater. Syst. Struct. 25, 1978–2009. 30–35. https://doi.org/10.1109/MERCon50084.2020.9185241.
https://doi.org/10.1177/1045389X13512908. Roddeman, D.G., Drukker, J., Oomens, C.W.J., Janssen, J.D., 1987. The wrinkling of thin
Fu, B., Sperber, E., Eke, F., 2016. Solar sail technology—A state of the art review. Prog. membranes: part I — Theory. J. Appl. Mech. doi 10 (1115/1), 3173133.
Aerosp. Sci. 86, 1–19. https://doi.org/10.1016/j.paerosci.2016.07.001. Santiago-Prowald, J., Baier, H., 2013. Advances in deployable structures and surfaces for
Hossain, N.M.A., Woo, K., Jenkins, C.H., 2007. Dynamic Response of Systematically large apertures in space. CEAS Sp. J. 5, 89–115. https://doi.org/10.1007/s12567-
Creased Membranes. In: 48th AIAA/ASME/ASCE/AHS/ASC Structures, Structural 013-0048-3.
Dynamics, and Materials Conference. American Institute of Aeronautics and Satou, Y., Furuya, H., 2018. Effects of Elasto-Plastic Behavior of Crease on Deployed
Astronautics, Honolulu, Hawaii. Shape of Space Membrane 1–10. https://doi.org/10.2514/6.2018-0449.
Huang, Q., Hu, H., Yu, K., Potier-Ferry, M., Belouettar, S., Damil, N., 2015. Macroscopic Satou, Y., Furuya, H., 2011. Mechanical properties of Z-fold membrane under elasto-
simulation of membrane wrinkling for various loading cases. Int. J. Solids Struct. 64, plastic deformation. J. Sp. Eng. 4, 14–26. https://doi.org/10.1299/spacee.4.14.
246–258. https://doi.org/10.1016/j.ijsolstr.2015.04.003. Secheli, G., Viquerat, A., Aglietti, G.S., 2017. A model of packaging folds in thin metal-
Jarasjarungkiat, A., Wüchner, R., Bletzinger, K.U., 2008. A wrinkling model based on polymer laminates. J. Appl. Mech. Trans. ASME 84, 1–11. https://doi.org/10.1115/
material modification for isotropic and orthotropic membranes. Comput. Methods 1.4037503.
Appl. Mech. Eng. 197, 773–788. https://doi.org/10.1016/j.cma.2007.09.005. Seffen, K.A., You, Z., Pellegrino, S., 2000. Folding and deployment of curved tape
Jenkins, C.H., Kalanovic, V.D., 2000. Issues in control of space membrane/inflatable springs. Int. J. Mech. Sci. 42, 2055–2073. https://doi.org/10.1016/S0020-7403(99)
structures. IEEE Aerosp. Conf. Proc. 7, 411–414. https://doi.org/10.1109/ 00056-9.
aero.2000.879307. Senda, K., Petrovic, M., Nakanishi, K., 2015. Localized wrinkle behavior near fixed
Jenkins, C.H.M., 2006. Membrane Wrinkling, in: Christopher H. M. Jenkins, Frank K. Lu boundaries in flat and cylindrical membranes. J. Spacecr. Rockets 52, 1074–1090.
(Eds.), Recent Advances in Gossamer Spacecraft. American Institute of Aeronautics https://doi.org/10.2514/1.A33213.
and Astronautics, pp. 109–163. Shen, Y., Zhang, J., Guo, H., Liu, R., Kou, Z., 2021. Deployment simulation of membrane
Jules, T., Lechenault, F., Adda-Bedia, M., 2019. Local mechanical description of an structures based on elastic-plastic behavior parameterization of the crease. J. Mech.
elastic fold. Soft Matter 15, 1619–1626. https://doi.org/10.1039/c8sm01791c. Sci. Technol. 35, 2083–2095. https://doi.org/10.1007/s12206-021-0424-4.

