Composites: Part B 32 (2001) 153–164

www.elsevier.com/locate/compositesb

Design guidelines for ply drop-off in laminated composite structures

A. Mukherjee*, B. Varughese
Department of Civil Engineering, Indian Institute of Technology, Powai, Bombay 400 076, India
Received 16 February 2000; accepted 4 June 2000

Abstract
The present investigation aims at developing a few guidelines for the design of tapered laminated composites. The tapering in laminated
composites is introduced by terminating (dropping-off) plies at different locations. The main objective in designing a drop-off is to reduce
stress concentration. At present some thumb rules are used to design the drop-off. In this paper, guidelines have been developed by studying
the effect of important parameters that determine the strength of the laminate. The numerical study shows that some of the thumb rules used at
present are rather conservative and may be relaxed to an extent. 䉷 2001 Elsevier Science Ltd. All rights reserved.
Keywords: A. Laminates; B. Delamination; B. Stress concentrations

1. Introduction stresses at the drop-off may initiate failure long before the
ultimate load carrying capacity of the laminate is reached.
Conventional metallic materials for the design of struc- Hence, the potential benefits in dropping plies may be
tural components like wing, fin, etc. of an aircraft are inhib- compromised through a reduction of the strength of the
ited by their high densities resulting in a reduction of laminate. Hence, the dropping-off of plies has to be done
payload. The evolution of laminated composite materials in a manner that does not affect the strength of the laminates
has opened new vistas in the development of aircraft indus- to a great extent.
try. The laminated composites offer a weight reduction Ply drop-off in laminated composites has been identified
thereby increasing the payload. A unique feature of lami- as a stress riser from the very beginning [1,7]. Since the last
nated composites is layered construction. The structural decade several experimental and analytical studies have
components are made by laying plies over one another so been reported regarding various aspects of this problem. A
that the required thickness of the laminate is achieved. The detailed review is available in Ref. [5]. However, it is not
laminate made in this manner is flat since the thickness of feasible for a designer to carry out a detailed analysis of the
each ply is uniform throughout. In practice, however, taper- entire structure. Therefore, it is necessary to develop a few
ing of laminates is necessary in many structural compo- thumb rules that the designers can follow to avoid premature
nents. The tapering is introduced by terminating plies at failure. In the absence of a systematic study such rules have
different locations. This is known as ply drop-off. Fig. 1 remained conservative and they constrain the design. More-
presents the geometry of a ply drop-off. The laminate tapers over, they lead to higher consumption of material leading to
from a thick section to a thin section as a result of the ply higher cost and weight. In this paper, the influence of three
drop-off. Plies may be inserted, on the other hand, at access most important design parameters, viz. number of plies
holes, lightening holes and at joints or connections to dropped-off at a single station, distance between successive
strengthen them. In all these applications use of ply drop- drop-offs (stagger distance) and lay-up of a dropped subla-
off results in significant saving in material and therefore, it is minate, on the strength of the laminate has been studied. The
cost effective. However, ply drop-off causes a discontinuity paper also discusses the variation of major interlaminar
within the laminate and therefore, it introduces structural stresses at the drop locations of the laminates.
difficulties like stress concentration at the drop station. Dropping-off of plies introduces out-of-plane (interlami-
This leads to failure of the components through delamina- nar) stresses in addition to in-plane stresses. However, the
tion and/or failure of resin. The formation of interlaminar interlaminar stresses are predominant only in the vicinity of
the drop-off and at locations away from the drop-off they
* Corresponding author. Tel.: ⫹91-22-578-2545; fax: ⫹91-22-578-3480. become negligible. Hence, it is possible to neglect the inter-
E-mail address: [email protected] (A. Mukherjee). laminar stresses at locations away from the drop-off and
1359-8368/01/$ - see front matter 䉷 2001 Elsevier Science Ltd. All rights reserved.
PII: S1359-836 8(00)00038-X
154 A. Mukherjee, B. Varughese / Composites: Part B 32 (2001) 153–164

Fig. 1. A typical ply drop-off.

