Calhoun: The NPS Institutional Archive

Theses and Dissertations Thesis and Dissertation Collection

1971

Study of flame heating of steel plate.

Johnson, Eugene Karns.

Massachusetts Institute of Technology

http://hdl.handle.net/10945/15919
STUDY OF FLAME HEATING
OF STEEL PLATE

by

Eugene K. Johnson
 STUDY OF FJ M E HEATING OF STEEL \" \TE

by

EUGENE KARNS JOHNSON

B.S., United States Coast Guard Academy

(1965)

SUBMITTED IN PARTIAL FULFILLMENT

OF THE REQUIREMENTS FOR THE DEGREES OF

NAVAL ENGINEER AND

MASTER OF SCIENCE IN

MECHANICAL ENGINEERING

at the

MASSACHUSETTS INSTITUTE OF TECHNOLOGY

May, 1971
 LIBRARY
NAVAL POSTGRADUATE SCHOOII
MONTEREY, CALIF. 93940

STUDY OF F] i : H IATU " OF STE !L PLATE

by

EUGENE KARNS JOHNSON

Submitted to the Department of Naval Architecture and Marine

Engineering on May 3A 1971 in partial fulfillment of the requirements
, 1

for the degrees of Naval Engineer in Naval Architecture and Marine

Engineering and Master of Science in Mechanical Engineering.

ABSTRA CT

The effects of welding and analytical procedures for predicting

welding effects are discussed. The results of previous investigations of
flame heating are analyzed. A flat plate was chosen as a model for this
investigation so that the analytical techniques, developed for welding,
could be applied to the flame heating. Line flame heating is employed
because this closely, resembles bead-on-plate welding. Three steels;
AISI 1020, ASTM A-2^2, and ASTH A-51/t were chosen as materials. Heating
was performed with and without water cooling and continuous readings of
strain and. temperature were taken at selected positions,

The results of testing are presented in the form of temperature and

strain plotted against time. Plots of experimental and analytical
temperatures and strain are presented for comparison. It was found that
flame heating with water cooling is more effective than without cooling,
and the direction of the bending is controlled by the cooling. When
water cooling is not employed, the flame heating is most effective on the
mild steel. However, the direction of bending depends on the initial
plate condition and is not controlled by the heating,

Comparison of the analytical predictions of temperatures and longi-

tudinal strains with experimental data shows good correlation. However,
transverse strains ax^ large for the flame heating so a one dimensional
analysis, assuming transverse stresses are negligible, is not satisfactory
for flame heating. Welding programs, as they are developed, can be
modified for flame heating. A full three dimensional analysis is needed.
to optimize flame heating.

It is recommended that further investigation of flame heating be

delayed until a three dimensional analysis is developed, There is enough
information available now on flame heating to make applications of this
bending or straightening process more effective.

Thesis Supervisor; Prof, Koichi Masubuchi

Title: Associate Professor of Naval Architecture

 ACKNOWLEDGEMENTS

I would like to thank the U. S. Coast Guard for sponsoring my studies

at M.I.T. and this thesis.

I also thank Prof. Hasubuchi and Prof. Argon for their assistance and

many helpful suggestions throughout the course of this investigation. The

assistance of Mr. A. Zona of the Materials Joining Laboratory was greatly

appreciated. The technical assistance of Mr. F. Merlis while preparing

for and conducting the experimental work was invaluable.

1 would also like to thank Mr. W. Laughton of General Dynamics for

allov/ing me to visit the Quincy Shipyard and observe how flame heating is

used in practice.

Finally, I thank my wife, Lynne, for her efficient typing of this

thesis.
 TABLE OF CONTENTS

Pajs ,o

Title Page 1

Abstract 2

Acknowledgements 3

Table of Contents 4

List of Figures 5

List of Tables 7

Nomenclature ,
8

I. Introduction 9

A. Background of Problem 9
B. Previous Investigations 13
C. Purpose of Study 1?
D. Selection of Parameters 18
E. Measuring Strain and Temperature 20

II, Procedures 24

A. Scope of the Research 24

B. Description of Specimens 24
C. Hearting Procedures - 26
D. Experimental Procedures 28

III. Results 40

IV. Discussion of Results 63

V. Conclusions 75

VI. Recommendations 77

VII. Appendix 78

A. Thermal Analysis 78
B. Analysis of Thermal Stresses and
Metal Movement 82
C. Material Composition and Physical Properties 87
]). Tabulated Experimental Results 90

VIII. References 103

 LIST OF FIGURES

_Fi£. Pa^o

1 Examples of spot flame heating as actually used 11

in shipyards

2 Flame heating techniques 12

3 Specimen from previous investigation 15

4 Illustration of the effects of plate thickness on 16

the relative seriousness of distortion problems
'and material problems

5 Curve of apparent strain versus temperature for 22

HT-1212-5B strain gage

2-1 Diagram of instrumentation 25

2-2 Photograph of torch mounted on automatic welding 2?

machine

2-3 Outline of experimental set-up 29

2-4 Thermocouple circuit 30

2-5 Strain gage circuit 31

2-6 Diagram of dimensions used to specify location of 33

heating and water cooling

2-7 Photograph of equipment set-up for recording 38

experimental results

2-8 Photograph of steel plate after one pass of flame 39

heating

3-1-a Experimental results for mild steel - Pass //I 41

3-1-b Experimental results for mild steel - Pass //l 42

3-2-a Experimental results for mild steel - Pass #2 43

3-2-b Experimental results for mild steel - Pass #2 44

3-3-a Experimental results for mild steel - Pass #3 45

3~3~b Experimental results for mild steel - Pass /;"} 46

 Fig- Pago

3-4-a Experimental results for T-l - Pass #1 47

3-4-b Experimental results for T-l - Pass #1 48

3-5-a Experimental results for T-l - Pass f/2 49

3-5-b Experimental results for T-l - Pass #2 50

3-6-a Experimental results for T-l - Pass #3 51

3-6~b Experimental results for T-l - Pass #3 52

3~?-a Experimental results for Cortcn - Pass #1 53

3-?-b Experimental results for Corten - Pass #1 5^

3-8-a Experimental results for Corten - Pass #2 55

3-8-b Experimental results for Corten - Pass f/2 5&

3~9-a Experimental results for Corten - Pass #3 57

3~9-b Experimental results for Cortcn - Pass #3 58

3-10 Analytical and experimental results for mild steel 59

3-11 Analytical and experimental results for T-l 60

3-12 Analytical and experimental results for Corten 6l

3-13 Photograph of mild steel plate after three passes 62

of flame heating

4-1 Illustration of application of line flame heating 69

A-l Model of flame heating 80

B-l Schematic representation of changes of temperature 83

and stresses during welding
 LIST OF TABLES

Page

2-1 Experimental procedures for mild steel 35

2-2 Experimento.l procedures for T~l 3&

2-3 Experimental procedures for Corten 37

^•-1 Predicted values of strain for the second pass of 70

heat versus measured strain
 NOMENCLATURE

Tx Experimental temperature at position indicated by number x

ATx Analytical temperature at position indicated by number x

EXx Experimental strain in x-direction at position indicated

by number x

EYx Experimental strain in y-direction at position indicated

by number x

AEXx Analytical strain in x-direction at position indicated by

• number x_

x t y,z Coordinates of the point

t Temperature of the point

t Initial temperature
o

s T.i me

Q Heat input

v Velocity

g Plate thickness

K Thermal conductivity

A Thermal diffusivity

f 1
Density

c Heat capacity

£x,.£y, 62 Stress in direction indicated by coordinate

oC Coefficient of thermal expansion

A Change in a quantity

8
 .

I. INTRODU CTION

A« Bac kground of Probl em

In the shipbuilding industry, the fabrication process widely used is

welding. This process is used for economic as well as structural reasons.

However, in the application of this welding process undesired effects such

as residual stresses and distortion, which may cause structural weakness,

are induced.

In the past, the primary emphasis in research has been on how to

prevent excessive residual stresses and distortion. Until recently the

means by which this was accomplished has been more an art than a science.

Observation of what occurred during particular welding processes lead to

the development of design standards to allow tolerance for shrinkage and

welding procedures to help minimize the residual stresses and distortion.

At present there is much effort being devoted, to the development of

analytical techniques for predicting the residual stresses and distortion

that will occur for any given welding situation v . The program that

has been developed at M.I.T. is a one dimensional analysis, Appendix B,

and determines the strains induced by non-uniform heating during welding.

Some experimental data, has been obtained and correlated with this program

a.nd at present further experimental work is being carried, on. Other

investigations exclusively on the heating effects of welding and prediction

-''
of temperatures in metal plates have been performed. '

Even if the best controls of the welding process s.re in effect, dis-

tortion will occur in a percentage of the welds made on a structure. The

amount of distortion that occurs may be beyond acceptable limits and when

this occurs it must be removed. There are many methods available for
 , 5

removing the distortion but the primary method used in shipyards for

ordinary carbon steels is flame heating. Flame heating is used because

it is both an economical and easy method to employ as compared to all

other available methods,

As practiced in the shipyards today, the flame heating technique for

removing distortion is an art. Usually the senior welder determines where

a distorted plate should be heated to remove the distortion. If the first

application of heat fails, a second guess is made and heat again applied

and so on until the distortion is removed or reduced to acceptable

standards. One technique is spot heating an entire panel as shown in

Figure (l). This method insures that the distortion is removed but is

probably an excessive use of flame heating.

The cost of removing distortion can be very high as experienced by

one shipyard where the cost of the distortion removal was half the total

welding cost. In this instance, the welders utilized the method shown in

Figure (l) for much of the distortion removal on 3/8 inch steel plates

used for bulkheads.

Other heating techniques used to remove distortion are

(1) Line heating of the panel, In this procedure, the area to

be straightened, is heated to approximately 1200 "F along narrow lines

which may be water quenched after the heating.

(2) Line heating of the backs of the fillet welds. This pro-

cedure is simila.r to (l) except the flame is applied to the back of

the weld

(3) Line heating approximately three to four inches from the

fillet weld along a single line. This method was developed, in the

German shipyards.

10
 r~r

o o G © & o Q © © o G o ©
o o © o CD © o o o © o o
o £> O © © o o o o Q © G ©
o o o G © © o <G Q G o o ©
Q o o o o © o © ©> © o o
O o o o o © © o O o o o o
"O © © o © © G O <D <© o o o
o o £? o d £? © © O O ID o o
O o o o o o o o O Q o o o
O o © o o o o o D O © o o
o <2 o o o o o o © o © G
O (9 o o G G o © © © o o
G o O o o o o o o o G o
o o o © o o o o o o o
o © o o o o o o o G o o o
o o o o o o o o o O G © o

Figure 1: Example of spot flame heating as actually used

in shipyard

11
a. Line Heating Panel

b. Line Hoo.ti.ng Back of Welds

c. Line Heating Parallel to Weld

Figure (2)

12
These methods are shown in Figure (2).

Richard Holt (5) t

an authority on flame heating in this country, states

that the three basic factors which influence flame bending arej

(1) The thermal expansion of the material with rise in temperature.

(2) The variation of the yield strength of the material with rise

in temperature.

(3) The behavior of the modules of elasticity at elevated

temperatures.

Those are the same factors which influence the stresses induced by welding.

Thus, if a satisfactory model for line heating with an acetylene torch can

be found to determine the temperature changes in the material, the methods

developed for analyzing the welding situation should work for flame heating.

A discussion of the model used for the heat source in the welding a.nalysis

and selection of a model for the flame heating is contained in Appendix A.

Appendix B is a survey of the methods used in analyzing thermal stresses

during welding.

B • .Pr eyi ou s In ye s t i gat ions

A search for pertinent literature revealed that very li ttle has been

published on either experimental or analytical studies of flame heating of

steel whether for bending or straightening. Two recent investigations have

been carried out at M.I.T. and a third at Battelle Memorial Institute.

(6)
Richard Walsh N ' performed a series of investigations to study defor-

mation changes resulting from flame straightening techniques on welded

plates and structures made of mild steel and KY-80 steel. The models

constructed represented structures found in shipbuilding. Spot heating

and water quenching were employed in removing distortion from the welded

models.

13
 The conclusions reached in this investigation were;

(l) Flame straightening procedures are two to three times more

effective on mild steel than rIY-80 steel.

(.?) Varying the position of flame straightening techniques from

plate mid-span to fillet weld area produces no significant differences

in reducing distortion.

Essentially, Walsh found that flame heating does reduce distortion and

determined some' relative values of distortion reduction for the model types

used. There is no indication that the hea.ting technique used could he

applied to a full scale structure with the same results. Walsh further

indicates that there may he a band of material strength where flame

straightening is most effective.

David Duffy (?)' continued the work started by Walsh. Ke chose as

materials: mild steel, U.S. Steel Corten, a.nd U.S. Steel T-l. He con-

structed a model for each o^ the three tyx>es of metal (Figure (3))«

Distortion was induced in the panels during the welding of the models.

These panels were then flame heated using the line heating technique.

Three separate line heats were performed on each panel with distortion

measurements taken after each heating.

Duffy concluded that:

(1) When using flame straightening techniques on panel structures,

it is necessary to use a water quench to achieve removal of distortion.

(2) There is no definite corridor of steel with yield strengths

where the use of flame straightening is more effective than in others.

Battelle Memorial Institute^ ' conducted experiments to determine the

effects of flame heating and mechanical straightening on base metal proper-

ties. Figure (>' ) from the Pat telle report shows that above approximately

Vl-
\0
f.
Distortion due to angular
w change in butt veld
6
O
H
&
O
&
o
w
0)
ft
w
o
•H
H
O
10

0)
>
Distortion due to
•H angular change in
«
rH
fillet weld
0)
W
I L

0.5 i-o 1.5

Thickness, inch

Figure 4: Illustration of the effects of plate thickness

on the relative seriousness of distortion
problems and material problems

16
3/8 inch plate thickness, buckling type distortion does not occur. This

figure also demonstrates that as plate thickness decreases the distortion

problem increases. The major effects observed in the flame straightening

were s

(1) Distortion in welded and unwelded plates was generally

removed with equal facility.

(2) 3/8 inch thick A51?, Grade A plate was more difficult to

straighten than l/2 and 3/4 inch thick A51? plate.

(3) In general A517t Grade A plate was easiest to flame

straighten and AB5~B plate (mild steel) was the most difficult.

(4-) Attempts to straighten by spot hea.ting were unsuccessful.

The major cause was attributed to the fact that the plates were

restrained on only one edge.

Conclusion number 3 is exactly the opposite of that observed by Walsh.

Duffy did not find flame heating more effective on any material.

The above investigations show that flame heating does remove distortion

and gives some measure of what results might be expected if the procedures

in the investigations are used. There was no optimization of the heating

procedure to most effectively remove distortion and from the results of

these investigations no generalization of the best method of heating can be

made. In fact, Duffy points out that not only is optimization of hea.ti.ng

procedures much needed but an analytical investigation of flame heating is

mandatory for this optimization to limit the cost of experimental data.

C« Pur noso^ of Study

The purpose of this study was:

(l) To experimentally observe the effects of line flame heating

on steel plate. This will be done by continuously measuring temperatures

1?
and strains induced, in the steel plates when flame heating is applied,

A better understanding of the flame heating process and how to use this

process for "bending or straightening are expected results. Also, the

question of how material strength affects the effectiveness of this

process may be resolved.

(2) Compare the experimental data to the M.l.T. developed one

dimensional analysis for welding, modified for flame heating. This

will determine if programs developed for welding can be used to

analyze flame heating,

(3) If the two dimensional analysis of welding presently being

developed at M.l.T. is available, the data will also be compared with

this program. If this program is not at the stage where it can be

used , a search for any other existing programs that may be used for

analysis of the flame heating will be made and. the programs utilized

if available.

D« Sele ction of Parameters

Material selection was made with the work performed, by both Walsh and

Duffy in mind. In order that a continuity be maintained in these studies

and because of the availability of materials, the three metals used, by

Duffy; low carbon steel AIST 1020 (mild steel), low alloy high strength

steel ASTM A- 2^2 (U.S. Corten), and quenched and tempered steel ASTM k~51k

(U.S. Steel T-l), were also used in this investigation,

The choice of system model was dictated by the decision to compare the

experimental and analytical results. The analysis of welding is developed

for infinite unconstrained flat plates. Accordingly, the unconstrained

flat plate i.s used, for this investigation. It is recognized that this model

is not found in production except in the case where flame heating might be

18
used as a bending technique. However, it is expected that the welding

program will evolve to the analysis of more complicated structures and if

it can bo shown that the simple analysis presently available is applicable

for flame heating, the analysis of flame heating can be carried forward

with the welding analysis.

The choice of plate thickness was made with Figure Ql) and existing

production problems in mind. In production much of the distortion that

must be removed occurs in internal bulkheads. The thickness of plating

used fox- these bulkheads ranges from l/'l inch to l/2 inch and up. The

Battelle report indicated that for material thickness of l/2 inch and

greater the distortion problem was not great and, also, that in the thick-

ness range of, \fh inch or less buckling type distortion occurs. Thus, a

thickness of 3/8 inch was chosen because in actual use this will give major

distortion problems and yet will not enter the buckling domain.

Material compositions and physical properties are given in Appendix C.

The method of heating used in this study was line heating which is

similar to a bead-on-plate weld. The torch tip and speed of heating were

chosen to give a 1_500 F maximum temperature, and a width of heating zone

approximately the sa,me as that seen in previous use of this heating

technique.

The results from the welding analysis and the experimental results, so

far obtained, show that the longitudinal strain (in the direction of the

weld bead ) is much higher tha.n the transverse strain (perpendicular to the

weld bead) and that the shear strain is very small. Because of this, and

also due to the cost of installing gages , transverse and longitudinal

strains are the only ones measured in this investigation.

In order that the accuracy of temperatures predicted by assuming a

19
source of constant strength with thickness can be determined and to observe

the bending strains induced by the flame heating, measurements of tempera-

ture and strain on both the upper and lower surface of the plate are made.

