ORNL-TM-2953.txt

  
  
 

 

O\ __;_;ZEOAK ""IDGE NATIONAL LABORATORY
R operated by s e
""‘;:V.'UNION CARBlDE CORPORATION . NUCLEAR DIVISION s
AU fo,. Ihe e | .

U_S.;ATOMIC ENERGY COMMISSION |
ORNL m 2953

  
    

- RECEWED BY oTiE S asimh T @

 

A NEW APPROACH TO THE DESIGN OF STEAM GENERATORS L
i FOR MOLTEN SALT REACTOR POWER PLANTS _; - )

arres

  

 

 

 

 

- Nfllltf Ti‘us document 'contmns information: of @ prehmmury nafure_-}-— L
‘and ‘was ‘prepared pmncnly for internal use at the Oak Ridge Nationol - = © 7 .
Laboratory. lt-is -subject to revision or correction ond therefore does S

 

- .not rePresent a final report.” - .

 

- ISTRBUTON OF THS DOCUNET IS URLATED

   

 
 

 

 

 

This report. was prepared .as an account of work sponsored by.the United | - . -~ .. o
States Government. Neither the United States nor the United States Atomic | -~~~ 7 = &~
Energy Commission, nor any _df their employees, nor any of their contractors, B : C '
subcontractors, or their employees, makes any warranty, express or implied, or | ) - o . gfl
assumes any legal liability or responsibility for the accuracy, completeness or | =~~~ = = - S .
usefulness of any information, -apparatus, product or process disclosed, or | = - ' : RN '
represents that its use would not infringe privately owned rights, '

 

 

 

 

)

 
 

 

ks 4 ‘f'

4)

ORNL-TM-2953

Contract No. W-TL05-eng-26

A NEW APPROACH TO THE DESIGN OF STEAM GENERATORS
FOR MOLTEN SALT REACTOR POWER PLANTS

A. P. Fraas

 

 

© | This report was prepared as an account of work
: .} sponsored by the United States Government. Neither
i the United States nor the United States Atomic Energy
"1 {1 Commission, nor any of their employees, nor any of
I their contractors, subcontractors, or their employees,
¢ { makes any warranty, express or implied, or assumes any
legal liability or responsibility for the accuracy, com-
‘pleteness or usefulness of any information, apparatus,
product or process disclosed, or represents that its use
.} would not infringe privately owned rights, - SN

 

 

 

~ JUNE 1971

OAK RIDGE NATIONAL LABORATORY
Oak'Ridge, Tennessee
Operated by
UNION CARBIDE COPERATION
- for the
U. S ATOMIC ENERGY COMMISSION

DISTRIBUTION OF THIS DOCUMENT IS UNLI
 

.

¢

<)
 

 

«j)

ABSTRACT ...... tisecessieocrsnrnonrosuy e Ceeeaieiecsiacnaras v
INTRODUCTION e eeesmenennanannnenns . ...
REVIEW OF BOILER DESIGNS PROPOSED FOR MOLTEN SALT REACTOR ..... ‘o
SYSTEMS
Conventional Shell-and- Tube Heat Exchangers e eneneeann eeneoe
Supercritical Pressure Unlts Ceececccsssaneannas ;...;........
Double-Walled Tubes with & Heat DA o.le.eueneen.. celiesanes
Flash Boiler .....;..,....g....,........} ........ e
Loeffler Boiler ....... esesienenns Cesseaanaa fi....,....; ...... ;
Triple Tube Boiler ...ccevess cieseans chedsectrsrence vecassseas
Reentry Tube Boiler ............ e enene s et eeraeens
' DESIGN FOR GOOD STAEBILITY AND CONTROL CHARACTERISTICS . .ovvewen...
Molten Salt Temperatures at Part Load . Ceetasseseneneea ces
Effects of Mode of Control on Steam Temperatures at eresssens
Part Load :
Heat Tragnsfer Instdbilities and Thermal Stresses ............
Flow Stability'Considérations Ceetieisietisetatrae s e
Heat Transfer Analysis ,...; ................. Leetersaceasanna
- Typical Calculatiofis e PR [
Heat Transfer CORFFficients .........e.ee... e
Temperature Differences'and Heat FLUXES +.oovvveennnnnns
| Pressure DroD v.veeeeevessoeneeeeens e ieeaaeeeens
SEstlmated.Perfonmance”Characteriétics eereeientaaresaeaanne
Effects of DeSign'Heat‘Load on Tube Length ....... PRI
Effects of De31gn Heat Load on Temperature .......ceoe..
Dlstrlbutlon , _
Effects of Part Load Operation ........ .......;.;..;..;;
Pressure Drop and ‘Pressure Distribution .,.; ........ e
- Effects of the Fraction of Heat Added in the e ean e
 Inner Tube 7
Effects of Slze Of Heat Increment Used .o.eeveveeeeennns _
Turbine Control Considerations .;..........,....;;.; .........

ILimitations on Rates of Change in Ivad .......ccv0eveeee
 

 

 

Startup and Rates of Change of

System

Possibility of Eliminating the Throttle Valve ......... .
Proposed Design for a Molten Salt Reactor Plant ........ ceoae
' Reheateré cereecratasiaeaas s ......;..;..........;...
General Description .......... e e ereenanaaaas ;.,.
Headering Problems ....;..Q....,..................,....;
Differential Therfiél Expansion {..:....' ...... ceseeene ves
‘Geometric and Performance Data .......... e eeeneenetees

_ Cosf Estimate- .............. ; ............................
Conclusibns .................................................
Récomméndgtions ........................................... .o

REFERENCES ..¢veeveeenncccnaasca cietsescensesassanus .

»

iv

Load with Proposed ......

Page

51

53

54
55 °
55

56
-

61
65

65
67 B

e

&)
 

 

o) @’ - C _

)

)

A NEW APPROACH TO THE DESIGN OF STEAM GENERATORS
- FOR MOLTEN SALT REACTOR POWER PLANTS

A P, Fraas

ABSTRACT |

A new type of steam generator has been dev1sed to meet
'the special requirements of high-temperature liquid metal and
- molten salt reactor systems, The basic design concept is such
that bolling heat transfer instabilities and their attendant
severe thermal stresses are avolded even for a temperature -
difference of as much as 1000°F between the feedwater and the
high-temperature liquid, thus giving good control characteris-
tics even under startup conditions. This is accomplished by
'employing a vertical reentry tube geometry with the feedwater
entering the bottom of the inner small diameter tube (~1/4 in.
diam) through which it flows upward until evaporated to dry-
‘ness. The slightly superheated steam emerging from the top
of the small central tube then flows back downward through the
annulus between the central tube and the outer tube. A por-
tion of the heat transferred from the high-temperature liquid
- to the superheated steam in the annulus is in turn transferred
- to the water boiling in the central tube. Design studies in-
‘dicate that this type of boiler not only avoids thermal stress
and salt freezing problems but it also gives a relatively com-
pact and inexpensive construction. Further, it appears to make
possible a simple plant control system with exceptlonally good
plant response to changes in load demand

 

" INTRODUCTION

It has been'apparenthinceVearly in the'molten salt reactor develop-
ment work that the high melting point of the fluoride salt sultable as

fuel for molten salt reactOrS”coupled with the thermal stress problems

"inherent in high—temperature liquid systems pose  some exceedingly diffi-
'“cult problems in the design of steam generators.! These are compounded -
by the complexities of two—phase flow and heat transfer problems under
'rboiling conditions, possible difficulties with b01ling flow instabilities,

and_thevproblems of obtaining good boiler operating characteristics for
 

 

a wide range of both full power design steam temperatures and pressures
and under the much reduced'pressare‘and temperature conditions inherent
in startup and part load operation, Many different attempts to design
boilers for molten salt reactor systems hare been made,! ® but each of
the approaches proposed has had some serious disadvantages. The startup
and part load control problems in particular have been so formidable, in
fact, that3nofattempt:hashbeec made tQVSO1Ve‘tfiem for many of the'designs
that have been proposed‘—-Oniy full load'desdgn conditions-hare been con-
sidered. It is belleved that the new. reentry tube concept proposed in
this report w1ll yleld compact, economical b01lers ‘that can be designed
| for any full power de51gn steam condltlons and yet w1ll glve good stability
and control characteristlcs over: the full range from zero power to full
power“condltlons,and further Wlll notppresent dlfflcult.thermal stress or
salt freezing :problems.', - - | .

A draft of this report substantially as it now stands was prepared
and distributed July 15, 1968 to'key people. ifi the molten salt reactor
project at ORNL' . They klndly rev1ewed the draft and suggested a number
of addltlons to help clarlfy thls new approach By far the most serious
reservatlons they had were concerned with boillng flow stabillty at low
loads. Thls requlred SOme sort of a test, and the report was held pend-
ing the availability of funds for a projected test rig. Because of lack
of funds this rig has hot yet been built. | ' _

It happened that basically similer requirements for a steam generator
arose last year in a program to develop a small isotope'power unit. These
requirements made it importaotlto test a low steam pressure, short tube
version of the reentry boiler. The results of the tests are now in, and
good performance was obtained.” Even though these results cover a limited
range and the steam generator proportions are substantially'different from
- those proposed here for a large central station, the basic concept appears
-to be validated for the low load portion of the operating.regime that was
most open to question. As a consequence, it was decided that the-report
should be issued.

Q.
 

<) g}/

REVIEW OF BOILER DESIGNS PROPOSED FOR
~ MOLTEN SALT REACTOR SYSTEMS

' Thesvarious boilers that have been considered for use with molten
salt reactors might-be.grouped in the categories outlined in Table 1.

The principal features and major advanteges and disadvantages of each

‘are summarized very briefly to help point up the problems and provide &

 framework for the subseqnent analysis of thetheatftransfer, thermal

stress, and controlrproblems.
Conventional Shell-and-Tube Heat Exchangers

A logical first candidate is a shell-and-tube heat exchanger with

the salt inside the tubes and the water boiling outside of the tubes.

Two_configurations-are'included in Table 1. - The hcrizontal'U—shaped tube
and casing'geometry_employedifor_the early pressurized water reactors

(see Ref. 8, p. 197) has the advantage that not only is differential ther-
mal expansion between the tubes end the casing readily accommodated'With-
out producing serious thermal stresses, but the difference between the hot

fluid'inlet-and outlet temperatures will not induce severe stresses in the

 header sheet as would be the case for @ simple shell-and-tube heat ex-
changer with a'U—tube'configuration; _The'majorvdifficulty with this con-
- ventional Ueshaped'casing_design'iS'tnat the temperature'difference be-

tween the water and the walls of the tubes carrying the molten salt is

. far greater than the temperature difference in the nucleate boilihg regime

so that an unstable vapor blanket would form”between the liquid and the

~hot metal surfaCe~(See*Figeil) This leads'to unstable, noisy operation

and severe thermal stresses 1n the tubes as a consequence of violent ir-

;regularitles in the heat transfer coefficient

. Supercritical Pressure Units

It has beenisuggestedsthat these difficulties might-be reduced
through the use of a supercriticel water system inasmuch as this would

reduce the temperature difference between the feedwater and the molten
 

 

4

Table 1. Types of Btesu Generstor that Have fleen Proposed for Use with Molten Belt Reactors

 

 

Type of Beat , L "' Shell-Side Tube-Side
Iten Exchanger . Gecnetry Flud Fluid Major Problems

1 Conventionsl Shell.
and-tuba, U-tube

Water Salt  Excessive temperature difference be-
tveen salt and steam gives unstable
boiling and/or possible freezing of salt.
Tempersture difference between salt in-
“let and outlet streams causes large
thermal ltrenel in hender sheet lnd

: casing.

Salt Excessive tem;perlture difference be-

_ tween salt and steam gives unstable
boiling apd/or possible freezing of
salt,

 

Water

   

2 Blull-md-tube vith U-tubes in
& [~shaped cssing for luhcritical
pressure steam

3 Shell-and-tube with
U-tubes in a U-
shaped casing for
supercritical pressure

H,0  Sinilar to above but problems less
severe at full load, but still serious
at part load and in startup; large

ateam preheater required to add 20% of
heat as preheat,
A Dpouble-walled . 8imilar to Item £ but with tubes fubri- . T Severe thermal stresses in porous
tubes vith a cated as in the detail shown below: Outer tube metal region would cause cracking
heat dan

. and indeterminately urse thermal
barrier.

 

 

5 Flash Boiler large number of tube-to-header joints
required. Not well suited to steanm

pressures above sabout 600 psi. Tends
to give large heat flux and local ealt

‘freeging near spray nozzle end.

6 loeffler Boller large stean drum required coupled with
heat exchanger and steam pump mekes
the equipment expensive. Operation

is inherently extremely noisy and
excites vibration.

 

 

‘? Triple Tube Boiler No suitable vapor separator is avail-
o able, and experience indicates that it
is unlikely one can be developed.
Outer tube diameter is inherently
large and thus requires & thick wall;
this léads to low heat flux and large
- walght and investment in tube material.

