bulk

solids Volume 5, Number 2, April 1985

hancll n9

Loads for Desig.n of stacking Tubes

for ·Granular Materials
S.S. 5afarian and E.C. Harris, USA

Abstract used each pair being staggered at 90 ° from the pair above
As yet, there is inadequate published information for the and the pair below. The stacking tube is used to allow
structural design of stacking tubes for use with granular smooth stockpiling of granular materials without causing
materials. This paper (the first of a two-paper series) sug­ excessive dust. Material enters the top of the tube from an
gests load conditions and load combinations that should be overhead conveyor and spills out of the tube by the lowest
considered i; n structural design. Where the method o com­ available opening that is not already surrounded by the con­
puting those loads is not familiar, the paper illustrates ical stockpile of material outside. When the level of material
methods which may be used for their computation. Also covers an opening of the tube, discharge from that opening
given are the results of qualitative tests performed by the ceases and further discharge onto the stockpile continues
authors on two small model stacking tubes. from the next higher opening (or pair of openings).

A second paper will address the design procedures for stack­ Diameter and height of stacking tubes vary depending on the
ing tubes of steel or of reinforced concrete. Illustrative design desired volume of the stockpile. For reinforced concrete
examples will also be given. stacking tubes, heights are as much as 150 ft (46 m) and
diameters usually range from 6 to 15 ft (1.8 to 4.6 m).
Stacking ubes are sometimes built directly on top of a
reclaiming tunnel (Fig. 2) containing a conveyor by which
1. Introduction
A stacking tube (also known as a "lowering tube" or "lower­
ing tower") is a vertical tube of either reinforced concrete or
metal, having openings in its walls at various elevations
(Fig. 1). Pairs of diametrically opposite openings are usually
Stocking
t er

./
D h

Material pile

Reclaiming
tuMel

Fig. 1: General view of stacking tube

Mr. Sargis S. Safarian, P.E., is President of S H Engineering, Inc., La ewood,

CO, and Mr. Ernest C. Harris, Ph. 0., P.E., is Professor of Civil Engineering
at the University of Colorado, Denver, CO, USA Fig. 2: Stac Ing tower on top of reclaiming tunnel

349
 bulk
s•ackina, blendina and reclaimina Volume 5, Number 2, April 1985 soHds
hanclHn■

