Professional Documents
Culture Documents
Stiffener Ship
Stiffener Ship
Stiffener is an effective way to reduce wafer warpage after carrier removal and
assembly-induced package warpage.
From: Modeling, Analysis, Design, and Tests for Electronics Packaging beyond
Moore, 2020
Related terms:
Stiffeners
The minimum required stiffener size is specified by the section modulus of the
stiffener as a function of stiffener spacing, stiffener span, design pressure, and
allowable stress (Figure 8.9).
Figure 8.9. Stiffener.
(8.12)
(8.13)
A stiffener is supposed to carry lateral pressure, which acts on the plate attached to
the stiffener, with a loading breadth equal to the stiffener spacing. Therefore, the
distributed load on the stiffener, q (in N/mm), can be calculated from the equation
(8.14)
By inserting Eqns (8.13) and (8.14) into Eqn (8.12), the following is obtained.
(8.15)
The classification rules contain this kind of equation for the design of beams under
lateral pressure.
6.5.3 Case 3: welding with side heating (140 °C) and tack welds
The stiffener welded with tack welds and side heating shows bowing and angular
distortion in the stiffener and bending of the web plate. The amount of bowing
distortion is at a maximum (5.2 ) at the stiffener midspan. As a result of heating, the
web plate tends to expand. However, the tack welding does not allow this expansion.
As a result, the web plate has bending distortion as well. Figure 6.11 shows the web
plate bent out of plane near one end of the stiffener. The web plate is bent on the end
of stiffener which was welded last. Table 6.3 lists the amount of bowing distortion
in this case.
6.11. Bending of the stiffener web plate in case 3 after welding with differential
heating (140 °C) of tack-welded specimen.
General Design
Dennis R. Moss, Michael Basic, in Pressure Vessel Design Manual (Fourth Edition),
2013
Case 1: Flat bar stiffener, flange of a tee stiffener, or outstanding leg of an angle
stiffener;
Case 2: Web of tee stiffener or leg of angle stiffener attached to the shell;
General
Stiffeners are employed to resist lateral loading of the plate and are usually made
from the rolled shapes integrally welded to the plate. Such stiffeners are distinct
from the other stiffeners used to prevent plate buckling. Typically, stiffeners run
continuously through the supporting frames. Otherwise they are referred to as
“intercostal” and require special connections at each end at each frame.
A summary of stiffener bending is given in fig. 7.95, the upper part showing context
in the stiffened panel, between frames and uniform loading. The lower part of the
figure shows the typical bending patterns and bending stresses in the stiffener/plate
combination.
Figure 7.95. Summary of stiffener bending
Figure 7.96 shows some of the typical stiffeners sections used. The US Unequal
angles sections are the prevalently available rolled shape in the US, but do not offer
a sufficiently wide selection, particularly in the larger shapes. The T-section, while a
particularly efficient shape material wise, is problematic and expensive in relation
to connections and frame penetrations and prevalent only in naval construction or
where special builders tooling is used. International hull construction uses both
the profile section (Europe primarily) and the JIS ship angle. Both are manufactured
primarily for ship construction, particularly efficient, and have distinct advantages,
especially large size. The latter are distinctive in that the tall leg is thin, with the short
leg thick, making an excellent flange.
Figure 7.96. Alternative stiffener sections
Tables 7.12, 7.13 and 7.14 respectively give dimensions and section properties of the
US unequal angles, profiles, and JIS ship angles. The latter two include combined
stiffener/plate properties as attached to a 20 in. wide, 1/2 in. plate. Figure 7.97
illustrates the notational conventions for the US unequal angle. The others are
similar.
Table 7.12. US Standard Unequal Leg Angles – with 1/2 in. plate at 20 in. width
attached
Table 7.13. Profile sections (HP)–with 1/2 in. plate at 20 in. width attached
Table 7.14. JIS (Japanese Industrial Standard) Ship Angles–(leg and flange unequal)
Figure 7.97. Notation for stiffener properties (US Unequal Leg Angles)
Before proceeding, a few comments on shear-lag and the effective breadth of the
plating should be made. As a flange attached to the stiffener, the bending stress in
the plate is not laterally uniform (maximum at the stiffener) and can be considered
as not fully effective. Figure 7.98 illustrates bending stress distribution in the plate,
a schematic of an equivalent uniformly stressed plate (effective breadth, 2be), and
provides notational reference.
Figure 7.98. Effective breadth of plating with stiffeners
Figure 7.99 is a plot of the effectiveness of the plate ( = be/b) as a function of the
ratio of the effective span (cL) to the breadth of the panel (b = s/2). Pending a more
complete discussion of shear-lag, c = 0.577 for mid-span locations and c = 0.423 at
the stiffener ends. What is important is that in most instances l > 4 s and most of
the plate is effective.
