II
IIII

I
l

‘I‘ I 1||‘? IV ;
IIWI III
III I III I

I

‘I
II

I
‘|
.I I

I

I I
I I

I
I

I I” II'
I L ;
I I

I

I
I

 

I—| —‘
Isa-104:. H
'mmmlt

THE APPLICATION OF ELECTRIC ARC WELDING

TO THE
M"“‘II”‘ ANT) ILHI
OF
my ,I, f’;'tg:f:§_ ‘ H . ,._ , :af

‘1.
.‘=' “I” ”< III -:.

THESIS FOR THE DEGREE OF C. E.
WILLIAM BARTLETT SPURRIER

THESIS

THESIS . o‘x. .. (I: .~ I
RI} I/L d" {'3’} c ‘J

 

I. 7 2"!-/4\/’~

f“
C

(“J

THE APPLICATION OF ELECTRIC ARC IELPIEG
TO THE
FABRICATION AND ERECTION
OF

LOW COST STEEL DECK GIEDER HICHEAY BRIDGES

Thesis for the Segree of C.E.

Killiam Bartlett Sourrier

gl'.

One bright spring morning several years ago, the author,
then employed by the Lichigan State Highway Department, was called
into the office of the Bridge Engineer and given his first assign—
ment on field construction. The department had under contract at
the time an all steel bridge of the steel deck girder tyoe with
steel sheet piling abutments instead of the usual concrete sub—
structure units. An attemofwas being made to fabricate special
sections of sheet piling by welding together half sections of
piling and other structural members in pieces thirty feet long and
things were not goin1 well. The steel fabricator had called the
office and asked tiiat so:ne one be sent down to see what a H——~ of
a time he was having and, if possible, make suggestions as to what
he could do to prevent the sections from twisting and warping out
of shape. Such was the authors first contact with welding and th
all steel bridge. The final outcome of this first assignment was
the apoointment of the author as welding inspector for the State
Highway Department.

During his time in direct contact with welders and weld-
ing, the author became thoroughly convinced that electric arc
welding was the coming thing in the fabrication and erection of
structural steel and there were tremendous possibilities for an
all steel, all welded bridge. Test s ecirnens with ultimate
strengths of 85000 to 75000 lbs. per sq. in. were not unusual and
welders who could not produce welds having an ultimate strength of
50000 lbs. per sq. in. of weld metal were not allowed on the job.
The all steel bridge, even though it was still in the experimental
stage, was a success bOth from an engineering and a financia
standpoint.

This thesis is an outgrowth of the exnerience the author
had in the building of this first all steel bridge. The first
portion of the thesis is given over to a detailed study of the
process of arc welding and the equipment involved. The strength
of welded joints is discussed and data obtained by testing welds
to failure included. The proner application of welding to struc—
tures in general and the choice of structural shapes most suited
to welding is exp ained and illustrated. Next comes a short dis—
cussion of the steel deck girder type of bridge sunerstructure and
the many advantages of this type over concrete beam and girder or
steel plate girder or truss construction for short hi: hvey spans.

The balance of the thesis covers in detail the all steel suo—
structure unit prOposed. The advantages and disadvantagesa and the
limitations to be placed upon the choice of this type of sub-
structure are set forth. lany of the difficulties encountered in
this type of construction are mentioned and their solution des—
cribed. And, finally, the cost of the all steel, all welded bridge
is compared with the cost of other types using actual cost as
tabulated in the Fourteenth Biennial Report of the hichigan State
Highway Commissioner as a basis for comparison.

These introductory parrmrcpns tell in a general way what

is included in this thesis and w; at the author attempts to establish
as reasonable conclusions to be drawn from the data submitted.

10910.2

(1)

Welding is defined as a localized consolidation of
metals by means of heat, ie., if two pieces of metal are heated
to the proper temperature, and then united by pressure, contact,
or fusion, they are said to be welded. There are rueny different
processes of welding, but they can all be included in one of two
broad classifications which in some cases over—lap, but are never—
theless essentia.lly different. These two general processes are
known as the plastic process and he fusion process.

Welding by the plastic process is accomplished by
heating the surfaces to be weldde until they are in a soft or
pla.stic condition ar 1d then forcing them together by external
pressure. No adcitional metal is aeded to the joint a.nd the
surfaces never reach the fluid stage. Uncer this heading are
classified forge welding of all tspes, electric resistance weld-
ing, and the thermit pressure process.

Welding by the fusion process is accomplished by
heating the surfaces to be welded to a fluid sta.te and then
adding sufficient metal from an outside source to complete the
joint. This group includes the simple thermit process, electric
arc welding, and gas torch welding which latter two processes
are the most adaptable to manual welding.

The field of application of welding by the veriou.s
meth ods is being steadily enlarged by research and experim.entaticn
and 1ew m thods and new equipment are being oresented each year
by the various manufacturers. Automatic welding machines are now
on the market that do their work with as little attention as an
automatic screw nachine. Telded steel is replacing cast iron in
the manufacture of machine parts of all shat es and sizes with a
considerable saving in time and eXpense. Simple structural
shapes and _3lates can be combined quickly and accurately by weld-
ing, to take the place of castings, thereby eliminating the costly
pattern shOp and foundry. In more recent years, welding has been
applied to the fabrication and erection of structural steel with
considerable success and welding equipment is now considered
necessary in an up—to-date fabricating shop. It is not likely
that welding will ever entirely supplant riveting, but the aoplica—
tion of welding particularly to angular and curved connections is
increasing steadily.

From such work is being derived a steadily increasing
fund of knowledge and experience which is eliminating fanaticism in
the application of welding and it is now recognized that the field
for weldino is not unlimited. Also it is being learned that each
type of welding ha it's own field of application in which it is
the most efficient and economical. Arc welding is not nearly as
successful as gas welding for cast iron and at present, electric
welding, is not to be recommended for joints in cast iron where

strength is essential. Neither is electric arc welding to be recom—

mended in the case of non-ferrous metals like copper, brass, or
aluminum in which oxidation of the parent metal takes place at a
temperature far below that of the electric arc. Such welding is not

(2)

Jossible, but it requires a highly skilled operator, and the
strength of the joints is not dependable. However, for the
welding of structural grade, and the more common alloy steels,
no process is superior to electric arc welding: in reliability
and ease of application. The necessary equipment is inexpensive
and highly mobile, and skilled Operators are now available in all
parts of the country at reasonable we cge rates. New types of
equipment are eliminating variations in the strength of weld, due
to the personal elen went and any operator who has received adequate
training and who has ha a reasonable amount of experience can
make welded joints which are uniformly dependable under almost
any conditions.

Since the purpose of this thesis is a study of the
application of the electric arc process, no further attention will
be paid to any of the other methods of welding.

It is a well known fact that, given the right type of
current supply, an electric arc can be formed between two pieces
of metal — known as electrodes - and that, as long as the supply
of electric current is ample for the length of arc held, the arc
can be maintained. The gap between the two electrodes is bridged
by a stream of electrons which complete the electric circuit. The
resistance offered by the surfaces of the electrodes to the passage
of these electrons results in intense he at at the surface of the
electrode, and a crater of molten metal is formed. If we sub-
stitute a steel wire of pr0per size for one of the plates, we can
raise the temperature of the surface of the plate and of the end
of the wire to the stage of fluidity, and the metal from the end
of the wire will flow on to and nits with the molte- 1etal at the
surface of the plate. By prOper manipulation of the wire elec~
trode, a.uniform depos sit of new metal can be placed upon the
surface of the plate in the form of a ridge, this deposit being
known as a "bead". Now, if we go a step further and proceed to ;
draw the arc at the line of intersection of two plates, the surfaces
of both can be raised to the fluid stage, and the bea u. of new 3
metal from the wire electrode placed at the intersection. The 3
metal from all three sources is thus fused into one common mass I
and in this way the two plates are joined together and the joint
reinforced by the addition of new metal from the wire. I

The equipment rec uired for arc welding is not particularly
complicated or expensive, and usually consists of the following
pieces of apparatus:- a welding generator with a suitable prime
mover, which may be either an induction motor or a gas engine, to-
gather with the necessary appurtenances for the control and
measurement of the flow of ca rrent to the arc, an electrode holder
suitable for use with the ordinary com erciel electrodes and the
necessary cable to establish connection with the generator, a
ground clamp and cable, suitable protective devices for the operator,
most important of which is the welding helmet with its specially
made dark glass eyepiece, and a supply of electrodes. All of this
equipment is made by a number of different concerns and no attempt
will be made to discuss the relative merits of the products of
these different companies. However, each of these pieces of equip—
ment will be described briefly with the idea of giving the reader

(3)

of this thesis s13n 1:1;ea of the thracter1°t1cs of the equipment,

and the conditions 0which 1Us be met by the manufacturer.

A steady, uniform flow of electric current to the arc
is one of the essentials for reliable and economical welding, and
the manufacturers of welding generators have given much attention
to the details of these machines and their controls. A good
welding generator must deliver current at the proper voltage and
ampere.ge for the work at hand, must not be overly sensitive to
variations in the resistance of the electric circuit which it
supplies, and must be able to withstand repeated short circuits.
This sounds like a rather large order for a generator, but the
modern welding generator will meet all these requirements with
ease and many others besides.

The present day generator is either a two or four pole
compound wound machine of the "buckin: field" type. The term
"bucking field" means that by some means or other there is set
up within the machine a field_ Opposite in direction to the main
field, and it is this opoosing or "bucking" effect that mhkes
possible high current flow at low voltages and permits adequate
control. There are many variations in the way in which this
bucking field is produced and used. One comoany uses a series of
inter,p oles. Another uses a split pole piece with windings on each
leg. Still another uses a pair of interpoles with tapoed windings.
The object of all these manufacturers is to obtain a series of
a1pere ge ranges without changing the value of the main field
appreciably and without moving the brushes from the neutral point
at which commutation is the best. In all cases, fine adjustment
of amperage produced is obtained by very limited variations in
field excitation. The final result is a generator whose output of
electricity at low voltage and high amoerage can be controlled by
the combination of a multi-point switch and a field rheostat.

fielding genera.tors are rated oy their nominal capacity
in amperes at full load and range from 100 amps which is the
smallest to 600 amps or more in steps of 100 amps or thereabouts.
Generators rated at either 200 or 400 amps are the most common and
will meet most requirements. The 200 amp machine will supply a
current range from 50 to 800 amps depending upon the voltage and
length 0 the arc. The 400 amo machine will supply from 75 to 500
amps and is of ample capacity for almost any type of welding using
either bare or coated rod. The smaller machine is hardly large
enough for steady use with coated rods.

The prime mover for welding generators may be eit er
an induction motor or a ge 3 engine. IYhere adequate AC current
is available, an induction motor is b] far the best driving unit
and the most economical. A motor rated at 20 H. P. with a con-
tinuous load temperature rise of 40 C. will drive a 400 amp gen-
erator and a motor of half that size will drive the 200 amp gen-
erator. For field welding, tne gas engine driven unit is by far
the best and in many cases is the only type that may be used.
A 400 amp generator will require a 50 H. P., six cylinder engine of
the heavy duty type running at 1500 rpm., or more under full load.
A 300 amp generator will require a 30 H.P. engine, also designed
for full load at 1500 rpm, or more, which may be of four cylinders.
Either engine will require a reliable governor to protect the engine

(4)

from over pinning when the load is suddenly removed as it is
whenever the arc is broken and also to bring the generator up to
Speed quickly after the arc is struck again.

As has been said before, the controls for the ordinary
welding generator consist of a multi—ooin: switch of sons type
to vary the welding current in stems of from SO to 100 amperes
a.d a field rheostat for finer adjustment. This arrangement
permits close and adequate control of the arc and also allows sons
variation in the relation of current to voltage, this being im—
portant when extremely long leads are required to reach the work.
modern day sets also have a reactanoe in series with the welding
leads to stabilize the arc and to prevent the surge that always
tends to occur when an electric circuit is shorted. An ammeter
and a voltmeter are standard equipment on all welders and are
necessary for the proper setting of the controls for any particular
piece of work. An experienced welder will eXperiment on scrap
material until he has the amperage and voltage at the arc exactly
as he wants it for the work at hand, an., once set, Xpeots-the
machine to be stable enough to maintain that setting without
further adjustment or attention.