E. Measuri n g Strain and Te mperature

To analyze what is taking place during the flame hea.ting process it

is necessary to measure the temperature and stress that is induced by the

heating. Stress cannot be measured directly, therefore, some type of

strain measurement is necessary. Since the flame heating technique used

is a continuous line hea.ting with the flame moving at a constant speed

across the material, continuous readings of strain and temperature are

desired.

The measurement of temperature at points on the plate is easily

accomplished. A thermocouple can be placed at the point of interest and

the temperature determined directly from this gage. In this investigation,

BLH type GTM-CA ( Chrome l/Alumel) thermocouples were an integral part of the

strain gages used.

The measurement of strain is more difficult, Due to the temperatures

that will be incurred during heating, it is necessary to select a high

tempsra.ture strain gage so that the probability of gage failure will be

minimized. The gages available for temperatures above 500 P a.re BLH type

1212-5B and 1212-5A, The A designates that the strain gage has a thermo-

sensing element included.

An uncompensated potentiometric circuit was used in measuring strain.

Since temperature compensation was not possible, the strains measured may

have varied non-lincarly as the gage temperature moved far from the

temperature at which the gage factor for this particular gage was

determined. This non-linear deviation can affect the accuracy of the

20
strain measurement.

The strain measured consists of several components. These component >

(a)
strains are normally superimposed to give a total measured strain .

e = e
m
+ e + kaT + ARAT
p

This equation shows that the total strain is composed of mechanical

(elastic) strain, plastic strain, thermal expansion, and component due to

the change in resistance of the strain gage with change in temperature.

The last two terms can be combined and referred to as apparent strain. A

plot of apparent strain versus temperature for the type 1212-5A strain

gage mounted on mild steel is shown in Figure (5). This figure clearly

shows the very large apparent strain that occurs with this type of gage.

Under these conditions it is doubtful that the superposition of the

different components of strain is exactly correct. However, lacking any

better method of handling this problem the superposition will be used but

some error will probably be induced.

This large value of apparent strain also introduces another problem

in measuring the strain. The visjcorder used will allow approximately

five inches for the measurement of strain from zero to the maximum value.

At the maximum safe temperature of 700 'F for which the 1212-5A gage can be

used, an apparent strain of approximately 27 000 micro-inch per inch will

j

be encountered. In calibrating the recording equipment, 15,000 micro-inch

per inch was set equal to five inches of displacement. The scale could be

halved to 30,000 micro-inch per five inches if larger values were

encountered., With this large a scale, any error 'when reading displacement

of the recording from the zero value can induce considerable error in the

value of the strain determined.

21
 1

0~
SO / \
at: /
/
i
i

: SEE DIES I ;B:!MIfi ._ _

V5 /
r
r
,:
'
!
' M "' 'i'l 11 ,T /
riji'lJt ijt\i\l uiuu t
(*r

/
_ K'l|- X2?pB Strain fage /
w .Lili
^CQ
/ j-i
I- L IillQ
t.-iSU S
i'-'J^in
y€]e;l
/ t<" t
^Aj.Di \
ly^t
ic\ c\ \
r /
ITlO^fl (Op. ^ /
ho-itk-ib-M oVen . ., . /

4 Ma:reh 19? y
cw
7
y
/
/
Is
^ /f
/
'

7 \r
/

>
J __

/
7
i_ /
4?
/
,'

/
c 7
,- ._ ._ (5

"3 '

y :: slope = 3E.>6 -c 3D jjg^eV^'

^° _L /L L
i-^N-
TJT
-1- '|
/ rr
_

3 -
„_
/ -

gt
1

^r

_^/
.

S£
<C
9*T /Z l .
. . .

to /
/f
7_
/
'

/
j

7c, /
5-r 1 /r
/

/
A7
/
f
/
/
/
/ I

/
/
A
'

7
j
T

II'
7\
j
30 -Y '

- '"'/ n
(0
- '

It
( c\
. au
6fr\ o/n
,}CtJ
1
LL/ri
-*+;. u
4^60
u,;j>
-'

, Q
. _ .. V;::,:',' n;iT?K f ^ L

_ II fi
1 i

.-,-,...-> t,

1
lj
^ II
 In the reduction of the experimental data., the recorded reading was

corrected by subtracting the apparent strain to give experimental strain

which consists of both plastic and elastic strains. For the above reasons,

the values determined cannot be utilized as absolute measures of the

strain encountered, but their relative magnitudes and trends can be used

in determining what occurs during the heating process.

23
 II, PROCEDURES

A • Scope of the Resea rch

A series of investigations were made of the strain and temperature

changes that occur in mild, steel, Corten, and T-l steel plates as a result

of line flame heating on the surface of the plates. Flat unconstrained

plates were used as models so that the results could be compared with

predicted, values, The heating value used in the flame heating was selected,

to agree with the data available for use in predicting the thermal strains

as stated in Appendix A. The speed, of heating was selected to give a

maximum temperature of 1500 F in the steel plates. Three heats were

performed on each pla.te. The first heat Has used for comparison with the

predicted values of strain and temperature. The other two hearts were used

to observe the effect of water quenching and. as a check on the first run.

The second heat can also be used to see if superposition of strains from

heats occurring at different locations is possible.

B« pescriptjon of Specimens

For this study 3/8 inch thick plates, one each of AIS1 1020 (mild

steel), U.S. Steel Corten, a.nd U.S. Steel T-l vrere used. The pla.te size

of 36 inch by 2? inch was chosen for all specimens. The yS inch length

was chosen to give a distance sufficient for the heating to reach a quasi-

stationary state before readings of temperature and strain were taken.

The 2? inch width was chosen to allow three passes of flame heating to be

made and minimize the effects from one pass to the next.

The plates were instrumented by BLH Electronics, Incorporated;

k2 Fourth Avenue; Waltham, Massachusetts as indicated in Figure (2-1 ).

•The pass lines indicated on the figure are representative of the three

2/1-
 V

IZ ->-

1,
l
i

^^
^-^»
1 |

6 * —
e O G
o 1
i

P o
-p
p Po -p
p
Ph o l
,o o
o
i

.
,Q ^a
^ 1
o
o o
0)
CO 1 |
bD

=tfc e
•H
$5 ^fc
•H
G
» K5 fc S
O
H fH O
H p O
M M
O
O ft

^ B I
H H
CO
o
fL,'
-H
e 8~s
to CO
o
»H
m

< ~> U r» >«

f9—
J-
>-
if 7
1

«-4

«, en
(0 wl w
tfl w M 1

H cd
Pi P-Il s^ C^v
=*fc =te =»fc

s s^ ^
o
M o
M O
M
E-i
M H
M fr>.
tH
CO CO CO
O o
Ft, s

X 1

^
>V"s
1

^> r

Figure 2-1 j Diagram of instrumentation

O
line heats performed on the mild steel plate, Gages were placed on the

tops and bottoms of the plates, the bottom gages being mirror images of

the top gages. The longitudinal gages (X direction) are BLH type 1212-5A

strain gages with BLH type GTM-CA ( Chrome l/Alumel) thermocouples included.

The transverse gages (Y direction) are BLH type 1212-5B strain gages.

These gages were attached to the plates by the BLH Rokide Process. This

process uses a ceramic cement which is applied by flame spraying. The

maximum operating temperature for this type cement is 1500 F. Since this

process has a poor humidity resistance, a moisture proofing was applied

after the strain gages were installed,

C« Heat ing P roced ures

A torch was mounted on the automatic welding machine in M.I.T.

Materials Joining Laboratory as shown in Figure (2-2). The automatic

welding machine was used to give a constant speed during heating. The

rig used for mounting the torch also allowed precise adjustment of the

vertical position of the torch so that the distance between the torch tip

and surface of the plate could be maintained at the proper value for most-

effective heating. This height adjustment was essential for the second

and third passes on a plate since distortion induced during the first pass

was present in the latter passes. The height adjustment was manually

performed and the hei.ght maintained such that the inner core of the flame

was as close to the plate surface as possible without impinging on the

surface. An oxygen-acetylene torch was utilized. The nameplate data on

the torch and tip is: Oxweld Torch, Type W-l?, Tip Size #10.

The steel plate was rested on fire bricks to minimize the heat

conduction away from the plate due to this contact, The fire bricks

supported the entire bottom surface except in the area of the bottom

26
Figure 2-2 j Torch 3d on automatic welding machine

27
 «

strain gages as shown in Figure (2-2) in order to minimize any bending

effects due to the plate not being uniformly supported.

Heating started at the leading edge of the plate and proceeded at a,

constant speed dovm the length of the plate. Pla.stic refractory was

placed around the strain gages on the heated surface to protect the gages

from the flame r

v, ash of the torch. This technique proved very successful

as no gages were lost due to the heating.

Oxygen and acetylene pressures were adjusted at the tanks by two

stage regulators. The amount of acetylene consumed was measured by a flow

meter in the acetylene line just downstrea.m from the acetylene regulator.

Oxygen and a.cetylene pressures were adjusted to give the d.esired acetylene

flow rate and the proper flame length.

D« Experimenta^l^r^c'pcedures

During the flame heating the temperature a.nd strains observed at the

different gage positions were recorded. The record.ed data was presented

as ta.pe outputs from the Honeywell Visicorder. The circuitry used for the

thermocouple circuits and the strain ga,ge cireuits along with the cali-

bration data are shown in Figures (2-3) i (2-4), and (2-5)

The Honeyviell Visicorder is limited to twelve channels so that input

from eight strain gages arid four thermocouples was all that could be

recorded for each pass. Therefore, data from two top gage positions and

two bottom gage positions was recorded for each pass.

Each plate had all gages calibrated before any heating passes were

performed on the plate. The calibration data was included as output on the

tape for each plate tested so that the output data could be read as

accurately as possible.

28
 FLAME HEAT

TEST SPECIMEN

J » STRAIN GAGES
- : THERMOCOUPLE

Figure 2-3: Outline of Experimental Set-Up

29
 -4
K CO
o
vn
:-i
i-J ( i vH
ca TO
c> 'Si
>H C5
H M
a to H
HH-

O > Q 1 i

X O

-P
•H
^ Eh ^
w < O
oo Ph « K-3 ^
K o- O PQ
1
•rl
o
M O3
X
O t-T
< ffl

o o Eh O u8
rH
H !l H
> O
P.
rs
Eh •
Eh
Cm o
CO o o
W
Eh
CO CO
0^
o
O P<
Ph o X H
Eh U
1
0>
CM 1
si
o it
c<^
Pm M S
q XO
CO CO M ..

CO 63
O
M CO Eh
fc,
O -tf
X Q
5= O Ph ^ C\!
^>
O
I
"3 0)
H
o g>
•H
Ph

30
 re;

o
-3-
CO
o a
O
rq
u^ p-«
p
O *H <
HI
CO =fc
Eh >
OM <

Q
Sr.
M
O M
O •»H
2
< Q 1^
1
a
CM
t-3M <
M « >3 oo
K W< =fr
O
X' PQ 0)
bo
o

i5
-p
to

••
m
PC CV
o
Eh
o <c
a>
Jh

X PF< 3
PQ to
o
pq
H •H
t-3 fc
•"3
<
O

o
C\J

—
\AxL--

H H (xl
o
< 52;

Eh
><
S < 6-1
CO £C CO
Z3 M .

Q CO

31
 Temperature and strain measurements commenced for each pass when the

flame reached a longitudinal position G\ inches from the transverse g;»;

r "

positions (Figure (2-6-b)). This aJLlowed the heating to he in a quasi-

stationary state before readings were commenced. Temperatures different

from initial temperature of the plate were not recorded at the gages until

the flame has passed this longitudinal position. Once the readings of

temperature and strain were initiated they were continued until the

temperatures fell to approximately 100 F and the strains had stabilized

to final values.

The location of the flame on the plate surface is specified by its

transverse position from the gage locations on the heated surface at which

experimental readings will be recorded. The transverse measurement is

taken as shown in Figure (2-6-b). Temperature and strain measurements are

also made at the gages on the opposite side of the plate from those speci-

fied in the transverse position.

Oxygen and acetylene pressures and acetylene consumption are recorded

for each pass. The flame length (l), as shown in Figure (2-6-a) s
is

recorded for each pass. Where water cooling is used, the water temperature,

rate of application and the distance the water is applied behind the flame

(d) (Figure (2-6-a)) are recorded. The water is applied to the surface as

a fine spray.

Also recorded are the effective values of heating (q ) a.nd a time

constant (TC) for the observed acetylene consumption. These values are

used in predicting the theoretical values of temperature and strain as

explained in Aonendix A. The values of qi and TC are determined from

e
(13).
Rykalin's data
Cooling water

Torch tip

(a) Diagram of torch tip and water cooling attachment

|
Position W
6*-
Line of travel of
flame
Commence recording-
Temperatures and Transverse distance
Strains when flame of flame from
reaches Position X~[~pos ition Z Positions tf & Z

(b) Diagram of dimensions used to specify location of

the flame for each pass

Figure 2-6

33
 Tables (2-1 ), (2-2), and (2-3) list the above values for each of the

three passes performed on each specimen. Testing was performed on mild

steel first, T-l second, and Gorton last. All three passes were made on

each plate before the next was investigated. The location of the flame

heating for the first pass was uniform for all three plates. The location

and type of pass was varied for the second and third passes, This was

done to observe the differences in readings with location and to see if

superposition of strains from pass to pa.ss is possible,

Figure (2-7) shows the equipment setup for recording experimental

temperatures and strains. Figure (2-8) shows the gage layout and a trace

left by one pass of flame heating.

yi
 1 i

u o O o
Eh r— h

o
o
10 o o o
C\J CM CvJ
D^ rH o- o- o-
cj
O

•d

pw m
•H

u ^—.1
a> • n
o
+>
-p
n5
***
rH
P \XI
P.0^
<H
< -P
«H
O v—

« P. "->
E fe
O Ov—
E-i

W
W
E-< c\>
to Q -P C en cn
rH tiD t!
rH cn cn
1 9
M H rH
CNi a

5 -P W M e\j
•
vn
•
m •
3 £
8 S2 Cn cn en
H &l
O ^->
• w
>> CO H
to X o -3r
o ffi C\2 CV> C\!

• ?H
,C
(0 o-
pV J>
C!
cn cn cn
Q cn
< O 4' rH rH rH
W
*H

CD
^->
O E ££
r3 P-i cn cn cn
to
o H M

M
o
w 8
w G e E
u O rH O C\J O en E
a> •H >< r4 X rH X O
> P W <H W «H
10 c5 4H
E = E : E
O MD OvO O =
H| 9
rH rH fH CO
tH\
f-l J-l

«H^ <H\.
rH r-l <n
oj vn N>A c\! VA

10
CO
rH CN cn

35
 o
O ,
o o O O
H * H rH iH

o
0)
l/J o O O
CM CM CM
CP r\ O- O- CN-
c5
o

Cn
-P -H
•H

<D , Ti
-P
c5
r~!
PH
PnCn
^ 1

1
1

1
t^
CM
•
< -p
tH
1 1 .-1

O v^

0) P4 ^->
E fa
0) O -3-

rH
I

E-<

3 0)
fa CM CM CM
cn cn 0^
E-" rH tj3 «H
CM CO fa C Cn cn Cn
I
CM 5 rH r-l H
CO

-P vn VA *n
3 O 95
s en
*
cn
• •

fi

M
O M
CO
O
oX £
£> CO •3"
Pi P-i CM CM CM
CO

^™s
• rH
w .C c^
-P c "^-v.
O o CA en cn
< o P
«H

o
C) <H IsM en cn en
P<
CO
O rH
fa ^
CM r-l
X 3
fa fa fa
*» *»
(0
H
CD
h
X e
O
cv)
X e
o
en e
X O
> fa H fa H K! U
<H <H «H
6 H e
o = o = o =
H f-l 00 Jm CO r ~fr
H <h\
en
<H\
en
!

<*-|\
rH
cm >n cm >n cm *n

M
re r-l CM en

36
 '
»

c
O C O O O
fH W rH rH

O
(D
W o O o
C0\ CM cm CM
O* rH
o

'd

-p 10
w CM
•h
bO
u
o • T*
CO -p rH ;C
8 Ph
Ph 0^
O-
-P
< -P w
«H
o "*— <D
x:
-p
H .
-—
Pu
co
-p
ro
S
CO O^-'
fri
-3"
ap
S?. rH
W
H
« -p
O
o 0)
rH
o
CM CM CM EJ
3 H 6 •H
H
C~\ I
Eh
CA
rH
C^v
H
0~\
H <0
1

CM tO a
ft

9
en •
-p
o
to
w
CO
dM
to
ir> *A VO
•
<D
>
W
< * P-) cA cA CA

£5
I
-P
10 -^ 0)

M ^ W M *
X CO 3- -3- -p
O ° £
(L)

CM CM CM

o
•
-p
to
c
:
^
r\
r>
•
c^
iN
CA
•
C^
C^
•
r
w!t\j
*A
< cS
<H
-P H rH r-I
10
C^
t
Tj 0)
, ,;— o
0)
*r-
CO O O
C) «H 9 Pm • • • Cu
Ph O rH
v_>
r-I CM CA fT* rH
to fe

0)
rH
a
IxJ
&
W
><
CxJ
C>
to c c
H o
•H
H
X 6
O
CN B
X O
CM
X E
o
ffi
CO bO
> -P cy m W r-t W Jh >,
to (. ,- *H 4h ^ *H X
O e E E o
§ o o = o = O =
rH
ft o
rH rH
z z Z -p
CM »T» CM ^ CM ^A
o
to • c
to o rO
«6 fe". r-I CM C\ «
CH *

37
 rH

•P
C
<D
e
•H

X
<D

bO
S3

to
o
R
o

-p

<p
H

38
"eel plate -fter one pass of flame heating

39
 III. RESULTS •

The results of experiments on test specimen are in the form of

strain plots and a photograph of the distortion induced in one specimen.

(

Appendix D contains tabulated values of the experimental data.

