0 Stimmen dafür0 Stimmen dagegen

3 Aufrufe93 SeitenDesign of Reinforced Concrete Silo Groups

Sep 21, 2019

© © All Rights Reserved

PDF, TXT oder online auf Scribd lesen

Design of Reinforced Concrete Silo Groups

© All Rights Reserved

Als PDF, TXT **herunterladen** oder online auf Scribd lesen

3 Aufrufe

Design of Reinforced Concrete Silo Groups

© All Rights Reserved

Als PDF, TXT **herunterladen** oder online auf Scribd lesen

- Finite Element Method for Professionals
- Shear Force and Bending Moment
- Abstract
- Design of ACP panel
- AFEM.Ch21
- material shape.ppt
- F1.xlsx
- 143258201 احمال الزلازل ياسر الليثي
- hse
- 130829_Course Syllabus_me5510 me6510_Fall 2013 (1).pdf
- Derivation of Link Equations
- Effective breadth-Espol
- AE2302 2 Marks
- Slide Chpt08
- 001 RCM Chapter 1 - Introduction
- EvolutionandDesignChallengesforFlexibles-ForIssue
- 2. Emergency Gate
- Lesson01 Computational Modelling
- MEStrength of Materials
- Recidual Force Equation

Sie sind auf Seite 1von 93

Fernando A. N. Silva

Bernardo Horowitz

João M. P. Q. Delgado

António C. Azevedo

Design of

Reinforced

Concrete Silo

Groups

Building Pathology and Rehabilitation

Volume 10

Series Editors

Vasco Peixoto de Freitas, University of Porto, Porto, Portugal

Aníbal Costa, Aveiro, Portugal

João M. P. Q. Delgado, University of Porto, Porto, Portugal

This book series addresses the areas of building pathologies and rehabilitation of the

constructed heritage, strategies, diagnostic and design methodologies, the appropri-

ately of existing regulations for rehabilitation, energy efﬁciency, adaptive rehabil-

itation, rehabilitation technologies and analysis of case studies. The topics of

Building Pathology and Rehabilitation include but are not limited to - hygrothermal

behaviour - structural pathologies (e.g. stone, wood, mortar, concrete, etc…) -

diagnostic techniques - costs of pathology - responsibilities, guarantees and

insurance - analysis of case studies - construction code - rehabilitation technologies -

architecture and rehabilitation project - materials and their suitability - building

performance simulation and energy efﬁciency - durability and service life.

Fernando A. N. Silva Bernardo Horowitz

• •

João M. P. Q. Delgado •

António C. Azevedo

Design of Reinforced

Concrete Silo Groups

123

Fernando A. N. Silva Bernardo Horowitz

Department of Civil Engineering Department of Civil Engineering

Catholic University of Pernambuco Federal University of Pernambuco

Boa Vista, Recife, Brazil Recife, Brazil

Faculty of Engineering, CONSTRUCT-LFC Faculty of Engineering, CONSTRUCT-LFC

University of Porto University of Porto

Porto, Portugal Porto, Portugal

Building Pathology and Rehabilitation

ISBN 978-3-030-13620-8 ISBN 978-3-030-13621-5 (eBook)

https://doi.org/10.1007/978-3-030-13621-5

This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether

the whole or part of the material is concerned, speciﬁcally the rights of translation, reprinting, reuse of

illustrations, recitation, broadcasting, reproduction on microﬁlms or in any other physical way, and

transmission or information storage and retrieval, electronic adaptation, computer software, or by similar

or dissimilar methodology now known or hereafter developed.

The use of general descriptive names, registered names, trademarks, service marks, etc. in this

publication does not imply, even in the absence of a speciﬁc statement, that such names are exempt from

the relevant protective laws and regulations and therefore free for general use.

The publisher, the authors and the editors are safe to assume that the advice and information in this

book are believed to be true and accurate at the date of publication. Neither the publisher nor the

authors or the editors give a warranty, expressed or implied, with respect to the material contained

herein or for any errors or omissions that may have been made. The publisher remains neutral with regard

to jurisdictional claims in published maps and institutional afﬁliations.

This Springer imprint is published by the registered company Springer Nature Switzerland AG

The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

Preface

Groups of reinforced concrete silos are structures commonly used in the food

industry, where it is usually necessary to separate the storage of different types and

sources of grain.

The grouped layout of silos has numerous beneﬁts when compared with

single-cell silos in which the emphasis is on creating further space for silage,

normally referred to as interstice—a space formed between the edges of the group’s

cells.

This economic beneﬁt, on the other hand, raises a structural problem for the

designer of this type of building, which is to assess the magnitude of bending

moments and hoop forces due to the structural continuity of the walls in the

interstice region of the cells.

Bending moments assume extreme values exactly when the interstice is loaded

and the other cells in the group are empty. In order to be able to achieve economic

and safe designs, it is, therefore, essential to understand the structural behaviour of

this type of structure, considering the loading imposed upon it.

The purpose of this book is to present a new calculation procedure of those

moments, easy to use and with satisfactory responses when compared to the

three-dimensional analysis using the ﬁnite element method (FEM), which today is

the state-of-the-art structural analysis of this type of construction.

To develop the formulation of the proposed analysis models, a parametric study

was carried out that allowed the adequate consideration of the variables involved.

The book is divided into six chapters.

Chapter 1 contemplates the characterization of the problem to be solved. It

provides a bibliographical review on the methods of calculation of the bending

moments due to the structural continuity in a group of silos available in the liter-

ature with brief comments on their foundations.

Chapter 2 presents the geometry of the groups of silos analysed and a review

of the aspects related to the applied loads. It is also deﬁned as the physical

parameters of the stored material and the procedures for calculating the design of

horizontal pressure diagrams to be used.

v

vi Preface

includes the deﬁnition of the method, the load cases considered, the ﬁnite element

meshes used in the analysis of the resulting structure and concepts related to the

application of the loads to the models generated. The theoretical formulation for the

imposition of constraint equations (required in the modelling with frame elements)

and the procedure used to calculate the stresses resultants in the solid element

modelling are also included in this chapter. In the end, a comparison of results

between the analyses carried out and conclusions about the use of the exposed

method is presented.

Chapter 4 addresses the three-dimensional analysis of groups of reinforced

concrete silos. Three types of modelling strategies are presented—solid elements,

shell elements, and shell and solid elements—with their respective ﬁnite element

meshes. In the modelling with shell and solid elements—here called mixed mod-

elling—the procedure to be used in the generation of the constraint equations,

necessary for the connection between the two types of mesh elements, is presented.

At the end, a comparison of results between the analysed models is presented.

Chapter 5 presents the interpretation of the structural behaviour and the for-

mulation of the proposed analysis models. The parametric study developed, which

enabled the establishment of the recommended calculation procedure, is also dis-

cussed. Finally, a comparison between the results obtained with the application

of the proposed analysis model and those resulting from the calculation methods

available in the literature and the Finite Element Method—FEM—are presented.

Chapter 6 summarizes the conclusions of the study carried out and contemplates

the recommendations about the analysis of groups of reinforced concrete silos.

Detailed and commented description of the steps required to use the proposed

model, with an example of application, are also presented.

Recife, Brazil Bernardo Horowitz

Porto, Portugal João M. P. Q. Delgado

Porto, Portugal António C. Azevedo

Contents

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

2 Geometry and Loading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

2.1 Geometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

2.2 Loading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

2.2.1 Janssen’s Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

2.2.2 Calculation of Overpressures . . . . . . . . . . . . . . . . . . . . . . 9

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

3 Two-Dimensional Linear Elastic Analyses—Slice Method . . . . . . . . . 13

3.1 Modelling Using 2D Frame Elements . . . . . . . . . . . . . . . . . . . . . 14

3.2 Applying Nodal Loads . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

3.3 Constraint Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

3.4 Lagrange Multiplier Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

3.5 Modelling with Plane Strain Finite Element . . . . . . . . . . . . . . . . . 20

3.6 Additional Shape Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

3.7 Calculation of Stress Resultants . . . . . . . . . . . . . . . . . . . . . . . . . . 23

3.8 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

3.8.1 Deformed Geometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

3.8.2 Horizontal Normal Forces . . . . . . . . . . . . . . . . . . . . . . . . 25

3.8.3 Bending Moments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

3.8.4 Bending Moments Envelope . . . . . . . . . . . . . . . . . . . . . . . 32

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

4 Three-Dimensional Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

4.2 Boundary Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

4.3 Modelling with Linear Finite Element . . . . . . . . . . . . . . . . . . . . . 37

4.3.1 Loading in the Cell . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

4.3.2 Loading in the Interstice . . . . . . . . . . . . . . . . . . . . . . . . . 41

vii

viii Contents

4.5 Modelling with Solid and Shell Elements . . . . . . . . . . . . . . . . . . . 44

4.6 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52

5 Structural Behaviour and Simpliﬁed Model Proposition . . . . . . . . . . 53

5.1 Load Bearing Mechanism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

5.2 Proposed Simpliﬁed Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54

5.3 Fitting with the Least Square Method . . . . . . . . . . . . . . . . . . . . . 59

5.4 Parametric Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63

5.5 Procedure Proposed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71

5.6 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73

5.6.1 Bending Moments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73

5.6.2 Horizontal Normal Forces . . . . . . . . . . . . . . . . . . . . . . . . 76

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82

6 Conclusions and Recommendations . . . . . . . . . . . . . . . . . . . . . . . . . 83

List of Figures

Fig. 2.2 Schematic drawing for deduction of Janssen’s formula . . . . . . . . 8

Fig. 2.3 Horizontal design pressure curves . . . . . . . . . . . . . . . . . . . . . . . . 10

Fig. 3.1 Cross section in silo groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

Fig. 3.2 Load cases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

Fig. 3.3 Modelling with 2D frame elements . . . . . . . . . . . . . . . . . . . . . . . 15

Fig. 3.4 Detail of the region of connection between two cells

of the silo group . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 16

Fig. 3.5 Mesh of the slice of the silo group using plane strain

ﬁnite element . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 20

Fig. 3.6 Four node ﬁnite element—Q4 . . . . . . . . . . . . . . . . . . . . . . . . . .. 21

Fig. 3.7 Deformation of the Q4 element without additional shape

function. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

Fig. 3.8 Data for the calculation of stress resultants . . . . . . . . . . . . . . . . . 23

Fig. 3.9 Deformed geometry of the silo group—Load Case I . . . . . . . . . 24

Fig. 3.10 Deformed geometry of the silo group—Load Case II . . . . . . . . . 25

Fig. 3.11 Deformed geometry of the silo group—Load Case III . . . . . . . . 25

Fig. 3.12 Deformed geometry of the silo group—Load Case IV . . . . . . . . 26

Fig. 3.13 Deformed geometry of the silo group—Load Case V . . . . . . . . . 26

Fig. 3.14 Deformed geometry of the silo group—Load Case VI . . . . . . . . 27

Fig. 3.15 Bending moments—Load Cases I e II . . . . . . . . . . . . . . . . . . . . 28

Fig. 3.16 Bending moments—Load Cases III e IV . . . . . . . . . . . . . . . . . . 29

Fig. 3.17 Bending moments—Load Case V . . . . . . . . . . . . . . . . . . . . . . . . 30

Fig. 3.18 Bending moments—Load Case VI . . . . . . . . . . . . . . . . . . . . . . . 31

Fig. 3.19 Bending moment’s envelope—Frame elements . . . . . . . . . . . . . . 32

Fig. 3.20 Bending moment’s envelope—Q8 element . . . . . . . . . . . . . . . . . 33

Fig. 4.1 Structure for three-dimensional analysis—cross sections

of interest . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 36

Fig. 4.2 Finite element mesh with linear solid elements . . . . . . . . . . . . .. 38

Fig. 4.3 a Horizontal normal forces—loading in cell and b variation

of the maximum normal horizontal forces. . . . . . . . . . . . . . . . .. 39

ix

x List of Figures

bending moments Mz along the height . . . . . . . . . . . . . . . . . . .. 40

Fig. 4.5 a Normal horizontal forces—loading in the interstice—

Section 3 and b variation of the maximum normal

horizontal forces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 41

Fig. 4.6 a Bending Moments Mz —loading in interstice and b variation

of the maximum positive and negative bending moments . . . . . . 43

Fig. 4.7 Mesh with shell ﬁnite element . . . . . . . . . . . . . . . . . . . . . . . . . . 44

Fig. 4.8 Mesh with solid and shell element . . . . . . . . . . . . . . . . . . . . . . . 45

Fig. 4.9 Connection of solid and shell elements . . . . . . . . . . . . . . . . . . . . 46

Fig. 4.10 Bending moments Mz in section 1 and 2—loading applied

in the cell . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 48

Fig. 4.11 Bending moments Mz in section 3 and 4—loading applied

in the cell . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 49

Fig. 4.12 Bending moments Mz in section 1 and 2—loading applied

in the interstice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 50

Fig. 4.13 Bending moments Mz in section 3 and 4—loading applied

in the interstice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 51

Fig. 5.1 Structural behaviour . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 54

Fig. 5.2 Deformed shape of the interconnection region—loading

applied in interstice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 55

Fig. 5.3 Normal stresses—longitudinal direction

(in kN/m2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56

Fig. 5.4 Proposed simpliﬁed model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56

Fig. 5.5 Scheme for calculating the spring constant . . . . . . . . . . . . . . . . . 57

Fig. 5.6 Detail of the Timoshenko beam model . . . . . . . . . . . . . . . . . . . . 58

Fig. 5.7 Calculation of transversal loading . . . . . . . . . . . . . . . . . . . . . . . . 59

Fig. 5.8 Comparison of nodal displacements . . . . . . . . . . . . . . . . . . . . . . 62

Fig. 5.9 Comparison of bending moments . . . . . . . . . . . . . . . . . . . . . . . . 64

Fig. 5.10 Parameters for geometric index H/D = 3.0 . . . . . . . . . . . . . . . . . 65

Fig. 5.11 Parameters for geometric index D/t = 20 . . . . . . . . . . . . . . . . . . 66

Fig. 5.12 Variation of parameter c . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

Fig. 5.13 Bending moments for the geometric index D/t = 20 . . . . . . . . . . 68

Fig. 5.14 Bending moments for the geometric index D/t = 30 . . . . . . . . . . 69

Fig. 5.15 Bending moments for the geometric index D/t = 40 . . . . . . . . . . 70

Fig. 5.16 Bending moments for the geometric index D/t = 50 . . . . . . . . . . 71

Fig. 5.17 Bending moments for the geometric index D/t = 60 . . . . . . . . . . 72

Fig. 5.18 Bending moments in Sect. 4 of Fig. 4.1 . . . . . . . . . . . . . . . . . . . 76

Fig. 5.19 Bending moments in Sect. 3 of Fig. 4.1 . . . . . . . . . . . . . . . . . . . 79

Fig. 5.20 Horizontal normal forces—plan of symmetry . . . . . . . . . . . . . . . 80

Fig. 5.21 Horizontal normal forces—support . . . . . . . . . . . . . . . . . . . . . . . 81

List of Tables

Table 5.2 Negative bending moments . . . . . . . . . . . . . . . . . . . . . . . . . . . 61

Table 5.3 Positive bending moments . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63

Table 5.4 Parameters for H/D = 3.0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64

Table 5.5 Parameters for D/t = 20 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

Table 5.6 Parameters c. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

Table 5.7 Negative bending moments . . . . . . . . . . . . . . . . . . . . . . . . . . . 74

Table 5.8 Positive bending moments . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75

Table 5.9 Horizontal normal forces—plane of symmetry . . . . . . . . . . . . . 77

Table 5.10 Horizontal normal forces—support . . . . . . . . . . . . . . . . . . . . . . 78

xi

Chapter 1

Characterization and Brief Literature

Review

groups of reinforced concrete silos due to interstice loading.