Normally the details of the structure will, to a large

extent, determine the arrangement of the parts and the location

of joints. The same care is necessary for welding as for riveting
to avoid inaccessible joints and the detailer should be familiar
enough with the possibilities of welding to know when and how to
pply welding. There is never an excuse for an inaccessible joint.
In general, work should be arranged in such a manner that short
leads are possible and a minimum movement of the welding machine
necessary.

One of the leads from the welding machine is connected
to the work by means of a "ground" clamo which can be securely
fastened at some point where it will not be in the way of future
work. So long as there is a continuous metallic circuit offered
by the work, the ground clamp seldom, if ever, has to be moved,
and is often fastened in place by a short tack weld.

It is necessary for the welder to protect himself care—
fully against the possibility of arc burn caused by the large
pr0portion of ultra-violet rays in the flame of the are. All
exposed surfaces of the body must be covered. Ordinary heavy
clothes are sufficient for the main portions of the body and
canvas gloves will protect the hands. The eyes and face are
usually protected by a hood or helmet with eyepieces of snecially
prepared dark glass which filters out most of the dangerous rays.
The effect of arc burn is the same as sunburn, but is more in-
tense and dangerous. Burned eyes are just about the most painful
things that a human being has to experience, and there is little
or no relief to be given the patient. He is temporarily blind
and a mild local anaesthetic is all that can be applied to the
eyes. Natural healing processes will heal the burn in 48 hours
if it is not too severe, but medical attendance is always necessary
at first. No person should watch the play of an electric are at a
distance of less than fifty feet without some eye protection and,

if it is at any time necessary to work close to a welder, suitable
protective glasses should be worn at all times.

The wire electrode, which has been mentioned from time
to time, is in most cases, a straight piece of wire or rod 14"
long, either bare or coated with a special past e or flux. Stan-
dard sizes are expressed in ter s of the nominal diameter of the
metallic portion of the electrode and are 3/53, 1/8, 5/32, 3/18,
1/4 and 5/8 inches. A heavy coating of flux may double the
actual diameter of the finished electrode, but the cesis nation of
size is always the diameter of the metal core. The metal in this
core is a low carbon st eel of setter than ave~ege grade anr of
such analysis that the metal deposited by the arc is practically
the same in composition as the parent metal. The America.n welding
Society has standardized the analysis of the rod to be user for
various types of welding and their Specification E—I—B for rods
to be used on structural grade steel is as follows:

Carbon .13 to .18t
Langanese .40 to .30"
Phosphorus less then .04?
Sulphur less than .041
Silicon trace

This steel should have an ultimate tensile strength of
from 55,000 to 60,000 lbs., per sq. in., with an elongation of 10
to 12 per cent. Then used by the average welder, the metal de—
posited will he.ve an ultimate tensile strength of somewhat more
than 40,000 lbs., per sq. in., and really fine welders will raise
this figure 10,000 lbs. It is very seldom however, that any
welder using bare wire will make a joint strorger than the parent
metal and this must be taken into consideration when specifying
the type of rod to be used for any piece of work.

host me nufacturers of welding rods also mal {e wa t is
known as a flux coated rod. The purpose of this flux coating is
to produce a reducing atmosphere, usually carbon monoxide, which
will envelop the arc flame and prevent the formation of oxides
and nitrides by reaction with the oxygen and nitrogen of the
atmosphere. The molten metal in the arc crater has an affinity
for oxygen and nitrogen, and if the are is maintained in an atmos—
phere composed chiefly of these two elements )21088 and nitrides
are formed within the deip ositec metal an d seriously impair its
strength. field metal deposited with a coated rod is normally
much finer in texture than 1 etal deposited with bare wire and
normally has tde fine silky appearance which is typical of
structural steel.

One fore ign manufactur r makes a rod that is coated with
spirally wound asbestos yarn which is impregnated with chemicals
and serves the same purpose as the coating of domestic rods. Laid

allel to the cord and within the coating is a fine aluminum wir
which serves as a cleansing agent for the weld metal and has a
tendency to 110 at the ox ires and nitrides to the top of the weld
where they are absorbed. by the slag. This rod will deposit weld

metal which compares very favorably in strength with that deposited

by the better grades of domestic rods

The steel core used in these coated rods is exactly the

(81

 

‘

‘J

sane analysis as bare wire or:
tics so the increase in the st
due to the effect of the coati
coated electrodes than xith be
lar:er, more oowerful velcing: acline is red . Also t:.e cos
of coated electrodes is nearly double that of bare iods, but if
there is any need for dependable strength, the coat ed rods are the
thing. This increase in cost is offset to some extent by the fact
that an experienced welder can Wor k much faster With the coated
rods. Another thing worthy of note in ciscuSs‘n' coated lectrodes,
is that the coating is a non-cond1cter of electricity and thus
eliminates W1C is knorn as side are, ie., the formation of the
are from any hart of the rod except the tip. This is of very area
importance to th— welder When Working in close quarters Where it
st impossible to avoid touching other parts of the structure in
an attemrit to reach the joint with the end of the rod. It is
possible With a coated roc to lay the rot on a grounded steel plate,
start the are at the exgosed end, and then v=atch the arc Ceoosit
a. senb ance of a bead on the surface of the slate Without further
attention. This is a very pretty trick to Watch and might have
some possibilities.

as the seas ohysicel character is-
n.:th of the v.eld metal is evicently
lore a jerag. is used With

n
r as a result,

6‘
C"
x.
A
K

3a

$- ‘1

14?“!

J

So much for the equipment and theory.

After the Work is all set up and all the necesser
equipment correctly arra.nged and adjusted, there is nothing left
to dobut strike the arc and go to work; at least at is the i1n
pression some vendors of Welding equipment seen to leave With th ei
customers. Therefore, let it be said now, that good Welding1‘s an
art in itself. Striking and holding the arc is a kn ack that re-
quires much 3 rectice. After the arc is established, it must be
held to the proper length and manipulated in th- prooer Way to
obtain the desired results. However, skill in arc welcing is not
beyond capabilities of any conscientious workman.

I‘

As has been said before, striking the are surely and
quickly is a trick that requires m1ch pre.ctice and. experience.
The end of the rod is touched to the Work with a slightly strok ing
motion of the e25 and then drCWn back about an 1/8 inch. If the
-rod is not drawn aWay q1ic1cly enough, it "freezes" to the work
and has to be brolen eve y by main strength. If drawn away too
quickly and too far, the arc is orol ten and the Whole orocess has
to be started. gain. Skilled Workman develop the e.bility to
strike an are by direct touch without th e strokino motion and this
is often necessary When Working in creased quarters.

After the arc is once started, it is necessary that the
length of arc be maintained uniform in length and as short as is
reasonably possible. An eighth of an inch or thereabouts is the
preper length, and any great variation from this length will effect
the strength of the weld, usually in an adverse manner. Holding
an arc of correct length is also complicated by the necessity of
"feeding" the rod into the are as the end of the rod melts off and
is deposited. Fortunately, holding an arc of correct length becomes
habitual after a reasonable amount of experience and practice, just
as swinging a golf club or a baseball bat correctly becomes a matter
of habit rathen th an of conscious effort for the exoerienced player.

This is indeed for unate for the welder cannot see the en? of the rod

(7)

after he has pulled his helmet down over his eyes, and therefore
has to depend entirely on a sort of "sixth sense" to determine
whether he is holding a short are. To the bystander, the crackle
of an arc of correct length is very distinctive, and, once heard,
easily recognizable afterwards. If the arc is too long, small
explosions occur within the are at roughly uniform intervals and
the pOpping sound of these explosions is an immediate indication
to a trained inspector that a long are is being used.

Variations in the length of are have a very direct
bearing on the strength of weld. The length of the actual flame
of the arc is some where near uniform for any combination of rod
size and amperage that is likely to be used. Therefore, if the
end of the rod is held close to the surface of the parent metal,
the flame penetrates deeply, a crater of molten metal is formed,
and deep and thorough fusion takes place. If too long an arc is
held, the arc flame plays on the surface of the parent metal, but
does not raise the surface to the melting point, and, as a result,
no fusion takes place, for the presence of a crater of molten metal
is essential for fusion. This word "penetration" is ever present
in a discussion of welding and welding methods for the success of
a weld depends entirely on the amount of penetration secured.

When welding two plates together, it is absolutely essential that
the intersection of the plates be firmly fused together and it is
for evidence of the existence or non-existence of this fusion that
the inspector seeks first. One characteristic of almost all poor
welds that the author has seen broken was lack of bond at the line
of intersection of the connected parts. Most good welders use
what is nown as a "stringer bead" placed at the intersection

line to insure proper fusion. This stringer bead is nothing but

a small bead carefully placed at the intersection to insure proper
bond between pieces.

The finished weld is built up of a number of separate
beads, each firmly bonded to the others, and the adjacent parent
metal. The number of beads used varies with the size of the weld
to make, the size of rod used, and, to some extent, the personal
likes and dislikes of the welder. Where a very large weld is to
be made, it is customary to place the stringer beads with a small
rod and finish with a large rod which will deposit metal more
rapidly. The author has met one welder who had the ability to
place a small bead with a 1/4" rod and preferred to do all his
work with this size. Normally a 3/8" triangular fillet weld which
is a common size is made by first placing a small stringer bead
with say an 1/8" rod and then finishing with a 3/16” or 1/4" rod
using a total of two or three beads or, as they are more commonly
called, "passes." A butt and bond weld of two 3/8“ thick plates
to the top of an I—beam with a gap of 1/3" between plates will
require three passes - a stringer bead at each bottom corner of the
gap and one large bead to fill the gap to the surface of the plates.

The weld metal in large beads is placed by a sort of
weaving motion of the rod which causes the end of the rod to
move in the form of a series of over—lapping triangles or ovals,
depending on the size of the bead to be placed. It is the ability
to, in this way, place the weld metal in a smooth and even ridge

with thorough fusion between subsequent beads that makes the

(8)

 

finished welder. If one of the pieces connected is thicker than
the other, it is necessary to play the flame of the are longer on
the thicker piece so that both pieces are raised to the temperature
-of fusion. If this is not done, there will be no bond to the
thicker piece.

The author's observation of welders and their work
had lead him to believe that the knack of welding is a thing
that some workmen can never acquire. Steady hands and nerves
are essential, and, lacking these, the workman will never be
able to hold a short, steady are for hours at a time. A good
welder must be conscientious about his work, patient, and
ambitious, for welding can not be learned in a day and many are
the trials and tribulations of an apprentice welder. He must
also be watchful of his work to avoid the formation of bad
habits and the adoption of faulty methods which in later years
will have a bad effect on his work. This latter is of par-
ticular importance for good welding is largely a matter of habit.

The fact that good welding is a habit is the one thing
that is now and will continue to build up confidence in welding
as a method of erection. Many engineers feel that because the
personal element is unavoidable in welding, and because there is
no non-destructive test of a finished weld, welding and welders
can never be relied on as we rely on properly inspected riveted
construction. Their position was perhaps well taken during the
early stages of welding when all work of this kind was more or
less experimental, but that day is past. One particular welder
worked almost continually for two weeks on a research problem
supervised by the author, and his welds averaged better than
40,000 lbs. per sq. in., ultimate tensile strength. This was
with bare wire. The author was also responsible for the
qualification of welders for the Michigan State Highway Departa
ment for two years, and the average of the strengths of the test
samples submitted by some twelve or fifteen welders was better
than 55,000 lbs. per sq. in., ultimate tensile strength. This
latter group of welds were made with coated rods of one kind or
another. To the author, this is not only an overwhelming argu-
ment for the use of coated rods, but also for the reliability of
such welding. These averages represent a safety factor of 4 and 5
respectively on the basis of the design strengths now in common
use. This superiority of coated rod welding over bare wire weld~
ing became so obvious to this department that they wrote into . ~
their specifications for welding the requirement that §;;_W81d1ne
shall be done with a coated rod of some approved type.

