ORNL-3500.txt

 

3 4456 O0L34Ll48 &

ORNL=-3500
UC-25 — Metals, Ceramics, and Materials
TID-4500 (23rd ed.)

 

 

 

FABRICATION OF THE HEAT EXCHANGER
TUBE BUNDLE FOR THE MOLTEN-SALT
REACTOR EXPERIMENT

R. G. Donnelly
G. M. Slaughter

-

 

 

OAK RIDGE NATIONAL LABORATORY

operated by
UNION CARBIDE CORPORATION
for the
U.S. ATOMIC ENERGY COMMISSION

 
 

 

 

LEGAL NOTICE

 

This report was prepared as an account of Government sponsored work. Neither the United States,

nor the Commission, nor any person acting on behalf of the Commission:

A. Makes any warranty or representation, expressed or implied, with respect to the accuracy,
completeness, or usefulness of the information contained in this report, or that the use of
any information, apparatus, method, or process disclosed in this report may not infringe
privately owned rights; or

B. Assumes oany liabilities with respect to the use of, or for dameges resulting from the use of
any information, apparatus, method, or process disclosed in this report.

As used in the above, '‘person acting on behalf of the Commission’" includes any employee or

contractor of the Commission, or employee of such controctor, to the extent that such employee

or contractor of the Commission, or employee of such contractor prepares, disseminates, or
provides access to, any information pursuant to his employment or contract with the Commission,

or his employment with such contractor.

 

 

 

 
ORNL-~-3500
Contract No. W=7405-eng-26

METALS AND CERAMICS DIVISION

FABRICATION OF THE HEAT EXCHANGER TUBE BUNDLE
FOR THE MOLTEN-SALT REACTOR EXPERIMENT

R. G. Donnelly and G. M. Slaughter

Date Issued

OAK RIDGE NATIONAL LABORATORY
Oak Ridge, Tennessee

oveon Saenee coreomon [ FTETANH

U. S, ATOMIC ENERGY COMMISSION
3 445k 0l34L48 g
INTRODUCTION ......
JOINT DESIGN ......

PRELTMINARY INVESTIGATIONS
Welding c.eeecaes

Brazing seeesesas

Technique Development

* S0 ¢ 0w

Alloy Selection ..ievuvs.

FABRICATION OF MOCKUP SAMPLES ..

omall Sample ceeiseecccorsasnans .
Large Sample ...
FABRICATION OF MSRE HEAT EXCHANGER TUBE BUNDLE ..
Welding and Brazing ....

Inspection .....

ACKNOWLEDGMENT .....

iii

TABLE OF CONTENTS

............ P A & 00 S0 O Passa e

e 0 T e Qs s s 4 & 8 & 8 & - 8 - . .0 . [
. . @ . " e & 9 & 8 . L 3 B % a8 e » s -

s 80 a . . L N ) - e 8 ‘. 8 4 8 2 4 s s * 8 W
" 8 & 08 4 & d o 5 & 9 B & & & 8 s a0 LN - eS8 -
[ ] ¢t o e s 0 * e 2 » s & & a8 a0 ® 0 0 0 & .
S o 8% 8 0 " B B . e ® & 5 & 8 &8 & - . * e 9 o s @

=9 0 ¢ 8 s o . ® o & 5 00 . + & ¢ & & 5080 * .
8 & 0 29 000 . & . 0 . . * s .

e s 80 e 9 &0 s . 9" - . . .

........ ® & ¢ » & 8 " 8 " e . . . .
* 8 & & 8 - * ® s .

& 8 o 0 s 00 0 . 008 0 . s &o s . - -
llllll . & * - LN ] . 2 & © & 9 ¢ 7@
& & 0 8 8 s 8 * 8 9 @ ® 4 & 6 95 9 08 ¢ 08 . . o8 .9
- » . . 5 & 5 0 8 v . * 888 * . e e

I 0NN P

10
12
12
15
17
17
26
31
31
FABRICATION OF THE HEAT EXCHANGER TUBE BUNDLE
FOR THE MOLTEN-SALT REACTOR EXPERIMENT

 

R. G. Donnelly and G. M, Slaughter

ABSTRACT

The INOR-8 tube bundle of the primary heat exchanger of
the Molten-Salt Reactor Experiment contains 163 1/2-in. diam X
0.042-in.-wall U-tubes welded to a 1 1/2-in.-thick tube sheet,
17 in. in diameter. Procedures were successfully developed for
welding and back brazing the closely spaced 326 tube-to-tube
sheet Jjoints in the unit, and the actual tube bundle was
fabricated without incident. The welded and back-brazed design
provides a double seal between the fuel and coolant salts and
was used because of the necessity for high joint integrity.