they can be included only in the vicinity of the drop-off. the normal stress (s 3) and shear stress (t 13), which are
Thus, it is possible to conduct a global analysis of the whole computed along the principal material directions (1,2,3).
structure and a detailed local analysis in the region of the The other interlaminar shear stress (t 23) is not considered
drop-off. Such a global–local approach results in consider- in the criterion. This is because the magnitude of shear stress
able saving in computational effort. In the present study the (t 23) at the drop-off was negligible when the dimension of
global analysis has been carried out by a special ply drop-off the laminate along y is several times higher than its dimen-
element. The ply drop-off element is capable of incorporat- sion along x. The interlaminar normal stress (s 3) and shear
ing closely spaced ply drop-offs within the element. Subse- stress (t 13) have been computed at different locations of a
quently, the detailed analysis (local analysis) has been section at the center of the dropped laminate.
carried out at the drop location. The above ply drop-off Since delamination is an out-of-plane phenomenon and is
analysis has already been validated for various ply drop-
off configurations. The details of the analysis and its valida-
tion are discussed in separate papers [3,6]. We shall use the
methodology in analyzing a number of tapered laminates.

2. Failure criterion for delamination

Failure in a composite laminate may occur in different

ways:

• fiber fracture;
• matrix cracking;
• delamination.

Out of the three modes delamination requires special treat-

ment. Delamination in laminated composites is an out-of-
plane failure mode caused primarily by interlaminar stres-
ses. The contribution of in-plane stresses to delamination is
insignificant. Therefore, failure criterion that is used for in-
plane failure analysis cannot be used for delamination
analysis. Hence, the delamination criterion is often devel-
oped from the three-dimensional form of failure criterion.
This is achieved by considering only the terms that corre-
spond to the out-of-plane stress components. The form of
failure criterion used here is developed by modifying the
three-dimensional Tsai–Wu failure criterion [4].
According to the Tsai–Wu strength criterion the failure
surface in stress-space is expressed as
Fi s i ⫹ Fij s i s j ˆ 1 …1†

where Fi and Fij …ij ˆ 1; 2; 3; …; 6† are strength coefficients

determined from lamina strengths. The Tsai–Wu failure
criterion has been modified in such a way that the resulting Fig. 2. Geometry and global mesh of the laminates. (a) nD ˆ 1 (not to
delamination criterion contains only the terms correspond- scale); (b) nD ˆ 2 (not to scale); (c) nD ˆ 3 (not to scale); (d) global
ing to the major interlaminar stresses. These stresses include mesh for all laminates.
 A. Mukherjee, B. Varughese / Composites: Part B 32 (2001) 153–164 155

independent of the in-plane stress state, the in-plane stresses dropped plies. In practice 0, 90 and ^45⬚ plies are used
may be uncoupled from the out-of-plane stresses. Further, if for laminate construction. The stiffness contribution of
strength is considered to be independent of the sign of the these plies are different along the laminate’s axis. When
shear stresses all components containing a linear shear stress these plies are dropped at different locations in a laminate,
term may be excluded. It may be noted that the interlaminar the rise of out-of-plane stresses at each drop-off may be
shear stress s 4 has been assumed negligible and therefore, different according to the stiffness of the dropped plies. It
the above criterion reduces to is necessary to reduce stress concentration at the drop-off to
prevent an early delamination failure. This is possible
F 3 s 3 ⫹ F33 s 32 ⫹ F55 s 52 ˆ 1: …2† through a careful selection of lay-up of dropped plies at
The stresses s 3 and s 5 in Eq. (2) are stresses computed in different drop locations. In order to establish design guide-
the principal material directions. The criterion in Eq. (2) can lines on this aspect a detailed investigation has been carried
be rewritten as out in the present study.
The study on the above parameters has been carried out
F 3 s 3 ⫹ F33 s 32 ⫹ F55 t213 ˆ 1: …3† through extensive numerical experimentation. Laminates of
The above delamination criterion is used in the present various geometrical and lay-up configurations have been
investigation. It consists of terms corresponding to the inter- used for this purpose. The global–local approach [3] has
laminar normal stress (s 3) and the interlaminar shear stress been used for analyzing the laminates. A thin long laminate
(t 13). The criterion in Eq. (3) can easily be incorporated into has been chosen (Fig. 3). The location of the ply drop-off
the analysis of tapered laminates. was kept away from both the load and constraint locations to
In order to predict the failure of resin pocket and resin avoid their interference. A global model of the entire
layers the maximum stress criterion has been used. Since structure (Fig. 3d) has been analyzed with an in-plane
resin has isotropic properties the principal normal stress is tensile loading at the tip and constrained displacement at
compared with the allowable resin strength to determine the the root. The laminate is assumed to be linearly elastic until
failure of resin: the point of delamination. The material used in these exam-
ples is T300/5208 Graphite–Epoxy unidirectional tape. The
s normal ⱖ s allowable : …4† triangular pocket generated due to drop-off is filled with
resin. The properties of the laminate and resin materials
The criterion in Eq. (4) has been computed for the elements
have been presented in Table 1. The region of the drop-off
in the resin regions. It is assumed that the resin elastic prop-
has been modeled in detail for failure analysis. The response
erties are linear until failure. It may be noted that sometimes
of the global model has been used as input for the local
the resin may fail much before the delamination in the
model.
laminate.