Any method of setting up work that is satigfgctory for
riveted construction will be suitable for wel ing an n many
cases simpler methods may be used. If the parts to be handled

nd 11 ht in weight steel C~clamps can be used to
§§3a§$§§§.a Heav§er pieces will require erection bolts but these

imited to the number necessary to carry the dead load 3
gfintfiz part at a high unit stress in the bolts. Perch angle:d |
can be tack welded in place either in the shop or in the -ie I
and members supported in that way until the finished wildstgre
made, at which time the perch angles can be knocked if ewwitha
sledge. It is often possible to hold the piece in p sot lds
the erection derrick and a few C—clamps until sufficien we 1 d
are made to carry the dead load. This is the most economics an-

(9)

satisfactory way and a skillful detailer can so design the parts
as to make this possible in the majority of cases. ~

The difficulties encountered in welding are none of
them serious and none of them in any way unsurmountable. First,
in order of importance, is the position of the welder in relation
to the work. No man can work well when in a cramped or otherwise
uncomfortable position when his work requires steadiness of hand.
For this reason it is always advisable to provide the welder with
a platform, staging, or whatever else is required for his comfort,
built in such a way that he can both reach and see his work easily.
If such platform or staging can be built in such a way as to
allow him the unrestricted use of both hands, so much the better.

Overhead welding is not very nice work for an inex-
perienced welder, and some inspectors will tell you that one sure
way of checking up on a new welder is to try him on overhead
work. If he finishes the assignment without burning himself he
must be a fair workman. The main difficulty with overhead work
is the shower of sparks and molten slag that fall from the arc
and this shower of hot Sparks is much worse if the welder can
not hold a short and steady arc. Some welders will also make
the statement that it is impossible to do overhead welding with
a heavily coated rod but a little observation of the work of a
skilled welder will soon prove the falsehood of such a statement.
However, it is always wiser to eliminate overhead work if
possible, and a competent detailer can usually find the way out
of such cases as are encountered. '

Joints inaccessible to the welder are a direct indict~
ment of the detailer who made the working drawings. No drafts~
man assigned to such work should be ignorant of the process of
welding and the equipment used and certainly no one is to blame
but himself if the erection crew reports an inaccessible weld.
This latter statement holds true for both riveting and welding.
However, it might again be wise to call to attention the fact
that the insulation on a coated rod makes it possible to work in
places that could not be welded with bare rod.

Difficulties due to the fit of the parts to be welded
are easily avoidable by correct detailing and careful inspection
When two surfaces are to be welded together, they should meet
smoothly, and, if necessary, be clamped or bolted together until
sufficient welding is done to hold the pieces in place. If the
pieces meet at an angle there should be contact along the entire
length of the line of intersection. If the pieces lap or lie in
the same plane, the adjoining surfaces should be parallel and
fit tightly together.

Sometimes when making butt and bond or some types of
fillet welds, insufficient clearance is allowed between sur-
faces, msking it impossible for the arc to reach the line of
intersection and bond the lower corner pr0perly. This difficulty
can be eliminated by careful detailing and inspection. EXper-
ience and some research work by the author indicates that the
minimum gap between plates of a butt and bond weld should be the I
thickness of the thickest plate plus 1/8" and that little or no .|
bond to the supporting structure can be expected if this allowance|

(10)

is not made. It is possible of course to allow a much wider gap,
but there is no economy in so doing, for the time and material
required increase much more rapidly than the strength of the weld.

Undercut or rounded edges on the plates will cause a
tendency towards weak welds and should be avoided if possible.
Even the slightly rounded edge of a universal mill plate will
make bonding the lower corners of the weld difficult, particularly
if bare rod is used. Some of this difficulty can be eliminated
by the use of coated rod in which the added heat and penetration
makes prOper fusion possible. The only real cure for this trouble
is the beveling of the edges of the pieces to be connected in such
a way as to leave a wide open Vee into which the weld metal can be
placed.

A 60 or 90 degree Vee with a gap of 1/16" or slightly
more between the bottom edges of the plates is the best possible
preparation for welding, but this is not always obtainable for
economic reasons. No quick, acourate and economical method of
beveling thin plates has yet been develOped and at present, bevel
cutting with an oxy-acetylene torch is the best method. This is
slow at best and leaves a film of oxide on the plate which is very
hard to remove and which has a detrimental effect on the weld.
Beveling with a milling machine is out of the question on all
fabrication jobs because of the time and expense involved. How~
ever, it has been proved conclusively that welds as strong or
stronger than the plates connected can be made without beveling
the plates, so this particular difficulty is more a matter of
theory than of practice.

One other difficulty encountered, particularly in
fillet welding, is lack of support for the bead. The welder is,
after all, dealing with a liquid metal which obeys all the laws
of gravity and, if not supported, will flow over the edge of the
plate. No welder'can place a 3/8" fillet weld on a projection
of only 1/4" and make it look right. The remedy for this diffiu
culty is obvious.

Difficulties due to expansion and contraction of parts
are the most troublesome and by far the hardest to deal with.
Careful design of the component parts of the structure based on
a practical knowledge of welding practice is the best remedy.
Long continuous beads are always a cause of warping and should be
avoided wherever possible. The allowable unit stresses for welds
now recommended by the American Welding Society are very conser-
vative, and any good welder can deliver welds with a factor of
safety of 4 or 5 based on these allowable stresses. Therefore,
the writer believes that it is both logical and safe to design
welded joints in the same way riveted joints are designed, ie.,
divide the stress to be carried thru the joint by the strength of
each unit of joint material. Stitch welding by which is meant
welding short beads with unwelded portions between adjacent beads
has been proved to be practical by actual tests of full size joints
and a weld three inchvskip three inch joint will develop an
amazing strength in longitudinal shear or pure tension.

(11)

In this way it is possible in most cases to avoid calling for a
long continuous bead.

If it is necessary to make a long continuous welded
joint, some of the eXpansion can be overcome by what is known to
welders as "back stepping." This consists of making the joint
in a series of short welds by putting in say 6" of weld and then
moving ahead a foot and putting in another 6" and finally going
back the full length of the joint and filling in the unwelded
portions to complete the joint. This way no one portion of the
connected parts has a chance to become greatly hotter than the
coolest portion, and the heated portions have a chance to cool
between passes.

When welding a joint between two fairly long members
there is always a tendency for the far ends to crowd together,
thus spoiling the alignment of parts and, in the case of butt
and bond welds, closing the gap between the members. This can
be controlled by means of small wedges or by tack welding. Tack
welding skillfully aoplied is one of the most useful kinds in
welding. Short bits of weld can be used instead of erection
bolts, thus eliminating the punching of holes in the shop. Light
parts can be held together by Cuclamps while tack welds are made
and the clamps then removed for use elsewhere. It is even pos—
sible in some cases for a helper to hold the parts together while
the welder puts in tacks. By means of tack welds, entire struc-
tures can be set up and the fit of joints checked and adjusted
before any final welding is done. Expansion and contraction can
be checked or controlled by tack welds. In short, the Oppor-
tunities for simplification of erection by preperly applied tack
welding is almost limitless.

If undue distortion does take place it can usually be
remedied by cold working or the application of heat and quenching
at the proper points. This latter method is very tricky, for
there is always the danger of releasing some internal stress, re-
tained within the metal after rolling, which will make the defor-
mation worse instead of better. For this reason, if no other, it
is always wise to avoid long continuous welds. It also should
not be forgotten that there are many instances in which welding is
not the practical way to make the joint. The writer witnessed one
attempt to use welding in the fabrication of special sections of
steel sheet piling that convinced him thoroughly that using beads
80 to 30 feet long to connect comparatively light members was
neither practical or economical.

Inspection of welds is still much a matter of routine.
For the fabrication and erection of structures there has not yet
been discovered a non-destructive method of testing that can be
applied out on the job. X ray or gamma ray pictures will show in
detail the inner structure of the deposited metal, but no one would
think of taking the necessary apparata out to a bridge site and
attempting to take pictures of every joint in the bridge. There has

(13)

been some experimental work done in the study of welded joints in
much the same way in which riveted joints are tested. The in-
vestigators believed that it would be possible to listen to the
sound transmitted thru a joint by means of an ordinary physicians
stethescope and in this way spot gas pockets, oxide inclusions and
other things that tend towards weakness in a welded joint. A

group of fairly capable welding inspectors were given some instruc-
tion in this method and then asked to pass judgement on a series of
specimens. Comparison of their judgements with actual strength
tests on the same specimens showed that the results of this type

of inspection were erratic and unreliable and that, without the

aid of apparatus of any kind, this group of inspectors could do
better work than with the stethescOpe. This does not mean for

one moment that this method may not be develOped into something
very exact and usuable. Nothing would be more acceptable to both
the inspectors and the welders themselves. The average welder
seems to take great interest in his ability to make good joints

and obviously wants to prove that he is as good a workman as any
other welder on the job.

For closed vessels, soap bubble or oil tests have become
standard and the results obtained check very closely with break
down tests on the same vessel. Many of the old time steel erectors
who are still hostile to welding, occasionally try the good old
"sledge hammer test" and many are the surprises that occur. On
one job a riveter was caught trying to knock off some welded perch
angles with a 20 lb. maul. As punishment for this offense, he was
instructed to keep at it until he broke off one or two of the
angles which he had damaged. He was a convert to the use of
welding before his job was finished and was also unable to lift
the sledge the next day. The writer also saw some welded special
sheet pile sections which had been pounded down to refusal with
a 2000 lb. gravity hammer at a 12 ft. drop in which the body metal
had failed without damaging the welded joints in any way.

The use of welding as a means of fabrication and erection
does have a very definite effect on the choice of sections to be
used. Just as in riveted construction, the shape of and method
of making the necessary joints is often the controlling factor,
but this is less true in welded structures than in riveted struc—
tures, as a great part of the economy possible thru welding is due
to the elimination of parts which, in riveted construction, are
used only as a means of connection. A typical case is that of the
stiffener angles in plate girder construction.

From a stress standpoint, the outstanding leg of a
stiffener angle is the portion designed to carry the load, and the
other leg serves only as a means of attaching the angle to the web
of the girder. Also if a stiffener angle is to be effective in
transmitting stress from the-bottomwflange to the web, it must be
milled and ground to an exact fit against the flange. Welded con—
struction eliminates entirely the use of an angle section and sub-
stitutes a plate of the proper size and thickness, thus reducing
the weight of material required one half. Instead of grinding the
plate to an exact fit against theébottom flange, it is cut a trifle
short and any variations in fit taken up by the welding. And the

(13)

bearing area is always 100 per cent of the sections.

Another typical case is the splicing of the top and
bottom chord section of trusses. Riveted construction requires
the use of additional splice members and increasing the number of
rivets at the joint to develop the entire section. Welded con—
struction requires only that the joint be fully welded with, in
case the stress at the joint is heavy, a small patch plate welded
along its entire perimeter.

In general, the choice of section for welded construction
is based first upon its adequacy to carry the stress imposed and
secondarily upon the way in which the joint may be made the simplest.
All elements not necessary for the transmissal of stress are elimin-
ated wherever possible. There must be, however, enough perimeter
to allow the amount of welding required to transfer the stress
across the joint. The ideal condition is attained when there are
large areas in contact, thus providing the maximum perimeter for
welding. Next best is for the two sections to butt against each
other so that eplice plates may be used. Least desirable is a
joint which requires a gusset plate. The economies obtained by
the use of welding are not in the cost of making the joint, but
in the reduction of the amount of material required, and the elimina—
tion, in so far as it is possible, of all punching and machining.
Modernization of cutting methods and the general adoption of the
oxy~acetylene torch as a method of cutting and shaping parts has
added greatly to the ease with which members can be fitted for
welding.

