Ninth International Specialty Conference on Cold-Formed Steel Structures

St. Louis, Missouri, U.S.A., November S-9, 1988

THE STRUCTURAL DESIGN OF LIGHT GAUGE SILO HOPPERS
by J. Michael Rotterl
Synopsis
Elevated light gauge silos usually have a conical discharge hopper at the
bottom. Although this hopper often carries much of the total weight of the
stored solids within the silo, it can often be cold-formed from thin steel
sheet because the structural form is very efficient. However, guidance on
the design of light gauge hopper structures is rare. The term "light
gauge" is used here to describe the class of cold-formed silo structure
which is not restricted by a nominal minimum plate thickness requirement
(eg 1/4" or 6 mm).
This paper addresses several aspects of the design of light gauge hoppers.
Current proposals concerning hopper loads are discussed first, and
recommendations are made. Appropriate structural analysis is then
presented. The potential failure modes of the hopper are identified, and
corresponding strength checks described.
1. INTRODUCTION
Silos made from cold-formed steel sheets are widely used in agriculture
throughout the world. Failures in silos are common, and improvements in
silo technology are clearly desirable.
This paper is concerned with the design of light gauge metal silo hoppers,
and those aspects of the rings and column support conditions which are
intimately related to the hopper (Fig. 1). It relates only to silos of circular
planform.
Conical discharge hoppers are generally subjected to only symmetrical stored
solids loading. However, the pressures on hoppers are less well understood
than those on vertical silo walls; the structural action is a little more
complex, and little attention has been paid to hopper design in the past.
These factors underlie the present review paper.
Cold-formed steel silo hoppers are susceptible to more modes of failure than
the larger industrial silo hoppers, because they often have bolted joints of
limited strength. The design of these joints requires a more careful
assessment of hopper loading patterns, so a significant part of this paper is
devoted to the definition of hopper loads.
Farm silos often differ considerably from industrial silos, and much of the
available design advice is concerned either with industrial and mmmg
applications (Wozniak, 1979; Trahair et aI, 1983; Gaylord and Gaylord, 1984;
Rotter, 1985) or with the cylindrical walls of light gauge silos (Trahair et
aI, 1983; Abdel-Sayed et aI, 1985; Rotter, 1986b).
1 Senior Lecturer in Civil Engineering, University of Sydney, New South
Wales, Australia.
529
530
2. LOADS ON HOPPER WALLS
2.1 Introduction
The chief loading on conical discharge hoppers derives from symmetrically
placed stored solids. However, the assessment of these loads involves both
the cylinder and the hopper. A detailed argument concerning appropriate
pressures for light gauge hoppers is presented elsewhere (Rotter, 1988), and
only the recommendations are given here.
The most commonly used theories for pressures in hoppers are those of
Walker (1966), Walters (1973) and Jenike et al (1973). J'vIost codes of
practice (American Concrete Institute, 1977; DIN 1055, 1986; Gorenc et aI,
1986; BJ'vIHB, 1987) specify either a constant pressure within the hopper or a
linearly varying pressure. J'vIost include a local high "s,,,itch" pressure near
the hopper/cylinder junction (the transition) to account for flow conditions,
but the details of the pressure distribution in the body of the hopper are
often thought to be relatively unimportant. However, light-gauge bolted
steel hoppers require a more careful assessment of pressure distributions.
2.2 Defining the Total Load on the Hopper
In elevated silos, the hopper supports the majority of the total weight of
stored material. The total load on the hopper is defined by the hopper
volume and the mean vertical stress in the stored material at the transition
(hopper/cylinder junction) (Fig. 2a). The latter depends on the height of
stored material above this point, and the proportion which is supported by
friction on the cylindrical walls. Cylindrical wall pressures are therefore
needed to define the loading on the hopper. The pressures on the walls of
the cylinder p, wall frictional tractions v, and vertical stress in the stored
solid q (Fig. 3a) are most easily assessed using Janssen's equation (1895)
p = Po (1 - e-Y/Yo) ( 1 )
v = /.Lp
(2 )
(3 )
in which Po = yR/2/.L, % = yR/2/.Lk, Zo = R/2/.Lk, R is the silo radius, y is
the distance below the effective surface of the solid, y is the bulk solid
density, /.L is the wall friction coefficient, and k is the lateral pressure ratio
(ratio between horizontal and mean vertical stresses in the stored solid).
2.3 J'vIaximising the Loading due to Bulk Solids
J'vIost silos are used to store a range of materials, so that the properties
may vary significantly from time to time. Other changes may occur as the
silo becomes polished or roughened by stored solids. The silo should
therefore be designed for a variety of different values of Y, k and /.L in
Eqs 1-3.
All loads are maximised when the value of y is maximised. The largest
values of wall pressure occur when k is at its maximum value and /.L at its
minimum value. The maximum vertical loads on hoppers occur when both k
and /.L take their minimum values. The smallest possible value of k is given
by the simple Rankine pressure ratio
k =
- sin <I>
+ sin <I>
531
( 4 )
in which <I> is the effective angle of internal friction of the stored solid
(usually in the range 28-33
0
for grains). More realistic values of k can be
derived from the relation first advanced by Walker (1966) and since adopted
by many others
k =
1 + sin
2
<1> - 2/(sin
2
<1>-g2cos
2
<1
4g2 + cos
2
<1>
( 5 )
However, the values obtained from Eqs 4 and 5 differ only slightly unless
the wall is very rough. For steel silos, it therefore is arguable that Eq. 4
should be used to determine k for the cylinder when the total hopper load
is being determined. Corrugated silos provide the one possible exception to
this proposal (Rotter, 1988).
2.4 Initial Filling Loads on Hopper Walls
The simplest theory of hopper filling pressures is that of Walker (1966). It
assumes that the stored material in the hopper carries no shear stresses.
The maximum pressure occurs at the outlet (Fig. 4b). This pattern is often