Trepan grooves were machined in the weld side of the tube
sheet so that low-restraint edge-type welds could be made,
This technigue greatly reduces the cracking problems associated
with welding thin-walled tubes to thick tube sheets. Welding
and assembling procedures were developed which provided welds
of high quality with a minimum of roll-over.

The 82 Au—18 Ni (wt %) brazing alloy used in this
gpplication is corrosion resistant, ductile, and produces
joints exhibiting satisfactory mechanical strength. A
unique Joint design incorporating a trepan groove and feeder
holes was used for the braze side, the purpose being to pre-
vent preferential flow of alloy on the relatively thin-walled
tubes.

After the determination of exact welding and brazing
conditions on several subsize and full-size mockup samples,
the actual unit was constructed. Nondestructive inspection
of the welds revealed no defects. Good general flow of the
brazing alloy was evident, and ultrasonic examination of
brazed joints showed only minor scattered porosity. The com-
pleted unit passed both helium-leak and 800-psi hydrostatic
tests.
INTRODUCTION

The Molten~-Salt Reactor Experiment is fueled with a molten salt
consisting of IiF, BeF,, ZrF,, and UF,;. This fuel-bearing fluid is
pumped through the reactor, fuel-circulating pump, and the shell side
of the primary heat exchanger.l A nonfuel-~bearing coolant salt,
LiF-BeFy, clrculates through the tube side of the heat exchanger and an
air-cooled radiator.

The containment material is the commercially available alloy,
INOR-8 (Ni—17 Mo—7 Cr— Fe, wt %).® This alloy is a nonage-hardenable
high-strength material which possesses excellent corrosion resistance
to the molten salts and good oxidation resistance. It also exhibits
good general weldability.3

In view of the general difficulties associated with repair or
replacement of a radioactively contaminated heat exchanger, an extensive
program was conducted to develop fabrication procedures which would
ensure a very high degree of reliability. This report describes the
development of the combination welded and back-brazed tube-to-header
Joint used for this component and the specific details of the procedure
used in fabricating it.

The heat exchanger is of the conventional U-tube design with the
tubes being 0.5-in. OD X 0.042-in. wall. All tube ends are Joined to a
1 1/2-in.-thick INOR-§ tube sheet, 17 in, in diameter, as shown in
Fig. 1. Design data for the unit are presented in Table 1.

JOINT DESIGN

The conventional welded tube-to-header Joint has performed satis-

factorily in a very large percentage of heat exchanger applications at

 

IMSRP Quart. Progr. Rept. July 31, 1960, ORNL-3014, p 3.

 

“W. D. Manly, et al., Progress in Nuclear Energy, Series IV,
Vol. 2 — Technology, Engineering and Safety, pp 164—79, Pergamon Press,
London, 1960.

3G. M. Slaughter, P, Patriarca, and R. E. Clausing, Welding J.
gg(lO), 393s—+400s (1959).

 

 
UNCLASSIFIED
ORNL-LR-DWG 52036R2

FUEL INLET

   
   

1/2-in-0D TUBES

    
  
 
     
  

THERMAL-BARRIER PLATE CROSS BAFFLES
TUBE SHEET

COOLANT INLET

 

 

16.4-in, OD x O.2-in. WALL x B-ft LONG

 

 

 

 

CCOLANT-STREAM COOLANT QUTLET

SEPARATING BAFFLE
FUEL OUTLET

Fig. 1. Heat Exchanger for MSRE.

I
Table 1. Primary Heat Exchanger Design Data

 

Structural material
Heat load (Mw)
Shell-side fluid
Tube-side fluid
Layout

Baffle pitch (in.)
Tube pitch (in.)
Tube
Outside diameter (in.)
Wall thickness (in.)
Active shell length (ft)
Average tube length (ft)
Number of U-tubes
Shell diameter (in.)
Overall length (ft)
Tube-sheet thickness (in.)
Design temperature (°F)
Design pressure
Shell (psig)
Tube (psig)
Terminal temperatures
Fuel salt (°F)
Coolant salt (°F)
Exchanger geometry

Effective log mean temperature
difference (°F)

Active heat-transfer surface
area (ft?)