-

Recirculating

 

Water
Insulation T f Final
¥ B Buperheat
Tube Sheet Bection

 

 

 

 

 

 

 

 

Peedwater
8 Reentry Tube Boiler i ' - _ : Salt |  Steam Concept hes not been tested at .. -
- . = ) pressures above 200 peia.

. O

§)

$5
 

 

 

 

   

                
 

 

 

 

 

 

 

   

  

 

 

_ - ORNL DWG. T1-5T2h
~<—— Interface evaporation ———>t<——— Bubbles >te<——— Film
< : 1' — >t J] —>of—— III'—->--<—1'V——>-<- V >4=<— V] —
K 8 ~|E . §sc| £ |2
1 L £EE fi£§ = B8EE| 2 S: -
w | Pure convection ~ heat "?fig’ 885/ H2 \ =8 &€ |==
| transferred by superheated wd=|825/ S|E EEgl = |E8
5| liquid rising to the. £3¢(=° BE® \VWSF 8 |gs2
£  liquid-vapor interface 2= 5 e =
2| where evaporation gEg> - -
| takes place | SES o
oo S = 8.2 o
5 B —~——
 poling W — o
0110 10 100 — 100

T,-T,, °F —>

: Fig. 1. Diagram Showing the Principal Pool-Boiling Regimes and Their -
Relative Position on a Curve for the Heat-Transfer Coefficient Plotted as a
- Function of the Film Temperature Drop (Farber and Scorah, Heat Transfer to
Water Boiling Under Pressure, Trans. ASME, p. 369, Vol. 70, 1948)" |
 

 

salt, reduce the fraction of heat added at low water temperatures, and
reduce the sharp changes in physical properties associated‘fiith_the phase
change from liguid to vapor.2°® In units of this type it is usually con-
sidered best to have the molten salt outside the tubes and the super-
critical watér and steam flowing inside the tubes. However, while there
is no‘sharp phase change, the very rapid changes in density and othér
physicalAproperties near the critical temperature under supercritical
pressfire conditi@ns lead to marked changes ifi the heat.transfer perform-
-~ ance and abrupt reductions in the heat transfer coefficient something like
those associated with the burnout heat flux encountered at subcritical
pressures.? In addition to these flow and heat transfer phenomena and
the boiling flow instabilities that would be associated with them, there
would be large, irregularly'fluCtuating thermal stresseé in the tube wall
~in the region near the feedwater inlet of thé.boiler unit, and these would
be.likely to cause tube cracking and failure. These problems are dis-
cussed later in some detail. |

- Although the use of a supercritical pressure system reduces the
severity of the boiling flow stability pfoblem.for operation near the de-
sign point, a great deal of difficulty has been encountered in ali of the
coal-fired once-through and supercritical pressure boilers in going from
zero power to part load conditions of at least 10%, and often to as high
as 30% power. These difficulties stem in part from the'large_density |
change as the water-steam mixture. flows through the boilers and in part
-from the reduced pressure drop at the lower flows which reduces the damp-

ing of the oscillations by'turbulencé losses.®

‘Double-Walled Tubes with a Heat Dam

‘M. E. Lackey suggested in 1958 that one means for.reducing the tem-
perature difference between the tube wall ahd the water to avoid film
boiling conditions wouid be to incorporate alheat dam in the form of a
sintered powder matrix between an inner tube carrying the molten salt and
the outer tube in contact with the boiling water.* This approach would

be quite effective at full'power where the average heat flux would be

<)
 

 

 

o)

x)

high, but wonld:presentfproblems_under startnpand low power conditions
-becausera'high heet flux is inherently associated with & large tempera-
ture drop through such a heat dam..iA,further disadvantage is that dif-
ferential thermal_expansion between the inner and outer tube walls would

~induce severe-thermal stresses in the sintered buffer material, and these

would lead to cracking of the sintered matrixland'unpredictable increases
in the thermal res1stance.87 , _
‘The cracking problem 1n the s1ntered materlal could be avolded hy

using 1nstead an air space asra heat dem. The OD of the inner tube could

~ be knurled, for example,-and the outer tube swaged down onto it. This

would have the dlsadvantage that dlfferentlal thermal expans1on between
the inner and outer tubeS‘would probably loosen the swaged joint and give
an 1ndeterm1nately large -probably excessively large — thermal barrier
at full load. If this did not happen, thermal stresses would probably.

| cause cracklng of the tubes.

‘Flash Boiler

In an effort tOVavcid"the difficulties outlined above, a flash boiler
was prOPOSed'in 1955, 1In a ‘unit of this type the molten salt would flow
outside of the tubes and feedwater would be 1n3ected in the form of a long
thin plume of,flne spray dlreoted‘along the axis of the boiler tube.

'.Experience in the development of noZZleS for diesel'engines indicates that
@ high penetration spray could be obteined with sharp-edged, single-
“orifice nozzles_having;a'hole.diameter of about 0.020 in.,° and that

no2z1es of this type would'break'the liquid:up into_droplets having ‘a

. diameter of the order of 0. 005'in 11 These droplets would form a long
- 8lender spray plume that’ would extend for perhaps 2 £t down the bore of
a 1/2 in. tube. Droplets would impinge on the tube wall at.a very low
'fangle of 1nc1dence, and’ would tend to skltter along the wall riding on
__”a'thin film of Vapor. Analyses indicated that the local thermal stresses
 associated with the cold. traoks left by droplets of this sort would be
- well within the elastlc llmlt and should not glve dlfflcultles with

thermal stresses.

*
 

A brief series of tests to investigate this concept was run by an
MIT practice school group.}2 These tests showed that there was a strong
tendehcy for a large fraction of the droplets to impinge on the tube wall

in the region close to the 1n3ectlon nozzle, and that this led to such a

' pronounced cooling effect that a frozen film of molten salt tended to form

on the outside of the tube in that reglon.la“‘The ‘tests had been initiated
on the premise that they offered an attractive way to provide for emer -
wgency cooling of the ART fuel dump tanks, hence the test work Was termi-
nated with the demise of the ANP proéram.” No further work on the concept
was carried out because it inherently requires a very large number of tubes
of rather short length so that the tube-to-header joint costs tend to be-
come excess1ve.' Further, the concept does not appear to lend 1tself well

‘to hlgh pressure and supercrltlcal pressure steam systems.

Loeffler Boiler

The Loeffler boiler concept used in @ few coal-fired steam plants =

has been considered.® Systems of this sort have been built and operated
for coal- fired furnaces. Operatlon apparently has been satlsfactory ex-
| cept for the extremely high nozse level assoc1ated w1th the b01ler and
dlfflcultles with the steam pumps required. The Loeffler concept entalls
admission of saturated steam to & heat transfer matrlx heated by molten
salt. The superheated steam leav1ng thls matrlx would be d1v1ded 1nto
two portions, one of whlch would flow to the turbine and the other would
be returned to a boiler drum where it would bubble through the water in
~the drum. ThlS approach has the dlsadvantages that 1t requlres large
and expensive boiler drums, 1mplos1on of the vapor bubbles in the boiler
drum makes operation extremely n01sy and 1nduces v1olent v1brat10n ex-
citing forces, the relatlvely poor steam-s1de heat transfer coefflclent
.and low enthalpy rise lead to a large number of tube-to-header JOlntS,
and steam pumps posing tough design and rellablllty problems are required
to recirculate steam through the boiler. On the other hand, the system
has the advantage that it lends itself readlly to startup and part load

operations, it wvirtually eliminates the possibility of salt freezing &s
 

 

w

)

a consequence of excessivescooling in the steam generator, and it greatly

eases the thermal stress problems by substituting a salt-to-steam heat ex-

changer for the salt-to-water boiler.

Triple Tube Boiler

 

_ When the writer solicited criticisms and-suggestions on‘the proposed
new boiler concept, M. E. Lackey pointed out that B. Kinyon and G. D,

_Whitman had proposed & somewhat similar boiler in 1960, (Ref. 6), and

S. E. Beall pointed out that a variation of this approach had been tested

as a means for cooling fuel dump tanks."_l3 The arrangement proposed by
B Kinyon employed three concentric tubes with boiling water flowing upward
;_through the inner annulus to a vapor separator and superheated steam flow-

ing down through the outer annulus The central passage would serve to

return the water from the vapor separator to a boiler water recirculating

| pump. This arrangement has the advantage that the superheated vapor in
the annulus between the outer tube heated by the molten salt and the tube
. containing the boilingpwater would act as a buffer both'to eliminate ser-

ious thermal stresses and to avoid excessive'metal temperatures adjacent

to the boiling water. The arrangement has the disadvantage that it re-

quires a fairly large tube diameter and hence & fairly large tube wall
thickness for supercritical water systems. Thus, the amount of heat trans-
fer surface area required tends to be large because the principal barrier
to ‘heat transfer lies in heat conduction through thick tube walls. Fur-

_'ther, the arrangement requires the development of a vapor separator that

- would fit within a small diameter, preferably that of the tube. Experience

in vapor separator development indicates that the velocities required for
good vapor-liquid separation are much lower than. those one would like to

- use in the tube for heat transfer purposes, and hence a rather bulky pro-
tuberance would have to be:employed at the end of each tube; this appears

%o lead to a set of extremely avkward mechanical design problems.
 

10
Reentry Tube Boiler

The boiler proposed in this report is somewhat similar to the one
described above.but differs in that it makes use of only two vertical,
concentric tubes in the form shown in Fig. 2. The water enters at the
‘bottom through a central tube having a dlameter of about 1/4 in. Pre-
'heatlng and b01llng occur as the water rises 1n thls tube untll evapora-
tion is complete, after which there is some superheatlng. The steam
emerges from the top end of the small dlameter tube, reverses direction,
and flows back downward through an annulus between the 1nner small tube
and an outer tube having an ID of ‘around 1/2 in. The molten salt enters
~at the bottom, flows upward around the outer tube, and out the top With
this arrangement there is only one header sheet separatlng the molten salt
from the atmosphere, and this header sheet is not subaect to a large pres-
sure differential. Thermal sleeves would be used at the header sheet to
minimize thermal stresses (see Flg. 3). There'would“bevno"high pressure
header sheets in this system, the tubes for the hlgh pressure feedwater
and the exit steam would be manifolded as in hlgh pressure coal fired
boilers rather than run into header sheets. The_steam annulus between
the inner and outer tubes would'act'as.a'buffer to isolate the relatively
low- -temperature boiling water region from the high-temperature molten salt.

This isolation would be so effectlve ‘that 1t would be qulte poss1ble to
| heat the unit to the molten salt operatlng temperature w1th no water in
the system and then slowly add water to 1n1t1ate boiling. CAs will be
shown later, it should be possible to design the unit so that it would
operate stably'over a wide range of condltlons from zero load to overload
with no difficulties from thermal stresses or freez1ng of the salt.

It might at first appear that the extra heat transfer films through
Wthh heat must be transmltted from the molten salt to the boiling water
might lead to a large increase in surface area requirements and hence in
the size, weight, and cost of the unit. However, it appears quite pos-
sible to design so that these disadvantages are more than offset by such
features as the absence of a high-pressure header sheet and the ability
to operate with high-temperature differences between the molten salt and

the boiling water so that the overall size, weight, and cost of the unit
2

 

 

 

11

- . ORNL DWG. T1-5T25

 

)

Steam

 

salt

 

 

 

 

Water

 

 

Fig. 2. Diagram,Showing Section-Thfohgh_?Ube: ]

0
 

12

5

  
   

 

 

 

 

 

 

\O 6 x
o O m m o
0 B .
mw . n“mw, . _ , g o pumw
A mwg, g mw mm 9 & m
2 S 2 B 2
o
{ A T T T T T T T N T T T T L T T T U T T T TT
y _ N
~ 3
WHENHIRRAR RN RNV NS .
ettt A AN\
,//Jafaafffdflflfffaidaflffaflflfflffl/l#/%f}?ffldlfllwfltfl/rr¢d1¢~

 

 

<

 

 

 MOLTEN SALT
AIR

-
,-“fi\:

Section Through the Header Sheet Region Showing Two

Typical Tubes with Their Thermal Sleeves and the Associated Welds.

Fig. 3.
 

 

 

 

 

 

»

|

)

15

are at least competitive with the corresponding values for eny other de-

"sign that has been proposed.

DESIGN FOR GOOD STABILITY AND CONTROL CHARACTERISTICS

The usual proCedure in'deVeloping a design for a steamrgenerator has‘
been to choose a geometry, establish the proportions for full povwer condi-
tions, and then — sometlmes -examlne the full range of control problems.
The inverse order seems at least equally logical and is followed here. The
writer. has_felt from the beglnnlng that some of the most difficult condi-

‘tions to be met are those associated with initial stertup and part load op-

eration. rThus, the first step'in.the evaluation work was to establish a

- typical set of molten saltiand steam‘temperaturesg and from these, using

bas1c heat balance cons1deratlons, deduce the ‘effects of dlfferent modes
of control for the various load conditions of 1nterest This approach gives

a'valuable insight into the full range of the over—all design problems.

