30426
30426
com
https://ebookgate.com/product/introduction-to-
structures-2nd-edition-w-r-spillers/
OR CLICK HERE
DOWLOAD NOW
https://ebookgate.com/product/java-foundations-introduction-to-
program-design-and-data-structures-2nd-edition-john-lewis/
ebookgate.com
https://ebookgate.com/product/storm-and-cloud-dynamics-2ed-edition-
cotton-w-r/
ebookgate.com
https://ebookgate.com/product/introduction-to-microfabrication-2nd-
edition-sami-franssila/
ebookgate.com
https://ebookgate.com/product/introduction-to-food-toxicology-2nd-
edition-sadras/
ebookgate.com
Introduction to Sociology 2nd Edition William Little
https://ebookgate.com/product/introduction-to-sociology-2nd-edition-
william-little/
ebookgate.com
https://ebookgate.com/product/introduction-to-remote-sensing-2nd-ed-
edition-cracknell/
ebookgate.com
https://ebookgate.com/product/introduction-to-probability-2nd-edition-
dimitri-p-bertsekas/
ebookgate.com
https://ebookgate.com/product/an-introduction-to-
neuroendocrinology-2nd-edition-michael-wilkinson/
ebookgate.com
https://ebookgate.com/product/introduction-to-fiber-optics-2nd-
edition-john-crisp/
ebookgate.com
INTRODUCTION TO STRUCTURES, 2"d Edition
"Talking of education, people have now a-days" (said he) "got a strange opinion that every
thing should be taught by lectures. Now, I cannot see that lectures can do so much good as
reading the books from which the lectures are taken. I know nothing that can be best
taught by lectures, except where experiments are to be shewn. You may teach chymestry
by lectures - You might teach making of shoes by lectures!"
James Boswell: Life of Samuel Johnson, 1766 (1709-1784)
Dedicated to
Frank DiMaggio
Professor Emeritus, Columbia University
ABOUT OUR AUTHOR
William R. Spillers
Professor of Civil and Environmental Engineering
New Jersey Institute of Technology
Newark USA
Woodhead Publishing India Private Limited, G-2, Vardaan House, 7/28 Ansari Road,
Daryaganj, New Delhi – 110002, India
www.woodheadpublishingindia.com
© W. R. Spillers, 2002
The author has asserted his moral rights
This book contains information obtained from authentic and highly regarded sources. Reprinted
material is quoted with permission, and sources are indicated. Reasonable efforts have been
made to publish reliable data and information, but the author and the publisher cannot assume
responsibility for the validity of all materials. Neither the author nor the publisher, nor anyone
else associated with this publication, shall be liable for any loss, damage or liability directly or
indirectly caused or alleged to be caused by this book.
Neither this book nor any part may be reproduced or transmitted in any form or by any
means, electronic or mechanical, including photocopying, microfilming and recording, or by
any information storage or retrieval system, without permission in writing from Woodhead
Publishing Limited.
The consent of Woodhead Publishing Limited does not extend to copying for general
distribution, for promotion, for creating new works, or for resale. Specific permission must be
obtained in writing from Woodhead Publishing Limited for such copying.
Trademark notice: Product or corporate names may be trademarks or registered trademarks, and
are used only for identification and explanation, without intent to infringe.
ISBN 978-1-898563-94-5
Foreword
3. Deflections of Stl1ldures 73
3.1 Vutual work for trusses 74
3.1.l Member stiffness 74
3.1.2 Member/joint displacement relationships 75
3.1.3 Virtual work for a truss element 76
3.1.4 Virtual worli:: for truss structures 77
3.1.5 Some truss examples 78
3.1.6 Williot geomeby 81
3.2 VJrtual work for plane frames 83
3.2.1 The moment-:anvature relationship 84
3.2.2 The virtual work equation 88
3,2.3 A note on the integral of the product of two functions 90
3.2.4 Virtualworkforplaneframes 91
3.2.4.1 Applications 95
3.3 More general applications 95
3.4 Computation of discontinuities 96
3.5 Real structures versus virtual structures 96
3.6 Exercises 98
9. References 213
Index 239
Foreword
Over the past 20 years massive changes have occurred in the practice of structural
analysis. Those of us who graduated from an engineering school in the 1950s were
taught relatively simple skills and hoped that we would never he asked to analyze
structures with more than three redundants. Today a structure with 300 redundants
is not considered large or thought to present any particular problem for analysis. In
fact, a highly reliable solution for such a structure is usually available at a modest
price. We have thus gone from a situation in which the analysis of a highly
indeterminate structure posed considerable difficulty to one in which analysis
capabilities are readily available and cheap.