Fuel-salt holdup (ft°)
Pressure drop
Shell side (psig)
Tube side (psig)

INOR-8

10

Fuel salt
Coolant salt

25% Cut, cross-baffled shell
and U=-tube

12
0.775

0.5

0.042

6

Approximately 14
163

16
Approximately 8
11/2

1300

75
125

Inlet 1225; outlet 1175
Inlet 1025; outlet 1100
Parallel-counter flow
133

259

6.1

24
29

 
low and intermediate temperatures. However, as has been reported,4
joints of this type, although initially sound, are subject to cracking
during cyclic service at high temperatures. A performance testing
program in which heat exchanger components were subjected to very severe
steady-state and cyclic-temperature service was terminated prematurely
as a result of the failure of several tube-to-tube sheet joints. ©Since
no microfissures had been observed during the metallographic examination
of a large number of as-welded joints, it was concluded that the initia-
tion and propagation of the cracks occurred during thermal cycling. It
appeared that these cracks originated at the unavoidable notch at the
root of the weld.

One means for circumventing this problem is to back braze and
thereby eliminate the notch.” The location of the major stress is
removed from the weld and relocated to a more favorable area near the
braze fillet. Back brazing also provides supplementary functions in
that it reinforces welds containing undetected flaws.

The welded and back-brazed joint design chosen for this application
is shown schematically in Fig. 2. The weld Jjoint makes use of trepan
grooving in the tube sheet so that low-restraint edge-type welds can be
used. This, in effect, greatly reduces cracking problems associated
with welding thin-walled tubes to thick tube sheets.

The back-braze Jjoint detail is similar to that on the weld side,
except that it contains three feeder holes and a wider trepan groove
to accept the brazing alloy.6 The trepan desigh was used here to
eliminate the problem of preferential runoff of the brazing alloy onto
the thin-walled tube which tends to reach brazing temperature before
the heavy tube sheet. With this design, the alloy cannot melt or flow
until the tube sheet has reached brazing temperature; at this time, the
alloy flows down the feeder holes and along the Joint. Since the alloy

 

“P. Patriarca, G. M. Slaughter, and W. D. Manly, Welding J.
36(12), 1172-78 (Dec. 1957).

°G. M. Slaughter and P. Patriarca, Welding and Brazing of High-
Temperature Radiators and Heat Exchangers, ORNL-TM-147 (Feb. 20, 1962).

®P. Patriarca, C. E. Shubert, and G. M. Slaughter, "Method of
Making a Tube and Plate Conmection,"” U. S. Patent No. 3,078,551
(Feb. 26, 1963).

 
UNCLASSIFIED
ORNL-LR-DWG 65682R3

     
   
 
 

TUBE

BRAZING ALLOY RING

WELD SIDE// \TREPAN

{a) BEFORE WELDING AND BRAZING

/TREPAN GROOVE WITH

 

  

[ T i T TR

 

 
  

 

7.

Ly

 

.

Fig. 2. ©OSchematic Drawing of Welded and Back-Brazed Joint Design.

R

 

(6) AFTER WELDING AND BRAZING
does not flow out of the groove, the observation of a braze fillet
around the tube serves as a built-in inspection technigue since it
assures that the alloy has been in the joint and flowed up the joint
by capillarity. The trepan design also eliminates "swapping" of
brazing alloy from one Jjoint to ancther, thereby ensuring that each

joint has a sufficient amount of alloy.
PRELIMINARY INVESTIGATIONS
Welding

An experimental study was conducted to develop satisfactory con-
ditions for making high quality tube-to-tube sheet welds. The desired
features of the welds were: (1) a minimum of 1-T penetration
(T = tube wall thickness of 0.042 in.), (2) no surface imperfections,
(3) no porosity, and (4) no cracks. A minimum "roll-over"” requirement
was also instituted in order to allow passage of a probe for ultrasonic
inspection of the brazed joints.