A • Pre sentation of D ata

The plotted results show the strains and temperatures encountered at

various positions versus time. Due to the large period of time involved

in heating and cooling the steel plate back to room temperature, time is

plotted logarithmically. Figures (3~3.a) to (3- 9b) show the experimental

values of temperature and strain measured for the different passes on the

specimen. The data is presented for the heated surface first followed by

the plot for the opposite side of the plate. This was done because of

limits in the computer plotting routine and also for a clearer presen-

tation of the data. All the plots are marked with the symbol T/l?

indicating the time at which the flame reaches the transverse line

connecting the gages.

Figures (3-10) to (3-12 ) show the temperatures and longitudinal

strains predicted by the one dimensional welding program and the experi-

mental temperatures and longitudinal strains observed on the heated

surface of the test specimen during the first pass.

Figure (3-I3) is a photograph of the mild steel specimen after three

passes of flame heating has been performed. This photograph gives an

indication of the amount of distortion incurred. The near edge is the

edge closest to the third pass which was water cooled.

ho
 •I

en
enr
s
O
to

TEMPERATURE, MECH STRAIN

2 >:10~ 3 In/in
( xlO °F ) ( )

FIGURE 3-1-ai Mild Steel Pacs #1 , Heated Surface

41
 I
r<c\] r* <^5

CO
a
CJ
CO

TEMPERATURE, MECH STRAIN

2
( xlO °F ) ( xl0" 3 in/in )

FIGURE 3-1-bi Mild Steel Pass #1, Bottom Surface

42
 Cv! *-* CM «H

CD
CJ
U_J
to
lu

TEMPERATURE, MECH STRAIN

2
( xlO °F ) ( xl(f 3 in/in )

FIGUK3 3-2-ai Mild Stool Tass #2, Heated Surface

43
 CO

o
(_>

CO

I I

TEMPERA T URE, MECH S T RAIN

2 -3
( xlO °F ) ( xlO irj/in )

FIGURE 3~2-b: Mild Steel Pass #2, Bottom Surface

¥\
 i

CO

o
t_>
UJ
CO
UJ
2:

TEMPERATURE, MECH STRAIN

2
( xlO °F ) ( xi(f 3 in/in )

FIGURE 3-3-aj Mild Steel Pass #3, Heated Surface

'+5
 f

CO

LiJ
CO

I I I I I I

TEMPERATURE, MECH STRAIN

2
( xlO °F ) (xl(f 3 in/in )

FIGUHK 3-3-bs Mild Steel Pass H 9

Bottom Surface

^6
 '

-< CVOJ

CO
— CD
.2"
CD
C_>
*-< U_!
>-> CO
-bS' fc. *>»
*—
H JL L-LJ
3»:

V,

TEMPERATURE, VEZH STRAIN

2 j:10" 3
( xlO °P ) ( in/in )

FIGURE 3-4-ai T-l Pass #1, Heated Surface

^7
 CO
CD
CD
C_3
LxJ
CO

TEMPERATURE, MECH STRAIN

2
( xlO °F ) ( xio" 3 in/in )

FIGURE 3-4-bj T-l Pass #1, Bottom Surface

'!-0
 t

CO
O
^r
CD
C_>
LiJ
CO

8
I I I

tj
I

TEMPERATURE, MECh STRAIN

2
( xlO °F ) ( xl0" 3 in/in )

FIGUHC 3-5-caj T-l Pass #2, Heated Surface

Z+o
 5

CO

CD
C_>
UJ
CO

sr:

I I I I I I

Ci

TEMPERATURE, MECH STRAIN

2
( >:10 °F ) ( xio" 3 in/In )

FIGUF:E 3-5'bj T-i Pass #2, Bottom Surface

50
 I

r»

CO
zz:

c_>
Lu
CO
\ *»

U-i

TEMPERATURE, MECH STRAIN

,2 o -3
( xlCT °F ) ( xlO * in/in )

FIGURE 3-6-a: T-l Pass //3, Heated Surface.

51
 .*.

CO

O
U.J
CO

J. I

CJ

TEMPERATURE, ^ECH STRAIN

2
( xlO °F ) ( xlO" 3 in/in )

FIGURE 3-6~b: T-1 Pass #3, Bottom Surface

52
 CD

CO

TEMPERATURE, MECH STRAIN

2
( xlO °F ) ( xlCf 3 in/in )

FIGURE 3-7-aj Corten Pass //l , Heated Surface

53
 X

CO

CJ
UJ
CO

I I I I I I I

TEMPERATURE, MECH STRAIN

2
( xlO °F ) ( xio" 3 in/in )

FIGURE 3-?-bi Corten Pass #1, Bottom Surface

5^
 CO
CD
CD
C_3
l_u
CO
UJ

X
W

I I I I I !

© ti O

TEMPERATURE, HECH STRAIN

2
( xlO °F ) ( xl0" 3 in/in )

FIGURE 3-8-at Corten Pass #2, Heated Surface

55
 I

CO
CD
o
CJ
LxJ
CO

O C

TEMPERATURE, MECH STRAIN

2
( xlO °F ) ( xl0" 3 in/in )

FIGURE 3-8~b> Corten Pass #2, Bottom Surface

56
 8

CD
CD
<_>

CO

TEMPERATURE, MECH STRAIN

2
( xlO °F ) ( xl(f 3 in/In )

FIGURE 3-9-at Corten pass #3, Heated Surface

57
 5

CO

CD
C_3
LU
CO

TEMPERATURE. MECH STRAIN

( xlO* °F ) ( xl(f 3 In/in )

FIGURE 3-9-bj Corton Pass //J t Bottom Surface

58
 I

r>

CO

O
LoJ
CO

H *b

<

An

TEMPERATURE. MECH STRAIN

2
(xlO °F ) (xl(f 3 in/in )

FIGURE 3-1 Oi Analytical Results and Heated

Surface Mild Steel Pass #1

59
 Jo

o
LxJ
cr>

U-J
5^

TEMPERATURE, MECH STRAIN

2
( xlO °F ) ( xlo" 3 in/in )

FIGURE 3-1 Is Analytical Results and Heated

Surface T-l Pass #1

60
 CO
CD
O
STL

C_>
Ul
CO
fc.

I I I I I I I

TEMPERATURE, MECH STRAIN

2
( xlO °F ) ( xl<f 3 in/in )

FIGURE 3-12 Analytical Results and Heated

i

Surface Corten Pass #1

61
 rt

09
0)

fi

c
t

P.

0^

C^

8,
•H

62
 IV. DISCUSSION OF RESULTS

In the experimental procedures, the speed of the flame was to be

maintained constant for all passes. The automatic welding machine was

used for this purpose. However, the speed adjustment on the welding

machine is very coarse especially at the low speed used for these experi-

ments. This made it necessary to preset the desired speed prior to each

pass by timing an interval of travel and making fine adjustments to get

the desired value of speed. The speed was preset prior to the first pass

on the Corten specimen but upon engagement of the drive mechanism for this

pass the speed adjustment changed and the speed, calculated from the

timing marks or the output data, was 2.8 in./min. versus the desired value

of 3.0 in./min. Comparison of do.ta for the first pass on each steel is

made difficult since the speed was different for the first pass on Corten,

The results from the first pass on the Corten specimen, and also the

other pass where water cooling was not used, show that for this steel the

highest temperature was recorded on the side opposite the heating. This

is contrary to what occurs for the other two steels where water cooling is

not used a.nd intuitively this appears wrong. There is no great difference

between the thermal conductivities of the different steels so it is

expected that maximum temperature for the Corten should occur on the

heated surface.

There are several possible reasons why this apparent inversion of the

maximum temperature occurred. These reasons, in the order that they were

investigated, are;

(l) Incorrect recording of the experimental data.

 (2) The surface belovi the gages on the heated surface was not

sandblasted clean prior to installation of the gages.

(3) The observed results actually occurred.

All the strain gage and thermocouple leads were tagged with channel

input numbers and the tags were left on after the experiments were

performed. Examination of these leads and the channel tags shows that the

hookup of the leads to the input junction box was correct.

The surface directly under the flame had a.n oxide layer loosened

during the heating. The surfaces under the strain gages were to have been

sandblasted clean prior to installation of the gages. Thus, the oxide

coating noticed on the flame path should not be present under the gages.

Examination of the area around the strain gages on the heated surface con-

firms that the sandblasting was performed prior to gage installation.

Therefore, the gages were properly mounted.

Since no equipment failure or incorrect experimental procedures could

be detected, the temperatures measured must have actually occurred. There

are several factors that could have influenced the heat flow from the

source out to the measuring positions. This would influence the tempera-

turns at the measuring points.

The experiments on the mild steel were performed, first. There was a.

period of approximately twelve hours from the finish of the experiments on

mild steel until the investigation of the T-l was begun. Thus, the experi-

ments performed on mild steel and T-l started with essentially the same

initial ambient temperature of the fire brick supporting the plates. The

investigation of the Corten was begun immediately after the investigation

of the T-l was completed. Thus, the supporting fire brick may have had. a

higher initial temperature than for the T-l and mild steel. This could

64
result in less heat loss to the fire brick from the lower surface of the

Corten plate..

The experiments performed on the Corten were done in the evening and

ambient room temperature was less than during experiments on the other

steels. The first pass performed on the Corten was ma.de at a, slower speed

and this could ha.ve affected the movement of the a.ir at the plate surface.

The combination of these two factors could give a. higher heat loss at the

heated surfa.ee than experienced with the other two steels.

Melting occurred on the heated surface of the Corten during the first

pass due to the slower speed of the flame. Thus, a metal transformation

occurred which was localized to an area, near the flame. This could have

influenced the heat flow because of heat absorbed during the melting-

transformation .

A combination of some or all of the above factors could have influ-

enced the heat flow from the source and. caused greater heat Dosses at the

heated surface and less heat loss a.t the lower surface than during heating

of the other two steels. This could result in the temperatures measured

at two and five inches from the source being higher on the bottom than at

the heated, surface.

Since all three steels are heated, in the same manner, the tempera-

tures in close to the flame will be essentially the same for all the

steels. During heat-up the metal near the flame will experience higher

temperatures on the heated surface and then equalization of temperature

across the thickness of the plate will occur during cool down.

All the strains induced in the steel plate are the results of the

unequal heating performed on the plate. As the plate is heated by the

oxyacetylene flame, the heated area will expand. This expansion will be
resisted by the cold rigid surrounding metal. When the compressive stress

exerted on the heated area equals the yield stress of the heated area,

which is less at the higher temperature, the heated area will experience

plastic yielding. Since in the experiments the plate appears a dull red

on both the heated and lower surface in near the flame, the plastic

yielding will take place throughout the thickness of the material. The

only direction that the plastic yielding can take place is perpendicular

to the plate surface. Thus, the very hot zone becomes plastica-lly

unstable and plastic upset will occur. The direction of this upset will

be influenced by many factors including the plate cond.ition prior to

hea/ting. Duffy (7)

x ' found that the initial conditions of the steel pla/te

and the boundary conditions determined the d.irection in which plastic

upset would occur.

Upon cool down the metal that experienced pla.stic upset will begin to

shrink as will the rest of the plate, until the initial temperature of the

plate is reached. The plastic strains that occurred will cause a, mismatch

of strains upon cool down and. result in residual stresses and strains.

These residual stresses and strains will cause the bending, The differ-

ence between the temperatures on the top and bottom of the plates away

from the source will cause a variation in the amount of plastic strain

experienced at the surfaces. This will also contribute to the bending

effect of the flame heating.

The exact analysis of how the bending is produced in the plates is

very complicated. Since the flame is moving, the temperature at a point

on the plate varies with time. Plasticity is time dependent, therefore,

it is necessary to know the time a, metal is at a temperature where plastic

yielding will occur as well as the variation of the material propert.i<~'"

66
 .

with temperature. The above discussion is very general but does serve to

point out the mechanism of hovf bending is induced.

Examination of the strains \\-pon cool down after the first pass shows

that for mild steel and T-l the plate is bent up around the flame path, and

in the longitudinal direction the center of the plate lifts and the ends

are bent down. Hence, the bending strains induced are opposite for the

transverse and longitudinal directions. The Corten experienced bending

exactly opposite that observed for the T-l and mild steel. Thus, the

plastic upset occurred in the opposite direction for the Corten.

There was no attempt made to insure that the steel plates were

perfectly flat. In fact, during the installation of the strain gages the

plates were baked and some bending was noticeable in all three plates but

especially in the Corten. This prior bending of the Corten probably caused

the direction of the plastic upset to be different and was further ampli-

fied by the temperatures being exactly opposite those experienced in the

other two steels

The difference in residual transverse stra,in between the top and

bottom of the three specimen at two inches from the flame path is:

0.99 x 10*"* in. /in. for mild steel, .36 x 10~^ in. /in, for T-l and

O.3O x 10 in. /in. for Corten. This indicates that flame heating without

water cooling 3s n^cre effective for bending mild steel. This agrees with

Walsh's observation^ ' that flame heating is more effective for

straightening mild steel.

The maximum transverse strains recorded during the first pass are:

10*"*
-4.33 x in. /in. for mild steel, --^.?6 x 10~^ in. /in. for T-l and

~5«^3 * 10 in. /in. for Corten. These high values of strain, when com-

pared to the difference in residual strains resulting upon cool down, show

6?
that flame heating without water cooling is 'inefficient as a. bending

method. Also, the direction of the resulting bending is dependent on the

initial plate conditions and is not controlled by the flame.

When straightening welded panels with edge constraints, the maximum

deflection that must be removed is usually near the m ids pan of the panel.

Also, the fillet welds used will cause deflect? on as shown in Figure (
l \— l).

The line flame heating could be used most effectively to remove this type

distortion by heating parallel to the weld but on the opposite side of the

plate from the weld. The flame path should not be directly opposite the

weld but moved in a slight distance toward the center of the panel. This

location of the flame line will give a maximum moment arm for the induced

bonding strains to cause deflection. The flame is not placed directly

opposite the weld so the edge constraint will not reduce the bending

effect.

The difference between longitudinal residua.1 stresses in the two

surfa'ces during flame heating is large, 1.05 x 10 in, /in. for mild

steel. If the direction of flame travel is the sa.me as that used for

welding, this should relieve some of the welding residual stresses too.

The distance from the weld could be adjusted to maximize this stress

removal.

If plastic behavior of the metal is assumed to be localized near this

hot zone trailing the fla,rne during heating, it should be possible to

predict the strains that will result if another pass is performed with the

measuring positions of the strains at the same distance from the flame.

This would be done by superimposing the strains that would, result at the

location on those already present. This was attempted during pass number

2 on T-l and mild steel. The predicted and experimental values of strain

68
Figure k-1 x Illustration., of how flame heating should

bo applied to remove distortion from welding

69
 •3 ON in » vn r-l
CM
>t
H•
-3
•
C\i
>~<
CM C\
m
W O o W o o
1

- ca
CM
X
£>-
rH
•
CO
c\
m
CM
X
CO &
W CM H W o
1 1 I

M
Eh
CO
VTN ON U-N CN
Q CM
>i
c/n NO
•
CM
>-«
CM
e
-3
*

W rH o w o o
1 1 1

3
CO
:r>
DO ON CM C^ ON
K CM
X
c~\
*
-3"
•
CM
X
CO
•
& W o
1
o H o
1
o
CO
CO
pi

o
o i-h* •»

r-l
On
H
vr>
H r-l
co
O
un.
CM
I fcl
W
CO
W ><
H O
•
o
•
r-l
>-l
Cxi O
»
O
«

00
a 1 I

EH
I

EH 9
«
O
• r>- - c> NO
H H CO H CM o
X • « X O
* •

£ H CM
1
H1
H 1
CM
1

CO

&
CO
U"N On CO ir*
H C^ <n rH o O
H >H •
o
• >h
o
•
O
•
<
> H 1 1
H 1 1

i
EH
O
M
o
w rH
On
C*n.
ON
o rH
0"A
CM
VT\,
O
X o
•
o
• X O
•
o
•
H w 1 1

nd -d
po
r(j
tJ CO
o -P a>
o r-l O H
•H 3 •H 2
tJ M Tj to
P ei <D rO
H
p.
c
££
h ^

70
 .

are given in Table (4-1 )« There is no correlation of the predicted and

measured values. This indicates that the strain induced during flame

heating is highly dependent on the plate condition at the time of heating

and that the plastic behavior is not localized to the hot zone near the

flame

The third pass on mild steel and T-l and the second pass on Corten

were made with water cooling employed. The strains of primary interest

are the transverse strains, since these cause "bending along the entire

length of the plate and the line flame heating would utilize those strains

in straightening or bending. The difference between the transverse

straAns on the heated surfa.ce a.nd bottom surface two inches from the flame

path is: 0.8? x 10/ in. /in. for mild steel and 1.02 x 10~^ in. /in. for

Corten, This shows a lo.rge increase i.n the bending strain induced in the

Corten. The bending strain induced in the mild steel is less than where

water cooling was not used. These values proba.bly differ considerably

from what one pass of flame heating would produce due to the large initia.l

strains in the plates from previous passes. The bending of all plates is

the same as described for pass number 1 on mild steel. The T-l pass was

performed on the side opposite the initial pass and the bending strain is

larger than for cases where no water cooling was employed.

Water cooling limits the thermal effects experienced by the plate.

The temperatures measured are less than half the temperatures recorded

with no water cooling. Thus, the very hot zone will be more constrained

than when no water cooling is used. The wafer cooling also prevents

temperature equa.liza.tion across the thickness of the plate. Si. nee all

the pla/tes bent in the same direction when water cooling was used, even

though the plates had. entirely different Initial conditions, the water

71
cooling forces the direction of plastic upset and hence, the bending.

Temperature equalization across the thickness of the plate was not per-

mitted since water cooling was applied directly after the flame. Thus,

plastic yielding could take place only in one direction.

The values of maximum strains recorded are much smaller when water

cooling is employed., Comparing the maximum strains with the residual

bending strains shows that line flame heating with water cooling i.s more

efficient than without, The maximum temperatures recorded with water

cooling are much lower. This shows that the heat affected zone is much

narrower and, for metallurgical reasons, the water cooling should be

employed during flame heating. The ability to achieve bending in a

specified direction is a desirable feature of the water cooling.