The groups of silos considered in the work are constituted of four cylindrical cells

of equal diameter and it is of special interest to investigate the magnitude of bending

moments at the ends of the regions between two cells, since it is in these locations

that the highest values of bending moments usually occur. Such moments are often

referred to as moments due to structural continuity.

Live loads are pressures exerted by the material stored inside interstice space

generated by the wall intersections of the cells of the group.

Literature offers many methods for calculating bending moments due to the

structural continuity in groups of reinforced concrete silos in an interstice-loading

situation.

Albigés and Lumbroso [1] provided expressions based on the analysis of a unit

height horizontal slice of the group. This slice is loaded with the resulting force of

the pressure applied to the interstice, with forces due to the walls intersection of the

cells of the group and with the elastic forces due to the differences in shear force

on the lower and upper sections delimiting the slice. The proposed formulation also

considers that the wall thickness and the length of the region common to two cells

are small in relation to the average radius.

Ciesielski et al. [2] presented a method in which the arch supports comprising

the interstice wall undergo certain tangential displacement. The solution to the prob-

lem is expressed by six equations that, due to the angular location of the section

studied, provide the values of the bending moment and normal force. The proposed

method considers that the interstice wall of groups formed by three, four or five silos

undergoes radial loading.

Timm and Windels [3] provided formulas to calculate bending moments based

on the hypothesis that the region connecting two cells prevents rotation, but per-

mits tangential displacement of the interstice wall supports, which, according to the

authors, behaves like an arch undergoing radial pressure due to the applied loading.

F. A. N. Silva et al., Design of Reinforced Concrete Silo Groups, Building Pathology

and Rehabilitation 10, https://doi.org/10.1007/978-3-030-13621-5_1

2 1 Characterization and Brief Literature Review

The formulas allow the calculation of bending moments, shear forces and tangential

forces per unit of length for different sections of the arch, as a function of its angular

location.

Safarian and Harris [4] suggest a numerical solution using computational matrix

analysis of a unit height section. The part of the section corresponding to the interstice

wall is modelled as an arch with very rigid elements in its supports. The loading

applied in a radial pressure, distributed along the element length.

Haydl [5] presented a finite element solution based on general-purpose software

[6], where a unit height section of the silo group is analysed using shell elements.

The region connecting two cells is modeled with three elements, applying specific

boundary conditions to their nodes to take into account the connection with other

silos. The solution also considers that the centre of the connecting region is situated

30° of the crown of the arch comprising the interstice wall and its thickness is four

times the wall thickness. The results are expressed in direct-application formulas of

values of the wall thickness/silo wall thickness ratio of 1 and 2.

Gurfinkel [7] proposes formulas for calculating the bending moments that are an

adaptation of formulas of Kellner [8].

As suggested by Gurfinkel [7] and Safarian and Harris [4], there is also the clamped

arch model. If the hypothesis considered is perfect fixed supports and the loading is

in radial direction, the bending moments calculated by this method are negligible.

Prato and Godoy [9] realized the importance of the three-dimensional nature of

the problem, but the modelling of the interstice region used by the authors is too

simplified, bearing in mind its strong influence on the applied forces range.

Balkaya et al. [10] also recognizes the importance of tri-dimensional effects on the

overall behaviour of silo groups but propose a 2D model based on beam elements to

represent the geometry of the overlapping region that clearly underestimates complex

phenomena that occur in that region.

With the exception of Prato and Godoy, all other aforementioned methods are

based on two-dimensional analysis that, by their very nature, overlook the influence

of the three-dimensional interaction of the different cells in the group, a decisive

aspect in the structural response of the silo groups under interstice loading.

Moreover, because of the different boundary conditions used in the analysis, the

numerical results provided by these methods are very conflicting. This fact leaves

the designer with no exact benchmark for choosing the most suitable method, an

even more relevant aspect when it is known that the divergence between the top and

bottom value of the bending moments obtained using these methods can be thirty

times or more, according to Horowitz and Nogueira [11].

On the other hand, because of the problems in modelling and interpreting results,

three-dimensional analysis using finite element method is not yet easy to apply in the

daily design of this type of structure. In this context, the book herein intends to offer

an alternative to calculate the bending moments in groups of reinforced concrete

silos, which provides quite satisfactory results compared with the three-dimensional

analysis but has the benefit of being easy to apply in the preliminary and final design

stages. The basis, validation, and demonstration of the effectiveness of the procedure

proposed are discussed in detail.

References 3

References

de carreau. Annales de L’Institut technique du bâtiment et des travaux publics, pp 1547–1562

2. Ciesielski R, Mitzel A, Stachurski W, Suwalski J, Zmudzinski Z (1970) Behalter, Bunker, Silos,

Schornsteine, Fernsehturme und Freileitungsmaste. Verlag von Wilhelm Ernst & Sohn, Berlin

3. Timm G, Windels R (1977) Silos, Sonderdruck ans Dem Beton-Kalendan. Verlag von Wilhelm

Ernst & Sohn, Berlin, Germany

4. Safarian SS, Harris EC (1985) Design and construction of silos and bunkers. Van Nostrand

Reinhold Company, New York

5. Haydl HM (1987) Bending of interstice walls in circular silos. J Struct Eng ASCE

113(10):2311–2315

6. SAP 2000 (2014) User’s manual, Wilson L Edward and Habibullah Ashraf

7. Gurfinkel G (1990) Reinforced-concrete bunkers and silos. In: Gaylord H, CN Gaylord (eds)

Structural engineering handbook, Sect 26. Wiley, New York

8. Kellner M (1960) Silos a Cellules de grande profondeur, Organe de la technique des travaux

publics et du ciment arm, pp. 612–622

9. Prato CA, Godoy LA (1989) Bending of multi-bin RC cylindrical silos. J Struct Eng ASCE

115(12):3194–3200

10. Balkaya C, Kalkan E, Yuksel SB (2006) FE analysis and practical modelling of RC multi-bin

circular silos. ACI Struct J 103(2):365–371

11. Horowitz B, Nogueira FA (1999) Stress resultants due to interstice loading in group of four

cylindrical silos. ACI Struct J 96(2):307–313

Chapter 2

Geometry and Loading

2.1 Geometry

The geometry of the silo groups of interest is shown in Fig. 2.1. They may have a

thick slab (or conical hoppers with ring beams) at the bottom, and a slab, possibly

on beams, at the top. Due to the thick slab close to the foundation, we will assume

the cylindrical shell clamped at the bottom slab. The objective is to compute bending

moments and hoop forces on the shell due to interstice loading, in order to be able

to design the horizontal reinforcement of the walls.

The group of silos showed in Fig. 2.1 is composed of four cylindrical cells of the

same diameter that forms an additional space for store materials, knows as interstice.

The acronym SOR means Shell Overlapping Regions, which represents the area of

intersections of the walls of adjacent cells. To show the quality of the model of

analysis proposed, an illustrative example with H = 30 m, D = 10 m, t = 0.2 m and

p = 42.46 kN/m2 is solved in Chap. 6. Regarding the diameter and heights considered

in the example, also investigated values cover a wide range of groups of existing silos

used to store granular materials. In fact, groups with circular cells with diameters of

4, 6, 8, 10 and 12 m and heights of 12, 15, 18, 20, 22, 24, 25, 30 m were studied.

Chapter 5 presents the results of the parametric investigation performed.

2.2 Loading

The loading considered is that resulting from the action of the ensiled material on the

walls and bottom of the cells. This action takes place through the following pressures:

• Vertical pressure;

• Lateral pressure and

• Vertical friction force.

F. A. N. Silva et al., Design of Reinforced Concrete Silo Groups, Building Pathology

and Rehabilitation 10, https://doi.org/10.1007/978-3-030-13621-5_2

6 2 Geometry and Loading

The responsibility for supporting the weight of the material stored inside a silo

is assumed by the bottom, through the vertical force coming from the weight of

the material on the existing slab, and by the walls, mobilized by the vertical force of

friction generated by the contact with the same ensiled material. In addition, pressures

arise perpendicular to the walls, usually referred to as lateral or horizontal pressures.

The horizontal pressures are loads of special interest because of result from them

the stress resultants necessary for the design of cylindrical shells that constitute the

silo walls.

The above-mentioned pressures are commonly referred to as static loads and there

are several theories available for their estimation. In the present work, the Jansen’s

theory of silo loads [1] detailed in Sect. 2.2.1—was used.

In silos design, it is also necessary to consider the pressure increases due to the

unloading process the ensiled material. During this process, the pressure gradient is

generally very high and the final pressures to be considered in the design often exceed

the calculated values for the resting material—static pressures. This increase in pres-

sures is sometimes referred to as a dynamic effect, although the term overpressure

is more commonly used in the relevant literature.

Several factors influence the occurrence of overpressures in silos. Among these,

the following stand out:

• Arching of ensiled material;

• The collapse of arches formed by ensiled material;

• A sudden change in the type of flow (funnel/mass).

2.2 Loading 7

There are, currently, two ways of estimating the final pressures to use in silo

designs. One of them is the direct calculation of the total pressures, using existing

theories available in the literature [2–10].

Another way is to multiply static pressures by overpressure coefficients [10],

which take into account the dynamic effects of loading and unloading process. In

the approaches carried out throughout the present work, the calculation of the design

pressures was made using the overpressure coefficients indicated by the American

Concrete Institute [11].

The pressures developed inside a silo are closely related to the physical properties

of the stored material, and the following properties can be highlighted as being of

particular importance:

• Specific weight;

• The coefficient of friction wall/stored material;

• Internal angle of friction of the stored material or angle of repose.

To establish numbers to represent each one of the properties reported above is a

very important task because the magnitude of the pressures exerted by the ensiled

material on the walls and bottom of the cells is strongly dependent on these values.

In design practice, it is common to use tables, available in the literature, that

suggest a range of variation to be observed. However, it is important to consider that

the values presented in these tables are approximate values and, therefore, should

be used with care and. Whenever uncertainties remain regarding the values to be

adopted the most recommended procedure is to carry out laboratory tests to obtain

properties desired. In the present work, the values of the physical parameters of the

stored material were the following, related to an existing structure.

• Specific weight = 8.1 kN/m3

• Coefficient of friction wall/stored material = 0.35

• Internal angle of friction of the stored material or angle of repose = 34°.

The value of the specific weight of the ensiled material was adopted according to

practical experience and the other values were obtained in the recommendations of

the American Concrete Institute [11].

The theory presented by Janssen in 1895 was the first to take into account the friction

of the ensiled material and the silo wall. It is based on the balance of vertical forces

acting on an infinitesimal horizontal layer of ensiled material, as shown in Fig. 2.2.

The theory considers that the vertical pressure in a given horizontal cross-section

is constant and the relationship between the horizontal and the vertical pressure is

expressed by the coefficient K constant for a given depth.

By verifying the equilibrium of the vertical forces acting on that layer, it is possible

to obtain:

8 2 Geometry and Loading

dq

q A + γ Ady − A q + dy − u pU dy = 0 (2.1)

dy

where q is the vertical static pressure, A is the cross-sectional area, γ is the specific

weight of the ensiled material, p is the horizontal static pressure, U is the cross-

sectional perimeter and μ is the coefficient of friction between the ensiled material

and the silo wall.

Simplifying Eq. (2.1), substituting p for Kq, taking into accounts that the rela-

tion (U/A) is the inverse of the hydraulic radius—R—and carrying out the pertinent

separation of variables, one gets at the following expression:

dq

dy = (2.2)

γ − u RK q

integration, that for y = 0 we have q = 0. Thus, we got the expression that relates the

vertical pressure to the depth (see Eq. 2.3).