3. Parametric studies 3.1. Effect of dropping plies at a single station

The effect of three parameters on the strength of the To study this effect three symmetrically dropped lami-
laminates has been studied. The first one is to observe the nates with plies dropped at a single station have been
effect of dropping many plies off at a single station. When it considered. The geometry and lay-up of these laminates
is necessary to achieve the tapering within a small span of are shown in Fig. 2a–c. In all the laminates, the lay-up of
the structure it may be required to drop a large number of the core and covering laminates has been kept the same.
plies at a single station. The drop-off causes a large step at However, the number of plies dropped has been varied
that point. The step size is changed by varying the number from 1 to 3. The dropped sublaminate consists of only 0⬚
of dropped plies in the laminate. The effect of this aspect on plies. In general, the lay-ups of the laminates can be repre-
the strength is examined. sented as (45/⫺45/0/90/0nD/90/0/⫺45/45)s where nD is the
The formation of a large step can be avoided by distribut- number of plies dropped and s signifies symmetric laminate.
ing the drop-off over a number of stations. The distance As nD increases from 1 to 3 the thickness of the thick
between two successive drop-offs (stagger distance) is section (H) increases; however, the thickness of the thin
selected in such a way that the tapering is smooth. The section (h) is the same in all the laminates. The thickness
drop-offs cannot be too close as it may lead to a large stress (H) of each laminate at the thick end is indicated in Fig. 2a–
concentration. On the other hand, it may not be possible to c. The triangular wedge created by the tapering is filled with
keep a large stagger distance when the span of the structure resin. Resin layers are also assumed as shown in the Fig.
is restricted. Though there are thumb rules for providing 3a–c. The laminates are fixed at the thick end and free at the
stagger distance (typically 6 mm), they proved conservative thin end.
according to this study. In the present work the stagger The global–local analysis is then carried out with each
distance has been studied as the second parameter. laminate. A typical global finite element mesh used for the
The third parameter considered is the lay-up of the analysis is shown in Fig. 2c. The local finite element meshes
156 A. Mukherjee, B. Varughese / Composites: Part B 32 (2001) 153–164

Fig. 3. Local finite element mesh for laminate. (a) nD ˆ 1; (b) nD ˆ 2.