The large values of transverse strain observed, during flame heating

are contrary to the assumption of negligible transverse strains in the

one dimensional welding program. Therefore, this program is not suitable

for predicting the effects of flame heating. However, the one dimensional

analysis was performed to check the validity of the model of the flame

heat source and to see if the program can be used, to estimate the longi-

tudinal straj ns«

At two inches from the flame path, the predicted temperatures agree

until the maximum temperature occurs. During cool down, the rate of

temperature change is much less for the predicted temperatures, This is

due to two factors:

(l) The analytical, model does not include heat losses to the

atmosphere and the supporting surface.

72
 (?) The temperatures are predicted for a plate symmetric about

the line flame heating , but in the experimental model this is not

the case.

This second reason is the major factor in causing a faster cool down in

the experimental specimen.

At a distance of appro Ximately five inches from the flame path, the

correlation is not as good for the temperatures. The slope of the two

curves during temperature build-up is approximately the same and the

maximum temperature is the same for both curves. The tiro curves are

separated by a constant time factor during the build-up of temperatures.

This is due to the fact that at increased distance from the flame, the

heat source will appear to be a point source. This change with increased

transverse distance is not incorporated in the mathematical model. The

cool down rates are d.ifferent for the same rea.sons as above,

The correlation of temperatures is sufficient to give an estimate of

the strains that could be expected.. If the strain for position two is

shifted by the time separating the temperatures at position two, an

estimate of the strains at this position can be obtained. One other

problem is encountered with the model, At points near the heat source and

directly under the source, the predicted temperatures will be very large.

This can be seen by examination of the temperature equation in Appendix A.

In the program, the maximum temperature was limited to 1500 F.

The mild steel results show the best correlation especially when

ARX2 is shifted. The predicted maximum values are close to what is

observed. The difference in residual strain is due to the d.ifferent cool

down rates for the analytical and experimental temperatures. The

correlation for the other two steels is 'not as good, but a. good estimate

73
 . .

of the strains that occur can be obtained from the pre. ram. The residual

strains differ as for the mild steel and for the same reason

This simple one dimensional analysis shows that computer programs

developed for welding can be used for flame heating. Two dimensional

programs for welding analysis use a finite element solution. Thus, the

solution for temperatures should be changed so that better correlation is

obtained throughout the plate

The same temperature solution was used to alter the two dimensional

program for welding which is being developed at M.I.T. The results from

this program were not plotted because the predicted transverse strains

were very small. The longitudinal strains were about the same as given

by the one dimensional analysis. The program has not been completely

perfected and great difficulty was experienced in trying to perform a. run.

It appears that the analysis used is correct , but there is an error in

the program in predicting the transverse strains.

From the analysis of flame heating, one does not desire the residua,!

stress for one surface. The variation of residual stresses across the

thickness of the plate is needed to -estimate the bending that will be

induced. The two dimensional program will not yield this information.

What is required is a full three dimensional analysis of the flame heating

process. Until this analysis becomes a working reaJLity, the analytical

programs developed for welding a.re of very little use for studying flame

heating.

74
 V. CONCLUSIONS

The conclusions of this investigation based on the preceding dis-

cussion are:

A • Experimen tal Observations

(1) Line flame heating without water cooling is most effective for

bending the mild steel. The amount of bending depends on the material

yield strength and variation of the yield strength with temperature. The

direction of bending without the water cooling is dependent on initial

plate conditions and not the flame location. Thus, the direction of

bending cannot be controlled by the heating.

(2) Line flame heating with water cooling is more efficient as a

bending process. The location of the flame controls the direction of the

bending and the heat affected zone is much smaller than without cooling.

(3) The large values of transverse strain necessitate a,t least a

two dimensional analysis for predicting the effects of flame heating.

B . Observations From Analysis of Results

(1) Line flame heating when used for straightening should be applied

parallel to the weld line, but on the opposite side and displaced slightly

toward the center of the pane] , Water cooling should be used to cause

bending in the direction which will remove the distortion i.nduced by the

welding. The heat is not applied directly opposite the weld to prevent

the edge constraint from restricting the bending effect of the flame heat

and to relieve longitudinal residual stresses induced by the welding.

(2) The one dimensional welding program gives a fair estimate of the

longitudinal strains that occur during flame heating for the heated

surface. However, transverse strains are assumed to be zero in this

75
program while experimental data shows that those strains are large. Thus,

the program is of no use for predicting the results of flame heating

although the comparison does indicate that analytical programs developed

for welding may be used for 13 ne flame heating if modified for the proper

heat source.

(3) Analytical methods that presently exist for welding cannot he

used to optimize the flame heating techniques. The desired result of

flame heating is handing in a specified direction, Present programs do

not evaluate the bending that will be induced by flame heating. Thus,

they give no useful data about the primary function of the flame heating.

76
 VI . RBCOMMBHD WIONS

It is recommended that effort:- in the study of flame heating be

directed toward solving the heat flow problem. Since a three dimensional

analysis is necessary for predicting bending strains, a three dimensional

analysis of the heat transfer will be necessary, An analysis of this type

will require the use of the computer and finite element techniques.

There exists the need in the welding field for the development of a

three dimensional stress analysis. Work is going forward in this area and

will not stop with a two dimensional analysis. Thus, a three dimensional

analysis will probably be available jn the future. The thermal analysis

developed for the flame heating can then be utilized with the three

dimensional stress analysis for predicting the effects of flame heating.

It is further recommended that experimental studies of the flame

heating methods be discontinued until the three dimensional analysis

becomes available, There is enough d.a.ta available from the studies so far

performed at M.I. T. to improve the flame heating practiced in industrial

applications. At most, the flame straightening is a problem of secondary

importance to industry and there exists no pressing need to carry forward

any further investigations of the straightening that particular heating

techniques will yield. However, once a, three dimensional analysis becomes

available, more experimental studies will be needed to determine the

accuracy of the analytical solution.

77
 A PPENDIX A

THERMAL ANALYSIS

The mathematical technique used in solving the heat transfer part of

the welding analysis was developed by Rosenthal^ ' in 1°^-1. The differ-

ential equation of heat transfer isj

2
h t }\ ft » Z l *t
If the welding is carried out over a sufficient length, a. state is

crea.ted in the welded piece which is called quasi-stationary. In the

quasi-stationary state, an observer stationed at the moving heat source

will notice no temperature change around the heat source, The differ-

ential equation of heat can be shifted to a new coordinate system whose

origin is at the heat source. If the X axis lies in the direction of

welding, replacing X by a new coordinate w where w - x-vs

leads to the equation:

p.2 + ilt + j* = -2Kv aw

|i
c^w 2y z Zz z
This is the differential equation of the quasi-stationary state of welding

and cutting.

In the yielding analysis, a line source of constant strength through

the thickness of the material and constant materia.! thermal properties are

assumed. The problem then becomes two dimensional. The solution of the

heat transfer equation is then:

L L
° 2 If K g

where r - \JvP-ty 7 and

-
Y.o is the so called Bessel function of the second kind

and zero order.

The thermal properties of metals are not constant with temperature

.variation and this variation with temperature may be considerable. In the

M.I.T. analysis of the welding problem this variation of thermal properties

is included "by an iterative method. The temperature is first solved for --

with assumed thermal properties and then thermal properties corresponding

to this temperature are used for the next solution. This process is

continued until the error in temperature calculated is a specified minimum.

The flame heating problem is different from the welding problem in

that the heat source is not a point source on the surface, but a dis-

tributed source of varying strength (Figure A-l~a).

The differential equation of heat transfer is the same, but the

boundary conditions imposed by the source are different than those of a

line source. Research of available literature shows that there is not a

closed form solution to this problem as is used in the welding analysis.

Nakada and Hashimoto^ have solved the problem in the form of Fourier

integrals which can be solved using a computer. They also developed a

method of direct numerical integration by digital computer. Both of these

methods require extensive computer space and a great amount of computer

time in order to solve the equations.

Since the strain analysis developed at M.I.T. (Appendix B) requires a

minimum of computer sp3.ee and time, the thermal solutions discussed above

seem incompatible with this program. The large space requirements a.nd cost

of these solutions for temperatures of an accuracy not matched in the

strain analysis make the above solutions unacceptable as models for the

flame heating,
(13)
Rykalin x has developed a simpler analysis of the flame heating

problems which should give the desired accuracy in solving the heat

transfer problem. The model he uses for the flame heating process is shown

in Figure (A-l-b). In this model he assumes an imaginary heat source at a

79
(a) Heat flux distribution of the standard
torch flame along radius r

(b) Model of flame heating with imaginary

heat source located a.t point

Figure A-l

80
distance, equal to a time constant times the speed of the flame, in front

of the actual source. The solution to the hea.t transfer problem then

becomes: r-^vw , &-q( TC)l

c
2 71 K j

q is an effective heat input for the imaginary source. A table of experi

mental values of q and TC (time constant) is found in Reference 13 f

page 19«

In modifying the M.I.T. welding program for flame heating Rykalin's

model of the flame heating is used. Since the two solutions are similar,

the modification of the program is very slight.

81
 AP PENDIX B

ANALYSIS OF THERM AL STRE SSES AND META L MOVEMENT

Because a weldment is locally heated by the welding arc, the

temperature distribution in the weldment is not uniform and changes as

welding progresses. This non-uniform temperature distribution causes

thermal stresses in the weldment, which also change as welding progresses.

(?)
A computer program has been developed at M.I.T. ' to analyze thermal

stresses during welding and the resulting residual stresses.

Masubuchi describes the welding process as follows . Figure (B-3J

shows schematically how welding thermal stresses are formed. Figure

(B-l-a) shows a. bead-on-plate weld in which a weld bead is being laid at a

speed v. O-xy is the coordinate system; the origin, 0, is on the surface

underneath the welding arc and the x~direction lies in the direction of

welding.

Figure (B-l-b) shows temperature distribution along several cross

sections. Along Section A-A, which is ahead of the welding arc, the

temperature change due to welding, T, is almost zero (Figure B-l-b-l).

Along Section B-B, which crosses the welding arc, the temperature distri-

bution is very steep (Figure B-l-b-2). Along Section C-G, which is some

distance behind the welding arc, the distribution of temperature change is

as shown in Figure (B-I-b-3)« Along Section D-D, which is very far from

the welding arc, the temperature change due to welding again diminishes

(Figure B-l-b-4).

Figure (B-l-c) shows the distribution of stresses along these sections

in the x-direction, 6" . Stress in the y-direction, 6"

t
and shearing
x y
stress, 7" » also exist in a two dimensional stress field (Figure B-l-a).
xy
 AT»0 Strcss^O

I. Section A -A

E
c
I-

2. Section B-B E
o
o

-jb°>
w?\sp~~
3. Section C-C
Residual
Stress

a. Weld

~^W' xpV^
4. Section D-D

b. Temperature Change c. Stress a

Figure B-l ; Schematic Representation of Changes of Temperature and Stresses

During Welding

83
 Along Section A-A, thermal stresses due to welding are almost zero

(Figure B-l-c-l). The stress distribution along Section B~B is shown in

Figure (B~l-~c~2). Stresses in areas underneath the welding arc are close

to zero, because molten metal does not support loads. Stresses in areas

somewhat away fi-orn the arc are compressive, because the expansion of these

areas is restrained by surrounding areas that are heated to lower tempera-

tures. Since the temperatures of these areas are quite high and the yield

strength of the material is low, stresses in these area,s are as high as

the yield strength of the material at corresponding temperatures. The

amount of compressive stress increases with increasing distance from the

weld or with decreasing temperature. However, stresses in areas away from

the weld are tensile and balance with compressive stresses in areas near

the weld. In other words,

I'
across Section B-B. This equation neglects the effects of 6"y and ?*y

on the equilibrium condition. Thus, the stress distribution along Section

B-B is as shown in Figure (B-l~c-2).

Stresses are distributed along Section C-C as shown in Figure

(B-l-c-3). Since the weld-metal and base-metal regions near the weld have

cooled, they try to shrink causing tensile stresses in areas close to the

weld. As the distance from the weld increases, the stresses first change

to compressive and then become tensile.

Figure (B-l-c-^l) shows the stress distribution along Section D-l).

High tensile stresses are produced in areas near the weld, while com-

pressive stresses are produced in areas away from the weld. The

distribution of residual stresses that remain after welding is completed.

as shown in the figure.

?h
 The cross-hatched area, MM', in Figure (B-l-a) shown the region where

plastic deformation occurs during the welding thermal cycle. The ellipse

near the origin, 0, indicates the region where the metal is melted. The

region outside the cross-hatched area remains elastic during the entire

welding thermal cycle.

The M.I.T. analysis of thermal stresses includes plasticity, variation

of coefficient of linear expansion and loading history. A plane stress

formulation is used, with stress and strains being considered uniform in

the thickness direction, and ^"h, the stress in this direction is assumed to

be zero. The analysis is then reduced to. one dimension by the assumption

that only the longitudinal stress is non-zero. The detailed equations and

computer program developed are found in Reference (2).

The assumption of constant stresses across the thickness preclude any

bending effects. The analysis does not include changes of stress-strain

characteristics and expansion characteristics due to metallurgical changes

which may take pla.ee during welding. The analysis first calculates the

temperatures that will be present in the plate a.t various times and then

uses these temperatures and times as inputs to the stress program. Thus,

the analysis is split into two separate pa.rts.

The line flame heating is similar to the bead-on~plate weld for which

the analysis was developed. The major difference is the heat source for

which a mathematical model has already been developed. Due to the finite

width of the source, the width of the hot zone around the flame will be

larger than in welding, This width and the fact that more heat must be

used to raise the zone to the desired temperatures will undoubtedly

introduce transverse strains, However, if these transverse strains are

still reasonably small the welding analysis should give estimates of how

8«3
the longitudinal strains act and reasonable values of residual thermal

stresses,

At present, a two dimensional plane stress program is being developed

by Andrews and Masubuchi. This analysis includes the complete two

dimensional stress equations and equilibrium equations. The material

assumptions and analysis are similar to the one dimensional program except

a finite element solution is used and the thermal and stress analysis are

combined. The program is currently at the stage where past welding data

is being compared with the program outputs to check the program. It is

still in the stage of being continually refined and debugged.

Only the results of the computer programs will be presented, in this

report. No attempt at developing a suitable program is bei.ng made. The

flame heating data will be compared with the analytical results to see if

the welding programs, modified, for the heat source, can be used for

analysis of flame hea.ting,

86
 » ^ ^ s *' »

O
o
m *
CO
r-l
^t
CO o
in O rH CM «
r-l O o o -^ • * Ov

O CM o 00
-3". CM Cn
O
CM
•
-3-
o • c>
*
r-l
CM 3
CM
in
•
en CM
O O• OmCn CM H CO o -3 • • CO
•

O
O CO
•
o •
CM
•
vO
cn 5 o
o CM -3-
o CO i-I CM *
H CM rH Cn • CO

O o CM cn o
O
o vO
* • ^ cS vO
•
co C^- rH CM
CM
O
O
ONH
CO CM r-l o cn • • CN-
Ph
ooo
o CM
•
m •
vO
•
O
vn c?
o
CM
o O- rH
o cn
en
rH CM •
\o CM CM • •

H Cn m On -3- O
to
w
o
o O
•
cn
•
CM
• uS
H 3
CM
r-l
•
p<
o H VO O m CM CM o CN • O-
p:
p.
sq cn cn On

o
_j- O CN
•
c^
VTN cS O
<
o o C^ vO
o
rH
en
H CM •
-3- CM CM • • F-
M
0':
>
u VO O CM
X Ph o
o O-
•
co
*
O
* vO
H 3
CM
On
•
PI Q -

vO OnCO
cn CM CM o en • * vO

o Ore*
O i-I

ooo o m O • •
CO
•
u^
vO ^t
CO CO
< o oo
CM
CM
0-N o CO
CM
rl CM
• vO
•
CM •

cn o CM IT\
• • « vO cS CO
O ON U"N
cn o vO H *
CM
• VO
•
C^ CM CM
s~^
H x—
— u c
O
cu
co '—
•H
10
'

•H
Nl 0)
-p * — •H
-p
CO
c Pw CO
P4
CD
e
^
')
CO
C
-p^
fl)
—
/- -. vO 3 <— 1 a
£ TTi rl r-l CM O cn
o H rO rH
H
o
o
p-i
X
05
e
r-l
OO t CM r-l
X
(d r! td
M
e^^ •> p. C 1
•rl E-< r-l Cv 1 V)

—» •H
o N^ H —
v./
N
X o fo 1
'

CM
o 0)
Pa
CM
04- H
^ CM
O
•k
/-

r-.
ov..-- 3
CO
x:
to
t:
•H
^-^
2
> — ..-'
J^ rH
«5
E
N^ w CM CM -3: CM rH C -p ^™% S
a O 1 v_y 3 ft CD -~^ ctf >> r^ G>
o H<< 1

10 CD Tj c •d «H CO -p rOl C jz
H
•H
M£S o H O o r-l CO w •H
>
rH|.H
V—
-P •H 3 5-1 ti P:
•H
W
CO H E-<
HI CO CO
-P
o -P
CO
-P
CO
tl-'

vO •r-l
o •H
-P >>
tH
O
fc CYi

C < < < CD CM ?H

« n o <H O -p fe
Pn P< O CO bO -d •r-l r-l •rl 3 •H •
£P
e
o
o H Pi
e §
r-l
CO
cC
rl V-X
X O -d
C
10
c
<H
CO
o fi H 0) •H -P a O CO O
i—! H
CO
Hi
Eh >* >^ CO CO C) p c^

ri rJ
O
<
•H •<-!