R γ

y = ln (2.3)

uK γ − u RK q

2.2 Loading 9

Finally, solving the Eq. (2.3) to find q, we obtain the expression of the vertical

pressure of the Janssen’s Theory, as it follows:

Rγ

− u RK γ

q= 1 − e (2.4)

u K

The equivalent horizontal pressure is given by the expression p = Kq:

Rγ

− u RK γ

p= 1 − e (2.5)

u

In Janssen’s theory there is no indication about the values to be adopted for the

variable K, usually referred to as the pressure coefficient. In the present work, Eq. (2.6)

was used.

1 − sin ρ

k= (2.6)

1 + sin ρ

where ρ is the angle of internal friction of the material. The vertical frictional force

was evaluated using Eq. (2.7).

V = (γ Y − 0.8q)R (2.7)

Equations (2.6) and (2.7) are included in the recommendations of the American

Concrete Institute [11] and the parameters need to calculate the static pressures and

the vertical friction force are presented in Sects. 2.2 and 2.2.1 of present work.

To calculate the hydraulic radius of the circular cells, the recommended value of

ACI [11] was used and for the interstice space, the procedure adopted was to consider

the hydraulic radius of a square of equivalent area.

For the calculation of the overpressures, the recommendations of the American Con-

crete Institute [11] were followed. This code recommends that the design forces and

pressures, resulting from the action of the ensiled material, must be evaluated for

silos with concentric discharge using Eqs. (2.8) and (2.9) as it follows.

qdes = Cd q (2.8)

pdes = Cd q (2.9)

10 2 Geometry and Loading

The maximum friction force on the wall occurs simultaneously with the minimum

vertical static pressure, and then, in concrete silos, we must adopt the Eq. (2.10) to

calculate it.

Vdes = V (2.10)

of the design pressures are those given by the American Concrete Institute [11]. The

numbers presented are minimum values recommended, however smaller values may

be adopted if the designer can attest that such values are satisfactory.

Figure 2.3 shows the variation of horizontal design pressure along the height of

the silo is shown for both cell and interstice. These values correspond to an existing

structure, with 10 m in diameter and 30 m in height, that is investigated in detail in

this study.

References

Ingenieure, pp 1045–1049

2. Caquot A, Kerisel J (1956) Traite Mecanique des Sols. Gauthier Villars Editeur

References 11

3. Pieper K, Wenzel F (1964) Druckverhaltnisse in Silozellen. Verlag von Wilhelm Ernst and

Sons

4. Geniev GA (1958) Voprosi Dinamiki Siputchei (Questions of the dynamics of granular mass).

Government Publication of Literature on Construction and Architecture, Moscow, URSS

5. Platanov PN, Kovtum AP (1959) Davlenie Zerna na Stenki Silosov Elevatorov, Mukomolno

Elevatornaia Promyshlennost

6. Theimer OF (1970) Betrachtungen Uber Druckverhaltnisse in Silozeneb. Deutsche Muller-

Zeitung

7. Walker D (1966) M. An approximate theory for pressures and arching in hoppers, Chem Eng

Sci J, pp 975–997

8. Reimbert M, Reimbert A (1980) Pressures and overpressures in vertical and horizontal silos. In:

International Conference on Design of Silos for Strength and Flow, Powder Advisory Center

9. Jenike AW (1977) Construction of concrete silos. Technical Report, Norwegian Society of

Chartered Engineers

10. Safarian SS (1969) Design pressure of granular material in silos. ACI J Proc, 539–547

11. ACI Committee 313 (2011) Standard practice for design and construction of Concrete silos

and stacking tubes for storing granular materials (ACI 313-11). American Concrete Institute,

Detroit

Chapter 3

Two-Dimensional Linear Elastic

Analyses—Slice Method

The initial procedure used in the structural analysis of reinforced concrete silo groups

is to remove a horizontal slice of unit height from the group of cells, as shown in

perspective in Fig. 3.1, and study it for the various expected loading cases.

Such a procedure is usually referred to as Slice Method and represents one of the

simplified processes available for the design of silo groups. The resulting structural

model is then, subjected to a two-dimensional analysis using computational matrix

analysis methods [1] or simplified manual methods [2]. In the work, numerical analy-

ses were performed using the finite element method taking into account the following

modelling approaches:

• Modelling with 2D frame elements;

• Modelling with plane strain elements.

From these analyses, one can obtain the stress resultants—horizontal axial forces

and bending moments—along all cross cross-sections of the silo group necessary

F. A. N. Silva et al., Design of Reinforced Concrete Silo Groups, Building Pathology

and Rehabilitation 10, https://doi.org/10.1007/978-3-030-13621-5_3

14 3 Two-Dimensional Linear Elastic Analyses—Slice Method

to design the cylindrical shells that comprise silo walls. Eleven loads cases were

considered, representing all possible combinations of load cases indicated in Fig. 3.2.

The applied pressures values are those indicated in Fig. 2.3 and the slice studied

was loaded with the maximum pressure values for the cell and the interstice, as

indicated in the figure.

The mesh idealized for the analysis of the silo groups with frame 2D elements

is presented in Fig. 3.3. Red lines linking nodes represent the region common to

adjacent cells (Detail A in Fig. 3.3). The origin of the global coordinate system is

located in the centre of the interstice space.

Each silo was modelled with 72 2D frame elements each one corresponding to

an arc length of five degrees. The height of the analysed slice was 1 m. The linking

region between two contiguous cells was treated as a rigid region, modelled as rigid

lines, interconnecting the nodes of the cell.

The generation of the rigid lines was implemented by imposing special Constraint

Equations—Sect. 3.3. These equations establish relations between degrees of free-

3.1 Modelling Using 2D Frame Elements 15

dom of the nodes of the elements that compose the rigid region, in order to allow the

consideration of its stiffness.

In the two-dimensional space of this analysis, three constraint equations are gen-

erated for each pair of constrained nodes. These equations define the three rigid

body motions in the global Cartesian System, that is, two plane translations—ux and

uy —and an out-of-plane rotation—rotz .

In addition, a minimum number of boundary conditions were imposed to suppress

rigid body motion of the structure. Such boundary conditions were implemented

through the imposition of prescribed zero displacements to certain nodes of the

model.

There are two usual concepts in the Theory of Finite Element Method associated

with the conversion of distributed load acting on the elements in nodal load, namely:

• Consistent loads;

• Non-consistent loads.

16 3 Two-Dimensional Linear Elastic Analyses—Slice Method

The first concept is associated with the conversion of distributed load into nodal

forces and moments. The loads, which come from this procedure, are referred to as

consistent loads or equivalent nodal loads, since the resulting forces and moments

in a point are equivalent to those arising from the originally applied distributed load.

The other concept presented relates to the conversion of distributed applied loading

in nodal force without moment. The resulting loads are generally referred to as non-

consistent loads also called as lumped loads. Such a load is preferable in straight

elements representing curves whose the rotation is a degree of freedom, since the

presence of spurious bending moments, which are not beneficial to the required

solution quality, is eliminated [3]. In the present work, the applied pressures were

converted into non-consistent nodal forces.

CD − Q = 0 (3.1)

DOF in D), D is the vector of nodal displacements of the model and Q is the vector

of constants.

In the present study, it was necessary to generate equations to allow the con-

sideration of the stiffness of the region that links two adjacent cells. To do this let

examine Fig. 3.4, where it is represented the finite element mesh of two adjacent

cells, with emphasis on the nodes—i and j—interconnected by a fictitious rigid line

whose existence is to be simulated.

The equations required to generate the rigid line in Fig. 3.4 are as follows:

of connection between two

cells of the silo group

3.3 Constraint Equations 17

uj − ui = 0 (3.2)

vj − vi − 2Lθi = 0 (3.3)

θj − θi = 0 (3.4)

in the y direction and θ i is the rotation of node i around z-axis.

In matrix notation, Eqs. (3.2)–(3.4) can be written as it follows:

⎡ ⎤ ⎡ ⎤

−1 0 0 100 0

⎣ 0 −1 −2L 1 0 0 ⎦ − [D] = ⎣ 0⎥ ⎢

⎦ (3.5)

0 0 −1 0 0 1 0

in which,⎡ ⎤

−1 0 0 100 T

C = ⎣ 0 −1 −2L 1 0 0 ⎦ and D = ui vi θi uj vj θj with vector Q being

0 0 −1 0 0 1

equal to zero.

There are two procedures widely used to implement constraint equations to a

finite element model, namely:

• Lagrange Multiplier Method;

• Penalty Method.

The Lagrange Multiplier method is the way in which the finite element program

used implements the mentioned equations [4].

The analysis of an engineering problem using the Finite Element Method results in

the solution of simultaneous algebraic equations, expressed in matrix form according

to the following equation:

Ku = F (3.6)

The aim is to impose to the system in Eq. (3.6) constraints equations expressed

as follows:

C T u = C0 (3.7)

where K is the stiffness matrix, u is the displacement vector, F is the nodal forces

vector and C is the coefficients matrix.

18 3 Two-Dimensional Linear Elastic Analyses—Slice Method

The Eqs. (3.6) and (3.7) can be written in the following way:

L

Kkj uj = Fk (3.8)

j=1

L

Cj uj = C0 (3.9)

j=1

with K = 1, …, L. Taking into account that L is the dependent of freedom that one

wishes to eliminate, it can be written that:

L−1

uL + Cj∗ uj = C0∗ (3.10)

j=1

in which

Cj C0

Cj∗ = and C0∗ = with k = 1, . . . ,L − 1 (3.11)

CL CL

The system represented by Eq. (3.6) is the necessary and sufficient condition for

the minimization of the functional Total Potential expressed as follows:

1

= uT Ku − F T u (3.12)

2

Therefore, the problem of imposing constraints can be presented in the following

form:

1

Min uT Ku − F T u (3.13)

2

with C ∗T u = C0T .

It is a problem with linear equality constraints, whose necessary optimality con-

ditions are:

• Feasibility;

• Stationarity.

The feasibility condition indicates that the solution of the problem formulated

in Eq. (3.13) must satisfy the imposed constraints and the stationarity condition

expresses the necessity of the existence of Lagrange Multipliers, as many as the

constraints, such that:

l

∇f u∗ + λi hi u∗ = 0 (3.14)

i=1

3.4 Lagrange Multiplier Method 19

where f is the Functional, u* is the solution vector, l is the internal number of equality

constraints, λ is the Lagrange multiplier and h is the Equality constraints.

Defining the Lagrangian Function as:

1

L(u, λ)= uT Ku − F T u + λ C ∗T u − C0∗ (3.15)

2

The above-mentioned associated optimality conditions can be written as follows:

∇λ L = 0 (3.16)

∇u L = 0 (3.17)

K C∗ u F

∗T = (3.18)

C 0 λ C0∗

Since the intention is to eliminate the degree of freedom L, the static condensation

of the above system for uL and λ is performed and one can write that:

⎡ ⎤ ⎡ u⎤ ⎡ F ⎤

K K∗L C ∗

⎣ K∗L KLL 1 ⎦ ⎢ ⎥ ⎢ ⎥

⎣ uL ⎦ = ⎣ FL ⎦ (3.19)

C∗ 1 0 λ C0∗

K K∗L ∗

C∗ F u

in which K = ;C = ;F = and u = , with K *L

K∗L KLL 1 FL uL

representing the column L of matrix K.

Eliminating the second and the last columns of the matrix in Eq. (3.19), we obtain

the matrix equation that follows:

⎡ ⎤⎡ ⎤

Kkj − Cj∗ KkL − Ck∗ KLj − Cj∗ KLL 0 0 u

⎢ ⎥⎢ ⎥

⎣ ∗

KLj − C KLL j 0 1 ⎦⎣ L ⎦

u

C∗ 10 λ

⎡ ∗ ∗

∗

⎤

Fk − C0 KkL − C0 FL − C0 KLL

=⎣ FL − C ∗ KLL 0

⎦ (3.20)

C0∗

This way, the final system to be solved, with the consideration of the desired

constraint equations, becomes:

20 3 Two-Dimensional Linear Elastic Analyses—Slice Method

L−1

Kkj∗ uj = Fk∗ (3.21)

j=1

with k = 1, …, L − 1; Kkj∗ = Kkj − Cj∗ KkL − Ck∗ KLj + Ck∗ Cj∗ KLL and Fk∗ = Fk −

C0∗ KkL − Ck∗ FL + Ck∗ C0∗ KLL .

The finite element mesh used to model the slice of the silo groups with plane strain

element is shown in Fig. 3.5. Detail A in this figure exhibits the level of refinement

adopted in the region of connection between two cells.

Each silo was modelled with 256 plane strain elements. Along the thickness of

the walls, the mesh has two elements. Two meshes were tested: one with four-node

iso-parametric two-dimensional solid elements (Q4) and other with eight-node iso-

parametric two-dimensional solid elements (Q8). These elements are available in

the ANSYS finite element library [4] and are referred to as Plane42 and Plane82,

respectively.

Fig. 3.5 Mesh of the slice of the silo group using plane strain finite element

3.5 Modelling with Plane Strain Finite Element 21

In both hypotheses considered, the angle describing the lengths of the elements

was five degrees. As well as the modelling with frame elements, prescribed displace-

ments were imposed on certain nodes of the model to avoid rigid body motion. In

the analyses developed with solid elements type Plane42, additional shape functions

were used. The detail of such type of interpolation function is discussed in next.

Figure 3.6 shows the geometry of a typical Q4 element with an indication of the

global Cartesian system (x-y) and the natural coordinate system (t-s).

The shape functions for the element shown in Fig. 3.6 are given below, defined in

natural coordinates:

1

Hi (t,s) = (1 − t)(1 − s) (3.22)

4

1

Hj (t,s) = (1 − t)(1 + s) (3.23)

4

1

Hk (t,s) = (1 + t)(1 + s) (3.24)

4

1

HL (t,s) = (1 + t)(1 − s) (3.25)

4

where H i (t,s) is the Lagrangian function associated to node I, in natural coordinates.