Figure 7.99. Effectiveness ratio, , of attached plating for stiffeners (Mid-span and
Ends)
Finally, it should be noted that, even if the effective breadth is small, the neutral
axis of the combined section is very close to the plate. This makes section modulus
at the flange relatively insensitive to the accuracy of the effective plate area. The
upper part of fig. 7.100 shows the neutral axis as a function of effective breadth for a
typical case. The lower plot shows the resulting section modulus, SMf as a function
of effectiveness, from 0.4 to 1.0. SMf is no less than 80% of its maximum value.
The point is that it is of relatively little consequence for stiffener design that shear
lag is ignored and the properties of the combined section are calculated on the basis
of the full width of the plating, s. For marginal cases, this should be re-examined,
however.
Figure 7.100. Stiffener neutral axes and section modulus variation with effective
breadth of plate
For design, the classification rule-based stiffener sizing criteria are given as formulae
for the section modulus of the stiffener and plate combination, SMf without consid-
eration of shear lag effects. The formulae are essentially based upon beam flexure
as outlined above. The dnv formulation uses a specific allowable stress and the ABS
formulation does not and is empirical. These are for minimum scantlings only. The
shell plating stiffeners tend to be larger in areas of high global stress both to increase
section (in lieu of thicker plating) and to improve stiffener buckling resistance.
ABS Criteria
As noted with plating, the ABS rules give two general cases. For the most general
case (Tanks), stiffener scantlings are given in Paragraph 9.5 (Stiffeners and Beams).
For the special cases governed by watertight integrity, in Paragraph 7.5 (Stiffeners
and Beams), the same expression is given, but the coefficient values are different.
The required, minimum section modulus is given as follows (metric, US units):
(7.63)
where [metric (US)] f= 7.8 (0.0041), s= stiffener spacing (m, ft), l= effective stiffener
span (m, ft), h = the specified design head (discussed earlier) taken at stiffener mid
span (m, ft). The constant c is specifically discussed below. The factor Q is material
constant adapted from the Ship Rules and is not given in the MODU Rules. It is also
discussed below.
The section modulus, SM, taken to the stiffener flange, is computed on the basis of
an integral section, taking both the stiffener and the full width (s) of the attached
plating. However, the rules note that, for short panels (l/s < 4), a lesser effective
breadth of the plate, se should be determined according to shear lag theory (at
l/s<4, se/s< 0.80). Because the stiffener-plate combination typically produces a quite
unbalanced section, accurate effective plate area has minimal effect on the section
modulus. Typically the stiffener flange is 3-times or more distant from the neutral
axis as is the plate, making bending stress in the plate of little consequence. Even
at l/s = 3, the effective breadth is still 0.65 and using the full plate width would still
be of little serious consequence. In any case, the combination of global in-plane
stresses from that of local bending is still required to be checked when the former
is high. Often, in such cases, there is sufficient global compression that buckling
will govern. When the buckling controls, the plate effectiveness is also reduced.
The factor c represents stiffener end fixity and the stress pattern of the specific
application. Although the Ship Rules have a number of specific cases represented,
the MODU Rules give prismarily two sets of factors:
•Tank boundaries and shell (9.5) : c = 1.00, both ends fixed or continuousc
= 0.90, terminated with clip or bracket
In the most general situation, shell plating, and tank boundaries (3-2-2/9.5), stif-
feners run continuously through the frames and are considered to function as a
series of fixed-fixed beams (c = 1.0). However, where this pattern ends (at a deck,
flat or bulkhead), the stiffener is terminated by a bracket or clip and is considered to
have less bending moment (c = 0.9). In the case of internal, watertight subdivision
(3-2-2/7.5), the factors reflect that strength being a damage issue, i.e. internal
flooding.
Further to the above, non-watertight flats and decks are addressed in Subsection 3,
Paragraph 7.3 (Upper Structure), where c = 0.6, clear of tanks, and c = 1.0 in way of
tanks. The latter is simply a reiteration of tank boundary requirements. Otherwise,
c = 0.6 is taken in association with a non-tank loading, h specific to the flat or deck.
The rules also give a number of comments regarding the determination of effec-
tive span, l. Unless effective brackets or clips are used, the effective span is the
centre-to-centre distance between supports. However, the effective span may be
reduced according to the effectiveness of the brackets meeting criteria specified in
the Rules. With brackets meeting requirements of table 3/5.2 (ABS MODU Rules,
2001), the effective span may be reduced by 75% of the bracket length.