E CO •
0) !>> <
6 p-i

rH CM

87
 o CO en o
o
o VT\ CO
o-
o CO
CM
CM o o CM CO

o
o en o o- rH
ON
0^
CO
CM
CM
CO c MD r-l O CM
rH CO

o CO en O
o CM
o cn
ON
o CO o-
O- CM
3 CO CM C^v o-

o
o
CO o NO en
o
en
CO
On
en
CM -3: en CO en CM
<H -3- en

O
o CO O H
rH
Cn
CO
O
O
o CM On
& rH CM
rH CM o-

O NO en en O
o
o vn
*
O cA
H
l-l
CO
CM
CO CM CO Cn VO

O rH O en
NO CO
rH
en
CO
o
O CM
cn
rH CM
NO
NO CM CO rH

o
0~\ O O- CO en ON
cn
00
o rH
H
en
CM
NO
-3 CM CO

Cn
o o rH en CM
o
o o H
H CO en
co CM
CM CM ON rH en NO

vn
o .ON CM en O
o o o
o
in O
H CO Cn
oo CM
en rH rH CM NO

•H CD O
10 •ri -P •H
£; w 05 Cm 10
p^ E
NO ed s
o en rO rH
1

o p<
O H rH o3 0) X
rH
X
Ph tO
o M f-
1 o to
1
w
:>"

-"*
£ rH
d- to •H trj

H X
-P
1
. •H
-p
E
rH
b0 CD en 0)
05 C Q> r°.l C JCJ
U o 0) to •H Eh
fH CO Pi >
-3" -p •H tH c
rH Cv to CO NO -P >> C •H
J-! O c -P P-
I 0) "ta nd •H -\ •H • c D
< P^ c H CO O to «H •r!
E p 0) f: 0) C CD ^-^
0)
Eh
o •H -P
to
P.
to oo P
05
oO
10
<

88
 o
o CO
O
m
rH O in CM
00

O
o CM O ON ON O
cn
* rH 6?
CM J* O NO rH CM
rH H cn CO

o
o CO in in \o .3 o
o
• CM CO
o CM CO CO in r-l CM
H CM rH en CO

o o in o- CM
en
o
cS o-
O \0 CO CM
CO CM CM

CM CO vo in
o
o CM
en
cn CO CM
ts- CM
CM -3- rH cn cv

cn ON -* CO .3-
o
o cn cn O
in. cn CM rH CM
in cm 3- cn o-

o
o
o CO
-3-
rH 3 5o
CM
3
-3- CM -3 rH cn NO

o
NO
-3"
UN O
co'
o
cn
c- ON O
cn
CM
cm •3" NO

in ON in
o in cn CO in
o
'CM
co
CM
ON
3" rH
On
CM
rH CM
NO

cn cn •=3"
£: C\)
cn CO cn
o ON cn NO H CM
in CM rH CM NO

rH
•H CD O
03 •H +5 •H
Pi m Q> CO
P< E
NO
o cn Pi
rH o X
X rH
y,
b0
fe CO C co'
•H E
H -p r-i
bo o CO
•H
O G) CO
B rH i Pi
CM -P ch c
d CO to NO >i O
u
CD
co
H rH H-
•H
1

•
•r 1

v
i r 3
< rH p3
N-^
X CO CH •H
E q r-l £ 0)
o P CD
oo
CO
to n
<
o

89
 I !I

CM o
NO ON
o o o .^
0\(vH vnt^Wi^OMAW
r-i cn^- no co no cm cn unvo c-no no o-
O^NNW O cn-
WvO
cn-md
C^- Un, r-t
onco co un cn cn
H ON ON
>H
oooooooooo I I I I I I I I
-I r-i —
I
1 l—l
i
I — H
I I
HI
ri
I
HHH
I I I
H
I
rH rH
I I I I

CM o vO
ooh
(^4- NNHW^HNOOCAOvO^.C^NNUNMvOHHC^N
ri h r\w o nr\4 iTiriCO CO C^O O rl H >A4 'HN N4
X OOOOOOOOOOOOOOOOO H -! i-l H rH H
I I I I I I I I I I I I I I

On O^s nO H CO 03 UN CN CN- u-n CO On CM nO CN On On O 00

Cn-
CN- CN- r-l
O ONNia^O
-4/ i~l CN-
CM OO O CM -4 un,vo CN- On CM -4 -4/ no CN- CO CN CO CO CN- CO C^-

W OOOOOOOOO I I I I I I I
r-i
I I I
rH
I
H iH
I I I
-IHH
I I
CM
I
HH — -I
I I I
r-i
I
1

CM
NO NO
OOO -!

1 O O
CO
O CM
OO
CN CN- r-i
r-l
NO
-4 VT\vO O- UN
O
CO _4 On CO
CM OOH
UnnO UNCO CO
CN CM CM CM UNnO
N ^ CMT^
CN Cn CNJ
VTi

X OOOOOOOOOOOOOOrHrHiH H -! H
-4f

HH H H
III I I I I I I I I I I
r-l
I I I I

CO r\H4 ON N W H O UNnO ON CO CO O, T\ W CO C"n N rl h h n vn

to
E-<
iH
>-<
OH OOONOOOCOCOONCN^.
C°i
rH On NO -4" CN O 00 CO INCn-VOvO C^N 0J
H oo ooo 1-1 cm cn cn cn cn cn cn cn cm cm cm cm cm h h h h rH hH i-*i

^
co
I I -I I I I I I I I I I I I I I I I I I I I I I

&3
«
H
< CO
X OOOn OHO .4-
CM i-I
OCO
r-l .4-
rH CN- CM nO nO
00 CM ON UN CO On
n^W CO >A On H NO
CN ON ^ ON CM CM H
r-l
("N CN CN
rH rH O -4On NO NO
CN-
CN-
r-i

X
E-<
<<
K
CO
<
Ah
OOOOOOOOI I I I I I I I I
H r-l
I
CM CM CM CM CM CM CM CM CM CM rH rH
I I I I I I
HI

H S
M
I I I I I I I

s cr,
H
W w w ONNNH
Ph
P,
<
Pk
X
w
Eh
CO iH
>h
OOOn CO H UN UN
r-l
O-
N On H CO
CM CN
On CN ON
U"^
r-l r-l r>- CN-
CM On
CN C\nO nO
CN- -4: r-l CO
rH
CN-
UNnO 0N.4
-4 CM rH O O UN ON . CN- NO
OOOOO rH rH CN. CN -4" -4; -4" CN CN ON CN CM CM CM CI CM CM — 1— rH rH 1—
Q I I I I I I I I I I I I I I I I I I I I I I
I

I
I

I I I

W
Eh
<:
•J
CNNO 4n0nD CN
t3
pq
<
rH
X
O ^ n N H W V\ O 4On j4
© CN.J- .4 .4
O44 r!
r-l
44444W
_H- -4" rH r>- CN-
rl
^4 CO
O On ©
-4/ rH rH
NvO
UN On
4- H
H K OOOOOOO I I I I I I
rl
I
rH CM CM CM
I I I I
Cn]
I
CM CI
I I
CN!
I
CM
I
CMCMCMrHHrHrHrH
I I I I I I I I

•
CM O- O-
04 >TiCvi
CN- CN- CO 00 On
0\N4 r-l
NW C^OOn ino
CN 4 O CN UN UNnO 04
C^C"\OON4
CN-
WiUNcO 00 o
C--O-CCC0C0C000CO
CM .4-
E-< rl r-i rH rH rH CM CM CM CM CM CM CM CM CM CM CM CM CM CM CM

CM OO-4-CNOCOONOtTNCN-C^CNO UN NO UN CN CM NO CM UN UN UN u"\ NO OO
E-< C^- C^- C^ CO On On H CM -d NO CO O
CM <T\ _3 UN NO CN- C^ 03 00 On On On On On CO
rH rH H
rH r-l CM CM CM CM CM CM CM CM CM CM CM CM CM CM CM CM

0^3- r! o n On
O
ip\ o. un O
UN nO .4- CN- O- Cn C- O O u^^0 no O UN .4 ^3" UN
CO
rH
E-<
C"n- CN- CO CO)
rH rH CM
UN CN CN CM CN- r-l ^ NO no NO NO UN
CN jj- .4 UN UN UN UN UN U ir, ,
u"n,
VP» U~>
CN CM r-l ON CO
UN UN U^^j- _j
H C^- -4" r-l
.h- jj ^}- •

r-i O r-l C^- ^ f>- NO O

UN CM rN O 4 CO 4;
-4' O UN Cn O On 00 CO
!NC0C0C0lNN0v04CNCNirlOC0CN-4 H
CN- CN- CN- CN- CN-
Eh C^- CO CO ON r-l
r
NO UN UN .4- O -4
I rH CM CN ^}- UN u^ un u^ un un, un un On un un un un
u-n ^J- .4 .4 ^)-

-IOOOOOOOOOOOOOOOOOOOOOOOOOO
4 \D w 4 \o o w 4
C\J O 4 o w4 4o
O'j CO VO CO (V vO CO vO CO
h-i
Eh
rH r-l rH r I r CM CM CM CM CM CN CN C
I N CN CN.4 .4- 3 ^- -4 U\nO
 o 4- »n w o -> vo o oo h
vovtnonc^o^,
r\HONNvOr>WHOHOHHH(^
CM
63 iHrHOOOOOOOOOOOOO
i r i i i i i i i i i i i i

CV
x ^W W O O H N H (M
...............
iT\4 f^irl ri 00
63 — — — —
(— I i — — !i — I i ! i I r-\ i I ( 1 i—l (— 1 i—| i— I r— 1 i ! r
- !

I I I I I I I I I I I I I I I

On O- O O- u"\\0 O J- O- On On NO CV CVnO
C^ 0~\ FA^ .^
CV CM CM r-l 0\ 0> tr^vD 3 -j- ^
63 HrHrHOOOOOOOOOOOO
I I I I I I I I I I I I I I I

\0 C%CM On CO nO On O O
CM
X
WP^.NO OW\D u^4 CM CO CM VN-d"
u"\ u~\ »J~nvD
i- I

vr-\-j-
. .
63 rH
r-( OOO OOO OOOOO
r-l r-l
I I I I i I I f I I I I I I I

^"^ rH
OO
NO On U~n. .3- LTN (N-vO r-l
CM CM 0~N.H On rH
O...............
O O O O rH
NO On C^ U\ C^t C*^ rH i- !

rtf >-l
0)
pj
63 HOOO OOO O©OOOOOO
I I I I I I I I I I I

1-1

OMA O (^ On mn O © 4 W On
UO -
rH
rl vO CO
ir\0\NHHOO\©ffi©©NNNN
...............
v^x X i-IOrHr-lr-lrHOOOOOOOOO
63
r-l I I I I I I I I I I I I I I I

CO
to
< CN-C00\{>-rHO-0\N£)C0
HO
C"n\0 CM vr\, f>- CO
P<
»-3
rH
>H
63
...............
CV O-
rHOHHrHOOOOOOOOOO
CM On On CO CO CO CO On On CO CO

63 I I I I I I I I I I I I I I I

EH
to
P)
1-3 H N CO 4-O- -4
On _4- jj
CO
CM
H ONVO
OOH O-
rH
O
CV!
U"\ CM
CM CM
nO CO NO
H r-l CM
r-i X ooooooooooooooo
S 63
I I I I I I

CN-OCr^r-J CN-vrN,^)-OCOO^OC^COON
O O
HH
CM O- C^-VO (tn,
-H; r-l CO NO v/n, o~N Cv! CM
EH CM CM CM CM CM r-l rH r-l H r-l rH H rH H

CM On O- CO vO CM CO CO CM
m r-! C^nO 0~\vO OO
HH
Eh Cn~ Cn- NO -3- CM CO vO .j- p~\ CM CM rH rH
CM CM CM CM CM r-l rH rH r-l H r-l r-l r! r-l

•r-l

VO NO CO C^ O U"N CM C r N CO u~\ U\ CO 03 CM On
O \.
rH ONNO O VO, rH O^ JN VTN^J- 0"N CM
Eh r\0^n^W WriHHHHrlH Cn! r-l r-l
rH r-l 6<
O
i
fl
!

r-l nOvC© r.O UN,Nr<NK)UNn© HIT)G

I I

o
rl
Eh OnvO lAr!
0"n CM
iT\4
<<~\
NW
CM CM CM
On
rH
N rH
H iH rH
riO!
r-l
CM
rH rH rH
HH CL>

fl
?
X
-P I

61
^
OCOOOOOOOOOOCOO
O
VO CM CO NO VO CO CO CO CO CO 00 CO CO
OJ
Pi
E
-H
r3

r-l
Eh
NO O- O- ON CM UN On
r-l r-l rH CM
CO-
C-> W CO rH
CM CM
£> CM
CM CM
» C<~n j^-
O
^ EH
0)
r-l

-P
CO

91
 r-l CM CM O
rH rH rH rH OwO CN UNnO inO\H CM MD NO VTlNH CM On CO CO vO
CNnO jH- Oa r-l
rH
>l
O CM .3 UN Cn CO Ox O rH rH CM
................................
r>r\r\0^.j CN rH O ON On On On CO VO UN. U-, UN ON CN i-l r-l

W OOOOOOHrlHrlrlrlHNNCMNWNNWHHH H H rH rH O
I I I I I I I I I I I I I I I I I I I I I I
r-l
I I
r-l
I I
r-l
I I I

H ONMVHHN^^rovO^wOMnOOJNCOrlOiTiChOOiriONCMNOCO
UN _4 CNOOOOiHC\lif-4-NOCOONHrHrHON.4^UNCNCNOOONONrHCMOCN
f-N.NO

W O OOOOOO OO OOOOOOO r- 1 rH rH rH rH r i H r-l r-l r-l rH H CM 0J CM rH

I I I I I I I I I I i I I I I I I I I I I I I I I I I I I I I I

O rH N ^ N r'\CMaH
H CN N
CM CM CM NO O CNi}
OO N On O CNnO O VT\ O
un On V\d^ unnO O CO
rH
>H
ON CN C* UN UN NO On O ON CN
--J
r-l r-l CN) CN ON CN 00 HO O On fN CO Un^} CO
.
W O OO r-l rH r-l rH r i r-l r-l rH CM CM CNl CM CM CM CNi CM CM CM CM rH CM CM CM i-H r-l rH r-l rH O
I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I

CO CO -3- rH CO NO -4" u~;UN jH CO rA H- O CM CO On CM NO On u^ UN. CNnO nO CM O CO CM NO CN jH-

rH
X CM CM ON OO r- ,'
CM ON CN ^
MD CN CO CO O O O CM CM CO r-l rH CM O OH CM ON CM On CO ON
M OOOOOOOOOOOOOO I I I I I I I I I I
r-l
I
rH
I
r-l
I
rHHrHrHrHOrH
I I I I I I I
r-l
I
Hr-!0
I I I
r-l
I
OOO
I I I

CM CN, CM CM O 3" rH H O ON VjN CN NO

O O
CM
>H
H
................................ CM CO
O O O O O O H CM C\ CN CN ON CN C\ CN CM
O
-4' Cn NO CN CM CM rH
CM CM CM CM CM CM CM rH rH r! rH rH rH rH rH
On CO NO -3 CM rH CM CO

I I I I I I I I I I I I f I I I I I I I I I I I I I I I I

$ rH On CN- CN On On On cn' NOCOCMUNCNCMCOCOCAOCNC\lrHCNNOrHNOCMNOUNCNCN.4-ON

CO
w
CN.'

x O O
................................
CO O- rH -4 CM CnnO 4 r-l if ON rH VTN. r-l CN UN, UN CN Cn NO O- NO vO UN CN CM CM CM r-l r-l

<
P-i
fcl OO OOOO O
r-l rH rH r I CM CM CM CM CM CM CN! CM CM CM CM CM CM CM CM CM CM CM CM CNJ CM
I t I I I I I I I I I I ! I I I t I I I I I I I t I I I I I I I

.-1

E-i WO
4 O NvO rlCONO NOMAM3N44 ONOvOOO 0\£) HOD ^4 NNri4 N On
CO CM r-l On NO VT\ rH NO CnnO lA ON rH On CO UNi4 O N CM
VTN, r>. CM O O ON Cn Cn 00 UN UN r- 1 CN-

Q frj O O O O O O H CN CN CN C\ CN CN CN CM CM CM CM CM CM CM CM CM CM rH
Cn! rH rH r-l r-l r-l r-l
M
M I I I I I I I I I I I I I I I I I I I I I I I ! I I I I I I I I

CM
CM
C\
^ O CO CN CnnO
CM rH Cn CM CM -3
CM O
NO .-4
On CA cn On
f>- r-l
m CN. .4
CN CN.
UN NO
CM CM CN (N
nO^CNCACMOOnCOCOnO-4
CN O -4" On .4 CN! -4- CN! NO On rH
Cn, H CM On CO
-4 _4

W O OOOOOOO rH H r-l W CM CM Cn! CM CM CM CM CM CM CM r-l rH rH r-l r-l rH r-l r-l O O

II I I I I I I I I I I I I I I I I I I I I I I I I I I I I I

o o o cnno cnno cm o o o o no cm unnc cn -4 co cn cn on h r.n^- 0NtnO4 CM On

r-l OOOrHrHCMCNV-NCNONrH CN-4' NO CN- 00 On O O rH r-l rH CM CN! CM CM r-l rH r-l O OMA
Eh r-l rH rH
r-l H H H CM CM CM
r-l r-l r-l Cn! CM CM CM C\ CN CN. C\ C\ CN CN, ON CN ON CN CN CN CM CM

rH
Eh
OOOCNNOOCN ON
O O O rH rH ON_4- CN CO CM O
UN C- On CM -4
CO rH NO CN
UN CN CO On O rH
iAC>NO O O
VTN r-l O O U\NO CN O CM On
CM CM CN CM CN CN CN CM H H C On VTN
i~I
rH i-l rH
r-l r-l r-! r-l r-l r- ! rH CM CM CM CM CM CM CN Cl CN C^\ CN CN Cn CN CN CN CN CN O^v CN CM CM

CM
O OOffi OCKOvOCNOnO^'JNCO
O O O rH CM -4 O CN UN NNCO uNNOJ^nOCOO-COH COUN^NOD
On rH CM CN CN CN CN CM CM rH O On CO Cn O CO
VTN
04
NO UN O
H H H H H N WCN444 UNUNU"\UNVAinvAmUNV(N4 4 4 4 4CM 4
r-! U~N
EH H ONPNt^iCN

CM O OO CO Cl CN UN On On CN CO CO O CN NO
CN O CN -4 NO
Cf ) NO CO NNOnHCNNH04 UN.
O O O r-l CNNO CM CO U\ On cnI O NO NO UN PNW H O On CO CN O CO N>AO
4 ^ UN UN NO UN IA ^4
Eh .4- u^N
r-l r!r-l ri HH CM 0^0^4 UN UN 4 4 4 4 4 fN^rNCN
l/N UN, vTi U"i ifN

fx}
S
1-! OOOOOOOOOOOO
4
(V O N4 O N 4 OOOOOOOOOOOOOOO
nO COO N4 O 4 4 O OO OO
N 4 W nO CO NO CO nO CO CM nO OD vO CO
M rH r-l r-l rH rH CM CM CM CM CM CNCNCNCNCnjj-jj-^-^-^^^MD NO CN ^ CO O
EH r-l

92
 Ono-vd u-\co u"»-j- u~\
NO NO NO
• *• •
U"n .3
I V
C^H H•
« *

oooo oooo
.

