You are on page 1of 52
AWRA TECHNICAL NOTE 14 _ December 1984 DESIGN AND CONSTRUCTION OF WELDED STEEL BINS by the Australian Welding Research As 8 alfred Sr eet, Milsons Point, 2061, Australia. Tele phone 9223711 . ractical andtheoretical guidance eee ae a ee aon of eolded Soe De «© fntended to asus, through is conten! and the feferences listed, design engineers, detallers, fabre sere Seton and plan operstors ‘¢ has been prepared from the latest information aval able from the iterature and from the industry @ has been prepared by AWRA Panel (6) “Structures” consisting of Mr BE Gorene {Convenor} ~ MacDonald Wagner Ply Lid Prof. A Abel = University of Sydney Dri Bennetts - The Broken Hill Proprietary Go Lid MrJCottam ——- The Broken Hill Proprietary Co Ltd MrGADay ~ Australian Institute of Steel Construction MrGFletcher ~~ New Zealand Heavy Engineering Research Association MrM Frost-Drury - Department of Housing & Construction /-~.MrB I Grishenko ~ Trensfield Ply Lt rot. H B Harrison - University of Sydney OrL Pham = CSIRO Division of Building Research DrMGLay __~ Australian Road Research Board Mr M McCormick - Newstee! Pty Ltd DrD Mansell - University of Melbourne MrA Marshall - Queensland Electricity Generating Board MrDH Nixon © ~Minenco Pty Ltd MrRJ Payne - Maunsell & Partners Pty Ltd Mr RC Sturrock - Maunsell & Partners Pty Lid OrG Sved = University of Adelaide DrPSwannell — ~ University of Queensland MrR Wedgwood - Department of Main Roads (NSW) Acknowledgement is made of the special assistance ven by Dr PC Arnold and Dr A G McLean in revie ction 1 which is largely based on their recearch wor This Technical Note will be revised in due course and comments would be most welcome, No extracts from this publication may be reprinted . without the Association’s permission, Editor: IW McGeachie Welding Research Engineer - AWRA National Library of Australia card number and (SBN 0 909539 33 2 CONTENTS Section Scope 1 Functional Design of Bins 2 Loads 8 Structural Forms 4 Structural Analysis 5 Design of Welded Joints 6 Detailing, Fabrication and Erection 7 Surface Treatment 8 Inspection of Bins in Service ‘Appendix A — List of References ‘Appendix B — Nomenclature index e SCOPE 1m recognition of the widely expressed need for quid- lange on the design and fabrication of welded steel bulk Solids containers AWRA Panel € undertook the task to prepare this Technical Note. Please note that the eral term “bins” will be used throughout this Tech: Bical Note to include bins, silos, bunkers, tanks and hoppers. ‘The main objectives in preparing the Technical Note are set out as follows’ © to gather together the most essential relevant mat- tial on: ‘* functional design for reliable flow ‘= determination of wall loads ‘= stress analysis of the containers © detailed design of the welded joints fabrication of the container structures # to promote the principle of safe and economical designby saing proven advancedmethodscombined Wwith cost-effective detailing @ toalertall concerned tocommon pitfallsin the design of bins and to suggest ways of avoiding problems in service Considering the large variety of applications, forms of ‘construction and storage capacities, it has been nec- essary to limit the scope 50 that the more usual types of bins Gan be more fully described. Included are bins Constructed from rolled steel plates, circular or rectang- tlarin plan, ring beam supported or elevated and having storage capacities not exceeding 3000 tonnes. Delailed objectives for each chapter of the Technical Note are as follows. AWRA TECHNICAL NOTE 14 ‘The functional design given in Section 1 almsat abetter understanding of the influence of the form of the con- ‘ainer on the flow of the buik solid. Precise instructions for the determination of optimal hopper and dis charge outlet sizes being of a very specialised nature are not included. However, the reader is directed to selected references on the subject. The determination of the wall loads in Section 2 is outlined against the backgroundof the various methods in common use in the design office. This chapter sum arises the lateral pressure formulations of the current Codes and other accepted methods, The reader is ref- grred {0 a number of useful current texts fr further study. Various structural forms of bins are describedin Section 3 forthe benefit of the novice designer and the student. Since the selection of the structural form precedes the Angle of Surface foughness Class Gy {imternal Frietion Vol.(a) Loads(b) 1 2 3 Wheat 75. 90+ ««80~=«(04BS20 BDA Corn 70 80 80 048 020 025 060 17 Barley 79 80 © 28_—S—«SB 20 HAD Four (wheat) 60 70 = 28D 205 83.8 Sugar 80 8S 29Ss«SZss«CKSCAHHOAD Quartz sand, dry 140 16028210. dO KO 1 Concrete aggregate,camm 160 180 31 048 035 045 060 12 Lime stone powder 110 ©««180,=S«80s—«iD02HsD OHS 1S Fly ash ao 1152558 OS | 12 Cement clinkers 480 180=83.—ssikCHHD HSA Cement 130 160.=«-28_s—siCOSD 0B OD OSD 18 Alumina 400 120«S« 27S SKS 18 Coal, anthracite 85 100 «90S Coal, brown 70 80 838 045 028 080 055 14 Phosphates 180 200 27 058 030 040 050 13 Petroleum coke 70 80 38 040 040 050 080 13 tron ore, haematite 220310 260370 38 040 040 050 80 13 Notes: (a) For volume calculation (E) Sprvalloed calculation fe} Allvaiues may vary by = 10% AWRA TECHNICAL NOTE 14 ries have been developed for the deter- Thinstion of loss due to the Dulk solids inside the bin The choice of the theory depends on the geometrical Configuration of the bins and on the stafe of the bulc Solid Table 7 gives the details of the subeivision of theories for the determination of the wall loads, Table7 Loads on Walls Shape of Bin State Theory Lowbins* Initial Coulomb retaining wall theory or tanks (Considers friction on walls) HID <1.0 Flow Usually only aslight joadincrease where pneumatic discharge is, used Bins with — Intermediate between the two H/D between adjining ratios, 1.01020 Initial Janssen, or similar static theory having (Ref 5) H/D> 2.0 Flow (i) Mass flow bins: Minimum Strain Energy (Ret 85, 190, 200, 228) (ii) Funnel flow bins: Factored Janssen (Ref 8) Minimum Strain Ener (Ref 85, 190, 200, 225) Bunkers _ Initial Jenike (Ref 85, 190, 200) Walker (Ref 335) Flow Jenike (ref 85, 190, 200) Walters/Clague (Ref 105, 340) * Squatbing are characterisedby relatively small ratios of wall eight, H, to the diameter D, This ratigisin the vicinity of 0.8 101.2 (eee also Reimbert, reference 260) Table 8 Loads in Funnel Flow Bins 2.3.2 Lateral Loads on Walls — Funnel Fi Mode of Operation oe The widely used Silo Codes, ACI 313-77 and DIN 1055- 4987 (with Supplement 1977) are only applicable tonal bins with flat or hoppered bottoms operating in the funnelflow mode. Both silo codes exclude Rely Se hesive bulk solids. The disparity belwoen ike moc: nilludes of lateral loads given by Various codes is quits noticeable as shown in Fig. 3. Inthe draft revision of DIN 1055, the lateral loads during flow are determined from the initial lateral foads multe plied by @ dynamic factor as in the AGI and McLean Methods. The dynamic factor in this revision depends on ‘several variables andincludes the effect of the rigidity of the bin wall, Steel bins usually have much thinner walls than reinforced concrete ones, and this is beneficial in that the lateral wall pressures are slightly reduced in stool bins. ‘Where a more rigorous analysis of loads is warranted, methods such as the following should be used: {a)_Jenike's minimum strain energy method of Ref 190. (b) Single boundstrain energy procedure developed by McLean et al (Ref 225). Table 9 gives design values of the dynamic factor, ¢ for steel bins, using the information from Ret 225, Table9 Value of the dynamic factor, c, ‘Aspect ratio, H/D 1.0 2.0 4.0 Depth ratio Value of 8 zd all 30° 40° | 30" 40" o=10 1.0 15 15 | 15 15 11 12 23 22°| 23 23 20 - 40 10 | 23 22 3.0 - + | 23 24 40 - et 1010 The values of the hydraulic radius, Rr are calculated as follows: ()) Ry = 0.250 for circular bin cross section {ii) Ry = 0.28 for square bins Gil) By a for long sides of rectangular bins of width a. Method of Calculation Ste . DIN 1055-67 ACI 313-77 McLean et al. 1. Determine: e 8. As specified From test From'test nM : * : K 05, Initial K=(1- sin/(1— sing) Ky= 0.4, 116 > 25° K, = 1.9, flow AS ACI, if 8 <25" 2. {nitial lateral load, pai Phi = WeR Cult Pri WeRnCo/a Pri WePKCelit ce (1=@-#20) * . Zo= Rnluk . * ‘8. Flow condition, Put= WeFinGe Pri Pr Par Pri, lateral load, Pay Using K, value where ovaries between where cis obtained 1.35 and 1.85 from graphs, Ret. 225 A comparison of the results obtained from the three methods given in Table 8 is shown in Fig. 3 for a grain bin. /8 Hy DIN 1085 Initial Loads DIN 3055 Janssen, ACI 32m, © 10 20 30 40 60 60 Lataral Load py —kN/m® Fig. 3 Comparison of Lateral Loads calculated by various methods 2.3.3 Lateral Wall Loads in Mass Flow Mode of Operation ‘The lateral loads on the upperpartof the bin operating in mass flow mode can be calculated by one of the folk ‘owing methods: {@) Jenike's minimum strain energy theory (Ref 190). {0) Single bound strain energy procedure developedby McLean et al. (Ref 225). {c) Simplified cynamic factor procedure by McLean et al. (Ref 225). (8) Procedures given in Ref 85. Because of the considerable computational effort re- quired, and absence of feel for the correctness of the intermediate results with methods (a), (b), and (d}, itis preferable to adopt method (6) fora preliminary apprais- BI of the loads. In essence, method (c) allows the flow loads to be calculated from: Pri = OmPi where cp is a dynamic coefficient obtained from graphs in Ref 225, and py is the initial lateral load obtained from Jansser’s formu. ‘Some typical values of cm are given below: Upper part down to 20: cnt Stor Hy targerthan2.0 S12 for W/O fess than 20, Lower part, =D to Thehopper transition: — H/D=2 c,=1.9 for 540" ‘M17 1019 for d=30" HyD=4 cqpn2.t 1022 Rg fae Vie Envelope curve of + Flow Lead Ba) A Pri Ps localised “owen Pass Peak flow load feultentoae) Initial normal logs [relevant for dester) Fig. 4 Lateral Loads under conditions of Mass Flow AWRA TECHNICAL NOTE 14, ‘Atypical diagram showing the distribution of flow loads s shown in Fig. 4. infuence of the circumferential wall stiffness has now been recognised. The recent draft of DIN 1055 Part 6 allows reduction of wall loads for thin-walled bins, 2.3.4 Loads on Bin Bottoms and Hoppers bin bottom may be flat or in the form of ahopper. Th method ot calculation of the normalloadg acting on the Surface of the bin bottom or hopper depencs on a umber of factors as sel outin Table 10, Table 10 Loads on Silo Bottoms Type of Mode of Governing Method of Bottom Operation Load State Calculation FlatBottom Funnel initial, «== ACI319-77, Eq 18. flow only Hopper Bottom (@) Hopperhalt Masstlow Upper part: McLean (Ret 225), ‘angle a less Flow Jerike (Ret 190) than cftical Lower part: Arnold etal (Ret 85) Initial Walters ef 40) Clague (Ref 105) Mclean (Ref 225) AGIS15-77 (Ref 8) BIN 1055 (Ref 55 & 86). Arnold ot al (Ret 85) ‘Jeniixe etal Ret 180) Note: * Criticalangie for mass flow canbe determinedtrom Ref 85 (©) Hopperhalf Funnel Initial ‘angie a: larger flow fran erica Loads on flat bottoms can be higher than estimated by ganssen’s equation pi — Pri/Ks, because of the dynamic behaviour of the bulk’ solids during the flow, especially with bulk solids prone toarching or bridging over the flow channel. The flat bottoms should be designed for: Ba = CuPyi/Kj, but not more than waz. where pyis the initial lateral load at the bottom of the bin, and’c, is the dynamic factor which ranges between 1.2’ and 1.7, as given in Table 6. ‘The hoppers designed for funnel flow receive the highest loading during the initial, stage (filling). The normal wall icad can be evaluated from: a= Pycostart Pysinta and the shear: P= 0.5pq, approximately. (see Fig. 4). Mass flow hoppers should be considered as follows: {@) Inthelowerzone, loads