Of the various specific application in the Ship Rules that uniquely give a value for c,
the one for bottom plating of a longitudinally stiffened ship gives c = 1.30. Where
otherwise 1.0, this factor empirically recognises that bottom stiffeners are part of
the overall hull section modulus and will sustain considerable global axial stress
over and above that of local bending. A non-documented, but very useful practice,
for preliminary design for offshore structures is to use a value of c from 1 to 1.3
for shell stiffeners of primary hull elements where they are known to have a high
global strength role. This is the case for parts of pontoons and columns. These have
a global loading pattern quite similar to that found in ship hulls: hydrostatic load
of stiffeners, participating in longitudinal strength. Using a higher value of c will
anticipate the added stiffener section area needed to sustain global hull bending
stress and buckling and improve the prospects that sizes chosen in preliminary
design will indeed be adequate.
The Ship Rules employ a multiplier “Q” to reflect the benefits of high tensile steels.
This can be used for offshore structures as well but has not been widely incorporated
in to the MODU Rules. Taken as unity for mild steel, various values are specified for
high tensile steels. Taking HT36 grade steels (Fy = 51 ksi), for example, Q = 0.72.
High tensile steels are prevalently used for the more highly loaded portions of the
hull.
As a rule of thumb, for hull plating, where mild steel thickness would exceed 5/8 in.,
high tensile steel should be used. If buckling controls, the threshold might better
be 3/4 in. In any case, for thickness in excess of 3/4 in., high tensile steels should
be used. The minor additional cost of the higher grade steel is more than offset
by reduced welding of thinner sections, not to mention larger, fewer modules for
a given lift limit. Conversely, high tensile steels should generally not be used for
thickness less than 1/2 in. as much as buckling will control and little benefit of the
higher yield strength will be achieved.
With regard to the ABS Rules and considerations of corrosion protection, Paragraph
3-2-2/1.4, Scantlings and Corrosion Control indicates that rule based scantlings
include a corrosion allowance in as much as it addresses reduction of scantlings.
It indicates that scantlings may be reduced should suitable corrosion protection be
provided. Conversely, this paragraph also indicates, that, should corrosion protec-
tion not be provided, stress-based calculation should consider the net section with a
suitable reduction. This would apply only to the stiffener's role in global strength.
dnv Criteria
As per the plating thickness criteria, the dnv stiffener design criteria is given in the
dnv MODU Rules Part 3, Chapter 1, Subsection 6, Plating and Stiffening. This criteria
is much more specific that the one given by the ABS. In Paragraphs B 200 (Stiffeners),
in the dnv notation, the following section modulus rule is given:
(7.64)
Like the plate thickness rule given previously, the dnv stiffener rule is essentially
the same form as that given by ABS, again with some differences. The terms, l and
s have the same meaning, if not units, and p is the pressure in kN/m2. The m in
the denominator is simply the denominator of the applicable beam equation. dnv
provides table B2 in Section 7, Girders and Girder Systems which is simply a complete
set of bending coefficients, m taken in the form: M = hsl2/m. For the fixed-fixed
case, m = 12 for the beam end and 24 for mid span, the same as used above under
the discussion of Loading and Stress. Again, as was done for plates, p is the allowable
stress. ks = 1 unless both ends of the stiffeners are pinned end boundary conditions
(in which case it is 0.9).
Simplifying, as was done with plating, the following can be written for the stiffener
ends:
(7.65)
(7.66)
This is essentially the same form as ABS, except that the ABS coefficient is 0.0041.
While the dnv coefficient is 20% less, it does not have the corrosion allowance that
is implicit in the ABS rule.
What is more notably different in the dnv Rules is the use of allowable stresses in
the formulation. For stiffeners, these are as follows (N/mm2):
Loading Condition
Gravity Only Combined Loading
Non-Primary Structure: 145 f1 190 f1
Primary Structure: 145 f1 – af 190 f1 – ( af + ac)
The “non-primary structure” category actually expresses local loading criteria alone,
without global stress adjustments. f1 is the material factor, which, for mild steel, is
1. As with allowable stress for plating, the global stress is deducted from the local
allowable for combined stress criteria. The definition of these deductions is as
follows:
To make the comparison more direct, the above is converted to ksi for mild steel (Fy
= 34 ksi) as follows:
Loading Condition
Gravity Only Combined Loading
Non-Primary Structure: 21.0 27.6
Primary Structure: 21.0 – af 27.6 – ( af + ac)
The mild steel equivalent values, for preliminary design, global stress are ae = 7.2
ksi and ae = 8.2 ksi.