‘Molten Salt Temperatures at Part Load

Several different approaches_can'be taken to the control of a molten
salt reactor steam poWer'plant. Perhaps the simplest'and most reliable ap-
proach is to make use of constant speed ac motors to drive pumps in both
the fuel c1rcu1t and the intermedlate salt CerUlt If this is done, the

temperature rise in each qalt clrcult w1ll be. dlrectly proportlonal to the

-"load so that the c1rcu1ts w1ll be 1sothermal at zero powver. The basic heat

transfer relatlons are such that the temperature dlfference between the two

- salt 01rcu1ts w1ll also be dlrectly proportlonal to the load and w1ll drop

frto zero at zero load. If there is no control rod movement in the reactor,

the zero load reactor temperature Wlll be the mean of the 1nlet and outlet

fuel temperatures at full load. - These effects are shown in Flg. La for op-

(*eratlon with constant: speed fuel and inert salt pumps ~ Note that both the .

mlnlmum and the mean temperatures of the 1nert salt rise as the load is
reduced — an undes1rable characterlstlc from the standp01nt of the design

of most steam generators., - Thls 51tuatlon can be changed by holding the

 
 

14

ORNL DWG. T1-5T27

1300

1200

1100

1000

1300

Temperature (°F)

1200

1000

 

0 20 ko 60 8o . 100
aefQ (%) |

Fig. 4. Temperature Distribution in the Fuel-to-Inert Salt Heat
Exchanger for & Series of Loads for Two Control Modes, i.e., &) constant
epeed pumps and a constant mean fuel temperature, &nd b) constant speed
purps &nd & constant reactor fuel inlet temperature.

‘-‘m
 

 

 

0 .

’)

 

15

reactor inlet temperature constant in which case the temperatures 'in the
fuel and inert salt circuits would be defined by the curves in Fig. 4b.
The fuel inlet temperature’ ought not be reduced below the value shown be-
cause'it is desirable to_maintain'the'fuel at least'100°F above its freez-
ing point. Other effects:could be obtained by varying the speeds of the
fuel and/or inert salt pumps,.but the resulting complications — particu-
larly at or near zero 1oad — are quite objectionable.

After analyzing a variety ‘of power plant control modes it was decided
that the simplest system would be the most reliable and should be used for

- the bulk of this study., The approach chosen is believed to be the simplest

possible, i.e., it assumes a constant mean fuel temperature and constant

speed pumps for both the fuel and the inert salt irrespective of load.

" Effects  of Mode of Control on Steam
Temperatures at Part Load

The inherent_effects'of typical control modes on the temperature dis-

'tributions“that will resultnas aiconsequence of fundamental'heat balance

considerations are.shown\insFigsi 5, 6, and 7 for the full range of load

conditiOns It can be seen'frOm“examination of these curves that the steam

'outlet temperature will rise as the load is reduced because the temperature

difference between the two fluid streams in a heat exchanger drops off with

. & reduction in the heat flux. This problem arises in the control of any

steam plant that is coupled to a high temperature reactor whose mean tem-

' perature is held constant 14 po avoid damage to the turbine, the steam
'-temperature could be reduced at’ part load by introduction of a desuper-

heater between the steam generatcr and the turbine The -available desuper-
heater units may not be: well suited to this particular application, but

‘the design of a suitable unit would be straightforward However, it is
' apparent that high steam temperatures under part load conditions would ser-

iously increase the creep stress problem in the steam system if it were

l-to operate with a constant b01ler discharge pressure.‘ The problem could

 be eased by scheduling the reactor mean temperature 80 that it would in-

crease with power output. One way of doing this would be to maintain the

reactor inlet temperature-constant and allow the temperature rise in the

 
 

16

ORNL DWG. T1-5728

TEMPERATURE ' (°F)

 

500 &
0

20 60 "8 100
AQ/Q (%) o

Fig. 5. Effects of Opergtion with a Constant Mean Fuel Tempera.tm'e
on the Temperature Distribution Through the Boiler for Typicel Loading
Conditions, Iocal temperatures are plotted as functions of the fraction
of the heat transferred to the water from the molten salt (i.e., 22/Q).
For this set of curves it was assumed that the effective boiler tube

- length would be varied with the load to maintein & constant temperature
and pressure at the superheater outlet.
 

 

x

"

%)

17

- ORNL DWG. 71-5_729

Temperature (°F)

60

 

. /e m

Fig. 6Q Effects of Operation.with & Constant. Reactor Fuel Inlet

- Temperature on the Temperature Distribution Through the Boiler for Typi-

cal Load Conditions. Local fluid temperatures are plotted as functions
of the fraction of the heat transferred from the molten salt (i.e.,

- M/Q). For this set of curves it was assumed that the effective boiler
~ tube length would be varied with the load at any given condition to main-

tain a constant temperature and pressure at the superheater outlet.
 

18

o

ORNL DWG. T1l-5T30 Q
1200 ' :

‘"

1100

1000

700

TEMPERATURE (%F)

300

 

200, 20 50 - 8 100 , ' :
AQ/Q .

Fig, 7. Inert Salt and Steam Temperatures as a Functicn of the | '
Fraction of the Heat Transferred from the Salt to the Steam for e Series

of Design Heat Loads with the Steam Pressure Directly Proportiomsl to - o~

the Eeat lLoad, : o : | . U
 

 

n

i)

19

_fuel salt to increase in direct proportion to the load. If this were done,

the temperatures in the inert salt and steam c1rcu1ts would vary with load

as indicated in Fig. 4b, This control'schedule for the salt circuits was
assumed in preparing a few .of the boiler performance estimates presented

later in this report to show that this arrangement might reduce or elimin-

~ ‘ate the need for desuperheating the steam under part load conditlons at

the expense of camplicating the reactor control prdblems
The curves of Figs. 5 and 6 were calculeted on the basis that the

boiler would be operated in'the conventional fashion with a.constant steam

discharge pressure of 4000 psi, and the pressure of the steam supplied to

~ the turbine would be reduced by a throttling valve.  However, there are a

number of advantages aSsociated with operating a steam boiler-turbine-
condenser-feed pump systEm.with no'throttle valve between the boiler and
the turbine so that the steam pressure is determined by the flow rate
through the critieal pressure drop orifice represented by the inlet nozzle
box of the first stage of the turbine. One of the more important of these
advantages in this‘instanee:is_that;'if the steam system were designed so
that the boiler discherge pressure weuld be directly proportional to the

load, the higher boiler temperatures would be associated with lower pres-

 sures, and the creep stresses in the boiler tube wall would not be exces-

sive. If this were done, to a first approx1mat10n the boiler pressure will

be directly proportlonal to the load, and curves for the steam temperature

- as a function of the emount of heat added on the water-81de will be as in-
‘dicated in Fig. 7'for_a”typical.ease. .th‘surprisingly, calcuwlations pre-
-sented later in the report show that‘the_tube-length to evaporate to dry-

ness is much the same'irrespective of the boiler pressure at a given load.

—However, the pressure drop through the boiler under part load conditions
fis much higher 1f the pressure is directly proportional to the load than

if the boiler were Operated at a constant pressure. Increasing the pres—'
sure drop at part load is edvantageous in that it will much 1mprove the
boiling flow stability. |

Heat Transferiinstabilities and Thermal'Stresses_

One can sense intuitively that severe thermal stresses might be induced

by the wide variations in the heat flux —-and hence the transfer coefficient
 

 

20

that can occur with changes in the difference in temperature between the- k&#)
metal wall and the saturation‘temperature of a boiling liquid (see Fig. 1). -
HoweVer, it is not obvious just how these thermal stresses may be related

to fluctuations in the boiling heat transfer coefficient, how large they'
may be, or why they may be more severe-at‘low loads than at the design
point, and M. Rosenthal asked that this short- sectlon be .added to clerify
the problem, particularly for a unit de31gned for supercrltlcal operation.
Tt should be noted that for some time it was thought that this problem

could be avoided by going to supercritical water pressures, but severe
cracking of tubes in coal-fired supercritical boilers showed that, un-
fortunatély, this is not the case.? Detailed investigations of-boiling

heat transfer relations in the supercritical pressure regime have shown

that large variations in heat transfer coefficient still occur, particularly
at high heat fluxes, i.e., if there is a large température difference be-
tween the metal wall and the bulk -free stream? (sge,Fig. 8). |

To illustrate the problem, consider a short section of INOR-8 tubing
with supercritical pressure water at 690°F inside and molten salt at 1150°F -
‘flowing outside the tube with a selt heat transfer coefficient of 1000
Btu/hr-ft? -°F. The thermal conductivity of the wall is about 12 Btu/hr.-ft-°F.
The thickness is 0.10 in., and hence the conductance of the wall would be
about 1440 Btu/hr-ft®-°F, end thus the temperature drop through the tube
wall would be about 70% that through the salt film on the outer wall. Two
very different operating regimes are possible. - Aésuming that the curves of
Fig. 8 define the heat transfer situation on the water side, the heat flux
-where the water-steam enthalpy ran 780 Btu/lb could be ~100,000 Btu/hr or
8t & nearby point downstream where the enthalpy reached 900 Btu/lb it could
- be 150,000 Btu/hr The resulting film and wall temperatures and tempera-
ture drops can be summarized in Table 2.

Changing the radial AT through the tube wall from.70°F to 105°F at
power, and to O°F at no load would lead to dlfferentlal thermal expansion
between the inner and outer surfaces and hence to both. c1rcumferent1al and
axial stresses that would be superimposed on the basic pressure stresses.
Power cycling and changes from one heat transfer regime to the other would
cause thermal strain cycling, and this could eventually.lead to cracking *

and failure. The problem would be much‘worse at subcritical pressures ./
 

 

¥)

1)

21
AR | ORNL DWG. T1-2175
1200 T T T 1 |
nrs |- ~ STEAM AT 3300 PSI> -

 

nso |- 6 = 340000 LBS/FT2-HR —
| - | D= 0033 FT
S - —— EXPERIMENTAL
noo |- |  CURVES (SHITSMAN)
- —— — MACADAM'S
CORRELATION (BULK
PROPERTIES)

 

  
 
  

—

1075 |-
1050 |-
1025
1000
975
950
925
900
875

850

 

 

WALL TEMPERATURE, °F

825

 

800
775

   

 

 

__ g i [
600 700 800 900 1000 1100 1200
' BULK ENTHALPY, BTO /L8

 

Fig 8. Deterioration in Heat Transfer Near the Critical Tempera-

eeture in Supercritical Pressure Once-Through Steam Generator Tubes Operat-
‘ing at High Heat Fluxes. '(M‘~E Shitsman, Impairment of Heat TTansmissiOn
'at Supercritical Pressures, Teplofizika Vysokikh Temperatur, Vol. l, No. 3‘

p. 267, 1963)
 

oo

Table 2. Effects of Heat Flux on the Radial Temperature
. Distribution Through an Element of Tube Wall in the
Inlet Region of a Simple Shell-and-Tube Molten Salt
Steam Generator at Supercritical Pressure Conditions

 

Water-steam enthalpy, Btu/lb ' 785 " 900
Heat flux, Btu/hr-ft? | 100,000 150,000
Salt free steam temperature, °F 1,150 . 1,150
Tube_Ou#er wall temperature, °F - 1,050 1;000
Tube inner wall temperature, °F 980 895
Water temperature, °F ' - 690 705
Salt film AT, °F | | 100 150
Wall AT, °F 70 105
Water film AT, °F | 20 190

 

where the change in heat transfer coefficient with enthalpy is both larger
and more abrupt (see Fig. 9). The problem could be eased somewhat by re-
ducing the molten salttemperature to 1000°F for startup and low power
conditions, but it would not be eliminated.

Observations of tubes transferring heat:to water at supercritical
pressures have shown that both of the regimes of Table 2 will be present
i. e., some sections of the tube will operate at a lower wall temperature
and a high heat flux while others will operate at 8 hlgher temperature and
a lower heat flux. Further, these regimes tend to shift back and forth
axially along the tube with changes in water flow rate. This 1eads-to
another type of thermal stress. Dilation of the hot regiou relative to
the cooler region leads to bending stresses in the tube wall, end these
stresses are likely to be severe because the shift from one heat transfer
regime to the other tends to be abrupt and the transition zone is short
.These stresses are analagous to those in thermal sleeves, and can be cal-
culated in the same way (see Ref. 8, p. 125). Unfortunately, data for
typical exial temperature distributions are not at hand'to provide a good
_basis for estimating the'magnitude of these stresses. o ’_

At first thought it would appear that at light loads the above prob-
lems would be eased because the average heat flux would be greetly reduced.
 