stored material is withdrawn from the stockpile around the 2. Friction load, parallel to the conveyor and acting in either
tube. In such cases, the tube will have not only side openings longitudinal direction. This would be the friction force due
for discharge onto the stockpile, but also a bottom, hopper­ to expansion or contraction of the conveyor support struc­
shaped opening for direct discharge from the tube onto the ture. The authors feel that if the magnitude of this force
tunnel conveyor. If the material to be handled is prone to is not specified by the owner, the structural designer
deterioration due to long-term storage and a bottom reclaim should assume a force of at least 10% of the total (dead
tunnel is not provided, the tube bottom is sloped so that the plus live) vertical reaction of the conveyor system on top
tube is self-cleaning, the material flowing easily out of the of the tube.
side openings nearest the bottom. For materials that are not 3. Belt tension, if the conveyor is so constructed that the belt
apt to deteriorate under long-term storage (limestone or tension must be transferred to the tube, rather than being
sand, for example) this precaution is not necessary and the resisted by the conveyor support structure itself.
tube bottom may be flat.
3. 2. 2 Wind Loads
If it is permitted by the applicable code or standard being
2. Design observed, the allowable stress for load cases that include
wind may be increased above that allowed for dead and
Two types of design are required:
other live loads alone. (Or, if the design is by strength
1. Functional design, which involves choosing the diameter, methods, the combined factored loading may be reduced.)
height, method of operation, opening size and opening Wind loads acting on the tube vary according to the degree
position, so that the tube will create the desired volume to which the tube is surrounded by the stockpiled material.
and shape of stockpile and so that material flow will be Specific cases that the structural designer should consider
satisfactory. are:
2. Structural design, so that safety (and reasonable 1. Static wind pressure on the projected area of the entire
economy) are assured under all loading conditions. height of the tube, with the tube empty and no stockpiled
This paper addresses only the latter concern - structural material against the outside of the tube. Various codes
design. and standards [1], [2] give design wind pressures for cir­
cular structures. If wind pressures for design are not
To date, no United States code or standard gives structural
specified by the applicable code or standard, the authors
design criteria specifically for stacking tubes. The most badly
suggest using the American National Standard ANSI
needed criteria are those concerning design loads. This
A58.1 [2]. This standard gives design wind pressures at
paper suggests design loadings and shows qualitative
various heights on the structure and considers structure
experimental evidence to support some of them.
shape, purpose and locality. In special locations, local
records may indicate that the specified loads are too low;
in these cases, the designer should seriously consider
3. Suggested Loadings using wind pressures higher than the minimum specified.
for Structural Design Wind force transferred to the top of the tube by the con­
veyor structure should be considered in combination with
Loadings which the authors believe should be considered in the force from wind on the tube itself.
structural design of stacking tubes are the following:
2. Static wind pressure on the portion of the tube that is
A. Dead load exposed above the level of stacked material. This
B. Live loads pressure should be considered along with the lateral
1. Conveyor loads pressure applied by .the stacked material. Perhaps the
2. Wind loads worst, but entirely possible condition would be wind on
3. Stored and stacked material loads the exposed portion of the tube, with the stockpile on the
4. Seismic loads leeward side being partially or completely absent. This
eccentric condition will result when material is removed
3 .1 Dead Loads only from one side of the stockpile. This load combination
Dead loads to be considered for stacking-tube design may be critical for vertical compression on the stacking
include the weight of the tube itself and the dead-load portion tube wall. Again, wind force from the conveyor structure
of the reaction of the overhead conveyor system, if that above should be combined with forces of wind on the
system is attached to the top of the tube. For purposes of tube.
computing horizontal seismic forces on the tube, the weight 3. Variation of wind pressure around the circumference.
of stored material contained within the tube should also be This will cause horizontal bending moments in the tube
considered as dead load. wall. The varying radial pressures have a resultant in the
direction of the wind, and this resultant is resisted by
3.2 Live Loads shearing forces in the wall, as shown in Fig. 3. Bending
moments due to these varying pressures are more likely
Stacking tube live loads are of four kinds: conveyor loads, to be significant in steel stacking tubes than concrete
wind loads, stored and stacked (outside) material loads, and ones.
seismic loads.
3.2.3 Stored Material Loads
3. 2. 1 Conveyor Live Loads
These should include loads both by material within the tube
Conveyor live loads that should be considered are: and by material stacked outside. Except as noted for seismic
1. Vertical reaction (at the top of the tower) to the weight of conditions under dead loads, above, all pressures and fric­
material carried by the conveyor. tional forces due to stored materials should be treated as live

350
 bulk
soHds Volume 5, Number 2, April 1985
ltan41I n■
Stackina, blendin9 and reclaimina

WINO
(F)