One might say, rather than equivalent, the dnv stiffener rule is consistent with the
ABS rule. Once the corrosion addition is factored in, it is found to be reasonably
close. While the dnv rules are much more detailed with the allowable stress formu-
lation, the same is accomplished in the ABS rules with the general requirement for
combined stresses not exceeding the allowable stress. In the end they are the same
except for the corrosion allowance.
Another important factor in the stiffener design appears in the dnv Classification
Notes 31.4 (Column Stabilised Units), Paragraph 6.4.2: Columns; Shell Plating and Stif-
feners. Therein is a discussion of curved plating that recognises that, a longitudinally
stiffened shell, with sufficient curvature, will share the stiffener load in the form
of circumferential stress in the shell. There is given a curvature reduction factor as
follows (dnv notation; consistent units):
(7.67)
where r = Radius of shell mid surface, I = Inertia of the stiffener and plate combina-
tion, L = Distance between frames (same as l), t = Shell thickness (same as tp), and s
= Stiffener Spacing. The quantity c is a factor reflecting the boundary conditions of
the stiffener as a beam. For fixed ends, c = 384; for pinned ends c = 384/5.
For large column diameters (> 60 ft), this factor does not provide much benefit
(<5% stiffener load born by the shell). However, for smaller columns, reduction of
the stiffener load by recognition of the shells participation is worth considering.
For columns and pontoons with radiused corners, along with plating reductions
previously discussed, one stiffener can be eliminated at each corner.
Shear Strength
While the shear strength of rolled shape stiffeners are rarely a problem, and are
not explicitly addressed in the rules, there can be difficulty at their connections,
particularly to frames. This is particularly the case where the stiffener are highly
loaded and high strength steels are used for plating and stiffeners and the frames are
of mild steel. Normally there is a double-fillet weld connection between the stiffener
and the frame web. This may at first seem to be a welding problem, but is actually a
problem in the frame web size in that the depth of the stiffener times the thickness of
the frame web cannot provide the shear strength required. This is a stiffener problem
in that it may be the basis choosing deeper or ever over size stiffeners. Thicker frame
webs are not always an attractive solution and over size stiffeners can benefit global
strength. Also, there are standard details to address this problem (clips and collars),
but these are costly.
82. Stiffener of I section girder, design variants, reduction factors, , partially after
Neumann3 and Stüssi and Dubas.10
Cutting out the inner corners, which is frequently performed (triangle or quarter
circle cutouts, Fig.83), comes from the era of rolled sectional girders made of
unkilled steel, in which the segregation zones in the transition area from web to
flange had to remain uninfluenced by welding because of their tendency to crack.
Later the inner corners were cut out on similar web- to-flange welded girders
made of segregation-free structural steel to avoid defects caused by bad fitting
and welding as well as unfavourable residual stresses at the inner corner with the
weld crossing. The relatively high notch stresses of the transverse weld on the side
of the web subjected to transverse force bending of the girder are also avoided.
Therefore the fatigue strength of girders with stiffeners with cut out inner corners
is substantially improved (in contradiction to the fatigue test results in Ref.77).
Transverse corrugations in the web (e) (in Fig.82) instead of transverse ribs also
perform relatively well. A short longitudinal weld (c, d) instead of a transverse weld
can also produce an advantage with regard to fatigue strength. The T section, which
is resistant to buckling, is preferred as a stiffener in the presence of large transverse
forces.
The reduction factors given in Fig.82 are also applicable to the transverse wall of box
girders.
Crane runway girders have not only to be transversely stiffened, the upper chord
must also be designed to be torsionally rigid (f ). Otherwise the transverse forces on
the girder and the transverse bending moments in the flanges which occur with the
high number of cycles of the individual crane wheels, cause early crack initiation.
and
The following conclusions were drawn from the finite element results for the global
structure, Fig.190 and 191, which are the basis of the subsequent notch stress
analysis. The changes in the direction of principal stresses in the tensile flange
caused by the stiffener are negligibly small. The stiffener presses into the flange
from above, with pure bending load to a lesser degree and transverse bending load
to a geater degree. Thereby the structural stresses on the inside of the tensile flange
are decreased and those on the outside increased. This seems to be advantageous
because of the superimposed notch effect on the inside whereas there is no notch
effect on the outside. The result for the tension plate with unloaded stiffener there-
fore is on the safe side. With transverse bending, there is an uneven distribution of
the tensile stresses in the flange over the flange width, increased stresses at the web,
decreased stresses at the flange edge. Cutting out the inner corner of the stiffener is
favourable in the case of transverse bending because the superimposed notch effect
is avoided in the area of the cutout where the structural stresses are increased.