the worst pressure distribution for welded hoppers without strong
transitional ring support (Rotter, 1986a) as it places the maximum load as
far from the support as possible. However, it may be unduly conservative.
Walker filling theory is used in some codes (Gorenc et aI, 1986; BMHB, 1987),
but not in others (DIN 1055, 1986; American Concrete Institute, 1977). It
should be noted that the frictional traction v has sometimes been omitted in
drafting code rules based on Walker theory (BMHB, 1987) leading to an
unsafe definition, since the hopper is deemed to carry less than the total
load on it.
Walker (1966) also presented a theory for discharge conditions. This theory
involves the general solution of the hopper equilibrium equation subject to
the condition (similar to the Janssen assumption) that the ratio of wall
pressure to mean vertical stress in the solid is invariant with depth (Figs
4b and 4c). Walker also proposed one method of determining this ratio.
Many writers have made modifications to this basic theory by determining
the ratio in different ways. McLean's (1985) modification of the Walker
theory is supported by some experimental evidence (Motzkus, 1974; Hofmeyr,
1986) and the finite element calculations of Ooi and Rotter (1987). The
general Walker theory leads to the pressure distribution (Figs 4b and 4c)
p = F q
(6 )
v g p ( 7 )
(8 )
in which p is the normal wall pressure, v is the frictional traction, q is the
vertical stress in the stored solid at height z above the apex, qt is the
value of q at the transition and H is the height of the hopper. Equilibrium
of the complete hopper requires that
532
n = 2(Fg + F 1 )
(9 )
McLean (1985) suggests that F = 1.0 leads to reasonable estimates of hopper
initial filling pressures, so Eq. 9 reduces to
n = 2g
( 10)
This theory is recommended here to define hopper filling pressures.
2.5 Mass Flow and Funnel Flow Pressures
The pattern of solids flow from a silo is known to affect both the pattern
and the magnitude of the pressures. Two simple forms of flow pattern have
been widely accepted (Jenike et aI, 1973) and are known as the mass flow
and funnel flow modes (Fig. 3b). The hopper pressures are normally
defined with one or other of these flow modes in mind.
Most published theories are concerned with the pressures during mass flow.
It is widely recognised that the pressures at the outlet decrease during
discharge, as only then can flow of the solids occur. A local high pressure
also develops at the transition (Fig. 4c), and most design guides recommend
that this "switch" pressure be considered. The magnitude of the transition
switch pressure only becomes really large when a very steep hopper is
used. Most cold-formed silos have quite shallow hoppers (-45
0
), so this
switch pressure is not a very significant item.
In cold-formed bolted hoppers, the critical point is usually a short distance
from the transition, and this distance is defined by the hopper structural
behaviour. In these circumstances, it is difficult to decide between the
different theories and codified rules, but the original Walker (1966)
discharge theory may provide reasonable values. It is certainly preferable
to uniform pressures when cold-formed hoppers with bolted joints are being
designed. It is given by Eqs 6-9 with
F =
1 + sin<\> cosE:
( 11)
1 - sin<\>
. -1 [1 -/ [ g2 ]]
Sl.n sin<\> l+g2
( 12 )
in which g is the hopper wall friction coefficient and <\> is the effective
angle of internal friction.
Funnel flow pressures are less well understood. Gaylord and Gaylord (1984)
and Rotter (1988) have presented strong criticisms of the commonly assumed
pressure distributions. The best current treatment is probably to assume
that all hoppers are subject to mass flow hopper pressures irrespective of
their flow patterns (Eqs 6-9, 11-12).
3. STRESSES IN HOPPER WALLS
3.1 Introduction
When the silo is continuously supported (not supported on columns), the
stresses developing in the greater part of a steel hopper can be determined
using the membrane theory of shells. Bending stresses in the hopper body
533
occur only at changes of plate thickness and are generally small. However
the stresses in the shell and ring at the transition are strongly affected by
the discontinuity there, are much larger, and require more computational
effort to establish precisely.
3.2 Hopper stresses under Symmetrical Filling, Discharge and Support
For symmetrical loading, the membrane theory of shells (Rotter, 1987b) leads
to two equilibrium equations in terms of the meridional stress resultant Nell
(force per unit circumference) and the circumferential stress resultant Ne
(force per unit width) for conical shells (Fig. 2b)
NCil = P z
( 13)
d
dz (z N
Cil
) = z (p + v)
(14 )
These may be solved for the general form of the hopper pressure
distribution (Eqs 6-9), to yield
( 15)
[
yH [z]2
3(n-l) II
1 [ yH ] [z]n+l]
+ (n+2) qt - n-l II
(16 )
It should be noted that, in general, the highest pressure on the hopper
does not coincide with the highest membrane stresses in the hopper wall
(Fig. 5).
The hopper is in biaxial tension. For welded hoppers, which can sustain
combined stresses in two directions, the meridional and circumferential
stress resultants should be combined to find the effective stress resultant
N
vm
using von Mises yield criterion
( 17)
Cold-formed hoppers usually have meddional bolted joints, so it is difficult
to exploit this additional strength. Moreover, the joints usually have lower
strengths than the sheets which are joined, so the circumferential stress
resultant (Eq. 15) must be compared with the joint strength. The
circumferential tension varies up the hopper, but to avoid changes of bolt
spacing, it is sometimes useful to work to the maximum value of
circumferential tension. If the silo has a significant cylindrical section, so
that
(18 )
the maximum occurs at the top of the hopper, where the circumferential
membrane stress resultant is
Ne = qt F H (19)
However, if the hopper occupies much of the silo, then Eq. 18 will not be
satisfied, and the maximum value of the circumferential force per unit length
534
Ne occurs within the hopper and can be determined from Eq. 15 with
z
II =
1
n-1
(20)
Finally, it is useful to note that the simpler assumption (American Concrete
Institute, 1977) of uniform pressure in the hopper with no friction on the
hopper wall always overestimates the maximum circumferential tension, unless
