This action might not be possible to undo. Are you sure you want to continue?

of the 6th International

Working Conference

all

Stored-product

Protection

=-Volurne I

**Developments in silo design for the safe and efficient storage and handling of grain
**

A. W. Roberts+

Abstract

This paper presents an overview of some recent developments in the technology of bulk solids hand! ing as it relates to-the grain industry. The paper focuses on silo and discharge equipment design, emphasising the need to understand the relevant properties of the grain and how these relate (0 silo geometry and discharge flow pattern to generate the load patterns exhibited in the silo walls, With the emphasis on grain conditioning procedures, such as aeration. to control grain qual ity, it is important to also take note of the inll uence of any such grain conditioning on the bu lk storage and flow properties of the grain. Any change in the flow properties of the grain can innuence, sometimes significant Iy, the pressures and loadings occurring in silo walls. The matter of structural integrity and safety of the silo and the efficiency and reliability of the discharge equ ipment, such as feeders used to control discharge rate, are of paramount importance. The paper includes a brief review of discharge flow patterns in bins and silos and presents an overv iew of the complex pressure loading patterns in silo walls. Single outlet, multiple outlet symmetric silos are discussed as well as silos with eccentric loading due 10 either non-symmetric loading, or silos with eccentric discharge openings. Brief mention is made of the relevant flow properties such as angles of internal friction, bulk strength and friction angles between the grain and silo walls. Particular mention is made of the influence of grain 010 isture content variations. interstitial air. and temperature variations on the pressures and hence stresses occurring in silo wadis. Methods of controll ing loads in tall, sl ngle and m LIlli outlet silos are discussed. Particular mention is made of the use of antidynamic tubes to control the flow pattern and, hence, control the magn itude of the wal t pressures generated during discharge. A recurring problem in bins and silos is the phenomenon of 'silo-quaking', a term used to describe pulsating loads which may be experienced duri ng discharge. If the frequency of the flow p u 1sarions is in harmony with any of the natural frequencies of the structure, severe dynamic loads and stresses can occur. The various mechanisms of 'silo quaking' will be outlined and methodologies to determine the magnitude and distributions of the dynamic loads will be given. The paper will be illustrated by results from tests performed on pilot scale model silos, as we II as by case studies dra wn from field experience.

design of safe and functional storage and handl ing plant which ensures efficient and reliable operation. A full understanding ofmodem technology of silo and handling plant design, which takes account of the relevant physical properties of the granular materials under the varying environmental conditions during storage, is essential. Over the past three decades much progress has been made in the theory and practice of bulk solids handling. Test pron,:dures for determining the strength and flow properties of bulk solids have been developed and analytical methods have been established to aid the design of bulk solids storage and discharge equipment Much progress has also been made in the understanding of the complex nature of loading conditions that may be generated in walls of storage bins and silos. There are now reliable theories and associated design procedures that may be applied, with confidence. to the solutions of practical silo design problems, Yet despite this progress, there are still a number of uncertainties particularly in relation to changes in environmental conditions in grain silos. notably those due to temperature and moisture as well as those due to unpredictable flow patterns. The latter often arises in the case of tall silos and in silos with multi or eccentric outlets. The purpose of this paper is to present an overview of tile present state of knowledge associated with silo design for the safe storage of bulk granular materials. The factors influencing the loads in the walls of silos are many and varied and, as is oftenthe case, the loads are a combination of several interacting effects .. The paper, focuses on silo wall loads indicating how these are influenced by the discharge flow pattern ..The complexity of loading patterns due to eccentric and multiple outlets is mentioned. The effects of moisture COil tent on gra in swellingis discussed with particular reference to the significant increase in silo wall loads that may occur. In addition, the influence on wall loads of pressures due to aeration and thermal expansion and contraction of silos is indicated. The use of ani i-dynamic tubes to control the wall pressures during discharge of grain in. tall single and multi-outlet silos is outlined. Finally the phenomenon of 'silo quaking', a term used to describe pulsating loads which may be experienced during discharge is discussed. If the freq uency of the flow pulsations is in harmony with any of the natural frequencies of the structure. severe dynamic loads and stresses can occur. The various mechanisms of 'silo quaking' are outlined and methodologies to determine the magnitude and distributions of the dynamic loads is g iven. .

Introd uction

Bulk materials handling operations

are a. key function in agricultural grain production. The capital costs of storage and handling plant are quite substantial, so too are the operating and running costs. With the emphasis on grain quality, it is not only important thatattentlon be given to the maintenance of appropriate grain conditioning during storage, but also to the

~ Institute for Bulk Materials Handling Newcastle, NSW 2038. Australia Research. The University of

Handling Plant Design-Basic

Concepts

As background information to the study of grain storage and handling, the basic concepts of handling plant design for the general c lass of bulk materials is brief! y rev iewed. General. remarks The procedures for the design of handling plant, such as storage bins and silos, feeders and chutes arc well established and follow the four basic steps:

259

Proceedings (i)

of the 6th International

Working Conference on Stored-product

Protection -Volume

I

Determination of the strength and flow properties of the bulk solids for the worst Ii kel y fl ow cond itions expected to occur in practice. (ii) Determination of the bin, stockpile, feeder or chute geometry to give the desired capacity, to provide a flow pattern with acceptable characteristicsand co ensure that discharge is reliable and predictable (iii) Estimation of the loadings on the bin and hopper walls and on the feeders and chutes under operating cond itions. (iv) Design and detailing of the handling plant including the structure and equipment .. The general theory pertaining to gravity fiow of bulk solids and associated design procedures are fully documented (Jenike J 961, 1964; Arnold 1982; Roberts 1988). Par the purpose of the present discussion, the salient aspects of the general philosophy are briefly reviewed. Modes

0

f now in bins and

Sil05

of symmetrical

geometry.

Following the definition by Jenike (Jenike 1961, 1964) the two principal modes of How are mass-flow and funnel-flow. These are illustrated in Figure I. In mass-flow, the bulk solid is in rnotion at every point within the bin whenever material is drawn from the outlet. There is flow of bulk sol id along the walls of the cylinder (the upper parallel section of the bin) and the hopper (the lower tapered section of the bin). Mass-flow guarantees complete discharge of the bin contents at predictable flow rates. It is a 'first-in, first-out' flow pattern; when properly designed, a mass-flow bin can re-mix the bulk solid during discharge should the solid become segregated upon filling the bin. Mass-flow requires steep, smooth hopper surfaces and no abrupt transitions or in-flowing valleys. Mass-flow bins are classified according to the hopper shape and associated flow pattern, The two main hopper types are conical hoppers which operate with ax i-symmetric flow and wedged-shaped or chisel-shaped hoppers in which plane- flow occurs. In plane-flow bins, the hopper half-angle a will usually be, on average, approxi rnately 8 to 10 larger than the corresponding value for axi-symmetric bins with conical hoppers, Therefore, they offer larger storage capacity for the

or

0

0

same head room than the axi-symmetric bin, but this advantage is somewhat offset by the long slotted opening which can give rise to feeding problems. The transition hopper, which has plane-flow sides and conical ends, offers a more acceptable opening slot length. Pyramid shaped hoppers, while simple to manufacture, are undesirable in view of build-up of material that is likely to occur in the sharp comers or inflowing valleys. This may be overcome by fitting triangular" shaped gusset plates in the valleys. Funnel-flow occurs when the hopper is not steeply sloped and the walls of the hopper are not smooth enough. In this case, the bulk solid sloughs off the top surface and falls through the vertical flow channel that forms above the opening. Flow is generally erratic and gives rise to segregation problems. In the case of cohesive bulk solids, flow will continue until the level of the bulk solid in the bin drops an amount H D equal to the draw-down. At this level, the bulk strength of the contained material is sufficient to sustain a stable rathole of diameter Vf as illustrated in Figure I b. Once the level def ned by H D is reached, there is no further flow and the material below this level represents 'dead' storage. This is a major disadvantage of funnel-flow, For complete discharge, the bin opening needs to be at least equal to the critical rathole dimension deterrnined at the bottom of the bin corresponding to the bulk strength at this level. However, for many cohesive bulk solids and for the normal consolidation heads occurring in practice. ratholes measuring several metres are often determined. This makes funnel-flow impracticable. Funnel-flow has the advantage of providing wear protection of the bin walls, since the material flows against stationary material, However it is a 'first-in last-out' flow pattern which is unsatisfactory for bulk solids that degrade with time. It is also unsatisfactory for fine bulk solids of low permeability. Such materials may aerate during discharge ·!hrough the flow channel and this can give rise to flooding problems or uncontrolled discharge, In the case of cohesive bulk solids, the disadvantages of funnel-flow may be overcome by the use of expanded-flow,as illustrated In Figure 2. This combines the wall protection of funnel-flow with the reliable discharge of mass-flow. Expanded-flow is ideal where large tonnages of bulk solid are

I""_'_D~

I

Total

Capacity

live

Funnel Flow

Flow Along

Central Channel

Walls

Fig.l

Modes of flow

(2)

Mass-flow; (b) Funnel- flow

Fig. 2.