I I I I I I I

CV CO CMsnONCO o
On O r-irH O CO CO r-l
X rl (M CJ W N N H H
I I I I I I I I

0~\ O On O- On On OO
On O On On On CO U"n C°\
OrHOOOOOO
I I I I I I I I

vO C^ 0^ On
^ ^}0\ NO
un, i~!
rH
X C~\ ^J- 0"N. (M OO
OOOOOOOO
I I I I I I

On On ON CO On NO O- u^
CV On rH OOO H
r\.j-

w OOOOOOOO
I I

no
0)

CM
CO On \D
O CnI UN, en CO
On CO On On CO O- C\
O
•H
-P CV rH
I I
H H i I
r-l
I
HI
r-l
I
r
I
I

O
O
O CM C~n CV vr\ <} O O
On
CV NO -3 CV 0~N CN CV \C)

to i-I H rH rH r-l rH r-l O

to I I I I I I I I

cn;
.3-On .3-
On o- u"n
UN,
vr\
.4
»?n.
O NO
vtn CV
CV

E"' x OOOOOOOO
-i)

to w I I I I I I I

rH
4-woovr^o\DO
_j C> H On NvC ^ r>-
E-" CV CV CV rH rH r-l r-l

rH
E~<
-^-
4
CV
^HO OOn UN
CM
CM CV
O
C- nO
r-l rH r-l
V.O
-^
rH
O
O-

• 0~\\O
rH On
OO vr\ O vO O
NO -^ O-
CM tN- -ch f>-
E-i CV CV CJ rH r-l r-l r-l

CV CanO O O VT\ O NO O
rH O-^-rHONO-VO-^-O-
CV CV CV r-l rH r-l rH

M
fe
OOOOOOOOn
O On
CO CO CO CO -:t CO
M C\' CA VC On CV NO CO O -

E-i r-l rH rH rH CM CV Cv.' ON

93
 I

cm
_4~
CM nO
Cnr^H
u-n.
lONH
u-\ .j-CM
rimo
CO rH ^4~ u"N
O- CM
-4" u"n.
-4" -4"
P N CM
u^ CO O NO [^ rl U> (V
On f>- CO no wnnO »/n.nO [>- -4"
H H
><
Oo'oOOOOOOrHrHr-lrHrHOOOOOOOOOO

X \o o o 4 \o o o
o\ o\ 0^.4- u-\in i/n, u^^j- cA m.4- .4- un,_4- o~\ in
r~{ OO HH CM -I C\.' CM CM CM CM CM CM CM CM CM CM CM CM CO CM CM
I I I I I I I I I I I I I I I I I I I I I I I

O 4 4 N^on^iCO NHinO)Ot^r»NHNH
vD CO On\T)
44^00n
(\1
CM ^-HCOW^r^WO(VCO^OH OnnO
On rH 0"N^t >-0-c}-

OOOOOrHOOOOOOOHCM rH rH H H H rH O
I I I I I I I I I I I I I I I 1 1
r-l
1 1 1 1 1

CM
rlOONOOvNt^nNO
N\D CMT^no O O O N
CMCMCOmrlONCMC-CMNONONOOU~N
rHN N t^f^O HN^WC^NNN
X ooodooodoooooooooooooooo 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

cMrHr-irHHcnr>-c^cMC>Ht^ONU^^-cMoo>^N^cncr\.4^0N
en iH OO rH OO -4" CM -4' O- CV en ^4- CMCAcMCMCMOrHOOrHCM
^ oooodrHrHcMcncncMr-iooooodooodoo
w III I I I I I I I I I I I I I I I I I I

CN
tfc

CO en
O- rH 00 m vA>AOD IAVO
u"n,
o:ONONNONHriOt^ CM CO 00
CT\
O nC
_j- u"N vtn.nO
On i-I
4NONririNON4
CM en en nO
CM C- "N _H; -4"
CO X OOHH O OOO O O
<
Ph
w rH O rH rH rH
I I
r-l H r-l rH rH rH rH rH rH

iJ
H
H
Eh On CO CO CM CM rH CN- U"N -V CO O-nO CO CM H Cn CM I4H !SCO
CO c^ C^CMN H N4 On O- -4" NO »AHH4 A^HONO COCO l
r-l
NUN,
Q
-3
>H
W OOOOO
I I I I I
r-l
I
rH CM
I I
CN Cn
I I
CM
I
HI
rH
I
r-l
I
r~l
I
r-l
I
HrHOrHOOOO
I I I I I I I I

NNNO\4HOOr\00 O- CM rH m vO rH O en O- CO 0~\ VTN.nO 0~\

CN O ON vO r\HHCOCOWv04H4\OWrlHOHOvONNN
vn
X OrHOOOrHrHOOrHrHrH iHrHrHrHrHrHOOOO
I I I I I I I I I I I I I I I I I I I I I I I I

CM
00>AOC04 m nO
o- o- o- co co o h n
vnvnOvnvncMCONONO CM CO ^J- CMO o-
r»4 mvninu>4cN.wNHHHocooD
vtn.

— — 1
I
1
I r-l 1 — I r— I r~i r— I r-l —
1 I HH —
1
I rH rH rH 1 — I r—

CM OOOiAir r\0\\0^u>OU>UN(V00v0vDNC04NOOm
H o in co co
a- h ri »Avnvnvn4nw(\:HriHOONco
1

iHiHrHiH
w 0^4 rHrHHrHrHrHrHH
r-l r-l rHrHH

CN
O U> On CM CM CM
C^ CO CO On
0^
NO Cn CM O
O On CO CM -4- ON -4 On
VT\ CNJ On U"n,nO U^ CM VTN
O- ^t NO CO CN rH f>-NO iT\riP> CM rIOCO
m
rH HH CVJ
r-l
r\4- 4- 4- CNJ CM CM ^rH r-l rH rH rH HHH

CN. OC MriC0Cr>C0(^OO\0C0OC0 0\4OCr>N0N vn.NO .j vn

r
E-i NO(>OnH N4 WON4 H4 N^rlNvfl innr^CVrl oco
rl H N n r»4 4( N(MC\!NHHHrlrlrlrl rl 5
VCN,

M
S
rHOOOOOO
NO CO O
CM
O'OOO
NO CO O
.4-
OO NOOOOOOOOOOOO
CO O NO GO O NO NO
CM -4" CM ^r -4' CM -4
M HH HHHN C\i W N C\) C^4 4 lA vnvO On CM .4-
fH (H CM

94
 1

vOO^^P^nJ- W\0 4-OONn UN H N C^ M^- CO OJ N rl W CM CO vOnO

CM
>H
W
•••••>••••••••••*••••••
OOOOr!00^v0OHWv0v0C000Orl0\0\OO0N0\C0a\NM^NCMO0\
O O O O O O O O rH Hi— — — r W N H H N W ri ri rl
tillI I I I I I
H r H rl O 1 i 1 i 1

1
i

1 1 1 t 1 1
O'.

1 1 1
r-l
1
r-1
I 1
C^v

1
i

f 1
r-l
f
CO

N O O On ^ O O O CMT\
CM O
VO-4 CO
rl WHOOOHN4
WW»Ar>.cO
U^
Q\ CM OnnO On On
CO
NvO N On On ON00 OnOnNKvO^) u>^ 4
-4" CN-nO -4 nO On On CN CO ^-
On
H OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO I I t I I I I I I I I

CMCM-4COCMCMvoOnOCOCOnC-4CMvo VOnO On4 O -4 CO CO On On CO rH nO

CM O rH O CO un 10 CN nO O- rl CM ^ un f>- CO \Q
r-l r-l CN CO CO On CN NO \0 CO CO CO CM H C-- CN- CN-
cn) r-l

W OOOOOOOOOOO H H H H H H r H rl H rl rH rH iH rH rH rl
I I I I I I I
r-l
I I
r-l
I I I I I I
r*l
I I
I

I I I I I I I I
r-l
I I I

CM
Cn- O -H NO
OCJ^nCVrlrlOrlt^CN;
Om-I CO O ON jj UN O On
UN NO NO
[n-
t>-
H H
On O
VT> »T) CNJ OJ rl Ox n
O On O On On O O O CN CO CO CO CN
On On
C°\ (^ ONNO -4' M3 CM UA
X OOOOOOOOOOOOOOOO OOHOHOO OOOOO
W I I I I I I I I 1 1 1 1 1 1 1
r-l
1
r-l
1
r~t
1 i 1 1 1 1

-4Cn-OCOcN-4nO-4 On CV 4 N CO On n lACO ISJ- Cn-C^-J- NCO H On CV O

[n.
N O [n 00
[n-
OVA4 UN^.tV^ rHAO O.Nn©4 VA(\! rl O OnNnO U>N rl H 4 H O
rH
>-(
C£J
................................
O O O O O O H r^4 104 4 4 cnc^^^N N N rl H H H rl H ri rl rl H C\J (\|
I I I I I I I I I I I I I I I I I I I I I I I I I I I

CMCMHCOCMOnCMCMnOCOO-^-CMOC^Ono COnO vo CM CN r! O r! O .4- O -4 On O

rl H
X O H O O iH rH O UN H -4 -4/ u-n\D -4 UN CO On CO CO CO On O £N CN CO NO uo
n un, O NO vO
*tfc
W O O O O O O O O rl rl ri rl iH rl rl rl rl rl rH rl H rH H rl rH rH r-l rl CM rH iH
l/i I I I I I I I I I I I I I I I I I I I 1 1 1 1 1 t 1 1 1 1

c/>
<
Ph
rl
O CN .4 n\0
O COnO
VT CO On CO O NO CO NO Cn- UN a> <n.' r-l CO On CM UN .4- rH CO CM CM CM CO C^- On CO rH
O
1 i—
>H
CN] CM CM VP1.4- CN--NO VfNHCO U"M*^OnC4 Cn! r-l O CT.nO ^a] r-l VT\ CO CM On
r<
W OOOOO
I I I
rl
I
W
I
C^\4
I I
4444
I I I I
C"\ C^i
I I
^ (^ N
I I
C\i
I I
Cvi
I
NN
I I
rl r! rl
I I I
H I
rl
I
OO I I
rl rl
I I
O
I

O O CO
CO
rl CM NO NO vonO -3O CMi^^a.') O- CO CO vo CO CO CO r-i vo .4 CO CO UN CO CN NO r!
rl CM CO VO rl -4" iH C> H O O CO NO -4" UN u>* CO CM On O
O^ On NO NO -4 rl CO CN rl O- NO
W O O O O O O rl
I I I I I I
CM CM
I I
CO CO CO CM CM CM
I I I
Cn! CM CM Cn! iH CM r-i rl rl iH rl rl
I
OO
I I
rl
I
O O
I I
I I I I I I I I I I I I I I

OOO CM -4 O n On O O r 00 (N
CL NO N^CO O N CO H NO O .4 CO O O O VO O Cn! CM -4
CM
E-"
O-r> Cn- ..
f>. o- CN CO O
»

r! CO a}
-4-
r! rH rl r-l r!
I

NO>H CO NO NO CO CO On On On On O O ON CO CO
.4-
H CM CM CM CM CM CM CM CM CM CM CM CM CO CO CM CM CM (N- Cn-
CM CM
no

CM
E-i
OO O
C^ IN C^-
CM CO NO vo vo On r-l CO O- NO O- CO CO
CN CN CO On O CN- CO rH NO CM -4
r CO .4- CN On rl CO _4 NO NO IN- CO CO On On On On
I
O O COOOOIAON4
O O On CO CO CN NO
rl rH rH rl rl r! CM CM CM CM CM CM CM CM CM CM CM CM CO CO CM CM CM CM CM

OO ^4 CO Cn- r-l CM CM O Cn 00 Cv) -CO On CM CO On NO CO CO O rl O CN-NO CN- O CO rH

vTn CM O
_4 CO H
rl
(H
CN- CN CN Cn- CO' H
rl
vo CO CO rH
rl CM CO .4
CN- CM
_4 vta
vOnO CO CO CO CO CN NO NO UN-4 CM r! CO
vo vo vr vf vo vo vo vo vo vo vo vo VO -4
"1 n, -4" .4 .4 CO CO CO
On NO VO

O OC0V04
CN CN CN 00 ON
N COnO vo O O rl O CO
H NO ^> CO ON COnO CO
-4 CO
_4 COnO CO CO
-4 r-l O-NO CN
On On CO CN-MD nO VO-d CM rl CO UN CM
CN rl _4 CO rl
On NO vo
O O O O
rl rl CM CO -4 ^} UN VO vr> UN LO ^\ VO VT\ VO VO Vf-, vo VO VO -3 -3" -^ -^ CO CO CO

W
S
rHOO O OOO O VO
CM -4 NO 00 O CM
OOOOO
CO O
OO OOO OOOOOOOOOOOOOO
j4 nO CO O CM .4 NO CO O CM
-4" NO CO -4 O NO CM CO -4 O
CN! .4-
M rl r! rH rl rl CM CM CM CM CM CO CO CO CO CO .4 -4 ^5-4 -4 UN NO VO (N CO ON Cn-
rH

95
 CM O- CO O jj 0~\ CM CO C\!
cm CO NO -4f -3" CM rH O OO
ooooooooo
I III I I I I

-d"
On
O no
O«•*•••»«
On O O C\
r-l
On On CO O- O- CO CO
i-l
CM
OrHOOOOOOO
I I I I I I I I I

CM On Os C-- VA C^ OS O CM O
£1 ooooooooo
I
C

I
C

I
• «

I
*

I I
t> «

I I
fc »

4- HN4
vOvO-j-oSr-lr-IOOO
c\; O IACO r-l
CM
X
Cxi
•••««••«
OOOOOOOOO
•

I I I I I I I

CM C^> On On CO
On v£) jj 0"\ r-l CM
NNC^H
O Cn r-l

Cxi OOOOOOOOO
I I I I I I

CO

•H W H W N ri O O
4 nr\^ _H P'N.
-P
X 3 H CM CO
r-!

oo fcl r- i r— I
I I
rH
I
i —
I
! r-l
I
rH
I
r-l
I
HO
I I

5!
m r-l On vO vO U"N.-j VTN CM CM O
to
<: OOOOOOOOO
1 1 1 1 1 1 1 1 1

r-l

t-«
CM ^} 00 CNCOONO-O-Nr-i
iH
X
NO CM OO iH O C^> ^)- NO
OOOOOOOOO
1 1 1

On CM CN O On NO NO C'"N O
CM
Eh
4
CM
r\rl OnvO
CM CM r-! r-l
lAW H O
r-l r-l r-i rH

CM On CM C'"N O OnvO NO 0^ O
Ch -d 0~\ rl
CM CM
OnvO VT-i CM rH O
CxJ r-l r-l r I r H rH rH

IAC0 C^iO 0-\NO ^j- Oo

O
H
rH
CM CO NO flOCO^
CM
CM CM CM CM r-l rH rH
0"n. r-l

iH W\C0 C^O C^vO -4 O O

E-i CMCONOfNOCO-cJ-CMO
CN CM CJ CM CM rH r-i rH rH

M
S
OO OO OO Oo o
CM O CO OT CO CO CO 00 NO
M O CM C^nO On CM rl O r-l
E-< r-irH H H r CM Cm -J v .'
\

96
 t-i
NC--(\iVANC0^44N-30DOONNONOONO\CDO'v0
ooHNCMtMln^^o^ooo^o4•l^^4 4u^^^ un no 4-4- 4-(N»AH4-COvO{M\0
r.w r\n onOncmxi 1

W O O O O O O O O O O H H O rH rH rH rH rH rH H rH rH rH rH rH rH O O O O
I I I I I I I I I I I I I I
r-i
I I I I I I I I I I
r-i
I I I I I I

NNCO ^4-4-
nc^HO WCO lA\0 OrlOr!C0\0\0lAO4vCC\!r\HC^0NHNOHv0C^
H HO C OH W N(^r\4 ^^^CvcdcOCOOOCOCOCOOCOCOCOffjCO
H
Cxi OOOOOOOOOOOOOO OOO OOOOO OO OOrHOOO OO I I I I I I I I I I I I I I I I I I I t I I I I

Hr!^HHCCHONa3Nri4
C^HOON ^(^ CM On CM wna»CO mw CM O- UN ON On CO -4" UN
1-!
>H
un\o CO O CM CN UN NO \o O- 00 On 00 00 o- o-\o NO UN
•••«•••••••••••••••••• un,_4 H OO c\; ON
Cxi OO OOOOOOOOO I I I I I I I I
r-l
I
rH iH
I
r-l
I
rH rH rH
I I I
r~l
I
i-i
I
r-l rH
I
r-l rH rH rH iH rH iH rH rH
I I I
O
I I I I I I I I I

vOvOHONNOrir,4 CNnO 4C0\00MAC0rlOC0CV4C^rl0N4OOC0iA0' Ox

rH COCOCO N»A4 W W WH O H W ^1(^4 lT\\0 CO CO OOONCVvONCOGOOnOO
W OOOOOOOOOOOOOOOOOOOO H H — H ri H r! H HN W
I I I I I I I I I I I I I I I I I 1 1
i