-

 

 

23

ORNL DWG. T1-5731

 

 

      

  
  

 

 

 

 

 

 

Nucleate - Convective ' ~ Convective
{ boiling . boiling -t fim N
Boiling =~ Liquid  Poling  steam
A starts - ~ deficiency |
High heat_ flux ~
'-'_ «—Burn-out limit
[ Medium heat flux
£ P T T
2 |
£ [
8 P l
o ' | |
2 | |
5 a 3 l
g Il |
3 N k |
T 1 |
o | |
| s \
o=} : Burn-out N
<-Subcooled water—>- -<———Steam-water—-————>- -s—Superheat—»
4 . Enthalpy . T L
- Saturated - - . Saturated
water e | ~steam

sttance from mlet —

-Fig 9, Effects of- Heat Flux on the Heat-Transfer Coefficient for

.,__-a.WOnce-Through Boiler Tube (Polomik et al., Heat Transfer Coefficients
~with Annular Flow During "Once-Through' Boiling of Water to 100% Quality
at 800, 1100, and 1400 psi, Jour of Heat Transfer, Trans ASME, p. 81,

vol. 86-2, 1964)
 

24

However,rthis is not the case. The bulk of the heat transfer oéCurs at
the water inlet end because the temperature difference there is inherently
high irrespective of load, and it is the high—témperature'difference that
gives the possibility of two drastically different temperaturé regimes.
Thus réducing the heat load simply reducés the length of the regiofi in
which severe thermal stresses may be induced — it does not eliminate the
problem, In fact, if the pressufefis reduced, the boiling point’bf the
water will drop and the témpéréfiure difference that can-be_induced in the
tube wall will be increased. | -

The above effects are COmplex‘and, in many respects, rathgr subtle,
but they are the reason for turning from the conventional shell-and-tube
heatxeXChanger geometry to the reentry tube construction proposed in this
report. - | '

Flow Stability Considerations

In a conventional coal-fired bbilér the pressure drop in'the boiling
region is rather low in recirculating boilers, but the overall pressure
drop is fairly high because the préssure drop through the superheafer is
substantial. In once-through boilers, particfilarly-in supercritigal pres-
sure steam plants, the bo;ler pressure drop is large, commonly 20% of the
boiler inlet pressure. This stems in part from efforts to get a high heat
r_transfer coefficient on the steam side, in part from the long tubes made
necessary by the relatively low average heat transfer coefficient on the
combustion gas side, and in part by stringent ofificing at fhe tubé inlets
to assure a flow distribution across the tube bank such that it will be
possible to avoid burnout in regions where the local heat flux may be high
as a consequencé of irregularities in the hot gas flow on the'cbmbustion‘
gas side o£ the boiler. These irregulafities in the local hot gas tempera-
'tUre and heat flux vary substantially with the heat load onlthe boiler,
the fuel used, and the peéuliarities and irregularities in the gas turbu-
-lence pattern in the combustion zone. ‘Fbrtunately, in a molten salt-heated
- boiler not only can the tube wall never exceed the temperature of the
molten salt so that severe ofierheating of the tube wall is not a yroblem,

but the molten salt temperature and flow distribution can be predicted

O
 

 

 

 

0

t

25

within quite.close\limitslinstead of being subject to the vagaries in the
large-scale turbulence that are characteristic of combustion zones in fur-
naces. As a consequenCe,'a steam.generator for a molten salt reactor can

be designed to give a much. hlgher average heat flux and yet a lower peak

“heat flux than can ‘be obtalned in a conventional coal-fired boiler. This
directly reduces the tube length and‘hence the pressure_drop. In addition,

it reduces the.need for*orificing to control the water flow distribution

through the boiler. As a consequence, it is believed that it will be pos-

;s1ble to de51gn for lower water pressure drops through steam generators for

‘molten salt reactors than are,ordinarily required for coal-fired boilers.

This will reduce both the power requirements for the boiler feed pump and
the tube wall thickness required for the feedwater piping.

Heat Transfer Analysis

Estimating the heat transfer performance of’the reentry tube steam

‘generator involves a set of implicit relations that make an explicit solu-

tion out of the question and even an iterative solution surpr131ngly tricky.

This stems from the wide varlety of combinations of condltlons that may

occur in the boiler depending on the steam pressure, temperature, and flow
'{'rate. The problems have much in common with, but are more ‘difficult than,

- those of steam generatOrs for hlgh -temperature gas-cooled reactors. 14 The

steam conditions were chosen to be essentially the same_as ‘those of Eddy-
stone Unit No. 2, which. was . used as the basis for an earlier study 15 fThe

feedvater temperature. and flow rate, the exit steam pressure, and the mol-

- ten salt inlet and outlet temperatures are ordlnarlly given. From these

'data it is pos51ble to estlmate a steam outlet. temperature ‘and from heat

balance considerations draw a set of curves such as those of Fig. 5 which

,show the temperatures of the molten s&lt and the steam as functlons of the
' fraction of heat addedrtoithersteam in the course of its transit through

the boiler. Figure 7*sh¢ws a similar set of curves'for'a series of lower

fpressures and lower loads, 1n thls 1nstance the pressure was taken as di-
o ,rectly proportional to the load a good approx1mat10n to the natural charac=-
teristics of 8 steam b01ler—turb1ne condenser-feed pump systemrln which no

- throttle valve is employed. Figures 5 and 7 illustrate one of the
 

- 26

difficulties in setting up an iterative calculation. Whereas:the.steam
témperature for the supercriticél condition of Fig. 5 is uniquely defined
as a function of the fraction of heat added to the steam, this is not the .
case for suberitical Steam_pressure conditions where the steam temperature
is essentially independent of the amount of heat addgd over a wide range
~of heat addition. In attempting an iterativersolution one can celculate
stepwise upward from the bottom of the tube using as his points of depar-
ture the given feedwater inlet conditions and the assumed steam outlet con-
ditions. The stepwise calculétions_can be continued to a point at the top
of the inner tube where evaporation would be completed or the steam super-
- heated somewhat. - It is also possible to assume & set of steam conditions
at the outlet of the inner tube and. make stepwise calculations frcm the
top downward to the feedwater inlet end accepting the superheatér outlet
temperature that results. The difference in the character of the relations
between the various load conditions of Figs. 5 and 7 lead to some conver-
gence problems in either case. These, in turn, make it'necessary to modify
the calculational procedure somewhat depending on the steam conditions.
Both methods have been used in this study, and both have been found to be
not only awkward but demanding in-that they.require good engineering Judg-
“ment to choose values that will'yieid convergence. However, no better ap-
proach has been found in spite of samé months of effort by both MIT gradu-
ate students who became interested in the problem and by the éufhors of
Ref. 7.

Typical Calculations

The steps followed in estimating the boiler tube. length, temperature

'_ distribution, and pressure drop on the water side_are summarized in Table'B,
and a set of typical calculations is shown in Table 4. The first set of
calculations was made for the 100% load condition. The first step in the
calculations was to staft at the bottom, or salt inlet end, and assume &

- decrement in the enthalpy of the salt. This was chosen to be 10% or the

' expectéd value for the boiler. For a given tube geometry the sfirface areas
of the inner and outer tube walls per unit of length are defined and

~hence the area per increment of length is readily calculated. For good
 

 

W

27 .

" Table 3. Calculational Procedure for Establlshlng the Tube
Length for a Given Set of De31gn Conditions
(see nomenclature in latter portion of table)

 

12.

Plot curves for the salt and steam temperatures as fUnctions of AQ/Q
from the inlet to the outlet (e. g., Fig. 7).

Specify the tube dlameters and Wall thicknesses.

Compute the overall heat transfer coefflclents for the inner and
outer tube walls (e.g., use Fig. 11).

For convergence in the iteratlon, the heat transferred from the salt
to the annulus steam near the superheater outlet must be greater than
the heat transferred from the annulus steam to the water in the cen-
ter tube per unit of length, i.e., ViAjAt)/I>UsA At,/L. This will
probably require an inner tube liner to provide a heat dam near the
bottom. To mlnlmlze the overall tube length, this liner should be
terminated as soon as this can be done and stlll maintain

Uy Ay 486, / L>U2A LIAD) / L.

Estimate the mean temperature of the salt, annulus steam, and boiler
water for the first increment of tube length using the ratio of the
heat added to the steam in the outer annulus to the total heat removed

- from. the salt and the curves of item 1 above (e g., Fig. 7)

Using the above as the starting point, follow the calculational pro-
cedure of Table 4 using constant increments in AQl, the enthalpy

change in the molten salt
Compute AL, that is -

_ AQl
UIA]_At]_ ; i
Compute AQs for the value of Al found in item 7 (AQa = UgAgALfltg/L)

Repeat steps 6, 7, and 8 for a better approx1mat10n if the changes in
mean temperatures ylelded by these steps dlffer by more than 20% from
the values estlmated in step 5.

Repeat the above for new 1ncrements of length using the same 1ncre-

ment in NJ; until EAQl Ql..'

If there is trouble with convergence, change the value of Us by chang-~

ing the length or effectiveness of the heat dam near the feedwater in-

let, It may also be advisable to change the proportions of the tubes.

For part load calcalations,rbe careful to assume & small temperature
- difference between the salt and the superheated steam for the first

increment in tube length if the temperature distribution along the
full length is desiréd, If the superheater exit temperature exceeds
by more than 100°F the- value assumed for the part load condition, new
curves similar to those of Fig. 7 should be prepared, If instead one
wants a rough 1ndicatlon of the active tube length at part load, com-
pute a case as if it were a design point and compare the resulting
length with that for the 100% load condition.
 

28

Table 3. Nomenclature

 

Outer tube effective heat transfer surface area, ft2
Inner tube effective heat transfer surface area, ft2

'ID of outer tube, in.

ID of inner tube, in.

Friction factor for salt

Friction factor for steam in superheater annulus

Friction factor for steam in inner tube

Mass flow rate of salt, 1b/ft2- sec _
Mass Tflow rate of steam in superheater annulus, 1b/ft2-sec
Mass flow rate of steam in inner tube, 1b/ft?-sec

Tube length (or distance from bottom of tube), £t
Increment in tube length, ft

Number of increment

 Steam pressure, psia

Steam pressure drop in superheater annulus in increment, psi
Steam pressure drop in inner tube in increment, psi

Heat removed from molten salt, Btu/hr- tube

Net heat added to steam in superheater annulus, Btu/hr tube

- Net heat added to steam in inner tube, Btu/hr- tube

Heat removed from salt in increment, Btu/hr
Heat added to steam in superheater annulus in increment, Btu/hr
Heat added to steam in inner tube in increment, Btu/hr

. Dynamic head in superheater annulus, psi

Dynamic head in central boiler tube, psi
Reynolds number for salt

- Reynolds number for superheated steam in annulus
" Reynolds number for steam in inner tube

Local temperature of molten salt, °F

‘Local temperature of superheated steam in annulus, °F

Local temperature of steam in inner tube, °F

Local temperature difference between salt and steam in super-
‘heater annulus, °F

Toocal temperature difference between steam 1n annulus and steam
in inner tube, °F .
Specific volune of steam in superheater annulus, t3/lb
Specific volume of water-steam mixture in inner tube, £t3/1b

 

O
 

 

 

 

 

; !l‘afile‘ 4.'._:Typ1cal Calenlations] Worksheet for a Single Boiler-Superheater Tube

 

Steam pressure, psia = 4000 ' Fraction of reference design load, % = 100 Outer tube OD, in.