Expanded flow

260

it is recommended that the haif-angte be chosen to be 3 less than the limiting value. In the case of planeflow. last-out' flow pattern which could lead to a reduction in grain quality. if the grain is at a higher moisture content. there is a minimum level Her which is required to enforce mass-flow in the hopper (Thompson 1984).an 'effective transition' (ET) forms and this defines the lower region of the flow channel during discharge.. symmetric bin shapes provide the best performance. In the earlier theory. For massflow in conical or axi-symmetric hoppers. they have no bulk strength. the bounds between mass and funnel-flow are much less critical than for conical hoppers. For this region.ll. Hence funnel flow. Typically for wheat. In this case mass. Genera. the design limit may be selected. if the transition of the hopper and cylinder is sufficiently radiused so that the possibility for material to build-up by adhesion is .-= 60 PLANE-fLOW .. Expanded-flow bins are particularly suitable. no account is taken of the influence of the surcharge head due to the cylinder on the flow pattern developed.J . the effective angle of internal friction S and the wall friction Since in the work of Ienike. needs to be noted. 1964).lIy speaking. Once the wall friction angle and effective angle of internal friction have been determined by laboratory tests. the advantages of mass-flow should not be overlooked even though a steeper hopper is required as in Figure la. = 30 ¢ = 20 for polished stainless steel or polished mild steel and ¢. lenike published a new theory to improve the prediction of funnel. Asymmetric shapes often lead to segregation problems with free flowing rnateri als of di fferent particle sizes and makes the prediction of waUloads very much more difficult.J "'" (a) Flat Bottom (b) Hopper Bottom ): 0 10 20 3J 4) SJ ffI 70 HOPPER HALF-ANGLE ex (Deg_) FIg..J '" e . The limits for axt-syrnmetric or conical hoppers and hoppers of plane-symmetry depend on the hopper half-angle 0.:. as illustrated in Figure 3.== 60° 6 == 50· 0= 40" fitI ~~~ I~ FUNNELflOW Ul Z 40. this led to new limits for funnelflow which give rise to larger values of the hopper half-angle than previously predicted. For this reason.. the boundary between mass" flow and funnel-flow was based on the condi lion that the -e.J Z 0 IIX 10 FLOW 10 0. The bounds fOI conical and plane-flow hoppers are plotted for three values of S in Figure 4.SO w . 3. this height ranges from approximately 0 . Q ... particularly in the region of the transition. Even though funnel-flow silos may be used to store and handle grain. is a viable option. the corresponding angles are Up = 35" and up" 27" respectively. During discharge. For plane-flow. 3D "'" l.. the dimension at the transition of the funnel-flow and mass.o) 0 (I) Free-flowing granular materials Since most bulk grains at low moisture contents are generally free flowing. 1 angle i. The concept of expanded. average values of the friction angles are . of discharge.0 D. then a half-angle 3° 10 4° larger than the limit may be chosen. In the case of conical or ax i-symmetric hoppers. it will exhibit cohesive strength. ~ . the grain will eventually settle at its final repose angle as indicated as indicated in Figure 3a. the grain retained in the 'dead' region of the silo may degrade with time and contaminate fresh grain loaded into the silo. However. In plane-flow hoppers. the silo should be self em pty ing. These limits are well known and have been used extensively and successfully in bin and silo design. For the mass-flow bin of Figure I<I.flow is definitely the preferred option. Q 0 Modification 10 mass-flow Iimits=more recent research Mass-flow and funnel-flow limits for symmetrical bins Established theory due to Jenike The mass-flow and funnel-flow limits have been defined by Jenike on the assumption that a radial stress field exists in the hopper (len ike 1961. For complete discharge. much huger hopper half angles are possible which means that the discharging bulk solld will undergo a signi ficant change in direction as it moves from the cyl inder to the hopper. For instance.Proceedings of the 6th international Working Conference on Stored-product Protection -Volume to be stored. particularly for high values of the wall friction angle.~ignincantly reduced. Once the head drops during the last stages. In functional form.4 Limits for mass-flow for conical and plane-flow charnels . flow in a hopper is based on the radial stress fietd theory. the hopper half-angle may be determined.75 D to 1. which may be accornpl ished by using a hopper bottom as illustrated in Figure 3b. For plane-flow... For instance. flow. wedged-shaped mass-flow hoppers.~ 25° for mild steel plate as rolled. Certainly. for storing large quantities of bulk solids while maintaining acceptable head heights. It is been known for some time that complete mass-flow in a hopper is influenced by the cylinder surcharge head.flow sections must be at least equal to the critical rathole dimension at that level. the disadvantages of a 'first-in. the required hopper half-angles (Figure I (a» are etc ~ 25" for the polished surfaces and ac = 17 for the rolled mild steel.. flow may be used to advanrage in the case of bins or bunkers with multiple outlets. =0 r($. More recent research has shown that the mass-flow and funnel-flow limits require further explanation and refinement. 0. Funnel-flow for free-flowing granular material 261 Fig.

6. bin shapes should be kept simple and symmetric. that is. in funnel-How bins the flow pattern is. the hopper half angle e. The parameter used! to describe the strength of a bulk solid is the Flow Function which is abbreviated as 'FF'. AS 3774-1990. Despite the widely varying approaches to the analysis of bin wall loads. the actual Flow Factor ffa applies and the major consolidating pressure at the outlet increases to the value corresponding to the actual opening dimension B. Referring to Figure 7b. Whereas the radial stress theory ignores the surcharge head. It then remains to determine the required dimension of the channel to ensure that arching does not occur and the required flow rate is achieved. 1964.:. Ooi and Rotter (1 989). Oc = 0 and the FF graph coincides with the horizontal axis. Slow flow In outer region Pulsating flow in central region Fig. namely the 'arc theory'. In most cases. In particular.is in itially filled from the empty condition. In view of its obvious importance it is a subject that has attracted a good deal of research effort.LE MASS. He deri ved a fundamental relationship for Her in terms of the various bulk solid and hopper geometrical parameters. For free-flowing grain. more difficult to ascertain. the . The characteristics of pu Isating dynamic loads is discussed j n Section 8. Benink has shown that the surcharge head has a significant influence on the flow pattern generated. This theory predicts the critical height H crat which the flow changes. Roberts (I 988a. the loading of the bin is not central and/or the bulk solid is prone to segregation. Roberts 19RB). p = Bulk Density g = acceleration due to gravity In practice. The FF represents the variation of the unconfined yield strength Oc as a function of the major consolidation pressure 0 [ and is illustrated in Figure 7a. the critical dimension for flow occurs at the outlet. where. a peaked stress field occurs as in Figure 8a. Benink developed a new theory. The flow pattern which a mass-flow bin exhibits . However. Intermediate How illustrating pulsating now regime.. the minimum B occurs when -01= Oc and is given by by Benink B . Arnold 19R2. 'Loads on Bulk. f FLOW limit on the basis of the wall friction angle $. and Wu (1990). Ooms and Roberts (1985). the stress acting in the arch is represented by 0'1 and is related to the maJor consolidation pressure 0 I by the' flow factor' ff. The subject of bin loads has been addressed in several design codes notabl y the recent Austral ian Standard AS3 77 41990. Grain Rows more quickly inthe central flow channel. Having selected the basic shape. whether the hopper is to be axi-symmetric (conical) or planeflow (wedge-shaped). to quantify the boundaries for the three flow regimes. H(a) = Arch thickness and hopper geometry factor parameter based on Hopper Half Angle (Jcnije 1961. but the later codes cover both mass-flow and funnel-flow bins.b). notably the H:D ratio of the cylinder and the effective angle of internal friction O. it is clear that the loads are directly related to the flow pattern developed in the bin or silo.is reasonably easy to predict and is reproducible. A brief selection of published research which deals with the subject of wall pressures is given in Arnold (1982). WALt FRICTION At. a flow channel parameter. The nature of the flow is such that pulsating loads may be induced. funnel-flow and an intermediate flow as illustrated in Figure 5. Fig. Benink (1989) has identified three flow regimes. Flow regimes for plane-flow ( 1989) hopper defined For the critical condition. the actual opening dimension B is made larger than Bmin in order to achieve the desired flow rate. especially if the bin has multiple outlet points. Hopper opening dlmenslon for mass-flow Symmetric mass-flow bins The principles embodied in mass-flow hopper design are illustrated in FIgure 7. When a bin .S. Solids Containers' (SAA 1990). is very comprehensive in the range of loading conditions included and in the types of bins considered. For the enlarged opening. the prediction of wall loads continues to be a subject of some considerable complexity. Unless there are compelling reasons to do otherwise. Roberts and Ooms (1983). Rombach and Eibl (1989). The earlier codes concentrated on funnel-flow bins. Intermediate-flow is illustrated in Figure 6. Silo Wall Loads In bin and silo design. = ujH{a) min opening dimension (3) os In a comprehensive study of flow in silos. In the revised theory the flow boundary is based on the condition that the velocity becomes zero at the wall. mass-flow.Proceedings of the 6th International Working Conference on Stored-product Protection -Volume 1 stresses along the centre line oflhe hopper became zero. is determined 262 The stress fields and normal wall pressures occurring in mass-How bins for the initial filling and flow conditions are shown in Figure 8. as follows: (2) HOP PER HALF -A tli LE a.