are determined as forfunnel flow hoppers. (b) In the upper zone having a height of 0.30 D, the normalloadson the walls can be many timesas high ag the initial loads because of the “switch” effect Which causes a radial stress field of high intensity. ‘The peak value of the normal wall load at the trans: ition, Par follows from Ref. 226: Prt= Cn Where © is a coefficient which can be read from epi in Hel, 225, and po, ts 8 given, above Re Ratnel slow hopper. The i istratea in Fg. Loads on the bottoms of tanks or shallow bins having HiD < 1.2 may be caleulated using the formulae w.H. It should be noted that: © Bin vibrators and flow promoting devices can drast- ically increase bin loads. © Special low-riction wall liners will also increase wall Toads. ‘© Impact loads due to falling material should be spec- lally considered. ‘AWM TECHNICAL NOTE 14 2.3.5 Vertical Wall Friction Loads Vertical wall friction forces must be correctly estimated for design of the walls of steel bins and silos. Vertical axial forces often govern the design of the walls of steel bins and silos since their relatively thin shells are very susceptible to buckling. Table 11 gives useful guidance for the calculation of the vertical wall friction forces. Table 11 Vertical Wall Friction Loads Mode of Loading Critical Operation State "Method Parameter Funnel flow Initial Not critical = Flow — ACI318-77, min 6 ganssen method — max. Equation 21 Mass Flow Initial Not critical = Flow —Jenike, (Ref 190) mind McLean(Ref225) max Pa Wy Py(2ZCe) Where the values of the coefficients are givenin Section 2.3.2, Pressure Distribution [A= Funnel Flow Channel B= Siationery Mass Fig. 5 Eccentric Discharge 110 2.3.6 Effect of the Eccentric Discharge/ Eccentric Filling : Bins with eccentric discharge openings ara subi ron-uniorm lateral loads. A method of estmnatifg ns loads custo hi affects ven in GIN 1988 Suppl ta7P ee also lenixe (195), (100), Wood (362), Johnston (202). ; aia Eccentric flow may be caused by — © Structural asymmetry. ‘© Outlet feeder operating unevenly. © Eccentric filing. The effects of eccentric low are manifested in the re~ distribution of pressures around the circumference as shown in Fig, 8. Eocentric filing also can Increase the wall loads in the bin. 2.4 Effect of Vibration Vibration may be caused by flow promotion devices, vibrating screens etc. The general effect is to increase density of stored material. impact caused by the falling material can lead to damage of the hoppers, unless measures are taken in the design. The main effect of Vibration is to reduce the wall friction coefficient, and thus increase the lateral loads, 2.5 Effect of Wall Liners Low friction wall liners are used to promote flow in the. mass mode. Their low value of wall friction coefficient results in an increase in wall loads. The possibility of deterioration of the liner surface should be considered In the calculation of the wall friction loads (refer Section 5}. 2.6 Loads due to Gas Pressure or Suction Internal pressure or suction due to alr or gas pressure differentials should be carefully considered in the design of steel bins, Suction created by the dust control fans may be large enough to buckle the relatively thin upper bin strakes, Suction due to discharge should also be considered for powdery bulk solids. Internal gas pressure generated by pneumatic dis- charge and by homogenisation should be considered in the design of roofs and walls — ‘Typical values: © Airslides Max pressure at silo bottom is equal to lateral wall pressure plus compressed air pressure, ‘The pressure is the highest at the bin bottom, and ceases to affect the bin at h = 1.3p,/v, where Py ‘equals the induced air pressure. © Homogenising Maxpressure at bottom should be taken as equal to Ws 2.7 Effect of Rapid Filling Rates denssen values of lateral pressure may be exceeded Buring ling ifthe buiksolid's relatively impermeableto fir. Powders such as cement are typical examples Of such materials. Reference 85 gives some indication of the pressure increases involved. 2.8 Air Pressure due to Pneumatic Discharge Air pressure due to pneumatic discharge affects the part of the walls nearest to the location of the air slides. The ait pressure should be added to the filling pressure, rgdually reducing from the full pressure to zero at @ eighth = 7-Spa/w, 7 2.9 Changes in Moisture and Temperature Cereals may swell when the moisture content increases and additional lateral pressure can arise. Special tests are required. The walls of stee! silos and bins may be affected by the temperature differentials between the wails and the bulk solid. A Sudden reduction in temperature of the was (rosty mornings) produces reduction in slo dar meter, leading to additional pressure on the walls, This presstreie highest when tho bulk sold has relatively ge buIK modulus, and when the temperature oycle 18 repeated frequently. Certain bulk solids may increase their temperature during storage and expand, This also gives additional pressure on the walls and bending stresses at the Support ring require special attention. 2.10 Equipment Loads It Is quite common to use steel silos for support of mechanical handling equipment. Conveyor belts, dust- control equipment, electrical switchboards and pent houses for housing of such equipmnent can substant- illyincreasethevertical and horizontl loads onthe sto structure. 