used for laminates with nD ˆ 1 and 2 are shown in Fig. 3a of shear stress at the thin end of the transition region is less
and b, respectively. than that at the drop-off. The shear stress value is positive at
From the local analysis the interlaminar normal stress the top of the dropped plies and it is negative in the bottom
(s z) and interlaminar shear stress (t xz) have been computed. interface. The interlaminar shear stress is developed because
In order to understand the effect of dropping many plies off of the relative deformation between the dropped plies and
at a station the interlaminar stresses along the laminate has the core and the covering sublaminates. The core and the
been presented here. The normalized stresses for laminates covering sublaminates tend to slide over the dropped
with nD ˆ 1 and 2 are shown in Fig. 4. Kemp and Johnson sublaminate as shown in Fig. 6.
[2] have studied the distribution of interlaminar normal In order to get a better understanding of the effect of
stress and shear stress for the laminate with nD ˆ 3. A dropping plies at a station on the magnitude and distribution
comparison of the results has been presented elsewhere of the normal and shear stresses the peak values and the rate
[3]. The correlation between the two results is very good. of decay of these stresses in the thick section of the lami-
The stresses peak at two points — the drop-off and the nates is presented in Table 2. The stresses correspond to a tip
thin end of the transition region. At the drop-off the normal load of 0.28 kN/mm. The rate of decay gives an idea of how
stress (s z) is compressive as shown by the negative sign. quickly the stress reduces as we go away from the drop-off.
The stress changes its sign and becomes tensile and peaks at The maximum normal stress and shear stress increase
the thin end of the transition. Fig. 5 presents the direction of as the number of plies dropped (nD) increases. The rate
normal stress at the key locations. As a tensile force is of decay is higher for the normal stress (s z) than that
applied at the thin end of the laminate, the tendency of the for shear stress (t xz). The concentration of normal stress
core and the covering sublaminates is to come closer and is high only near the drop-off. The shear stress,
results in compressive normal stress at the drop-off. While at however, reduces slowly. It is also observed that the
the tip, the tendency of the normal stress is to split the core
and covering sublaminates resulting in a tensile s z. If the Table 1
applied stress is compressive the stress pattern will be Material properties
tensile at the drop-off and compressive at the thin end.
It can be seen from Fig. 4a that the compressive s z is Material T300/5208 Resin
Graphite ^ Epoxy tape
maximum at the junction between the covering and dropped Elastic moduli (MPa) E1 ˆ 138 × 103 ; E2 ˆ 3.45 × 10 3
sublaminates. The absolute maximum normal stress is at the E3 ˆ 9:65 × 103
top of the dropped sublaminate. The interlaminar shear Poisson’s ratio n12 ˆ n13 ˆ 0:3; n23 ˆ 0:6 0.36
stress (t xz) shows peak values at the drop-off at the top Shear moduli (MPa) G12 ˆ G13 ˆ 5:516 × 103 ;
G23 ˆ 4:14 × 103
and bottom interfaces of dropped sublaminates. The value
 A. Mukherjee, B. Varughese / Composites: Part B 32 (2001) 153–164 157

Fig. 4. Stresses at the drop-off. (a) Normal stress (nD ˆ 1); (b) shear stress (nD ˆ 1); (c) normal stress (nD ˆ 2); (d) shear stress (nD ˆ 2).

concentration of interlaminar stresses near the drop-off Failure analysis has been conducted on the three
increases with the number of dropped plies at a laminates using the failure criteria discussed in Eqs. (3)
station. However, when the number of plies dropped and (4). The tip strain at delamination initiation has been
at a station reduces then the stresses spread over a obtained from the failure analysis. The location of failure is
larger area and that reduces the risk of a sudden failure also identified. The strain at delamination computed for the
at the drop-off. three laminates is shown in Fig. 7. The strain at failure is

Fig. 5. Formation of positive and negative normal stress. Fig. 6. Formation of positive and negative shear stress at the interfaces.
158 A. Mukherjee, B. Varughese / Composites: Part B 32 (2001) 153–164

Table 2
Effect of nD on the peak value and distribution of interlaminar stresses

nD Normal stress (s z) Shear stress (t xz)

⫺2
Peak value Rate of decay ( × 10 ) Peak value Rate of decay ( × 10 ⫺2)
( × 10 ⫺2) (N/mm 2) (N/mm 2/mm) ( × 10 ⫺2) (N/mm 2) (N/mm 2/mm)