Imperfections in bin walls. which give rise to over-pressures. th is only occurs if the bi n is com pletely emptied and then filled again. This is indicated by the upper bound Pn curve for the cylinder in Figure 8b. the switch travelling up the hopper becoming locked in at the transition as in Figure 8b. Tn view of the difficulty of using the Jenike strain energy method. that is in the cylinder. Switch Pressure J<::JE--. the effective transition previously defined. the stress field in the hopper switches to an arch stress field.No Flow Condition (b) Critical Arching Dimension Fig. Stress Field Radial Initial Filling (b) Flow (c) Normal Wall Pressure Fig. ma-y be due to manufacturing and/or constructional details such as weld projections or plate shrinkages in the case of steel bins or deformation of form work in the case of concrete bins. In the arched stress field. controls the flow channel in the lower region of the bin as ind icated in Figure 9.Proceedings of/he 6th lnternational Working Conference on Stored-product Protection -Volume 1 major principal pressure is almost vertical When flow occurs. Deterrni nation hopper opening dimension for rnass-fl ow. the stress fields and corresponding pressures shown in Pigure 8b wi II remai n. Jenike has used strainenergy methods to analyse these over-pressures during flow in the cylinder.S_ Pressures acring in mass-flow bins. The stress field does not revert to that of Figure 8a.--. the peaked stress field remains" although imperfections in the cylinder wall which give rise to localised flow convergences cause overpressures to OCCUf in the cylinder. Wall pressures in symmetric funnel-How bins In the case of symmetrical funnel-flow bins or silos. Above the effective transition the now occurs along the walls similar to that occurring in a tall mass-flow bin. It is to be noted that when the bin discharges and the flow is stopped._7. Above the hopper. the load is transmitted to the wall of the hopper with the major principal stress acting more in a horizontal direction. design codes generally employ over-pressure factors to account for flow conditions in the cylinder. 263 .J. An over-pressure occurs at the effective transition similar to the switch If 0'1 corresponding to opening dimension B (a) Flow .

52 5 6 29.. are normally computed by the application of over-pressure factors on Janssen. for example. The loading conditions that may occur are shown in Figures 13. This shows the effects of wheat grains after prolonged exposure to high humidity conditions.0 6.5 14.58 19. pressure of28 kPa. Apan from the bending stresses which are superimposed on the average hoop stresses at each location. When gra in is subject to swelling. the distorted shape of the silo may give rise to buckling problems.b. if not completely constrained. the pressures acting on the wall in !he region of !he outlet are of lower magnitude than the pressures that act in the section of the bin or silo where there is no flow. The moisture content of the grain in the various cells was varied and movement of the piston due to swelling of the grain over varying periods was noted. In the limiting condition the wall friction acts downward as indicated in Figure 13. Loads in Grain Silos due to Grain Moisture Increase Possible J 2 and loading condluons Moisture Variations in Grain Free expansion due to moisture increase Grain. the procedure for design involves defining the lowest position at which the effective transition is likely !O occur. The results indicate.si 10diameter y= bulk speci fic weight of wheat Table 1 Free expansion of humidified wheat Storage time (hours) 1. the conservative approach to design is to assume that the possibility exists for only one of the outlets to operate at a particular instant.Q 95_0.0 24. The subjectof eccentric loads in silos is addressed in a comprehensive way in AS3 77 4 (SAA 1990). when subjected to increases in moisture. that an increase of moisture content from 9% to approximately 23%. It is to be noted that where there are several symmetrically located outlets.65 2 3 4 13. Since the effective transition is not stationary but tends to move during the discharge process. Under normal conditions of storage moisture variations of this magnitude are unlikely.0. Assuming the pressure is uniform across the cross-section.:>8 8.. Consequently bending moments ar-e induced in the walls leading to a distornon of the silo's sy mmetrical cyl indrical shape.75 5. The flow pressures in funnel-How bins. the corresponding expansion of the bulk grain is 1. on the basis of such results the increased loading on the walls of silos may be estimated. A typical set of results is shown in Figure 11 which applies to a.79 264 .) Initial (%} Volume increase 0.16 Sample I Moisture content (w.2%. imperfections in the walls can give rise to increased pressures. whilst significant._62 4. may not be as severe as the mass-flow switch pressure. For moisture changes of lower order the degree of swell ing is correspondingly lower.32 4. such as wheal. J (4) where +sign appl ies 10 the Janssen case of Figure 12 -sign applies to the swell ing case of Figure I 3 11" coefficient of friction as the wall Pv= vertical pressure Pn= lateral pressure D. 9.7 6. the results being based on samples conditioned in a humidity cabinet. the upward movement being resisted by wall friction. Referring to Figure 10. it tends to expand upward as well as outward.74 (%) Final (%) 7.83 16. The wall friction may not be fully mobilised and the resistance at the wall to upward expansion may be less than for the Iirniting condition. the following differential equation arises dPn 4pKj _± __ p dz: D n = yK. However the results indicate the order of swelling that may occur during constrained storage. In ~the upper cylindri. 3.Proceedings of the 6th International Working Conference on Stored-product pressure in mass-flow.07% after 4 days. As indicated the moisture content increased from 9. although the magnitude.8%.0. Changes in moisture content can occur during aeration of grain in bulk storage when the humidity of the circulating air increases.26 20.91 10. The loading patterns illus tra ted in Figure 10 are based on this Standard. The amount of grai n swell ing is i\I ustrated in Table I..51 9_54 16.54% to 20. Mul tip Ie outlet bins and silos and with eccentric discharge silos Multiple outlet silos and silos with an eccentric discharge point give rise to complex loading patterns as a result of the flow channels that develop above the outlets. the pistons being free to move in the vertical direction.40 17. The procedures are given in AS 3774 (SAA 1990)_ Protection -Volume 1 Constrained expansion due to moisture Increase In a storage bin situation the volume increase due to swelling is constrained. the corresponding increase in volume due to swelling being 17. will undergo an increase in volume.0. To gauge the possible volume expansion under such constrained ccndltionsvtests were conducted on wheat contained in cylindrical cells with rigid bottoms and s ide walls with a loaded piston bei ng located on the top of the contained samples.83 11.cal section of the bin above the effective transition. depicted in Figure 9.

9. Fig. 10. Circu mfere ntia 1 pressure variation due to operati on of one eccentric outlet. Pressures in.Proceedings of the 6th International Working Conference on.. Reduced Increased Uniform PressureSymmetric Discharge l~ . 265 . Stored-product Protection -Volume J Locus of Pressure Peaks During Discharge ssen Effective Transition Lowest Position of Effective Transition Fig.. -----I}" h.funnel Row silo.