Educational institutions have for the most part been slow to respond to these
changes. While we did institute computer programming courses quickly and in
some cases graduate courses in computer-aided structural analysis, surveys have
shown that the undergraduate curriculum - the backbone of professional engineering
- is roughly the same as it was 20 to 30 years ago. To the extent that we teach
engineering, not computer programming, this lack of response to the advent of the
computer may have been appropriate. But in the long run the computer will surely
have its impact upon the way we teach structures. While not presuming to know
how matters will eventually turn out, it is the thesis of this text that because of the
computer (if for no other reason) structural engineers must know more today.
The question is, of course, how to know more. As far as this text is concerned that
question is answered in two ways. In terms of depth, an attempt has been made to
discuss three-dimensional problems more than has been common in the past. In
terms of scope, the text moves through statically indeterminate structures and on
some plastic analysis. In order to do this it has been necessary to omit some (in this
context) redundant topics such as the conjugate beam and the three-moment
equations.
William R. Spillers
WRS
1
1.1 MODELING
Structural analysis, the subject of this text, is for the most part concerned with
finding the structural response (the lateral deflection of a builihng under wind load,
the reaction of a bridge to a moving train, ...) given the external loads. In all but
the most trivial cases, real structures, that is structures without the simplifications
commonly associated with analysis, turn out to be imposs11>ly complex. And what
is finally aoa1y7.ed - the structural model - may appear at first glance to be quite
different than the real structure
Constructing a structural model of a given physical situation involves
discarding certain features and emphasizing others in an attempt to develop a
•reasonable' representation. In doing so the engineer must exercise judgment in
knowing what to discard and when he bas reached a wo:rkable model. This brings
up the difference benveen engineering and analysis.
This text is concerned with analysis, not engineering. Given the structural
model and the type of analysis to be performed, actually performing the analysis
should be a matter of routine and not involve engineeringjudgment. However, even
with analysis, engineering judgment is required at two points. It is first of all
necessaiy to use engineering judgment to construct the model, given the real
structure. At some later point in time, given the analysis, the engineer must use
judgment to decide - for whatever reasons - whether or not his or her results make
sense.
It is not posmble to over-emphasize the importance of these two steps
Eventually the engineer must accept the analysis and move forward with the
process of design and construction. If an error of analysis leads to a design failure,
he or she cannot simply shmg their shoulders and walk off. The engineer is legally
and ethically responsible .for producing a design which functions adequately. In
practical terms the only way this can be done is through developing an
2 [Ch.I
understanding of the stmcture to the extent that he or she knows bow the analysis
will turn out before it is actually done. The curious part is that this undelslaocting is
developed through performing analyses and thus one of the fiK:ets of this text
Modeling may proceed on many levels:
the primaty analysis of the Rio-Nrteroi Bridge (see the frontispiece) was probably
petformed on a stmctma1 model which was a beam, rqucsented by a single line.
Jn Older to learn to model structures properly. it is imponant fur the
engineer to observe structures and tly to understand bow each structure func:tioos.
(This is eqoivalent to makjng a structural model in your mind as you pass a
structure.) When the functioning ofa structure is obrious, so is its sttuctural model.
The truss bridge schematic of Figure 1.1 is a case in point. Hele the primary
sttuctun1 elements are the pamllel trusses that transfer the loads from the bridge
itself to the abutments. The typical load path involves a load on the bridge deck
which is transrerred to the stringers which are SllJJP9lted by the floor beams which
fuune into the truss joints. The upper and lower latera1 systems are concemccl with
lateral load (e.g. wind) transfer and bracing against buckling.
A similar analysis can be made of the industrial building of Figure 1.2.
Schematically, the roof loads are transfem:d by the porlins to the roof trusses that
are supported by columns. The bracing systems again are primarily concerned with
lateral load and buckling
There is a foll spectrum of stroctural complexity. While the two structures
just mentioned function in rather ob\ious manners. a point load applied to a
spherical shell or even a cable net (see Chapter 6) can produce a complex set of
reactions which can be difficult to anticipate. In these latter cases it is even more
important that the engineer develop some way in which he or she can understand
the structure for which he or she is responstl>le.
Sec. 1.1] Modeling 3
2 --
----
~_,
-""---
.... _ _ _ bok:ooka
-
_....,. ___
_....,._ Oaolpller-
~~
llCl.lllltll'-
"'
lllO
150
"'
:=
:l.OllO'
:......