Porosity could be readily eliminated by careful cleaning of the
tubes and tube sheet. They were scrubbed with both trichlorethylene
and acetone, and subsequently assembled with clean white gloves.

Root cracking was effectively eliminated by providing a very tight
fit-up between the tube and tube sheet. This was obtained by flaring
the joint on the weld side of the tube sheet to a maximum depth of
1/8 in. The procedure was as follows: (1) the tube sheet part of the
joint was flared to a 0.530-in. ID with a special punch and (2) the
tube end was extended through the tube sheet and flared to a
0.532~in. OD with a similar punch. With a 0.002-in. interference fit,
the tube end had to be tapped level with the tube sheet surface for
welding. This procedure ensured a very snug fit of the tube and tube
sheet at the weld area.

The welding parameters (i.e., current, travel speed, arc distance,
etc.) were adjusted to provide the desired 1-T minimum penetration
without excessive roll-over. A travel speed of 7.3 in./min was found
to provide consistently sound welds and allow satisfactory operator
control. Samples were made at various welding currents and the depth
of penetration and amount of roll-over noted. The welding conditions

finally selected for use on the MSRE heat exchanger are listed in Table 2.
Table 2. Welding Conditions for MSRE Tube-to~Tube Sheet Joints

 

Blectrcde material
Electrode diameter
Electrode taper

Electrode-to-work distance

Electrode position

Tungsten plus 2% thoria
3/32 in.
30-deg included angle

0.035 in. (determined by feeler
gage)
0.005-in. outside Jjoint interface

for full 360-deg rotation
(+ 0.002-in. concentric)

 

Inert gas Argon (99.995% purity)
Gas flow rate 12—-13 cfh
Welding amperage 390
Electrode travel speed 7.3 in./min
Weld overlap (full amperage) 3060 deg
Weld overlap (amperage taper) 120-150 deg
Welding position Flat
Brazing

Alloy Selection

 

The 82 Au-18 Ni (wt %) brazing alloy was chosen for this study

because of its generally good brazing characteristics on INOR-8 in dry

hydrogen, its satisfactory corrosion resistance to molten fluoride

salts,’ its good ductility, and its relatively low brazing temperature
(1830°F, which is below the recrystallization temperature of INOR-8).

In order to obtain strength information on this type of Joint,

Miller-Peaslee~type shear-test specimens,8 as shown in Figs. 3 and 4,

were furnace brazed and tested at room temperature and at the reactor

operating temperature of 1300°F.

in Table 3.

 

The results of these tests are listed

"E. E. Hoffman et al., An Fvaluation of the Corrosion and Oxidation

 

Resistance of High-Temperature Brazing Alloys, ORNL-1934, p 16

 

(Oct. 23, 1956).

8F, M. Miller and R. L. Peaslee, Welding J. 37(4), 144s—50s

(April 1958).
UNCLASSIFIED
ORNL-LR-DWG 78478

% in. 01A7 e—tin. —=

 

 

 

 

{in,

 

 

 

 

 

 

 

 

BRAZE

S5 i 3 41
6 in. R mln\ D32 14 in. R

N

J —0.125 in. BRAZED JOINT LENGTH

 

 

 

 

 

MACHINE

Fig, 3, Miller~-Peaslee-Type Chear-Test Specimen.
 

  

Testing Temperature

UNCLASSIFIED
Y-44199

1300°F

 

Fig. 4. Typical Miller-Peaslee-Type Shear-Test Specimens. Note the clean shear at
the brazed joint on the specimen tested at 1300°F and the elongated and rotated joint on

the specimen tested at room temperature.

In both cases, separation was in the brazed joint.
10

Table 3. Results of Miller-Peaslee Shear Strength Tests
on Brazed Joints?

 

 

 

Test Shear Strength (psi)P
Temperature Min Max Av

Room 54,400 67,000 59,000

1300°F 15,300 17,000 16,100

 

“INOR-8 specimens brazed with 82 Au—18 Ni (Wt %) at
1830°F for 10 min.

bFive specimens tested at each temperature.