2.11 Wind Loads The distribution of wind loads aroundthe circumference of cylindrical free standing silos is shown in Fig. 6. Table 12 gives the pressure coefficients to be applied to the stagnation pressure q,- Plddd Pressure Suction {a) Distribution of pressure around circumference of a cylinder [ol t ‘a 4 |! | | (b) Assumed height of bin for pressure distribution calculations Fig. 6 Pressure Distribution AWRA TECHNICAL NOTE 14 Table 12 Pressure Distribution around Cylindrical Structures Position on Pressure costficient Position on 884 ficient Cpe Surface: r Siinreisctone” | Surface: Smooth degrees | n/o=10 | 025 | wo=10 | nO>25 ° +10 | +10 | +10 | +10 10 +09 | +09 | +09 +08 20, +07 +07 | 407 +07 30 +04 +04 | 4035 | +035 40 ° ° ° 80 -05 | -04 05 60 -o95 | -o8 1.05 70 125 | =44 -128 80 ~12 | 108 -13 90 =10 085 =12 100 -08 085 08s 120 -08 035 04 140 -04 | -039 0.25 160 ~o4 | -03 0.25 180 -04 | 08 025 Source: NZS 4203:1976 Overall drag force on a bins given by: F= A.Coae where Gp is the drag coefficient, the value of which should be taken 2s given in the following table. Cross Section Ratio H/D 25 70 ro) Drag Coefficient oO ar Os Oo 218 08 eo) 4 1a@~40 20 14 14 The length H should be the height of a vertical cylinder except that for elevated bins where there isa free flow of air about both ends H shall be taken as half the length ‘when calculating H/D. In the calculation of the load on the periphery of the pliner, the value of the coefficient of internal pressure Cishallbe taken ntoaccount For operrendedcylinders Chi shall be taken as: ‘0.8 where H/D is not less than 0.3, 05 where H/D is less than 0.3 \Wind uplift on the low pitched roofs of circular bins can be'as high as 1.5 a9 at the windward edge. ‘The resultant uplift force on such roofs should be Pyp= 0.80-8,.4, acting at 0.150 eccentricity in the windward direction. When the bin is elevated from the ground and the Support structure is not enclosed, the table should be read at H/D values reduced to 50%. a Wind uplift load on bins roots should be determined in gccordance with AS 1170; Part 2 for rectangular bins. ‘The Danish loading code specifies a local uplift co- efficient ot 1.8 for conical roofs, n AWRA TECHNICAL NOTE 14 2.12 Explosion Hazard Certain combustible bulk solids such as grain and coal vith dust emission are fiable to explosion. The ways in Which the explosion hazard may be minimised are: 1. Design the dust contra! system for high volume idling, 2. Roof plates to be only tacked. & Explosion hatches to be incorporated. These hatches are attached in such a way that the fixing will break when a certain overpressure occurs. 2.13 Earthquake Loads Earthquake loads should be considered for bin location in the earthquake zones 1 & 2. AS 2121 gives the method of determining the forces on the substructure but does not include provisions for bins. Earthquake forcesaftect both the substructure andthe bin walls and hoppers. Generally, the load is assumed to be carried partly by the bulk solid and partly by the wall. ‘More guidance on earthquake loads can be found in Ref. 322 and in API 650, Appendix E. 2.14 Foundation Settlement Loads Loads due to the differential settlement of the found: ations can be significant ang in some cases a 100% load increase can be experienced, eg. settlement of one column out of four columns. itshould be realised thatthe bin walls generally have little capacity to deform in sympathy with the foundations. Where doubt exists about the competence of the foundations, itis recom ‘mended that the column reactions be increased 30%, 2.15 Feeder Loads Loads on discharge conveyors or feeders are ata maximum during the initial condition. Consult Ref. 258, for determination of feeder loads. 3 sTRUCTURAL FORMS 3.1 Layout Configurations Bing can be ground supported, skirt supported or elev- ated depending onthe ouloading system fromthecest point of view, ground supported bing resuit in the lowest structural cost but this may not be a governing factor in the overall design. The type of bin bottom has a great influence on the method of supporting the bin. Flat bottomed bins favour ground suppor although they can be readily built with elevated support. Hoppered bins are usually constructed as elevated structures. Much more fundamentally, bins can be classified into thin-shell structures and stittened flat plate structures. A typieal example of a thir-shell type 's a configuration consisting of cylingtical main body with conical hopper and oot. Stiffened fiat plate bins usually take the form of a rectangular main body with @ pyramidal or wedge- shaped hopper. Thin-shell bins offer the advantages of high structural efficiency, relatively low first and maintenance cost, ore reliable flow Conditions and better appearance. ‘The stiffened plate bins of rectangular plan have none of these advantages bul they do fake less space inthe @y plant and require less steep hoppers if wedge-type hoppers are used. Figure 7 illustrates the various layout configurations. (b) Entablature (a) Ground Supported Supported MMM (c) Elevated and Independantly Supported *(d) Single Unit (e) Row ®) Cluster (g) Row of Rectangular Fig. 7 Layout Configurations | | | itis usually convenient tojudge the economy of abin by aking the cost per tonne of the Dulk solid stored. The Unt Soa is expected to be lower for larger capacity ins. Larger storage capacities canbe providedina variety of ways: '* Using the largest feasible size of asingle bin. Welded Steel tanks of 40 000 to 60 009 tonne capacity have been built ‘© Using horizontal storage concepts such as storage sheds. ‘© Combining anumberofcelisina single structure. This feasiblewith rectangular and hexagonal cells using fened plate. The decision on the bins layout is usually dictated by economy, foundation conditions, available space, the eed to Cater for various types of bulk solids, and the avallable mechanical handling systems. In the final is, however, itis the total cost and reliability of the whole inetallation that must be considered and there~ Tore its not possible to give definite guidelines. 