Eq. 18 is not satisfied.
The meridional membrane stress resultant often reaches its maximum value at
the top of the hopper, where a bolted joint may be made to connect with
the transition ring. The membrane stress resultant here is also needed in
design of the ring, and is given by overall hopper equilibrium
(21 )
When very light rings are used, the meridional stress resultant falls below
the value given by Eq. 21, but a conservative design of the hopper top
and transition ring is obtained if this effect is ignored.
3.3 Stresses in Symmetrically Loaded and Supported Transition Rings
The transition is a point of greater complexity. The top of the hopper is
in meridional tension, which applies a radial force to the transition (Fig. 6).
This radial force induces a circumferential compression in the transition,
which can lead to plastic collapse or buckling. The radial force per unit
circumference on the ring can be derived from Eq. 21, and induces a
circumferential force in the transition junction of
P = NcI> R sinf3
(22)
When the silo is continuously supported, the circumferential compressive
stresses may be deduced quite accurately using hand methods, for example,
as derived by Rotter (1983, 1985). The phenomena are demonstrated here
by means of an example silo. In large industrial silos, the form of the ring
is relatively well defined. By contrast, a considerable yariety of different
types of ring are used by different cold-formed silo manufacturers. The
example chosen here is intended to illustrate typical stress patterns, without
describing the product of anyone manufacturer in detail. The example silo
is not proposed as an adequate design.
An example silo, with its hopper and supporting ring, is shown in Fig. 7.
The ring consists of a cold-formed channel, which is welded to the top of
the hopper. The cylinder wall is supported from a cold-formed angle, so
that the two components can be assembled separately. The silo is
supported on twelve columns, equally spaced around the circumference. The
stored solid is assumed to have a density of 9.0 kN/m
3
, an effective angle
of internal friction of 28
0
, and a wall friction coefficient of 0.364. The
hopper was subjected to a Walker (1966) discharge pressure distribution
(Eqs 6-9, 11, 12). The silo was analysed using the linear finite element
program LEASH of the FELASH suite developed at the University of Sydney
(Rotter, 1982; 1987c). The corrugated wall was treated as an orthotropic
shell with properties determined using the relationships of Trahair et al
(1983). The stresses are all taken as tensile positive. Two versions of the
535
design were studied: one with a cylinder of simple rolled sheet steel (Design
A), and the other made from corrugated sheeting (Design B). The choice of
cylinder wall type will be seen to have a bearing on the design of the
hopper.
The structure was analysed first as if it were continuously supported all
around the circumference. This illustrates the pattern of stresses arising
in and near the ring as a consequence of the ring compression. The
results are virtually independent of the form of the cylindrical wall. The
circumferential membrane stresses are shown in Fig. 8a. The compression
varies considerably through the ring, being only sensibly constant in the
annular plate element of the channel. Hand methods of analysis are chiefly
aimed at determining the value in this element alone. It is also clear that
circumferential compressive stresses arise in the hopper and cylindrical wall,
that these differ in magnitude from the value in the ring, and that they
decline in a non-linear but rapid manner away from the ring. Thus, a
calculation based on an effective section must be interpreted with care, as
the stress distribution is very different from those found in steel frame
members which do not distort.
The meridional membrane stresses near the top of the hopper are shown in
Fig. 8b. When very small rings are used and the hopper is relatively
thick, the meridional tension falls below the value defined by Eq. 21,
because some of the hopper load is supported by transverse shearing in
the hopper (Rotter, 1987c). The reduced tension leads to a slightly reduced
compression in the ring, but very high bending stresses occur at the top
of the hopper. The meridional bending stresses are shown in Fig. 8c,
where the very high very local maximum at the transition junction is clearly
seen. Alternative hand methods of determining these bending stresses were
advanced by Barthelmes (1977), Fuchssteiner and Olsen (1979), Gaylord and
Gaylord (1984) and Rotter (1985). Some authors (e.g. Gaylord and Gaylord,
1984) have argued that the bending (discontinuity) stresses should be
determined and allowed for in the design. However, unless the silo is to be
loaded and unloaded so many times that a fatigue failure is possible (Rotter,
1985), the bending stresses are not directly implicated in a definable failure
mode, as noted below. The tedious calculation to determine these stresses
is therefore not normally warranted.
3.4 Stresses in Hoppers of Column-Supported Silos
The top of the hopper in a column-supported silo is subject to a stress
state which is closely related to that of the transition ring or junction.
The hopper must therefore be designed with the ring and support condition
in mind.
Column-supported silos present a much more difficult problem than
continuously supported silos. They were first described by Ketchum (1907),
and the procedure which he suggested, although often in serious error, was
not re-examined until the 1980s. Column-supported silos have been the
subject of a number of recent studies (see Rotter, 1988) and several hand
methods for estimating the stresses in both the transition ring and the
cylindrical shell have been advanced. Many of these hand methods are both
complicated and give rather inaccurate results. More reliable stresses are
obtained from finite element calculation (Rotter, 1982).
The problem of the column-supported silo is essentially three-fold: first the
536
ring at the transition must be examined for bending or combined bending
and torsion: second, non-uniform axial forces develop in the cylindrical wall
above the ring, and these may lead to buckling of the cylinder: and
thirdly, the hopper is in non-uniform meridional tension, and this
non-uniformity may lead to rupture of the hopper. This paper deals only