In addition to the strength data, long-time diffusion studies
of INOR-8 lap joints brazed with the gold-nickel alloy have also been
conducted. Specimens were aged for 1000, 3000, 5000, 7000, and 10,000 hr
at both 1200 and 1500°F in air. Micrchardness traverses were made on
all specimens to determine the extent and effect of diffusion between
the braze and the base metal. Although evidence of diffusion can be
seen upon metallographic examination (Fig. 5), there was no detectable
hardness difference between the diffusion zone and the unaffected base
metal. Also, no detectable hardness change wags observed in the base
metal, the diffusion zone, or the braze after aging at either 1200 or
1500°F, Thus, aging at these temperatures for up to 10,000 hr would be
expected to have an insignificant effect on the joint strength and

overall base-metal properties.

Technique Development

 

In order to assure complete brazing of the tube-to-tube sheet joints

and filling of the feeder holes, it was deemed necessary to have an
excess of brazing alloy present at each joint. From Fig. 2 it can be
seen that there must be enough brazing alloy te (1) fill the Joint,
(2) £ill the three feeder holes, and (3) form a fillet at the junction
of the tube and tube sheet face., Thus, taking into account the appro-
priate maximum and minimum tolerance limits, maximum volumes were cal-
culated for the Jjoint, the feeder holes, and the fillet. The results

of these calculations are listed in Table 4.
*

UNCL ASSIFIED | / * 7. UNCLASSIFIED § AN
ey T TSR g e e

- UNCL ASSIFIED
S Y4059 002,

- e

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

As-Brazed Aged 10,000 hr at 1200°F Aged 10,000 hr at 1500°F

Fig. 5. Metallographic Studies of Aged INOR-8 Joints Brazed with 82 Au-18 Ni (wt %). Only a
slight change in the appearance of the diffusion zome is evident after aging at 1200°F. At 1500°F,
a definite boundary for the diffusion zone is not apparent and a general carbide precipitate is
visible in the INOR-&.

TT
12

Table 4. Summary of Volume Calculations

 

Maximum Fillet Volume 0.7 x 10=2 in.3
Maximum Joint Volume 7,1 x 1072 in. >
Maximum Feeder Hole Volume 4.2 X 1072 4in.?2

(total of three holes)

 

Maximum Total Volume to be Filled with Alloy 12.0 x 1072 in.?

Volume of One Split Alloy Ring 13.8 x 1072 in.?
(3/32-in. diam)

 

Wire preforms, 3/32 in. in diameter, were found to provide an
adequate supply of alloy, and the necessary size of trepan could be
machined around each joint without encroaching on & neighbor,
Consequently, one 3/32-in. split ring was positioned in each trepan
groove before agssembly with the tubes. The rings were positively
retained in the grooves by upsetting the edges of the tube sheet face
around the outside of the trepans with a special machinist's punch.

In order to determine the limits in which the brazing conditions
could be allowed to vary from the nominal and still create a satisfactory
braze, several small seven-tube assemblies of the type shown in Fig. 6
were assembled, welded, and back brazed. OSpecimens were heated to the
brazing temperature of 1835°F at three different rates of temperature
rise, 135, 270, and 405°F/hr. All specimens were well brazed with no
variation in quality. In addition, another specimen was brazed at
1785°F, This specimen revealed incomplete filleting and joint penetra-
tion; thus it was apparent that temperature gradients over the actual
unit should be controlled to ensure that the lowest temperature is

always above 1835°F.
FABRICATION OF MOCKUP SAMPLES
Small Sample

A proof test of all welding, brazing, and assembly conditicns and
techniques was carried out on a subsize 12-in.~-diam assembly containing
98 joints with some short U-tubes. The unit is shown in Fig. 7. The

header plate was drilled to the heat exchanger configuration to simulate
 

UNCL ASSIFIED
PHOTO 56641

 

Fig. 6. Seven-Tube Braze-Cycle Test Specimen.
14

UNCLASSIFIED
PHOTO 58284

4w WO Y4 ‘\"l\ " )

. uxwuwfiflfififiw\
| \"‘I \l\ \fl, |
* -h“-.'“",_ N BB E

 

Fig. 7. Proof-Test Assembly Containing Short U-Tubes.
15 .

actual heat flow conditions during brazing. As an economy measure,
the trepanning, drilling of the feeder hcles, and application of
brazing alloy was limited to only half the tubes. Dummy stainless
steel tubes were tack-welded in the remaining holes.