= .65. Total tube length, ft = 32.97
Steam temperature leaving inner tube, °F = 745 Superheater temperature out, °F = 1045 Outer tube ID, in, = .50 Pressure drop through inner tube psi = 38,3
8alt temperature in, °F = 1200 : Feedvater temperature in, °F = 560 Inner tube 0D, in. = .25 = Pressure drop through ennulus, pei = 14.4
Salt temperature out, °F = 950 Water enthalpy rise, Btu/lb = 886 Inner tube ID, in. = .23 Total preesure drop through boiler-superheater, psi = 52.7
Heat load per tube, Btu/hr =Qy = 650, Water flow per tube, lb/sec = 0,204 .
(@) Fractional change in salt enthalpy in increment  my/a .10 .10 .10 .10 .10 .10 .10 .10 .10 .10
- (@) Effective surface area of outer tube, £t3/ft nL JA52 0 .152 .15 L1520 L1522 .15 152 152 .152 .152
(3 Effective surtace area of inner tube, 2/t AL _ .052  .052  .0s2  .0s2  .052  .0s2  .052  ,052  .052  .0B2
' O Oversl hea.t transfer coefficledt for outer tube, T 667.6  667.6  667.6 667.6 667.6 667.6 667.6 667.6 667.6  667.6
: @ Werall heat transrer coei'ficient for 1nner tu?be, o N s 300 - 300 300 1145.5 1145."5- 1145.5  1145.5 1145.5 - 1145.5 1145.5
-'r - ‘ - - o R SRR RE o o SR
Heat transfer thraugh outer tu'be per | un:lt of length per 13U1A1/L 0. 1.4 10L.4. 1014 10L4  100L.4 - 101.4 1014 1014  10L.4  10L4
| °F, Btufir. R _ _ o _ o » L : - i '
(D Heat transfer throuh imer tube per unit of length per Uh/L S 157 157 157 59.6 - 59.6  59.6  59.6 5.6  59.6  59.6
°F, Btu/br.ft.°F . : . : . S _ .
_ : _ _ : o
. (8) Molten salt _temperatm-e, °F 7_(mean for 1ncrement) - ty Mg 7 1190 . 1164 1139 2 1087 1062 1038 1012 987 962 O
(® stean temperature in outer tube, °F (mean for increment) t;  see (D) 1030 1016 958 907 872 80 87 M2 T 752
@ Steen; tanpei-atm in 1nher tube, *P (mean for increment) ¢, see @ .585 575 605 630 660 692 708 720 728 72§
@) Temperature drop from salt to steam in snmulus, *F AT -0 160 148 181 205 215 212 n 220 217 210
@) Temperature drop from stesn in sunulus to steam in imer AT, (3) - @ 445 - 4AL 353 2 212 158 119 72 42 23 :
. tube, °F ‘ T S -
@ 1nerement in lemgtn, ft o & &/ (OO 4,006 4331  3.542 3.127 2,982 3.024 3.038 2.914 2,95 . 3,052
@ Distance from vottom, £t \ ‘ L.y 4,006 2337 11.879 15,006 17.988 21.012 24.050 26,964 29,918 32,970
(@ Heat ndded to steam in inner tube, Btuwhr & OO 27,991 29,988 19,628 51,623 37,672 28,474 21,547 12,503 7,395 4,184
@9 Heat added to steam 1n outer tube, Btufnr oM M - 37,009 35,002 45,372 13,377 27,328 36,526 43,453 52,497 57,605 60,816
@ mtio of incremental entbalpy change fn superheater to  Ma/%. @ /& .0%9  .0539 0698 .0206 .0420 0562 .0668 .0808  .0886  .0936
total for stesm s
Ratio of enthalpy change in superheater to total for . ), —1-A22/& .05%9  .1108 1806 = .2012  .2432° ,2994  .3662  .4470  .5356 6292
steam )
©® matio of incremental enthalpy change 1n boller to total M © /a 0431  .0461  .0302 .0794  .0580 .0438  .0331 .0192 .0ll4  .0064
for asteam
Ratio of enthalPy change in boiler to total for steam ), A/Q 0431 . .0892  .1194 .1988  .2568  .3006  .3337  .3529  .3643  .3707
0P00000000 90000 ® ®

Mean va),lue of 2A/Q in snnulus for interval (read ta in
Mg, 7 ‘

Mean V31ue of 24/Q in boller for interval (read t3 in
Pig. 7

Ratio of enthalpy change in salt to total
Fluld specific volume ;.n-uperhea.ter, £t? /10
Fluid specific volume in boiler, £t*/1b
Mass flow rate in superheater, 1b/sec. £t?
Mase flow rate in boiler, 1b/sec. £t?

Dynamiec head in superheater, psi

Dynemic head in boiler, psi N

(a =.25)

(a = .20 ana ,23)

R@lds number in superheater
Reynolds mumber in boiler
Prictlon factor in superheater
Friction factor in boiler
Incremental pressure drop in superheater, psi
Inc;etuenta.l pressure arop in boiler, psi

Pressure drop from tube inlet, psi (cuter annulus)

"Pressure drop from tube inlet, psi (inner tube)

Table 4. (Continued)

1-@a- @
2
Q.4
2
n=n -
ey SR/
v @t; in Fig. 13
vy €ty in Pig. 13
64 . 204/, 00002
64 . 204/, 000236
or ,204/.00029

2 @) @26

s @ @) 2ezme
Re,  Ref, 8, p. 291
Res  Ref. 8, p. 291
£2 Ref. 8, p. 29
2y Ref. 8, p. 294
= @D O

2 0@ O

P2 Pap + 22::1; AP
nooITEe

972

.022

.10
.185

. 0202
200
864

.798
1.611
179,867

238, 503

.0168

2.685
5.131
52.709
5.131

.916

. 20
.180

. 0200
200
864

772
1.660
180,871
243,578
. 0168
. 0158

2,809

5.680
50.024
10.811

+854

’lm

.30
170
.0208
200
864

.690
1,684
185,802
250,232
.0167
L0157
2,041
4.682
47,215
15,493

 

809 .78 729
159 .28 .279
.40 .50 .60
247 35 a2
.0215  .0230  .0260
200 200 200
703 703 703
587 522 A4
1162  1.226  1.439
191,010 192,715 192,715
243,482 253,812 294,922
L0166 .0166  .0166
L0157 L0157 .0152
1.524  1.292  1.190
2,852 2.870  3.307
45.174  43.650 42.358
18.345 21.215  24.522

.667  .%593
217,343
0 .80
116,100
0290 0318
200 200
703 703
.48 328
1.450 1,503
191,858 170,400
299,136 314,880
L0166 .0172
L0151 0150
1.054  .B22
3.326 3,285
41168 40.124
27.848  31.133

 

« 509

.35

» 207
1.57

148,855
317,265

.0178
0150
« 544

3,494
39,292
34,627

.368

1.00

. 0350
200
703

.160
1,599
115,628
317,265
.0189

. 0150
.461
3.660
38,748
38.287

 

 

ot

 
 

 

 

31

compatibility with both the molten salt and the water, the tubes were con-

sidered to be of Hastelloy N. This is a high nickel alloy that has a rela-

tively poor thermal-conductivity, an important factor in a supercritical

~ boiler because of the fairly thlck wall reqUired in the outer tube — in

- this 1nstance 0.075 in., It was found advantageous to increase the thermal
‘resistance of the inner tube wall.at the lower end by making use of a
rdouble-walled'tube. Indthis_regionuthe 1/4-in. OD central,tube with a wall
" thickness of 0.010 in._was_assumed to be lined with a{smaller'diameter tube
,“ ofrthe same.wall thickneSs-with a radial clearance of 0. 002 in. between
',them The gap would be vented and, in view of the local metal tempera-
‘tures, would be filled with steam rather than water,

The reason for using the tube liner can be seen by env1s1oning the
heat transfer 51tuation at the bottom end of the reentry tube.  Unless the

~heat transfer rate from the salt to the superheated steam per unit of

length exceeds that from_thersuperheated_steam to the feedwater, the.tem—

perature of the superheatedrsteam willractually drop as it flows toward

the outlet. Inasmuch as the local temperatures of the salt' the super-

I

heated steam,'and the feedwater are fixed, and so, too, are the surface

:_areas and the surface heat transfer coefficients, a simple solution is to

interject some extra. thermal re51stance between the feedwater and the

superheated steam. This region need not be very long because the steam

cexit temperature is relatively insensitive to the steam temperature well

above the exit. FUrther,_to keep the overall tube length down it is de-

“51rable to maintain a large temperature difference between the salt and
'.the superheated steam Thus one . of the places where Judgment is required

when making the stepwise calculations is in ch0031ng the point at which to

'terminate the heat dam formed by the 11ner of the 1nner tube

It should be noted that introducing the above liner in the inner

p:tube w1ll act to 1mprove the b0111ng flow stability characteristics of
rrthe reentry tube b01ler. th only will it increase the pressure drop in
iithe inlet region but 1t will also increase the fluid velocity there, both
"thelpful faotors., e |
 

32
Heat Transfer Coefficients

'Heat transfer cOefficients for the molten salt were computed'using
the Dittus Boelter relation and physical propertles supplied by J. W Cooke.
From these the chart shown in Fig. 10 was prepared In computing the over-
all heat transfer coeff1c1ents for the 1nner and outer tube walls, the
molten salt velocity was taken as constant for all of the conditions con-
51dered and hence the heat transfer coefficient for the fluid film.on the
salt sxde was constant The temperature drop through the tube walls would
- of course, be directly proportional to the heat load, and hence these two
-represent constant conductances. The heat transfer coefficient for the
superheated steam flow in the annulus varied directly with the 0.8 power
of the ‘steam flow rate and hence this was the dominant factor at low flow
rates. The heat transfer coefflcient for water under nucleate or annular
film'boiling conditions is almost independent of heat 1oad and is very J
.{high hence it had a relatively small effect on variations in the overa]l_ |
heat transfer coefficient with load.
~ The calculated values for the overall heat transfer coefficients
through the inner and outer walls are presented in Fig. 11 as a function
of the steam flow rate. The principal resistance 'to heat transfer at the
lower steam flow rates is that in the surface films in the superheated
g»steam annulus, whereas at high loads the principal res1stance 1s repre-
'sented by the temperature drops through the tube walls '
An 1nherent error in the calculation procedure stems from the use of
a constant heat transfer coefficient on the water side throughout the
length of the boiler. Fbr the flow rates employed here, annular film
boiling would prevail through the greater part of the boiler until the
vapor quality reached about 90% after which the heat transfer coefficient
would drop rapidly. However, this effect is small at or near supercritical '
pressures, and the heat transfer coefficient in the first portion of the
superheater is about as high as in the "boiling" zonel® (see Fig. 12).
This in effect compensates for the reduced hesat transfer coefficient in
the last portion of the boiler. These variations have relatively small
effects in the region of interest here and were neglected for the purposes

of these calculations.

O
 

Heat-Transfer Coefficient, h, Btu/hr:ft3.°F

 

. Temperature, °F

ORNL DWG. 71-5732

equivaient diameter, 0.1
0.2
0.5
1.0
2.0

2 34 5 010 15 20 30 4 50 . 100 150 200 300 400 500
- " Flow Rate, G', lb/sec.ft? - -

' Fig. iO; Heat-Transfer Coefficients for NaEF,; at 1350°F Under Turbulent Flow Conditions.
(Values for physical properties used in June 1968: cp = 0,37 Btu/1b. °F, p = 15 lb/hr--ft,
k = 0,232 Btu/hr. £t. °F, density = 131 1b/£t2.)

 

1000

ee

 
 

 

34

OENL ING. T1-5733

_Overall Heat Transfer Coeffictents (Btu/hv.ft2.%F)

103 16" 10° | 108
Stean Heat Joad (Btu/hr.tube) :

Fig. 11. Effects of Heat Load on the Overall Heat Transfer Coefficients for the Inner and Quter
Tube Walle for an Quter Tube of 0.66 in. OD and 0.50 in. ID, an Inner Tube of 0.25 in, OD and 0.20 in. ID,
& Molten Balt Mass Flow Rate of 2500 Ib/sec-ft2, and & Tube Wall Thermal Conductivity of 147 Btu/hr. £t. °F.
The molten salt heat transfer coefficients were cbtained from Fig. 10 nd those for stesm from Fig. 12 or
ng- H5.7’ po 320 Of Ref. 8- . ) | )

 

O
 

 

 

ORNL DWG. Ti-5T34%

 

12— — T
| ~ Pressure 4500 psi

 

 

'o—25x106|b/hr-ft2 R | o
0 =21 % 106 Ib/hreft2 | a\

b
o

 

 

--16x 10‘5 Ib/hr-ft? | \(

 

 

 

 

 

By Bu/heeft2eF x 1070
>

Bz o %
—_° AL

O

 

 

4}£%? ;aft;”‘.’Ta 4.-£f""’A _ \\- | NQ‘

 

\

 

™ °

 

 

'Qfif— -l - o D
o gd-HrT e TN |
et L L T

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Do o,

 

D

 

 

 

3 4 5 6 7 8 9
| | L Surface temperature °Fx 102

- Fig. 12. Heat-Transfer Coefficients for Several Mass-Flow Rates

10 )

of Water in a Uniformly Heated Once-Through Boliler Tube Operating at

4500 psia (Dickinson and Welch, Ref. 16)

11

s€
 

- 36
Temperature Differences and Heat Fluxes

The appropriate cur#es of ‘Figs. 5, 6, and 7 were used as the basis
for estimating the_effective local temperature differences in any given
increment of tube length. From these témperature differences, togethér
with the heat transfer-coefficients and. the afiount of heat added in the
increment, it is pbssible to calculate the lehgth of the first increment
of outer tube length and the amount of heat transmitted écross the inner
tube wall. These exchanges of heat in turn make it possible to calculate
the tempersture in each fluid stream at the beginning of the next incre-
ment of’the length. Note that the average local temperature differences
for each increment as used in the calculations were'estimatedéand gen-
erally differed & little from the values indicated by the compléted cal-
culations for the increment. The'difference wés usua;ly sufficiently
small so that iteration was not necessary. To facilitate the use 6f
Fing 5, 6, or 7, the summations of the amounts of heat added or subtracted
to each fluid stream up to and including each increment were also tébulated

togetheriwith the fractions of the total heat added to the water-side stream.