with the last of these three.
The meridional membrane stresses in the hopper of Design A are shown in
Fig. 9a. The non-uniformity of hopper meridional tensions extends
approximately half-way down the hopper. The meridian of the column
support is much more highly stressed than that of the midspan.
The corresponding stresses for Design B are shown in Fig. 9b. These
stresses are significantly higher than those for Design A. This is because
the cylindrical wall of Design A plays an important role in redistributing
the forces from the columns, but the corrugated wall of Design B is very
flexible in vertical deformation, so it cannot fulfil this role. As a result the
ring in Design B is more highly stressed, sustains larger deformations and
leads to greater non-uniformity in the hopper meridional stresses. The
column-support condition affects the circumferential stresses in only a small
zone at the top of the hopper, and most of the hopper is stressed as
defined by Eq. 15.
The circumferential variation of meridional membrane stresses in the top of
the hopper is shown in Fig. lOa for the three conditions of continuous
support (C), Design A and Design B. The corresponding variation of
circumferential stresses is shown in Fig. lOb. The column support causes a
major change in hopper meridional tension, and a significant change in
circumferential stress at the top of the hopper. However, the difference
between a rolled steel and corrugated cylinder wall causes a less easily
anticipated but large difference in meridional tension (40%). Safe hopper
design clearly depends on more than the loads acting on the hopper itself.
4. CRITICAL STRENGTH ASSESSMENTS FOR THE HOPPER
4.1 Introduction
The conical hopper of a silo is susceptible to several different failure
modes, including plastic collapse, meridional seam rupture, and transition
joint rupture. The transition junction, which is intimately related to the
hopper, is susceptible to plastic collapse and buckling, and may initiate
either hopper rupture or cylinder compression buckling. The transition
junction is referred to here only insofar as it affects the hopper design.
4.2 Plastic Collapse of Hoppers
Because the hopper is in biaxial tension, its resistance exceeds the value
determined by simply restricting the effective membrane stress to the yield
stress, provided the hopper seams are strong enough. A clear distinction
must therefore be made between fully welded hoppers and bolted hoppers.
The strengths of fully welded hoppers have been the subject of a recent
study (Teng and Rotter, 1988), in which it was shown that the strength
estimated using the results of membrane theory underestimates the real
strength by about 10%. A typical collapse mode at the top of the hopper is
shown in Fig. lla. The plastic collapse design strength N
evm
for a hopper
537
strake of thickness t with top edge radius R may be closely approximated
(Teng and Rotter, 1988) by
R [0.91 +
tFy R
(23)
in which N
cvm
may be compared with the strake top edge von Mises
effective stress resultant under working loads N
vm
(Eq. 17) and Fy is the
steel yield stress. Because the structure displays a strongly stiffening
plastic behaviour, it is not easy to define the collapse load. Larger
strengths may therefore be assumed if very large deformations of the
hopper are permitted. The critical segment of the hopper is generally the
one with the largest upper-edge radius-to-thickness ratio. For continuously
supported silos, the pressures during discharge are likely to provide the
critical design loading.
The high bending stresses which may be calculated at the transition
junction under elastic conditions lead to early yielding. This yielding has
only a small effect on the load-deflection curve, and a plastic hinge soon
forms, which moves down the hopper as the yielding spreads. Under very
large deformations, the plastic hinge moves far from the junction. High
transition bending stresses are therefore thought to be important only in
silos which may fail in fatigue.
The strength of bolted hoppers is also restricted by plastic collapse, but
the full plastic collapse strength may not be achieved because of premature
joint rupture.
No reported failures of hoppers by plastic collapse are known. The simple
explanation is that the hopper becomes steadily stronger and stronger under
large deformations until a secondary failure occurs by joint rupture. By
then the hopper may be badly deformed, but many users are unperturbed
by the deformations of a hopper local plastic collapse mechanism.
Nevertheless, it seems desirable that hoppers should be designed to prevent
plastic collapse.
4.3 Hopper Meridional Seam Rupture
A light gauge hopper is often fabricated from a number of pie-shaped
segments, which may be field-bolted together. The joints run down the
hopper meridian. This joint must transmit tensile forces arising from
circumferential tension, and should be designed (Fisher and Struik, 1974)
for the highest local value of circumferential tension which can occur in the
hopper. This is given by Eqs 15, 19 and 20. A few failures have been
reported in which the meridional seam in a hopper ruptured, leading to loss
of the silo contents.
4.4 Hopper Circumferential Seam Rupture
Bolted hoppers are sometimes built up from plates which do not extend over
the full height. Joints between these strakes run circumferentially around
the hopper. Each joint should be designed for the maximum meridional
tension which may be transmitted through it. For a continuously supported
silo, this tension can be determined from Eq. 16, using the appropriate
height z. These joints will usually be most highly stressed under initial
filling conditions.
538
1.5 Hopper Transition Joint Rupture
Both "'elded and bolted silos are susceptible to rupture of the hopper
transition joint under the meridional forces which must be transmitted to
the supporting ring. Several failures of this kind have occurred. In most
silos, the majority of the weight of stored solids rests on the hopper, and
the whole of this force must pass through the hopper transition joint. The
transition junction joints are particularly difficult to fabricate, as a conical
shell must be joined to one or two cylindrical shells, and perhaps a ring.
Where the circumferential joint at the transition is given a strength equal
