Field Welding

Download as pdf or txt
Download as pdf or txt
You are on page 1of 16

Field Welding to Existing Steel Structures

DAVID T. RICKER

Field welding to existing members is becoming increasingly


common. During the recycling of older structures, new
performance requirements often require addition of
reinforcing material to increase load-carrying capacity, to
restore areas eroded by corrosion, to strengthen fireweakened members or perhaps alter the appearance of a
member by changing its shape for aesthetic reasons. One of
the many advantages of a steel-framed structure is that it can
be reworked more readily than structures of other materials.
Field conditions are often far from ideal and it is
necessary to ascertain the effects of the field work on the
existing structure, especially the common method of
attachmentwelding.
Over the years, various rules-of-thumb have been
applied: no welds where calculated stress exceeds 50% of
working stress; no transverse welds across tension members;
no welds where calculated stress exceeds three ksi unless
shored; make all welds parallel to stress lines, etc. Most of
these made sense at the time and were the result of practical
experience plus common sense.
Safety, economic considerations and the endless search
for understanding and refinement require us to delve deeper
into the subject of field welding to existing steel structures.
These items merit attention:
1. Weldability of existing and new steel
2. Selection and design of the weld
3. Anatomy of the weld
4. Heat input
5. Position of the weld
6. Surface conditions
7. Weather conditions
8. Nature of the load
9. Nature of the reinforcing
10. Shoring and stress relieving
11. Reinforcing connections
12. Effect of field alterations on the entire structure
13. Fire hazards and precautions
14 Testing and inspection

In pursuing these topics, it is necessary to address three


parties in the discussionsthe designer (engineer), the
fabricator and the erector (welder). These remarks are often
directed to all three simultaneously, but occasionally only to
one or two. It is hoped the temporarily uninvolved party
bears with us until his turn comes again. Some very basic
information is included for the benefit of those lessexperienced who desire to increase their understanding of the
subject. Some information is repeated in certain chapters to
benefit those who seek specific knowledge or who cannot, for
some reason, read the entire article.

David T. Ricker, P.E., is Vice President, Engineering, of The Berlin


Steel Construction Company, Inc., Berlin, Connecticut.

*This paper was developed as part of ASCE Committee on Steel


Building Structures.

FIRST QUARTER / 1988

1. THE WELDABILITY OF STEEL


By 1900, steel had almost replaced completely wrought iron
as a building component, although the latter continued to be
used for certain applications mainly associated with its
relative imperviousness to atmospheric corrosion. Cast iron,
once a popular material for columns and the fittings used in
heavy timer construction, was fast dying out. Steel became
the dominant building metal. Prior to about 1910, there was
little standardization in the industry. Each steel producer
used his own recipe and rules. This resulted in a wide variety
of chemical and mechanical properties. Some of this early
steel is very weldable and some is not. But the fact that it
may have been made after 1910 does not guarantee
automatically its weldability.
The obvious first test for weldability is to examine the
existing steel work to see if welding was used during the
original fabrication and erection or if the structure has been
successfully welded onto previously.
Any steel whose weldability cannot be confirmed by such
positive evidence, or whose chemical and mechanical
properties cannot be verified with mill certificates or other
documentation, should be tested. This can be done by cutting
or drilling samples from redundant parts of the existing steel
and having them mechanically tested for ductility and
chemically tested for carbon content and other ingredients.
The test samples should be taken from the thickest parts of

2003 by American Institute of Steel Construction, Inc. All rights reserved. This publication or any part thereof must not be reproduced in any form without the written permission of the publisher.

the members, since these areas usually exhibit the greatest


property fluctuation (Fig. 1). If possible, test the specific
members on which the welding is to be performed.
A simple on-site test can be made by welding a lug of
weldable steel to the existing member and beating it with a
hammer (Fig. 2). Refer to the Fillet Bend Test as described in
AWS Spec. D1.1, paragraphs 5.38.3 and 5.39.3.2 If the weld
deforms without fracturing, the steel can be considered
weldable. If the weld separates from the base metal at the
junction of the weld and base metal it indicates the base
metal is subject to hardening, often the result of high carbon
content. The separation exhibits a gray granular surface and
usually follows the profile of the heat affected zone,
appearing as a shallow crater in the base metal. This is not a
very sophisticated test (about as refined as the hammer test
on welded shear studs), but it is better than nothing and is
useful in emergencies or when a testing agency is not readily
available. It is not recommended if other methods of
determination are available. The use of low-hydrogen
welding electrodes and preheating can improve the
weldability of most base metal.
If it is suspected the existing material is wrought iron,
welding should be avoided if possible. If welding is
mandatory, the welds should be oriented so tension loads are
not exerted on the member in a direction perpendicular to the
direction of rolling. Wrought iron tends to contain slag
inclusions which form in layered laminations and does not

Figure 1

Figure 2

possess good trans-lamination strength. It is recommended


that a competent, experienced welding authority be consulted
before attempting to weld on wrought iron. The same
recommendation pertains to welding to cast iron.
Three of the ingredients of structural steel have
considerable effect on its weldability: carbontoo much of
which results in high hardenability and eventual loss of
ductility, excess phosphorus which increases brittleness, and
high sulfur which often results in porous welds.3 The
American Society for Testing and Materials (ASTM) lists
bounds for these and other ingredients of steel by which we
can determine its weldability.
If the chemical properties of the existing steel are known,
the weldability can be determined by the carbon equivalent.
There are quite a few carbon equivalent formulas in
circulation today. A carbon equivalent of less than 0.48 in
the following formula in Ref. 8 generally assures good
weldability.
Carbon equivalent = C + Mn/6 + Cr/5 + Mo/5 + V/5 +
Ni/15 + Cu/15 + Si/6
where C
Mn
Cr
Mo
V
Ni
Cu
Si

= carbon content (%)


= Manganese content (%)
= Chrominum content (%)
= Molybdenum content (%)
= Vanadium content (%)
= nickel content (%)
= copper content (%)
silicon (%)

Another comprehensive carbon equivalent formula in Ref. 7


is:
Carbon equivalent = C + Mn/6 + Ni/20 + Cr/10 + Cu/40Mo/50-V/10
If the results of this formula are below 0.40 the material is
readily weldable. For values between 0.40 and 0.55, use of
preheat or low-hydrogen electrodes is suggested. Values
above 0.55 indicate an increased likelihood that cracks may
develop unless special precautions are exercised
precautions which are not likely to be feasible under field
conditions.
When a carbon equivalent formula is used, the laboratory
analysis report should be required to list the quantities of
each of the elements in the formula, even if the percentage
reported is zero.
Much of the steel used in buildings in the first 60 years
of the 20th century was ASTM-A7, a medium carbon steel.
This is generally accepted as being a weldable steel, but it
had a wider range of permitted carbon content than its
ASTM-A36 eventual successor. The interim ASTM-A373
steel generally is considered to have good weldability.
Although weldable with most electrodes in the E60, E70
classes, certain levels of carbon and other ingredients in
ASTM-A7 steel may require low-hydrogen electrodes and

ENGINEERING JOURNAL / AMERICAN INSTITUTE OF STEEL CONSTRUCTION

2003 by American Institute of Steel Construction, Inc. All rights reserved. This publication or any part thereof must not be reproduced in any form without the written permission of the publisher.

preheating. The following carbon equivalent formula in Ref.