0
0
----plilbml
3
100 0
''"
,
Biii-
_ ------.
---
lllO
lllO
II'
7
-- "'
7
.. u.w ......,._ 40
(II
.. __.. .....
==
125
.....
----
, 75
lill"l' 125
........,,..
==
cm.
"'
---
IO
ll 1511
CllqlalloalOd_,,._ lllO
a.-
...
411 ...
•
-...
Dem 411" 0
. . . . . -.. .
411 0
--
o-
=~.i:r.,
l'.
tS =:.-•fartliolJpoof-
lllO
•
.......
--*'- _,
,__
~ 411 1~
_
l6. - 2SO
17.
u.,.
........... _...
l2S
2SO
lOO ,......
It. -
lllO
..-_
··-~
...........
_ _ •_______ __ ........................,., .. ___ _
..... _ -
_ _ ...,.,. _ _ .,ilacllalldfos_lloo_1614
~.._.--10•
""·~----140744
-m:Jl:Jl]--~~-t0-4W1-~~~--1~
SfMIWE HSp.4SlmucKS
li'IGUREt.J ...........
(Rqlrinted wilh pamiaion of1hc Amarbn Amodalicia. ofsaate Highway mil T1wporbdion
Ofticials)
(cl) Some sttuaural materials such as some types of plastics~ with time
and exposure to sunlight independent of their load level This can be impo11ant in
the design of fabric roofs.
(e) Problems of corrosion and &tigue can be significant in struc:tma1 design
situations.
(t) While it is normally assumed that unloaded stroctul3I elements are stress
free, mau11racturing pmoesses can in some cases be quite detrimeldaL Typical
examples of this include stn:sses induced through the 'rolling' of wide flange
sections, punching for rivets and bolts, and welding which is known to crack thick
steel plates when not done properly.
(4). Linear ve1'8US non-linear analysis. Without CX)IDment beyond this section, this
text is concemecl with linear analysis in which the equilibrium equations of an
element are written in its undefonned configumtion. (A compatable
approximation is made when length change is computed from member
displacements.) Put more simply, while in real lite stmctures are in
6 Introduction and Review [Ch.1
1.2 NOTATION
This section is concerned with vector notation and the rules of vector algebra. It is
assumed that the reader is already famihar with the concept of force, moment,
displacement, and (small) rotation as vector quantities from elementary physics
and mechanics courses. For that reason a fairly rapid pace will be maintained;
applications will he stressed.
From the point of view of this text vector notation is simply a short hand
and a vector equation, a simple means of representing two scalar equations when
dealing with plane problems or three scalar equations when dealing with space
problems. The following discussion will center on the three-dimensional case; the
plane case will be relegated to examples.
A vector A is descnbed by listing its three components. That is commonly
done in three different ways:
(a) ijk notation
(1.1)
Here i, j, k are the usual coordinate unit vectors or base vectors and Ax , Ay , Az are
called the components of A.
Sec. 1.2] Notation 7
r
FIGURE 1.5 Construction of a unit vector n along a line from point C to point A
A=[~] (1.3)
These notations all have therr uses but for the most part the ijk notation is regarded
in this text to be unnecessarily cumbersome.
Some of the common vector definitions and operations are:
A=O =>
Vector equality. Two vectors are said to be equal when their components are equal.
That is
Ax= Bx
A=B => A1 = B1 (1.4)
Az =By
the magnitude of A, written IA I. is simply the length of the line from the origin
to the point. That is
(1.5)
lnl = 1 (1.6)
Note that any vector divided by its magnitude is a unit vector. That is if n =Al IA I
then In I = 1.
Vector addition
Ax+ Bx = Cx
A+B=C Ay + By = Cy (1.7)
A,, + B2 = Cz
Scalar product.
A·B = C (1.8)
I I = (A ·A )112
Note that A or (A ·A)= A I I2.
Vector product
Cx = Ay Bz - A,, By
AxB=C Cy = Az Bx - Ax B. (l.9)
c. = Ax By - Ay Bx
An= A· n (l.10)
In mechanics it is common to use a unit vector to represent the slope (or direction)
of a line. For example, given two points A and C and their position vectors RA and
Re , a unit vector pointing in the direction from C to A can be wntten as
(l.11)
Sec. 1.2] Notation 9
Eumple 1.1 Given points A and C, consttuct a unit vector D alona a liDe from
C to A. Given:
RA• (4,4, 1)
Re: • (5, 2, 2)
Step I. Compute the components of RA - Re
RA - Re .. (-1, 2, -1)
This expression follows directly from Figure 1.5 since the vector RA - Re
has the desired direction and dividing a vector by its magnitude produces a
unit vector. Example 1.1 shows the calculations that are required to
compute a typical unit vector.