Plate girders have heretofore consisted of a web plate,
two or more cover plates, top and bottom angles with or without
side plates, stiffener angles, and splice angles and plates if
needed. Of these parts, only tie web plate, the outstanding legs
of the top and bottom angles and the outstanding legs of the
stiffener angles are used in welded construction. A welded plate
girder consists of a web plate with two heavy plates welded to the
edges and stiffener plates welded to both the web and the top and
bottom plates. In this way the required flange area is concen—
trated as far as possible from the neutral axis and attains a max-
imum of effectiveness. Side plates are not used and, if one plate
will not suffice for the flanoe cover plates are added. 'As has
been said before, the connecting leg of the stiffener angles is
eliminated entirely. The girder is designed exactly as for rivet—
ing and the amount of welding proportioned in the same manner as
rivet spacing. Flange splicing consists of butt welding the plates
together and perhaps adding a short splice plate and web splicing
is done in the same manner. '

That such a girder is strong and most efficient is borne
out by tests made by the Westinghouse Electric Company on three
similar beams. Bean 1 was the usual riveted type and weighed
798 lbs. When tested to failure by center loading failure took
place by buckling of the top flange, the yield point being 55000
lbs., and the ultimate load 68900 lbs. The section modulus was
61.1. Bean 2 was identical with Beam 1 except that the parts
were welded together and the section modulus 60.2. This beam

failed by crimping of the top flange, the yield point being 65000
(14)

lbs., and the ultimate load 7730' lbs. The bending moment was
BTOOO ft. lbs., and the maximum fibre stress 53800 lbs. per sq.
in. Beam 3 was designed strictly to take advantage of all the
possibilities of welding and weighed only 656 lbs. Under test,
the beam failed by buckling of the tOp flange, the yield point
being 55000 lbs. and the ultimate load 78000 lbs. Although

the weight was less, the section modulus of this beam was 62.2
and the maximum fibre stress attained was 52700 lbs. per sq. in.
Thus it is shown that although the welded beam weighed 15 per
cent less than the riveted beam, it was considerably stronger.
And failure was not due to weakness at the welds.

The individual welds used to connect the component
parts of a welded structure normally fall within one or another
of the following classifications which are in common use among
welders and designers:

Triangular or Fillet

Butt

Butt and Bond

Plug or Rivet Welds

Combination Plug and Butt and Bond Welds

HUGH)?

Triangular or fillet welds are, as the name indicates,
roughly triangular in form and are used principally where the
parts to be connected either overlap or meet at right angles.
The stress to be transmitted is usually shear, although in some
cases this type of weld is used in tension. Small fillet welds
can be made with but one pass, but this method should be limited
to not cover & inch depth of throat. For larger welds between
thick plates two or more passes should be used arranged as
shown on Plate 1.

Butt welds are used to connect plates which have their
edges “butted together" and the stress to be transmitted is
usually tension. Actually the pieces should not come in contact
with each other as some allowance has to be made for the addition
of weld metal. Good practice includes the beveling of the abut~
ting edges, but this is an eXpensive process and is seldom used
except for pipe welding or the connection of small sections. The
number and sequence of passes for the welding of plates of various
thicknesses is shown on Plate I.

Butt and bond welds are, as the name indicates, used to
connect two pieces and at the same time, bond both to a third, and
is the weld used in the so-called "battle-deck construction." A
gap is left between the tOp plates and the weld made by filling
this gap with weld metal. This is also a means of avoiding the
EEVEIing of plates mentioned in the paragraph above for, if the
gap left is of proper width, it is possible tEAmahe a ggod jgint

u -v . n s. This 'oint show u a wa"s 6 male
gigfieggosgraigrgdgggsgéegg the strength of the weld depends largely
on the stringer bead placed at the intersection angles. The
number and sequence of passes to be used is shown on Plate I and
is the same as for simple butt welds. .

(15)

   

‘11-, ""F'g "" ""","_.."'__‘I'i-'."“fi‘F’ff-‘b .1;
METNOD '05 WELmNe ,t-Qf-1.N.T5 _ «
VAWOUS THICKNESS-ES 'o-F METAL ’

 

 

 

 

'

 

 

NE 7 E

II

 

 

 

 

.L
4.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

:X- Y

 

 

 

 

 

 

 

 

 

 

 

 

Plug welds are the counterparts of rivets and are
used in a sim ar manner. One of t‘e pieces to be corrected is
punched and then welded to the other by filling the open lole
with weld metal, taking particular care to bond the pieces at
their line of contact. So far as the actual W€ldifl5 is coacprned
plug welds are a variety of butt and bond weld and the number and
sequence of passes shown on Flat 3 I for butt and bond welds will
apply. Tests have shown that pldg weld s less than 13/13 incn in
diameter are not successful azd that 13/16 inch diameter is a
better size. The ad' ded wel:1ing in a 17/15 ilCh diahe ter plzg does
not increase the strength in proportion a11d therefore the larger
ping welds are uneconorical.

It is possible to combir e plug and butt a. bond welds

and this type of weld was st 1died in detail by the i: hway Dept.
The plates are prepared for welding by first pup chin ng a series of
holes of say 15/15 ir ch diameter in a straight line 3r d then
cutting the plate alon g a lip e thro.cl tb e centers of these Open
holes. The Isailt therefore is to increase the length of weld
possible for a plate of any given width by malt ng tr e edE e of the
plate serrated as is the perforated edge of a postage stamp. This
however is an eXpensive process and tests inlicate that the strength

f SlCh a weld will actually be less than strength of a properly

made butt and bond weld.

“a
T?
'—
L.

The strength of welded joints has been the subject of

any investigations, and many papers have beene written on the
subject. Therefore, the discussion in this th sis will be limits
to a report of the findings of an investigation made by the aithor
for the “ichigan State Hi ohwey Department. Tlie inve etion
dealt mainly with the butt and bond weld and the possible effect
of position of the weld upon its strength. It must be borne in
mind in studying these findings that all welding was done with bare
wire and that the strengths obtained are therefore low. Subse~
quent testing of welds made with coated electrodes proved beyond
doubt that a strength of 55000 lbs. per sq. in., in tension, is
easily obtainable and that good welders will normally make welds
stronger than the parent metal. The Highway Department required
that all welders be pre—qualified and test sam .les breaking at
70000 lbs. per sq. in. were not urusual. In fact it becahre necesu
sary to reduce the net width across the weld to prevent failure

of the test specimen through the parent metal which after all
proved only that the weld was stronger than the parent metal. The
following plates and the explanations thereon are from the original
report.

One problem which is easily solved by welding but very

difficult in riveted onstrnction is that of joints in a skewed
structure. Normall v all joints in a steel lstficture are at .
rigzht angles for tie seeirate bending of an odd shaped connection

plate to a predetermined angle of skew is entirely too costly for
general practice. Also there is a limit to the ability of steel

(15)

 

 

BUTT & BOND WELDS PLATE A

 

 

 

 

 

 

 

 

 

 

 

 

 

 

,1
50000 ”‘
~1 ' 1L4 FT V7
3. t" r” _J
a} -n h_ ,
’t
'1‘ r f L:
.‘ >-——+
Q ~— (enéfli
\l +—L _:I———.-
1 :1
’,.«r‘>-~ _~ ”—I A N
430001 ,1? a .14} fiH
.. r-
‘1 __.1 Iii. 1 I,
I 1.“! ‘. 1
to H 1.1 i”? ”1
2 r“ ' 1
2: V“ 1:11.111 b"
w 4-1 ~—
r-- Hi
J i 1
33000 1 L.1‘} . L_J
1n 1...
Q i- r
E —‘
C‘ s
V .
b
L.

 

 

J." I in u .3.“
4. ‘3’ ’ 2 ‘E 4
GAP BETWEEN PLATE:

This plats shows the effect of the gap between the plates
1pzz. H10 strength of s simple butt and bond weld stressed in
tensio.. parallel to the plstss. with the exception of the welds
with a 3/6” asp tbs strengths are remarkably uniform die perhaps
to tnc fact tost the wsld failed in all cases. Why the 54’8” gap
voids acie so wssk is hsrd to explain. One possibility is that
t ere is,for each combination of plate thicknesses,ons particular
:3; in which the concentration of heat in the weld metal reaches
a maxilla and results in so undue amount of oxidation. Welds in
a narrow gap are quickly msdc and the metal does not absorb a
great amount of best. fields in a wide top are made slowly and,
sltno a great amount of heat is obsorbed,tners is ample opportun-
ity for dissipation of this host. Between thsss two extremes
t‘c e may be s point st which large anoints of heat srs absorbed
withoit adequsts Opportunity for dissipation and,ss a result,
citreu.c oxidation tskss plscs

One thing not shown by this plste is that thsrs was littl
or no bond between the plates and tho beam flange where the gap
was only l/4'. The boss fell off diring handling but the strength
of the weld compared fsvorsbly with tbs other specissns

. BUTT z. BOND WELDS
t W

 

 

 

 

 

   

 

 

     
  
     
    
   
   

 

‘ [-1 "1
3
Z’Qflgm J. ‘ “—1 —-+ E.
.Q ‘
‘ H E
' .1." ,49‘4‘ H [XE—31f A
‘ ’ \\ ,___4 [I r— _
:1 w E 7T7
’""-" ‘ L
"w H l
k. ‘ L- ,
! +—

HJ’E

 

 

- L fh-A
l

7'? M

_ I

.n 0.-

   
     
 
 
   

.33] 'J i l“ g" 3'
" ,. 4,‘ g z - Z '
'U GAP BETWEEN PLATE: -.
S . 3...." . _ _
I ' - 1-.” "' “5.1, is". _
h“ '.?m"~.~iee plots, shows the effect of the gap between "V‘é‘i’jw. .

" ~ 7 - ~ . - ' d‘i“ .
erer—F‘stue stre‘u‘ tn of a simple butt and bond sold ""559 A
”tgfixsizgu perpen icular to the platen The 99°91'99““ Md“ '1?“ "
. 3 3:31,, a,1/4' gap between plates failed in 13_ggx‘dgr%alé.13;g%f°
soae‘ mad-e. with s 3/8“ gap. It is crbvious—Lst'ugt MN .F“‘i3§£%%n§gp°
r‘ea-Jlts"xi'.*x.u..a Joint with little or no stxength under uni.sfi_:~‘£_~‘:,-
éfrgf-htreag; It would seem that a gap 1/6" “gar than ”I", _:.§*<{.E}"
‘ xa'egfg‘o‘! the‘ thicker plate is the most .315! 10.1sn't1592f°i°n%:%§$19
'\ @w strength of Joints made with a 5/8 gap 101 :99; uh" 4‘1; “.- ,
wanna. say as for welds stressed by tension Fatal?°1,.1~° t" P'~9‘-£.9{
' u ' ' ' " ‘ ‘

. A”

’

fl
I

'4 , g

- ~ .' .
t-._ I“ ’i ' ‘ :1 .
‘t I ' ‘s V - . _.
_. ' I .- .r: ‘ .7
‘ ‘
. ' . ‘r

.W.

   
 
  
 
  

   
   
  
  
   
  

. 1‘ .1
1* ‘11

 

 

 

 

 

 

 

. ,s.":l.‘-F ."0 .h ' '1 ' ' I ’ I,"
11 . 7mm (PARALLEL \ ' ’j .-
To PLKTES . 3' ; ”$1.1.
275:” '3‘ - . ' ‘ ' ‘ . ‘ LEGEND - q, 1.151" -
. 13:4- - , . .~ 2.“; pum BUTT 8.‘BO'NLC‘_ ' °‘ '
Fl; ' v ............ —- -,- -- WNW-1% PLUG ‘ ' I
7‘; ' _,-_--...'.*-.. mm 1'53" PLUG
J ‘ . '
5‘”: l {:1
.1 A
l -—- .— ‘7";- . '
, {‘1 I; ' _ j, .1 ' IR"! 1.".
Y “g '4 ’1’ g. ‘. .IO.‘ ’h 1. » .
z‘f ' . '3 ,-f-“ "
f§$>f€ ‘1“,"I‘AW‘J r I"
k... \ J ' '

    

 

 

  

5.x .
. ‘+' h l‘ ' r‘ I“ .
1 4-1 ’ ' ”I . '
A *1 . TENSION , if”???
. 135?; piRPENDICULAR 33"
.31. f. ’ T-O PLATES - 5'
. J: ”4% ‘ 1 ' it.
i; _. 1-fiii - V i- " " ‘ ‘
. Etc. - - -'. ‘4."
:32ij -
t "P *
.fiii“ ° N . 'ofis.
s'fif a”, .. s l;
.1, q.-. . “I =,..
s. 5.11 11;”. ‘ N ”'l I. '
:3", . ‘ . ' “ ‘ ~15” "f: ‘V c.
13‘.