--' 0 > ~ w z . Expansion (Minutes) 1000 of wheat due to swelling underconstrained storage pressure applied: 28 kPa.t'--··--···---. Putting the minus sign into equation (4) and solv ing yields Pn='YD(e4I'Ki76_I) 4).5 10 100 TIME Fig. 11. where the loading is symmetrical.5 Moisture Content o w :::!: ::::l D 9% (Datum) 14% J":: -··········-·····.-I---~~. Fig.J (5) grain swelling (6) Grain swelling may occur if moist air is circulated through stored grain via aeration ducts as illustrated in Figure 14. Janssen pressure due to static loading. Physically the effect of grain swelling is likened to a piston trying to push the column of grain upward (Roberts 1988). it is usua I to apply an over-pressure factor to the pressure Pn of equation (5). 'Piston' equation/or P"=~(I-e-4pKjYv) and p " =Pn K.Proceedings of the 61h International Working Conference on Stored-product Protection - Vol urne 1 1.« :::c 0 CJ 0. Effect of upward swelling producing forces at wall reverse friction Janssen Equation Putting the + sign into (4) and solving yields the Janssen equation For silo design.1 (7) 266 . 13. Fig. 12.5 #- a -0.

(7) and (8) are shown In Figure 14. The pressure curves in Figure 14 apply to a I 5 m diameter by 25 m tall wheat si 10 and have been plotted for a constant val ue of Kj = 0. 267 .. The large magnitude of the pressures given by equation (7) other than for the Janssen pressure given by equation (5) is clearly evident. Reduction ill wall pressures in lower region of silo walls. 15. If the floor of the silo is not entirely rigid. Pv = ')'Z (8) and Pn=Kj/'z (9) The graphs depicting the wall pressures. produced in the silo where the forces in the silo wall resisting upward expansion of the grain can become quite considerable. This would be particuIarly the case for tall silos where the height is several times the diameter. 14. This is depicted in Figure 15.l 0. 150 200 (kPa) Wall pressure d istri bu I ions in 15 m diameter by 2S rn high wheal si [0. a 'piston effect' is. 0 20 25 +---LT"""-:L.T"""--r--~ o Fig. ~ (kPa) Equation (7) assumes the wall friction is fully mobilised. In this case the vertical pressure Pv is equal to the hydrostatic pressure Some general comments There seems no doubt that an increase in moisture content of the grain can cause a significant increase in the silo wall pressures. It is possible that K.. Janssen Reduction in Pressure due to Non Rigidity of Base Normal Pressure Fig. increasing the pr6ssures beyond those plotted.Proceedings of the 61h International Working Conference on Stored-product Protection - Volume 1 5 E N 10 15 s: <l.. The key to this behaviour is the dlrection and magn itude of the friction force at the wa [I.. When the friction force acts downward. An intermediate case of interest occurs when the wall resistance reduces to zero which implies that the vertical wall support just equals the upward force at the wall due to swell ing. will be higher for the grain swelling case. then the frictional forces in the wall in the lower region of the silo would act upward and the pressures exerted on the silo walls in this region would reduce.4. The extent to which the frictional forces at the wall change in magnitude and direction over the full height of the silo arc virtually impossible to predict. 50 100 Normal Pressure 11. based on equations (5).

s N 5 X=2m :c b= L. Some typical results for various R81Rr ratios and various val ues of X (Figure 16) being shown in Figure 17. I:I~ r=---:--~----. Boundary Layer Annulu H Central Segment Fig . there is every reason to believethat the moisture content of the stored grain would not be uniformly distributed throughout the mas!'>. as the dimension X increases.16. made in the foregoing analysis.5.J X=3m X=5m X".25 m.5 times the Janssen pressure. the estimated increase in pressure is of the order of 2 10 I.7m 15 20 o o Fig. Silo pressure model showing boundary layer annulus to account for grain swelling. With the presence of a boundary layer the pressures acting at the wall are reduced considerably.Proceedings of the 6th International Working Conference on Stored-product Protection -Volume 1 R RT . the presence of the boundary annulus significantly influences the wall pressures.. there is little variation in pressure with increase in X. while at the depth z .. Based on the assumptions O~-------------------' --. and allowing for the relative elasticity of the wheat and the steel si 10 shell. and the greater becomes the pressure. Simplified theory for estimating silo wall pressures due to grain swelling It Isevident that the large size of actual silos would be such that pressure distribution at any cross section would not be uniform. the greater become the proportion of the grain mass which must be constrained by the silo walls and hence the greater become the pressures on the walls.the geometry andIocation of the air supply ducts suggest this would be the case. Pressure distributions.0. as the thickness of this annulus increases.10m. It is to be noted that based on experiments to examine grain swelling. RBI Rr ~ 0. For comparison purposes the Janssen curve and the pressure curve for reverse friction or shear of Figure 14 are also included in Figure 17. Furthermore. The change in pressure with change in dimension X are shown more clearly in Figure 18. The silo in this case is 15 m diameter by 25 m high and stores wheat.10 . At the base of the silo. the pressure increases quite significantly as X increases At the depth z ~ 20 m the increase in pressure is less pronounced.2 times the Janssen pressure whi Ie with a boundary layer thickness of say X ~ 2 m. 268 .. then the closer does the pressure distribution approach that depicted in Figure 14. The differential equations for the grain pressures have been derived and numerical so luti ons obtained(Roberts 1988a). As indicated.0.5. The model used by Roberts (I 988a. this surface being sloped I'll an angle 13 to the vertical and intersecting the top surface of the silo at a radius Rr It is assumed that the conical central portion of the grain mass is able to expand upwardly due to grain swelling and induce pressures in the outer.. A conical rupture surface is assumed. wedge-shaped annulus comprising the boundary layer. In general.. The outer boundary layer annulus is constrained by the silo walls.L. In reality the behaviour of the grain during swelling is likely to produce a boundar)' layer effect. the results indicate that the maximum pressure at the base of the stlo with zero boundary layer is close to 200 kPa. 50 100 150 200 PRESSURE~ (kPa) Influence of boundary layer on si 10 wall. The peak pressure with zero boundary layer is 4.b) to analyse this case is illustrated in Figure 16..4 and RBIRr . that is at z . These values are based on the pressure ratio K . 17. the wall pressure at the base of the silo is approximately 1.

h P =-p_ <I hf) ~' where I'!.. ~-.. Silo aeratlon-> effect where Ph = 0. are briefly discussed.. namely aeration and temperature effects... --3 _.... .. ..T -------_Bs o_C RT = 0. --..:: (I_ Ph -»..._ . other environmental factors may influence bin and silo wall loadings.. during aeration and concluded that body-centred cubes would occur most frequently.p_ Maxi m m aeration pressure flo = Effective maxi· urn material depth.. . These produce a maximum increase in value (12) (b) Contraction Fig.. 18_ Variation of s .:. ( (0) e -4jlKQYo ) (II) 'Y~ Bulk specific weight D ~ Silo diameter !J.. . Jenkyn (1978)con5 idered likely packaging configurations.. ~. occupied by the material of approximately 25% with a corresponding decrease in density.. 269 ... ~. The aeration pressure is given by I'!. Dimension 7 X (m) 8 at selected depths....Proceedings of the 6th International Working Conference on Stored-product Protection -Volume 1 --- RR· = 0. Expansion and contraction of silo..._---4 .10 wa II pressures with boundary layer. or air pressure Pressures due to aeration must be added to the pressures generated by the bul k materialitsel f. Jenkyn recommends a design pressure given by P<les = Other Factors Influencing Bin and Silo Loadings In addition to moisture. Two of these factors. . it does provide an indication of the influence of a boundary layer in tempering the wall pressures generated as a result of grain swelling.5 8 . . While the foregoing analysis is an approximation. i' z == 10 m a:: a. o o 1 2 56 BOUNDARY LAYER DIMENSION Fig. -. _ . Coefficient of friction h-Head K 0 '"' Lateral pressure cae ffic:ient. Under aeration conditions Ko may approach unity. 19.1 w a:: (/) (f) => ui .