Mo=rx:F (1.12)
(1.13)
10 Introduction and Review [Ch.1
<Molw: = ry Fz
- rz Fy
{Mo)7 = rz Fx
- rx Fz (l.14)
(Mo)z = rx Py - ry Fx
r,
'z
E+-----'--Y
FIGURE 1.7 x-Component of M. shown using the right-hand rule
and noting that the term (n x r) is perpenchcular to both n and r and has
the magnitude R of the shaft in Figure 1.8. The scalar product of the force
F with this term then becomes the product of the component of F tangent
to the circle times the radius R as suggested above.
Given a rigid body under a small rotation about an axis (see
Figure 1.9) the displacement 8 is given by
(1.16)
d,_,..,
(1.17)
12 Introduction and Review [Ch 1
n m
l°,r, xF, + l°,M 1 =0
1=1 J=l
TREOREM I.I If the forces sum to zero and the moments sum to zero about some
point 0, then the moments about any point O' must also sum to zero.
Proof Let p; represent the position vectors which locate the forces F1 with respect
to the new O' and r', the translation of the origin (see Figure 1.10). It follows that r,
=p, - r'. Using this transformation the theorem follows directly as
f r,xF,+ f M •0
.
1
••• 1•J
.
L (p, - r'} x r, + L M
,.. ..
, 1 ... o
...L p, x F, - r' x
,..L F, + ,L.. M, • 0
Sec. 1.4] Rigid Body Medlania 13
Equations (1.18) and (1.19) are the very basic equations of equilibrium of a
rigid body. They state that the fortes must sum to 7.el'O in three dilections (the x, y,
z coonlinate directions) for three-dimension problems and in two directions in
plane problems. The moment equations are somewhat mote diffiadt. They imply
that the moments must sum to 7ao about the three OOOldinate axes in space and
about a single axis (oot of the plane) in two dimensional problems. Theorem 1.1 is
conceptually useful. It argues that once the three equations of moment equilibrium
have been used, it is not possible to obtain additional information (indepeadeot
'equations) by taking moments about another point of the same rigid body.
...._.. L2 Gi\rea die tnllS lbowll lldaw, find die RllCdcm F., F,. F,.
.,, 1. Compote aait ..... COlll(>'Jlllild
Sap z.
Wlife equilibtillln eqaadoas
Sap J. Solve for Fi. F,_ F,
Uait vec:ran:
-FJ./3 + F,J.[3 • 0
-FJ{i - F,J{l + 10 • 0
-FJ./3 - FJ{i - F1 • 0
In terms of these variables, the pnnciple of virtual work for a rigid body
follows directly·
THEOREM 1.2 The virtual work of a rigid body in equilibrium under a small rigid
body motion is zero.
Proof:
n "' " ,,.
L F,. 6, + L M1 • m = L F,. (60 + m x r,) + L m. M1
1•1 1•1 t•I 1•1
= 611 • (i F,) + ro • (f
·-· •1
M1 + i
•••
r, x F,) = 0
In this theorem, the left-hand side is simply the sum of the work done by the forces
moving through their respective displacements and the moments under the constant
rotation; the right-hand side is shown to be zero, using the equilibrium equations,
Eqs (1.18) and (1.19).
l!quibbM!m eqaatkmJ.
~Fi• !p,. 4
" J
s '
nF1+5F2•3
.. so•
l":1""jj"
AllttP91ivt: Qlmpuu F1 diNG!Qi by laluac 11111111'1111 illOV.t Ille llJll*' $llJIPGft.
80'
Fl• 13 (IPJll)
Sec. 1.4] Riaid Body Medlanla 15
of all, to assume that these rra::tions are paralld to the bars themselves and can be
written using unit vecton as indicated in the example. This point will be argued
more carefully later.
The most obvious way to proceed is the write the equations of force
equilibrium as
or
F1(•1lr + F21•J,. + FJ(a,j,. + (P). - 0
F~•J, + F1.•J, + F.J.a~, + (P), =0 (scalar form)
F .(a,), + Fl.•2), + F.J.,a,), + (P), • 0
Example 1.2 shows in detail how these equations can be solved in a specific case.
Jncidemally, this example also shows that the moment equation about the x-exis
can be used dUect1y to oompute F3. This example will be tetumed to in the next
chapter.