Pressure Drop

The pressure drop across each increment of tube length was also cal-
culated in Table 4. To facilitate these calculationé, Fig. 13 was prepared
to show the density of the fluid on the steam side as & function of its
temperature. For the subcritical pressure boiling conditions the density
in any given increment of boiler tube length was estimated by considering
the fractional change in density in the boiling zone as eQuai to the frac-
" tion of heat ‘added to the steam fip to ‘that point in the boiling region. |
No attempé_was made to allow for the effects of the two-phase flow friction
féctors, but it is believed that the overall effects of these would amount
to only abodt a factor of two in the boiler region, and, of course, there

-would be no effect on the pressure drop in the superheater region.
37

U _ . - \ ' o ORNL DWG. T1-5735

 

 

" 10.

 

4,00
2,00

1.00

SPECIFIC vm (£t3/1p)

0.20

- 0,10

 

 

 

200 koo 600 -

‘ | - Fig.13. Specific Volume as & Function of Temperature for & Typlcal

u - ~ Set of Boiler Discharge Pressures and Temperatures,
 

38

Estimated Performance Characteristics

-

The prime objective of this studyiwas_to investigate the effects of
both the design heat load and the design steam tempersture"and pressure

| cenditions on the boiler tube iength. Once thege effects are established

for a single tube it is easy to estimate the effects of the choice of the

full-power steam pressure and temperature on the size and eost of a full-

scale steam generator. Further, once these effects are established for a

single tube, the effects of the mode of control on the part load tempera-

ture and pressure distribution can be inferred readily, and inferences can

be drawn with respect to the startup and part load control characteristics

and possible boiling flow stability problems.

Effects of Design Heat Load on Tube Iength

A study of the fine structure of the heat transfer relations indicates

that; if a prime criterion is considered to be completion of boiling prior
to exit from the inner tube, the.boiler tube length is'determined bj the
maximun load conditions anticipated. That is, the tube length required
for boiling to dryness will be reduced as the design heat load is reduced.
This effect is shown in Fig. 14 for the two control modes shown in

Figs. 5 and 7, that is, a constant steam generator discharge pressure ir-
respective of heat load and a steam generater discharge pressure directly
proportional to the load. In all cases the calculations were earried out
to determine the tube length required to meet that particular design con-
dition.

Effects of Design Heat Load on Temperature Distribution

A more detailed insight into the effects of design heat load on &

reentry tube boiler is given by Fig. 15 which shows the calculated tempera-
ture distribution in the boiler as a function of distance from the bottom,

"and indicates the length of tube required to evaporate the water to dry-
. ness in the inner tube for a wide range of steam conditions. In this in-
stance the boiler exit pressure was proportionai to the load. Note how
little tube length will be required for the 1% load condition. Note, too,

O

i*

@)
 

 

*

39

ORNL DWG. T1-5T36

700

500

8

TUBE LENGTH (ft)

g |
PRESSURE DROP (psi)

200

. . . _ 0 .
0 ' 40 -8 - 120 160 200
| DESIGN LOAD IN % os' REFERENCE DESIGN CONDITION

 

' Fig. 14, Effects of Ioad -on the Tube Length Required to Handle Loads

‘Iess the Design Heat Ioad of 650,000 Btu/hr per Tube, The assoclated steam-
- side pressure drop is also plotted The steam generator discharge pressure
- for the lower design loads was taken as & constant 4000 psia for one.set of .
;curves , &nd as directly proportional to the load for the other set.

100
 

TEMPERATURE (OF)

ORNL DWG. T1-5737

 

0 | 10
DISTANCE ABOVE BASE OF BOILER (ft)

Fig. 15. Effects of 100, 75%, 50%, 25%, 104 end 1% Heat Loed on the Axial
Temperature Distribution for Operation Witk the Steam Pressure Directly Proportional
to the Heat Load, = : : :

i
 

 

 

w

4

that the curves for the tube length andlpressure,drop in'Fig; 14 rise
steeply between the lOO and 200% load conditions, thus implicitly Jjusti-
fying the choice of flow rate for the nominal 100% load condition. Ac-

‘tually, the choice'of steam»flou rate for the 100% load condition in this

case was arrived at by a series of preliminary calculations with just this
p01nt in mind. '

- The effects on the temperature distribution of the design heat load
per tube for operation-W1th the system pressure held constant irrespective
of the load are shown in Fig. 16 Comparison of Figs 15 and 16 indicates
that the design steam pressure has only a mild effect on the required tube -
length at a given heat load, in spite of the fact that the reduced pres-

sure greatly depresses the temperature in the boiling region

N -

Effects of Part Load'Operation o

If a boiler tube length for; say, the-lOO% reference”design load.con-

. dition of Fig. 15 were chosen as the basis fOr a design, it is evident that
at part load the steam would be superheated to a temperature higher than

- that 1ndicated for the particular cases shown in Fig. 15 because these were
ecalculated on the basis that the inner tube would be termlnated at the

point at which b01ling was completed The superheating effects would be

. particularly large at low steam flow rates where the steam temperature in

the annulus would run close to the local molten salt temperature through
the upper portion of the tube until it dropped when chilled by heat extrac-

~tion for boiling the water rising in the inner tube. | . This effect is shown
_fiin Fig. 17 The calculations were carrled out by recognizing that vir-

| “tually all of the heat transferred from the salt to the superheated steam
‘in any given increment would flow directly into the boiling ‘water in the

 inner tube. Thus,,aalaffirst:approXimation,‘it vas assumed that UpA;ity =
. UsApAt, which then defined tp-and t;. A check showed that this does in

fact give a good approximation andrno iteration was needed in moSt'cases.

| The reason for this is that the enthalpy change in the outer annulus at “the

lO% load condition is only about 5% of the enthalpy rise in the inner tube.
' Tt should be noted that to get good convergence the detailed calcula-
tional technique of Table A was used for the cases of Fig. 17. This
 

mmam °r)

42

* ORNL DWG. T1-5T38

 

0 ' ) 10 20 30
DISTANCE ABOVE BASE OF BOILER (rt.)

Fig. 16. Effects of Design Heat Load on the Axial Temperature Distribution
for 0peration ¥With & Constant Steam Discharge Pressure of LOOO psia.
 

 

 

43

ORNL DHG. T1-5T39

»

1100

1000

800

700

 RMFERATURE °F).

. 500

 

300 b _ _
o . o 10 20 30
DISTANCE ABOVEBASEOF BOILER (£8)

o _ Fig. 17. Comparison of the Axial Temperature Distributions for Operation
‘ J - &t 100% and 10% of full power with the Steam Pressure Directly Proportional to
B ' the Heat Load and the Same Tube Tength in Both Cases.
 

44

represented an improvement over that used for Figs._lA through 16 and is
the reason for a small difference in tube length that may be noted between
Figs. 15 and 17. The heat transfer calculations for Figs. 14 through 16
could be repeated to give better épproximations. This was not done for
this preliminary study because it was felt that the effect would be small,
- and, in any<event, unimportant so long as thé tube length required for
evaporation to dryness in the inner tube is less at part load than at full
load.

Pressure Drop and Pressure Distribution

Figure 18 gives a set of curves similar to those for Fig. 16 to show
the pressure réther-thafi the tempeiature distribfition through the boiler
forvthe load conditions of Fig. 5, i.e., a varying heat load with a con-
stant boiler pressure. Note that the much smaller flow passage area in
the inner tube gives a much higher mass flow rate than in the annulus for
the superheated steam, and this offsets the effect of the higher fluid den-
sity in the inner tube to such an extent that the overall pressure drop
through the inner tube is substantially gresgter than that through the outer
annulus, o |
. fThis effect is desirable from the standpoint of boiling flow stability.
Note also that the pressure drop at low loads becomes very small — too
small to contribute to boiling flow stability. Under these conditions the
principal factors contributing to boiling flofi stability would be the sta-
tic head of the ligquid column in the inner tube and capillary effects,
which would be substantial inside the small diameter inner tube. The
latter are believed to represent a littlé recognized adfiahtage of small
diameter boiler tubes. | | |
| Figure 19 presenfis a set of éurves to show the effects on the pres-
sure distributioh of varying thé deéign heat load with the system;pressure
directly proportional to the hea£ load instead of constant as for Fig. 18.
It is interesting to'note that the pressure drop through the boiler falls
off much more rapidly with a decrease in load if the system steam pressure
is held constant than if the steam pressure is made directly proportional
to the load. - Thus there should'bé less tendency toward difficulties-with

i

O,
 

 

( 45
* -
ORNL DWG. T1-5Th0
&
3
&
K
&
s &
R - | | nrsmmzasmmszmnom(rt) '
- R ‘Flg. 18. Effects of Design Heat Load on the Axial Pressure Distribution for Opemtion

wvith a consta.nt Steanm Diseharge Pressure of 4000 psia.

"%

 
 

FRESSURE DROP FROM BOILER TUBE INLET (pst)

46

ORNL DWG. T1~57Th1

 

10 ' 20 : 30
DISTANCE ABOVE BASE OF BOTLER (ft) '

- Fig. 19. Effects of Design Heat Load on the Axial
Pressure Distribution for Operation with the Steam Pressure
Directly Proportional to the Heat Load.
 

 

4T

boiling flow instabilities under light load conditions if the system is
designed so that the boiler discharge pressure varies linearily with the

- power output.
Effects of the Fraction of Heat Added in the Inner Tube

- Perhaps the mOst inportant-arbitrary assumption made in order to pro-
vide a starting pointtfor-the‘iterative calculations was the fraction of
the heat added in the inner‘tube; A series of calculations was carried
out to investigate this effect'for the load condition found to be most sen-
sitive on this score, i.e., 10% of the reference design load. The results
| of these calculations are plotted in Fig. 20. These curves indicate that
the results are relatiVeijdinsensitive to the initial assumption for rea-
sonable values of QB/Ql,ii.e;;-Qg/bhlless.than 80%. Thus terminating the

center tube when boiling is completed gives a well-proportioned design.

Effects of Size of'Heatfificrement Used

Initially ‘the calculations were carried out using 1ncrements in the
amount of heat transferred,equal to 20% of the total for the tube. These
coarse increments led to Quite'serious convergence problems, hence the
size of the 1ncrement wag reduced to 10% of the total heat load. This
gave good convergence except at loads of 10% or less where the temperature
of the superheated steam v3501llated with a perlod double the increment
‘used for the calculation 1n the manner shown in Fig. 21 ThlS effect was |
largely eliminated by reducing the- 1ncrement of heat transferred to 5%
'1nstead of 10% Figure 21 shows that doubling the number of calculational
"increments not only reduced the amplitude of the 080111ation but 1t also
* reduced the period of oscillations, This further confirms the belief that
the oscillation is an artifact of,the'calculational procedure.

. Turbine Control Considerations

As will.be-shoun‘lster;‘the steam generatorldesign proposed here in-

herently gives a nearly constant superheater steam outlet'temperature
 

PRESSURE DROP. (ps1)

 

ORNL DWG. T1i-5Tk2

Lo

25

20 .

TUBE LENGTH (£t)

| : 10
50 60 70/ 80 90 100
e
%4 linitiel
Fig. 20, Effects of Initial Choice of Fraction of Heat Added
to Steam in Inner Tube on the Tube Length Required and the Associated
Pressure Drop.

O

O,
 

 

 

 

~ TEMPERATURE (°F)

49

" ORNL DWG. T1-5T43

 

0 W 8 1. . T
- n:s'mm ABOV‘E‘.BASE OF BOILER (rt)

Fig. 21 Erfects or Size of Increment in the Stepwise Calculation on the Amplitude

© end Frequency of the Apparent Oscillation Encountered at & Heat Losd 104 of the Rererence

Des:lgn COndit:lon w:.th a Steam Pressure of 4000 psia.
 