7 is useful in determining preheat requirements:
Carbon equivalent = C + Mn/6 + Ni/15 + Cr/5 + Cu/13 +
Mo/4
The preheat temperatures suggested for several ranges of
carbon equivalent are:
Carbon equivalent Suggested preheat temperatures*
0 to 0.45%
preheating optional
0.45 to 0.60%
200-400F.
over 0.60%
400-700F.**
If excessive carbon equivalent values are indicated, then
further investigation of the existing steel should be
considered, such as destructive testing. Before welding to
existing steel, its weldability must be established positively
by one of the methods mentioned. The AISC publication,
Iron & Steel Beams 1873 to 1952 is helpful in identifying
and dating hot-rolled sections prior to 1952. It also lists some
allowable stresses of that era.
From the foregoing, it should be evident the higher the
carbon content the less adaptable is the steel to welding and
for unweldable situations other means of fastening should be
considered. Other means usually refers to bolting.
The past service record of the structure may be very
helpful. Were members subjected to fatigue, impact,
vibration, stress reversal, cyclical loading, a corrosive
atmosphere or extremes of temperatures? A record of this
past history, together with the intended future use, helps to
determine the anticipated extended life of the rejuvenated
structure and the means to accomplish it.

3. Try to avoid cutting across stress lines with the weld,


especially if the member is carrying load. If
unavoidable, then consider shoring or reducing the load
to a safe level. If undercut is a potential hazard, design
the weld so that any undercut will occur in the new
material where it can be accounted for (Fig. 4). If an
undercut occurs, especially in a member loaded in
tension or subject to fatigue, it should be ground out,
filled with sound weld metal, and ground smooth. If
possible keep the grind marks parallel to the stress lines
or reduce the grind mark profile by finishing up with
light pressure.
5. Avoid biaxial and triaxial stresses at welds or in the
immediate vicinity by altering the geometry of the joint
(Fig. 5).
6. Avoid overwelding, both excessive size or quantity,
since this may accentuate distorsion and shrinkage.
7. Avoid abrupt geometric discontinuities which are located
at welds. This is a major source of weld failures,
especially in tension-loaded members (Fig. 7).

2. SELECTION AND DESIGN OF THE WELD 4,5


The following suggestions may aid in deciding on weld
selection and placement:
1. Use fillet welds in preference to groove welds where the
design permits. Groove welds are more costly and often
result in overkill. Fillet welds are made easily and are
dependable.
2. Arrange the welds if possible so they are made in the
flat or horizontal position (Fig. 3).

Figure 4

Figure 5

Figure 3

FIRST QUARTER / 1988

*If AWS specifications prevail it is recommended that AWS Spec.


D1.1latest edition (Table 4.2) be followed.
**Some codes recommend that maximum preheat not exceed 450F.

2003 by American Institute of Steel Construction, Inc. All rights reserved. This publication or any part thereof must not be reproduced in any form without the written permission of the publisher.

Steel
Specification
ASTM A7
A9
A140
A373
A36
A53
A500
A501
ASTM A7
A373
A36
A53
A242
A441
A500
A501
A572 Gr.50
A588

Minimum Preheat and Interpass Temperatures for Common Building Steels


Welding
Process
Material Thickness
Min. Preheat Temp. - F

SMAW-other
than low
hydrogen
electrodes

SMAW-with
low hydrogen
electrodes
SAW
GMAW
FCAW

Up to 3/4 Incl.

None

Over 3/4 thru 1-1/2

150

Over 1-1/2 thru 2-1/2

225

Over 2-1/2

300

Up to 3/4 Incl.

None
50

Over 3/4 thru 1-1/2

Over 1-1/2 thru 2-1/2

150

Over 2-1/2

225

Figure 6

Matching Filler Metal Requirements for Common Building Steels


Steel
Specification

Min. Yield
Point (ksi)

Min. Tensile
Strength (ksi)

Matching Electrode
Specification

ASTM A7
A9
A140
A36
A53 Gr.B
A500
A501

25-33
25-33
33
36
35
41
36

50-65
50-65
60-72
58
60
58
58

SMAW
AWS A5.1 or A5.5
E60xx or
E70xx
SAW
AWS A5.17 or A5.23
F6x-Exxx or
F7x-Exxx
GMAW
AWS A5.18
ER 705-x
FCAW
AWS A5.20
E6xT-x
E7xT-x
SMAW
AWS A5.1 or A5.5
E7015, E7016
E7018, E7028
SAW
AWS A5.17 or A5.23
F7x-Exxx
GMAW
AWS A5.18
ER705-x
FCAW
AWS A5.20
E7xT-x

ASTM A242
A441
A572
A588

42
42
42
50

63
63
60
65

Min. Yield
Point (ksi)

Min. Tensile
Strength (ksi)

50
60

62 Min.
72 Min.

50
60

62-80
70-90

60

72 Min.

50
60

62 Min.
72 Min.

60
60

72 Min.
72 Min.

60

70-90

60

72 Min.

60

72 Min.

Figure 6a

ENGINEERING JOURNAL / AMERICAN INSTITUTE OF STEEL CONSTRUCTION

2003 by American Institute of Steel Construction, Inc. All rights reserved. This publication or any part thereof must not be reproduced in any form without the written permission of the publisher.

8. Consider a joint design that uses the least volume of


weld. Least weld usually means least heat and least
distortion (Fig. 8).
9. If appropriate, use a partial-penetration weld in place of
a full-penetration groove weld (for column splices, for
example, and other static compression welds).
10. For static loads, consider intermittent fillet welds in lieu
of continuous welds. Do not use intermittent welds for
cyclical loads which may result in metal fatigue since the
allowable stress for this type is heavily penalized. See
AISC 8th Edition Manual Appendix "B".6
11. After the desired weld has been determined, make sure
the proper weld symbol is used on the drawings.
12. Orient welds so contraction strains are imposed on the
base metal in a longitudinal direction, to diminish the
possibility of lamellar over stress.
13. Follow AISC Spec. 1.5.3 and AWS Spec. D1.1-Table
4.1.1 with regard to matching weld metal. Complete
penetration groove welds perpendicular to tension stress
lines must have weld metal which matches the base
metal. Most other welds can be made using weld metal
with a strength level equal to or less than matching weld
metal. Weld metal one strength level stronger than
matching weld metal will be permitted for most welds.
But it is not recommended, because higher strength welds
have less ductility and the potential for cracking
increases. (Fig. 6).
14. Avoid arc strikes and splatter in highly stressed areas.
These cause hard spots which can initiate cracks in the
base metal, especially if the carbon content is high.
15. Do not attempt to weld to cast iron members. Special
procedures are required which may be impossible or at
best difficult to accomplish in the field.
3. THE ANATOMY OF A WELD 3
Most field welding in building construction is done by the
arc-welding process. This, in effect, creates a small electric
furnace at the adjacent bounds of the material to be joined. It
takes two electrodes to make an arcone is the steel building
and the other the welding electrode (rod). The welding rod,
the smaller of the two by some degree, is designed to melt
and be consumed into the molten pool in combination with
the part of the structure which also melts.
Immediately adjacent to this molten pool is the heataffected zone (HAZ), a region about 1/8 in. to 3/16 in. in
width between the molten mixture of deposited weld rod plus
base metal and the heated but unchanged base metal (Fig. 9).
In the HAZ, changes occur in the molecular and grain
structure of the heated steel, and ductility and hardenability
are affected. But oddly enough, relatively few weldassociated problems start here. Cracks usually do not start in
the HAZ. They may start in unsound weld metal near the
HAZ and travel into the base metal or viceversa.3