1c .4‘ '
”'41 It is obvtoas that tbsge m nothmg 1,6 be 3311‘ Qiby
mm? 12111111101) of I. plug weld within a 0111,13 41%;! bgnd'gsld wlprc‘ktse‘
to 1?: evened r'r'. 1ens16~n parallel 1’0. tbsQPlfitre‘v‘mM" ““h-f; ; 1;
atreia‘is 1611’s ‘fiensionfisrpsndiculu to “the plates the eddigio .
335a apzwg- ‘weld increases 1th 31:10:15“! of the join-t, firsligiag '0'-
Wei :‘epuatlon of: plates 1m 1m; $390 ’of ma: 1‘s 16,0 96 -.;,,. _ .,

r \ '* .., '4
s1vs.‘ r or?! 1‘ U!!!- - 7 ‘ ‘° ‘ V: t
( .8.‘ f5... -0 in. ! $~ ‘. _’ - ‘ -5 . . . _‘. W“
K 1.. ' 3." - ." 1’ i" 1
in“?! _ 1 ., , . :2 . ' '1 .

r '_ “—V“
I

l
.-

PLUG" WELD s 1 T'PLA'rTE’b gI

 

 

 

 

 

 

    
  
   

 

.

 

 

‘11 HI

I L

 

..
- I.
u- I
- V .
_ I ' l‘5_0 .
- .. 0'
.5

"“ Fez/W
\
\

s
"z

4"“ -__

 

V ~\:VL‘H
F-

 

 

 

 

 

 

 

 

 

 

 

 

 

 

I ’ .
_ {J 1 A
«.-' I
.u
11 ’ 1.1 1
. 5-1.. m ,’ .1 .1
EH _J ’ I
\n I '
. ' .. ~l
' ' #41 C:
I
1 r; . « a
I P Q
. Li. r4 1 . - \r-
1 u; - ‘ . - .
‘E: F1 , 3'..- r‘ .fi‘e‘:
. .t'RY'C ‘ I" k 32‘3“"? 1 '
s‘ 12“. 2 *‘1 :1,” ' » I . r
1- 115:? b J, 1
1 . ,er
_ 'I ' 3.1 ‘ g ”1 L1" ' . 1
u- , 4 . . ’ .‘b. . i 1.. '
.f“ 11;“ " H' 1. _1Ifi:,‘.’ .
_ fivmgmfi ' ' L“ ‘ ‘ “‘3‘”... .
- - ‘ ' L_ " '4. *4.
. .1 _£¥.. 15. 11' 1;! lé" . 11f . - ,-\_ -'
‘I6 16 I6 '6' 3-m- . '6 ' MU ' .
SHEAR TENSION
, J; " The strangth of pldé welds 1e surprisinglj h"‘ N. ' i
s wu>ClJu considers the small amount of weld mstsl 1° 1: ClaC-d "1

the stress is spp11ed as a compreaslv-¢ force sLtingI in the? _34
,m- of the plates 11 is evident that even -a 13 18' Plus '31d 1

- ”Chub be sssigned a far greater tllosahle stress thlil is stand- 1

19.10: a 3/4" x1vs1 1n single shssr. fibers tbs stress is 1

" :awpiicd as s tensils stress in the plans of the plates 1kg . ;

F ' hurenbth or the wsld 1s much lower and 1h1s type of goxnt should ‘.fl

' g *fi‘avd1dsd ‘ . 1

“:31,- .. , _ . ‘1
-.Q ‘ ' .

’hktir . - I
'g a - ‘.. ' ' . ‘

v- v
Q
.§'
8'- ..
Q
WVAL

L

.4

_‘___._

*‘AL

L

M}
‘3.
3
k
K
Q
- ‘3
I 0
‘S'
‘51.

In

-‘...1-

1;;

a"; fin MA 7P 5 mew

‘h

 

 

'¢

[H11

A

Ff.

1

v

 

 

LONG.

The data above 1: for 3/8” mm Verde-:93 "hyme-
This type of wild .ge v5 nyue
ed

7‘4-

The ee.fety factors chain are based on e;1q§abflgg?
e8 Q! 2000 1be./11n.1ncn in longitudinal aLear aue.23
n then in trencveree ehear.

e
A;

 

LI

 

 

TRANa

INKH

IooobJD

wauflmos firw' tum

IN-

umber

L

J

a; 77 MA 72’ 5 resivs TH

:i-

_L

—+

 

8000 .1

“eggrial of the same thickness.
@tramg and dependable regardleee of how the stereo ia'ayp%3
a: emaild he need in preference to an) othér tips of-wegg

8F

o‘lr m.

 

LONG.'

[L J]

I

 

 

It.woa1d epcetr th.at a:

allowable stress of 2500 lbs.

safe {6:  we1de of thle else.

LE ‘T.MJEV

‘1 LL- -_.. L-

 

  

 

 

 

 

 

 

 

*1";

 

 

 

 

TRANi'

lper Pin.1n¢h woulejg:

 

A...

  
  

 

    

to stand the strain of bending to a very acute angle. In welded
structures, as has been said before, all that is necessary .cr

the making of a joint is for the members to overlap or intersect
at an angle, preferably of more than 45 degrees, and no connection
plate is needed. The only reason for preferring an angle of
intersection greater than 45 degrees is that it is difficult for a
welder to reach into deep and narrow V-shaped opening between
members.

Plates II and III show in detail how certain of the
joints in a plate girder or truss structure may be made. The
arrangement of members shown is typical, but by no means the
only way in which the joints may be made. Each structure will
present it's own peculiar problems to be solved by the detailer
and he will have to be guided by his own experience and ingenuity
in their solution.

It will be noted that the joint between stringers and
floor beams lends itself very easily to continuity. Under ord-
inary conditions, butt welding the top flanges at their inter-
section will suffice, but any doubt as to the efficacy of such
construction can be relieved by the addition of a Splice plate
as shown. Also note that a fairly heavy seat angle is used both
to carry shear and also to aid in erection.

The connection of floor beams to plate girders is a
simple problem and the solution is self-evident. Note that
the knee brace shown consists of only two plates and that one
plate serves both as a part of the brace and as a stiffener
plate for the web of the girder. The connection of floor beams
to a truss is more difficult because of the lack of perimeter
in contact. However, the choice of truss members with wide
flanges and their arrangement in such a way as to provide flat
planes will increase the amount of welding possible and in most
cases there is ample Opportunity to more than exceed the amount
necessary according to the shear at the joint. There will be
cases in which a gusset plate may be advisable and, while gusset
plates are not contemplated in welded structures, there is cer-
tainly no reason why they should be considered as "out entirely.”

For heavy truss construction the wide flanged beam
sections with the flanges square instead of tapered are the
most suitable. There are many sections of the same depth, but
differing in weight and it is possible to choose these in such
a way as to make a very neat and attractive structure and at
the same time maintain a maximum of strength. The joints are
made by coping both flanges of one member so that it will fit
within the flanges of the other, thus eliminating as far as
possible, secondary stress due to eccentric joints and cross
bending. Plate II shows a typical joint of a heavy truss in
which the tOp and bottom chords are made of 8 by 8 inch column
sections and the web members of lighter 8 inch beam sections.
If additional chord section is required a plate welded to the
upper edges of the beam flanges will provide added strength

without adding greatly to the cost.

(17)

 

 

 

 

au: 4';

 

7W0 2x219-

7%

ueHT’cousTRUCnoN

 

 

 

 

 
 
 

. 1 I .
. [f X/é‘x-jl -70,
ihké’weé (€30 ‘I7 afl/y

 

«"B'xéé'r

 

/8-I floorfieam C. m I
L's-4.3:,
Jr'zfoxapI axéauxjjfie //ZI
’ 7
-:::1

, _ .._ _' ‘ t .
,-..5T 9. s N GER‘ CON-N earn; N

   
  
 
  

a; >9 DATE 2le ’ .

_p/afc f0 jé’I/e/w
53” fII‘MII/
I )7

  
  
 

' 3
A .5
Wilt/‘70” LEC’QI‘I

a...

 

-——-— —_—._.-..

1 ' f" ""’“""‘1’I":““‘" “"
I l
I I.
i l
I.

  

 

 

      
    

I
I
i
r“ 5/Iffener: -E'IL
Ew—‘J ._ -~-e;{-L--—_~_:J. -+—:_

ELEVATION

 

 

 

 

"--.--.._ -.¢—-—-- *—~~~-

SECTiON A'A

TYPICAL PLATE GIRDER

 

Lighter trusses may have chord sections composed of
two channels back to back with either bar lacing or batten
plates, and web members composed of light beans or tee sections
or angles. The required distance back to back of the chord
channels is governed primarily by the radius of gyration required
for the compression chord but there should be no difficulty in
adjusting this dimension to fit the depth of beam sections avail-
able. In this type of construction, joints would be made by
inserting the web members between the channels.

In both cases, gusset plates would be almost completely
eliminated. Splices in chord sections would be made by butt
welding with or without splice plates. For trusses with in~
clined end posts, the haunch joint would be treated in the same
way as a chord Splice. The joint between the top and bottom
chords at the ends of the truss is usually made by running the
bottom chord through and fitting the top chord into it. The
two members are welded together at all points of contact and
splice plates welded to the outside of the members to further
stiffen the joint. The bearing plate is welded to the bottom
chord using stiffener plates if necessary.

Cantilever brackets are made in much the same way as
are knee braces using a plate cut to the preper shape as a
basis. Attachment cf such brackets is a very simple matter
for there are no reasons for avoiding welds in tension. If the
tension at the upper edge is high, it is usually possible to
add a tension member in such a way as to place the welds in
Shear.

So far this thesis has been essentially a discussion
of the application of welding to bridge structures in general.
The theory of welding and the apparata necessary has been dis«
cussed in detail. The actual making of a weld has been dis~
cussed and the difficulties encountered explained with the
remedy for each such difficulty described. The strength of
welds of different types has been touched upon lightly. The
choice of members most suitable for welding has been outlined
and typical construction illustrated. All of this information
has been given to the end that the reader might have a clearer
understanding of the problems involved in the design and erec—
tion of the all steel, all welded structure to be described in
the balance of this paper.

Modern highway practice has established beyond reason-
able doubt that the steel deck girder type of superstructure
is the simplest and most adaptable for highway apans up to
perhaps 90 feet in length and in relatively level surroundings.

This type consists of steel girders resting on abutments and70r
piers supporting some kind of floor slab. Diaphragms trans»

verse to the center line of the roadway at about fifteen foot
spacing are used to provide prOper stiffening for the compression

flange and to aid in distributing the loads. The slab may be

(18)

62.035; don Dm20_mm0
mmDFUDMmemunm WEDNEO quo Aman

 

          

s Essa

     
     

«Gs Kn. bbxhvso
S musket sodas

. . . {K
ansoabxc anusxh medwk. ”Quip \ xsk s

\% WE QKQQVxQ

 

 

 

 

 

N %x W . N til
7 JEN». UK.N\.U§QU \3Q‘\\\S KOBE \Q\\m \bmxh. n\§\h\§ %D\K

as item

 

]

send. twin. m :

 

 

 

[El-”:0

assesses i
E

 

 

either concrete or pre-febricated steel. If concrete is used,

it is customary to place one girder under the outside edge of

the walk, but where steel is used, the sidewalks may be bracketed
out from the outside roadway girder. It is quite customary to
make the sidewalks easily removable so that the roadway may be
widened without replacement of the original slab. For both con—
crete slab or prefabricated steel slab construction it has been
found that a girder spacing of approximately five feet is the most
economical and easiest to erect.

The advantage of this type of superstructure over con-
crete girder or arch construction and steel truss structure are
easily recognizable. For highway loadings the dead weight of a
truss and floor beam structure is entirely too great in relation
to the live loads to be carried. Concrete structures require
complicated falsework across the entire span and, of course,
form work.