Fig. 4: Uniformly distributed stockpile, surface sloping at angle of repose ,p

Fig. 3: Distribution of wind pressure on tube and shear stresses in wall
lf the outer lateral pressure for this condition is larger than
loads. The following stored-material load conditions are sug­ the silo pressure computed for the material stored within
gested, and these are based partly on the qualitative the tube, then the inside pressure resisting inward move­
experimental studies reported later in this paper: ment of the tube wall may be taken as high as the passive
pressure for the inside material. The authors suggest a
1. Stacking tube full, but the outside stockpile absent. This limit for this passive pressure equal to twice the active
condition is, perhaps, onl,y remotely possible, but it might pressure value computed by the Ja n s s e n or Re i m b e r t
occur if all the supposedly self-opening gates on the sides equations (Fig. 4).
of the tube were frozen or rusted shut, or if the tube were
4. Tube either full or empty (whichever is more serious) and
filled with moist material that had become frozen or
one sixth of the outside stockpile removed in a 60 ° sec­
cemented together so that it would not flow as expected.
tor, as shown in Fig. 5. This lateral load condition,
For this load case, the radial outward pressures would be
illustrated clearly by the qualitative experiments, causes
computed as tor a silo, using for example either the
an overturning moment. This moment, in turn, causes
Ja n s s e n or Re i m be r t equations for lateral pressure
large vertical bending stresses which must be combined
[5), [6], [7]. with the stresses due to the simultaneous vertical loads.
For computing the vertical force in the tube wall, friction (A method of analyzing for this condition is illustrated
due to the stored material in the tube should also be con­ later.)
sidered.
2. The exact opposite of Load 1 above, with the stockpile 3.2.4 Seismic Loads
completely in place all the way to the uppermost opening
Seismic forces acting on a stacking tube are illustrated in
and uniformly distributed all around the ou side of the
Fig. 6. They consist of:
tube, but with the tube itself empty. This would be possi­
ble only if the tube had a bottom discharge opening into 1. A lateral force £1 due to seismic action on the weight of
the conveyor tunnel below. This load condition would the conveyor and the material it carries.
cause a hoop compression in the tube wall. 2. A lateral force £2 due to seismic action on the weight of
3. The same as Load 2 above, but with the tube full also. the tower itself and on the weight of material stored in the
This condition would cause lesser horizontal hoop com­ tower. The authors recommend using the entire weight of
pression or tension in the tube walls, but would result in the inside stored material to compute the seismic force,
the largest value of vertical load in the wall due to rather than the 80 0/o used for silo design in ACI 313-R77
downward friction of both the inside and outside material. (5).

Fig. 5: Determination of the design lateral loads (one sixth of stockpile ongle 8 = 90· (CC rup • f)
er
removed) a = )' Do X ( area BC E)

P = Ox an 6)
rup

Active moss (causing overturning)

351
 bulk
s1acklin9, blendlina and reclaimina Volume 5, Number 2, April 1985 solids
handUna

3. A lateral force E3 due to seismic action of the outside Combination E:

stockpile, computed as a product involving either Q3 - Tube full
(Fig. 6) or Q (Fig. 5), whichever controls. For cases in - Outside stockpile absent
which the stockpile is uniformly distributed all around the - Wind on full height of tube
cylinder, the authors would assume that the seismic force - Tube dead load
from material stacked against one side is resisted by - Conveyor dead load
passive pressure of the material on the opposite side. - Longitudinal force from conveyor (if it causes higher tube
However, for the condition in which one quadrant of stress).
material has been removed, this seismic force could be
significant. Combination F:
(if there is a bottom reclaim tunnel)
- Tube empty
- Outside stockpile complete.
Combination G:
E

Same as A, but with earthquake rather than wind.

Combination H:
Same as C, but with earthquake rather than wind.
:r

Combination I:
Same as E, but with earthquake rather than wind.