3.2 Cylindrical Bin with Conical Hopper ‘The basic conceptisto achieve a structural form which is capable of resisting the lateral loads, principally in tension, Cylindrical bins have the advantage of efficient design. Unavoidably there is always some bending and axial compression in parts of the bin but these can be Kept fo 2 minimum by judicious design, Because of the needto have the hopper outletat some height above the outloading conveyor or vehicle, the following three ‘methods of support are used: {) Skirt support. The cylindrical shell continues below the hopper junction and rests on a ring footing Entabiature structure, The cylindrical shell starts from the hopper junction where its supported by the entablature girders. At) Engaged support columns. In this frm of const ction the support columns are directly attached to the cylincrical shel. Figure 8 shows the three methods of support. For relatively small steel bins method (i) is the most econ- ‘Omical, although it restricts the space for the location of outloading equipment. ‘The choice of the hopper slopes depends on the desired flow mode, as explained in Section 1. Mass flow hoppers result in greater overall height than funnel flow ones. AWRA TECHNICAL NOTE 14 (Concrete sub-wall al dl, 1 nr) ¥ T Section Avs (ii) Entablature for elevated | i | | i | on t Gil) Engaged support columns Fig.’8 Methods of Support 13 AWRA TECHNICAL NOTE 14 3.3. Cylindrical Bin with Flat Bottom The typical flat-bottomed bin or tank has a height to diameter ratio between 0.8 and 1.5. The walls of a tank are supported viaan annulai ing. directyby the eating conereteraftor entablature. Thewalls of the larger tanks Ze designed for resistance to vertical buckling com bined with hoop tension. Themain advantage ofthistype of bin js in its relatively low structural cost. The main difficulty to be overcome with the flat-bottomed bins is, how to minimise the dead storage volume. This may not bbe a problem with low cost, non-degrading bulk solids because self-forming hoppers form around the several ‘outlets. The alternative is fo form the bottom to achieve the appropriate self cleaning slopes, Figure 8 shows some typical examples of flat bottom bins. 3.4 Rectangular and Polygonal Bins Rectangular bins offer the advant layout but this ls usually offset by Inbroased eeerehat costs. Unlike circular Bins which ‘camry ail ihe fonds predominantly in tension or compression, ular ins are predominantly stressed in bending. For this. reason the walls of rectangular bins are provided with stiffeners in one or both directions depending on the aspect ratio of the wall panels. The vertical walls of the bunkers also fulfil the second function - that ofa deep wall gitder spanning between the supporting columns. special form of rectangular bins (bunkers) uses wedge shaped hopperswith slot outlets forimproved discharge characteristics. One advantage of the wedge-shaped hoppers is that they are less prone to arching and can therefore be less steep than the pyramidal ones, ‘The essential components of arectangularbin orbunker are the following: (a) Upper wall panes, (©) Hopper wall panels, (6). Waist girder in form of horizontal rectangular frame, (@) Top rectangular frame, {e) Hopper outlet discharge framing, (Ridge girder where multiole hoppers are used. ‘The plate stiffeners may be arranged vertically or hori zontally and the choice depends on the ratio of heightto ‘width of the walls. Figure 11 shows the typical arrangements of stiffeners and other details, Fig. 9 Flat Bottom Bins Oftenitis desirable toclean out the bottoms completely, and this requires suitable mechanical clean-out devices ((otating screw conveyor), or pneumatic discharge sys- tem (air slides). Figure '10 illustrates one form of construction, AU “A, Fig.10 Flat Bottom Bin with clean-out device ‘The problem of additional loading due to eccentric flow conditions inherent in the design of this type of bin Is discussed in Section 4 — Stress Analysis. (a) Vertically (©), Horizontally stiffened hopper stiffened hopper Fig.11 Rectangular Bins @ 4. structuraLanatysis 4.1 General “The choice of the method of analysis depends on the {ype of structure used and on the accuracy of the sok ttlon desired, The structure of a bin can be divided into the main body, the support structure and the root. Each Of these paris requires a different approach to analysis Since each comprises diferent types of structural elements. ‘The steel supporting structures can be designed as — Rectangular frame joints; © Rectangular frame with triangulated bracing: ‘© Combination of rectangular and triangulated; Circular shell or skirt support Thedesign ofthe framedstructuresis usually carriedout Inacoordance with tho Steel Structures Code, AS 1250. should be realised, however, that the likelihood of the ccurrence of full design live load from the contents of the bin is very much higher than in normal building Structures and therefore a higher factor of safety may be Feaured Of parculr concen should be the possibilty Of differential settiement of foundations or of aiferentia! deflection of the supporting framework The structural analysis and design of the main body of a inis not at present regulated by any codes of practice. Some use can be made of the rules for liquid containers, guch es API 650 (Ref 10), BS 2654 (Ret 45) or AWWA Standard D-100 (Ref 40), but these may only apply to Some aspects of the design of cylindrical shells. Further Sources of design rules may be found in various codes forunfired pressure vessels, such as AS 1210 (Ret. 20), 97885500 (Het 51), Avery usetulfeature of these codes iS that they allow for 2 ited amount of plastic deform- ation inthe regions of severe discontinuity and sopermit 4 more economical design to be achieved. 