1
1

1 1 1 1 1 1 t 1

CM
4 ^) [n W 4" O iA N \D 4"
4-
OOOOCMvNOri4 O 4 H VO 1S4
0U'\riO \OW > O On O
W ri4- 4 O 4 CO
O- il CM
r-i
4 On On N
\0
O On On O
O NO
0"\ Cv
UN. .4- CM r-i
>~t
ti oo I
o o
I I I
r-i
I
0.4-
I I
.4-
I
vrv^
I t
.3-
I
r^nntnnN n w w w h h h h h h o o h
I I i I I I I I I I I I I I I I I I I

rl H U^4- CM CO UN NO NO NO O On H CN UN, O NO CNnO On On O VTN CM ON CM NO NO UN. On

CO O n^4 CN-N0 -4 -4 UN
CV! r-l
CM CM O O rH CM rN. CNNO rl......................... CM CM rH O O CO NO UN 0~\ CM O O ON 1
^tfc X
Crl O O O O O O O rH rH CM CM CM CM CM CM CM CM CM CM CM CM CM CM rH rH H rH O
CN) r-l r-l i-l
co r 1 1 1 1 i 1 1 1 1 1 t 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

cn
<
r-i
I
-
CM
W r! U"\lT( N 4^
CM CN O NO
r-l r-i
^4 0^ N N (M 4 O CO N Cn (A
U\ CO CO U\ ON O
4 CO 4
On CO UN 3
-4
C^CAH 4- [n
O On OO CO ONC^- -4" r-l
CM
rH UN.
CN! H
CM rH 00 00 00
CO ^ H
Eh >->
Cxi OOOOOO rH CM CN-4! -4" -4 -4 CN ON ON CM CM CM OJ OJ CM rH H rH rH rH rH rH OOO
1 1 1 1 1 i 1 1 t 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 t 1 1

CV O CO N4 ^O ON N O. Wi4 O HC0 4 4' NH OH NON4COVD ^^(A

O O CO MD O-O-nO
CM
X CO 00 O- Cn- OnOOnOOOOOOOOOHOnOOnOOOnOnCOCOO-COCO
Cxi O O O O O O O rH CM CM CM CM CM CM CM CM CM CM rH CM H CM CM rH rH H H rH
I
r i r-i i-i r-l
I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I

rH O O UN On On CN CN un vr\ CM O- CM On UN .4- f>- 00 Cn- CO <>• CN CO O vO ON ON 00 ON O- CNnO

.4-
OO
rH
r-i
O
rH
O
H
O rH CM 0N_j- NO
rH r-l
NOnO
H rH CM CM -4CM
r~\ r-i r-i r- 1
CN!
UN NO O- CO On
CM CM CM CM CM
O
ON
O
CN
H H rH CM CM rH O O On
CN CN ON ON ON ON 0^ ON ON CM
r-l

O O U^ On On O-NO NOn H O- -4 O O- UN NO ON U)4 O NO CM ^V ON CN CO CN O- CNnO

O O CO
0"-,
rH OOOOOriNN4iONCOOOjCN
rH rH rH H rH rH rH H rH CM CM CM r-i r-i r-l
r-i

C\! Cn:
C- 00 On
CN! CM CM
O
CN ON ON
r-i
C^N
r-i
ON ON C\
CM CM rH r-l O On
0"N CN CN CN CM
O

CM OO^C\COin4
O O O O CN O- O
CN! ON o- CM O CN LP, O- UN O r-i 0- CN UN r>- O OnOJ^O OvOCO O

H r-l H CM
r-i r-l
r-l
r-i r-i
.4- C2 On UN O Cn UnnO vO NNCn-\OnO Vv4 4 CM O ir, CO O-
C\ CN _4" UN UN. U U\. U\ Un. Un UN U> UN UN UN UN, UN UN .4- _4 _4 ^) ON C\
1
N WO

CM
O
O O
O
UN On On CO
O O CM
O r CN O- -4 O- CN UN. 0N\O CM vO O- CN O
CM O NO rH
-4" UNvO CO
\D O- 00 00 CO CO 00 CN-NO UN.
r-i On
Cn
r-l -4/
CO UN
NO Ci
CN.
O CN
O On O-
r-l

E<
i

rlHr:HririHWC'N.44iniri UN. UN U\ IT,

-4"
VJN IP, UMA UN
_4"
^^ .4-
IT,
r-i
VT\4 ^4 _4" _4 CN CN

Cxi
£.
rHOOOOOO OOO OOOO OO OOOO OOOOOOOOOOO
N4vDC0OC\'4\0C0OW4n0C0ON4vDC0O0!4n0C04On0NC04O CN
Hi r-l rH rH rH rH CM CM Cn! CM CNj CN ON ON CN ON -4 i} 3- -4 -4' U-,vO NO O- 0- CO On

9?
 O CNNO ur>,\0
rH \0v04WWWWOO
C~- CN. CM u"n.

W ooooooooo
I I I I I I I I

rH xnvnr\w cm nwoo
w ooooooooo
I I I I I I I I I

OCN-Cr\CMvrN.r-lvr\._3-u-N.
r-l NinnHOOONCM
W OOOOOOOOO
I I I I I

C^-vOO- CM CO CO
C*"n. U~\nO
r-I iHrHrHOrHrHOOO
.........
x
W CM CM CM CM CM CM CM rH CM
I I I I I I I I I

ri
CMO O
rMn r> h NO
4 NO
isints
CM
>H
m
»••••«•••On
rHiHrHOOOOOO
^> O- -^

1 1 1 1 1 1 1 1 1

©
c
+•>
CM
X CON^nriOOrirl
.........
rj
O
O W OOOOOOOOO
I I I I I I I

$1
i-IOOCMONrHCOvOO
CO CM vO^^riOHOnc^i
<
Pn
ca ooooooooo
till
Hi

Eh
CO -4 r-l ON O- CO CM On -H-
CM OCO n^^.^-j- woS
pel r-l r-l r— I r— I r-l rH rH rH r-l
I I I I 1 I I I I

r-l
e-i
CN vr> -4
r> vt\ on,
rH
on
O
o CN-
r>-
CM On CO
\o cm rH
CM CM Cv! CM r-l r-l H rH rH

rMT\4 onh nwoco

r-l O- \T\ pS O On O-NC CM r-l
E- CMCMCMCMrHHrHrHrH

CM vO rH CM CN CO On CM On CO
E-i H
CM On NO CM On O- VD CM
riOJWWHHrlHH

NO r-lCM r>- CO On CM On CO
CM CN! O NO
n CM O > t>- NO CM rH
E-< (^ CM CO CM rH rH r-l rH H

M
S
OOOOOOOOO
CO no -d -3 -4 -4 -4
r-l
E-"
O
rH
W4 O- PNNG CM CO
-

rl <H rH CM CO CO CN 0"\,
O
-3- -3-

98
 -3- OO
O- in in rl ri CO
cn O O O O O
in On inNO OnnO CM On CO -H- CM O-nO CO CM c-l NO CM cncO H
CM
>-*
W
».••••••••••••••.••••••••
cn in
ri CM u '-wo CO On On CO 'A
OOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
CM CM o- O- O- CO On On On On CM- NO O-

I I I I I I I I I I I I I I I I I I I I I I I I I I I

CM
vOCO
H
no
CM rH CM CM
r> en NO -r) NO cn r- in On CO CO in -3" CO CM On no
-3" »n UN. ^n<V CM O O i-i rH CM r-1 O ri O i-l O O N
h nco
r-1
On
O O
-j-
r-l
r-(
r-|
en O.
r-l O
x
W OOOOOOOOOOOOOOOOOOOOOOOOOOOOOO I I I I I I I I I I I

- 00 On U"\\0 cn O cn rH C*\ O- .;}- OMA -3 CM nO CM NO O -3- CM cn CM in On ri en CM ri ri

CM
>*
CM HOH CM CM CM O Cn O- On r-l eM CM CM O O CO CO O- CO O- NO in NO NO NO NO NO NO
fy OOOOOOOOOOOrirlririrlriOOOOOOOOOOOOO
I I I I I I

O O NO CM- vtn. cn O NO O- CO CM -d- CM -^ -3 -3" ^n CO CM On O-

^n cn ^ m
rH C^ CO CO -d -CM -d CO NO NO
CM r-l rH r\ H rH rH CM Cn^J NO O- CO On CO CO MD C^-NO .j- vn_j- _j- .j- ^J-

K rl H H
f I I
rl rl
I
rl H
I I
rl r!
I I
HH
I
rl r! rl
I I I
HI
rl
I
HI
rl r!
I I
HHHHHH
I I I I I I
rl
f
HH
I
ri rl
I I
I I I

NV04 niPiCNrlOCNO-d O-O-^NOCOCn^-eno-CNJCnCMO OnvO

en
>1
OOH4COncOHVAOO\Or!Or!r\4 u\ir\\OvOvO ^^44 44" NH
O- rH -3" rH

W OOOOOrHCM-^--^--cJ-CMOOOOOOOOOOOOOOOOOOO
I I I I I I I I I I I I I I I I I I I I I I I I I I I

^ Cn CO rH NO CO On ^ rH O U^ r> CM O in CO CM O O CM cn en ^ -3 CO rH NO On u~n.
Cn Cn OOOOCMinCN-.jCMCMri^-CniHOOOOriOriOOOOOOrHOri
..............................
-ch

^fc-

cn
><;
H OOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
ii i i i i i i i i i i i i i

1
rH
i
-
en
On On On
ev
O- O- CM
ca^^oco^- vO O O H On rH r!
rico oncc wnw
C^- r-l cnvO ON O CM CO cn CO
o o h^ ^^^ir^
o o CM -H;
c\]
r-l
1-
C>~\
iri r~i
O NOcn O -j-
EH >-<
W OOOOririCMCnCMrlOOOOOOOOOOOOOOOOOOOO
I I I II I I I I I I I I

CO r-4 ^V rl jj- Cn CO rH cn OO
On en O CD OnvO rl UN H 4
O- r- O- ^J CM •<> r-l Cn- Cn-
o CMrHOrHCMCMCMOCO
!

Cn CM ^- Cn C^n O- IT"! rH rH Cn- \C O- CO NO >n^J- Cn (n en .3f o-

M rH A rH i-*l rH OO <H rH O O O O O O O O O O O O O r H rH rH rH H O O .

I I I I I I I I I I I I I I I I I I I .1 I I I I I I I I I I

cm
Eh
o o \o On cm
OO O O ri cnvo
rl r I
Hn0 4
cm en cm no
CM CM Cn^} in NO NO NO NO in un,
m
en co ^j- cm cm .h o>no cn On co cn u\ r \ en
^J- en CM CM CM rH rH r~! m O O O
rH
r-l r- 1 r- 1 i— I r-l rH r-lri rH ri i— rH rl r rH i-( rH r rH i—l rl — r-l r-l rl r! rl rl rl
I I I i I

o o cn in nC' no o no r-i no no no r-i co cm o

no no c^ -3 cm en o- in en cn o o o^
CM O O O O O O rH r-l CM' CO cn .j- in NO NO in ITN V\ jj- (n CM CM CM CM rl r-l OO O O
EH rH rH r-l rH rH rl rH rl r! rH r r! rl rl r-l r— r— r rl r^ rl rl rH ri rl rH rl ri rH rH
! I I I

en
EH
no co en gn\c
O O H CM .3 hvc
00 4
cn:
r
co cn
CM
o on
in
^j cm co en cm co cn.H o> on ^ onno o cn o- en
rH O CO O- O- O-NO NO m.:} cn Cn CM CM CM rl O O
H -3
r-l
r H ! ri r-l r-l rH OJ
^J-I

f^nr'\ CN.' NH rl ri rl HH rl HH rl r I ri rl rl ri ri ri

o oo o ^ on in o co no ri ri in r>. er ^
no co en on o- cn in in Cn no no en o-
en
f>-! ri r-\
r-i
r-i
r-\
r->.
cm
r-i r-
1
cn
,
cn-
ri CM CM
_-v
en co ri o
^j- o On o. r>- o- no no no
C^\ CM CM CM
j$ en cn cm
r->, ri r-i i->, r~i rl H
m r~\ rl r! ri rl
^)
cm
r-i
r i

ri ri ri ri
ho o

W
S ri OOOOOOOO
CM O -j-
OOO OOOOOOOOOOOOOOOOOOO
O
NO CO O O O O CN.' ^+ NO CO O! ^J NO CO CM ^} NO CM CO -^ NO CM CO NO
ri ri ri ri ri ri cm o.! cv cm cm en cn e- \ cn ^} ^j vt\no no cn- c^. o\ cm in co
EH ri rl rl

99
 CM GHU^ 00W 4CA 0- O -3
UN, f>.
un.\q 00 On
O CA 0~\
O rH CM r-l
UN,ON NO CO
On unnO O-nO
4 4 CO
C^
CN- CO
Un> C\! CO r-l 00 U\ CM CA 0~N.nO
On CO 00 C^ CO C-O-nO va
(N-
CA4
CM On
On
W OO-OOOOOOOr-l
I I I I 1 I I I I
r
I
I r-l
I
OH
I I
rl
I
HH
I I
r-l
I
r-l
I
r-l
I
<-|
I
H
I
r{ rl
I I
HH
I I
rl
I
HI
rl
I
HHO
I I I

O vCO^COOOincviAHNinr^O Wf^4
O C C\_3 4 NO CO VA CM H OOCO NHOCinM^^^
CM iH CM ON N^ri ON CA4
NCON
4 C\ CJ O CO
CM
x
C£)
O .•..••..*«••.....•••**.•*••••.
r-l

OOOOOOOOOOOOO
r-l

H rH rl ri iH rH ri
CN-

ri O r-l r-l r-l r-l r-! r-l

Cn!

r-l r~t r-l

r-l

r-1
I I I t I I I I I I I I I I I I I I I I I I I I I I I I I I I

4 CM CO O CN NO On 4 U"\ On O CA CM On VA
UN, CN- VA 4 O-nO ri NO CO 4 CO UN, (NJ CN- CN- CN-
CM OHr;»A44'V)^03a.NNO\0tD0>0Na.Or:WHHHH0\avO>C0iniAO
W O O O O O O O O O O ri ri ri ri ri H CMCMCCCMCICMriri H r- 1 r-l r-l r-l r-l
1
r-i
I I I I I I I I I I I I I I I I I 1 1 1 1 1 1 1 1 1 1 1 1 1

CM O 0\CN-OmCOvDC^H\OWHvONNONH CAvO O 4 4 CA O ri UN CA CA On CM CO CM
O O ri O ri CA UN NO nOnOvOOnCACMOOOCM4 vAMj O CO On CM CA 4 UN UN NO Cn-

W O O O O O O O O O O O O O O O O O O O O O O O O O rH H rH
I I I
H I
r-l r-l r-l
I
I I I I I I I I I I I I I I I

4N HNO
W H O ^ H r\ O
UN.VO CA NO CAU^OCMOOnCACMCN-COOncn-4 4 VT\WCO ChC^vO O N
UN,
IN O ^4 WC0 4 HC0NDC^Of>\04 C\'Of>-Or-iOOOOON
rH
>-<
ta
................................
ooooo H(vr»4in^u-^ j-^ nr\n^w wnnnh nhhh h h o
i t t i i i t i i i i i i i i i i i i i i i i i i i i i i

C^O^^t^(^J^^^^.ocOO^C^^4^^r^^nr>W^OO^ CAnO CAvO On fM^O o-

rH
=$fc
-
r-l
X
Of^rMf^OCMr^ nn CA CM O On CN-VO N4 0"inNOOMBO\nOO
................................
CO
0"NON^ C^-

00 W O O O O O O CM Ci CA CA CM CM CM CM CM CM CM CM CM
r-l ri H H H ri O O
0~\ C>, 0~v. r-l r-l
CO I I I I I I I I I I t I I I I I I I I I I I I I I i I I I I I I

O CO VA O- 4
W HOH
CO VA O0
H O NO
N CO CM On CA VO O- CO UN, CM
W\4 ^4 (^ vAnO IAN OO
CO On CO On CA CO -4 va CM CO
N4 O 0nCO4
r-l r-l -4-
E-i rH ri CA CA CA C^.4 4 CM On
nO
O
O &3 o o o o 6 o d ri n n4* 4' 4 4 4 4 4* cA ca ca cm cm cm cm cm cm AHH HH o I I
I I I I I I I I I I I I I I I I I I I I I I I I I

r-^
UN
O ri CO
UN NO
CM CM CM
On ON ON CC 0\ VT, H 4 CO
CM CN O
r-l
O NO On CO
r-l CN.' CN-
CO 44 CA CA CA
On CM CA CM CM CA CM
4 CnI
r-i
CO UN NO
On O
r-i
CM CO -NO NO NO
P"\ 0^\ nN ON
U> CA
£1 O O O O O O O O O O O r ri ri ri ! ri CM CM CM CM CM CM CM ri CM CM ri ri ri r-l r~\ r-\
I I I I I I I I I I I I I I I I I I I I I I I I I I I I

cm
04CONnOOOWWNC°(On4
cc co on o ih cm ca vanc
r>- cm- r>-
C°>4 CN^COOCOvOWINrlvOriNrilNONOCO
o ri c ca 4 no no mocoo\0\oooo\oooo4 cn-
E-" ri ri CM CM CM CM CM CM CM CM CM CM CM CI CA CA CA CM CM CM CM
r-\ r-l r-\ r-i r-i

CM 04CO CM NO CO
CO CO CC On
CM 4 O 4 vOrivOriO CM v000CN-OOCn4
CM -4 Un.nO On O
U^ CO VO CO On C- On O CO
CI CA unno no cn0DC0COON0n0n0nC0C04
r-l
E-i C^ r>- Cn- r-\ r-\
ri ri ri ri ri ri CM CM CM CM CJ CM
r-\ CM CM CI CM CJ CI CM CM CM CM CM Cn!