The steel deck girder span requires no false work and
very little form work, in fact none at all if a steel floor is
used. If a concrete slab is used whatever forms are necessary
may be supported from the bottom flanges of the girders. Since
the steel companies have adOpted the wide flange section, and
particularly since the addition of the 33 and 33 inch girders,
it is possible to build up to 60 foot span length without going
to plate girder fabrication and for the longer spans plate
girders are less costly than truss construction. A single girder
will not weigh over 10 tons and can be easily handled with the
very common gasoline engine powered erection cranes.

During this period of rapidly increasing highway traffic
the ease with which a steel deck girder superstructure may be
widened is a primary consideration. Widening a truss Span requires
the moving of at least one of the trusses and an entirely new floor
structure. Widening a concrete arch or girder span requires the
erection of falsework to carry forms, and in many cases there is
no way of bonding the old work to the new. Widening a steel deck
girder superstructure requires only the addition of one or more new
girders and the extension of the floor construction, provided of
course hat the abutments and piers are adequate in width. Such
widening may be accomplished without interrupting the flow of
traffic which is most important on heavily traveled trunk highways.

Another advantage which is of evident importance is the
adaptability of this type of superstructure to skew construction.
Skew trusses and monolithic concrete structures present major
problems in design and erection. Skew construction for steel
deck girder spans requires only that the ends of the girders be
offset along the top of the supporting structures and the necessary
changes in the floor slab. If it is desired to take advantage of
the economies possible by the use of continuous beam design, butt
welding the ends of the girders and the addition of a plate welded
across the top flanges at the piers will amp-y provide for any
negative moment develOped.

(19)

Another important economical consideration is the ex—
tremely low dead weight. This factor is reflected not only in
reduced cost of the superstructure, but also reduction in the
cost of the substructure units. For a 60 foot span, the weight
of a steel deck girder superstructure designed to carry R—lS
highway loading is approximately 150 lbs. per sq. ft. of floor
surface. A concrete T—beam floor for the same Span and load
will weigh more nearly 300 lbs. per sq. ft. of floor surface.
For the deck girder span it is possible to out the abutment
cross section down to the minimum necessary for stability against
over-turning without creating excessive soil pressures. This
means less excavation.

In comparing steel deck structures with short trusses
it is necessary to take into account the additiOn of a pier which
might not be necessary if truss construction were adopted. How—
ever, the cost of two 60 foot steel deck girder epans with a
central pier is less than the cost of one 180 foot low truss due
not only to the lower cost of fabrication and erection, but also
to the reduction of the size of the abutments. It must be noted
also that whereas the truss structure cannot be widened without
replacing the entire floor structure and closing the bridge to
traffic during such widening, the deck girder structure can be
widened by the addition of girders and floor slab without inter-
fering with the flow of traffic.

The all steel substructure units which the author would
propose to use with the steel deck girder superstructure des»
cribed above consist essentially of a row of deep arch web of
steel sheet piling with a suitable cap structure. The piling
forms a wall approximately 10" out to out to restrain the fill
which, if prOperly designed and driven into stable soils, will
also be adequate to carry the vertical loads imposed by the
superstructure. Wingwall corners and any special sections re~
quired for skew construction may be fabricated from part sections
of steel piles and such other structural shapes as may be suitable.
It should be noted that welding does not work out satisfactorily
for the fabrication of these special sections because of the ex-
treme length of weld required. The effect of the heat absorbed
is to release any internal stresses which are left in the metal
by the rolling process and it becomes virtually impossible to
produce a straight piece of piling without bend or twist.

For the abutment section, the cap structure would con-
sist of a channel, a Z bar, and two plates, arranged as shown on
Plate IV. The channel and plate on the stream side of the piling
provides a certain amount of beam action at the top of the piles
and tie the top of the piles together. Also these members form a
c0ping which breaks the monotony of the row of piles and gives a
sense of solidity to an otherwise light appearing structure. The

 

 

 

 

 

 

 

 

 

 

P/Uy Wei/(2’5

   

face P/afc’

    

 

 

 

 

 

 

 
   
   

FRONT ELEVAT-Jd'fii} "

:5? '

fiupers, rucfure /.5 shown '
m dubbed/Me‘s "'5‘?

 

 

 

 

 

 

 

 

_ ._ _' I . _‘
- - _ - ._ .4 DETAML
..,.,. _‘ .
t, LE1 ‘ _
. ‘ r’.‘ :' 3”; "$.39
- ~ - ~ — B um;

\ 5".

n _'v . .‘ .4 '

, ‘ o -. ‘51:“ 2’s

‘ 5 o
3 - ' '

f :14. ' g. . .

I 5* W

 

‘
fing/‘V‘Zaf?

5kg ffiaLZ‘

 

 

 

 

 

 

. .é'
_ «TYPICAL CORNER SECTIONS ‘ (a
.L" "' FOR ' '

.
‘. I II I

 

O

{\
H3

 

 

 

 

 

 

 

 

 

 

 

flnly/e- l
5

e

: HLf‘

1| H

L ‘ , ru‘

' ‘3‘

U

" 4‘ \9'
o ‘ \1
., ~82;
. ' I \E'
O, _ l \‘1

- " Q
a- i ”\M' '
n K '-_

 

 

‘. yfylmcm SECTION

#2. D.

'- 1' ~. '

‘ UL: TAM L.
a - RF"
"3?1:E3_§EE,TIO N

 

k

 

.r ‘R -
“?
'

|
" H": P/a
' If s We 45

 

 

 

 

 

 

    

 

 

 

 

 

 

 

 

 

 

 

 

 

A 5 -- —-:J
r112"?
u ":l s I
| ' I a
I ' ‘ '
| .Cl' I l
I | H |
' to" l '
l . I '
l I' l '
l I ‘ I I: |L_I_|_J|
: pp in ' l a
l I I' ‘ ' ' '
I I |‘ P ‘ I l
' no.| .r u 1 ~ '5
. WV?“ 7 -
1' ' ¢ 5 1 O
‘l LLI Li-_{.“ ' .. l’_‘ ,- -/_
EL E '\~"/—\T x 'ON’ _
_ I I
d
' ': vl
.. -,_ .4
‘ =€-;:—:-£‘-.:=: - '. - '. :E—:-:-.:'” '3'!“- h
5' R . /
II \\ >¢ pl
1’ \\
_____ 12:: ._ 212%‘2331 \
h m- -_._ __-- , t _»
- ‘1
SECTION A-A hi
"1
-QETN’L. a
9" "

DlF‘Q' '

the rack of the pilin1g provides stiffr1ess at the top
of the piling and also forms a recess into which tie backwall

the suoerstructure fits thus fo1ning cu «off for
the earth fill behind the abutment. The cap plate is welded
to these other members and forms a solid top for tlze abutmeR t
to which tt1e beam bearir g plates may be Relded. Note that the
top flange of the chennel and the top surface of the Z bar are
set an inch above the cut—off line of the piling to provide
opportunity to weld all contacting faces of the piling to the
cap structure and thus further stiffen the entire unit.

The pier section is u1cre conplicated particularly
where the stream hanks a: c steep and proper grade on the road—
way requires a high pier. A:ain, t e fouzidation of the pier
consists of a sincle row of deep arch web steel sheet piling.
In this case, however, it serves as a support for a column and
girder construction of such height as is necessary to place
the beam bearing plates at the proper elevation. The to es of
the piles are made rigid by an angle and plate construction
both plug and fillet welded to t1 e piling. This rigidity of
connection is acsolitely necessary for the plate forms one side
of the column anchorage. The end of the column is inserted in
the opening between piles and plug welded to the plate on one
side and to the face of the piling on the other, using as mar y
welds as are necessary to carry the column reaction, and filler
plates where necessary. A cover plate is then added to prevent
debris from collecting within the piliR1g wall. The plate on
the outside of the piles enould be Qiite deep both to stiffen
the piling and also to prevent stream born debris from catching
on the faces of the piles.

‘ tie top of the columns a beam section consis tine
of angles and platm is used to support the searing plates for
the siperstructure girders. The deep side plates are plug
welded to the colimn flanges and serve as the web plates of a
three sided bor girder. The anzles, which n'ay be shop welded
to the plates, add to th 1e girder section and also serve to
widen the support for the cap plate. After these portions of
the pier structure have been assembled and welded, the cap and
bearing plates are added and the pier is COmplete.

C

Architecturally, the appearance of these substructure
units is not n.npleasing to the eye. he outer surfaces except
for tt1e aoitueR+ piling, are smooth and unbroken. If it is
desired an apron plate may be plug welded to the face of the
abutment piling below the channel to hide the piling. Varying
the method of placing the fill will create entirely different
appearances. It would be possible to add the apron plate hen—
tioned and carry the fill up almost to the hottom of the channel
and thereby hide the piling entirely with a sodded slope. This
treatment would be particularly pleasing for bridges located
within a municipality where th e banks of the stream are and-
scaped. The appearance of the pier could be altered by adding
channels under t11e top plates at both levels to form c0pings.
Probably the most izportant detail in regards to the appearance
of the pier is that t1 e face plates on the piling should be
of such depth that low water will not expose the piling.

(31)

The advantages of the steel substricture unit over
its counter part of reasonry are numerous and all tend towards
lower cost of construction. Le t us consider them in the order
in which the bridge contractor piisies his work. First, cones

«L

the laying out of tae job. For masonry construction, the out~
line of the sic-footing or cofferdam must be staked out in its
relation to the center line of roadway and the stationing there~
of. This is a orelatively simple operation if the angle of
crossing is 90° at for skew bridges is often a difficult and
tedious joc. After excavation has been completed ar1c the subu
footing poured it is again necessary to get out the transit

and lay out the outline of too footiQ . Laying out pier ccffer~
r”ems and footings is an ever grea ter task if the streezt is deep
or swift moving.

In contrast to this, the layout for an all steel sub»
structure unit requires only the establishment of a single line
for the center of the piling. The instrument man sets up at
the proper station and tirns off the angle of skew. Two measure
ments along this liR e establish the location of the acut11ent
corners, and the job is done. The ends of tne pier piling can
be established by accurately lining up th e pile driver leads
and following th 8 driving of the first pieces of sheet piling
to assure that they re1tain plumb.

Next COuES excavation. For masonry constriction, a
large volume of earth must be removed fiom the cofferdams to
obtain proper deptn of footing and assure that the footings
rest on a stable soil. Sheet piling cofferdams are usually
necessary for pier excavations and very often for abutment
excavations. Excavation is further comolicated by the neces—
sity for bracirg the coffeide m walls and pumping out the water
that leaks into the hole. In rainy weatr1er conditions within
the cofferdan are often so bad that the work has to be held up
until the rain ceases. If the high flood water elevation is
misjudged, pier cofferdams may collapse due to excessive out—
side pressure or be flooded. Extensive excavation during
freezing weather is difficult and, normally, constriction is
postponed until reasor1ably warm weather arrives.

The excavation for a steel abutment takes one of two
forms depending upon the tOpOgraphy of the site. If the grouni
is level a trer1ch four feet wile a Rd deep enough to permit the
top of the piling to be driven to correct elevation will eiffice.
If tie banks of the stream are steep, a shelf out into the bank
at the proper elevation is required. If the roadway approach—
es are on a fill section, it may not be necessary to make any
excavation. The excavation for a pier.consists only of re—
moving logs, boulders, and other debris from the site of the
pier.

The construction of a masonry foundation unit is, of
course, essentially a field problem and very little work can be
done except at the site. Large amounts of lumber, aggregates,
cement, and other materials must be tranSported to the site an

stored until needed and in aynumher of cases storage of aggregates
J I’ :11]

(22)

is a major problem to be solved. Erection equipment usually
consists of a derrick, a concrete mixer, wheel barrows or
trucks for the handling of aggregates and mixed concrete or
perhaps a tower and chute system, and perhaps a power shovel,
if the job is large. Concrete fo~ms have to be built and
properly braced and reinforcing steel wired into place. After
concrete is poured there is a delay of one to three days before
the forms may be stripped for possible use on another similar
urlit n

In contrast to this, the only field equipment neces—
sary for the construction of a steel unit is a derrick with
pile driving leads and a welding machine. The raw material
to be used is steel which is not subject to damage by rain or
high water. Special piling sections are fabricated in the shop
and arrive at the site ready to drive. Other parts are made up
according to the plans and are ready, when delivered at the site,
to take their place in the finished structure. As soon as the
piling for a unit is driven, the welders start he attachment
of the cap stricture and trim, and in a few days the pier or
abutment is complete and ready for the superstructure.