5. Pressure and Force Computation

Direction 5.1 Lateral Pressure of Stacked Material

of E.Q.
Lateral pressure applied to the walls by the outside stockpile
Fig. 6: Seismic forces acting on stacking tube may be computed using Re i m b e r t's approach for retaining
walls [8], assuming that the surface of the pile is inclined at
the angle of repose for the material. By this approach, the
active unit pressure at height Yper linear horizontal width of
wall is
4. Load Combinations for Design
(1 - 2e/1r)3
Load combinations that are reasonably possible should cer­ Pa = 'Y · Y ---'----'----'- (1)
tainly be considered by the designer. However, such com­ (1 + 2 e I1r)2
binations as full wind and full earthquake force are extremely
The total active pressure per unit width of wall at the bottom
unlikely, and most designers would not consider them. Com­
of the wall is
binations the authors would consider are the following:
Combination A: 'Y · R 2 (1 - 2eI1r) 3
pa = � (2)
- Tube full 2 (1 + 2el1r)2
- External stockpile complete
- Conveyor live and dead loads The .passive pressure of the stockpiled material (useful for
- Tube dead load load combinations that tend to cause overturning) is given by
- Wind on conveyor A. and M. Re i m b e r t [8] as unit pressure at height Yequal to
- Wind on exposed part of tube (1 - 2el1r)2
- Longitudinal force from conveyor (if it causes higher tube 'Y. (3)
Pp = y (1 + 2eI1r)
stress).
Combination B: Similarly, the total force due to passive pressure of the
Same as A, but without wind. stockpiled material is
Combination C: 'Y · R 2
(1 - 2 e I1r) 2
Pp
= �. ---'-----
(1 + 2eI1r)
(4)
- Tube full 2
- Stockpile complete except for a 60 ° sector
- Tube dead load
- Conveyor dead load
- Conveyor live load (if it causes higher tube stress) 5.2 Friction by Stockpile Material
- Wind on conveyor
- Wind on exposed part of tube According to the Re i m b e r t experiments, friction forces on
- Longitudinal force from conveyor (if it causes higher tube the outside of the tube (the downward components of
stress). pressures from the stacked material) are independent of the
magnitude of pressure against the walls. To compute these
Combination D: forces, the Re i m b e r t s suggest assuming the angle of fric­
Same as C, but without wind. tion of the material against the walls to be equal to the angle

352
 bulk
soHds Volume 5, Number 2, April 1985
bandH •
Stackin9, blendin9 and reclaimin9'

of repose. Thus, the outside frictional force per unit area at 6. Seismic Forces
depth Y is
Seismic forces for structural design may be estimated in
= p · tan e (5) severaf ways. A relatively simple approximation is shown
here. More precise methods, if available and practical,
and the total frict,ional. force per unit width of wall at the bot­ should be used in lieu of the approximation. For reinforced
tom of the wall is concrete stacking tubes, Appendix A, Special Provisions for
Seismic Design, of ACI 318-83, Building Code
= P · tan e (6) Requirements for Reinforced Concrete [3], should be
observed where applicable.
The total vertical force per unit width of wall will include the
value given by Eq. (6) above as well as the vertical force due The direction should be the same for all the seismic forces
to dead load and other live loads. shown in Fig. 6; this is consistent with the assumption of first­
mode action. Loads £1 and E2 should be determined follow­
ing the method of the Uniform Building Code (UBC) [1].
These forces would then be given by
5.3 Friction from Inside Stored aterial
£1 = Z · 0.1 Q1 (10)
Vertical friction from the material stored inside the tube may
be computed by either the A e i m b e r t or J a n s se n methods £2 = Z- 0.1 Q2 (11)
[6], {7], (8) for tubes of 8 ft (2.4 m) diameter or larger. For
in which Z is the seismic zone factor. Values of Z are given
tubes of smaller diameter, the authors suggest that the total
downward friction force from the inside stored material be by the UBC as 1.0, 0.75, 0.5 and 0.1875 for earthquake
zones 4, 3, 2 and 1 , respectively.
assumed at about 75 0/o of the weight of the inside stored
material.
To determine load £3 due to the material stockpile with a 60 °
sector removed, the authors suggest the following:
1. Assume that material in the shaded region shown in
5.4 Pressure and Overturning Moment from External Fig. 6 (Area ABO) is effective and that the force £3 acts at
Stockpile with 60 ° Sector Removed the centroid of triangle ABO. Similarily, assume that
material in the Area BCE (Fig. 5) is effective and that the
Fig. 6 shows a method of computing this force and the over­ force E acts at the centroid of triangle BCE.
turning moment it causes. According o R e i m b e r t (8), the
2. Consider the volume of effective material to be the Area
angle of rupture is
ABO multiplied by the tube outside diameter D.
CXrup = (45 - Q/3) (7) 3. Assume that the magnitude of force £3 is given by