~~-[-::::I!:]·[:-:]-[_::: ::.1-:1
-r-0.6' 1
-4.96•J
There are two points to be made in connection smss resultmts here in this
section. The first bas to do with the number and fype of smss resultants associated
with any cut:
Space beams. When a space beam is cot an internal member force vector and an
internal moment vector are both generated (six scalar quantities).
Sec 1-4) The Concept of Stre11 Re!lultant!I 17
Plane beams. When a plane beam is cut, an internal member force vector and a
scalar moment are both generated (three scalar quantlties).
Both of these statements can be argued from the equations of statics for a
cantilever beam For example, since the cut space beam must be in force
equilibrium, the possibility of arbitrary loads requires the exJ.Stence of an arbitrary
internal force vector associated with the cut Sunilarly, moment equilibrium
requires the eXIStence of an internal moment vector associated with the cut. For
plane beams this argument degenerates to the existence of two scalar force
components and one scalar moment component at any cut.
There is one final point to be made here which gives rise to the concept of
a local coordinate system. It is not enough to simply compute the internal forces
and momems under discussion, they must be computed in a proper coordinate
system if they are to be useful when member stresses are to be computed. For
example, it is common to divide the axial force by the member amt to obtain the
'axial stress'. In this case, strength of materials requires that this axial force be the
component of the internal force vector which is parallel to the member centerline.
Just any component will not do!
Figure 1.12 attempts to make this point graphically. The straight beam
first of all tends to blur the issues since it is so simple. but the cwved beam makes
the point clearly. With respect to statics alone it would be adequate to simply
compute the global components Fx, Fy. and M but from the structures point ofview
that is simply not adequate. To be useful it is necessary to compute the components
T, V, and M of the local coordinate system. The figure attempts to generaliz.e this
discussion to the case of space beams.
The idea of a rotation matrix comes about naturally through the necessity
of computing stress resultants in special coordinate systems. That is the point of
Example 1.4. Given the reactions, it is a relatively simple matter to compute a set
of stress resultants at the upper point (say the cut). Force equilibrium requires
horizontal and vertical force components of 4k and 3k respectively. The moment
at A can be computed by writing the moment equilibrium equations about point A
as indicated in the figure.
The point is that it is necessary to make one additional step and compute
the shear and thrust. Since the shear and thrust are statically equivalent to the
components in the global coordinate system, both sets of forces must simply be
different components of the same force vector at A. It follows that the rotation
matrix (see Appendix 1) can be used to compute these force components as
indicated in this example The a1temative of computing the shear and the thrust
using projections is also indicated.
Example 1.5 shows the d.iffi.culttes which arise when working in three
dimensions. Given, for example, the reactions at point A, the equations of statics
can be used to compute the forces and moments required at B for equihorium. It is
then again necessacy to compute the internal stress resultants at thil! point in the
local coordinate system. That is done using the rotation matrix.
~ensional problems are sufficiently complex to make a
more organized approach using a rotation matrix a necessity. Attempting to use
scalar projections (which is of course identlcal to using a rotation matrix) tends to
18 [Ch.1
~ ->~
\ Lad'>~..
components
\
·1 • .
Componenu of r.
r, • -10, SID 20°'" -3 41'
r, • -10. (1- cos 20") • -0.61' => Ma .. (-1615, O, 341)
20
r, • -10. iiO ir. ran 10• • -o 615'
R • [
cos(-10°) 0 -sin(-10")] [ cos 20•
0 1 0 SID 20°
-11n 20• cos 20• ~1
11n(-10") 0 cos(-10°) O 0
0925 0336 0174]
- [ -0341 0937 0
-o 163 -0.059 0 985
20 [Ch.I
[-3.35']
. . .=
.. -9.39'
0.59211
~beadlaa-,
belldin&
•M'a•RMa• S.llk'
r-9k~
36..1 k'
1.5 EXERCISES
~·
,. 1'
;.r...r;:
11 Find die reacriodl for die lpllCe truss shown.
p
u
11 Find tbe reaclions for the dmdar ucb.,.... (Fiad diem ill 1er1111 of
*-addlnllL)
.....
U For die lldix of Bamplc 1.5 under its own wdlfd. ...aane poiat A to
be free 81 apin compare the Illas zaaUaa1I aa B. (Use a 4' x 8"
..,.)
Saladoa Bun:Jso 12.
Slep I. Compute the weight of the beam sepient.
Let , be the wagllt per foot o( . . . . . . . . . . . . lllllle - 10·.
*' •• J!'""x-7,----Y
l;,....;.....,
!Y'
22 Introduction and Review [Ch.I
Wxy•t
r""w-·mi·~
..