FIRST QUARTER / 1988

Figure 7

Figure 8

Figure 9

2003 by American Institute of Steel Construction, Inc. All rights reserved. This publication or any part thereof must not be reproduced in any form without the written permission of the publisher.

The temperature of the molten pool is near 3,000F.


(Steel melts at roughly 2,800F. Fig. 10). The HAZ is in the
1,250 to 2,700F. range. The base metal temperature ranges
from about 1,250F. on down to the ambient temperature.
Some other significant temperatures are: 1,333F., above
which austenitic changes take place; 1,500F., the yield
stress is near zero and steel loses its magnetic properties.
This is one reason why many codes limit the temperature of
heat applications to about 1,200F. maximum (for heat
cambering or straightening for example). At 500-600F.,
steel attains its greatest strength, but at the cost of some of its
ductility.
The significance of all this to the subject at hand is that

when we weld to an existing building member it is probably


carrying a load. If so, we must not jeopardize or impair its
ability to do so by melting away too much of the loadcarrying area. That makes a lot of sense. More on this later.
Weld metal is usually more pure than the metals it joins.
A properly executed weld is almost always superior in
mechanical properties to the base metals.
Much of the field welding today is done by the shielded
metal arc process (SMAW), using manual stick electrodes.
Another popular method is flux-cored arc welding (FCAW),
a semi-automatically fed hollow-wire electrode with flux
inside the hollow.
One does not have to be an engineer or metallurgist to
lay a good weld, but a welder who develops an understanding
of the basics of the subject can improve markedly the quality
of his work.3 Failures in service rarely occur in properly
designed and executed welds. Most fractures start at a notch,
groove, discontinuity or other type of stress-riser.

4. HEAT INPUT 3,3.1

Figure 10

The welding process, by definition, produces heat. The


members to be connected must be within a certain
temperature range prior to the start of welding, the weld must
be applied as rapidly as conditions will permit and the
welded area must be allowed to cool properly. The process,
routine under shop conditions, is often more complicated
when applied to existing members under field conditions. But
the same sets of rules apply. The AWS Spec. D1.12 has
established easy-to-follow preheat rules. Table 4.2 in the
same specification shows minimum preheat and interpass
temperatures for various types and thicknesses of steel.
Maximum preheat should generally not exceed 450F. The
preheat temperature must extend a minimum distance of 3 in.,
or the thickness of the material if over 3 in., in any direction
from the weld and must be maintained adequately in advance
of the weld. When two different strength steels are being
welded together, the preheat for the higher strength steel
governs. On multi-pass welds the proper interpass
temperature must be maintained. The purpose of preheating is
to slow down the rate of cooling. Welds cool principally by
conduction. Thus, a thicker member will draw heat away
from a weld faster than the same weld on a thinner member.
This is one reason why thicker plate welds are more prone to
cracking and the required preheat temperatures are higher
than for thinner material (Fig. 6).
Welds also cool by convectionheat lost into the air.
Thin plates are more susceptible to cooling by convection.
Rate of cooling should not be confused with time required for
cooling. Thick material may have a faster rate of cooling but
the elapsed cooling time will be longer because many more
BTU's must be dissipated than on a thin plate which has
much less heat input. When welding is done transversely
across existing members, it may be necessary to perform the

ENGINEERING JOURNAL / AMERICAN INSTITUTE OF STEEL CONSTRUCTION

2003 by American Institute of Steel Construction, Inc. All rights reserved. This publication or any part thereof must not be reproduced in any form without the written permission of the publisher.

welding in several stages to allow cooling time in between.


This procedure results in residual stresses, but these are
usually of little consequence.
The temperature of the molten weld pool is about
3,000F. or greater. At these temperatures, the internal
structure of the metal has changed. The extent of the change
depends on the maximum temperature, the length of time this
temperature is sustained, the composition of the metal and the
rate of cooling. Steel maintained at a high temperature for an
extended length of time produces large grain size and results
in a cooled steel which is harder and less ductile. For this
reason it is important the cooling process not be delayed.
As the molten area cools the motion of the atoms, as they
skittle around seeking their rightful place in the grand scheme
of things, becomes slower and slower until the freezing point
of the metal is reached. At this stage the random motion of
the atoms ceases and they have all joined hands and assume a
fixed location with regard to each other, forming a solid
metal. As the temperature further decreases, the atomic
motion continues in the form of vibrations. Atoms arrange
themselves in space lattices, each comprising a unit cell. A
crystal (or grain) of metal may contain millions of unit cells.
Rapid cooling usually results in the formation of many small
grains because the unit cells do not have time to link hands
with more than a few hundred thousand neighbors. A finegrain texture results. Slower cooling permits more unit cells
to join together into larger (but fewer) grains. The larger
grains have greater spaces between them and the
intergranular stresses are higher.
When steel is welded, a melting process that has been
done previously is repeated. This is another of the many good
qualities of steel. It can be re-melted and still regain its
strength. The re-melting causes relief of the internal stresses,
agitation of the atomic structure, but eventual recovery of the
familiar properties we expect of steel. For these reasons, a
new weld bead can be laid on top of a weld bead which may
be many years old, and it will be effective and reliable.
A very important factor is the amount of heat applied
both from preheat and from the actual welding process. This
should be kept to a minimum and the welding process
completed as quickly as practical. To avoid excessive heat,
use arc welding process, intermittent welds if conditions
permit and skip around. And on multi-pass welds allow the
weld to cool between passes.
Why dwell on this matter in discussing welding to
existing members? When we create our little electric furnace
during the welding process, we affect both the existing
member and the new reinforcing piece. Although steel
actually gains strength up to about 600F., it then starts to
taper off, and at 1,500F. it has hardly any strength at all.
When the member being welded is carrying a load it is
important to know the extent to which we diminish its ability
to maintain the load. (The same can be said of any
loadcarrying member whether we attack it with torch, drill or
weld rod). When we create a molten weld pool we must know

FIRST QUARTER / 1988

how much of the cross section is affected and to what extent.