If. Use of Anti-dynamic Tubes to Control Silo Pressures In tall grain silos. the trernrnie tube divides the tall funnel-flow silo into IWO squat silos in series. This is nol the case. it docs so by expanding radially outward and upward. For example. a ty pical elasticity modules for wheat IS Ew = 2. The loading pattern will be more complex and the resulting pressure changes around the peri phery wi II not be uniform. This is an upper bound value since the modules Ew for bulk wheat is less than that given above. the elasticity of the wheat relative to steel needs to be taken into account. Load cells fi tted in the wall of the silo enabled these measurements to be taken. Reimbert's anti-dynamic tube. one daily expansion and contraction could produce a residual increase in wall pressure of 0. The work was mitiated in order to provide a simple and low cost sol ution to contrail ing the pressur-es in a number of badly cracked concrete gra in silos approxi rnately ten times the scale. When the silo contracts it does so by contracting radially inwards and downward. the actual process of expansion and contraction coupled with the effect of cyclical loadings on the grain and influence of moisture variations is an extremely complex problem to analyse. This is i IIustrated in the test resul ts of Figure 20. The effect j s to cause dynamic pressures to be generated. the overall changes from summer to winter can have a significant effect. by the usc of an anti-dynamic tube. as depicted in Figure 19. hypothetically. Since the temperature cycles on a daily basis. This can be II problem in the case of silos storing bulk materials over long periods of Lime without discharge. 20. This has the effect of significantly increasing the pressure in a similar way to that of increased moisture. This allows the grain to settle. In this way. variations in grain packing densities and variations in moisture throughout the bulk grain mass could give rise to non-uniform pressure distributions in eccentric loadings. an increase in wall pressure of 18 kPa would occur. Australia (Ooms and Roberts 1985. mass-flow of grain with flow along [he walls occurs over a substantial height of the silo above the effective transition.. Measurements of daily temperature fluctuations and pressures generated in walls of steel silos have indicated quite clearly a significant increase in pressures during the night period when temperatures wer-e lower than during the day. Temperature variations The variations methods use to analyse the wall pressure in steel silos generally assume a rigid silo. period indicates a cumulative increase in pressure or I I kPa or 25% of the maximum static wall pressure. the latter effect causing the friction force at the wall to switch from the initially upward direction to the downward direction. extrapolation of the data for a twomonth. me 1. these pressures being in the order of two to three times the static pressures generated after the silo is filled from the empty condition.. Apart from the daily temperature fluctuations. it is possible. ifany. When the silo expands due to temperature increase. 270 o Anti-Dynamic Tube Lowest Observed Effective Trans~ione _ 45 0 5 10 Mean Normal Pressure 1'5 (kPa) Fig. the cumulative effect of expansion and contraction over II period of days and weeks is significant. It must be noted that the foregoing discussion assumes uniform temperatures around the periphery of the silo. the combined effect of the progressive settling of the grain and the action of the reversed friction force at the walls could have II most significant influence on the structural safety of the silo. This will not occur in practice. As shown by Reimbert (see Thompson 1984).2 m diameter by 3 . placed centrally in a symmetrical si 10. m tall model nat 5 bottom test silo and a sample set of test results for the normal wall pressures. The design of the bottom ports and tu be sizing in relation to the silo outlet dimension are important in order to promote automatic choking of the lateral flow at the bottom until the tube empties. The modulus of elasticity of wheat varies significantly from one wheat type to another [ 16] . during contraction. In practice. ex tends almost the whole height and has a series of holes or ports to allow grain to enter the tube at various levels. Mode! silo srudtes usmg anti-dynamic tube. The arrangement ensures thaI at no time does the effective transition intersect the wa 115 of the si 10 and hence the pressures never exceed the values corresponding 10 me static or initial fining condition. Figure 20 shows. A variation of the Reimbert tube is the tremmie lube which has no holes in the walls and extends slightly less than half the height of me silo.0 I = I % Using [his ratio. However the quoted example indicates the possible order of magnitude of increase in wall pressure due to repeated expansion and contraction. in a 15 m diameter steel silo. weekly and monthly variations in atmospheric temperatures cause the shell of a silo to expand and contract with some settling of stored materia] during expansion with little recovery. For example. the wall pressures never exceed the static values. Daily.1 kPa_ While this is in itself insignificant. . a temperature increase of 20° could cause an expansion in the order of 3 m m. the silo contained wheat and was completely rigid so that the contraction of the silo were prevented as the temperature dropped. Research using this type of anti-dynamic tube was conducted at me University of Newcastle. this bei ng 41 % of the maxi mu m static wall pressure_ However. Ports in the bottom of the tube allow grain to flow laterally to the silo outlet. schematically. there will be differentia'] expansion owing to the position of the sun relative to the silo walls. Roberts 1988b). Based on the properti~s of indi vidual ~ains. as a result. No valves are necessary.07 x 10 kPa grves (EwfEs) = 0. to control the flow pattern so that funnel-flow always occurs without flow along the walls.Proceedings of the 61h International Working Conference on Stored-product Protection -Volume I The foregoing assumes symmetrically located air injection points and a uniform air distribution. In effect. Furthermore the analysis is greatly simplified. the effective transition occurs low down the silo walls. The top half of the silo discharges first followed by the bottom part once the level drops below the top of the tremmie tube and the tube empties. as well as decreasing on a longer-term basis.

. In terms of sl 10 loads. the solution of one problem which leads to an improvement in plant performance exposes other Taft mass-flow and funnel-flow bins The flow characteristics in mass-flow bins arc illustrated in Figure 22.21.I~~ 2 (a) Single Eccentric Outlet Fig. The silo quaking problem Pulsating Loads in Bins . (a) velocity profiles and pressure distribution. It is to be noted that the anti-dynamic tube described here is SUitable only for free flowing. 2 .Stage 3 Discharge ~~I ~ ~~~..'Silo Quaking' As is often the case. They should not be used for cohesive bulk solids. It should also be noted that the draw-down loads on antidynamic tubes can be quite high. expanded-flow and intermediate-flow bins. there is a minimum level Her which is required to enforce 1 .Proceedings a/the 6th International Working Conference The anti-dynamic tube described above has been successfully used to control the flow in silos having eccentric and multiple discharge points. and Craig and Roberts (! 994).Stage 2 Discharge. the walls of such si los are subject to significant bending stresses in addition to the hoop stresses. I994). a re-occurring problem is the phenomenon of 'silo quaking'. funnel-flow. 011 Stored-product Protection - Volume I problems which require further research and development. severe dynam ic loads and stresses can occur. The computation of these loads and design of location rods has been discussed in Ooms and Roberts (1985) and Roberts (1988b). 22. cohesionless bulk solids such as grain.Without the tube in place. Roberts (1993. a term used to describe pulsating loads which may be experienced during discharge If the frequency of the flow pulsations is in harmony with any of the natural frequencies of the structure. 271 . 3 .Stag:e 1 Dlscharqe. (b) variable density and dilation. Typical applications are shown in Figure 21 .For MessFlow erreb!e D'ensHy Diletion (Die remmetic) Canso) idati ng Pressure Fig. Mass-flow bin. As discussed in Free Rowing granular materials. Applications (b)Two Eccentric Outlets (c) Truck Loading Side Chute - of the ann-dynamic tube for eccentric discharge . The discussion that follows provides an overview of the 'silo quaking' problem as it relates to massflow. The problem of 'silo quaking'is discussed in Roberts et al (1991).