The displacement field inside the element is then interpolated as follows:

element—Q4

22 3 Two-Dimensional Linear Elastic Analyses—Slice Method

1 ui (1 − t)(1 − s) + uj (1 − t)(1 + s)

u= (3.26)

4 +uk (1 + t)(1 + s) + uL (1 + t)(1 − s)

1 vi (1 − t)(1 − s) + vj (1 − t)(1 + s)

v= (3.27)

4 +vk (1 + t)(1 + s) + vL (1 + t)(1 − s)

The deformed geometry of a Q4 type element is indicated in Fig. 3.7, with the

sides deforming as straight lines.

The additional shape functions modify this behaviour of the Q4 elements, allowing

a parabolic deformation along its sides, which is important in problems with bending.

Additional Shape Functions are obtained by adding terms to incorporate modes

that describe a state of constant curvature [3] to the functions that interpolate the

previously defined displacements.

The displacement expansions thus characterized are expressed by:

1 ui (1 − t)(1 − s) + uj (1 − t)(1 + s) + uk (1 + t)(1 + s)

u= (3.28)

4 +uL (1 + t)(1 − s) + u1 1 − s2 + u2 1 − t 2

1 vi (1 − t)(1 − s) + vj (1 − t)(1 + s) + vk (1 + t)(1 + s)

v= (3.29)

4 +vL (1 + t)(1 − s) + v1 1 − s2 + v2 1 − t 2

Since the above expansions contain six shape functions, Q4 elements with Addi-

tional Shape Function are sometimes referred as Q6 elements in finite element theory

books. In the present work, the Q4 element used was the ANSYS four-node linear

solid [4], with the option of adding the Additional Shape Functions.

Q4 element without

additional shape function

3.7 Calculation of Stress Resultants 23

The stress resultants—bending moments and normal horizontal forces—in the vari-

ous sections of the silos were calculated taking into account two important properties

of an analysis performed using the Finite Element Method, which are independent

of the refinement of the mesh considered [5]:

• At a given node, the sum of the nodal forces of all the elements sharing this node

is in equilibrium with the applied external loads and

• Each element is in equilibrium with respect to its own nodal forces.

It follows from the two properties mentioned above that the nodal forces obtained

from the finite element analysis are always in global equilibrium with the applied

loads and, thus, the calculation of the stress resultants computed from these forces

ensures the equilibrium in each cross section analysed.

Figure 3.8 indicates the data required for the calculation of the stress resultants.

In this figure, two elements of the mesh located along the wall of the silo are showed,

both inclined at an angle α with respect to the horizontal. Nodes 1, 2 and 3 are also

indicated with the respective nodal forces, which define the section where the stress

resultants of interest will be calculated.

calculation of stress

resultants

24 3 Two-Dimensional Linear Elastic Analyses—Slice Method

The comparison between results obtained in the analyses is presented. The reference

number of the indicated load case corresponds to that shown in Fig. 3.2.

The following figures show the deformed geometry of the silo groups studied for

each of the load cases analysed, considering the two types of modelling strategies

already discussed.

These deformed geometries allow to observe that the structural behaviour of the

silo group, for all the Load Cases, are quite similar, either modelling the slice with

frame elements or with plane strain elements. It can also be observed that, in the

case of Load Case VI—Load on the interstice—the deformed geometry assumes

a configuration that causes the emergence of high values of the stress resultants,

especially the bending moments (see Sect. 3.8.3). Such situation does not reflect the

actual structural behaviour of the silo groups analysed, since in the Slice Method the

structure is considered plane, without any influence of the bottom and top slices that

comprise the slice studied. Such a consideration disregards a very important aspect

the loading-carrying mechanism of this type of structure—i.e. the three-dimensional

interaction between the various cells of the group (Figs. 3.9, 3.10, 3.11, 3.12, 3.13

and 3.14).

3.8 Results and Discussion 25

The values of the normal horizontal forces along the sections located in the walls

of the silos remained constant, for load in the cell, in a value of 350 kN/m (tensile),

in all modelling that were used. This value was the maximum verified for all Load

Cases analysed and it always occurred in the cross sections of the loaded cells.

26 3 Two-Dimensional Linear Elastic Analyses—Slice Method

For the load acting on the interstice, the maximum horizontal normal force was

of 88 kN/m (tensile) and occurred in the wall common to the cell and interstice.

Figures 3.15, 3.16, 3.17 and 3.18 represent the bending moments in several cross

sections of a silo, considering all the Load Cases analysed and the three modelling

3.8 Results and Discussion 27

approach considered. In these figures, the average surface of the silos wall is plotted

in black colour with divisions indicating the angular position of the several cross sec-

tions studied. As complements to the each figure, tables summarizing the maximum

bending moment values are shown, both positive and negative.

The sign convention adopted was that positive bending moments elongate the

inner fibres of the silo wall and ls and negatives bending moment are those that

elongate the outer fibres of the silo wall.

Negative bending moments are plotted outward of the wall of the silos and positive

ones inwards.

As shown in the figures, the values of the bending moments vary throughout the

studied sections and one can observe that there are sections with positive moments

and sections with negative moments.

It was observed, in addition, that the maximum values of the bending moments

occurred, in all Load Cases and for all models investigated, in the cross sections

immediately adjacent to the region of interconnection between two cells.

As well as the deformed geometries, it is observed that the behaviour of the

diagrams of bending moments is quite similar for the three modelling approaches

studied. This fact indicates that the methodology for calculating stress resultants from

the nodal forces—used in the modelling with plane strain elements—produced very

satisfactory results, aspect that highlights the efficiency of the adopted procedure and

beyond demonstrate a great practical interest in the application of the Finite Element

Method in the analysis and design of reinforced concrete structures.

28 3 Two-Dimensional Linear Elastic Analyses—Slice Method

3.8 Results and Discussion 29

30 3 Two-Dimensional Linear Elastic Analyses—Slice Method

It is further noted that the values of the bending moments for the model with Q4

elements are practically identical to those of the Q8 elements, thus indicating the

good performance of the additional shape functions employed in the Q4 elements.

This fact signals to the possibility of using linear solid element with Additional Shape

Functions in the three-dimensional analysis of the silo groups studied, presented in

the next chapter.

A behaviour divergence was observed between the modelling with frame elements

and plane elements in the Load Case IV, motivated by the difference of stiffness of

the two models in the region of interconnection of two cells of the group. On the

other hand, it was also observed that, in spite of the divergence observed, the resulting

bending moments for this Load Case are very small when compared with the other

Load Cases, in all the modelling approaches studied.

3.8 Results and Discussion 31

One can observe, also, the high magnitude of the bending moments for the Load

Case VI, compared with other Load Cases studied. Such values come from the fact

that the wall of the interstice is not behaving like an arc, due to the displacements of

the points of intersection between two cells. Such displacements are the responsible

for the generation of the high and unreal bending moment’s values that do not reflect

the true behaviour of the structure, due to the applied loading.

As for nodal displacements, the values of the bending moments obtained using

the Slice Method analysis for all Load Cases studied do not adequately characterize

the structure response and an approach that considers tri-dimensionality is really

required.

32 3 Two-Dimensional Linear Elastic Analyses—Slice Method

Finally, Figs. 3.19 and 3.20 represent bending moment’s envelope for one of the

cells, considering the modelling with Q8 elements and for frame elements, for all

Load Cases studied, with the exception of Load Case VI, due to the impossibility of

visualizing the values of the bending moments for this case vis-à-vis the others.

References 33

References

1. Stalnaker JJ, Harris EC (1992) Bending moments in walls of grouped silos due to structural

continuity. ACI Struct J 89(2):159–163

2. Filho JF, An introduction to the study of silos. Master Thesis, Engineering School of Engineer,

São Paulo University (in Portuguese)

3. Cook RD (1994) Finite element modelling for stress analysis. Willey

4. Ansys (2014) User and theoretical manuals. Swanson Analysis Systems, Inc

5. Bathe KJ (1996) Finite element procedures. Prentice Hall

Chapter 4

Three-Dimensional Analysis

4.1 Introduction

The study of reinforced concrete silo groups using the three-dimensional finite ele-

ment analysis demands from the designer hard work to formulate of the problem.

Aspects such as the characterization of the geometry, the determination of the loads

and choice of the appropriate boundary conditions, as well as in the interpretation of

the obtained results, are examples such difficulties. Despite these difficulties, three-

dimensional analysis is an important tool for understanding the structural behaviour

of this type of construction.

Regarding the geometric representation, the choice of a finite element mesh that

adequately incorporates the characteristics of the physical model is an important

issue and deserves a special attention.

The research considered three different modelling approaches, which are the usual

ways of modelling of this type of structure, namely:

1. Modelling with solid elements;

2. Modelling with shell elements;

3. Modelling with solid and shell elements.

The Model 1 and 2 are widely adopted in design of silo groups and Model 3 is less

frequent. Model 1, which uses less simplified approaches regarding the structural

behaviour of the silo group, will be considered as the benchmark model for the

comparisons between the other two modelling approaches.

Each one of the modelling strategies studied will be presented in details and the

comparisons performed are deeply discussed.

In all modelling approaches, the diameter of the circular cells used was 10 metros

and the height adopted for the silo group was 30 m. The finite element package used

was ANSYS—Engineering Simulation & 3D Design Software.

F. A. N. Silva et al., Design of Reinforced Concrete Silo Groups, Building Pathology

and Rehabilitation 10, https://doi.org/10.1007/978-3-030-13621-5_4

36 4 Three-Dimensional Analysis

demands more than to study all the possible Load Cases, but the identification of

which load combination lead to the highest stress resultants values necessary to the

design and detailing of the structure.

In the analysis of the silo group investigated, one can identify critical situations

in some cross sections along the geometry of the group such as [1]:

• Maximum horizontal tensile normal force in the wall of one cell in a silo group

happens when this cell is fully loaded and no increase of this force takes place

when the others cells of the group are also loaded. Furthermore, this force came,

usually, associated with some flexion—bending moments;

• The wall common to the cell and the interstice may be subjected to compression,

with maximum value occurring when only the interstice is loaded;

• The maximum values of the bending moments occur when the interstice is fully

loaded and the other cells are empty, in cross sections located in the wall common

to the cell and the interstice. In cross sections at the ends of the wall common to

two contiguous cells—cross sections 1 and 4 of Fig. 4.1, where the larger values

of the bending moments appear.

From the two-dimensional analyses performed previously, it was possible to con-

clude that in the three-dimensional analysis of reinforced concrete silo groups must

consider at least two specific loading situations, namely:

• Loading applied to the cells;

• Loading applied to the interstice.

The combination of these two load conditions will allow the calculation of the

stress resultants value in the cross section indicated in Fig. 4.1.

Taking into account the shape and loading symmetry of the structure, one can use

only one-eighth of it to generate the models for the three-dimensional analysis.

three-dimensional

analysis—cross sections of

interest

4.1 Introduction 37

the emphasis was given to the cross sections indicated in Fig. 4.1, where the highest

values occur. Cross sections 1 and 4 are of interest because they lie in regions with

geometric singularity and cross sections 2 and 3 are chosen because they are located

over one of the symmetry planes of the structure.

The sign convention used to represent the negative and positive bending moments

in the several figures presented below was the same one adopted in the bi-dimensional

analysis, as it was discussed in the previous chapter.

For the horizontal normal forces, the sign convention adopted to their represen-

tations are: compression forces are plot outward the silo wall and tensile forces are

plotted inward. The boundary conditions used are discussed in Sect. 4.2.

• In the bottom, taking into account the rigidity of the existing slab as well as the

connection of the wall with the foundation, the boundary condition imposed was

the clamped support;

• At the top, considering the existing slab acting as a rigid diaphragm for membrane

loads allied to the structure symmetry, the boundary condition imposed was simple

support.

The use of shape and loading symmetry conditions of the structure generates two

planes of symmetry on which appropriate boundary conditions have been imposed

[2].

Finite element mesh, with linear solid elements, was used to model for three-

dimensional analysis of the structure is shown in Fig. 4.2.

The element used was the iso-parametric tri-dimensional solid of ANSYS [3].

Eight nodal points define this element, each one having three degrees of freedom—-

translation in x, y and z directions. Additional shape functions were also include

improving the performance of such finite element, in the same way as it was carried

out in the plane analyses previously discussed.

In this modelling, one-meter high elements were used and, like in the 2D mod-

elling, a division with two elements were considered along the thickness of the walls.

The mesh refinement at the intersection region between two cells was the same

as that used for the 2D modelling, discussed in Chap. 2, and the angle describing the

element length was a maximum of seven and a half degrees.

38 4 Three-Dimensional Analysis

The model resulting from this modelling approach has 4774 nodes and 3180 ele-

ments and, considering the boundary conditions imposed at the bottom, at the top and

at symmetry planes of the structure, one has a system with about 13,000 equations.

The stress resultants in cross sections of interest were calculated using the same

procedure adopted for the modelling with plane finite elements.

The results of the analysis performed for the loading applied in the cell and in the

interstice are discussed in the following section.

Figure 4.3a shows the diagram of normal horizontal forces in several cross sections

of the silo wall, for a given transverse slice. This is a typical diagram for any cross-

section of the structure in which are represented the silo wall and the values of the

horizontal axial forces in several perimeter sections—graphically represented in red

colour.

4.3 Modelling with Linear Finite Element 39

Fig. 4.3 a Horizontal normal forces—loading in cell and b variation of the maximum normal

horizontal forces

One can observe tensile forces in all sections, with a maximum value of

370.5 kN/m occurring in the section located at 67.5° with the horizontal. Such value

was the maximum one for the whole structure.