(Traditionally, a similar equation is used, but with a "2" in the (or Q in Fig. 5) (12)
denominator rather than a "3" .) In Fig. 5,. the plane of rup­
ture is drawn through poin r t O at the tube center ine, intersec­ where Q3 (or Q) is the weight of the effective material.
ting with the stockpile surface at point A. Next draw line AB,
its slope with the horizontal being equal to the angle of The total base shear due to seismic forces is
repose. Point B on this line is at the tube centerline. From
point B, draw line BC, its slope from the vertical being equal (13)
to the angle of rupture. Point C is at the intersection of line
BC with the stockpile surface.
and the overturning moment about the base is (Fig. 6)
The total! weight of the trapezoidal mass of material (shown
crosshatched) is (1 4)

Q = 'Y · D0 • rea BCE (8)

If £3 H2 computed from Fig. 6 is less than that determined
where D is the outside diameter of the tube. The design from Fig. 5, then the value obtained from Fig. 5 should be
value of force Q is used instead.
Seismic forces and the overturning moment can act in any
P,up = Q · tan fJ horizontal direction .. Vertical seismic forces are also possi­
where ble, but they are normally considered to be of less impor­
tance in design than the horizontal ones.
0 = 90 - (a rup + e) (9)
The authors would use the same approach with respect to
This design force is assumed to act horizontally through the seismic forces for stacking tubes of either material - steel
centroid of triangle BCE, and knowing the amount and loca­ or reinforced concrete. As for load combinations that include
tion of the force, the overturning moment at any elevation wind, design codes either permit higher stress levels or
may be computed. Mass BCE is assumed to be significant in require tower load factors for the design of structures subject
causing the overturning moment. The semi-active Mass ABC to combinations that include seismic loads. (The authors do
is usually considered to have a negligible effect on the not allow such stress increase for the design of anchor bolts,
overturning moment however.)

353
 bulk
S�ackjn9, blendjn9 and reclajmjn9 Volume 5, Number 2, April 1985 soHds
hanclUna

7. Qualitative Test of Model Tubes

To help establish suitable design criteria, the authors per­
formed simple tests on two cardboard model stacking tubes
(Fig. 7). These were loaded with sand to create a complete
surrounding stockpile. The smaller tube was also loaded with
powdered coal. In each case, the surrounding material was
then removed from one 60 ° sector to create an unsym­
metrical load condition.

eE

=Nio

=�

-
� (29mm)

� (19mm)

=.n------
-
-N
E
E
�
in

Fig. 7: Model stacking tubes A and B Fig. 8: Stacking tube (test Model A)

Fig. 8 shows the dimensions of Model A, the smaller model. ing (ovalling) of the tube. The same procedure was followed
This model was first loaded with dry coal of 1/8 inch max­ for the test with sand, the sand having a maximum grain size
imum size. The coal was fed into the top of the tube and of 3/16 inch, but the majority being about 1/32 inch. The
allowed to flow out of the side openings to form the surround­ sand was partly dry and partly damp (Fig. 10).
ing cone of material. After the cone was complete to the Fig. 11 shows the dimensions of Model B, the larger model.
height of the top of the upper openings, coal was gradually This model was tested with sand only. Figs. 12 and 13 show
removed from one 60 ° sector around the tube to create the Model B with sand removed from one sector and two
worst unsymmetrical load condition, as shown in Fig. 9. opposite sectors, respectively. Measurements were made at
Material was then removed from the opposite side of the two stages as sand was removed from one side. These mea­
cone to create the worst condition causing horizontal bend- surements are shown in Fig. 14.