to that of the thinner joined plate, it can be shown that transition joint
rupture is most likely to precede hopper plastic collapse in steep hoppers
with rough walls. Light gauge silos seldom satisfy these conditions, but
they also rarely have joints of strength equal to the joined plate. The
transition joint is usually the most critical design detail for the hopper.
Hopper plate thicknesses are often increased near the transition. The
reason given is generally that high bending stresses occur at this point in
the structure. As noted above, the high bending stresses do not lead to
failure, but the increased plate thicknesses do make it easier to design a
strong joint. Thus it is possible that a satisfactory remedy for transition
joint rupture is in common use, though the reason given is in error.
4.6 Transition Junction Plastic Collapse
The hopper/cylinder junction (transition junction) is in circumferential
compression. The compressive stresses are high because only a small part
of the shell carries the large force required to equilibrate the radial
component of the hopper meridional tension. Whilst buckling failures are
possible if a wide thin ring is used, the most likely failure mode at the
ring is by plastic collapse (Rotter, 1987a).
The plastic collapse strength of the ring and junction under continuous
support conditions was defined by Rotter (1985). Plastic hinges form in the
hopper, cylinder, and skirt (if the design includes a skirt), whilst the ring
itself is quite uniformly yielded under circumferential stresses (Fig. llb).
As the loading rises, the hopper plastic hinge moves from being adjacent to
the junction down into the hopper, redistributing the high elastic bending
stress at the top of the hopper. Plastic collapse of the junction is affected
by the hopper thickness, but the plastic collapse modes of the hopper and
ring do not interact significantly (Teng and Rotter, 1988). A small
transition ring, or no ring at all, generally means that junction plastic
collapse will control the silo strength. Where the ring is made larger,
hopper plastic collapse is likely to control.
The plastic collapse strength of the junction, with or without a ring, may
be closely approximated by
(24 )
in which Pc is the circumferential force in the junction at collapse (cf the
working value derived from Eq. 22), BT is the area of the ring at the
539
junction, t
c
' ts and th are
hopper respectively, and the
may be assessed as lec =
0.975/ (Rth/cosf3).
the thicknesses of
effecth-e lengths of
0.975/(Rt
c
)' les =
the cylinder, sltirt and
adjacent shell segments
0.975/ (Rt
s
) and leh =
For continuously supported silos, traditional simple transition junction design
techniques (Wozniak, 1979; Trahair et ai, 1983; Gaylord and Gaylord, 19841
are generally conservath-e, but sometimes they are "e17 conservative.
Useful savings may thel'efore be made by designing each component to its
real strength. Nevertheless, it should be realised that the transition ring
is ineffective if placed only a short distance above the hopper/cylinder
intersection (Rotter, 1985).
4.7 The Column-Supported Silo and its Transition Ring
One of the most difficult tasks in silo design is to achieve an economic
solution to the problem of an elevated silo on columns. The column
supports introduce high local \-ertical compressive forces into the shell.
These lead to high vertical compressions in the cylinder, and high local
meridional tensions in the hopper. In addition, the transition ring is
subjected to axial compression, bending about two axes and torsion. The
problem of stress analysis of these components was mentioned above, but
very little work has been undertaken to establish rational failure criteria
for any of them.
It has been shown above that the meridional
higher in the hoppers of column-supported
continuously-supported silos. No previous study
failure criteria for the hopper when the silo is
particular, no rigorous calculations are known.
stresses are locally much
silos than in those of
appears to have discussed
supported on columns. In
Three criteria are therefore proposed here: one for ring plastic collapse,
one for hopper plastic collapse and the third for hopper l'upture. It is
proposed that the ring plastic collapse strength should be assessed as the
same as that of the continuously supported structure (Eq. 24). This
proposal is based on the observation that the junction collapse mode
involves large bending deformations of the shell elements meeting at the
tl'ansition, and that significant redistribution may therefore be possible.
It is also proposed that hopper plastic collapse should be assessed as if the
hopper were continuously supported (Eq. 23). The reasons are similar to
those for junction collapse, but the additional observations may be made
that the hopper plastic collapse occurs within the hopper, away from the
region most influenced by the column supports. Further, the non-uniformity
of the meridional membrane stress extends further into the hopper than that
of the circumferential membrane stress. The strength of most designs
(shallo\-:, smooth walled hoppers) should not be affected very much by
locally elevated meridional membrane stresses because the shape of the
biaxial yield criterion indicates insensitivity to this parameter.
By contrast with the two above criteria, it is proposed that hopper
transition joint rupture and joint rupture should be expected
when the maximum meridional membrane stress (Fig. lOa) attains the yield
stress or the strength of the joint. This proposal recognises that local
rupture of the hopper near the column could lead to complete rupture of
the hopper. It is also based on the observation that there is little scope
540
for redistrib:,tion e c ~ u s e the support condition (ring) is in bending, whilst
the. h.opper IS stret?hmg. The scope for redistribution of the high hopper
merIdIonal stresses IS thus very limited. A joint of high strength is most
easily achieyed by using thicker steel sheeting for the hopper.
Unfortunately, no simple hand method of predicting the local high meridional
tensions in the hopper is known, so a finite element analysis is currently
required ""hen designing this joint to the proposed criteria. The rings on
column-supported silos require separate and careful analysis (Rotter, 1982,
1984, 1985), which is beyond the scope of the present paper.