Inspection of the completed assembly revealed no weld defects and
excellent brazes in all Joints except one, this one Jjoint being
completely unbrazed. Metallographic examination of this joint revealed
small metal chips in the Jjoint and a complete lack of wetting by the
brazing alloy. The apparent cause of this condition was inadequate
cleaning of the parts before assembly. Therefore, the importance of
cleanliness to the brazing operation was greatly emphasized. OSpecial
handling and assembly procedures were instituted, and a full-time

inspector was assigned to the job.
Large Sample

It was, of course, essential that the fabrication of the actual
heat exchanger core be an unqualified success on the first attempt.
Sufficient material in the form of INOR-8 tube sheet forgings and tubes
was available for only one heat exchanger core, and replacement would
have meant about a one-year delay and a large financial expenditure.

In addition, the size of the unit made it necessary to ship the heat
exchanger to an outside vendor where it was to be brazed in a newly
fabricated retort. With these facts in mind and because of additional
uncertainties remaining (such as the effect of large mass variations
during brazing, retort integrity, assembly of long U-tubes, ete.), it
was decided to build a full-size sample heat exchanger core to gain
further assurance of all stages of fabrication. The sample, which is
shown completed in Fig. 8, contained nine full-length U-tubes and

54 welded and back-brazed joints. These joints were positioned in the
center of the tube sheet and at 3, 6, 9, and 12 o'clock positions
around the periphery. All additional tube holes were rough drilled and
a full complement of dummy stainless steel tubes was fitted in order to
simulate the expected gas and radiant heat baffling effects. For
economy, the tube sheet was machined from rolled plate instead of from

a forging as in the actual heat exchanger. Aside from these few points,
16

 

a
w&
L
ms
vy
o
<
Cn_._
QX
Za

 

 

Completed Sample Heat Exchanger Positioned in Welding-

Fig. 8.
Inspection Fixture.

 
17

the sample was machined, assembled, welded, shipped, brazed, and
inspected in exactly the manner planned for the actual unit.

Inspection of the completed sample revealed all welds to be free
from porosity and all brazes fully filleted. Ultrasonic and metallo-
graphic examination of the brazed joints revealed only minor scattered
porosity. A photomicrograph of one of these Joints is shown in Fig, 9.
The excellent weld contour, good weld penetration, and flow of the

brazing alloy to the root of the weld are evident.
FABRTICATION OF MSRE HEAT EXCHANGER TUBE BUNDLE
Welding and Brazing

Because of the success with the full-size sample, assembly of the
actual heat exchanger tube bundle was not delayed. The tubs sheet had
previously been machined and all tubes bent, degreased, and sealed in
polyethylene bags.

After a final degreasing and inspection of the tube sheet, all
brazing alloy rings were inserted and locked in place by upsetting the
face of the tube sheet around the edge of the trepans. The tube sheet
was then mounted in the assembly-welding fixture and the supporting
structure of rod and baffles installed, as shown in Fig. 10. Several
tubes with compound bends were also installed at this point since the
final bends had to be made after being threaded through all the baffles.
The U-tubes were assembled and welded one row at a time starting
adjacent to the diametral flow separator and working outward. The
welding conditions used were those presented previously. The setup for
welding is shown in Fig. 11. The torch assembly, specially made for
this application by the Union Carbide Nuclear Division, Y-12 General
Machine Shops, was provided with both coarse and micrometer adjustments
for centering over a joint. The torch was also adjustable in the
vertical direction. Circular movement was provided by a magnetic
plate cam and knurled-pin follower located above the torch next to

the micrometer adjustment.
18

UNCL ASSIFIED
Y-46953

all

%

Tube W

 

 

 

Fig. 9. Tube-To-Header Joint from Sample Heat Exchanger. Note
excellent weld contour, good weld penetration, and excellent flow of
brazing alloy to root of weld. 27X.
- e e

 

19

UNCLASSIFIED
PHOTO 39258

Fig. 10. Tube Sheet with Baffles, Support Rods and Compound Bent
Tubes Installed.
20

UNCLASSIFIED
PHOTO 39318

 

Fig. 11. ©Specially Built Torch Positioned for Welding.
21

Special precautions were taken to minimize contamination which
might later adversely affect the brazing operation. All through the
four weeks that were required for assembly, welding, and radiographic
inspection, the tube bundle was encased in a polyethylene bag and the
tube sheet surface covered with a polyethylene sheet when continuous
access was not required. Along with these precautions, the welder and
mechanics wore clean, white cotton gloves, and personnel access to the
rocm was limited.