We have already mentioned that steel reaches its maximum
strength at about 600F., at which time it starts to trail off.
At about 650F., it has the same strength as at 75F. (room
temperature). The temperature of the steel dissipates rapidly
as we move away from the weld so that in thick (over 3/8 in.)
material 650F. is attained about an inch away from the pool.
In thinner material welded at slow weld speeds, this distance
is about 1 in. The danger area seems easy to calculate. But
remember, the welder is advancing his weld pool. How fast
does the molten pool cool down to 650F. behind him to the
point where we know it will be strong enough to again carry
its share of the load? This depends on the mass of the
material and heat input, and can vary from a few minutes to
over half an hour. If welding is done parallel to stress lines,
this cooling time has less significance. But if the design
requires a weld perpendicular to stress lines, the welder must
be aware of the amount of cross section temporarily taken
out-of-action at any given time. Temperature crayons or other
suitable means should be used to monitor the temperature of
the worked-on pieces.
Weld cooling is as important as heating up. When welds
cool too rapidly or too slowly, crystalline changes occur.
Internal stresses can increase to a troublesome point,
sometimes causing local failures (otherwise known as
cracks).
For low-carbon (0.10% to 0.30%) steel, cooling rates are
relatively insignificant. However, a slower cooling rate
results in more toughness and ductility, and rapid cooling
gives harder, more brittle steel. For members subject to
tensile stresses this is important. For compression members
ductility is less important.
Another method used to delay the cooling rate is post
heating. Post heating prolongs the cooling cycle and allows
the hydrogen and other harmful gas bubbles to dissipate from
the molten weld. Post heating is seldom required, except
under extenuating circumstances, such as if preheating were
not performed. Post heating requires heats to about 1,100F.,
far greater than the 450F. maximum for preheating. Post
heating in the field often presents logistical problems.
Preheating, much more effective than post-heating and
generally less expensive, should be given priority.
Any discussion of the heat of welding must include the
root pass. The root pass, especially when welding on thick
material, is the most critical part of the weld. It is subject to
the greatest rate of cooling (by conduction) and, since it is
the thinnest part of the connection (at this stage), is subject to
cracking (Fig. 11). If a root weld crack is detected it must be
removed by grinding or air-arcing, since it is likely to spread
into subsequent layers of the weld.
As a weld cools, stresses build up similar to the residual
stresses in any hot-rolled shape. The established allowable
stresses recognize the almost universal presence of these
welding residual stresses. Residual stresses have little
influence in failures unless there is a flaw, notch, abrupt

2003 by American Institute of Steel Construction, Inc. All rights reserved. This publication or any part thereof must not be reproduced in any form without the written permission of the publisher.

have a long history of reliable performance. The surfaces of


heavy sections are not susceptible to the same deficiencies as
the cores. Consequently there is no problem in welding to the
surface of these members, provided the rules for weld size,
preheating and post heating are observed.
5. POSITION OF THE WELD
There are four weld positions: flat, horizontal, vertical and
overhead (Fig. 3). Flat and horizontal welds are the easiest
and most economical, and usually the most reliable and best
quality welds.
Figure 11

Figure 12

Figure 13

change in cross section or other stress riser present, which


concentrates their effect. If the cooling weld is prone to crack
it will usually have done so before the steel passes below
200F. However, some cracks are known to have started
several weeks after welds were made. All weld beads should
be inspected carefully after cleaning.
There has been some concern in recent years as to the
wisdom of welding to heavy wide-flange sections with parts
exceeding one and one-half inches in thickness, especially
when these members are used in tension applications (Fig.
12). The problem is that cross sections of these heavy
members are sometimes non-homogenous, the result of
slower cooling of the thickest parts which causes pockets of
large-grain structure and high residual internal stresses.
Segregation of the various ingredients of the steel may also
occur.
If groove welding, rough grinding, flame cutting,
undercuts, notching, arc strikes or other similar work is
performed in the core area serious stress increases can occur
and may spawn cracks. These cracks are sometimes
subsurface and not easy to detect. Partial penetration welds
in heavy tension members should be avoided because of the
presence of stress concentrations in the root area.
However, heavy sections used in compression situations

Unfortunately, in welding to an existing structure, the


members are often at awkward or inaccessible locations. This
presents a real challenge to the designer to devise a scheme
and select reinforcing members most convenient to the welder
and at the same time efficient to their performance (Fig. 13).
Fillet welds can be made in all positions with relatively
little difficulty. Flat and horizontal groove welds are also
fairly easy to accomplish. Vertical and overhead groove
welds are more costly because of the greater volume of weld
metal deposited and the relatively slow rate of deposition.
Different types of welding electrodes are available for
welding in different positions. Fast-fill electrodes melt
rapidly and are best suited for flat or horizontal welds where
a high rate of deposition is called for. Fast-freeze electrodes
solidify rapidly, hence are often used for vertical and
overhead welds. Fill-freeze (also called fast-follow)
electrodes are a compromise between the fast-fill and fastfreeze rods. A properly trained, experienced welder knows
the correct type and diameter of electrode to use without
advice from design drawings. The designer should indicate
the proper classification of electrode to use. For example:
low-hydrogen rods for high-strength steels (E7018, E7019,
E7028); E8018-C1 or C3 electrodes when a good color
match is required for weathering steels; E7018 on weathering

ENGINEERING JOURNAL / AMERICAN INSTITUTE OF STEEL CONSTRUCTION

2003 by American Institute of Steel Construction, Inc. All rights reserved. This publication or any part thereof must not be reproduced in any form without the written permission of the publisher.

steels where a color match is not important; low-hydrogen


electrodes where adverse weather conditions are anticipated
(such as high humidity or low temperatures). Most
manufacturers of welding electrodes publish comprehensive
guides on good welding practice, including electrode
selection. Readers are urged to obtain these useful guides.

6.

6. SURFACE CONDITIONS
Existing members are often in deplorable condition,
corroded, covered by many layers of paint, oil, grease or
other contaminants. These fouling elements should be
removed prior to welding. Severely corroded steel must be
cleaned to bare metal and any feathered edges removed by
burning or grinding to present a proper weld surface. Most
other surface contaminates can be removed with flame,
brush, scraper or grinder. If the contaminates cannot be
removed because of inaccessibility or extra tight adhesion,
then the first surface passes should be made with E6010 or
E6011 electrodes. A slow travel speed allows time for gas
bubbles and contaminates to rise to the surface and boil out
where they can be cleaned away prior to the next passes.
E60XX and E70XX electrodes can be used at the same weld.
If the existing members are galvanized, galvanizing should
be ground off in the area to be welded. Good ventilation
should be provided so the welder will not inhale any of the
noxious gases produced by welding inadvertantly through a
galvanized surface.
7. WEATHER CONDITIONS
When the welding must be performed in sections of the
structure exposed to weather it is wise to heed this advice:
1. Do not weld if the temperature of the steel in the vicinity
of the weld area is below 0F. Steel, which possess great
toughness (ductile strength) at normal working
temperatures, becomes brittle at lower temperatures.
2. When the steel temperature is between 0 and 32F.,
preheat and maintain the steel at 70F. during the
welding process, except where thickness or type of
material requires greater preheat. Refer to AWS Spec.
D1.1 Table 4.2 for preheat requirements.
3. It is just as important to observe preheating requirements
in the summer because of the danger of electrode
coatings absorbing moisture from the air, with
subsequent absorption of hydrogen into the molten weld
pool. The use of low-hydrogen electrodes helps solve this
problems.
4. Do not weld on wet or frost-covered surfaces because of
the danger of entrapping hydrogen from the moisture.
The presence of hydrogen in steel increases its
brittleness.
5. Frigid weather conditions may require post-heating in
order to slow the cooling rate. This allows time for
harmful gases to escape the melted area and helps assure

FIRST QUARTER / 1988

7.