A number of practical matters, relating to the use of steep hoppers, column
braeing and ground-supported skirts are discussed elsewhere (Rotter, 1988).
5. SUJ'vlJ'vlARY
In this paper, a review of design advice for light-gauge cold-formed steel
silos has been presented. The review has made specific recommendations on
the pressures which should be used in design and has described
appropriate stress analysis of the structure. The failure modes which
control the design have been defined, and simple rules for some of these
have been presented.
For column-supported silos, it has been shown that the hopper must be
thicker than it is for continuously supported silos. It has also been shown
that the form of the cylindrical silo wall can affect hopper stresses
markedly. Design criteria for the hoppers of column-supported silos have
been proposed. No comparable existing criteria are known. J'vlore detailed
information is given in a longer report (Rotter, 1988).
6. ACKNOWLEDGEJ'vlENTS
This paper forms part of a major research program into the loading,
behaviour, analysis and design of silo and tank structures being undertaken
at the University of Sydney. Support for this program from the Australian
Research Grants Scheme, the University of Sydney and cooperating
commercial organisations is gratefully acknowledged.
APPENDIX.- REFERENCES
Abdel-Sayed, G., J'vlonasa, F. and Siddall, W. (1985) "Cold-Formed Steel Farm
Structures Part I: Grain Bins''', Jnl of Structural Engineering, ASCE, Vol.
111, No. STlO, Oct., pp 2065-2089.
American Concrete Institute (1977) "Recommended Practice for Design and
Construction of Concrete Bins, Silos and Bunkers for Storing Granular
J'vlaterials", ACI 313-77, Detroit (revised 1983).
Barthelmes, W. (1977) "Ermittlung del' Schnittafte in kreiszylindrischen Silos
mit kegelrmigem Boden", Bauingenieur, Vol. 52, pp 423-435.
BJ'vlHB (1987) "Silos: Draft Design Code for Silos, Bins, Bunkers and Hoppers",
British Materials Handling Board and British Standards Institution, London.
541
DIN 1055 (1986) "Design Loads for Buildings; Loads on Silos", German
Standard, Sheet 6, September.
Fisher, J.W. and Struik, J.H.A. (1974) "Guide to Design Criteria for Bolted
and Rivetted Joints", Wiley, New York.
Fuchssteiner, W. and Olsen, O.W. (1979) "Ein Problem der Stahlblechsilos",
Bauingenieur, Vol. 54, pp 17-21.
Gaylord, E.H. and Gaylord, C.N. (1984) Design of Steel Bins for Storage of
Bulk Solids, Prentice Hall, Englewood Cliffs, New Jersey.
Gorenc, B.E., Hogan, T.J. and Rotter, J.M. (eds) (1986) "Guidelines for the
Assessment of Loads on Bulk Solids Containers", Institution of Engineers,
Australia.
Hofmeyr, A.G.S. (1986) "Pressures in Bins", MSc(Eng) Thesis, University of
Witwatersrand, Johannesburg.
Janssen, H.A. (1895) "Versuche uber Getreidedruck in Silozellen", Zeitschrift
des Vereines Deutscher Ingenieure, Vol. 39, No. 35, pp 1045-1049.
Jenike, A.W., Johanson, J.R. and Carson, J.W. (1973) "Bin Loads- Part 2:
Concepts; Part 3: Mass Flow Bins", Jnl of Engng for Industry, ASME, Vol.
95, Series B, No.1, Feb., pp 1-12.
Ketchum, M.S. (1907) "Design of Walls, Bins and Grain Elevators", 1st edition,
McGraw-Hill, New York (2nd edn 1911, 3rd edn 1919).
McLean, A.G. (1985) "Initial Stress Fields in Converging Channels", Bulk
Solids Handling, Vol 5, No 2, April.
Motzkus, u. (1974) "Belastung von Siloboden und Auslauftrichtern durch
kornige Schuttguter", Dr.-Ing Dissertation, Technical University of
Braunschweig.
Ooi, J.Y. and Rotter, J.M. (1987) "Elastic Predictions of Pressures in Conical
Silo Hoppers", Research Report R555, School of Civil and Mining Engng,
Univ. of Sydney, Oct.
Rotter, J.M. (1982) "Analysis of Ringbeams in Column-Supported Bins",
Eighth Australasian Conference on the Mechanics of Structures and Materials,
University of Newcastle, Aug.
Rotter, J.M. (1983) "Effective Cross-sections of Ringbeams and Stiffeners for
Bins", Proc., International Conference on Bulk Materials Storage Handling and
Transportation, Institution of Engineers, Australia, Newcastle, Aug., pp
329-334.
Rotter, J.M. (1984) "Elastic Behaviour of Isolated Column-Supported
Ringbeams", Journal of Constructional Steel Research, Vol 4, 1984, pp 235-
252.
Rotter, J.M. (1985) "Analysis and Design of Ringbeams", in Design of Steel
Bins for the Storage of Bulk Solids, edited by J.M. Rotter, Univ. Sydney,
542
March, pp 164-183.
Rotter, J.M. (1986a) "On the Significance of Switch Pressures at the
Transition in Elevated Steel Bins", Proc., Second International Conference on
Bulk Materials Storage Handling and Transportation, Institution of Engineers,
Australia, Wollongong, July, pp 82-88.
Rotter, J.M. (1986b) "Recent Studies of the Stability of Light Gauge Steel
Silo Structures", Proc., Eighth International Specialty Conference on
Cold-Formed Steel Structures, St. Louis, Missouri, Nov., pp 543-562.
Rotter, J.M. (1987a) "The Buckling and Plastic Collapse of Ring Stiffeners at
Cone/Cylinder Junctions", Proc., International Colloquium on Stability of Plate
and Shell Structures, Ghent, April, pp 449-456.
Rotter, J.M. (1987b) "Membrane Theory of Shells for Bins and Silos",
Transactions of Mechanical Engineering, Institution of Engineers, Australia,
Vol. ME12 No.3 Sept., pp 135-147.
Rotter, J.M. (1987c) "Bending Theory of Shells for Bins and and Silos",
Transactions of Mechanical Engineering, Institution of Engineers, Australia,
Vol. ME12 No.3 Sept., pp 148-159.
Rotter, J.M. (1988) "The Structural Design of Light Gauge
Research Report, School of Civil and Mining Engineering,
Sydney, May.
Silo Hoppers",
University of
Teng, J.G. and Rotter, J.M. (1988) "Plastic Collapse of Steel Silo Hoppers",
Research Report R568, School of Civil and Mining Engineering, University of
Sydney, April.
Trahair, N.S., Abel, A., Ansourian, P., Irvine, H.M. and Rotter, J.M. (1983)
Structural Design of Steel Bins for Bulk Solids, Australian Institute of Steel
Construction, Nov.
Walker, D.M. (1966) "An Approximate Theory for Pressure and Arching in
Hoppers", Chern. Eng. Sci., Vol. 21, pp 975-997.
Walters, J.K. (1973) "A Theoretical Analysis of Stresses in Axially-Symmetric
Hoppers and Bunkers", Chern. Engng Sci., Vol. 28, No.3, March, pp 779-89.
Wozniak, R.S. (1979) "Steel Tanks" in Structural Engineering Handbook, 2nd
edn, Section 23, Eds. E.H. and C.N. Gaylord, McGraw-Hill.
543
Roof
Ring
Cylinder
Ring
Skirt
Conical
Hopper
FIG.1 TYPICAL COLD-FORMED ELEVATED SILO
S
u
p
p
o
r
t
i
n
g