After welding and radiographic inspection, the tube bundle was
assembled in the shipping container (Fig. 12) and sent to Wall Colmonoy
Corporation, Detroit, Michigan, for retort brazing in dry hydrogen.
Furnace runs were made on the empty retort to clean the imner surface
and obtain an oxlde-free enviromment for the actual brazing operation.
The tube bundle was then removed from the shipping container and
positioned in the furnace pit on the retort base, as shown in Fig. 13,
with the protective bag still intact. After leveling and positioning
the various thermocouples, the bag was removed and the retort top
lowered into place (Fig. 14) and seal-welded at the edge-weld preparation.
The retort was then leak checked, vacuum purged, and filled with dry
(=80 to -85°F) hydrogen at a rate of 145 cfh. After purging with
hydrogen for 1/2 hr, the furnace top was set in place (Fig. 15) and the
brazing cycle started.

The rate of temperature rise was approximately 300°F/hr up to an
equalization temperature of 1650°F where it was held until all thermo-
couples were within *25°F, The initial part of the cycle was charac-
terized by large thermal gradients between the center and edge of the
tube sheet (approximately 150°F) and between the heavy tube sheet and
the top of the U-tubes (approximately 450°F) despite the fact that heat
was applied in the pit section of the furnace only. As the cycle
progressed, regulation of the upper and lower zones of the gas-fired
furnace reduced these gradients to about 50°F approaching the 1650°F
hold. After 20 min at 1650°F, the temperature was increased to the
brazing range of 1850—1885°F and held for 1 hr. All thermocouples on

the tube sheet end of the bundle registered within this range during
Fig.

12"

22

UNCL ASSIFIED
PHOTO 38936

 

 

Shipping Container for Tube Bundles.
23

J UNCLASSIFIED
J PHOTO_ 60220

R \.4

 

Fig. 13. Tube Bundle Positioned in Furnace Pit on Retort Base
with Protective Bag Still in Place,
24

4 UNCLASSIFIED
PHOTO 60221

!

 

Fig. 1l4. Retort Top in Place for Seal Welding.
Gas-Fired Furnace

25

UNCL ASSIFIED
PHOTO 60269

 

 

 

Top in Place for Brazing Cycle.
26

the brazing cycle. A slow furnace cool at about 300°F/hr was main-
tained to below 300°F where an exothermic gas purge was introduced.
The retort was removed from the pit, the seal weld cut, and the bundle

visually inspected.

Inspection

A photograph of several braze fillets 1s presented in Fig. 16,
these being typical of all visible joints. No distortion, discoloration,
lack of filleting, or other adverse conditions were noted. The bundle
was agalin bagged, assembled in the shipping container, and returned to
Oak Ridge for further inspection, tests, and assembly into its shell.

Final inspection included a general dimensional check on the over-
all bundle (no detectable change) and dye-penetrant inspection of all
tube-to-tube sheet welds. No flaws were revealed. The brazed Joints
were then inspected by a newly developed ultrasonic technique.9

With the tubes filled with water, the ultrasonic probe, which
contains a sending and receiving crystal mounted at a preset angle to
each other, was rotated 360 deg around the inside of the tube and
indexed down the tube in steps to fully cover the brazed area. The
device, set up for inspecting the sample heat exchanger, i1s shown in
Fig. 17. This inspection revealed somewhat more porosity in the brazed
Joints of the actual heat exchanger than was observed in the large
sample. Nevertheless, judging from signals from standardized defects,
it was believed that this porosity was scattered and of small size
and that the back brazes should certainly be an effective secondary
seal in case of inadvertent weld failure in service., TFigure 18 shows
the completed tube bundle mounted on the fixture.

After welding the tube bundle into its shell, helium leak and
800-psi hydrostatic tests were conducted. No leaks were found. The
completed heat exchanger is shown in Fig. 19 ready for installation in

the MSR system.

 

°K. V. Cook and R, W. McClung, Welding J. 41(9), 404s—08s
(Sept. 1962). T
 

 

Fig. 16.
Vigible Joints.