As the settling lime increased. the loads acting in the cylinder are examined.0 Ll rn 0. With H::> H cr. The amplitude increased relative to the average values at higher locations in the bin. It is quite clear that during undisturbed storage. the pulsations . The same effect was observed for the three HID ratios examined. the normal wall pressures and shear stresses at locations 5 and I 4" At location 5. 111e phenomena depicted in Figure 26 was observed at all locations.flow hopper is constructed of stainless steel. the locations at which the measurements are reported are 5. The velocity profile is further developed in the hopper as shown. Funnel-flow and expanded-flow bins During funnel-flow in bins of squat proportions.5 metre tall cylindrical section constructed of mild steel. illustrated in Figure 24. A series of tests were conducted for different stored heads in which the normal pressures and shear stresses at several locations were recorded during filling. with flow along the walls.2.9 and HID .0 D.2 metre diameter with a 3. Pulsating loads can also be experienced in intermediate-flow bins of the type shown in Figure 6. Three head heights of the wheat in the cylinder were examined: H=3. 10.0 are virtually the same.4lm 2.. 272 Figure 28 shows for HID. A similar action to that described above may occur in tall funnel-flow bins or silos where the effective transition intersects the wall in the lower region of the silo. Records of the undisturbed settl ing were taken over prolonged lime periods. For the purpose of the present discussion all load pulses. As a result. As the material flows. simultaneously. during undisturbed storage and during emptying . the flow in the cylinder is uniform or 'plug-like' over the cross-section. both normal pressure and shear stress. For instance. In the region of the transition. further dilation of the bulk solid occurs. the long rise Lime is no doubt associated with the Lime taken to establish the arched stress field in the hopper. Pulsating loads. are capable of measuring both normal pressure and wall shear stress. this effect continues on after thc bin IS filled. located as shown in Figure 25. As a resul t of the di lation. particularly if the slope angle of the Iransition is too steep.2. Asa result some reduction in the radial support given to the stationary material may occur. Pilot scale mass-flow bin handling wheat e Studies of pulsating flow ha ve been conducted in the laboratories of The University of Newcastle using a mass-flow test bin (Roberts et aL 1991). can 011 Stored-product Protection -Volume I occur in such bins. The bi n is illustrated in Figure 25. the stored mass approaches. asymptotically. dilation of the bu lie sol id occurs as it expands in the flow channel.7. The cycle then repeats.9 Pressures pu tsosions during seuling after filling A dynamic 'slip stick' effect was observed in both the normal pressure and shear records during filling and during settling after filling. The influence of a pulsating ftoweffect is depicted in the records for location 5. The central flow channel leads to more rapid flow in this region giving rise to pulsating flow due to the 'slip stick' nature of the friction generated at the boundary of the central and outer flow channels. the con ical mass. 13 and 14. Pressures during emptying Flow Zone (Dl1f1ted Condition) Stationary r"---I Material Slip as Pressure P ReI fixes n Fig_ 23. Funnel-flow bin. the flow being characterised by a 'ratchet' type effect. as for all other locations. Typically. Fourteen load cells. it is possible that the vertical supporting pressures decrease slightly reducing the support given to the plug of bulk solid in the cylinder. AI this location.75 D to 1. at location 14. By way of illustration. This shows that the settling characteristics for the two cases HID = 2. As the flow pressures generate in the hopper. which is near the hopper transition. this height ranges from approximately 0. its critical state consolidation condition by a pulse type settling action. The cycle is then repeated..9. 12. there is flow along the walls of a substantial mass of bulk solid above the effective transition. the dynamic wall pressure and shear stress have each taken approximately one minute to reach their maximum values.. the effect being most pronounced in the shear stress record where the pulses are most evident. as Figure 28b shows. It is 1.Proceedings of the 6th International Working Conference mass-flow in the hopper. it dilates leading LD variations in density from the static condition. with the pulses at each location occurring at the same time intervals. the pulse period also increased in an exponential manner as illustrated j n Figure 27.then the stationary mass may Slip momentarily causing the pressure in the flow channel to increase as a result of the 'squeezing' action rather like that of a collet used to clamp a rod -of steel in a lathe.9. By way of example Figure 26. say (9 ~·o). If the hopper is fairly steeply sloped. This causes the plug to drop momentarily giving rise to a load pulse. Referring to Figure 25. A dynamic 'slip stick' effect is depicted in both the normal pressure and shear records during filling. This is depicted pictorially in Figure 22b. where there is no flow along the walls. the flow starts. typical test results using wheat are presented. 24. Funnel Flow Fig. Expartded-flow bin. The bin is fitted with load cells which measure.44m 2. as depicted in Figure 23. Expanded-flow bins are commonly used to store bulk solids in large tonnages. even though the amplitude of the pulsations in the normal pressure and shear stress at location 5 is small compared with the average values. to converge due to the influence of the hopper and the velocity profile is no longer uniform. shows sample records for location 14 for the ratio HID = 2.

0 SECTlON 14 13 12 11 --2150 1·850 1550 --·1250 B 0 0 B---- SECTION A -100 -300 -500 o Fig. 26. Mass-How test bin. 25.. 1i III 0 Filling..- 8 4 0 '" "" '" ~ til III 2 ~ VI .. Normal pressure and shear stress records for wheat at location I4 of test bi [J 273 ..- s "" '" -.ling ---- _ Fig.Proceedings of the 61hlnternationai Working Conference on Stored-product Protection . -.Volume 1 0 ._+-01----Undisturbed Sett..

pi ane occurring at It height He.K A (b) Equation (13) may be used to estimate the peak dynamic pressures due to silo quaking. A u. ". 6.. While there may well be several shock planes.. HID = 1. as the friction changes from limiting or static 10 kinetic values.0. the case of a single shock.. F D is the total shock load and acts over the cross-sectional area A. it is assumed that Apwy 'tails off' to approach the flow pressure at the top surface in the cylinder. loads are imposed on a plane at some location in the cylinder wall. J+ w.oc A" -fiK% increment (13) Where the maximum sure is given by An ~""II'O- b. Robertsf 1993) developed It simplified theory to determine the magnitudes of the dynamic loads due to 'silo quaking' .bulk density. are quite pronounced...pwo in lateral wall pres(14) - f. say. the period of the pulsations approximately 5 seconds. As y ~ h. Pressures and shear stresses for wheal during emptying HID = 2. Shock loads 80 100 120 PULSE TIME (mins. (a) Results for location 5 (b) Results for location 14.. 6 4 2 0 0 20 40 fh . Urn] P '" perimeter of flow channel. The behaviour is illustrated in Figure 29. Load pulsations test bin. in the cy Iinder is considered as the worst possibility as far as loading is concerned. 27..9.J ::::I I=::: 60 HID ~ 2. F D is given by FD kd{:[I-e = -~K% are Jlf1~-h. m h = head. (20) . averaging and W s . there was no ev idence of any pulsations during discharge..9 HID ~ 2... The loading conditions are depicted in Figure 29. It is assumed that the amplitude of the shock..a.} (15) The two weight components W-pPRh (16) (17) Fig.0. 28..0 Dynamic loads in tall mass and funnel~lIow bins c.. m hs ~ surcharge head..2. After each shock occurs. m R = effective radius. but when the head was reduced D 2(1 + m) (19) were experienced for the case of to HID = 0.pwy is the additional waH pressure applied to the initial or static wall pressure Pni as indicated in Figure 29 (b).9. the bulk solid is considered to behave as a plug which imparts step loads due to 'slip-stick' motion as may arise from variations in density and in wall friction. kN p . the dynamic wave will decay with time according to the bulk elastic properties and damping characteristics of the consolidated bulk solid . That is. Above this plane. P P R hs where 'Y = P g = bulk specific weight. Thi s eon274 Similar dynamic effects Surch arg e head h =s m+2 H.Proceedings of the 6th Internat ional Working Conference on Stored-product P rotection ~ Vol urne I 12 10 8 Perimeter P = (2)I-m (RT (L)l-m (D)m+l (18) a:: w ED :iE ::::I Z W firms the conclusion that self excited pulsating flows in massflow bins which give rise 10 the 'silo quaking' phenomenon are associated with surcharge heads greater than a minimum value of.) Fig.")" = ..p.Pwy normal to the wall will decay exponentially with distance measured on each side of the shock plane..Dynamic load factor Effective radius R := Effect of surcharge head HID . kd . pressure 6. during undisturbed storage of wheat in (a) It is postulated that due to 'silo quaking' shock..