Figure 4.3b shows the variation of the maximum normal horizontal forces in each

cross section of the silo, along the height of the structure. It can be observed that there

is a growth in the value of the normal force until its maximum value (370.5 kN/m)

which occurs in the cross section located at 28 m of depth. From this depth on, there

is a decrease in its value, which reaches 96.1 kN/m in the cross section located at the

bottom of the cell.

Regarding the bending moments, Fig. 4.4a shows the variation of the bending

moments Mz —longitudinal direction—for several perimeter sections of the silo wall.

Similarly to normal horizontal forces, the bending moment diagram shown in

Fig. 4.4a is typical for the other cross sections of the silo group. One can observe

that there are parts with positive bending moments and parts with negative ones.

The highest value occurs at the lower end of the common wall to two contiguous

40 4 Three-Dimensional Analysis

Fig. 4.4 a Bending moment Mz —loading in cell and b maximum bending moments Mz along the

height

cells—cross section 4 of Fig. 4.1. In addition, Fig. 4.4b shows the behaviour of the

maximum bending moments Mz along the height of the silo group.

It should be observed that the behaviour of the diagram in Fig. 4.4b is similar to

that presented for normal horizontal forces. In fact, there is an increase in the value

of the bending moment until its maximum value (15.27 kNm/m) is reached in the

cross section located, identically, to 28 m of depth. From this point on, a decrease in

value of bending moments begins to occur until, in the cross section located at the

bottom of the cell, its value reaches 4.37 kNm/m.

In addition, it is important to note that the maximum bending moment occurred in

the cross section close to the maximum value of the applied horizontal pressure—see

Fig. 2.3 in Chap. 2.

4.3 Modelling with Linear Finite Element 41

Fig. 4.5 a Normal horizontal forces—loading in the interstice—Section 3 and b variation of the

maximum normal horizontal forces

Figure 4.5a shows the normal horizontal forces in several perimeter sections. The

diagram presented in this figure corresponds to the cross section located at 28 m in

depth, in which the maximum value of the normal horizontal force of –159.49 kN/m

(compression) was observed, occurring in section 3 of Fig. 4.1.

This compression arises because the wall common to the cell and interstice is

functioning as an arch, whose supports are the regions common to two contiguous

cells. As a function of the three-dimensional effect, displacements restraints of the

arch supports generate compression force in that wall. This behaviour, when com-

42 4 Three-Dimensional Analysis

pared to the study presented in the previous chapter, exposes the divergence between

the two analyses and shows the inefficiency of the two-dimensional approach for the

loading applied in the interstice.

Figure 4.5b shows variation of the maximum normal horizontal forces in Section 3

(symmetry plane) along the height of the silo group. It is possible to see that part

of the cell height is in compression and part is in tension. In fact, up to a depth of

18 m the silo wall is in tension with the maximum value of 21.48 kN/m occurring in

the cross section located at 12 m in depth. This behaviour shows that in this region

the elastic restriction to the displacement prevails to the detriment of the applied

loading. From 18 m depth on, the silo wall is in compression, reaching its maximum

value of –159.49 kN/m in the cross section located at 28 m in height. Such behaviour

indicates that in this region the action of the applied loading prevails.

Regarding the bending moments, Fig. 4.6a shows the variation of the moment

Mz in several perimeter sections, for a slice located at 17 m in depth. In this cross

section, the highest values of negative and positive bending moments were observed

for the whole structure, –52.25 and 31.691 kNm, respectively. These values occurred

in sections 3 and 4 of Fig. 4.1.

Figure 4.6b exhibits the variation of the maximum positive and negative bending

moments along the height of the silo group. It should be noted, additionally, that, in

contrast to the loading applied in the cell, the maximum bending moment does not

occur close to the cross section where the applied horizontal pressure is maximum.

The mesh resulting from the discretization of the structure with shell elements is

shown in Fig. 4.7. The cells were modelled with thin shell elements developing along

the average surface of the walls. The characterization of the interconnecting region

between two contiguous cells was made using appropriate boundary conditions in

the symmetry plane that contains such region.

The ANSYS quadrilateral flat shell element was used with Additional Shape Func-

tions. Four nodal points each one having six degrees of freedom—three translations

and three rotations, define this element. The considered element combines membrane

and bending behaviour in its displacement field.

It should be noted in Fig. 4.7 the refinement of the mesh close to the section that

delimits the intersection of the silo wall and the region of interconnection of the two

cells—sections 1 and 4 of Fig. 4.1. Three elements of equal length were arranged,

in order to get stress resultants, using quadratic extrapolation of the values provided

for the centre of the elements, from the finite element analysis.

The model resulting from this discretization has 1147 nodes and 1080 elements

generating a system with 5500 equations, approximately.

4.4 Modelling with Shell Elements 43

Fig. 4.6 a Bending Moments Mz —loading in interstice and b variation of the maximum positive

and negative bending moments

44 4 Three-Dimensional Analysis

For the loading applied in the cell, it was found that the diagrams of horizontal ten-

sile forces and bending moments Mz exhibited the same behaviour as those indicated

in Sect. 4.3.1, and the obtained values were 363.45 and 19.49 kNm/m, respectively.

Similarly for the loading applied in the interstice, the diagrams of the normal

horizontal forces and the negative and positive bending moments Mz obtained are

of the same pattern as those indicated in Sect. 4.3.2 with values of –151.05, −55.11

and 27.62 kNm/m, respectively.

It is a mixed model, where the silos walls were modelled with thin shell elements and

the region of interconnection between two cells was modelled with three-dimensional

solid elements, as shown in Fig. 4.8.

In this model, it was necessary to impose constraint equations to correctly establish

the connection between the two types of elements used in the model, since the shell

elements have six degrees of freedom per node and the solid elements have only

three. This situation occurs in sections 1 and 4 of Fig. 4.1.

4.5 Modelling with Solid and Shell Elements 45

Figure 4.9 shows details of this connection, indicating its geometry, non-deformed

and deformed shape, for the loading acting in the interstice.

Figure 4.9b, c show three generic nodes of the model—i, j and k—that will

be used in the generation of the constraint equations that define the rigid regions

interconnecting them. This procedure is similar to that presented in the 2D modelling

with frame elements discussed in the previous chapter.

The expressions in Eqs. (4.1, 4.2 and 4.3) define the equations required to generate

the rigid region connecting the nodes i and j:

t ◦

ui − u j − sin 20 θz j = 0 (4.1)

2

t

vi − v j − cos 20◦ θz j = 0 (4.2)

2

t t

wi − w j + sin 20◦ θx j − cos 20◦ θ y j = 0 (4.3)

2 2

where t is the thickness of silo wall, ui is the translation of node i in the x direction,

vi is the translation of node i in the y direction, wi is the translation of node i in the z

direction and θ zi is the rotation of node i with respect to z-axis.

46 4 Three-Dimensional Analysis

Analogous procedure must be done for nodes j and k. Figure 4.9d shows

the deformed shape—magnified for visualization effect—of the connecting region

between shell and solid elements, where the efficacy of the constraint equations

imposed can be observed.

The constraint equations previously related can be automatically generated in the

finite element analysis program used—ANSYS—through the CERIG command.

4.5 Modelling with Solid and Shell Elements 47

Regarding the level of refinement of the mesh, the same detail shown in Figs. 4.2

and 4.7 was adopted.

The mesh shown in Fig. 4.8 has 3410 nodes, 2580 elements and a system with

close to 11,000 equations.

For the two load cases considered, the diagrams of the normal horizontal forces

and bending moments Mz exhibited the same pattern as those indicated in Sects. 4.3.1

and 4.3.2, with the maximum values occurring in the same cross section location.

For the loading applied in the cell, the maximum horizontal tensile force was

363.62 kN/m and the maximum bending moment was 14.25 kNm/m.

In the case of the loading applied in the interstice, the maximum horizontal com-

pressive force was –160.96 kN/m and the negative and positive bending moments

were, respectively, –52.43 and 30.03 kNm/m.

Figures with the variation of the bending moments along the height of the silo group

for the sections of Fig. 4.1 and loading applied in the cell and in the interstice are

shown. (Figs. 4.10, 4.11, and 4.12).

One can observe that the modelling that combines elements of shell and solid

elements in the mesh—called mixed modelling—is the one that exhibited the closest

results to the model with only solid elements. Such behaviour can be observed in

the several figures presented and, in the case of loading case of special interest for

the present research—loading applied in the interstice -, the values of the bending

moments in the sections of the wall common to the cell and the interstice are quite

similar (see Fig. 4.13).

The results thus obtained indicate the possibility of using the mixed model in

finite element modelling in the parametric analyses, described in the next chapter.

48 4 Three-Dimensional Analysis

Fig. 4.10 Bending moments Mz in section 1 and 2—loading applied in the cell

4.6 Results and Discussion 49

Fig. 4.11 Bending moments Mz in section 3 and 4—loading applied in the cell

50 4 Three-Dimensional Analysis

Fig. 4.12 Bending moments Mz in section 1 and 2—loading applied in the interstice

4.6 Results and Discussion 51

Fig. 4.13 Bending moments Mz in section 3 and 4—loading applied in the interstice

52 4 Three-Dimensional Analysis

References

1. Safarian SS, Harris EC (1985) Design and construction of silos and bunkers. Van Nostrand

Reinhold Company, New York

2. Cook RD (1994) Finite element modelling for stress analysis. Willey

3. Ansys (2014) User and theoretical manuals. Swanson analysis systems Inc

Chapter 5

Structural Behaviour and Simplified

Model Proposition

The understanding of the load-bearing mechanism in the silo group analysed in this

research, due to the loading acting in the interstice, can be intuitively understood

observing the three figures presented below.

Figure 5.1(A-1) shows, in plant, a group of reinforced concrete silos where there

is no structural continuity between the various cells, nor a top slab. There is only the

provision of a material with plastic characteristics in the region of connection of two

cells, to avoid the leakage of the content deposited inside the interstice. The loading

coming from the stored material imposes to the silo a deformed shape quite similar

to a cantilever beam clamped in the bottom—Fig. 5.1(A-2). The clamp is ensured

by the stiffness of the bottom slab and by the foundation of the whole silo group.

Figure 5.1(B-1) represents the plant of a group of reinforced concrete silos with

concrete top slab, but still without structural continuity between the cells. Under the

loading action from the material stored in the interstice, the silo group assumes a

deformed configuration similar to a beam clamped in the bottom and simply sup-

ported at the top—Fig. 5.1(B-2). The simple top support is provided by the stiffness

of the top concrete slab on its own plane and by the symmetry of the structure.

Figure 5.1(C-1) indicates the plant of a group of silos where there is structural

continuity between the various cells as well as a top slab. Due to the action of the

loading in the interstice, the silo group deforms as if it was a beam on elastic founda-

tion, clamped at the bottom and simply supported at the top. The elastic foundation

represents the effect of continuity between the various cells, which manifests itself

through a restrain to the displacement of the wall of the interstice.

Figures 5.2 and 5.3, resulting from the three-dimensional analysis performed,

corroborate the intuitive interpretation of the loading bearing mechanism of the silo

group discussed above. In these figures, respectively, the deformed shape of the

interconnecting region between two cells, as well as the contour plot of the normal

stresses (longitudinal direction), are represented for a given cross section of the silo.

F. A. N. Silva et al., Design of Reinforced Concrete Silo Groups, Building Pathology

and Rehabilitation 10, https://doi.org/10.1007/978-3-030-13621-5_5

54 5 Structural Behaviour and Simplified Model Proposition

The deformed shape shown in Fig. 5.2 shows that the silo, in fact, behaves like a

beam (circular cross section of diameter D and vertical axis of length H), where the

effects of shear deformation have influence that needs to be considered (Timoshenko

beam theory).

Figure 5.3 further characterizes this behaviour, where one can observe parts in

compression and parts in tension in cross section, aspect that confirm the behaviour

predicted by the classical beam theory (Fig. 5.4).

especially the aspects related to the stress resultants and the load-bearing mechanism,

it is proposed to approach the analysis of reinforced concrete silos groups using the

following models:

1. A Timoshenko beam on elastic foundation to model the cylinders vertically (lon-

gitudinal bending);

2. A circular arch, clamped at both ends, subjected to transverse loading and

imposed displacement δ to model the interstice walls horizontally (transverse

bending).

5.2 Proposed Simplified Model 55

56 5 Structural Behaviour and Simplified Model Proposition

5.2 Proposed Simplified Model 57

calculating the spring

constant

area) e k (spring constant of elastic foundation)—are modelled by dimensionless

parameters α, β and γ that affect the values of I g , Ag and k t , respectively, as indicated

below:

I = a Ig (5.1)

Ag = β Ag (5.2)

k = γ kt (5.3)

where I g is the gross moment of inertia of the silo cross-section, Ag is the gross

cross-sectional area of the silo and k t is the theoretical spring constant of elastic

foundation.

The dimensionless parameters α, β and γ are functions of the geometric indices

of the structure (D/t and H/D). Their values will be evaluated through a parametric

study, where least square fitting of displacements of the Model 1 was applied to

fifteen different geometries of silo groups. The silo groups were analyzed with the

finite element method using the mixed model (with shell and solid elements) already

discussed in Chap. 4.

The gross moments of inertia and gross cross-sectional areas are obtained from

the silo geometry. The theoretical spring constant of the elastic base is calculated for

the flat structure modelled with frame elements, as indicated in Fig. 5.5.

58 5 Structural Behaviour and Simplified Model Proposition

In this figure, it is schematically represented the mesh with frame elements of the

silo group, in which unit displacements are applied to the nodes of the elements of

the regions of interconnection between two cells. The reaction forces—F—in these

nodes generate a resulting force that will be considered as the elastic foundation

reaction. The numerical value of the spring constant is given by:

2F cos 45◦

kt = √ (5.4)

2

Model 1 is presented above.