Another great advantage is the ease with which this
type of structure can be salvaged in case of need. Once in
place a masonry foundation unit with the possible exception of
one constructed of rubble masonry, is there to stay and there
is no way of salvaging any of the materials therein for use
elsewhere. The same is true of superstructure of masonry.

The steel deck girder superstructure with pre—fab—
ricated steel floor is almost 100% salvageable. Railing and
sidewalk are easily renovable in sections. Also the floor.

The girders may also be removed as soon as the diaphragms are
cut loose. True, it is necessary to burn out the welds, but

a good man with a cutting torch will cut the joints quickly and
neatly with a minimum of damage to the parent metal.

The upper part of the pier and abutment units in
similarly removed. Perhaps it may be necessary to scrap the
whole of the abutment cap construction and cut the piling off.
Nevertheless, the piling is only shortened slightly and may
either be used in another bridge or sold for use in cofferdans.
Research by steel producers indicates that copper bearing steel
has a life expectancy of forty years so the first usage is not
necessarily the last.

host naturally there are some very definite limitations
to be placed upon the use of the steel substructure unit. This
type of substructure is not adaptable to use with heavy super—
structures or for live loads in excess of highway loadings.
After all, the load bearing capacity of steel piling is due al—
most entirely to skin friction with the soil and direct bearing
on the end of the pile is of secondary importance. Therefore
it is necessary that the imposed vertical loads be light and
that the soil conditions be favorable to the develOpment of
skin friction to the greatest extent possible.

(33)

A ccnplete and caref1l study of soil con ditions at

the site should a.lvays be made. Test borings should be made and
samples of the various strata obtained so that he leg of borings
will present a couplets and accurate picture of t? e soils through
which tt e piling will pass. It is convenient to plot this in~
format ion as a vertical section showing the elevations at which
the vari o1s types of soil are enco11ntered and tr eir nat1re. From

such a chart it is easy for the de si-ner to de teimine how deep the
piling should extend and whether s1fficient lateral,stability be
dEVELOPBd. ‘(JIH

This matter of lateral stability is of utmost importance
for the load on the piling itself is largely due to earth pressure,
and therefore ability to withstand direct vertical loads is only
one of the req1irements for a stable structure. The top of th
piling wall acts as a wtil ver been and is so designed. The
length of the loaded section of the bee m depends upon the dep h 01
he soft soils enc01ntered. The her the bean will carry it load
depends on the beam itself and also on t‘ne amount of restraint t
the fixed end. Therefore it is necessary that the end of th
piling penet ra te v.ell in to a stable soil and for all practical
purposes the length of the piling so restrained shoum exceed the
lengm in unstable soil. It is proba_ble that an upper stratum of
coarse material such as gravel and. a lower stratum of clay presents
the ideal cor ditio_ f or both lateral stabili y and lca d bearirg
capacity. however, gravel alcn— ashes a suitable foundation and
the sti fer clay loans are also suitable. Clay of course will
develop a great amount of lateral stability, but drivirg piling
into stiff clay is a difficult process. he i.1iortent point is
that simply driving he piling down to a hard strat1n is not
sufficient.

8

Tie backs of one kind or another are always usable and

11 many cases are necessary One of tre sinplest types consists
of a fe'w sections driven on a line at right an igles to the piling

all a.nd attached theieto by a T section of piling. Several of
these may be used in a loni wall. Another type consists of
sections of piling driven down some distance back of the main
wall and connected by rods jacked through the intervening soil so
as to not disturb it. A well constructed "dead«mar" will also
serve. If such means will not develop adequate lateral stabilit‘,
it will be best to abandon the is es of using the steel substructure
and t1rn to a masonry structure.

The use of the steel substruc ours is also limited pretty
much to the shorter spans because of the limitations of the steel
deck girder superstr11ct1re. Such an abutment will no carry th
heavy concentr stior1s of a truss struct :re or of a long pla ate girder
span. Tith girders Spaced at about five foot oer ters, t.e reaction
are small and there is ample opportunity for th eir distribution
th ougho m1t the wall. The cap structure shown will act as ae been
of considerable strength and distribute the loads to the pil
sections best able to carry t? em and prevent localized settlen.ent

from distorting the entire structire.

The difficulties encountered in the erection of the
steel substructure unit are not many and their solut iors simple.
Probably the matters of length nd alignment are the most cowluon
and troublesome. If the lengths of t1-e members in the cap
structure are based on plan din:e1sions it is necessary that the
length of the driven wall vary but little from the dimension an
the plans if excessive field alterations are to be avoided.
Preper and. adequate welding of the cap nen1bers to tie pilir1g
also requires tl1st the alignment of tie wall be reasonably close.
There piling is used for wing walls, it is necessary that all re—
turns from the s 1perstruct ure fit as planned. Backwall clearances
at the expar1sion end of the superstructure also depend on accu-
rate slignment. And last, b it far from least, the alighment of
abutments and piers in multiple span bridges must be such that
girders out to length in the shop will fit.

At first glance the solution of the problems of length
and alignment would seem to be to set up all the piling in its
proper location before any of the pieces are driven. However,
'f the piling sections are 30 to 40 feet long, cs is often the
case, bracing such a wall becomes a very difficult and expensive
process. The next thought is to drive the corner sections,
taking all due precautions to locate them accurately and keep
them plumb, and t11en fit th intermediate sections between. The
difficulty e1 countered in this method is th1at of fitting the
last section. If tt 6 wall is driven from both nds towards the
center, it is very probable th1t the closing section would not
fit and it would be impossible to make any adjustments in the
line of driven piling. It has been the writers experience that
the best solution of this problem is eternal vigilance on the
part of th foreman and the inspector. Drive the center piece
of piling first, keepin a it in alignment and pl1no. Then as
each succeeding piece is driven, ch1eok he overall length of
the wall.

If there is a tendency for the sections to crowd te-
gether, thus shortening the wall, put short pieces of t inch
bare welding rod into the interlocks as tt e pile is driven, and
t1 us keep the joint as open as possible. If the wall tends to
lengthen, use a clamp cc mpos ed of a piece of 1%“ cold rooled
round bar bent into a U shape with the ends threaded and a
piece of l x 3 inch flat drilled to fit over the open ends of
the bar. The bar can be put through he handling hole in the
adjoining piece of piling or thr01oh a similar hole birned in
the web of the pilin g and. the position of the flat adjisted to
keep a constar 1t pressure on the piece being driven. By th
use of arch a device, it is possible to crow d the interlocks
tight together and thus shorten the length of wall. Constant
checking of the length of wall will easily determine whether
the sections are going down according to th1e plans and as much
as &“ variation can be obtained in the fit of each pair of
interlocks.

Proper alignment of the face of the wall is purely a
matter of bracing an 1r--'e.lers. In all cases, the piling shoi.ld
be driven between one pair of walers and tv 0 pairs abc1t five
feet ap art vertically is better practice. These walers must
be thoroughly braced and should be composed of at least 10 x 10

(2'5)

timbers. Holes burned through the webs of the pile sections
will permit bolting the timbers together and to the driven sec~
tions. Then, if an instrument is set up and the plumbness of
the driving leads constantly checked, it is relatively easy to
drive the piling to a true line. Lining up the giling before
dr iving has proved impracticable, for the ease reasons mentions
hereto fore.

After all piling has been driven, tlere is at ll the
problem of the fit of the cap men bers. In general, this problem
is best solved by working from both ends towa.rds the center, and
fitting the center sections in he field. This, of coarse, neces-
sitates that tt e cer ter sections be ordered t least 3 inches
longer than plan dimensions we culd ind cate. A aood torch operatoi
can burn off an; excess neatly and the small gap left can be filled

with weld metal and th.n ground smooth.

The alignment of the sidewalks or safety walks can also
create a problem if too much faith is placed upon the accuracy
with which the girder sections are constructed. There is no goo
reason for the web plate of a plate girder being off center, but
the web of rolled sections may be off as much as a" and still be
within the allowable tolerances of rolling practice. Therefore
in lining up the superstructure girders, it is best to use th
web as the controlling point. Variations in flanges can be
nullified by proper allowances in the fit of members. This
explar ation is made becaie e the fitting azd align: ent of the

curbs and sidewalks depends rwrzally on the accu1rc c"y with which
the sidewalk brackets are lo ated.

Q

If the diaphragms are ac urate 1y bailt, the girder
webs will be pare.llel and symmetrical asset the center of the
roadway. The sidewalk brackets are attached to the girder
web and will also be in good alignment. ny variations in the
relation of the curb line to tie center of the top flange of the
outside girder will then :e unimportant. The author was called
noon in one case where there had been careless workmanship in

he fabrication of he diaphragms and the insgec tor had allowed
this fact to be overlooked. An overwide diaphra;m hm been
placed near the center of the span Mn the disc .re.oan v overcome

by deforming the girders to fit. As a resalt, t.e citeide
girder was decidedly rainbow shaped, and the sidewalk brackets
correSpondingly out of 1i: . The aroirt of cutting and patching

d-
:71’
(D
(b
p:
'3)
S
{n

K
C
L
1+
0
f—h
m
m
(D
H
Ho
C.-
;‘ 2
C']
{In
P10
“‘0
1 1, ’D
*1
(1
L.
H
d-

<(
0

As will be 110 ed on 1 wingo on Plate IV the back—
well of this type of str acture eons sts of a plate welded to
connection plates which are in turn welded to the er do of the

sirder recs. The hottor edce of this bec‘ rail plate fits into
the slot formed by the cap p1ate and the Z-bar of tt e as Ht ent
ard it is of course important test there sh01ild be no failire
of the expansion provision due to binding of parts. Since the
alignment of the piling will be imperfect and since also it
mil 1 be obviously itp:.esible to nake snv sijum ent of the piling
any variations in alignment have to be corrected in the adjustment

of the backwall plate. It is wise t

her. - ' u s— 130
of short plates about ten feet long and place them ll their
proper position before any of the connection pl to: are welded,
either to the backwall plates or the gir' ft r the back~

c.
(“x

lab
on
I

A 9
wall.is in place, the expansion slot should be filled with a
flexiole mastic which will prevent any dirt or s as
getting into the joint.

The method employed to attach the sidewalk brackets will
depend to a great extent upon the possibility of future widening
of the structure. If the roadway and sidewalks as planned are
likely to be adeqrate for many years, it is best to weld the
brackets in place and take advantage of the strength of weld metal
in tension. If there is likely to be a need for widening, it is
better to bolt the brackets and so design the location of holes
that they can be used in the fastening of future diaphragms. In
case the attachment of diaphragns requires more holes than are
used in the attachment of the sidewalk brackets, the extra holes
may be temporarily filled and the appearance of the structure
improved by the use of bolts similar to those used to attach the
bracket. This will avoid the burning of any holes necessary but
not punched when the girders were fabricated. It is not recomu
mended that rivets be used for they are difficult to remove and
not satisfactory for the transmission of the tension at the tOp
of the bracket.

There is also a likelihood that the railings may not
line up well although careful fabrication should reduce this
difficulty to a minimum. The attachment of the railing posts
to the brackets should be by bolts to facilitate their removal
in case of widening or accidental damage. Likewise, the panels
of railing should be bolted into place. Any variations in fit
can be easily overcome by slotting the holes used for attachment.