8.

9.

a more ductile weld. The use of low-hydrogen weld


electrodes lessens the amount of preheat and post-heat
required.
Protection from wind should be provided while welding
is done, so shielding gases will not be diluted and blown
away, to retard the loss of preheat and interpass heat and
also for the comfort of the welder in frigid weather. A
steady hand is important to making good welds.
Do not artifically cool welded areas by water spray or
forced air. This invites brittleness and loss of ductility as
compared to the base metal.
Protect welding electrodes from adverse atmospheric
conditions. Abide by exposure time and reheat
recommendations provided by the manufacturer.
If the location of the completed joint is such that it may
become wet, even by wind-blown spray, or dampened by
condensation, the remainder of the joint not involved in
the structural weld should be sealed with additional weld
to prevent moisture entering between the joint plies.
(Other sealants are also available for this purpose).
8. THE NATURE OF THE LOAD 4

Four factors should be considered when welding to an


existing member:
1.
2.
3.
4.

Type of load to be applied


Manner in which the load is applied
Duration of the load
Stresses in the existing member at the time welding is
performed
The types of loads are: tension, compression, shear and
combinations thereof. The loads may occur as static, cyclical
(causing fatigue), vibration, reversible (extreme compression
to extreme tension) and impact.
The history of the member, when known, aids the
engineer in judging if and how to apply the weld. For
instance, a transverse weld on a tension member which has
been subject to fatigue loading over a long duration should be
avoided because of the possibility of adverse granular
changes. (Tests are available for determining the nature and
extent of granular deformations.) Likewise, members subject
to impact loads at extremely low temperatures should be
welded with care, using low hydrogen electrodes. On the
other hand, compression members statically loaded can be
welded with less concern for future performance.
There are a great many other combinations of load
conditions which affect the choice of welds. The following
deserve special attention by the designer:
a. Welds tranverse to tension stresses
b. Welds on members with a long history of fatigue
loading
c. Welds on members subject to extremes of stress
reversal

2003 by American Institute of Steel Construction, Inc. All rights reserved. This publication or any part thereof must not be reproduced in any form without the written permission of the publisher.

When these conditions are encountered, the various rules


and precautions should be observed as set out in codes and
specifications. Proper allowable stress reductions should be
applied for fatigue, impact and other penalty conditions.
Refer to the AISC Manual of Steel Construction, 8th
Edition.6
Some codes require that when a building is reinforced for
one reason or another, the entire structure be upgraded to
meet all current code requirements. For example: placing a
new air-conditioning unit on an existing roof may require
some minor reinforcing. But the current applicable building
code may require a sizeable part of the roof be strengthened
to meet increased snow-drift design requirements, even to
requiring new foundation work. This can be a shock to an
unsuspecting owner.
9. NATURE OF THE REINFORCEMENT
We have previously discussed various aspects of the weld
itself. Let us examine the things being welded. Sometimes the
reinforcement is relatively simple, such as an increase in the
size or length of an existing weld. Sometimes we must add
additional material, either to restore the member to its
original strength or to increase its strength to support a
greater load. The purpose of reinforcement may be to
increase the area of the member cross section or to add to its
stiffness so it will have an increased allowable stress.
Occasionally, material is added for aesthetic reasons, such as
plates to a wide-flange column to change its shape from an H
to a square or rectangle. Sometimes entire sections of a
member must be replaced.
After the member is reinforced, the end connections may
require upgrading to carry greater loads (touched on later).
It is often necessary to add A36 material to high-strength
steel members. If the addition is made for strength,
differences in working stresses must be considered. If new
material is added to increase stiffness or lessen deflection,
then the size and arrangement of the material is of prime
importance. ASTM-A36 material is appropriate to use with
higher strength steels in either case.
If material is added to a column, such as channels or
cover plates to improve bending strength, a tight fit bearing
at base and cap plate is not necessarily required. But if the
new material is added to increase the bearing area, then good
bearing is required. This, however, does not call for a shrink
fit. It is common practice to cut the material short by about
in. or so (depending on material thickness) and to fill the gap
with weld metal to accomplish the proper bearing (Fig. 14).
It is recommended that the member be relieved of as
much load as possible prior to reinforcing. In some cases this
is impractical or impossible. However, the successful
performance of multitudes of structures reinforced while
under stress confirms the fact the procedure is safe and
feasible in most cases.

10

When new material is to be added to an existing member


under stress, some engineers require the area of the new
piece to be computed using a stress equal to or less than the
allowable stress of the base material minus the actual stress
in the base material at the time the new steel is added. This is
a conservative approach. When cover plates are added to a
column, the anticipated allowable stress can be based on the
properties of the built-up section.
Example
Original allowable column stress = 16 ksi
Anticipated allowable column stress = 18 ksi (a result of
lesser L/r ratio after reinforcing)
Actual stress at time of reinforcement = 12 ksi
Stress to be used in calculating the area of the new steel
is 18 12 = 6 ksi.
To increase this usable stress in the new steel it is
sometimes desirable to relieve some of the load on the
original member prior to adding reinforcing. This can be
done by removing live load or by shoring. In the example
above, if the actual stress at time of reinforcement can be
reduced from 12 to 8 ksi, then the usable stress in the new
steel will increase from 6 ksi to 10 ksi.
The reverse is true if the anticipated alteration increases
the column L/r value: for instance, if column flanges must be
narrowed to clear an obstruction but are then reinforced with
narrow plates to restore the necessary area. The following
example illustrates:
Example
Original allowable column stress = 16 ksi
Anticipated allowable column stress (due to an increased

Figure 14

ENGINEERING JOURNAL / AMERICAN INSTITUTE OF STEEL CONSTRUCTION

2003 by American Institute of Steel Construction, Inc. All rights reserved. This publication or any part thereof must not be reproduced in any form without the written permission of the publisher.