M
e
r
i
d
i
o
n
a
l

T
e
n
s
i
o
n

'
\

R

(
a
)

E
q
u
i
l
i
b
r
i
u
m

a
n
d

N
o
t
a
t
i
o
n

M
e

z

N
c
p

(
b
)

A
x
i
s
y
m
m
e
t
r
i
c

S
t
r
e
s
s

R
e
s
u
l
t
a
n
t
s

F
I
G
.
2

H
O
P
P
E
R

G
E
O
M
E
T
R
Y
,

L
O
A
D
I
N
G

A
N
D

N
O
T
A
T
I
O
N

c
:
n

.
.
.

.
.
.

545
~
0
LL
QJ
c:
c:
:J
LL
VI
QJ
"'C
0
L
~
0
Vl
LL.
UJ
0
0
.c ~
~
~
0
0
....J

u...
LL
--- ---
....J
--
VI
c(
~
VI
9::
ItI
L
w
z
~
a..
0
z
c(
Vl
> ~ c
UJ
a::
::::>
---
Vl
Vl
---
UJ
a::
a..
---
a::
UJ
c-
D
Z
-
:::J
>-
w
--
L-
,.,..
QJ
"'C l.:i
c:
u...
~
>-
W
c:
VI
QJ
L-
:J
VI
VI
QJ
L-
a..
..
ItI
>-
qt
~ ~ ~ ~ ~ ~ ~ ~ ~ ~
~ = 2 o
f.l=0.4
cD=30
o
(al Hopper
H
546
o 0.2 0.4
p/'6'H
(bl Pressure from
Hopper Contents
Walker
Filling
0.2 0.4 0.6 0.8 1.0
p/qt
(e) Pressure from
Surcharge
FIG.4 THEORETICAL HOPPER PRESSURE DISTRIBUTIONS
I

1
.
0

"
'
-
N

.
.

;

0
.
8

"
"
'
-
<
t
>

=
3
0
0

-
-
-
.
.
:
.
.

.
.

-
-
-

0
.
6

'
"

'
'
'
.

/
,
/

0
.
4
)

)

,

c
:
:

o

0
.
2

Q
J

E

o

o

0
.
2

0
.
4

0
.
6

0
.
8

1
.
0

W
a
l
l

P
r
e
s
s
u
r
e

(
p
/
o
R
)

(
a
)

P
r
e
s
s
u
r
e

D
i
s
t
r
i
b
u
t
i
o
n

0
.
1

0
.
2

0
.
3

0
.
4

C
i
r
c
u
m
f
e
r
e
n
t
i
a
l

S
.
R
.