4.2 Structural Analysis of Shells of Revolution The structural analysis and design of axtsymmetric sieel containers for buik solids is discussed in some detailin Ref,322. Inthis reference, the principalsections are devoted to the analysis of the stress distributions in steel bins and to their design against yielding and buckling, Other sections referto loads, supporting struc- tures, analysis for earthquake loading, fatigue, brittle fracture and lamellar tearing. In the section on design against yielding, orteria are Suggestedior the working stress design against genera) yielding in bins analysed by membrane theory, and againet gonoral and ical yielding in bins analysed by elastic bending theory. Cylindrical and conical shells subject to axt-symmetric loads can be analysed by one of the following methods: @ Membrane theory; Elastic shell bending theories; Finite element methods. ‘The choice of the method depends on the distribution of the lateral loads along the exis, on whether the part of the shell being analysed is sufficiently distant from the discontinuities and on the accuracy of solution desired. Membrane theory is the easiest to perform and tts accuracy is very good provided the following conditions are met: © The intensity of lateral pressure should not vary too abruptly along the xaxis. © According to Ref. 322, the part of the shell being analysed must not be closer to any geometrical or stress discontinuity than the distance Lo. Lo=2.4 VRAsinG} where angle @ is the angle between the axis and the normal to the shell (80 degrees for a cylindrical shell. AWRA TECHNICAL NOTE 14, Typical discontinuities are the changes in the shell lanmont a tho cylinder to, cone junction, the shel boundaries (eg. junction to the bottom plate) and re~ inforced openings for nozzles or manholes. Another type of stress discontinuity occurs at the point of applic- ation of concentrated forces (brackets, lugs). ‘Since membrane theory assumes that the shell has no bending resistance, it can only estimate the stresses in the plane of the shell ie. membrane stresses. To deter- mine the magnitudes of bending and shear stresses in the parts of the shell subject to Severe load gradients or ‘lose toa shell discontinuity, itis necessary touseashell bending theory or, alternatively, afinite element method. At present it is only practicable to use the elastic shell bending theories which give accurate estimates of the ‘membrane and bending stresses provided that no plas- tic deformation takes place anywhere in the shell It would be beneficial if some limited plastic deformation was permitted at the discontinuities but this presents {ermidable computational ificuties The practical way of overcoming this would be to make use of the elastic theory and permit some overstress. 4.3 Membrane Stresses In Cylindi and Conical Shells Membrane theory is described in references 90, 135, 235 and 315. Formulae for membrane stresses are given in references 90, 265 and 327. Some selected Solutions for typical loadings and geometries are given in Fig. 12. 4.4 Bending Stresses at Discontinuities The magnitude of the bending stresses due, to, the bending moments acting out of the plane of the shellcan ‘be estimated by using an elastic shell bending theory as described in references 90, 135, 235 and 315, The typical discontinuities which can give rise to significant bending stresses are: © The junction between the cylinder and the cone; © The junction between the cylinder and the flat bottom, © Change in shell plate thickness; © Attachment of ring stiffeners; © Attachment of bracket or lugs; '® Openings for nozzles and manholes. olutions for some typical loading conditions given in Figure 13 are computed from references 265, 135 and 15 [AWRA TECHNICAL NOTE 14 Fig.12 Membrane Forces in Cylindrical and Conical Shells Fig. 13 Bending Moments and Shears in Cylindrical Shells 1. Internal Pressu 14. Gradually Varying internal Pressure TTL, Hoop forge at dopmnz: \ = Pom Ber \ TY Resalepecementa cennz Vj "e} Yerdional forces are zero 1 B= 1288// Rt r p/p = 0.6040. R 6 = Bho P| M reset i w = PRYVELforZ> 078 VAT A Smax = 6 Mo it? Dat = 1812 pA 2. Hydrostatic Preseure Hoop foree at <3) Hoop force at B: Peeps Polécose eristonatiore at A Pr Ro; Beaaraispacemen a AS sesmenue Hoop force at A: PpapRoieosa force at B: Pe icose Friction Forces Hoop force: Yoon Meridional force: P= QRol? sina fia spiecament: 5. Gravity Loading Wy = self weight of hopper in KN (total) Ws Uniformly distributed load trom fi inkN' (otal) W= concentrated load et epexin kN jg produces Py = 0 Meridional force at A Pao ti + We 4 Ws\/2Roxcose 6, Trlangular Pressure Distribution Hoop force at a: By=pyRo/eosa Merigional torce at Bn 05ap; ane, for @< 0.38 Rial displacement at A: we PeRolet 7. Filled Hopper Hoop force: Pes tanction of Pn Meridionat force at & Pry Wel2Ro.n.cosa 6 2. Moments and Radial Forces wtEoge pin N per mm Min Nimm perm Maman = Qo! 2 88Giveein =ay02608:26* 3. Internal Pressure = og020 Ry on Bre Mme = 0.18%B/6, = Sos7Bans tz wae mex = SMP wen DB SeaRRA # 4, Moments and Radial Forces at edge 22209 STARE Rot 2 = r/4p 5. Internal Pressure Q,= Myf, Mu (O/2BAIAMA, + to YB) ym area of ting gitder cf height ofing girder Stsing Acommon situation in the design of cylindrical bins with conical bottoms that requires careful consideration is, the ring stiffener region at the cone junction. Using the. method of sections, itis relatively easy to determine the appropriate stress resultants, as shown on Figure 14. ‘The problem is then how to translate these into @ prac tical design because — () The loads cannot be estimated with great accuracy; eg, the lateral loads acting on the upper part of the cone; The stresses obtained correspond to the predic tions of elastic shell theory which cannot take into account the beneficial eect of localised yielding at the surface of the shell plate. A practical solution to this problem is to use the bending and membrane. stresses, obtained from the elastic, theory, and then to apply a 33% reduction tothe bending stresses so obtained. The alternative is to increase the ible stress value to 1.0F,, The latter method is. Specified in AS 1210, Supplement 1 which reflects the practices in the pressure vessel industry. ‘Assumptions: (@) No rotation takes place atthe joint‘. {b) Displacements wy, and ws; are for siiging ints, fixed against rotation. Condition of equtiium: REX + X= 0 were X; and Xq are the shear forces per nit of fength. and A= Pesina Conaitions of compatibility of asplacements: Wyo Wa = Woo ~ Was OK + Xa) 0 Wao — WarK ~ Wao ~ Was Os #3)" 0 where: ‘wi9™ dsplecement ofthe upper well due to py ‘zo displacement ofthe rng gider due to R {use negative sign} Wo displacement of the cone due oP Wi = displacement of the upper wall ue 10 X wwii displacement ofthe ring grder due 10 0) + Xa) = 1.0 Way = displacement of the cone due 10 Xs. Solution: % AB X= O/D wher [A= Wao ~ Wag+ Wostwro~ woo}llis + Won) B= Wh toys + Wey )— Woy — Wt Cm Wag = vo + WaslWg0~ Wa0l(as + Wor) wh Awe + Wai) Way = Way Use Fig. 19 to calculate the bending stresses due to X; and Xe 4.5. Finite Element Methods of Shell Analysis Finite element methods provide a tool of analysis for shells of any complexity and toading concltions (Ret 870, 230). The programs currently available, such as. SAP, EASE, MSO/NASTRAN, ars based on a linear el- stig finite élement thoory and there is no provision for Introduction of local plastic yielding in the highly stress- ed regions of the shell. tis, however, possible to iter- atively change the Youngs modulus isa Tangent mod, Ulus and repeat the run. The comparatively high cost of running the finite element method programs and inter pretating the results of analysis preclude the use of this method for any but major bins. Considerable cares needed to select the most approp- riate types of finite elements and to devise a mesh which is neither too coarse nor too fine. The designer should seek specialist advice from the computer service bur- ‘eaus oF from engineers specialising in the field. Figure 18 (a) shows a possible mesh for a bin supported on columns which are integral with the shell AWRA TECHNICAL NOTE 14 ‘Amajor simplification can be introduced when the shell is axisymmetric and free from local discontinuation. In this finite element approach the shell is subdivided into ngsegments as shown in Figure 15(0) The description of euch a program can be found xample of uch an snalyeis i found in ef 36° he Ethers a onal sent oe TH - ar lament @ ©) Triangular and quadrilateral shell bending elements Annular elements for ‘axisymmetric loads Fig. 15 Finite Element Subdivision 4.6 Semi-Empirical Design Methods Fromthe rich experience gathered from pressure vessel technology, it can be concluded that certain semi. ‘empirical methods of design of the cylindrical shell end. ‘closures (conical shells) have a satisfactory safety re- cord. The Unfired Pressure Vessel Code AS 1210 gives formulae for designing the cylinder to cone transitions, without ring girders which may apply to relatively small bins within certain restrictions given tn Rule 3.10 of that code. Figure 18showstwodesignstorthe knuckle region. Another semi-empirical methods specifiedin BS 1515, Part 1, Rule 36. « e R=Rsing po (ax erieere)= Bre \p (2) Torisoherical __(b) Straight transition (knuckle) traneition Fig. 16 Forms of Transistion from Cylinder to Cone w AWRA TECHNICAL NOTE 14 4.7 Stress Analysis of the Conical Hopper Transition ‘The junction of the main body of the cylindrical shell and the conical hopper causes several load effects which give rise to a very complex stress ficld, Firstly, the membrane tension forces are deflected severely and the resultant inwardly directed forcemmust beresisted by the ring girder, orifthe ring girderis not used. by the shell in the immediate vicinity of the knuckle. Secondly, the disparity of the deformation in the three elements, cy- linder,cone and ring, results in bendingstress and shear in the shell plates which usually exceed the membrane stresses. Finally, the ring girder usually doubles as the support girder (bow girder} ‘Some of the commonly used designs for the knuckle Tegion are shown in Figure 17 and ete described as follows: (@) Support brackets; these are usually attached tothe Gyfinder and produce very sevare sttesses in the shell (Only suitable for relatively small bins). (©) Stub column supports have the advantage of pro- viding a direct path for the vertical forces. The ‘Supporting beams are usually arrangedas shownin Figure 17 (f) so as to achieve equal cisinbution of the bin reactions. (©) Engaged columns offer a most direct transfer of vertical reactions into the substructure. (@ Triangular cross section waist, girder fulfils the function of compression ring and bow gitder in one unit. Careful design for torsional stressesis required. (©) Stiffened bow girder offers the advantages of simple fabrication and good performance. The, bottom flange of the integral girder doubles as the com- For preliminary desion purposesit may be conveni Uso the following simpliving assumptions Wobe vertice by.a more ngorous analysis): () Determine the axial compressive force inthe ring rom: Pa=Pysing R where PIs the meridional longitudinal tensile force inthe cone. The effective cross-sectional area of the compression ring should include the whole area of the ring directly connected to the knuckle, and the areasof the shell plates limitedtoawidth of 10 times their respective thickness. ) Determine the bending and torsion moments in the ring girder using the bow girder theory. Figure 18 ives coefficients for bending and torsion moments, (li) Combine the stresses from steps (i) and (i, The negative bending moment stresses may combine with the ring compression stresses in some designs. (v) Check the maximum possible bending stress inthe Gylinder and the cone assuming full fixity at the ring idler Figure 19 gives the appropriate expressions for the calculation of the shell bending stresses. Actual stresses will usually be lower than these since the assumption of full fixity may not be realistic, and since plastic yielding may reduce the bending Stresses. Alternatively, a method for the calculation of stresses from Feet. 142 can be employed, (¥) The meridional compressive stresses in the cylind- ‘foal shell should be increased 60% to account for the stress concentration over the columns. (See Ref 50). Pression ing The lop flange.end the acjoiing shell M 18 Biste form the uper oouttday of he