1 ⫺1.22 3.00 ⫺1.23 1.05

2 ⫺1.78 3.50 ⫺1.66 1.16
3 ⫺2.19 5.40 ⫺1.97 1.47

highest for the laminate with one ply dropped and it symmetric about the x–y plane. In Fig. 8a the two stations
decreases as the number of plies dropped increases. The are situated at a distance ‘dd’ (stagger distance) apart. The
failure strain has dropped by 7.5% when two plies are effect of distance dd on the laminate’s strength is studied
dropped in place of one ply and it drops by 12.3% when here. For this purpose failure analysis is carried out on the
three plies are dropped instead. The strain at failure laminate for different values of dd. The value of dd is varied
decreases at a decreasing rate with an increase in the by gradually shifting the location of the second drop-off
number of plies dropped. Therefore, the number of plies towards the first drop off. The first drop station has been
dropped must be decided on the basis of the reduction kept stationary. When the two drop stations are away from
in strength that the designer can accept. For the least each other the zones of stress concentration do not interact.
reduction only one ply should be dropped at a time with Therefore, they behave as independent drop-offs without
a staggering. However, if higher reduction in strength can influencing the failure loads of one another. This distance
be allowed more plies can be dropped at a station. The forms the upper bound of the study. Any stagger distance
reduced drop-off strength should not be less than that of beyond this value does not influence failure load at all. As
the thin end. the stagger distance reduces the zones of stress concentra-
tion start interacting. Therefore, the failure load of the lami-
nate reduces gradually. In this study the effect of such an
3.2. Distance between successive drop-offs (stagger
interaction is examined until the lower bound of the
distance)
stagger distance is reached. The aspect ratio of the
A gradual drop-off by introducing many drop-off stations transition region is 1:3. Therefore, the lower bound of
is preferable to a single drop-off station. The selection of stagger distance is three times the thickness of drop-off
minimum distance between successive drop-offs (stagger
distance) is important to achieve the desired tapering. A
common thumb rule is to keep a staggered distance of
6 mm. In order to examine this rule a laminate having two
drop-off stations has been studied. The geometry of the
laminate model used is shown in Fig. 8a. The laminate is

Fig. 7. Strain at failure for various nD values. Fig. 8. Geometry of drop-offs. (a) Staggered drop-offs; (b) lower bound.
 A. Mukherjee, B. Varughese / Composites: Part B 32 (2001) 153–164 159

Table 3 3.3. Effect of stagger distance on interlaminar stresses

Lay-up sequence for various sets
In order to study the effect of stagger distance, the varia-
Set number Lay-up sequence …u1 =u2 =…u3 =u4 †D1 =…u5 =u6 †D2 =u7 =u8 †
tion of interlaminar normal stress and shear stress have been
1 (⫹45/⫺45/(0/90)D1/(90/0)D2/⫺45/⫹45)s studied at the first and second drop-offs of the laminate (Fig.
2 (⫹45/⫺45/(90/90)D1/(90/90)D2/⫺45/⫹45)s 10). The stresses have been computed just below the first
3 (⫹45/⫺45/(0/0)D1/(0/0)D2/⫺45/⫹45)s and the second drop-offs, i.e. at z ˆ 0:505 and 0.250 mm,
4 (⫹45/⫺45/(⫹45/⫺45)D1/(⫺45/⫹45)D2/⫺45/⫹45)s
5 (⫹45/⫺45/(90/90)D1/(0/0)D2/⫺45/⫹45)s
respectively. Fig. 10a–f shows the normal stress and shear
6 (⫹45/⫺45/(0/0)D1/(90/90)D2/⫺45/⫹45)s stress at the first drop-off for sets 1–6. The stresses for two
extreme values of staggered distances, i.e. 0.762 and
2.540 mm are plotted. All other values of staggered
distances lie between these two bounds. Stagger distances
(Fig. 8b). The staggered distance dd has been reduced beyond 2.540 mm can be considered independent of
from 3.048 mm (12 times the drop thickness) to 0.762 mm drop-offs.
(three times the drop thickness). Eight stagger distances The shear stress remains largely unaffected by the stagger
between these two bounds have also been examined. The distance. Towards the lower bound the shear stress is
lengths of thick and thin sections have been kept constant marginally higher. The normal stresses for dd ˆ 0:762 and
for all values of dd. 2.54 mm have little difference at the drop-off station …d ˆ
The triangular pockets created at each station by the 0†: As we proceed towards the tip of the second drop-off, the
covering plies and dropped plies have been filled with stresses, however, diverge. This is due to the interaction
resin. Two plies are dropped at each drop station of the between the two drop-offs for small dds. The largest stresses
laminate. The thickness of each ply is 0.127 mm. The thick- are obtained in case of set 3 that has the severest drop-off
ness of the symmetric half at the thick end is 1.01 mm and with two 0⬚ plies dropped at both stations. The normal stres-
that at the thin end is 0.508 mm. ses in case of set 5 that has two 0⬚ plies dropped at the
The main purpose of this study is to determine the mini- second station, are also of comparable magnitude.
mum stagger distance between two drop-offs so that a The divergence in stresses is most drastic in case of sets 5
smooth transfer can take place. The minimum stagger and 6, where 0/0 and 90/90 layers are dropped subsequently.
distance for various stacking sequences can be different. However, the largest stresses for the two dds were close to
Therefore, six combinations of stacking sequence have each other; indicating that the failure stresses are not very
been analyzed. The lay-up of laminate in different sets is sensitive to stagger distance. Therefore, there is a scope for
presented in Table 3. In each set the lay-ups of two dropped considering the smallest possible stagger distance as three
sublaminates have been varied. The orientations of dropped times the thickness of the dropped sublaminate. However,
plies at the first and second drop stations have been speci- this point must be confirmed by studying the second drop-
fied by subscripts D1 and D2, respectively. The lay-ups of off in the laminate. The normal and shear stress variation at
the covering and core sublaminates have not been changed. the second drop-off for all the six sets have been presented in
Fig. 8 may be referred along with Table 3 to understand the Fig. 11a–f. The stresses at the second drop-off are substan-
lay-up sequence in each set. tially higher than the first drop-off. The effect of second
A typical finite element mesh for local analysis is shown drop-off is more severe because the laminate is thinner at
in Fig. 9. The distance of the thick end from the root of the that location than at the first drop-off. Only in set 6, which
laminate is 4.572 mm and is constant for all values of dd. has the (0/0)D1/(90/90)D2 dropping sequence, the stresses at
The local mesh is fine near the two drop-offs and other the second drop-off are lower than that at the first drop-off.
locations the mesh is coarse. Both eight-noded quadrilateral This is due to lower stiffness of the 90/90 sublaminate.
and six-noded triangular isoparametric elements have been The stresses are very close to one another for the whole
used. The largest finite element model has 624 elements and range of dds. The divergence in s z at the thin end of the
1851 nodes. drop-off in case of dd ˆ 0:762 mm is absent here as