L/r ratio of the final reinforced section) = 15 ksi


Actual stress at time of reinforcement (the result of load
relief by means of shoring) = 8 ksi
Stress to be used in calculating the area of the new
reinforcing steel = 15 8 = 7 ksi
In this example, it is shown to be beneficial if the
original section is left intact until the new reinforcement is in
place and welded. However, this is not always possible
because of accessibility requirements for welding.
In the examples just used, we assumed no part of the
reinforced member will be stressed beyond the allowable
working stresses as set forth in the AISC Manual of Steel
Construction.
Before reinforcing steel is added to an existing member,
a decision must be made as to whether to keep the working
stresses within the normal allowable stress range or to allow
some of the reinforced member to exceed the normal
allowable stresses and approach the yield stress. The service
conditions should be considered. Some designers are willing
to permit certain members or parts of them to operate above
normal working stress. The increased deformation from
above-normal stresses should also be considered. When
AISC established the ratio of working stress to yield at 0.6 it
simultaneously set a fairly comfortable limit for deformation,
a limit most construction materials can handle. The excessive
deformation of one or more members may have an adverse
effect on other parts of the structure and also on the
connections. It may be wise to keep the reinforced member
deformation from exceeding that intended for the original
member. This is a matter of engineering judgment. For a
distorted member it must be decided whether to freeze (lock)
the piece into its damaged position or attempt to return it to
its original shape. This depends heavily on the nature and
location of damage and the effect on the rest of the structure.
This accents another good feature of steelheated steel
regains its strength after cooling.
When welding a cover plate to a beam or column it is
wise for the welder to start at one end and work in one
direction toward the other end, or to start in the middle and
work toward both ends. Do not start at the ends and work
toward the middle. Since the cover plate and main member
are rarely the same size, their heat absorption rates vary,
resulting, if the ends are restrained, in a cover plate that
looks like an inch worm.
There may be a temptation when cover plating a
compression member to take advantage of this inch-worm
effect to reduce compression stresses in the original member.
There are many unpredictable variables, such as the amount
of end restraint, exerted on heated members by the rest of the
structure and the rates of heat absorption and cooling of
member components. It seems unrealistic to rely on such
nebulous factors to rationalize a stress reduction.
In a case where a column is reinforced with a pair of
cover plates, it helps assure the final straightness of the

FIRST QUARTER / 1988

column if cover plates are welded simultaneously and directly


opposite each other. Testing at Lehigh University9 of coverplated WF columns produced these observations:
1. Heat generated during welding increased the length of
the column and was followed by a shortening of the
column when the member cooled. When the column
length is held constant (or nearly so) there is a
corresponding change in load. This load change is
transmitted in turn to those other adjacent members of the
structure involved in the restraining action.
2. Welding changed the residual stress distribution in the
original WF column flange tips. In the sections tested
(W831 with flange cover plates 73/8) the residual
stress changes were beneficial. However, in heavier
sections the change in residual stresses resulting from
welding may not work to advantage. Further study is
required on this subject.
3. Tests showed the ultimate stress of the reinforced
members was not reduced by welding.
4. The influence of welding is confined to a very small area
in the weld vicinity. Material properties were not
affected enough to reduce the strength of the section.
After a cover plate is welded to another member it is
assumed the weld deformation is so slight there is virtually
no differential movement between the two elements.
Visualize the loading progressions of a cover plate with zero
initial stress welded to an existing WF member already
carrying some load. If the allowable working stresses of both
pieces are kept within elastic limits the mated elements will
act in perfect unison, each load increment increasing the
stresses and strains in direct proportion. However, if the load
is increased so stresses in the WF members rise past the yield
point into the plastic range, then, since the two elements are
locked together by the welds, the plastic WF will tend to
stretch the cover plate along with it. The cover plate, on the
other hand, resists the effort to yield because it is still in the
elastic range. At this stage, the behavior of the combined
elements, when subjected to further load increment, depends
greatly on the proportion of the areas of the two elements. If
the area of the added plate is relatively small, then a small
load increment will result in the plate also entering the
plastic range. The resulting yield eventually will be arrested
by strain hardening of the member, unless the welds fail or
some other type of failure occurs first. If the area of the plate
is relatively large, it takes a correspondingly larger load to
get it into the plastic range. Early literature10 suggests a safe
condition exists as long as the maximum stress in the original
member is not over one third greater than the average unit
stress of the combined section. It was assumed the average
unit stress would not exceed the allowable working stress.
The question to be asked is this: Is it wise to allow parts
of the structural frame to approach the plastic stress range?
In light of the difficulty in analyzing existing loadings, with

11

2003 by American Institute of Steel Construction, Inc. All rights reserved. This publication or any part thereof must not be reproduced in any form without the written permission of the publisher.

the uncertainties of residual stresses and changes in residual


stresses caused by welding, how much does an engineer dare
to flirt with the yield point? If he keeps all his stresses within
the normal working stresses (0.6 Fy) he should not have a
problem either with strength or deformation. If he decides to
exceed 0.6 Fy, the amount is up to his engineering judgment.
If the reason to add the cover plate is because the normal
allowable working stress would otherwise be exceeded, is it
wise to go to the trouble of adding a cover plate and still
have the original part of the member exceed the normal
allowable working stress? One of the most important
decisions facing the engineer involved in alteration work on
existing steel structures is the selection of the working stress
limit.
AWS Spec. 8.122 lists guide lines for welding together
the components of built-up members with reference to stitchweld spacing and unsupported width and length limits for
plates. The integral parts of a built-up member are assumed
to be connected so they act together, and that a given strain
will produce a uniform stress change in all parts. Contrary to
some belief, it is not necessary the new reinforcing material
match the parent material.
Example:
Original member of A572 Gr.50 steel has an allowable
bending stress = 0.6F = 30 ksi
Actual stress at time of reinforcement = 18 ksi
Usable stress in calculating the area of the new
reinforcing = 3018 = 12 ksi.
A36 reinforcing steel, which has an allowable stress
greater than this, will do the job nicely.
If, on the other hand, to increase the effectiveness of the
reinforcing, the decision is made to relieve the existing
member of most of its load so the actual stress at time of
reinforcing was reduced to, say 2 ksi, then the usable stress
in calculating the area of the new reinforcing becomes 302
= 28 ksi. Since this is greater than the allowable stress for
A36 steel, a higher strength steel must be used.
The load capacity of a non-composite beam supporting a
concrete slab can be increased by adding shear connectors to
make it composite. This is accomplished by core-drilling
holes in the existing concrete slab over the beam at specified
intervals, welding the studs and filling holes with non-shrink
grout. If this method is contemplated, potential deflection
should be investigated. For lack of a better method, welded
shear connectors should be tested with the standard AWS
"hammer test" (AWS Spec. 7.7), especially if the original
beam material is unknown. Adding a bottom-flange cover
plate will increase the beam capacity, especially when used
in conjunction with the added studs (Fig. 15). Even without
the studs, the bottom-flange cover plate can be an effective
means of reinforcing. When the load capacity of any beam is
increased, the existing connections should be investigated to
determine if they, too, need upgrading.
When the area of a tension member is increased by

12

adding angles or plates, reinforcing must be adequately


connected to the gussett plates or whatever type of end
connection exists. The resultant connection must be strong
enough to transfer the new loads (Fig. 16).
If the new cover plates or angles are of a length which
requires they be spliced in the field, the groove-welded splice
should be made and allowed to cool before the new material
is welded to the existing member, so groove weld shrinkage
will not affect stresses in the connected pieces.
If welding is performed across a member, transverse to
the stress lines, when the weld cools it will shrink and
tension stresses appear. If the member was initially in
tension, the weld will return to a tension condition and the
resulting shrinkage stresses may be lesser or greater than
original stresses. If the member was initially in compression,