 

50

irrespective of load, This, coupled with the very small water inventory,
makes e_method of power plant control impliecit in Fig. 7 lcok ettractive.
The inherent characteristics make'possible simple and reliable control
~equipment and instrumentation; and the plant response characteristics ap-
pear to be much better than for conventional fossil fuel plants; In view
of its pramising characteristics but highly unorthodox nature, R. B. Briggs
urged the writer to dlSCUSS this control approach with turbine des1gn
engineers, Talks with several different engineers at both Westinghouse and
Generai Electric in each case seemed to lead them to conclude tentatively
that there is no apparent reason why it shouldn't work, and they suggested
that an evaluation of permissible rates of change of load be made using a
typlcal chart supplied to operators of power plants. A chart of this type
from a turbine operators manuval prepared by-westinghousel7 is shown in

Fig. 22. '

Limitations on Rates of Change in Load

It should be mentioned that a major limitation on the operetion of a
large steam power plant is the rate at which the load can be'changed. One
set of limitations in coal-fired-plants is associated with the furnace and
boiler; tnese not only have an enormous thermal inertia but the boiler
perfbrménce is sensitive to changes in temperature distribution in the fur-

nace as a consequence of changeslin the flame‘geqmetry'with,changes in

load. A'second and usually more stringent set of limitations is associated.

with the turbine. The close running clearances between the rotor blade
tips and the casing make it imperative that these two components;change
temperature together. However,_the temperature cf the thin rotor‘blades
operating in the high velocity steam_follows the steam temperature with
virtually no phase lag, but the thick-walled casing does not. Further,
there gre major circumferential asyrmetries in the casing, partly because
it is split horizontally through the centerline and the upper and lower
halves are joined by a'heavy bolting flange, and partly because the steam
inlet and outlet passages enter the casing from the bottom side; Thus,xto
avold catastrophic interference between the rotor and the casing as a con-

sequence of differential expansion between the rotor and the casing or

O

O
 

 

 

“C

a

q

51

because of distortion in the casing, it is generally necessary to limit
the rate of change of thefsteam temperature to a low value. Obviously, it
would be highly advantageouswto'the“electric power plant)system operator
if he could change the load on & unit rapidly Thus it is interesting to
use Fig. 22 to evaluate the time required to change Joad in a plant having

- the characteristics of Fig 7

Startup and Rates of Change Of Load with‘Proposed System

Considering first the startup conditions, it is helpful to quote from
Ref 17, p. 8, the steam pressure and temperature conditions to be estab-

lished prior to starting,the turbine- generator unit:

For drum type boilers, establish about 25% of rated pres-
sure and at least 100°F superheat, but not more than 800°F
total temperature, at the turbine throttle valves

For once- through (drumless) boilers, a reduced pressure
start is recommended.  This applies to both "cold" and "hot"
starting of turbine units. For "cold" starts, the boiler should
be stabilized before opening the: throttle valves and the inlet
temperature held to 800°F maximunm.

Thus the turbine -generator should be started with a low steam pressure at

" the turbine throttle valve if one is used. If a throttle valve is not
_employed, thenrthe steam pressure delivered_by the steam generator must

~emphatically be at a low value. From the standpoint-of the steam pressure,

1t appears from Fig. 7 that the control scheme proposed should be satis-

,factory : However, the steam.temperature will be too. high, and hence a de-
_superheater would be required to hold the steam temperature to 800°F at the '

'r-fturbine. Such units are commonly used in conventional coal fired plants,

Once the turbine has been brought to equilibrium at a load of perhaps

110% at a pressure of 400 psi (see Fig. 7), the time required td increase
. the steanm temperature up to 1000°F (1 e, » close to the normal operating
- temperature of about 1050°F) can be estimated from Fig 22 to be. 40 min, a
' substantial period;. but one consistent with conventional practice Once

 the system is up to operating temperature, however, changes in load can be

made relatively guickly.- If-one wishes to go from the 10% load condition
of Fig. 7 to the 50% load cOndition, this will entail increasing the
 

52

ORNL DWG. T1-5Thk

 

 

Ao

 

 

 

 

/
7
7
://
L/// _
1
é
|

   

 

G

 

/)

7

Z

/

|

A AE
9—

g.

i

'
NG TEMPERATURE —o-

 

 

 

 

 

 

 

TEMPERATURE —

 

 

 

 

 

 

 

 

 

 

 

| Ala_zo.__azaqmmod__ — /Q/ ovovJ

 

~ SAULANIN - NOLLIONOD WVALS H0/ONV QVOT 39NVHO OL 3WIL

 

DECREASING

 

 

 

 

 

 

 

 

 

 

ol

,‘ , >
8 8 £88§°88%
. _ .ao._w Im.o

JSVIUONIITEV3YO34a

, 1Sd =~JONVHO
AUNSSIYd  FILLIOUHL

EXAMPLE: 80 MINUTES ARE REQUIRED TO CHANGE FROM 5%

LOAD 800 PSIG-800°F TO 100% LOAD 2400 PSIG-I000°F

P-68969 -

Recommended Time for Changing Load and/or Steam Conditions,

(Courtesy Westinghouse Electric Rorp.)

Fig 22.
 

 

 

 

)

»

53

pressure by 1600 psi but the temperature by only about 30°F. According
to Fig 22 the turbine could tolerate making this change in only 11 min,
Similarly, if the turbine were operating at full load, the load could be
cut in half in as little as 20 min., i.e,, substantially faster than the

- common rule of 2%7min Further; 1if the steam generator can be designed

S0 that the superheater outlet temperature increases SOmewhat with a reduc-‘
tion in load, even faster rates of change can be tolerated. For example,
if the full load steam pressure and temperature were . 4000 psi and 1000°F
and the 50% load conditions were 2000 p81 and 1050°F, Fig. 22 indicates
that the load change could be made in about 1l min. As will be shown later,

it happens that the reentry tube steam generator can be proportioned to
o give this characteristic and hence would make-. possible unusually rapid

changes in load.
Possibility of Eliminatingrthe.Throttle Valve

. Steéeam throttle valves are a major source of operating trouble in
steam power plants. The valves are relatively large, must operate at _
about 1000° F, they are subject to heavy loads by the high steam pressure

"drop across them, and lubrication condltions are highly unfavorable Thus
it is not surprising to find that throttle valve sticking is one of the |

principal causes of - forced outages in steam power plants.'
Throttle valves are used to make it possible to change the steam flow

- rate much more rapidly than can be- accompl:l.shed by controlling the boiler :

and furnace in conventional coal fired plants. The most critical consid-

eration is to avoid a serious“overspeed of the turbine in the event of an

" abrupt and complete loss'of~the electrical load, The inventory of super-

. heated water in the . reentry tube b01ler 1s vastly less -than 'in'a conven-

tional boiler; it is estimated to be only about 3 1n.3/tube at full load,

and about 1in.? at 10% load on the steam generator NOte that as a con-
sequence of mechanical, electrical, and fluid friction losses, ete. in the

]turbine—generator unlt, 1n the event of a complete loss of. electrical load

the steam flow required to keep the turbine up to speed would be about 10%

 of the full power output of the steam generatOr. Thus an abrupt loss in

electrical load would entail a reduction in the superheated water inventory
 

54

in a reentry tube boiler by about 2 in.3/tube, or 4 in.2/Mw(t). This full
load steam flow rate is about 3 1b/sec for 4 Mw(t), whereas 4 in.3 of
superheated water is only about 0.1 1b. Thus the full load steam flow

- rate would consume the surplus water inventory in only about 1/30 sec if
the feedwater supply were abruptly cut to the 10% load level. The inertis
°of the massive rotor in the turbine-generator unit should be. sufficient
“to - keep the‘bvershoot in rotor speed to a low value, Inasmuch as.the most
diffiéult‘condition to meet is that for an abrupt ldss‘in electrical load,
it appears that control of the power plant could be accomplished by con-
trolling the feedwater flow rate with one or moré relatively'small valves
opefating at about 650°F rather than the large, hot steam throttle valves
normally employed. This should give a greatly increased reliability.

It is of interest to note that this approach to the control of a Ran-
kine cycle plant has been analyzed and investigated experimentally and is
a very similar but much smaller system designed for a nuclear electric
space power plan.t.18 The analyses and tests gave highly encouraging re-
suits'fdr a system in whiéh.the thefmai ifiértig-OT the boilef wés small as
_in_fhe'system proposed here. Further; the variable pressure apprbach is

coming into use in conventional plants.l®
Proposed Design for a Molten Salt Reactor Plant

A conceptual design for a molten‘Salé7reactor'With its intermediate
heat exchangers and fuel pumps integrated into & common pressure vessel is
presented in-a'companion-report.20 'The net electrical output in that.
study is 1000 Mw(e) and the fuel and inert salt temperatures are those

- given in Fig. 4a. Plant layout studies favored-the use of six steam gen-
~erators each of which‘would be directly coupled to one of six fuel-to-NaFF,

heat exchangers. The steam generators requiréd for this plant provide a

'good illustrative example for use here to show how the sihgle tube analyti-

cal work presented earlier in the report can be applied to the design of =&
full-scale steam generator as well as’the proportions=to be expected in a
" finished unit. ‘ B

O
 

 

 

B

3

0

55

Reheaters

Most modern steam-plants emploj reheaters, in part because there is a

| direct increase in overall cycle efficiency of about 5%, and in part be-

cause they give a further indirect increase in cycle efficlency of a few

percent by eliminating the moisture churning losses 1n the lower stages of

- the turbine, In addition, reheaters can be designed to eliminate the pos-

sibility of turbine bucket erosion in_the.lower turbine_stages by reducing

the moisture content in thet‘region to almost nothing. In conventional
coal fired steam plants the length of piping required to connect the boiler
to the reheater and the relatively large amount. of tube surface area re-

quired because the reheater must be located in a relatively low-temperature

- zone to avoid burnout difficulties have combined to make the cost of in-

cluding provisions for reheat rather high for conventional coal-fired
plants. However, it is believed that in a mOlten salt reactor plant the

turbine can be located much closer to the steam generator?and,'because

there is no danger of tube burnout from excessive locsl temperatures, &

much higher average heat flux through the reheater tube walls can be main-

tained so that the tube surface area requirements are quite modest., Fur-

‘ther, the design of the tube matrix for the reheater is quite straight-

forward because there are no flow stability or two-phase flow problems in-
volved. Thus a reheat cycle appeered_highly desirable for this design
study, and the two reheat stages specified for_Eddystone'Unit No. 2 have

been included in the proposed steam generator for. a mOlten:saltgreactor.

. General Description

The dominant consideration'in choosing the steam'generator'0verall

_geometry was to make use of the statlc head in the ligquid column in the
o b01ling region to help provide for flow stability in the boiler. This
_'_meant that the boiler tube should be vertical with the’ feedwater entering
irat the bottom While it would ‘have been desirable from the standpoint of

thermal convection 1n the molten salt system to admit the molten salt at
the top and have it leave at the bottom, a brief glance at ‘the temperature

distribution diagrams shown in Figs. 4, 5, 6, and 7 is sufficient to show
 

 

56

that the salt sfiould move counterflow relative to the superheated steam in
the outer annulus. Figure 23'presents a somewhat similar disgram including
the two reheater stages and shows that'they,'tbo, should be inICOunterflow.

It was thought at first that it would be desirable to make use of
separate casings for the boiler and the ' two reheaters, but the additional
plping, manlfolding, and associated temperature distributlon prdblems led
. to the conclusion that it would be best to mount the reheater tubes in an
" annulus surrounding the b01ler tubes in the conflguratlon shown schema-
tically in Fig. 24 .

'fieadering Problems

As mentioned earlier the headers envisioned for the boiler region
would entail the thermal sleeve arrangement shbwn in Fig. 3. This avoids
-the use of a high'pressureeheader sheet'ifi'eontact with the molten salt;
rather, the tubes would be manifolded below the pressure vessel containing
the molten salt. The same arrangement could be employed for the reheater
tubes. In either case any tube could be blocked off readlly  -outside the

steamvgeneratorAcasing in the event of a leak or other mslfunction. Inss-

much as the reheater tubes are free of the complexities associated with the

~ inner tube used for the boiler, it would also be possible to assemble them
in bundles of perhaps 12 tubes with the tube bundle headers inside the
casing of the steam generator to minimize the number of penetrations
fhrough the shell, This looks preferable if a unit has more than 100 tubes.
Provision of adequate space for header sheets is always difficult
when designing heat exchangers with closely spaced tubes. Preliminary lay-
out studies indicate that this is difficult but might be done with the con-
figuratioh’of Fig. 25 without severely distorting the‘Outer_casing or hav- |
ing a large volume of inert salt that serves no useful function but in-
creases both the capital investment and the,time required for the system
to respond ta control actions. The outlet'spigets from bundles of 16
tubeS'cduld be passed through a header sheet in the bottom head of the
vessel with a thermal sleeve to avoid large local stresses. This would
glve an adequate ligament thickness (about 0.5 in.) in the header sheet
'after allow1ng space for the thermal sleeves, This will take up most of

<:;9,i
 

”»

»

"

»

-Tepii:eratum (°F)

57

~ ORNL DWG. T1-5T45

'1206
1100
1000
900
=
T00

600

 

-
| f AQ/Q ®

: Fig. 23, Temperature I)istribution in the Bo:ller a.nd Reheater as
a Function of the Fraction of the Heat Removed from the Molten Salt

Under Full Load Conditions.
 

58

SALT POMP ORNL DWG. T1-5Thé

 

 

 

REHEAT STEAM
OUTIET MANTFOLDS

' 4 in. dis.
5 TEEL 45 £t tall

 

 

 

 

 

SALT IN

 

 

Fig. 24. Proposed Configuration for One of
& Set of Six Steam Generators for a 1000 Mu(e)
Molten Salt Reactor Power Plant. Detailed data
are given in Table b,

O
 

 

 

o

P

| ]

n

59

ORNL-LR-DWG 73420

    
   
 

~1),-in. TUBE -
(vEg)

t-in. TUBE -
N7 4)

 

 

 

 

 

 

 

 

 

 

%-in. TUBE
Wz 4)
Y5-in. TUBE_
CAN
. - _ " . i . "
W W W W WD LG

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 Fig. 25. Schematic Diagram Shoving the Bifurcated Tube Header

Arrangement Used as a Basis:;fbr"_ the Propbsed Tube Bu;ridle and Header

~ Sheet Configuration. . The nominal tube id (d,) is indicated for each
 step in the manifold to show that the flow velocity is kept constant.
 