190. Finite element mesh and deformation of stiffened girder subjected to transverse
bending.
191. Structural stress distribution in area of weld joint between transverse stiffener
and tensile flange; subsequent notch stress analysis at points marked by triangles.
The results of the subsequent notch stress analysis, Fig.192, were converted via
equation (67) and (13) into fatigue strength reduction factors, P, for pulsating
tensile load taking the maximum tensile notch stresses from the weld toe (for t)
and the weld root (for r) into account, Table 21. The reduction factor, f, designates
the structural stress increase on the outside of the flange without any notch. All the
reduction factors refer to the maximum nominal bending stress of the girder, i.e.
the fatigue strength of the girder without stiffener.
192. Boundary stresses for contour model of stiffener-to-flange joint with fillet
weld for I girder subjected to transverse bending, tensile forces in flange and
compressive forces in stiffener, stiffener with cutout, notch position near cutout; K
factors indexed: t, toe; r, root; f, flange; u, upper; l, lower; m, median.
Reduction factors
Component Stiffening Type of Loading f 1 r
weld
I girder None None Pure bend- 1.02 – –
ing
I girder Transverse Fillet weld Pure bend- 0.95 0.55 0.64
stiffener ing
without
cutout
I girder Transverse Fillet weld Transverse 0.74 0.71 0.57
stiffener bending
without
cutout
I girder Transverse Fillet weld Transverse 0.73 1.08 0.75
stiffener bending
with cutout
I girder Transverse Double bev- Transverse 0.68 0.96 1.21
stiffener el butt weld bending
with cutout
Tension plate Transverse Fillet weld Tension 1.00 0.51 0.71
stiffener
Tension plate Transverse Double bev- Tension 1.00 0.52 0.78
stiffener el butt weld
The following general conclusions can be drawn for design and dimensioning from
the lowest (tensile) reduction factor in each case of Table 21. The transverse stiffener
is associated with a reduction = 0.55-0.73 for pulsating load (compared with =
0.25-0.75 for alternating load248). The high values can only be put into effect in the
case of pure bending if measures are taken to improve the weld toe e.g. by grinding
or dressing (in this case even > 0.73 seems possible for pure bending) or if the
transverse forces are relatively high (with crack initiation on the outside of the flange).
Deep penetration welding or double bevel butt welds are recommended to prevent
crack initiation at the weld root in the case of transverse bending. Cutting out the
inner stiffener comers improves the notch conditions considerably in the case of
high transverse forces. Cutting out is also recommended because of the excessive
gap and the limited accessibility at the inner corner. The tension plate with transverse
stiffener can only be correlated to pure bending loading of the girder.
The above reduction factors correspond with comparable minimum values in the
German standard for crane design, DIN 15 018,152 see Ref.248. In the IIW design
recommendation,141 the range of stress Δ = 80 N/mm2 with N = 2 × 106 (number
of load cycles endured) is assigned to the transverse stiffener on girders as well as
on tensile plates; the failure probability is Pf = 0.023, the reduction due to residual
stresses has been taken into account. On the other hand, with the reduction factor
P = 0.55 obtained by the investigation described above, after conversion from
Pf = 0.10 to Pf = 0.023, and after adjustment for residual stresses, the range of stress
corresponds quite well to the IIW design recommendation, see Petershagen.229
The top shell consists of two layers of plain weave for the suspension box cover and
for the areas that are in direct contact with the monocoque (shaded in black). The
top is bonded to the top surface of the assembled pair of side shells.
The tooling approach can be done by either an OML mold with rubber mandrels
or using a hollow bladder as an IML approach. Both approaches are schematically
represented in Fig. 5.64.
Figure 5.64. Tooling approaches used for hat stringers (A) OML and (B) IML.
Rohacell HERO foam [47] is an example of a typical type of foam used in such
applications. Specific examples where these are used are external doors and hatch
openings. Depending on the foam type, they can swell in areas of high moisture so
its recommended to examine the application to ensure that no adverse impact to
moisture is possible.
One of the challenges in this case can be the compaction of the plies in areas where
the foam parts terminate, as we need to ensure that the layers are draped smoothly
in those areas. FiberSim can be used to determine the best lamination approach and
splicing including the design. An example of the results that FiberSim can provide
is shown in Fig. 5.66. This is the results from a flat panel with several hat shape
stiffeners. As shown, there is no large indication with the splices used that any
manufacturing issues might be encountered.
Figure 5.66. FiberSim results showing no distortion generated from the different
splice designs considered.