⎡ −12E I 6E I −12E I 6E I ⎤

L 2 (1+φ) L 2 (1+φ) L 2 (1+φ) L 2 (1+φ)

⎢ 6E I E I (4+φ) −6E I E I (2−φ) ⎥

⎢ L 2 (1+φ) L 2 (1+φ) ⎥

⎢ −12E I

L(1+φ)

−6E I

L(1+φ)

−6E I ⎥ (5.5)

⎣ L 2 (1+φ) L 2 (1+φ)

12E I

L 2 (1+φ) L 2 (1+φ) ⎦

6E I E I (2−φ) −6E I E I (4+φ)

L 2 (1+φ) L(1+φ) L 2 (1+φ) L(1+φ)

inertia Izz, ϕ = 12EI/GAs L2 , G = E/2(1 + ν) and ν is the coefficient of Poison.

Figure 5.6 shows the discretization used for the application of the Model 1.

The nodal forces applied (see Fig. 5.6) are those resulting from the load portion

of the interstice that acts in the silo wall, as shown in Fig. 5.7.

5.2 Proposed Simplified Model 59

were modelled with frame elements of an arch length of 5°. The pressure applied

on each arch was converted to nodal force at the respective node (Fig. 5.7) and the

nodal translations resulting from the use of the Model 1 are applied as prescribed

displacements in the supports of the clamped arches (interstice walls) as shown in

Fig. 5.4b. From the solution of Model 2 the stress resultants—bending moments and

normal forces—necessary to design the structure are obtained.

geometric configurations studied, was made using Nonlinear Least Squares Method,

as indicated in next expression.

n

2

min u mi − u ti (α, β, γ ) (5.6)

α,β,γ

i=1

60 5 Structural Behaviour and Simplified Model Proposition

where n is the number of nodes of the model, um is the nodal displacements vector

from the three-dimensional model and ut is the nodal displacements vector from the

theoretical model.

The displacements um are those resulting from the three-dimensional analysis

using the mixed model (model with shell and solid elements) and the displacements

uti are those calculated from the proposed Model √ 1. The displacements from the

three-dimensional analysis are multiplied by 2, since the designed beam model

develops in a plane located at 45° from the centre of the region of interconnection

between two cells (Sect. 4 of Fig. 4.1).

The solution of the minimization problem formulated above was obtained using

a procedure developed in Matlab [1], where the nodal displacements are calculated

for Model 1 and the optimization is performed with an internal least squares rou-

tine—leastsq, which uses the Levenberg-Marquardt algorithm.

The results obtained for a silo group with four cylindrical cells of 10 m in diameter

and 30 m in height are presented as an illustrative example of the proposed adjustment.

The data required for analysis are given below.

• Longitudinal Modulus of Elasticity—E = 21,000,000 kN/m2

• Transversal Modulus of Elasticity—G = 9,130,434.8 kN/m2

• Gross Cross Section Area—Ag = 6.7876 m2

• Moment of Inertia of the Gross Cross Section—Ig = 88.4368 m4

• Elastic Foundation Spring Constant—kt = 130,970.7 kN/m

Table 5.1 presents the nodal displacements (projected in the bending plane of the

cylinder—beam) of the three-dimensional analysis with the Finite Element Method

(FEM) and the displacements calculated using Model 1, with the dimensionless

parameters α, β and γ adjusted by minimum squares method.

Figure 5.8 (graphical representation of Table 5.1) shows that the nodal displace-

ment curve obtained with the use of the proposed Model 1 is quite similar to that

resulting from the three-dimensional analysis performed. Such behaviour attests to

the efficiency of the adjustment proposed for the calculation of the nodal displace-

ments. The values of the dimensionless parameters α, β and γ associated to the

displacements of the Model 1, adjusted by the least squares method, were 0.3513,

0.1464 and 0.4978, respectively.

Tables 5.2 and 5.3 show the values of the bending moment’s for the three-

dimensional analysis and those resulting from the use of the proposed analysis pro-

cedure—Models 1 and 2. Negative bending moments occur in Sect. 4 of Fig. 4.1 and

the positives ones in Sect. 3 of the same figure. The sign convention adopted is the

same as the three-dimensional analysis of Chap. 4.

The values shown in these tables are shown graphically in Fig. 5.9a, b. Figure 5.9a

shows the negative bending moments and Fig. 5.9b the positive bending moments.

In these figures, one can observe that the bending moments of the adjusted model,

to a large extent of the height of the silo, are higher than those resulting from the

three-dimensional analysis. In the case of negative bending moments, this difference

was 18% and it happened because there is a deformability of the arch supports that

is not considered in the model designed for the calculation of the bending moments.

5.3 Fitting with the Least Square Method 61

H (m) FEM Proposed model (mm) H (m) FEM Proposed model (mm)

(mm) (mm)

1 0.125 0.126 16 1.353 1.348

2 0.249 0.251 17 1.354 1.348

3 0.372 0.374 18 1.341 1.354

4 0.491 0.495 19 1.313 1.305

5 0.607 0.611 20 1.269 1.263

6 0.718 0.722 21 1.210 1.205

7 0.823 0.826 22 1.136 1.132

8 0.920 0.923 23 1.045 1.044

9 1.010 1.012 24 0.938 0.940

10 1.090 1.093 25 0.814 0.820

11 1.162 1.163 26 0.674 0.685

12 1.223 1.223 27 0.517 0.535

13 1.274 1.272 28 0.344 0.370

14 1.313 1.310 29 0.159 0.192

15 1.339 1.336 – – –

H (m) Negative bending moments H (m) Negative bending moments

(kNm/m) (kNm/m)

FEM Proposed model FEM Proposed model

0 – 0 16 −51.92 −62.43

1 −2.33 −5.71 17 −52.43 −62.68

2 −7.06 −11.36 18 −52.43 −62.35

3 −11.75 −17.00 19 −51.91 −61.43

4 −16.37 −22.50 20 −50.88 −59.90

5 −20.89 −27.84 21 −49.33 −57.75

6 −25.26 −32.96 22 −47.24 −55.96

7 −29.42 −37.78 23 −44.60 −51.54

8 −33.31 −42.23 24 −41.41 −47.44

9 −36.88 −46.30 25 −37.64 −42.65

10 −40.14 −49.98 26 −33.40 −37.20

11 −43.06 −53.25 27 −28.92 −31.48

12 −45.64 −56.07 28 −24.12 −24.41

13 −47.83 −58.42 29 −16.34 −17.14

14 −49.63 −60.28 30 −6.04 0

15 −51.01 −61.62 – – –

62 5 Structural Behaviour and Simplified Model Proposition

0

1

2

3

4

5

6

7

8

9

10

11

12

13

Height (m)

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

E

29 A

30

0 1.4e-4 2.8e-4 4.2e-4 5.6e-4 7e-4 8.4e-4 9.8e-4 0.00112 0.00126 0.0014

Displacements - m

E - Finite Element 3D A - Adjusted Model

Regarding to positive bending moments, the mean difference was 8%. It is important

to note, however, that, despite the differences observed, the results obtained are

always on the safety side and, as will be discussed in the following sections, they are

quite satisfactory when compared with the simplified calculation methods available

in the literature.

The parametric study described in the following section will present factors to be

used in determining the final design bending moments.

5.4 Parametric Study 63

H (m) Positive bending H (m) Positive bending

moments

moments (kNm/m) moments (kNm/m)

FEM Proposed FEM Proposed

model model

0 – 0 16 29.76 32.45

1 1.36 2.97 17 30.03 32.58

2 4.07 5.91 18 30.02 32.41

3 6.75 8.84 19 29.73 31.93

4 9.40 11.69 20 29.14 31.14

5 11.98 14.47 21 28.25 30.02

6 14.48 17.13 22 27.04 28.57

7 16.86 19.64 23 25.51 26.79

8 19.08 21.95 24 23.65 24.66

9 21.13 24.07 25 21.45 22.17

10 23.00 25.98 26 18.89 19.34

11 24.68 27.68 27 15.97 16.36

12 26.15 29.14 28 12.56 12.69

13 27.41 30.37 29 7.84 8.91

14 28.44 31.33 30 0.35 0

15 29.23 32.03 – – –

The objective of the parametric study is to investigate the behaviour of the dimen-

sionless parameters α, β and γ , regarding the variation of the silo geometry, and to

propose an approximate expression for their calculation.

The variation of the silo geometry will be measured by two distinct indexes, which

are conventionally referred as geometric indexes of the structure, listed below:

1. Diameter/Wall Thickness Ratio—D/t;

2. Silo Height/Diameter Ratio—H/D.

The implementation of the parametric study considered silo groups with diameters

and heights indicated in Sect. 1 of Chap. 2. The dimensions referring to the arch

describing the region of interconnection between two cells (40°), the thickness in the

centre of this region (30 cm) and the wall thickness of the silos (20 cm) were kept

the same used before, since such values are typical in silo groups construction.

Table 5.4 presents the variation of the dimensionless parameters α, β and γ for

several values of the geometric index D/t, keeping the H/D = 3.0 index constant.

The data in Table 5.4 are plotted in Fig. 5.10.

The behaviour observed in Fig. 5.10 is typical for other values of the studied H/D

geometric index (2.5, 3.0, 3.67, 3.75, 5.0 and 7.5) and in it one can observe that the

64 5 Structural Behaviour and Simplified Model Proposition

D/t Parameters

H/D = 3.0

α β γ

20 0.4464 0.1983 0.7927

30 0.3498 0.1604 0.6490

40 0.3532 0.1505 0.5664

50 0.3513 0.1464 0.4978

60 0.3385 0.1580 0.3885

5.4 Parametric Study 65

geometric index H/D = 3.0

H/D Parameters

= 20

α β γ

2.5 0.4258 0.2004 0.7764

3.0 0.4464 0.1983 0.7927

5.0 0.4585 0.1916 0.8139

7.5 0.4408 0.1881 0.8192

parameter γ is more sensitive to the variation of the geometric index D/t than the

parameters α and β. In fact, for D/t = 20, the calculated spring constant is about

80%, decreasing linearly to about 39% when the D/t ratio reaches 60. Within this

range, the other dimensionless parameters present behaviour with little variation in

their values.

The behaviour related to the geometric index D/t = 20, in Fig. 5.10, was slightly

different from the others due to the influence of the deformation of the cross section

of the silo, an aspect that interferes with the load bearing mechanism of the structure.

Table 5.5 shows the variation of the dimensionless parameters α, β and γ for

several values of the geometric index H/D, keeping the D/t = 20 index constant.

Figure 5.11 shows the plotting of the numbers in Table 5.5. In a similar way

to that shown previously, the present figure is characteristic for the other values of

the geometric index D/t studied (30, 40, 50 and 60) and in it one can observe an

approximately constant behaviour in the values of the dimensionless parameters α,

β and γ . Such behaviour is discretely disturbed at the point corresponding to the

index H/D = 2.5, a fact that represents the greater influence of the top and bottom

slabs of the silo group, for this geometry.

The results of the study allow to conclude that the dimensionless parameters α

and β show little variation in relation to the two geometric indexes considered. More-

over, taking into account that Fig. 5.10 exhibits a characteristic behaviour of those

parameters for the silo groups studied, it is possible to consider them as constants, in

their observed mean values, respectively, 0.35 and 0.15. This consideration implies to

adopt, for implementing the calculation of the nodal displacements by the proposed

66 5 Structural Behaviour and Simplified Model Proposition

inertia of the cross section as the flexion inertia of the elements and 15% of the

gross area of the section as their shear area. The value of the shear area considered

corresponds approximately to the total area of the connecting region (web) between

two cells.

Regarding the dimensionless parameter γ , it was observed that the geometric

index D/t (see Fig. 5.10) is the one that most influences its variation. In the case of

the H/D index (see Fig. 5.11), there was little interference.

The figures and tables presented allow concluding that the dimensionless param-

eters of the proposed simplified model—α, β and γ —are more sensitive to the

variation of the geometric index D/t than in H/D. Among these parameters, it was

also observed that the one that experiences greater influence of the D/t index is the

parameter γ.

In this way, the proposed analysis model will take into account a single dimen-

sionless parameter—γ, which will have its behaviour evaluated for several values of

the D/t index.

In order to assess the behaviour of this parameter, the minimization problem

discussed in Sect. 5.3 was slightly reformulated as follows:

n

2

min u mi − u ti (α, β, γ ) (5.7)

γ

i=1

α=0.35

β=0.15

The first column of Table 5.6 shows the parameter γ values that solve the mini-

mization problem described above (optimum values), for several values of geometric

5.4 Parametric Study 67

D/t Value

Optimum Adopted

20 0.8255 0.84

30 0.6815 0.71

40 0.6038 0.58

50 0.4899 0.45

60 0.2895 0.32

index D/t. The second column indicates the values resulting from a linear fitting per-

formed to the optimum value. The equation of the linear fitting procedure is shown

below.

Figure 5.12 exhibits the value of Table 5.6 together with the linear fitting. In

the sequence, it is shown figures with the results of positive and negative bending

moments for silo groups with geometric index D/t equal to 20, 30, 40, 50 and 60,

applying the proposed analysis models.

For the calculation of nodal displacements of Model 1 (Timoshenko Beam on

Elastic Foundation) the following values of the dimensionless parameters were con-

sidered:

68 5 Structural Behaviour and Simplified Model Proposition

• α = 0.35

• β = 0.15

• γ according to linear fitting equation describe above.

The positive and negative bending moments obtained with proposed analysis

models are shown in Figs. 5.13, 5.14, 5.15, 5.16 and 5.17 together with the results

obtained from 3D finite element analysis.

Exception made to the positive bending’s moments corresponding to the geometric

index D/t = 20, the other results obtained are in safety side.