Fig. 9: Coal pile removed from one side (test Model A) 7'' (179 mm)

354
 bulk
solids Volume 5, Number 2, April 1 985
II Ml f ■
Stack•na, blendina and reclaimina

-Ill
N 'It
� eE
E
E
i
-I") -IO

eE It'>

1 4 " ( 3 56 m m ) 24" (61 0 mm)

Fig. 1 0: Dry sand pile removed from one side (test Model1 A)

,n Fig. 1 2: Sand removed from one 90 ° quadrant (test Model B)

(\J

,,' ' ''

,, ',

Fig. 1 1 : Stacking tube (test Model B) Flg. 1 3: Sand removed trom opposite sides (test Model B)

355
 bulk
S1ackina, blendina and reclaimina Volume 5, Number 2, April 1985 soUds
handun■

5 . 5"
( 14 0 mm)

.E
N

-II')

SURFACE DURING
MEASUREMENT

EE EE
CD
�
IO
-IO
N N
ai I')
N
N
/
/

,/

. . . �.' f = 3 6. 4
°
..

5.5"
( 1 4 0 mm)
1 4.5" (368 mm)

33" (8 3 2 mm)

Fig. 1 4: Semi-wet sand pile removed from one side (test Model B)

8. Observations and Conclusions Acknowledgements

When material was gradually removed from one 60 ° sector The authors gratefully acknowledge the cooperation of their
of the stockpile, a steep but unstable slope developed in the staff members at SMH Engineering, Inc., Lakewood, CO,
material at either side of the sector. The height and slope of and the University of Colora_do, Denver, CO.
the mass on either side decreased as material fell into the
area from which material was being removed. Material
directly opposite the region from which material was being References
removed fell in both directions around the tube (if the
material was dry). This flow would continue until the surface [1] Uniform Building Code. International Conference of
slope was at the angle of repose for the material. Building Officials, Whittier, CA 1979.
Damp material did not flow around the tube so readily, and [2] American National Standard Building Code
the pile of material opposite the area of removal remained Requirements for Minimum Design Loads in Buildings
higher. This would suggest that the overturning moment on and Other Structures, ANSI 58.1-1972. American
the tube would be higher for cohesive materials than for non­ National Standards Institute, Inc. 1972.
cohesive. Since stacking tubes are used outdoors, it is [3] Building Code Requirements for Reinforced Concrete,
reasonable to expect that the stacked material will be damp ACI 318-83. American Concrete Institute, Detroit, Ml
and that its cohesiveness will be high. Thus, the designer 1983.
should be conservative in estimating the overturning
[4] Specification for the Design, Fabrication and Erection of
moment due to unsymmetrical loading from the stacked
Structural Steel for Buildings. American Institute of
material.
Steel Construction, Chicago, IL 1978.
The tests showed that the total lateral pressures were
[5] Recommended Practice for Design and Construction of
lowered when material was removed from two opposite sides
Concrete Bins, Silos and Bunkers for Storing Granular
of the tube, so that this load condition was not so serious as
Materials, ACI 313-R77. American Concrete Institute,
that with material removed from one side only. The dif­
Detroit, Ml, rev. 1983.
ference in material level next to the tube was so low as to
suggest that the tendency to cause horizontal bending (ovall­ (6] S a f a r i a n , S.S., and E.C. H a r r i s : Design and Con­
ing) was negligible. struction of Silos and Bunkers. Van Nostrand Reinhold,
New York, NY 1984.
[7] R e i m b e r t , M., and A. R e i m b e r t : Silos - Theory and
Practice. Trans Tech Publications, Clausthal, Federal
9. Structural Design Republic of Germany 1976.
A future paper will address the actual structural design of [8] R e i m b e r t , M., and A. R e i m b e r t : Retaining Walls.
stacking tubes, and design examples will show features of Trans Tech Publications, Clausthal, Federal Republic of
design for both steel tubes and reinforced concrete tubes. Germany 1974.

356