Fig. 9. Local modeling of the laminate with two drop-offs.

160 A. Mukherjee, B. Varughese / Composites: Part B 32 (2001) 153–164
 A. Mukherjee, B. Varughese / Composites: Part B 32 (2001) 153–164
Fig. 10. Stress variation at the first drop. (a) Set 1; (b) set 2; (c) set 3; (d) set 4; (e) set 5; (f) set 6.

Fig. 11. Stress variation at second drop-off. (a) Set 1; (b) set 2; (c) set 3; (d) set 4; (e) set 5; (f) set 6.

161
162 A. Mukherjee, B. Varughese / Composites: Part B 32 (2001) 153–164
Fig. 11. (continued)
 A. Mukherjee, B. Varughese / Composites: Part B 32 (2001) 153–164 163

Therefore, higher stagger distance is necessary when ^45

laminates are dropped. Higher interlaminar shear stresses in
case of ^45 laminates is responsible for the reduction of
failure load. Sets 1, 3, 5 and 6 are not sensitive to stagger
distance. Set 6 has a higher failure load than sets 1, 3 and 5.
The difference between set 6 and set 5 is that in case of
set 6 the stiffer plies (0) are dropped first. This observation
reinforces our guideline that the plies should be
dropped in decreasing order of their stiffnesses. Set 1 that
has a (0/90) combination at both drop-offs has not generated
failure load higher than that of set 5. Therefore, it can be
concluded that, combining plies of different orientations at
the drop-off may not be effective in increasing the laminate
strength.