5.4 Parametric Study 69

of the deformability of the arch’s supports used in the model for the calculation of

the bending moments) by establishing a 10% reduction factor for the negative bend-

ing moments values. Applying this factor to the analyses performed, the maximum

error verified regarding to the three-dimensional model was 12%, for the silo group

corresponding to the geometric index D/t = 20, the minimum error was 5% for the

geometric index D/t = 60 and the mean error was 8.8%.

The results of the parametric study showed that the positive bending moments

obtained with the proposed model exhibited smaller differences regarding to the

three-dimensional analysis than the negative bending moments.

70 5 Structural Behaviour and Simplified Model Proposition

Regarding to the positive bending moments, the value obtained from the proposed

model can be directly adopted for design purposes, without the need for any reduction.

By doing do, the maximum error committed is 10%, corresponding to the geometric

index of the silo group with D/t = 50, the minimum error is 2.5% for D/t = 30 and

the mean error is 7%.

The only case in which the bending moments from the three-dimensional model

were greater than those of the proposed model was for the silo group with geometric

index D/t = 20, only for the positive bending moments. The maximum difference

found in this case was 5%, which is perfectly satisfactory for design purposes.

5.5 Procedure Proposed 71

The proposed procedure for the calculation of the bending moments in grouped of

reinforced concrete silos consisting of four cells of equal diameter for interstice

loading are as follows.

1. Calculate the lateral design pressure due to the silage [2];

2. Calculate the cross-section area and moment of inertia of one cell in the group,

to obtain Ag and Ig;

3. Calculate reference stiffness of elastic foundation as indicated in Sect. 2 of

Chap. 5 to obtain kr ;

72 5 Structural Behaviour and Simplified Model Proposition

beam and calculating the nodal displacements (translations), considering the iner-

tia of each element as being equal to 35% of the gross inertia of the cross section,

the shear area equal to 15% of the gross area and define the participation of the

spring constant, calculated in item 3 adopting the linear adjustment described in

Chap. 5;

5. Build the transversal model of a clamped arch and calculate the bending moments,

as indicated in Sect. 2 of Chap. 5;

6. Apply a 10% reduction factor to negative bending moments as indicated in

Sect. 4 of Chap. 5. Positive bending moments do not need correction. The result-

ing moment diagrams are those that will be used for the design of the structure.

5.6 Results and Discussion 73

The comparison of results between the methods available in the literature and the

procedure proposed in the present work is presented below.

Tables 5.7 and 5.8 summarize, respectively, the values of the negative and positive

bending moments that are represented graphically in Figs. 5.18 and 5.19. In addition

to the values of the bending moments, the figures also indicate the variation of the

applied horizontal design pressure.

The figures and tables presented are for a silo group with 10 m in diameter and

30 m in height. The geometry of the group is that indicated in Chap. 2, and the

diagram of the horizontal design pressures for the interstice is the one introduced in

Sect. 2.3 of Chap. 2.

The calculation of the bending moments of the proposed model considered the

indications listed in Sect. 5 of this chapter.

It was observed that the results obtained through the application of the simplified

model proposed are the ones that are closest to the results of the three-dimensional

analysis, benchmark for the research developed.

In fact, the behaviour of the bending moment diagram resulting from the three-

dimensional analysis shows that the bending moments (positive and negative) are

initially small, increase until reaching their maximum value in a depth of 16 m and.

From that depth on there is a decreasing until it reaches, in the bottom, values of the

same order of magnitude as those occurring at the top of the structure. This behaviour

is characteristic of the bending moments diagrams that come from the application

of the proposed model, aspect that assure the efficiency of the formulation designed

for the calculation of the bending moments.

The same behaviour does not occur with the other methods for calculating the

bending moments available in the literature, which results in values of bending

moment diagram directly proportional to the applied design pressure value.

In addition, it should be noted that the simplified methods found in the literature

give very conflicting results. The fixed arch method (arch clamped at both ends),

suggested by Gurfinkel [3] and Safarian and Harris [4], presented the smallest values

of the bending moments (positive and negative), always contrary to the safety of the

structure.

74

H (m) Negative bending moments—kNm/m H (m) Negative bending moments—kNm/m

Method Method

Gurfinkel Haydl Timm Ciesielski FEMa Proposed Gurfinkel Haydl Timm Ciesielski FEMa Proposed

and and

Windels Windels

0 0 0 0 0 – 0 16 −22.13 −87.23 −229.76 −72.58 −51.92 −59.00

1 −1.63 −6.44 −16.94 −5.35 −2.33 −5.44 17 −22.98 −90.56 −238.43 −75.35 −52.43 −59.15

2 −3.32 −13.09 −34.47 −10.89 −7.06 −10.84 18 −23.80 −93.87 −246.91 −78.03 −52.42 −58.76

3 −5.06 −19.94 −52.49 −16.59 −11.75 −16.20 19 −24.57 −96.83 −254.94 −80.57 −51.91 −57.80

4 −6.83 −26.92 −70.89 – 22.40 −16.37 −21.44 20 −25.13 −99.74 −262.62 −83.00 −50.88 −56.27

5 −8.63 −33.99 −89.49 −28.28 −20.89 −26.52 21 −26.01 −102.52 −269.93 −85.31 −49.33 −54.14

6 −10.44 −41.14 −108.32 −34.23 −25.26 −31.38 22 – 26.70 −105.22 −277.03 −87.55 −47.24 −51.43

7 −12.16 −47.93 −126.20 −39.88 −29.42 −35.96 23 −27.35 −107.80 −283.83 −89.70 −44.60 −48.11

8 −13.52 −53.30 −140.32 −44.35 −33.31 −40.18 24 −27.98 −110.28 −290.35 −91.76 −41.41 −44.18

9 −14.81 −58.38 −153.72 −48.58 −36.88 −44.03 25 −28.36 −111.76 −294.26 −93.00 −37.64 −39.62

10 −16.04 −63.19 −166.39 −52.58 −40.14 −47.50 26 −28.66 −112.94 −297.37 −93.98 −33.40 −34.46

11 −17.19 −67.73 −178.34 −56.36 −43.06 −50.57 27 −28.93 −114.01 −300.20 −94.87 −28.92 −28.73

12 −18.28 −72.05 −189.70 −59.95 −45.64 −53.20 28 −29.18 −115.01 −302.80 −95.70 −24.12 −22.46

13 −19.32 −76.12 −200.42 −63.34 −47.83 −55.39 29 −29.41 −115.91 −305.19 −96.45 −16.34 −15.70

14 −20.31 −80.02 −210.70 −66.59 −49.63 −57.09 30 −29.63 −116.77 −307.44 −97.16 0 0

15 −21.24 −83.71 −220.41 −69.66 −51.01 −58.30 – – – – – – –

a Finite element method

5 Structural Behaviour and Simplified Model Proposition

Table 5.8 Positive bending moments

H (m) Positive bending moments—kNm/m H (m) Positive bending moments—kNm/m

Method Method

Gurfinkel Haydl Timm and Ciesielski FEMa Proposed Gurfinkel Haydl Timm and Ciesielski FEMa Proposed

Windels Windels

5.6 Results and Discussion

1 0.82 4.10 8.75 2.70 1.36 3.11 17 11.60 57.63 123.15 38.04 30.03 33.82

2 1.68 8.33 17.80 5.50 4.07 6.20 18 12.01 59.68 127.52 39.39 30.02 33.60

3 2.55 12.69 27.11 8.38 6.75 9.27 19 12.40 61.62 131.67 40.67 29.73 33.05

4 3.45 17.13 36.61 11.31 9.40 12.26 20 12.78 63.47 135.64 41.90 29.14 32.17

5 4.35 21.63 46.22 14.28 11.98 15.16 21 13.13 65.24 139.41 43.06 28.25 30.96

6 5.27 26.18 55.95 17.28 14.48 17.94 22 13.48 66.96 143.08 44.20 27.04 29.41

7 6.14 30.50 65.18 20.13 16.86 20.56 23 13.81 68.60 146.60 45.28 25.51 27.51

8 6.83 33.92 72.47 22.39 19.08 22.97 24 14.13 70.18 149.96 46.32 23.65 25.26

9 7.48 37.15 79.39 24.52 21.13 25.17 25 14.32 71.12 151.98 46.94 21.45 22.66

10 8.10 40.22 85.94 26.55 23.00 27.16 26 14.47 71.87 153.59 47.44 18.89 19.70

11 8.68 43.10 92.11 28.45 24.68 28.91 27 14.60 72.56 155.05 47.89 15.97 16.43

12 9.23 45.85 97.98 30.26 26.15 30.42 28 14.73 73.19 166.39 48.31 12.56 12.84

13 9.75 48.44 103.51 31.97 27.41 31.67 29 14.85 73.76 157.63 48.69 7.84 8.98

14 10.25 50.93 108.82 33.61 28.44 32.65 30 14.96 74.31 158.79 49.05 0.35 0

15 10.72 53.27 113.84 35.16 29.23 33.34 – – – – – – –

a Finite element method

75

76 5 Structural Behaviour and Simplified Model Proposition

Tables 5.9, 5.10 and Figs. 5.20, 5.21 indicate the result of the comparison performed

between the values of the normal horizontal forces.

The tables and figures presented contain the values of the three-dimensional anal-

ysis with the finite element method, the results obtained with the use of the proposed

analysis procedure and the numbers from the application of the simplified methods

available in the literature.

The results presented are for the sections located in the plane of symmetry (middle

of the interstice wall—Sect. 3 of Fig. 4.1 of Chap. 4) and sections located at the

Table 5.9 Horizontal normal forces—plane of symmetry

H (m) Horizontal normal forces—kN/m H (m) Horizontal normal forces—kN/m

Section 3—plan of symmetry Section 3—plan of symmetry

Method Method

Gurfinkel Haydl Timm Ciesielski FEM Proposed Gurfinkel Haydl Timm Ciesielski FEM Proposed

and and

5.6 Results and Discussion

Windels Windels

0 0 0 0 0 – 0 16 −95.256 −57.096 −65.660 −66.810 13.400 41.300

1 −7.026 −4.212 −4.844 −4.929 1.820 6.747 17 −98.869 −59.274 −68.165 −69.360 8.550 35.526

2 −14.291 −8.568 −9.853 −10.030 5.280 13.546 18 −102.381 −61.390 −70.587 −71.830 2.370 28.092

3 −21.767 −13.050 −15.008 −15.270 8.300 18.861 19 −105.714 −63.378 −72.285 −74.170 −5.230 18.956

4 −29.393 −17.622 −20.265 −20.620 10.770 23.974 20 −108.896 −65.286 −75.079 −76.400 −14.320 8.074

5 −37.109 −22.248 −25.585 −26.030 12.610 28.352 21 −111.929 −67.104 −77.170 −78.530 −24.980 −4.626

6 −44.916 −26.928 −30.967 −31.510 13.920 32.044 22 −114.871 −68.868 −79.198 −80.590 −37.320 −19.204

7 −52.331 −31.374 −36.080 −36.710 15.100 35.697 23 −117.693 −70.560 −81.144 −82.570 −51.430 −35.677

8 −58.186 −34.884 −40.117 −40.820 16.630 39.754 24 −120.396 −72.180 −83.007 −84.470 −66.920 −53.678

9 −63.741 −38.214 −43.946 −44.720 18.470 43.591 25 −122.017 −73.152 −84.125 −85.600 −82.980 −72.751

10 −68.995 −41.364 −47.569 −48.400 20.110 46.627 26 −123.308 −73.926 −85.015 −86.510 −101.400 −93.134

11 −73.949 −44.334 −50.984 −51.880 21.160 48.741 27 −124.479 −74.628 −85.822 −87.330 −131.340 −115.210

12 −78.662 −47.160 −54.234 −55.190 21.480 49.805 28 −125.560 −75.276 −86.567 −88.090 −159.490 −138.980

13 −83.106 −49.824 −57.298 −58.300 20.960 49.699 29 −126.550 −75.870 −87.251 −88.780 −87.590 −164.350

14 −87.369 −52.380 −60.237 −61.300 19.510 48.305 30 −127.481 −76.228 −87.892 −89.440 30.040 0

15 −91.392 −54.792 −63..011 −64.120 17.020 45.537 – – – – – – –

77

78

H (m) Horizontal normal forces—kN/m H (m) Horizontal normal forces—kN/m

Section 4—support Section 4—support

Method Method

Gurfinkel Haydl Timm Ciesielski FEM Proposed Gurfinkel Haydl Timm Ciesielski FEM Proposed

and and

Windels Windels

0 0 0 0 0 – 0 16 −101.767 −42.822 0 −45.400 0.043 25.794

1 −7.507 −3.159 0 −3.349 1.007 5.315 17 −105.649 −44.555 0 −47.130 −4.435 19.969

2 −15.271 −6.426 0 −6.812 2.908 10.696 18 −109.402 −46.035 0 −48.800 −9.914 12.638

3 −23.260 −9.787 0 −10.380 4.503 14.559 19 −112.964 −47.534 0 −50.390 −16.618 3.755

4 −31.409 −13.217 0 −14,010 5.690 18.336 20 −116.364 −48.965 0 −51.910 −24.397 −6.724

5 −39.654 −16.686 0 −17.690 6.446 21.379 21 −119.605 −50.328 0 −53.350 −33.572 −18.866

6 −47.996 −20.196 0 −21.410 6.890 23.791 22 −122.749 −51.651 0 −54.760 −43.770 −32.728

7 −55.920 −23.531 0 −24.950 7.258 26.240 23 −125.765 −52.920 0 −56.100 −55.359 −48.331

8 −62.177 −26.163 0 −27.740 7.775 29.188 24 −128.652 −54.135 0 −57.390 −67.213 −65.298

9 −68.112 −28.661 0 – 8.414 32.012 25 −130.385 −54.864 0 −58.160 −79.804 −83.172

30.380

10 −73.726 −31.023 0 −32.890 8.959 34.134 26 −131.764 −55.445 0 −58.780 −91.538 −102.220

11 −79.020 −33.251 0 −35.250 9.132 35.442 27 −133.015 −55.971 0 −59.340 −102.459 −122.760

12 −84.057 −35.370 0 −37.500 8.784 35.813 28 −134.170 −56.457 0 −59.850 −92.366 −144.890

13 −88.805 −37.368 0 −39.620 7.778 35.132 29 −135.229 −59.030 0 −60.320 −51.878 −168.480

14 −93.361 −39.285 0 −41.650 6.057 33.289 30 −136.224 −57.321 0 −60.770 −2.842 0

15 −97.660 −41.094 0 −43.570 3.484 30.203 – – – – – – –

5 Structural Behaviour and Simplified Model Proposition

5.6 Results and Discussion 79

intersection of the interstice wall and region of a connection to another cell (Sect. 4

of Fig. 4.1 of Chap. 4).