3. A..5 . seconds. It has been determined. this gives fi = 7..() + .Proceedings of the 6th International Working Conference on Stored-product Protection - Volume I T = Shock period 1-------_ -e(a) Dynamic Pressures A Pws Increment due to shock (amplitude) due to Shock T b.Ps. 37.actual surcharge head hs"" Effective surcharge head If the load F D is suddenly applied at each pulse period..p wy Initial Pressur.- p.)+ ~. as II (Figure 8) decreases. Assumi~ A. Under severe conditions.. then the natural frequencies will be fi = 7 Ih 225/h 37S/h . m = 1 for axi-symmetric How D ~ diameter or width of flow channel Hs .5.tf = 1. The nurn ber 0 f cycles of dynamic load due to 'silo quaking' duri ng each discharge is obtained from the equation Number of pulses per discharge cycle . Wa ve motion=brief discussion period Til ~ 0.0 for plane-flow L = length of flow channel. where m ~. Dynamic loads induced if] silo. . When the level drops to the critical height Her> h ~ 0 and the shock pulses virtually disappear. 10 m. (2]) where A. For instance if h . the •81 P stick' effect is Iikely i to be more pronounced..(21 a {J 0 +. h decreases and the amplitudes of the shock pressure normal to the wall. Time to discharge to critical height Her Pulse period T It is also noted from equation (17) that. Therefore.0.!:. the natural frequencies of the stored bul k solid will increase.) (J (23) . ru ·I------------------------------~t (b) Shock. 29.wave veloci ty such that A=~ (22) E .. respectively. as the level in the bin drops. that for a certain dry bulk solid material.1:.0. In IIIis case s«> 1.. also decrease. for example. Waves at Location z Fig. Apo and ll. Period of pulses As shown by Roberts (1993).. then k. ranges from 250 to 400 mfsecond. the pulse period may be estimated from the following relationship Ii" = ~~ i = 1. the period will increase as Effect of reducing head due to discharge The natural frequencies of the stored material are given by It is assumed that the location of the shock plane remains substantially at the critical height Her during discharge.. the fundamental 275 T =. such as when the moisture content of the bulk sol id is high. 22...5 . = 300mlSIXond..(I. (Hz) for i ~ 1.J:.3.133 damping increases.5. With low damping.elastic modulus of bulk solid P ~ bulk density The wave velocity for the bulk solid will depend on the properties of the bulk solid including the consolidation conditions and moisture content.5. Hz.

the variation of the expected pulse period T with the average velocity v for various values of Aey "" Dy is shown in Figure 30.D . which is a typical value. The following data apply: O-I. Squat funnel. kgls A = cross-sectional area of cylindrical section of bin F D . the pulse period is relatively insensitive to average velocity for velocities above 0. the actual pulse period ranges from 2 to 7.4 kPa In order to compare the predicted and measured results for the load cell locauon 14 uf Figure 25. Of .O. N to = ti me for motion of upper mass to be initiated dEy = expansion of consolidated mass in vertical direction due to dilation The parameter oey is difficult to determine precisely.4 m (assu med) h~ 1..5 Q-5 tlhR= F(OI4) =. the measured value of Apyo ses 1. Pulse period Assuming that t. The acceleration given by equation (24) will always be less than the gravitational acceleration. 15 .I m/second2. For the acceleration a . except in this case the flow pulsations are generated in the lower hopper section.01 0. It is dependent on properties of the bulk material including the degree of consolidation and the particle size distribution. since FO < WT.9 kPa which is in reasonably close agreement with the predicted value.€ ranges from 5 to 7 mm for the wheat. that equation (19) is a reasonable predictor of the pulse period.012 Pulse period versus velocity a = I mlsecond2 (25) (26) Hs-O Jl . ~ coefficient of internal friction 0 f bul k sol id Normally tan <1'(2' f1.0. Equations (15) to (20) apply in this case with the following substitutions: -.4 rnrn Dy. reference is made to the experimental silo investigation described in Section 3..008 Ve. respectively. using equations (19) to (23).discharge rate. intermediate and expanded flow bins Example=Pilotscale wheat silo of Figure 25 To illustrate the application of the theories presented.2 '" 1. Q) o ·c o 3 2 --t·x···-o-y'·.0. This is depicted in Figure 31 .'.dynamic force defined by equation (10) lVT ~ total weight of upper mass. 2' sin 0 (26) 276 .5 seconds and T = 4.577 K".85 tlm3 Amplitude of Normal Pressure (27) The variables given by equations (23) to (27) are defined as = acceleration of upper mass during pulse motion. From Figure 28. rn "5 ID o o Fig. mlsecond2 v = average velocity of bulk solid in the cylinder during discharge. equation (13) is used noting that y .tan 30° .9 kN Equation (15):FD = 9.··' -mth- a.locity (m/s) 0. mlsecond Q .pwy = 2. On this basis it is concluded.Proceedings of the 6th International Working Conference Of! Stored-product Protection -Volume I 6 .2m H-3.5 seconds averaging around 5 seconds.1.9 seconds. where (24) 0. This gives t.3.002 0. Here the analysis is similar to that presented for tall bins.3 rn p -0.. From Figure 28.0 m. o Dy = 5 mm Dy -". T =.2.4 H cr ."" 3rnn) Dy = 2 mrjl 4 "'C n.006 0. As indicated. the estimated pulse periods are.width or diameter of flow channel and f1.06.44m The effecti ve dynarn ie mass of bulk solid in the central flow channel of the hopper section is the total mass in this section less the frictional support due to shear at the boundaries between the central and outer flow channels. 3(1.3. This indicates the velocity dependency characteristic of the pulse period.1._ ~ +_----+-----L-----~--~ . The pulsating mass is as indicated in Figure II.6 kN Equation (14):APwo =.004 0.004 a Equation (16):W = 18.

As each silo fills and empties.pwo = 56 kPa Equation (13) has been used to compute the locus of the shock pressure amplitude. As a result. Vibration of structure h s Shear S h Fig. the internal diameter is 21. while the four outer feeders were not opera Led. hence.. Roberts 1993. The hopper splits into three outlets. the pressure profiles will vary around the periphery of the bin. The maximum discharge fate is 2700 t/hour through vibratory feeders.12. In one series of tests the three feeders along the centre line parallel with the reclaimconveyor were operated. 1994). . there will be dynamic coupling between adjacent silos which could impose significant loads on the concrete connecting these si los. which. By way of i\lustration. and 05 = where <h . two case study examples of pulsating !low in bins are presented.3 m. the observations could be equally applicable to grain.static angle of internal friction angle of internal friction Silo quaking problem-two effective A critical factor in the operation of the silos is the in!luence of the dyna mic characteristics of the overall structure. while complex. Also shown are the static and .flow pressures. 12700 t. six around an outer pitch circle and one located centrally. reinforced concrete coal silos.6. in Figure: 33. severe shock loads were observed at approximately 3 second intervals during discharge. which are based on . The modes of vibration. the lauer induced by non-symmetrical loadings of coal in the silos as well as variations in ground stiffness in the zone of the supporting piles. the half angle a. stand 58 m above the ground. d iffcrcnt frcqucnc ics for adjacent silos. After several years of use. this being the angle with respect to the vertical. the conditions were right for severe 'silo quaking' to occur.57 wnempty The natural frequency of the full silo is only 057 of the empty silo. The natural frequency for the fundamental mode is given by cu. Multi-outlet bins Silo-quaking problems have been known to occur in bins with multiple outlets. Conflrmauon of the mechanism of silo quaking was obtained in field trials conducted on the bin. When the level in the bin had dropped to approximately half the height. Computation of shock lauds cuJull = 0. 277 The shock plane is assumed to be located at a he ight of 26 m above the base of the hopper. particularly at the junction of adjacent silos. Adjacent silos are connected by concrete along the vertical lines of intersection.. which are shown schematically in Figure 32. Not ing that the silos are supported on columns on a base. the effective transition was well down towards the bottom of the bin walls and the critical head Hm was of the same order as the bin diameter and greater than D F. T = 2. the following values are calculated: W=77626 kN Equation (15):FO = 50 130 kN Equation (14 ):.=& Noting the variation in mass. being 40°. Tall silos with chisel-shaped hoppers Severe silo quaking problems were experienced in an installation of three. the shock loading being most severe when the silos were sabstantially full. The silos. is supported on piles. The discharge rate from each feeder was in the order of 300 tfhou r. Figure 13 applies to the side of the bin w here the chisel-shaped hopper has its highest intersection point with the cylinder wall. While the various stiffness values of the structures are not known. In view of the significant variation in the silo mass from the full to the empty condition. with sections of COncrete being dislodged.6. a simplified dynamic model of the silos is shown schematically in Figure 34. the shock loads had diminished significantly. the height above the base is 52 m. While the examples refer to coal bins. The reported period of the shocks was 3 -5 seconds. are = in general agreement with the approximate values of 3-5 seconds as reported for the actual silo. The hopper geometries provide for reliable flow permitting complete discharge of the bin contents.. there will be different mass contents and. The head of coal above the shock plane is h = 22 m. When the bin was full or near full. consider the large coal bin shown in Figure 35.4 seconds These val ues. in tum. There will be vertical and lateral stiffness due to the columnsarid piles as well as vertical and lateral stiffness due 10 the concrete connecting adjacent silos. plane-flow hoppers lined with carbon steel plate. the natural silo will change as follows: frequency of each case study examples Field observations confirm the foregoing description of the mechan isms of 'silo quaking' (Roberts et al. The bin has seven outlets. it is possible that the load poise frequency of3 to 5 seconds could excite one of the silo modes. In v iew of the curved transi lion of the chisel-shaped hopper wi th the cylinder.4 mand the wall thickness is 0. Substantial flow occurred along the walls. the results being illustrated. Using the procedures presented in Section 4. there is a significant variation in the natural frequencies . Coal was discharged by means of seven vibratory feeders onto a centrally located conveyor belt. and since the reel ai m hoppers were at a critical slope for mass and funnel-fl ow as de term ined by flow property tests. 199 I.Proceedings a/the 6th International Working Conference on Stored-product Protection - Volume I Estimation a/pulse period The appl ication of equations (19 ) to (23) gi ve the follow ing predicted values of the pulse period: For discharge rate Q = 2700 t/hour. Effective mass in pulsating flow. For illustration purposes.31. 5 mm for the coal. The silos have chisel-shaped. With all the outlets operating. severe damage started to occur. the lower hoppers. being lined with stainless steel. would involve a combination of vertical and sideways swaying motion.