4. Design guidelines

Fig. 12. Tip loads at the onset of delamination. We shall summarize the design observations we have
made in this paper:
there is no drop-off ahead of this one. That reinforces the • When in a laminate a large number of plies are to be
observation that stagger distance can be kept as low as three dropped, the drop-offs may be staggered with the number
times the sublaminate thickness.
of plies dropped at any location kept to a minimum. The
The normal and shear stresses reached maximum values
strength of the drop-off should not be below the thin end
in case of sets 1 and 3. In both these cases 0⬚ plies have been
strength. The strength of the drop-off reduces decreas-
dropped in the sublaminate. A 0/90 drop-off has marginally
ingly with the increase in number of plies dropped.
less stress than a 0/0 drop-off. This demonstrates that the
• The stagger distance can be kept as low as three times the
orientation of dropped plies is a very important factor. 0⬚
thickness of the drop-off. The stagger distance should be
plies lead to the largest stresses and ^45 and 90 plies have at least eight times the thickness of the drop-off when 45
considerably less severe effects. Therefore as far as possible plies are dropped.
the ^45 and 90 plies should be dropped. We shall investi-
• The plies should be dropped in decreasing order of their
gate the drop-off of 0⬚ plies now.
stiffness. The stiffest plies (0 plies) should be dropped at
We shall observe the stresses in sets 5 and 6 to devise a
the thick end and the softest plies (90 plies) should be
strategy for drop-off of 0⬚ plies. In sets 5 and 6 the same set
dropped at the thin end. This ensures a smooth transfer of
of plies, 0/0/90/90, has been dropped. In set 6 the 0⬚ plies are
load and reduces stress concentration.
dropped at the first drop-off and in set 5 they are dropped at
• The strength of the laminate does not improve if, to tailor
the second drop-off. The stresses in set 6 are much lower stiffness, plies of different orientations are dropped at the
than that in set 5. The relatively thicker laminate at the first same station. For example, a 0/90 drop-off and a 0/0
drop-off is responsible for a better performance of set 6.
drop-off have similar strength reduction. Therefore,
From this experiment we can conclude that, when we design
combination sublaminates (e.g. 0/90), if dropped should
a staggered drop-off the plies should be dropped in decreas-
be dropped at the thick end.
ing order of their stiffnesses, i.e. stiffest plies should be
dropped first and the softest plies last. We shall observe
the effect of these stresses on the failure load of the
laminate. Acknowledgements
The tip load intensity at the onset of failure for the lami-
nates in sets 1–6 for the whole range of stagger distances This work has been supported by ARDB, Ministry of
from 0.762 to 2.54 mm is plotted in Fig. 12. It can be seen Defence, Government of India, under the grant no. Aero/
that set 2 has the largest failure load. This set has only 90 RD-134/100/10/93-94/779. The authors gratefully acknowl-
plies dropped. These plies are softest and they have the least edge help by Mr V. Ravindra in preparing the manuscript.
effect on the failure loads. Set 4 is next in failure loads. This
set has ^45 plies dropped and they are next in the order of
References
softness. Therefore, the failure load is second to the 90 plies.
The ^45 layers also exhibit sensitivity to stagger distance. [1] Curry JM, Johnson ER, Starnes Jr. JH. Effect of dropped plies on the
The failure load reduces sharply as the stagger distance strength of graphite–epoxy laminates. AIAA J 1992;30:449–56.
reduces from eight times the dropped laminate thickness. [2] Kemp BL, Johnson ER. Response and failure analysis of a graphite–
164 A. Mukherjee, B. Varughese / Composites: Part B 32 (2001) 153–164

epoxy laminate containing terminating internal plies. In: Proceedings [5] Varughese B, Mukherjee A. A ply drop-off element for analysis of
of the AIAA/ASME/ASCE/AHS 32nd Structures, Structural Dynamics tapered layered composites. Compos Struct 1997;39:123–44.
and Materials Conference, Part 1, Orlando, Florida, AIAA Paper No. [6] Varughese B, Mukherjee A. Analysis of tapered laminated compo-
85-0608, 1985. p. 13–24. sites with non-symmetric lay-up. J Reinf Plast Compos 1997;16:
[3] Mukherjee A, Varughese B. Development of a specialised finite 631–60.
element for the analysis of composite structures with ply drop-off. [7] Wilkins DJ. A preliminary damage tolerance methodology for compo-
Compos Struct 1999;46:1–16. site structures. In: Noor AK, Shuart MJ, Starnes Jr. JH, Williams, JG,
[4] Tsai SW, Wu EM. A general theory of strength for anisotropic materi- editors. Failure analysis and mechanisms of failure of fibrous compo-
als. J Compos Mater 1971;5:58–89. site structures. NASA CP 2278, 1983. p. 67–93.