It was observed that, as it happened with the bending moment’s diagrams, the

literature methods available present results that do not capture the actual structure

response.

Indeed, in all of them the resulting normal force are compressive, and there is no

part in tension, contrary to behaviour that the three-dimensional analysis shows.

80 5 Structural Behaviour and Simplified Model Proposition

forces—plan of symmetry

5.6 Results and Discussion 81

forces—support

In addition, one of the simplified methods of the literature [5] indicates null

stress for the sections located in the support, a fact that does not represent the actual

structural behaviour observed in the three-dimensional analysis.

Although the proposed analysis procedure has not been developed for the calcula-

tion of the horizontal normal forces, the diagrams of these efforts due to its application

present behaviour quite similar to that from the three-dimensional analysis. Some

considerations, however, must be observed.

Firstly, in the case of normal forces in the symmetry section, it was observed that

the proposed model shows an increase of the tensile region of the silo wall. In fact,

the region of the height of the silo where a tensile force is developed is increased from

18 m in the three-dimensional analysis to 21 m in the proposed model. In addition,

82 5 Structural Behaviour and Simplified Model Proposition

the values resulting from the proposed model are always higher than those of the

three-dimensional analysis.

On the other hand, the magnitude of the tensile force of the proposed model in this

region (50 kN/m) correspond to only 14% of the values that are usually considered

in the design this type of structure—internal pressure x internal radius—which, for

the silo considered in the comparison, represents 350 kN/m.

Regarding to the symmetry section, it was still verified that the compressions

resulting from the proposed model are smaller those from the three-dimensional

analysis, however, the maximum value in the two analyses performed occurred at

approximately at the same depth, with close values −160 kN/m.

For the normal forces in the arches supports (Sect. 4 of Fig. 4.1), there was also an

increase in the tension region of the silo (from 15 m in the three-dimensional analysis

to 19 m in the proposed model). Similarly to the normal forces in the symmetry

section, the values resulting from the application of the proposed analysis model are

also larger. However, the order of magnitude of the tensile forces of the proposed

model (35 kN/m) corresponds to only 10% of those usually considered for design

purposes, already discussed previously. In the compressed region, especially in the

depth of 27 m, a behaviour difference of the diagram of the proposed model was

observed in relation to the three-dimensional analysis. Such divergence comes from

the vertical flexion of the walls, an aspect that is not considered in the proposed

analysis model. The flexing effect is more expressive in the region of connection

between two cells (this fact did not occur in the normal forces in Sect. 3 of Fig. 4.1).

For the design purposes, it is possible to use the horizontal normal forces resulting

from the application of the proposed simplified model, taking into account, in the

sections of the supports, the value corresponding to the section located at 87% of the

height of the silo. From this point on, it is recommended to consider a linear variation

up to zero, as indicated in Fig. 5.21. The horizontal normal forces in the symmetry

section can be used in the manner that results from the application of the proposed

analysis model without any restriction.

References

2. ACI Committee 313 (2011) Standard practice for design and construction of concrete silos and

stacking tubes for storing granular materials (ACI 313–11), American Concrete Institute, Detroit

3. Gurfinkel G (1990) Reinforced-concrete bunkers and silos. In: Gaylord H, Gaylord CN (eds)

Structural engineering handbook, Sect. 26. Wiley, New York

4. Safarian SS, Harris EC (1985) Design and construction of silos and bunkers. Van Nostrand

Reinhold Company, New York

5. Timm G, Windels R (1977) Silos, Sonderdruck ans Dem Beton-Kalendan. Verlag von Wilhelm

Ernst & Sohn, Berlin, Germany

Chapter 6

Conclusions and Recommendations

Taking into account the analyses performed throughout the research it is possible to

draw the following conclusions.

(1) The results presented attest to the efficiency of the proposed formulation for the

calculation of the bending moments in reinforced concrete silo groups due to

the interstice loading;

(2) The two-dimensional analysis methods available in the literature do not present

reliable results;

(3) If a three-dimensional analysis is required, the Finite Element Method is prefer-

able. In this case, solid elements, shell or mixed modelling may be used. The

use of this method involves quite complex models that demand, besides time

(computational and human), a strong theoretical knowledge of the technique

used;

(4) The three-dimensional solid modelling is one that admits less simplifying

hypotheses regarding the representation of the physical model. On the other

hand, it is the analysis that requires more computational effort and more time

of the designer engineer in the interpretation of the results, aspect that makes

difficult its use in the daily design practice;

(5) The technique of stress resultants calculation from Finite Element solution (used

in solid modelling) proved to be efficient and can be used as a practical appli-

cation of the finite element method in reinforced concrete design of structures;

(6) The three-dimensional modelling of the structure integrally with shell elements

is the simplest and least demanding computational effort and time to interpret

the results from the design engineer. This modelling, although satisfactory from

the point of view of the structural design, presents an excess of stiffness in the

region of intersection between two contiguous cells, especially in the sections

near the intersection. In addition, this modelling requires the consideration of

rigid connection in the regions of connection between cells, an option not always

available in the structural analysis software package used in the design offices;

F. A. N. Silva et al., Design of Reinforced Concrete Silo Groups, Building Pathology

and Rehabilitation 10, https://doi.org/10.1007/978-3-030-13621-5_6

84 6 Conclusions and Recommendations

(7) Three-dimensional analysis using mixed modelling (walls modelled with shell

elements and intersecting regions modelled with solid elements) is the most

cost-effective;

(8) The proposed simplified model presents quite satisfactory results and can be

used for initial design or even for the definitive design of the reinforced concrete

silo groups. The implementation of the proposed model is very simple and the

results obtained are always in the safety side, within the silo groups range

considered in the research;

(9) The horizontal normal forces of the proposed model can be used in the design

of reinforced concrete silo groups, provided that the respecting the recommen-

dations presented in Sect. 5.6.2 of Chap. 5 are adopted.

- Finite Element Method for ProfessionalsHochgeladen vonAnonymous lV8E5mEO
- Shear Force and Bending MomentHochgeladen vongeorgekenjiputra
- AbstractHochgeladen vonAaron Wong
- Design of ACP panelHochgeladen vonSibiSam
- AFEM.Ch21Hochgeladen vonPradyumna Nahak
- material shape.pptHochgeladen vonHatem Jafoura
- F1.xlsxHochgeladen vonDanish Ahmad
- 143258201 احمال الزلازل ياسر الليثيHochgeladen vonTarek Abulail
- hseHochgeladen vonkkfemin
- 130829_Course Syllabus_me5510 me6510_Fall 2013 (1).pdfHochgeladen vonabdullahaafaq
- Derivation of Link EquationsHochgeladen vonSami Syed
- Effective breadth-EspolHochgeladen vonJorge Ricardo Alcivar
- AE2302 2 MarksHochgeladen vonKevin Morris
- Slide Chpt08Hochgeladen vonC V CHANDRASHEKARA
- 001 RCM Chapter 1 - IntroductionHochgeladen vonMossSkosana
- EvolutionandDesignChallengesforFlexibles-ForIssueHochgeladen vonIrvin Alberto
- 2. Emergency GateHochgeladen vonQC&ISD1 LMD COLONY
- Lesson01 Computational ModellingHochgeladen vonIñigoAlberdi
- MEStrength of MaterialsHochgeladen vonالعلم نور
- Recidual Force EquationHochgeladen vonGirish Deshmukh
- Neutral AxisHochgeladen vonJimmy Ioannidis
- 7.10.1Hochgeladen vonRahul Shrvastava
- 5Hochgeladen vonDadi Aziz
- Bending Moment Lab (CES511)Hochgeladen vonAlif Akhmizan
- AS-II_2-MARKS.pdfHochgeladen vonVasthadu Vasu Kannah
- Assistant Engineer (Mechanical), Group-B Maharashtra Engineering Service in Water Resources.pdfHochgeladen von24sandeep24
- u 02530863088Hochgeladen vonTobiasOliveira
- Axial Force, Shear Force, Torque and Bending MomentHochgeladen vonYo_Rocks
- sfHochgeladen vonHilary Watson
- Shear Force Bending Moment Test ReportHochgeladen vonZahed AbdAziz

- 61Hochgeladen vonMarcelo Protz
- ACI 209.2R-08Hochgeladen vonThanh Vo Phi
- SomeAspectsRCCUseinBrazil 06JulyRF2015Hochgeladen vonMarcelo Protz
- A Review of Different Forms of Sulfate AttackHochgeladen vonMarcelo Protz
- Artigo 1 - An Evaluation of Fibre Reinforced Sprayed ConcreteHochgeladen vonMarcelo Protz
- 13thICAAR - 2008.pdf.pdfHochgeladen vonMarcelo Protz
- Dynamic Increase Factor -Review of Strain Rate Effects for Concrete in TensionHochgeladen vonMarcelo Protz
- b195 - DMA - A Simple Device for Measuring Unit Water in RCC MixturesHochgeladen vonMarcelo Protz
- 13thICAAR - 2008.pdfHochgeladen vonMarcelo Protz
- ASTM A20.A20M 2013.pdfHochgeladen vonjoao carlos protz
- Astm f1030Hochgeladen vonMarcelo Protz
- USSD Construction Risk ManagementHochgeladen vonMarcelo Protz
- ASTM A 934 A934M 2013.pdfHochgeladen vonjoao carlos protz
- ASTM A 74 2013.pdfHochgeladen vonMarcelo Protz
- 2b_02_sulfato ataques diferentes formasHochgeladen vonMarcelo Protz
- 0097_0145_000078Hochgeladen vonMarcelo Protz
- 72Hochgeladen vonMarcelo Protz
- 70Hochgeladen vonMarcelo Protz
- 67Hochgeladen vonMarcelo Protz
- 66Hochgeladen vonMarcelo Protz
- 63Hochgeladen vonMarcelo Protz
- 62Hochgeladen vonMarcelo Protz
- 60Hochgeladen vonMarcelo Protz
- 40Hochgeladen vonMarcelo Protz
- 31Hochgeladen vonMarcelo Protz
- 29Hochgeladen vonMarcelo Protz
- 27Hochgeladen vonMarcelo Protz

- A Swirler Stabilized CombustionHochgeladen vonGabriel Sousa Campos
- Calvi 2013Hochgeladen vonRoberto Gentile
- Troubleshooting_guide Activated Sludge ProcessesHochgeladen vonRoberto Goncalves
- ECE.486Hochgeladen vonphoebezz
- 207 6-16-38 GuniteautomaticslackadjusterHochgeladen vontruckman1000
- Btech Hallticket NewHochgeladen vonVishnu P Sudhakaran
- Mine HoistsHochgeladen vonAnuar Anchelia
- PARTE 31Hochgeladen vonelvis
- DF40Hochgeladen vonsathish_sbit
- Influence of Heat Treatment on Duplex Stainless Steel to Study the Material PropertiesHochgeladen vonIJSTR Research Publication
- PROFITank Sample Structural CalculationHochgeladen vonAndrea Castillo
- 1st Floor Deodorizer Platform Analysis ReportHochgeladen vonukta
- Bell-John Deere 315 SK Workshop Service repair.pdfHochgeladen vonRonaldWiseman
- WW5Hochgeladen vonJhon Nuñez
- Docdownloader.com Hazop Amp Hazid Report Cng Gas Plant Tambak Lorok Rev bHochgeladen vonshahin Tamouzadeh
- TRANSPORTATION SYSTEMS.pptHochgeladen vonCollenne Kaye-Lie Garcia Uy
- Sathi Paran 2015Hochgeladen vonDIMariaAUnited
- T. J. E. Miller-Brushless Permanent-Magnet and Reluctance Motor Drives (Monographs in Electrical and Electronic Engineering) -Oxford University Press, USA (1989)Hochgeladen vondeepak_gupta_priti
- Decking Fixings Guide 667Hochgeladen vonneelfred
- Screw Conveyor DesignHochgeladen vonSaron Yayneshet
- CNC Turning Centre ManualHochgeladen vonranjithkraj
- A2LA ISO 17025 Certificates and Scopes of ApprovalHochgeladen vonBrandon Troc
- Challenge-2003-12- Rocket From the Pocket - Spinning the Reel - High JumpHochgeladen vonsanits591
- Brochure - 14.458 DC Spring Apllied BrakesHochgeladen vonkpr_02161371
- FWEC1A Operation Manuals EnglishHochgeladen vonLevente Soos
- ansys poradnik cz.3.docxHochgeladen vonjuhy
- 221251664-operation-manual-direct-drive-pdf.pdfHochgeladen vonMagno
- 2014 N13 Engine with SCR Engine Manual.pdfHochgeladen vonHarlinton
- laser cuttingHochgeladen vontoto161616
- CHAPTER 7 Heat TransferHochgeladen vonaimanrsln

## Viel mehr als nur Dokumente.

Entdecken, was Scribd alles zu bieten hat, inklusive Bücher und Hörbücher von großen Verlagen.

Jederzeit kündbar.