Figure 15

Figure 16

ENGINEERING JOURNAL / AMERICAN INSTITUTE OF STEEL CONSTRUCTION

2003 by American Institute of Steel Construction, Inc. All rights reserved. This publication or any part thereof must not be reproduced in any form without the written permission of the publisher.

the welded area stresses become zero in the molten stage and
then go into tension as the weld cools and shrinks. This puts
part of the member out of action, or means the remaining
unaffected cross section must take the load, or local yielding
may take place which will return the entire cross section to
compression. In such a case it is likely the stress distribution
across the cross section will be highly variable. As
previously stated, it is preferable to avoid transverse welding
in highly stressed members, if at all possible.
Pre-tensioned reinforcement can be employed on
occasion. Examples of pre-tensioned rods are fairly common.
However, pre-loading of cover plates and similar reinforcing
prior to welding is often very difficult under field conditions.
And any cost advantage usually is cancelled by the added
labor and equipment requirements.
10. SHORING AND STRESS RELIEVING
As just indicated, shoring before reinforcing can make
reinforcing more efficient since it can increase the usable
stress range in the added material. However, sometimes it is
impractical or impossible to shore. The decision as to
whether or not shoring is feasible must be made prior to the
determination of the size and extent of reinforcing.
Shoring may be desired to reduce the deflection of an
existing overloaded member. In this situation, some positive
camber may be forced into the member prior to reinforcing.
In a case where a plate is added to the bottom flange of a
beam, the flange may go into compression when the camber
is forced in, and this should be accounted for when sizing the
cover plate.
Incidentally, the act of welding a cover plate to the
bottom flange of an unloaded beam tends to make the beam
arch upward because of the heating and cooling produced by
welding. This is one reason to keep welding as symmetrical
as possible.
Caution must be exercised when reinforcing a beam
known to be composite is contemplated. It must be
determined whether the original construction was shored or
unshored before concrete work was done. If it were shored
prior to the pour, so that both dead and live load were
resisted by the composite action, there should be no problem
jacking it up to its original no-load position. This no-load
position necessarily may not be represented by a straight
beam. In composite construction it is common to start with a
cambered beam. If design drawings of the original building
are available they indicate the type of composite construction
and whether beams were cambered. (Engineers note - this
information should be placed on the design drawings in
addition to the job specifications. As years pass, the
specification book is often lost, but drawings tend to be
around a bit longer. Someone in the future may appreciate it.)
If the original composite poured construction was unshored,
then it is wise not to attempt to raise the beam. Some relief
can be gained by removing as much of the live load as
practical.

FIRST QUARTER / 1988

The AWS Spec.2 recommends that if the dead load


produces over 3 ksi of calculated stress, the member should
be shored, prestressed or relieved of its excess load prior to
welding. This is a very conservative, safe approach, but is
not often followed in the real world, where welding may be
required under trying or dangerous conditions in a condensed
time frame.
The sound judgment of an experienced engineer in
assessing the effects of cutting, welding and drilling on an
existing structure may save his client needless expense and
lost time by eliminating unnecessary shoring, cribbing,
jacking, scaffolding and associated work.

11. REINFORCING CONNECTIONS


Whenever a member such as a beam or bracing is reinforced
to increase its load-carrying capacity, end connections must
be examined to see if they, too, must be strengthened. The
original connection was devised with something specific in
mind. It will probably fall into one of the three framing
categories classified as Type I-rigid, Type 2 - simple or Type
3 - semi-rigid.
Unless the alteration design clearly states the type of
connection is to be changed and the main members properly
altered to accommodate, then it is wise to keep the
connection as it was originally intended and any reinforcing
work should be so tailored. Failure to do this may result in
the introduction of unwelcome local stress concentrations.
There are many ways to reinforce or improve an existing
connection. Here are a few:
1. Remove and completely replace the original connection
with a new stronger connection.
2. Add weld length or increase weld size. There is no cause
for concern when adding new weld to existing weld as
long as the old weld is sound. The difference in strength
between E60- and E70-class welding electrodes is small.
And the difference is even less in the finished weld,
which is a mixture of pure weld metal and base metal.
3. Remove old rivets or A307 bolts and replace with A325
or A490 bolts, or ream holes and use larger diameter
bolts (worth mentioning even though it does not pertain
to welding).
4. Extend the length of framing angles by welding on
additional lengths (Fig. 17).
5. Add web framing angles to an original seated connection
(Fig. 17).
6. Add weld to existing riveted or A307 bolted connections
(Fig. 18). When this is done the capacity of the old bolts
and rivets is normally discounted and the weld figured to
take the entire load. This is a conservative approach. The
AWS Spec. D1.12 states that where rivets or bolts are
overstressed by the total load, they may be assigned the
task of supporting the dead load, provided they can do so
without over stress. In such cases the reinforcing weld is

13

2003 by American Institute of Steel Construction, Inc. All rights reserved. This publication or any part thereof must not be reproduced in any form without the written permission of the publisher.

added to support the live and impact loads. Despite this


provision, consideration should be given to the following
circumstance. Assume the original fasteners take the
dead load and the new weld carries the subsequent
superimposed loads. However, under the increased load,
local yielding may take place in other parts of the
connecting material. Since welds yield to a lesser extent
than rivets or bolts which are attempting to plow their
way through the connected material, the stiffer weld may
end up taking more than its anticipated share of the final
total load. If there is any doubt as to the final load
sharing between the mechanical fasteners and the

14

weld, it is safer to design the weld to take the entire


load.
7. Add weld to existing high-strength bolted connections.
High-strength bolts, if torqued properly, can be figured at
their friction value when used in the same connection
plane with added weld, the amount of added weld being
sufficient to carry the difference between the total load
and the load taken by the high-strength friction value
bolts. If the original high-strength bolts were actually
figured at bearing value with painted faying surface, it
may be assumed the connection has already eased into

Figure 17

Figure 19

Figure 18

Figure 20

ENGINEERING JOURNAL / AMERICAN INSTITUTE OF STEEL CONSTRUCTION

2003 by American Institute of Steel Construction, Inc. All rights reserved. This publication or any part thereof must not be reproduced in any form without the written permission of the publisher.