..pdsy•t prdhecu(l-cosl}
• ,r IC'""' - 1111 •> Ir'"
'"' 387 X 101 )( 1.01$ X (~S: -m 20•) m '1:16.1
1- ~ ... 0..2019'
1371 -
Step 3. Compute the reactions of B in the global coordinate system.
·"<~-
e~~
Forces sum to zero: W.,. Fa= 0
'· .. -w - {0, 0, +1371)
U The c:anrilever beam shown has a parabolic shape with .enicll ta1lpnt
atpoiat A rather tbml the eirmlarshapeofBample 1.4. Fiad T, V, M.
,,,,.f._r
-.,!
T 1.34'
l.
14 Compale 1be lactioal F 1 ml F 2 far 1llc .,._ uu. tllown. (See
~1..3.)
. I . Fe
z
---·O"
y
2
This relatively long chapter 1s concerned with a topic which is basic to any
introductory text on structures: the analysis of statically determinate structures. A
statically determinate structure is, first of all. a structure for which the internal
member forces can be computed uruquely using only the given external loads and
the equations of statics. These computed internal member forces are then
mdependent of the materials of which the structure has been constructed, the effects
of temperature and settlement, ...
In order to develop a perspective for statically determmate structures, a
parallel with the equations of linear algebra is useful. Frrst note that the equations of
statics as they are used m this book, no matter how they are constructed, are hnear
equations in the unknown forces and moments. (This 1s due to the fact that these
equations are always written in terms of the given initial configuration rather than
the deformed configuration which would be more correct) Now, given a set of
linear equations, Ax= b, there are three alternatives with regard to solution type as
indicated in Figure 2.1. For each of these alternatives there 1s a structural
classification:
Examples.
/ / /
l$tatlcaHy_ determinate
plane truss
71 l
s,1at1callv lndeterrmnate
panelnlss ~
FIGURE 2.1 Classification or Structures
already been cited in Figure 2.1 and another will be discussed later m this
chapter- Unstable structures are, as a class, somewhat counter-productive since
they are neither designed nor built.
This chapter moves slowly with mcreasing complexity through the analysis of
statically determinate trusses, frames, and some simple cases of shells. A
subsequent chapter will present methods for the computation of displacements in
statically determinate structures which then form a basis for the analysis of statically
indeterminate structures.
""ourvecl.
FIGURE 2.2 Pin-ended Member
pon-encled member
Since the bar in Figure 2.2 ts pm-ended, no moments are shown at Its ends.
The bar is then acted upon by two force vectors. The equations of equilibrium will
now be used to show that (a) these forces are equal and opposite and (b) they must
he parallel to a line passing through the pinned ends of the member. Since the forces
must sum to zero for any rigid body in equilibrium, it follows that
Wnting moment eqmhbrium about the left end of the bar requires that
Since the force on the end of a pm-ended bar must he along the line between the
pins, on any free body diagram of a bar which has been cut, the internal force must
also lie along this line and can therefore be represented by a scalar ti.mes a umt
vector. This scalar is called the bar force and is taken to be positive when the
member is m tension.
For trusses there is a simple, necessary test for statical determmancy. The
node equihbrium equations first provide a set of independent equations; there are 3j
of these for space trusses and 2j for plane trusses, where j is the number of jomts.
The number of unknowns in either case is equal to the sum of the number of
reactions r and the number of bar forces b.
Figure 2.4 shows how these relationships apply to two specific examples. What this
analysis does not show is whether or not a structure is geometrically unstable. In
fact, an example below shows a case m which these relationships are satisfied, but
the structure is unstable ma rather unpredictable manner. But this is not common.
One final comment concerrung real trusses. Whtie it is uncommon today to
find a structure which actually contams hinges, trusses remam ~ common model for
many classes of structures. Figure 2.5 is a case in point. This bridge has heavy
riveted joints which are probably closer to fixed joints than they are to hinges, but
the structure itself has been analyzed and designed as if its members were pin-
connected. The fact that structures hke this can be modeled as trusses has a lot to do
with member stiffness and less to do with member connections. If the members
themselves are sufficiently flexible (think of a strmg) then it does not matter how
28 Statically Determinate Structures [Ch 2
Number ol JoiDll • J • 5
Number of membtrs • I> • 5
Number ot reacdOGI • r • 4
Spacetrua
(Three~support
Numborof j - • 1 • 8 3J•b+r
Num'* of memben • b • 12 3 x8 • 12+ 12
Number of nacdanl • r • 3 x 4
FIGURE 2.4 Counting equations and unknowns for a statically determinate truss
they are connected: the moments that develop will be small. In other cases such as
wooden domes, it 1s the connections themselves that are flexible while the members
themselves may be quite stiff. In either case members can not to be subjected to
significant bending if a truss model is to be valid.