16
P.K. Kamaliya et al. International Journal of Solids and Structures 254–255 (2022) 111902

Shinde, S.D., Kumar, S., Upadhyay, S.H., 2021. Investigation on material combination Wang, X., Ma, J., Law, S., Yang, Q., 2016. Numerical analysis of wrinkle-influencing
technique to enhance the anti-wrinkle and anti-vibration characteristics of the factors of thin membranes. Int. J. Solids Struct. 97_98, 458–474. https://doi.org/
planar membrane reflector. Proc. Inst. Mech. Eng. Part C J. Mech. Eng. Sci. 235 (21), 10.1016/j.ijsolstr.2016.07.004.
5675–5683. Wang, X., Law, S.-S., Yang, Q., Yang, N., 2018. Numerical study on the dynamic
Sipos, A.A., Fehér, E., 2016. Disappearance of stretch-induced wrinkles of thin sheets: a properties of wrinkled membranes. Int. J. Solids Struct. 143, 125–143.
study of orthotropic films. Int. J. Solids Struct. 97_98, 275–283. https://doi.org/ Wang, C.G., Mao, L.N., Du, X.W., He, X.D., 2010. Influence parameter analysis and
10.1016/j.ijsolstr.2016.07.021. wrinkling control of space membrane structures. Mech. Adv. Mater. Struct. 17,
Thiria, B., Adda-Bedia, M., 2011. Relaxation mechanisms in the unfolding of thin sheets. 49–59. https://doi.org/10.1080/15376490903082862.
Phys. Rev. Lett. 107, 1–4. https://doi.org/10.1103/PhysRevLett.107.025506. Wang, X., Yin, L., Yang, Q., 2019. Numerical analysis of the wrinkling behavior of thin
Torisaka, A., Ogawa, K., Miura, S., Parque, V., Miyashita, T., Yamakawa, H., 2019. Study membranes. Arch. Appl. Mech. 89, 2361–2380. https://doi.org/10.1007/s00419-
on in-plane and out-of-plane deformation considering elastic plasticity of membrane, 019-01583-4.
in: AIAA SciTech Forum. American Institute of Aeronautics and Astronautics, Inc., Wong, Y.W., Pellegrino, S., 2006. Mechanics of materials and structures PART II:
San Diego, California, pp. 1–23. https://doi.org/10.2514/6.2019-2255. analytical models. J. Mech. Mater. Struct. 1, 61–93.
Tsuda, Y., Mori, O., Funase, R., Sawada, H., Yamamoto, T., Saiki, T., Endo, T., Woo, K.-S., 2010. Thickness effect on wrinkle-crease interaction for thin membrane.
Yonekura, K., Hoshino, H., Kawaguchi, J., 2013. Achievement of IKAROS-Japanese J. Korean Soc. Aeronaut. Sp. Sci. 38 (5), 421–426.
deep space solar sail demonstration mission. Acta Astronaut. 82, 183–188. https:// Woo, K., Jenkins, C.H., 2012. Analysis of crease-wrinkle interaction for thin sheets.
doi.org/10.1016/j.actaastro.2012.03.032. J. Mech. Sci. Technol. 26, 905–916. https://doi.org/10.1007/s12206-011-1247-5.
Urbinati, L., 2020. Inflatable structures for space applications. Politecnico di Torino. Woo, K., Jenkins, C.H., 2013. Effect of crease orientation on wrinkle-crease interaction
Walker, S.J.I., Aglietti, G.S., 2007. A study of tape spring fold curvature for space for thin membranes. J. Spacecr. Rockets 50, 1024–1034. https://doi.org/10.2514/1.
deployable structures. Proc. Inst. Mech. Eng. Part G J. Aerosp. Eng. 221, 313–325. A32183.
https://doi.org/10.1243/09544100JAERO209. Xia, Z.M., Wang, C.G., Tan, H.F., 2018. Elastoplastic folding behavior of membrane
Wang, C.G., Du, X.W., Tan, H.F., He, X.D., 2009. A new computational method for ribbon based on plane strain beam theory. Int. J. Solids Struct. 143, 167–174.
wrinkling analysis of gossamer space structures. Int. J. Solids Struct. 46, 1516–1526. Zak, M., 1982. Statics of wrinkling films. J. Elast. 12 (1), 51–63.
https://doi.org/10.1016/j.ijsolstr.2008.11.018.

17