27

J] UNCLASSIFIED
PHOTO 39430

{

o,

Braze Fillet Area of Heat Exchanger Typical of All
28

            
    
    

i

    

3 @-
" .@@@@@ B

1000000000000
F4 2000000000000 \ 1
[F000000000000000 6,
(/5000000000000 000 0
LT el 0 000 9 ¢

..1 ..
g

UNCLASSIFIED
PHOTO 59492

   

      

 

Ultrasonic Probe Device Set up on Sample Heat Exchanger

To Inspect Tube-To-Tube Sheet-Brazed Joints.

Pig, 17
 

Fig. 18.

29

UNCLASSIFIED
PHOTO 39431

 

Completed Tube Bundle Mounted in Fixture.
 

UNCLASSIFIED
PHOTO 39597

 

Fig. 19. Heat Exchanger Ready for Installation into the MSR System. -

 

 
‘31

SUMMARY

A novel welded and back-brazed joint design was used on the tube-
to-tube sheet in this heat exchanger because of the necessity for long-
term relisbility. The double sealing would give more confidence in these
tube-to-~tube sheet joints, which on many heat exchangers are the areas
most prone to failure.

The development of both welding and brazing procedures was
necessary in order to determine the optimum fabricating techniques. In
this line, sample components were welded and brazed under various con-
ditions, then inspected and tested. The strength of the brazed joints
and the effect of long-time aging on brazed joints were also investigated.

The results of these investigations culminated in the fabrication
of a heat exchanger tube bundle which required no repairs on either
welds or brazes on any of the 326 tube-to-tube sheet joints and gives
every indication of being able to fulfill its purpose in the reactor

system.
ACKNOWLEDGMENT

The authors gratefully acknowledge the work of L. G. Bryson in
the preparation of test assemblies and assistance in fabricating the
heat exchanger, the planning and coordinating efforts of A. Taboada,
W. B. McDonald, and C. K. McGlothlan, and the engineering advice of
C. H. Wodtke. Many thanks are due the Y-12 Shop personnel for their
heartening cooperation and personal interest throughout both the
developmental and construction stages of this project. Thanks are also
extended to the Nondestructive Testing Development and Metallography
Groups of the Metals and Ceramics Division for their invaluable
assistance. The advice of the Wall Colmonoy technical staff and the

patience of the shop personnel were both greatly appreciated.
7-26.

30-49.
50.
51.
52.
53.
54,

55-59.

33

ORNL-3500

UC-25 — Metals, Ceramics, and Materials

INTERNAL DISTRIBUTTON

. Central Research Library

Reactor Division Library
ORNL — Y-12 Technical Library
Document Reference Section
Laboratory Records Department

R.
C.
Je
W.
R.
H.
M.

V.
E.
Ge
We
H

R.
Ge
L.
R.

60. C. E.
61l. H. G.

. Laboratory Records, ORNL R.C.
* Kl
. Je

Cook
Cunningham
Donnelly
Fox

Frye, Jr.
Gall
Gilliland
Hemphill
Hill
Larson
MacPherson

EXTERNAL DISTRIBUTION

78. C. M. Adams, Jr., MIT
David F. Cope, ORO

D. E. Baker, GE, Hanford
Ersel Evans, GE, Hanford

79-80.
g1.

82.

g83.

84.

85.

86.

87.
88-594.

J. L. Gregg, Cornell University

J. Simmons, AEC, Washington
E. E. Stansbury, University of Tennessee
Donald K. Stevens, AEC, Washington
Research and Development Division, AEC, CRO

Given distribution as shown in TID-4500 (23rd ed.) under Metals,
Ceramics, and Materials category

62
63.
64 .
65.
66.
67,
68.
69.
70.
71.
2.
73.
T
75,
76
77

W.
W.
R.
Se
Je
M.
G.
Je
Je

A.
R.
A.
J e
C.
R.

0

TID-4500 {23rd ed.)

+ Manly

« Martin

« McClung

. Rabin

. Redford

« Skinner

. blaughter
Swartout

. Tallackson

. Thurber

. Welinberg

. Worsham

Burr (consultant)
Johnson (consultant)
Smith (consultant)

moluchowski (consultant)