Finally there are practical reasons for not usmg actual hinges in structures
such as the fact that (a) they are expensive to make and (b) they do not work well
(unless they are carefully maintained).
FIGURE 2.S - Chesapeake and Delaware Canal Bridge (Kirkwood-Mt Pleasant, Delaware, USA)
(Courtesy Gerard Fox, HNTB, New York City)
30 Statically Determinate Structures [Ch. 2
/1\IZl\IZl\l/f\
Warran truss
~
K truss
VVV0\NSlsJ
Howe
Scissors
~
Pratt truss
FtnK
or like compression members. Plane trusses can in fact be found in many possible
configuration depending on the functional requirements of buildtng shape
and bridge span and upon material considerations. The discussion is
endless and not especially relevant to this text.
This section is concerned with developing methods of analysis for
plane trusses that can be performed by hand and that add to the
understanding of how trusses function. It is noted in passing that the
most systematic way to approach a truss is to write the joint equilibrium
equations as simultaneous linear equations in the unknown bar forces
and react10ns. These equations can then be solved at low cost usmg a dtgital
computer. While elegant and systematic, this is not a frwtful way to proceed when
working by hand since humans do not perform this type of task well. The approach
taken in this section will be to explain some basic principles of truss behavior and
then show how these pnnciples can be used to solve trUsses without solving large
systems of equations.
lOP - 22R;J{i. • 0 • R1 • {2 jP
-~-·· A 1"06F,
'
08F, F,
( 1) Trusses with three reactions. Smee there are three equilibrium equations for a
plane rigid body, when a plane truss has three reactions it is always possible to
compute these directly. As a practical matter, if these reactions are to be
computed it is generally a good idea to compute them first.
Example 2.1 indicates a typical structure in which it is require to
compute three reaction components R 1 , ~ , Ra . If a moment equation is
written about the left support, Ra can be computed directly; then
horizontal and vertical equilibrium can he used to compute R 1 and R 2 •
(2) The two-bar truss. It should by now be clear that the plane two-bar truss
ts statically determinate unless the bars are collmear. (It 1s then geometncally
unstable.) This imphes that: (a) Any two-bar Joint in a plane trUSs may be analyzed
directly and that (b) the two bar'forces of an unloaded two-bar joint must be zero.
<This is true smce the equtlibrmm equattons become homogeneous. That is. tf the
system Ax = b possess a uruque mverse matnx A.1 , then b =0 and x =0.)
Both of these points are illustrated in Example 2.2. As indicated, joint B
can be solved directly as a two-bar truss for bar forces Fa and Fb smce joint c is
unloaded it can be argued without further analysis that Fe= Fd =0 .
___ __ __......L__
____.,/ Unloaded Joint
!p
Loaded 101nt
(f:.Jy•P
(3) Collinear bars at a three-bar joint. If two bars at a three-bar joint are collinear,
the force in the remaining bar may be computed directly. If the joint is unloaded the
remaining bar force must he zero. (See Figure 2.7 and Example 2.3.)
~F.
n,
lbwtorcacomponanl)
(4) Bar force components. If any component of a bar force is known, the bar force
itself may be computed drrectly. (See Figure 2.8 and Example 2.3.)
Example 2.4 discusses a final two-bar truss. Note the two approaches used:
either the straightforward application 0f force equdibrium or a c0mbined approach
using both moment and force equilibrium may be used.
Fh • 0 (unloaded 1om1)
~ 60' -
Sec 2 2] Statically Determinate Trusses 33
I· 4'
·I
Alternative method of soluuon·
Moment equation (about lower suppon)
h S + Jt
F x " x 2 • 7 x 4 + 2 x 3 • 34-> F1 =6 BS
Jomt 2
+1.
_ _J__H L = 14 H • 32.47
3247 ~
14
~~
--M
Jomt4
13 '14 ......... N
1-2 -
.\#,.+Nu"' -J 2 47 -12.816 M + _;_:___
14.10S N .. -3247
Mv- Nv'" -9
M N
--+ --= -2206 ~N=-010
12 816 14 705 14 705
lt,f I 8R8
i2ii6 - 14 705 N"' -Z
M • 31.547 x (-1) N • 3 595' x (-1)
Jomt5
4.5
p = 2 x 31 54 x 12.816
• 22 ISk
~rr
I I 14• 1r ,,..