(
N
e
/
o
R
2
)

(
b
)

C
i
r
c
u
m
f
e
r
e
n
t
i
a
l

M
e
m
b
r
a
n
e

S
.
R
.

,

/
,
,
/
'

/
,
/

0
.
5

0
.
2

0
.
4

0
.
6

0
.
8

1
.
0

1
.
2

M
e
r
i
d
i
o
n
a
l

S
.
R
.

(
c
)

M
e
r
i
d
i
o
n
a
l

M
e
m
b
r
a
n
e

S
.
R
.

F
I
G
.
5

T
Y
P
I
C
A
L

S
T
R
E
S
S

R
E
S
U
L
T

A
N
T

D
I
S
T
R
I
B
U
T
I
O
N
S

(
H
O
P
P
E
R

O
N
L

Y
l

C
l
1

"
'"

-
l

/cone
JI' Meridional
Tension
Ring

Skirt
tsupport
(a) Junction Local Geometry
548
Cylinder
!
VertiCal
Compression
---..... Radial Force
/1
provided by
Cone Skirt Ring Compression
Meridional
Tension
(b) Static Equilibrium at the Junction
FIG.6 TRANSITION JUNCTION EQUILIBRIUM: CONTINUOUS SUPPORT
o
o
L.f'l
(Y'l
o
o
L.f'l
,
3500
o=9kN/m
3
<Pj=28
\-1=0.364
(a) Structure
549
r 100x100x6 L
1mm Smooth Wall (A)
1mm Corrugated Wall (8)
50x50x6 L
1.6mm Smooth Wall (A)
1.6mm Corrugated Wall (8)
See detail

Wall
(b) Ring Detail
FIG.7 EXAMPLE SILO
T

1
0
0
M
P
a

L
-
.
.
.
.
.
.
.
.
J

(
a
)

C
i
r
c
u
m
f
e
r
e
n
t
i
a
l

M
e
m
b
r
a
n
e

S
t
r
e
s
s

c

1
0
0
M
P
a

(
b
)

M
e
r
i
d
i
o
n
a
l

M
e
m
b
r
a
n
e

S
t
r
e
s
s

T
:

t
e
n
s
i
o
n

C
:

c
o
m
p
r
e
s
s
i
o
n

3
0
0
M
P
a

L
.
.
.
-
.
.
.
.
-
.
.
.
J

(
e
)

M
e
r
i
d
i
o
n
a
l

B
e
n
d
i
n
g

S
t
r
e
s
s

(
H
o
p
p
e
r

o
n
l
y
)

F
I
G
.
8

S
T
R
E
S
S
E
S

A
T

H
O
P
P
E
R

T
O
P

(
C
O
N
T
I
N
U
O
U
S

S
U
P
P
O
R
T
)

c
.
n

c
.
n

o

551
X
T
Support
Position
200MPa
L--............
X
Support
Position
200MPa
I I
T
(a) Design A
(b) Design B
FIG.9 MERIDIONAL MEMBRANE STRESSES IN HOPPERS
OF COLUMN-SUPPORTED SILOS
3
0
0

~

2
0
0

L

V
I

V
I

Q
J

L
-
V
)

1
0
0

o
I

,
.
.
.
.
.
.
.
.
-
-
:
:
:
:
:
0
"
"
"

-
-
I

5
0
,
,
-
-
~
-
-
-
-
r
-
-
-

-
-
-
~
-
-
-
-
r
-
-
-
,

T
e
n
s
i
o
n

o

I

I

\

I
I

r
o

a
.
.

L

V
I

-
5
0

V
I

Q
J

L
-
.
.
.
.
.

V
1
-
1
0
0

C
o
m
p
r
e
s
s
i
o
n

D
e
s
i
g
n

B

-
5
0

I

I

I

-
1
5
0

L
.
'
_
_
-
-
-
'
-
_
_
_
_
.
.
L
.
.
.
-
_
_
.
.
.
.
L
.
.
.
_
_
-
-
-
'
-
_
_
_
_
I
.
-
_
_
.
.
.
.
.
.

-
1
5

-
1
0

-
5

0

5

1
0

1
5

-
1
5

C
i
r
c
u
m
f
e
r
e
n
t
i
a
l

C
o
o
r
d
i
n
a
t
e

(
d
e
g
r
e
e
s
)

(
a
)

M
e
r
i
d
i
o
n
a
l

M
e
m
b
r
a
n
e

S
t
r
e
s
s

-
1
0

-
5

o

5

1
0

C
i
r
c
u
m
f
e
r
e
n
t
i
a
l

C
o
o
r
d
i
n
a
t
e

(
d
e
g
r
e
e
s
)

(
b
)

C
i
r
c
u
m
f
e
r
e
n
t
i
a
l

M
e
m
b
r
a
n
e

S
t
r
e
s
s

F
I
G
.
1
0

C
I
R
C
U
M
F
E
R
E
N
T
I
A
L

V
A
R
I
A
T
I
O
N
S

O
F

S
T
R
E
S
S

A
T

T
O
P

O
F

H
O
P
P
E
R

1
5

0
1

0
1

"
"

553
(a) Collapse at Top of Hopper (b) Collapse Mode of Junction
FIG.11 PLASTIC COLLAPSE MODES OF HOPPER AND TRANSITION