O4 OOO O CM On o UN. CA CM CO NO
O4
CC C- CM va CN Cn On CM-NO -4 CN- VA CM CAnO VA CM
ri CN- IS- C- CO O
CM C^. NO O- nO CA CM- CM CA CA CA O CN. CM r-\ On CO nO nO CM On nO CA r-\ CM- r~\
E-< t-i H, r-\ CM fA_3 vj"t un.n£) nO nO no NO nO NO NO no UNU^iAmu>^ 4 44 CA CA

r-l O j4 CO CJ O O ^J- ri CM NO U"> CM ri UN, o: VO UN

C'\ O ^ CO CO Cn- CA Cn-
00 CM (A NO .4- CM On
O o
^A CA OJ CO NO UN, CM
On CO NO NO CM On nO CN r~i [>- r-l O
O
E-h CN- Cn- CN- CO CTn ri CM C> O-i CM ri
r-i ri r- CM CA j^ UN
1 U-, NO NO NT; NO NO NO NO NO ir\UN,lT\irN,»A4 -4 -4 CA CA 4

H t-\ O O O O O vO
O O vO O O CJ O O O O O O O C O O O O O O O O O O O no OO
S
M CM CO O
-4" -4 CO O CM .4 NO CO O CM -4
Cn' CO O CM -4 vO CO O NO CM 00 On vT- -4' -4'

r-<
ri ri CM CA CA ON CA CA
ri ri r-l CM CM UNVO VO CM CM -4 4444 C^ CN- r-i
rri

100
 CM
H CM V~\
P"v. CO
ON O- CO vO vO
-H-
-3"

oooooo
I I I I I I

Vn CM Cn CM (X!
CM cS O- r>- VT\ U-\i}-
X oooooo
P0
I I I I I I

VTN-cV CO CM u"n,vO
CM O 00 CN-VO vr, C"n.
rH OOOO O
I I I I I I

H vtn \r> o-
u"n ^3-
CM
X
W
•
rH rH rH
••*«
o- vo vo >a
HH
_3

r-1
u--»

1 1 1 1 1 1

H O
On
C~\ O- On
C*- C"n
CO VO VO _^-
CN-
>H
w OOOOOO
tS 1 1 1 1 1 1

0)
3
c
•H
-P
r:
»
rH
J-^N^OH
CM rH r-! r-1 CM -j-

oo X t

oooooo
* r t * •
&q
1 t 1

£
CO H CM vO VO O- VO
CO
<
fL,
i-!
->H
W
•«•«•
CO vO VA C""\ 0"N rH
oo oo oo
*

I 1 1 1 1 1
£?-.

H
E-<
ft;
o
o r-1
oo
CN. CM
tf"\-d-
PS
O- rH
rH CM
0~\
O
X
w H1
r-1
1
r-1
t
r-1
1
(H rH
1 1

cm
rH
ca
(A OnvO
r-! co vo vn
H CM
cm
E~< CM CM rH rH rH H

Cv! rH Cn. C'nvO rH CM

E-t r\HCOvC VP\ CM
CM CM rH r-i HH

On O VT\ \T\ VT\ rH

rH
En
r>-
CM CM CM
m
r! CO vO
HH
C\
r-1

rH
Eh
On
rv.
O
VO
>A
p-i
if"; mH
CO vO CA
CM CM CM rH rH rH

W
S O OOOOO
CM O O O O O
M 0"MT\00 C r-1 .:' i

E-< rH rH ri CM CM CN

101
 uw^OD u^ N O O O W u^^
CV
>^
W
•••••••.....•'•••
I
M^
-4 -4 -4

I
Ox O^OD
>-r

ooooooooooooooooooooooooo
•

I I I I
oo
O- O-nO <]

I I I I I
c^ 0\\C 0\ V^ (>

I I
ri

I I I I
c\'
Cnu-n.no O- CO 00

I
C^C^vO

I I I
03

I
C-
(N- un, un,^j-

I
00 00

I I

O CV N N ri CO vO On n On
nwrioo
-4 (\i W N N H v> 4 n^ vn
ritvooHHHN 00wc^nr^tnc^c^^^^ C^- (\) U"\
cm ^-d-
M ooooooooooooooooooooooooo
I I I I I I I I I I I I I I I I I I I I I

UN ^riHOCv
iri WO\£) ^h ^\o o iahonh o \o ^^^a ^j-
(V 4 4J- «"\^0v0 4 W O r\4- C^WHOOHHHNC'iCMN^ fAi

W ooooooooooooooooooooooooo
I I I I I I I I I I I I I I I I I

Vn UN NO vr\0 W N UnnO NO NO Cm -4 U~n.nO

C N CN CM CM CN 0^ CN u~, UNM3 M3 O- O- O- CN-
^OsNHOON^OP'iav NO NO UN,
CM CM-O- 00 CO CM- CM- CM-
X
pq HHH
I I I
ri rl rl
I I I
HO
I I
ri
t
H I
rl rl ri rl ri
I I I I I
ri H H H r! ri r! H rH rH
I I I I I I I I I I

vo 4 h (^4 o co co f^i C' r» n ^vo \o o.^, (v ts r>4 od on n

CN
>H
O O C- CM O On H O U\nO O On C\ CM 0"N C\ (^ W O Q\ On On
O.........................
r-i r-i

W OOOOOriCMCM-4 CV r O r
I I I I I
rH H rH
I I
ri
I
ooo 1 1
I r-1
1 I 1 1 1 1
r-i
1 1
r-i
1 1 i 1

CM - O C\ On On CN ITn UN On On CO NO UN, CM C^N C\ On 00

c\^ vn u^vo o\o4ooo N44 CM rl ri ri J-
UN, On NO O Cn U"N, .4-
sfc cn. vo r>. ri ri CM CM CN CN

Vl W O O O O O O O rH A rH ri O O O O O O O O O O O O O O
CO I I I I I I I I I I I I I I I I t I I I I I I I I

W
rH 0~N
vO C- On ON C"\C0 m
CM (A On
CM -4 VO ON CM un, CM CM CM- CO CN
H N NO
O N N\C C
un On O
lOCO 0^4 On r\ O iA
CM unkq CM CN Cn CN 00 On
4
O
O •&! r-i
I
rH
I
HI
r-1
I
CM CM CM CM CM CM CM
I I I I I I I
CM H
I
rH CM
I
r-
I
!

I I
CM CM CM CM CM CM CM rH rH
| | I I I | I I I

rH On CM On
O C^- NO CM CO CM- CN CM- CM UN CM NO CO O NONCACr\4 00 IT\ CM
CN
X CO On rH CM O CN CM -4 On CM NO CM r-i O O H O O O rH CM rH NO On
W OO
I I
rH
I
r-i
I
HI
ri rH rH
I I I
r-i H I
CM CM rH CM CM CM CM CM CM CM CM CM CM rH rH
I I I I I I I I I
I I I I I I I

• O Cn. un Cn vnvC O CM CN un un, un u\_4- CO -4 CM vr\ On

CI CM CM CM CM CM CM CM H HHO
00 03 CO
O O O oo O H H C-
»tn, CM-

HH
CM On On On r-i 03
E-< ri ri rl r! rl rl rl r! H H rl rl rl — — rl H ri ri i I i I

CM CO 00 03 O Cn u-\ 03 CN vtnnO C- O CM C\ un un un. un _4 o3 -4 CM MD u^ r>-

E-" On On On O O O O
rl ri i— rl
r! r! rH rl CM CM C! CM CM CM CM CM
rl rl — ri — rH rl rl i— rH ri rH
r-i r! H O On CO
r-i rl rH rl
I i I i I I

O- r! NO O Ci CN CM rH .4 O
O H CM UN CM O C -4 O CM- CO NO rri CN O CM VTN UN -4
H rl ri H H
CN CO On On OnH rl O- CN O On CM-M3 M3 £>- un,^ o'N CM O On
C^ CM C" \
0~\ CM W O, rl ri ri ri rl ri i~\ rl ri ri

CN fM- rl NO O NO NO C\ ri 00 CM O
COW ,OON^C
O O .4 O CM- ri CO NO C\ O CM U~N UN, _4
E-< 00 On On OO rl C\ X
ri rl ri r! ri CM CN. C\ CM
On
CN.'
Cm- C^
CM rH rH r! rl rl
NO NO UN.
<-\
4
r~l
f/N. C\!
r^, i~l
O
rl
On

K
S riOOOOOOOOOOOOOOOOOOOOOOOO
4 O CMO O O O
NO 03 CM .4 NO CO CM -4 vC 00 CM .4 NO CM 03 -4 00
M r- 1 r-lri ri rl CM CM CM CM CM CN. CN N CN.4 -4/ UNM3 (>_ s
.

102
 !

rH
U\
H
UN. r-!
CO 4
Cn VO CO Un. C^- CM -4"
VO O- CO ON CO On
-4" -3-
HO O r-i CM O! CM
cn 00 PlHvO
Cn -4 O^
UN. Un,
^ r^i
On O O O- cn
O rMT\HO
ooooooooooo
I I I I I I I I I I I
HHH
I I I
r!
I
HO
I I
H HO O OO
I I I I I I

un,On cn H cn On rH vO N P^ N N CM O H cn
_4 -3"
4- N H H r! O O O H O r! W W O! ^4'
un ^> On O 4 en
On -4
rH
X 4 O VO CO ir» O-vC CM r- o
ooooooooooooooooooooooooo I I I I I I I I I I I I I I I I I I

VO
OO -cV cn CO n^c^ONW N C^OO
U\ O-CO CO On
CM CM J- CM O
{>- -4- VT\IP|NH C^
U^ rH O
CM CM Cn.^- CM r-l cnij ^} LT\ij- 0"\ On CO ij- w
OOOOOOOOOOO r
-! HH -1 rH H H H O O O O O
o
to
crj

w
i i i i i i i i i I I I I I I I I I I I I I

CT< 0"^ O CO U\ Un,^}- N C^ CO H N rl H CO

C~\ C\ \T\ On NCO CO -;! 4 lA
OnO iHCOvO O O O H rH Cn O O CM Cn Un, O- r>_ On CO r-i Cn OS in UN, U
-p
O rH rH O OOO OOO O OO O OO OO C* rH CM CM CM CM CM W
I I t I I I I I I I I I I I I I I I I I I

o
£1
-P
o Osco O On no
O Cn C"N UN, o
H -3 On U a CO CO vO On CO On CO cn r>- -4" r-'.\Q s:
CM 6'n cnvb UN. UN, On UN, f>- r~\ VO H
CV -3 O
ONCO r>4 rH CO VO to
OOOOO
I I I I I
-i cm -3-
i i i
in in in
i i i
m in
i i
.3
i
-3-
i
cn
i
c<"^
i
en
i
cm
i
hi
rH
i
Hhoo
i i i i
O
r-l

-P
co
Cn en on o o-vo vt\On cn on^ en enco cn.3- On o en in en cn.H- h
=& CM
X r-l CnvC VO On CO ffi W 4 N en
ON O rl O O On ^4 r\ r^- Cv CO NO
en

o o HO O H H cm en en cn ^j- ^- ^v ^j- en cn en cn cm r : hooo

I I I I I i i i i i i i i i i i i i i i i i i i

3
p< n
CD

ck r\ ca On o vo o co rH en en co vo c^^ envo on u> cm

o o o co co r -i
ovo 00 00 vo en 0\ u\ un cm o o in o o- H
cv) r-i
I
cr,
CV! cm en o r-i .3- r-i

O
o 6 6 6 O O O W r'\4 4 4 4* 4 ^(AcAn cn rH rH rH O O O
I I
r'l
I I I I I I I I I I I I I I I I I I
rM
-P

co
x:
•p

CM
UN.C0 On
U\, CM ON Cn O CM CM
Hn^ririN^UNCC^NOri O- rH rH
^J-. r-l H CO N rl
UNVO O- On CO O\vOv0 4
C^- UN. -;> CI CO
cn UN,
N
O -p
X OO O rH rH CM CM CM CM CM CM
W
r-i Cn.' Cn 1 CM CM CM CM CM CM cO
I I I I I I I I I I I I I I I I I I I I I I I fH

in
rH
4
cn-
CO r, u-, on o-
[n.. 00
co co on
m un cn cn\o CO CO co
UN,
h cm cn^v u^\o co j>-
r>-
on
VO CM
o
CN!
oj
Hvo
m.^-
^
O- ^r UNCO
cm On ^
r-i H rH r-l rH H r-l r~! r-l H CN) Cv! CM CM CM CM H r- I 1—

CO

rH ^7 co
Cn- CM-
en NH n vo vo
un.
CO CO CO On On O H
jj-cn un.no 00 on On co vo
CO cn.^ U>vO f>- CO On O
^ vo o m cm u>oo
H 4 4 On^ O C^j
ON
r-l
r-l H rH r-l HH H r-i H r-l CM CM rH
Cn] CM CM r-l H ch
£1
CO

4 O ^oofMNio H O- o co od cn Cn CO
co cn- cn on vo vo un CO -4 CO
CM C^- CC> On OH -4 Cn O CO envO
Cn.^- ^
CO On On CO CO VO u h -4 o \ un, r>.. un.
rH H
r-l r-l OJ U, Un UN
UN, U>jj Cn CM u-N \f\ UN, UN, C°. r-l i~ I

o
w
>^
CM 00 N4 O
O Cv' 00 On 00 O- rH
^ r-l CO 0~\ U\ On O- U",
m^ cn cn c-- CO CO -4 Cn: VO ^> 00
O O
Cn- 00 On
r- 1
r-l
iH
Cv' VC:
r-: r-l cm
-d'
e
C^\ On CM
? ^ ^j- ^j- \s^ u ,
UnnC vO VO VO
in ^^^^^4 _r]- O- r-l a} Un
cm rH rH <H
O
P
;j

Eg OOOOOOOOOOOOO
4 O O H OOOOOOOOOOO
O
VO CO OOOO CM -:^ VO CM VO CO rr \0 .-$ Cv'
O
c
M CN.'

r I r-l r-l rH r-l

i

CM CM CM CN,' OJ C' i
Cv'
fi C Cn n U i C- On CM UN.C0
r-l i-l r-l a;

103
 REFERENCES

(l. ) K. Masubuchi, F. B. Simmons and R. E. Monroe, "Analysis of Thermal

Stresses and Metal Movement During Welding", RSIC-820, Redstone
Scientific Information Center, Redstone Arsenal, Alabama
(July 1968).

(2.) J. B. Andrews, M. Arita and K. Masubuchi, "Analysis of Thermal

Stress and Metal Movement During Welding", Report No. 71-1,
Massachusetts Institute of Technology, Department of Naval
Architecture and Marine Engineering. (December 1970 ).

(3« ) R« Tanbakuchi, "Comparison of Measured Metal Temperatures During

Arc Welding with Temperature Calculated Using a Finite Heat
Source", The University of Wisconsin, Ph.D., (1967)1 Engineering,
engineering, mechanical.

(4.) V. Pauclic, "Temperature Histories in a Thin Steel Plate Welded

with Tungsten Inert Gas Process", The University of Wisconsin,
Ph.D., (1968), Engineering, engineering, mechanical,

(5« ) Richard E. Holt, "Flame Straightening Basics", We ldi ng Enfi'ineer,

Vol. 50, (September I965 )•

(6.) R. A. Walsh, "Investigation of Distortion Removal in Welded

Structures", Thesis XIII-A, (May 1970 ), M.I.T.

(7. ) David A. Duffy, "Distortion Removal in Structural Weldments",

Thesis XIII-A, (May 1970 ), M.I.T.

(8.) H. Pattec, R. M. Evans and R. E. Monroe, "Experimental Flame

E.
and Mechanical Straightening and its Effects on Base Metal
Properties", Project SR-185» Battelle Memorial Institute,
Columbus, Ohio, (July 19o9).

(9.) W. M. Murray, "Stress and Strain Relations", Part of a series

of lectures presented at Wayne State University during the NSF
Summer Institute on Experimental Strain Analysis for College
Teachers, (July 25 to September 2, I966).

(10.) Daniel Rosenthal, "Mathematical Theory of Heat Distribution During

Welding and Cutting", Welding Jou rnal , 20, (5) Supplement 220-S to
23^-S, (19^1).

(ll.) T. Nakada and S. Hashimoto, "Heat Conduction in a Semi-Infinite

Solid Heated by a Moving Heat Source Along the Boundary",
Bull JSME 6:21, 59~69, (February I963)'

(12.) N. Y. Olccr, "On the Theory of Conductive Heat Transfer in Finite

Regions", Internal. J. Heat Mmss Transfer. 7, 3, 307~31^i
(March 19&J-).

104
(l3« ) N. W. Rykalin, "Calculation of Heat Processes in Welding", Moscow,
(I960).

(l^. ) K. Masubuchi, "Control of Distortion and Shrinkage in Welding",

WRC Bulletin, (April 19?0).

(15. ) J.N. Goodier, "On the Integration of Thermo-Elr.stic Equations",

Phil. Ma r?. . 23, (1937), P. 1037.

(l6.) Alexander Mendolson "Plasticity: , Theory and Application",

McMillan, New York, (1968).

(l?.) C. F. Lucks and H. W. Dean, "Thermal Properties of Thirteen Metals",

ASTM Special Technical Publication, No. 227, 0-958).

(18.) L. L. Lengyel "An Analytical Correlation of Heat Flux, Nozzle

,

Diameter and Melting Time as Applied, to Nozzle Burners", Apjol*. Sci._

Res. (A), 13, V5i 393-4-00, (196^).

(19.) H. S. Carslaw and J, C. Jaeger, "Conduction of Heat in Solids",

(2nd Ed.), Clarendon Press, Oxford, (1959).

(20. ) "Aerospa.ee Structural Metals Handbook", Mich. Properties Data

Center, Air Force Info. Center, Belfour Stulcn, Inc.

(21.) U.S.S. C. - "U.S.S. T-l Steel, A Proven Engineering Material", (1959).

(22.) U.S.S.C- - "U.S.S. Corten High Strength Low Alloy Steel", (i960).
 —
Ljaulora —
SH ELF BINDER
SSS Syracuse, N. Y,
-— Stockton, Colif.

beaJyy
l n
off]
'
9 of -/'*me

?l
D IS p
LAY

Thesis
J593
127351
Johnson
Study of flame
heating of steel plate.
thesJ593
Study of flame heating of steel plate.

3 2768 002 10799 7

DUDLEY KNOX LIBRARY