• 2..,a: '' •• •
Tnm
mJDJ.tl ACea•...... ~T..-
unstable stmctore possesses the possjbllity of a rigid body lllOlioa which implies
polmtial collapse. While not quite in the spirit of the material of this text, there is a
beautiful tlleorem. of linear algebra (see Appendix 2) that clmcribes this
fon:e/dMplacement duality of geometric insaabllity. Tnms•amct into suuctma1 terms
it slates that
Fig. 2.11 attempts to show the aeaml coafiguation of the W'JCbert ttuss
which uses a rather P""'futr 1oaking c:enml panel with mudanl ttusses oa eitber
side. While the W-Jdlat tnlSI is statically determinate, beamse ofits configmadoo it
behaves like an indeterminate SllUClUR to the exrent that the load P on one span
produces a n:aaion R.. on the odD' span. In any case the analysis of the Wichert
truss is intaating. First, note that the 1nlSS of Fig. 2.11 satis&oa the aeccswy
condition for Slati<:al determinacy siacc thcre are 36 bars, four Jmetions, and 20
joints and therefore
38 Statically Determinate Structures [Ch. 2
2 J =b +r or 2 x 2 0 =3 6 +4
Still there is no obvious starting point for the analysis since there are four reactions
and no two-bar joints to be solved. A common procedure in situations like this is to
begin to take the structure apart in order to use internal structural conditions. (This
will be done again later for the case of the three-hinged arch.) That is done in Fig.
2.11 and it appears that the decomposed structure may be manageable. It contains
two rigid bodies which supply 2 x 3 = 6 equilibrium equations and there are six
unknowns, Ri. Rt, H, V , R, S
Two specific cases will be presented here in detail. First, Example 2.8
describes a Wichert truss which will now be shown to be unstable. Rather than
using the six equations mentioned above, this example uses four equations in the
four unknowns R , S, Ri. R 4 , which are obtained by simply avoiding the terms H
and V That can be done by writing the moment
R1 ;:_ "' JO T R1 ... JU - HIP "' 0 (mumc:ms .iboul ff, t' oo llae left)
20
I'
s1,;1. )(JO+ R. IC JCI ... u (moments about H, v Oii the nsltt)
R, - P + R/."! + S/~'i + R. - 0 (\'tt11cal forces wm to 1ero)
R!vl -~/vi -u (111.oru:unuil lor- IUm 10 zerv)
In malnlt form
ltc111mmal ....,_.,. R • S
equations about their point of application and by putting the two pieces tosetber
before writing the equatiolls of horimntal and vertical equilibrium. The fact that the
suuetme is unstable can best be seen from the matrix form of the equilibrium
equations in this example. Since row three in the matrix is a lintar combination of
rows 1 and 2 these equations me clearly singular, In kinematic terms, the pntiadar
dimeDsioas of this problem allow a kind of 'squashing' rigid body motion in the
suucture which is not obvious.
A more fypical Wichert ttuss is described in Example 2.9. The analysis
begins with the computation of the four unknowns R, S, lt1, a.,· as described in
Example 2.9 except that horizontal equiJibrinm is immediately used to set R = S.
Having these reactions, the method ofjoints is used to compote the bar tomes.
Other documents randomly have
different content
of her (since to weary of one so true, intelligent and
tender was even more of a punishment than a
misfortune) but most particularly, with regard to his
craving for "gowd and gear." He reminds me of a
similar piece of humanity described in three short
stanzas which were found by Mr. Macmath written on
the fly-leaf of a little volume printed at Edinburgh
about 1670, and which I found in Child's Ballads:
"He steps full statly on the street,
He hads the charters of him sell,
In to his cloathing he is complete,
In Craford's mure he bears the bell....
437.
440. "Afterwards."
Bottom. Why then may you leave a casement of the great chamber
window (where we play) open, and the Moone may shine
in at the casement.
Quince. Ay, or else one must come in with a bush of thorns and a
lanthorne, and say he comes to disfigure, or to present
the person of Moone-shine....
The Play.
Lysander. Proceed, Moone.
Moone. All that I have to say, is to tell you, that the Lanthorne is
the Moone; I, the man in the Moone; this thorne bush,
my thorne bush; and this dog, my dog....
If I could dwell
Where Israfel
Hath dwelt, and he where I,
He might not sing so wildly well
A mortal melody,
While a bolder note than this might swell
From my lyre within the sky.
459.
ebookgate.com