girder. Ring Girder Moment Apart from the support arrangements shown in Figure for Vertical Loading 17 (a) (b) andi(c) twould not be correct to design the ring graers {a sogtion trom ine shell othe eine Ths wes Ssmonstrated by Rotter (Re! 80) in an erie Investig. My = Wi x coef. ating the behaviour of column supported bins, The Mya = W.Ra x coeff. anajsis using a finite element method. should give Fe My = W.Ra x coeff, tesuls which are far superar to those cbiained by 2 Simplfied anaiyeig assuming that no interaction ocovrs Between the bin shell and tne ring girder. Further read Number of supports (n) Ingonthe subject can be found inteterences 140, 160, 4 6 8 185, 170, 175, 270, 277, S22, 345 and 365. Coariciente ‘The integral girders shown in Figure 17 (e) are never- reine Bia esses by ‘using Simpl rng guder — Moment aA -0.03415 -0.01482 -0.00827 IRSSee thetbendny ements era tomorel Moments — MomentatM 40.01762 40.0761 40.0416 can be calculated as shown in Figure 18, Torsion at C 0.00530 0.00151 0.00063 2. > . 10: 6 1- Transition knuckle 3 4 y 2- Shell reinforcing plate i: 3 Support bracket 4 : : 4= Ring girder bee 5- Stool o\ .S 8= Shell stitener 77 Engaged column 8- Column extension 8 Twenguler waist girder 10 Suifened bow gitéer ©) © @ o Fig. 17 Methods of Bin Support and Arrangement of Ring Girders 1B ¢ ‘The common practice of disregarding the shell bending stresses on account of plastic yielding is grossly un- conservative, although surprisingly few bins are dam- Sgedasaresuttof thispractice Butatthe other extreme, A.rigorous analysis using elastic shell bending theory Gan produce uneconomical designs with thick plates, a farge amount of welding and a danger that thick plates sould be susceptible fo brittle fracture (Ref 63) and lamellar tearing (Ref 64). The compromise. solution would appear to be in allowing some plastic yielding to fake place and to limit the yield strain to, say 1 percent (fefer to 110, 125, 140, 145 and 205). 4.8 Buckling Resistance of Cylindrical Shells Cylindrical shells used for upper part of bins are yuk. nerable to local shell buckling which may be caused by © axial compressive forces due to friction between the bulk solid and the walis; © external wind pressure, internal pressure drop, or non-uniform lateral wall loads; © localised compressive membrane stresses due to loads from brackets 4.8.1 Axial Compression The prevalent mode of buckling due to axial compress- ions In the form of diamond pattern buckling. The bulk solid inside the bin provides some elastic lateral res- traint to the shell andthe hoop tension itself is known to Increase the buckling resistance of axially compressed. cylindrical shells. The critical buckling stress for cylindrical shells with relatively small imperfections is given by (Ref 90): c= Keks(0.605Et/R) The expression inside the parenthesis is the critical theoretical buckling stress of an ideal cylinder. The coefficient k, is sometimes termed “knock down factor’ because it ig a measure of a strength reduction due to imperfect shape of the walls of the cylinder. The value of ig angee betwoen 018 and 022 for typical tee! bine lef 322), depending on the ratio R/t Rt ke 500 0.21 1000 022 1500 018, 2000 016 ‘The coefficient ksisa strength enhancement coefficient which depends on the magnitude of the circumferential tension induced by the lateral pressure of the bulk solid, Tests have shown (Ref 275) that under favourable cir- cumstances the value of k, may exceed 4.0. However, it Iscuggestedthat much lower values be used because of the uncertainties in the prediction of the lateral loads. under flow concitions. The following values should be used for the preliminary design: oR/Fyt ka 0 10 200 12 400 14 600 16 800 18 1000 20 ‘The circumferential stress.o,~pyifi/t,correspondstothe Initial loading condition. AWRA TECHNICAL NOTE 14 4.8.2 Unsymmetrical Loading Unsymmetrical loading may occur as a result of: (a) Eccentric flow caused by off-centre openi symmetrical hoppers or eccentric filling. (b) Uneven draw-olf especially where multiple charge hoppers are used, (©) Markedly non-uniform flow mode where instantan- eous load peaks occur in a random pattern, ‘There are no simple manual calculation procedures to deal with unsymmetrical loading and recourse to FEM analysis is necessary. Suggested references are 280, 155 and 360. 195, uns is: 4.8.3 External Pressure 4.8.3.1 Wind Loads Cylindrical bin walls are very vulnerable to wind induced buckling. The enilcal time fs when the bin is empty oF partly empty. The distribution of wind pressure atound the circumference is shown in Figure6, Section 2, using data from NZS 4203 (1876) The windward part’of the wall is in an unstable condition and becomes progress- ivelymoreunstable asthe buckle develops. The problem Is complicated because of the variably of wing press. Ure over ihe wall surface and also Because of the dynamic effect of wind gusting The problem can be simplified by assuming uniform external pressure. In this case, reference 352 gives the following estimate of the pressure causing buckling: Pe =O.74E(L/RY |. (R/t) 28 where Lis the height ofthe cylindrical shell between the ring stiffeners or between the tension ring at the roof level and the first stiffener. Awell proven method of determining the need for wind stiffeners in the form of ing girders Ie given in BS 2564 whichis based onexperiencein the olltankconstruction industry. The procedure takes several steps, as follows: () Determine the maximum allowable height of the ‘shell without wind girders: Hp and D in m in in mam Vin m/s where K=95 000/(3.563V, + 5800q4) wis the design wind velocity, and isthe internal drop in pressure in KPa. i) I the thickness of the shell within the distance H, varies, an average thickness should be calculated rom: to VG +2) (il) Determine the equivalent stable height of the shel: Hem zh Vetin/ A (iw) The number of wind girders is determined as follows: Hp

You might also like