bearing. However, with the recent relaxation of torquing


requirements for bearing value high-strength bolts, it is
suggested the bolts, if used in the same plane with welds,
be torqued to friction-value requirements, painted faying
surfaces or not.
8. Add a seat to a web framed connection (Fig. 19).
9. Add a second angle to a single angle web framed
connection (Fig. 20).
10. It is common in renovation work to remove or relocate
members, leaving open holes behind. AWS Specification
D1.1 paragraph 3.7.7 pertains to the restoration of
abandoned (unused) holes in steel members. Abandoned
holes may be left open provided they are not required to
be filled in for structural, aesthetic or weather-tightness
reasons. If the holes must be restored by welding, rules in
the previously mentioned specification must be followed.
Members subject to dynamic tensile stress deserve
special attention in this regard.
Looking into the innards of an ancient clock with all its
wheels, gears, levers, ratchets and cogs, often kindles a
feeling of admiration for the craftsman who devised those
works. But a common structural joint is no less imposing. In
fact, it is more so. In a clock, keen observation comprehends
what is happening. In a joint, many things are happening
which are not visible. Parts are being compressed, stretched,
twisted, bent, often in two or three directions at once. To say
that one completely understands a connection is unrealistic.
Most connections have predictable patterns of performance.
But it is beyond practicality to assign exact magnitudes of
stress and strain to the various pieces or parts of the joint. A
good joint must be designed so wide ranges of stress and
strain do not cause its eventual failure. We know yielding
takes place in parts of most connections. Another one of the
many beauties of steel is its maintenance of strength after
yieldingin fact, it increases in strength because of strain
hardening.
A field-reinforced connection is no less impressive. Here
we may have some bolting and/or riveting combined with
welding. The same rules which apply to new connections
must apply to reworked connections. There may be enough
stress concentrations to send a purist into a cold sweat, yet
the connection as a whole, if properly designed, performs
well.
12. EFFECT OF FIELD ALTERATIONS ON THE
ENTIRE STRUCTURE
We have talked about individual welds and the effects of
field working individual members. It is often wise to step
back for a brief overview of the effects of alteration work on
the entire structure. If we stiffen a beam here and there has it
affected the columns? We cut off a cantilever beamwhat
did we do to the back span? We remove X-bracing here and
add moment connections. Can the columns absorb the added

FIRST QUARTER / 1988

moments? Can the foundations take the change in loads?


Often, in the name of expedience, alterations must be
redesigned and reworked on short notice in the field by the
tradesmen. Such work can be very ingenious and at times a
life-saver. But it should be reviewed by the designer-ofrecord to assure its suitability to the overall structure. A
designer will generally welcome suggestions from the field.
13. FIRE HAZARDS AND PRECAUTIONS
Existing state and local fire codes and regulations and safety
rules must be observed. When welding is done, or cutting
with oxyacetylene torches, grinding or air-arc gouging, redhot metal droplets or sparks may fall on combustible
materials.
Older buildings are often dirty and littered and contain a
higher proportion of combustible materials than modern
structures. If these conditions exist, protection should be
provided in the form of fireproof blankets and readily
available fire extinguishers, or other suitable means. If an
explosive atmosphere is suspected, welding should not be
performed until proper ventilation and all necessary
safeguards are provided. It is often wise to maintain a fire
watch after the work shift is ended and until the next shift
starts work in case latent embers become active.
14. TESTING AND INSPECTION
The owner may wish to have his structure inspected either
during or after work is performed. Inspection during the
welding process is more effective. If a problem is discovered
it can be corrected with less disruption of the welded area.
The inspector should inspect all welds visually for signs
of poor quality or non-conformity. Visual inspection after
welding detects only major surface imperfections, and does
not detect below-surface defects, which can be the most
serious. Non-destructive testing, such as ultrasonic or
radiographic, is one means to examine the interior of a weld
area. However the non-destructive proof of the existence of a
weld flaw does not measure its serviceability. (Only a
destructive test will give this information.) The nondestructive test is merely someone's opinion as to what is
good or bad. The correlation between the flaw and
serviceability is the important issue.
A flawed weld may have adequate quality to be
acceptable. A perfect weld is hardly ever required in building
work. Purity is a matter of interpretation. A flawed but
acceptable weld does not mean the work is sloppy or
careless. It means the degree of contamination, porosity, nonfusion, inclusions or other non-conformities is not great
enough to prevent the weld from performing safely during its
service life.
The engineerand not the inspectorshould be the
judge as to the serviceability of the weld. These factors
should be considered in judging weld quality:

15

2003 by American Institute of Steel Construction, Inc. All rights reserved. This publication or any part thereof must not be reproduced in any form without the written permission of the publisher.

1. Service conditions, stress level, nature of the stress,


working temperature
2. Material properties
3. Risk of defects happening
4. Risk of defects not being found
5. Consequence of failure
If the weld quality is questionable, it is recommended the
inspector consult a comprehensive text on the subject, such as
Chap. 8 of the AWS Welding Handbook, Vol. 5, 7th
Edition,4 prior to passing his conclusions on to the engineer.
Various tests can be performed on the existing steel to
determine certain characteristics. In addition to the
aforementioned chemical analysis to establish weldability,
and the tension test to confirm ductility, tests are available to
determine fracture toughness, cyclic rupture, grain flow (by
photo macrography) and susceptibility to hydrogen
embrittlement. It is recommended a qualified testing
laboratory be contacted to discuss the type of testing most
effective and the size and quantity of samples required.
SUMMARY
1. Establish the weldability of existing and new steel
members.
2. Determine the load condition of existing members to be
field-worked.
3. Be aware of what welding can do to existing members
and plan reinforcing and welding so there will be no
harmful consequences.
4. Account for fatigue, impact and other factors that affect
the design of the new reinforcing and welds.
5. Design the new reinforcing material based on the usable
stress, the difference between established allowable

16

stress in the existing member minus the actual stress at


the time reinforcing is performed. Shore only if
necessary.
6. Make sure connections are reinforced as required to
satisfy new load requirements.
7. Account for effects of the alteration work on the balance
of the structure, including foundations. Check to see that
all applicable building code requirements are satisfied.
8. Make sure welds are properly executed.
If all these things can be accomplished successfully, the
revitalized structure should provide many added years of
service life.
REFERENCES
1. James F. Lincoln Foundation The Procedure Handbook of Arc
Welding Cleveland, O.
2. ANSI/AWS Specification D1.1 Latest edition.
3. Jefferson T.B., and Graham Woods Metals and How to Weld
Them James F. Lincoln Arc Welding Foundation.
4. American Welding Society Welding Handbook Volumes 1
through 5, 7th Ed.
5. Blodgett, Omer Design of Welded Structures James F. Lincoln
Arc Welding Foundation.
6. The American Institute of Steel Construction, Inc. Manual of
Steel Construction 8th Ed., 1982, Chicago, Ill.
7. George E. Linnert Welding Metallurgy Vol. 2, American
Welding Society.
8. Bjorhovde, Reidar, Unpublished notes University of Arizona,
Tucson, Ariz.
9. Nagaraja Ras, N.R. and Lambert Tall Columns Reinforced
Under Load AWS Welding Journal, Vol. 42, April 1963.
10. Spraragen, W. and S. L. Grapnal Reinforcing Structures Under
Load AWS Welding Journal, Vol. 23, Research Supplement 65-S,
February 1944.

ENGINEERING JOURNAL / AMERICAN INSTITUTE OF STEEL CONSTRUCTION

2003 by American Institute of Steel Construction, Inc. All rights reserved. This publication or any part thereof must not be reproduced in any form without the written permission of the publisher.

You might also like