So much for the details of the fabrication and erection
of the all steel, all welded bridge. The drawings which follow
immediately are included to indicate definitely the types of
structures with which the cost of the all steel bridge will be
compared and represent in general the type of structures being
built by the hichigan State Highway Department in the years
1929 to 1931. The comparison is based on the cost per square
foot of roadway which seems the best unit to use. In as much
as there are few of the all steel bridges in existence, it is
impossible to obtain accurate figures on their cost based on a
large number of jobs. The structures selected, however, rep—
resent typical construction, one an ordinary stream crossing
and the other a highway overpass at a two track grade separation.
The bridge was erected over the Looking Glass River three miles
north of the Lansing City Airport and is of two 55 foot spans
with 23 foot clear roadway and 2 foot 6 inch safety walks. The
grade separation is at the intersection of State Trunk Line Mull4
and the Grand Trunk Western Railway 2% miles east of Grand Rapids
and consists of two spans of 88'-10&" and one of 48'w8g" with
42 foot clear roadway and 2 foot 5 inch safety walks. The author
was placed in supervisory charge of the erection of the bridge
and is therefore familiar with the problems to be solved in his
type of construction and this job in particular.

(37)

    

x. ,.I ‘ " - '. . c. _
1”. l - ,2 ‘ ,' T f...“ Erin‘s,- 2" f._
”'2le “ 3 2 I“: h
5‘: (mqmdk ' ' " ‘
all , 5! Que flab] 6'9»: Mar: 17] 521r‘fac'é 7

r.- 'M
. ”.t
t .. n

U

          
    

     

  

. I

 

   

          
       
   
     
 
   

2 I

 

m W L15 01W

‘ fsTEEt: DECK G'I'RDER . .- -
suEERsTRucw-UEE _ ._ -.
cwenzvrsun noon » '

 

 

 

STEEL DECK GlRDER ,7
‘SUPERETRUETURE - ‘
‘ arm GRID FLOOR

_- 3W (Agile! We“ _

    
   
     
     
 
    

f/éd ..

 

 

STEEL DECK GIRDER .2:

SUPERSTRUCTURE ' , 1
3ATTLE4 DECK noon ' “a"

 

 

 

_- CONCRETE DECK GIREER
. . ‘ ' SUPERSTRUETERE

  

.7 TVF’iCAL cauETRuchN
.. MICHJGAN STATE HIGHWAY DEPT. _

.__.,., ~.-—--o - "W

4

     
 

 

fly/0:400 0/ 5 mm} of A €00,115? 7!!

 

(caviar/ark
0! approx. I?’
c/rs.

 

 

 

 

 
  
 
 

" TSEM‘FERAVITY can NTERFO’RT
TVPsCAL ABUTMENT SECTIONS

 

 

 

 

 

 

 

 

 

 

GRAV FTY COLUMN 8. 'G1RDER q

\_ ‘1"
TYPICAL PIER SECTlONS

 

 

Tie total cos t of the bridge as reported in the Four—
teenth Biennial Report of the Eichigsn State Highway Department
was 310473. 45 for 4380 sqxare feet of bridge floor or 33.90 per
square fo . The cost per square foot of ten other structures
of the steel_deck girder type erected on masonry substructure .
inits during the same year and representing nearly 33000 square
feet of floor area averaged 33.50 or more than double the cost
of the all— steel Job. alow steel truss with a floor area of
3252 sq1are feet cost 38. 80 per squaie foot. A reinforced con-
crete T~beam structure with a floor area of 4550 square feet cost
310.77 per square foot. These figures indicate very forcibly
the great decrease in cost of construction possible by the use of
this type.

The total cost of the all steel grade separation as re-
ported in th e sane volume was 332000 for 4284 square feet of
bridg e floor or 33.14 per square foot. Two other grade separations
of t? e steel deer girder type on masonry foundations, a total floor
area of 2370 square feet, cost 319.00 per square foot. Two grade
separations of the half t rough steel plate girder type on masonry
foundations, a total floor area of 5880 square feet, cost 319. 83
per square foot. It is interesting to note that, since the cost
of a steel deck girder superstructure would not vary greatly from
one job to another, tie difference in cost between the various
types of structure represents the savir g possible by the use of the
steel s1 cstructure construction.

Another met? od of comparison. 'The total cost of the
83000 square feet of floor area of the steel deck girder spans on
masor ry suostruotures was 3311, 571.68. Using the unit price

fi ure 00m uted above for the same t1-e of s1 erstruct ure on steel
P JP P

substructure units, this total cost would be reduced to 3128, 00.00,
a saving of 3182, 871. 68. here is no doubt that son 6 of the

str1uct ures included in this group were not of such nature as to

make the use of steel s1ostrictur1e unit advisable. Revs rtheless,
a saving of even ahalf of the 3182, 871.88 would be a major item
and would represent at least two structures of the average t5 pe.

The one treat and all important q1es stion which has
arisen in regards he type of structure preposed in his thesis
is that of usable life. Since the superstr11cture would he
essentially the same regardless of what type of substruct1ure
units were used, it seems fair to disregard that phase of the
problem. We are all used to thinking of masonry as a practically
permanent type of constiuction, and t1at steel is subject to
oxidation to tr e extent that a life of perhaps 20 years is all
that can be reasonably expected mithout major 1611aCement- Hot-
ever, moue n im provements in the neta1111’y of steel are Chd.elhv

Hum

his ’Concept Iégluly. inc use Of ST%11 8310Unt8 O. rceppfr 3.3 an
DIN v1 0'
alloy has practically doubled the exrected life of 53311 expose
to the weather.
alternate wettir g and dr;1g is rcw considered Its-
most harmful typ e of exposire and t s exuus11e‘1sy1 evitab:e* :fifir
a steel pier unit. He must not, ho ver, forge tie na 11:3a1ce-
of perhaps a different application o1 the same materis a1 whic1
prove tiat with reasonable maintenance th e lifefio1 a steel»
struct1.re under, hese conditior s is much over so years. -ue

(38)

' w-
L

1
..

)flfVH-

0,01va

Brooklyn Bridge in New York City vas built over fifty years ago
and, while it is now very inadequate to meet the demands placed
upon it, the original steel cables are still in use. The huts
bridges being built in the Ear Francisco area are certainly not
eXpected to have a usable life of only 80 years for the probability
is that they may not be paid for by that time. The Carnegie

Steel Company in their handbook on the use of steel piling make

an extensive report on the condition of piling removed at
Pittsburg which had been in place or twenty years. This piling
was a 12 inch 35 lb. section,known at the time as I—lOl of
ordinary cart n steel with a carbon content of only .01 percent

and was drive en feet into hard mud and gravel with the tops
extending 18 inches above normal water level. Sections were cut

at the top of the pile, at the water line, at the riddle of

the pile, and at the bottom of the pile and the loss of metal die
to corrosion and other causes carefully determined. At the water
line there was a loss of 14.9 percent or less than ne percent
annually over the time the pile was in place. scenes of mining
activities 'n the Pittsburg region, the water of the Ionongahela
River contains more or less free sulphuric acid which during the
low water months runs es high as 100 parts per million. This,
therefore, was a severe test of the piling and proves definitely
that even plain carbon steel has a usable life of more than

twenty years. Copper bearing steel can therefore be expected

to have a life of at least forty years or more and past exper-
ience with masonry has shown that in the length of time mentioned,
the best of concrete structures will need very major repairs.

0

:3 C)»

There is no questioning the iact that maintenance
charges on the all steel bridge will be higher than for masonry.
Kot that there will be major items for the replacement of struc-
tural menbers, but that a steel structure has to be painted quite
frequently if it is to endure. The cost of painting the average
steel deck girder bridge on concrete abutments is shown by the
report of the nichigan State Highway Department for the year
1982.to be approximately $850.00. Suppose for purposes of dis-
cussion we double this figure to take into account the additional
surfaces to be painted, in the case of the all steel bridge,
making the charge for painting $500.00. Taking the Specific
case of the bridge over the Looking Glass River north of the
Lansing City Airport, it has been shown that the saving in first
cost of this structure over an ordinary steel deck girder bridge
on masonry abutments would be approximately 3 0,000. Certainli
this amount would represent a large number of paintings, so many
that it is safe to assume that the increased cost of maintaining
the all steel structure will never become a major item in the
choice of bridge type for a site where the all steel bridge can
be used.

It has been the purpose of this thesis to set forth in
some detail, the possibility of fabricating and erecting an all
steel bridge by the process of arc welding, and to prove that
the type of bridge proposed is practical from both an engineering
and a financial standpoint. To this end, the process of electric
arc welding was discussed in considerable detail in order that

the reader should have a clear understanding of the methods to be

(as)

used and the results that could be reasonably expected. Th

data included in regards the strength of welds made with the

use of bare wire electrodes si 018 that such welds will permit

a factor of safety of four which is common for steel construction
It must also be borne in mind, in this connection, that the use
of bare wire electrodes has been almost discontinued in favor es
the heavily coated, shielded arc electrode which will practically
guarantee a.1 increase of twenty- -five percent in the strength of
the weld over similar welds made with bare wire.

The type of structure preposed was then described in
great detail with drawings added to show typical details of con-
struction. The advm ntages and disadvar tases of the new type
were di SCJSSGd and compared with mas cnry construction. many of
the details of fabrication and erection difficulties were set
fcrth at 1d the ap;:1opriate sol ition explained. The limitation.
to be placed upon th -e use of the new type were set forth. And,
finally, the cost of biilding and maintainins tt is new type of
bridge structure was compared with similar costs for masonry or
riveted steel structures.

The author believes that the following conclusions may
be arrived at ar 1d ttat there is ample justification for so doing

A. That are weldin” as a means of fabricatins and

E: ..
erecting steel structures is a dependable process and that its
use will permit many economies in design and erection.

B. That a bridge design based on the use of steel
sheet piling for supporting abutments and piers is sound from
an ensir1eering standpoint.

C. That the limitations which have to be placed upon
the use of suc11e structure and t11e difficulties to be ex; acted
in fabrication and erection do not preclude t11e general adoption
of the proposed methods.

D. That the strUcture preposed can be built at a
considerable ea Villg of both time and money.

- E. That the life of such a structure will compare
favorably With the life of br dges constructed on masonry abut-
ments a1d piers.

Acvvowrrn sisirs

bealu 1.1.1

The author wishes to acknow le
man3 for criticisms a.r nd sugzestions con C
and center ts of this thesis. To Cyrus A. Lelick, formerly
Bridge Engineer of tt 6 Iichigan State Highw e3 Decartment, for
the opportun1ity to study the design and construction of welded
bridges and to do research wori on the strength of cert ain
types of welded joints. To A. C. Benkelma an, Eformerly of the
Pesearch Pi s ion of t e Iich can State highway Depar
for mich help in the conduct of such research work and many
suggestions as to th ebest means of analyzing and par '
the data obtained. To ma13 of he autt1ors associates at the

dg his indebtedness to
es rnin~ tb1e preparation

(30)

Fridge Department of the Xichigan State Elislway -ecert1eut, and
to the 1any 1crrecen1at1"e- of tEe u.anafact1ring 9015:1160 who
have giver so freely of their time that the author might have a
more complete knovledse of tte possib1ilities of electric arc
welding. W'ith 1t this help it we 11d rave been dif ficult indeed
to prepare this thesis.

()'

 

BIBLIC HAPPY
A Symposium on Arc Telding
collected papers presented at meetin e of t‘e
A1erican fielding 8001ety.

Arc Welding Data

booklets publish d by the Testing101ee E.& E. Co.
Engineering Eews-Fecorfl
articles w1itten diriag the last ten years by
various author1 and prir ted therein.
F01rteenth Biennial Report of the State Highwa" Commis-
sioner of iichigan for data as to tte actual cost of
c3u°t11ctio1 of various type S of br gee.
Traie Pitlice ions and Catalogues
particularly those of the GCJETPI llectric 00.,
Lincoln Electrlc Cc., Westinfh01se Electric &
11g. Co., and Harnischfeger Corp.

14 46

E :

allfl 1..., 1‘ . 1‘ V

£01.; in]. .. _u. . - 1-

unit"

 

 

 

I

l
|

5

177 590

I'll
'1
'l
IIIII
||II||
II
Ill
I||I|||l
I
I'll
IIIIIII
IIIII
|||
ll
ll
ll
Ill:
III
I‘ll!
||||
ll
I'll
I||

93 03

3 1

 

|HIHIUIWIIIHIHUHII

I
4