~ . ---~-- i ----total 40 ~ .low press. Tall coal silos 50~~~~~~~~~~~~ j-I~itial j-e-F.Proceedings of the 61h International Working Conference on Stored-product Protection -Volume I C hise I-S haped Hopper Plan of Silos (Reduced Scale) Fig. '1' 10 o Fig. 33. . i = ' . 30 I 20 ···'T·····orSrload.. 9ress. 32. Oyn.ie··~···"·····.. 278 .. Pressure distributions 50 100 150 200 250 (kPa) 300 350 Normal Wall Pressure for bin.: ESlimaledi Location : ····.ri .j·'·" i \...·········t·········l .. .

.king'.. 34. This gave rise to the triangular prism shaped dead region in the central region. In a second set of trials.. in this case the stationary material in the bin formed a conical shape. I) was operated.. that is H m< Her (=DF) or H m« D. If the frequencies of the flow pulsations are ill harmony with any of the natural frequencies of the structure. The paper has focused on silo design. the load pulsations were barely perceptible. Multi-outlet coal bin. ~ 1---- D----I ... grain quality. while the four outer feeders were operated.Proceedings of the 6th International Working Conference on Stored-product the strains measured that H m< H cr- Protection -Volume I so when the flow pattern was controlled ~ Swaying Mode ... 279 . the three central feeders were left stationary. increased moisture con tent which produces grain swell ing can cause substantial increases in silo wall loads. With the emphasis on grain conditioning procedures. the motion down the walls was greatly restricted and. Particular mention has been made of the use of ami-dynamic lubes to control the flow pattern and. severe dynamic loads and stresses can occur. Vertical Mod:e Fig... hence. When the bin was full (or near full). asa result. ~I Ph:m Feeder. Simp1i lied dynam ic mode lor silos. A re-occurring problem in bins and silos is the phenomenon of 'silo qua.. Dynamic strain measurements were made using strain gauges mounted on selected support columns.. in the case of mul ti -outlet si los and si los with eccentric discharge the loadmgs are much more complex leading to bending and buckling stresses in the walls. 35. with the effective transition occurring well up the bin walls. of Bin (Reduced Scale) Fig. While the loading patterns in the case of symmetrical si Ios are well defined. such as aeration. Daily and seasonal fluctuations in temperature which cause expansion and contraction of silo walls. control the magnitude of the wall pressures generated during discharge.. the measured dynam ic strains with H m » Her were in the order of 4 times . Methods of controlling loads in tall. Under these conditi ons.The load pulsations were just as severe in this Case as WaS the case with all feeders operating. In this respect.. a term used to describe pulsating loads which may be experienced during discharge. particularly in the case of steel silos. Concluding Remarks This paper has presented an overv lew of some recent developments In the technology of bulk solids handling as it relates to the grain industry. Single and multi-outlet si los have been discussed. This induced funnel-flow in a wedge-shaped pattern as indicated in Figure 35. can also cause substantial silo wall pressure fluctuations leading to increased loads during periods of contraction and structural fatigue.. it is important to also take note of the influence of any such grain conditioning on the bulk storage and flow properties of the grain. to control. The various mechanisms of 'silo quaking' have been outlined and methodologies to determine the magnitude and distributions of the dynamic loads have been reviewed. The same was true when only the central feeder (Fdr.. with substantial mass-flow along the walls . emphasising the need to understand the relevant properties of the grain and how these relate to silo geometry and discharge flow patterns to generate the load patterns exhibited in the silo walls.

The Institution of Mechanical Engineers. IIow and handl i ng. A. then the future efficiency of bulk materials handling as a key function industry and agriculture is assured. Oslo. N ewcastle. The Uni versity of Newcastle Research Associates (TUNRA). Australian Standard AS3774-1990. Some aspects of grain silo wall pressure research-influence of moisture content on loads generated and control of pressures in tall multi-outlet silos. Bulletin 108. The Univer-sity of Newcastle. References Arnold. The Netherlands. Handling and Transportation.W . Jenike.1. Roberts. It is gratifying to acknowledge the increasing industrial awareness and acceptance throughout the world and particularly in Australia of modem bulk materials handling testing and plant design procedures. Enschede. May 11-24.M. A. 0. prov ided Un iversi ties and research establishments are encouraged to pursue this research.• 1966. 1. USA Jenike. 50. significant advances have been made in research and development associated with bulk handling systems for granular materials. USA. A. DOC-WHIIThesis. The Institution of Engineers. The University of Wollongong.J. Van Nostrand. A. W. A. 5. 280 . Standards Association of Australia (SAA) 1990. Engineering Experiment Station. 151-170. In: Third International Conference on Bulk Materials Storage. and Roberts. in recent years. W.. In: Third International Conference on Bulk Materials Storage. Static and dynamic analysis of the flow of bulk materials through silo. and Roberts. Silo quaking.229-236. Arnold. Newcastle. 1991. 48-52. of Utah. Jenike. 7. 65.W. Modem concepts in the design and eng ineering of bulk solids handling systems.c. Storage of particulate solids. In: Proceedings Conference. Australia. Powder Advisory Centre. Australia. The Inst itu [ion of Engineers. Gravity flow of bulk solids. The reduction and control of flow pressures in cracked grain silos. 1989.K. Ooi. The Unlv. 9. Powder Tech . Thesis. Elastic and plastic predictions of storage pressures in conical hoppers. A.W. P. A pulsating load problem during discharge in bins and silos. 1993. Beni nk. 203-207. U. Proceedings of the 2nd Reliable Flow on Particulate Solids Symposium 983-1004. J. 1990. Rombach. Proceedings Bulk 2000.W and Ooms. USA. Australia. Roberts. I. Load on bulk solids containers Thomson P.Proceedings of the 6th International Working Conference on Stored-product Protection -Volume I It is quite evident that. 1978. Handbook on Powder Science and Technology. W. Civil Engineers. 1988b.. Oorns. P. Roberts.W. W.. 1961. A. E. 1989. A. Bulk solids: storage. A. A Theory of flow of particulate solids in converging and diverging channels based on a conical yield function. Australia. A. Journal of Agricultural Engineering Research.• Scan. Bulletin 123. G. II. of Utah. R. Roberts.H. Numerical simulation of filling and discharging processes in silos. Jenkyn. Flow and stress analysis of cohesionless bu Ik materials in silos related to codes. Norway. Calculation of material pressures for the design of silos.988a. M.L 1989. Chicago.12. and Eibl.. Wu.K . 2. Bulk Solids Handling. Handling and Transportation. 1964 Storage and Flow of Solids. Mechanics of self excited dynamic loads in bins and silos. The University of Twente.M. Proceedings Institution. Stress distributions in loaded wheal grains. 1.C. and Wiehe. and Roberts. Wall loads in large steel and concrete bins and silos due to eccentric and other factors. Roberts.T. McLean. A. 1985. Australia. and Rotter. I. S. 1984. In: Proceedings 2nd International Conference on the Design of Silo for Strength and Flow.Y. Y.W. M. TUNRA Bu Ik Solids Research. It is also noted that areas for ongoing research are continu ing to be idcn tified. 1009-1016. The Univ.W 1982. A. (5). Engineering Experiment Station. 1983.G. U. PhD.

Are you sure?

This action might not be possible to undo. Are you sure you want to continue?

We've moved you to where you read on your other device.

Get the full title to continue

Get the full title to continue reading from where you left off, or restart the preview.

scribd