.60

the area of the elliposidal head in the region where its radius of curva-
ture is large. The space just outside this region but inside the shell
will serve as the inlet plenum for the molten salt which would enter tan-
gentially.through two pipes. The 96 spigots from the tube bundles could
be manifolded outside the header sheet using a bifurcated tube arrange-
‘ment similar to that of Fig. 25. Three steanm pipes-could.be coupled to
each set of 32 tube bundles. The feedwafier could be suppliéd through two
penetratibns in the largest bifurcated colipling in each of these three
sets. _ _ :

The layout envisions installation of small header drums for the re-
heater tube bundles in fhe annulus just above the salt inlet plenum. ‘The
annular baffle between the reheater region and the boilef-superheater re-
gion fiould extend downward no farthef than the header drums for the re-
heater tfibe bundles.- | _ .

- The layout problems would be eased by reducing'the numbe? of boiler
tubes in each steam generator unit and employing a large number of units.
The ratio of tube matrix cross-sectional area to header sheet area could
be increased in this way and manifolding problems of the tubes for the

feedwater and superheated steam could be eased.

 

Differential Thermal Expansion

There would be no problem associated with differential thermal expan-
sion between the boiler tubes and the steam generator casing beéause the .
top ends of the tubés wouid be free to float axialiy. However, the re-
heater tubes will tend to run at & scmewhat different temperature than the
casing and hence some means for providingfieiibility shofild be employed.
Probably the easiest way to accomplish this will be to install the tubes
on & spiral having a stéep pitch, i.e., perhaps a total twist of 60 deg
in the length of the regenerator region. This could be chosen to pyovide
sufficient flexibility in the tubes so that the differential thermal ei--
pénsion between the tubes and the shell would be accommodated elastically.
The pitch would be steep enough so that the lateral loads on the tubes
 would probably not be large enough to make it neceésary to employ addi-
tional structure to resist the forces associated with the small amount of

erossflow that will occur.
 

 

L0 kA e LA 3l R o L Lo AR

W

»

ol

'_Geometric and Performance Data

The calculations and prlncipal geometric and performance data for the
complete steam generator are summarized in Table 5. Recent revisions in
the physical properties of.NaBF4, particularly in the viscositj, made it
desirable to change the salt massrflow rate through the steam generator
from the value of 2500 1b/sec:ft? used'in_the-Calculations of'Table 4 and
those for Figs. 11 and 14 through 20. This, coupled with the changes in
physical properties, led to a change in the tubemlengthirequired.

The heat load for the boiler-superheater and each of the reheaters
was used to calculate the number of tubes for each matrix, with the molten

salt streams for each sectlon treated as flowing independently of each

other with no lateral heat transfer, Note in therflrst portion of Table 5
that- steam bleed-off in the turbine through seals and for feed heating re-

duces the steam flow rate to the reheaters 50 that it is substantially less
than the steam flow rate to the boiler.

'CoSt'EStimate

A rough estimate of the cost of the steam generator of Teble 5 can be
obtained by summing the estlmated costs for the three ma jor items, i.e.,
the shell, the tubing, and fabrication of the tube-to-header joints. This
has been done and the results are summarized in Table 6. Note that the
overall cost amounts to about $7.05/xw(e), which compares with around
$20/kw(e) for the comparable'portion of & conventionalrcoal'fired boiler.

(These cost estimates were made on the basis of unit costs as of 1968. )

- Thus it appears that the combined effects of a uniformly high heat flux,
:fsmall diameter tubes with relatively thin walls, and ellmination of the

need . for pressure vesselswsubject to high pressures.much.more than offsets

" the high cost per pound of Hastelloy N.

- The prOportions glven in Table 5 were not 1terated to mske both the

'etube length and the molten salt pressure drop for the boiler-superheater

and the two reheaters the same., This could and should be done, but the

' combined effects)of uncertainties'in'the calculated heat transfer coeffi-

cients and friction factors are 1arge enough to make this an unwarranted

refinement at this stage.
 

 

62

Table 5. Summary of Design Calculations for Steam Generator Units

for a 1000 Mw(e) Molten Salt Reactor Plant

| u

 

 Steam System

Superheater outlet temperature, °F
Superheater outlet pressure, psia
'Reheater No. 1 outlet temperature, °F
" Reheater No. 1 inlet temperature, °F
Reheater No. 1 outlet pressure, psia
Reheater No. 2 outlet temperature, °F
Reheater No. 2 inlet temperature, °F
Reheater No. outlet pressure, psia
Generator output, Mw
Auxiliary power requirements, Mw
Boiler feed pumps and boiler auxiliaries, Mw
- Fuel pumps, Mw. '
NaBF, pumps, Mw
Miscellaneous, Mw :
Net electrical output, Mw
Overall thermal efficiency, %
Fraction of heat load to boiler-superheater
Fraction of heat load to reheater No. 1
Fraction to heat load to reheater No. 2
Fraction of weight flow to boiler-superheater
Fraction of weight flow to reheater No. 1
Fraction of weight flow to reheater No. 2
Heat added to steam flow in boiler superheater, Btu/Ib
Heat added to steam flow in reheater No. 1, Btu/1b
Heat added to steam flow in reheater No. 2, Btu/lb

Inert Salt (NaBF;) .

Entrance temperature, °F
Exit temperature, °F
Entrance pressure, psi
Exit pressure, psi
Flow rate, lb/hr
Flow rate, lb/sec
Flow rate, ft2/sec
NaBF,; physical properties at lO75°F
Specific heat, Btu/lb :
Thermal conductivity, Btu/hr.ft.°F
 Viscosity, 1b/hr.ft
Density, 1b/ft3
Melting point, °F

MONOHH

Shell
Number of units
- Shell diameter, in.
Shell overall height, ft
Shell wall thickness, in.

1050
4000
1050
786
1043
1050
705
251
1043
66
50

5.4

10

1
977
i 4
0.775
0.118
0.107
1.000
0. 84
0.68
886.0
161.0

- 180.2

950
1200
180
51

13.92 x 106

3860
33.6

0.36
0.27

3.4

115
725

45

40

0. 50

e
 

e

B

o

63

Table 5 (Continued)

 

Boiler
 Boiler heat load, Btu/hr

Tube OD, in,
Tube ID in. ,
Tube length, ft '
Tube pitch (equilateral triangular), in
Salt mass flow rate, lb/sec 52
Dynamic head, psi S
- Reynolds number -
Friction factor -
Pressure drop, psi - -
Shell-side flow passage area/total Cross- sectional
area

~ Shell-side flow passage equivalent diameter, in. -

Number of tubes per square inch, in.”

Number of tubes per unit -

Shell-side surface area, ft?

Rehegter No. 1

Tube OD, in.
Tube ID, in. |
Tube pitch (equilateral triangular), in.

' Log mean temperature difference, °F

Heat load per steam génerator, Btu/hr

Steam mass flow rate, 1b/sec:ft?

Overall heat transfer coefficient, Btu/hr-ft2 °F
Surface area per steam generator, ft? SR
Surface area, ft°/ft of tube

- Number of tubes -

Tube length, ft o
Salt flow rate, ft3/sec

'Salt mass flow rate, lb/sec £t2
Salt flow passage area, £t2

Total cross-sectional area in tubes, ft°
Total cross-sectional ares in tube matrix, ft2

Shell-side flow passage/total matrix cross- sectisnaln
. area ,
_Shell-side flow passage eqUivalent diameter, 1n
-Steam pressure drop, ps1l _ :

‘Salt pressure drop, psi

Reheater No. 2"
- Tube OD in.

Tube ID, in., ’

Tube pitch (equilateral triangular), in.
Log mean temperature difference, °F
Heat load per steam generator, Btu/hr

9.73 x 108

0,65 -

- 0.50
. 35,0
0. 775 .
.. 1350
1.5 '
60,000
- .0.025
45 ,
0,366

0.465

191
- 1493,
8900

1.00

0.80.

- 1,035
150

1,478 X 108

200 -
380 |
- 2390

0.236
365

33.6
33.6
11350

o127
S2.328
L W145

.18

20
84

1.00

0.90
1.01 -

- 200
1.34 x 108
 

64

Table 5. (Continued)

 

Reheater No. 2 (continued)

Steam mass flow rate, lb/hr.ft? _
Overall heat transfer coefficient, Btu/hr ft2 °F o215

70

Surface area per steam generator,: 2 3300 -
Number of tubes 632
Tube length ft 22
Salt flow rate, ft3/sec 3.6
Salt mass flow rate, 1b/sec-ft? 1000
Salt flow passage area, ft2 0.414
Total cross-sectional area in tubes, ft2 - . 3.63
Total cross-sectional area in tube matrix, f£t2 - 4,04
Shell-side flow passage/total matrix cross-sectional = 0.10
area

Shell-side flow passage equ1valent diameter, in - 0,11
'Steam pressure drop, psi - : - g

 

Table 6. Rough Cost Estimate for the Steam Generator of Table 5

 

 

. Ttem flAmount ) :M'i Unit Cost = Cost
Shell (fabricated) 10,000 Ib $8/1b° ¢ 80,000
Baffle - 2,000 1b .. $/m ' 16,000
Boiler tubes 60,000 ft, 0.25 in. diam $3/ft = 180,000

- o 60,000 ft, 0.65 in. diem $7/ft =~ 420,000
Reheater tubes 25,000 ft, 1.0 in. diam - $8/ft 200,000
Material Cost | . . $ 896,000
Tube installation 2490 tubes © - $110/tube 273,900
(including welding - o ‘
of tube-to-header
Joints and inspec-
tion
Total Shop Cost © $1,169,900

 

 

O
 

 

 

ad

e

65

Conclusions

 The proposed“reentry tube boiler appears to offer a long?sought solu-

tion that satisfies all of the magor requirements for the steam generators

of 11quid metal and molten salt reactor power plants It appears to be

well suited to the generation of steam at any desired pressure and tempera~
ture condition w1th good stability and control characteristics throughout |
the range from zero to full power, .It_lends,itself_to_designs in which
the thermal stresses throughout the unit can be‘kept within the elastic
range under all operating conditions. The heat flux is uniformly high so
that the inventory of'structural metal or liquid metal and molten salt is

- near minimal, the number of tube-to-header joints is relativelycsmall

 there is no high pressure header sheet in contact with the high-temperature

liquid, there appears to be no unusual fabrication problems, and the capi-
tal cost appears to be competitive. The only apparent disadvantage is
that the concept is novel and has”not'been'tested.

The performance characteristics of reentry tube boilers presented in
this report represent rough preliminary approximations based on a set of
simplifying assumptions. Test experience~with'a unit'employing at least
three full-scale tubes would provide a firm foundation for the design of a
large unit. ‘

-~ Recommendations

A test unit consisting of one or & few tubes should be built and
tested In the test program particular attention shOuld'be given to the

-1nvest1gation of possible boiling flow instabilities under startup, near-

zero power, and part load operating conditions to see to what extent ori-

-ficing may be required at the feedwater inlet. The test program should
 also 1nclude an investigation of the overall stability and control charac-

 teristics under steam conditions ranging’from 100 psia to 4000 psia.

 Most of the problems and uncertainties associated with the reentry

.tube boiler could be investigated with a relatively S1mple single tube unit

fitted with clamshell heaters, Such a unit could be built and tested ex-
peditiously and inexpensively, and would lend itself nicely to a detailed
 

66

investigation of the temperature distribution along the tube under a wide
range-of conditions as well as facilitate modifications such gs the inser-
" tion of inlet orifices. - | _ , |

An electrically heated single tube test will probgbly be very much
1ess convincing to most people than a long endurancetestiof'aunit'havihg
3 to 19 tubes heated by a molten salt. Such a test unit should.give a
good demonstration of the freedom of the system from éorfosidn,_mass trans-

fer, thermal stress, and boiling flow stability probléms, f

O
 

 

»

10.

11,

13.

 

67
-':References

A. P. Fraas, Flash Boilers for Fluoride Fuel Reactors, USAEC Report
ORNL~-CF-55- 6~79, Oak Rldge Natlonal Laboratory, June 16 1955.

P. R. Kasten et al., De81gn Studies of lOOO-MW(e) Mblten—Salt ‘Breeder
Reactors, USAEC Report ORNL-3996, Oak Rldge National Laboratory,

,Ausust 1966 TP 77-87.

D. Scott and A. G. Gr1nde11 CQmponents and - Systems Development for
Molten-Salt Breeder Reactors, USAEC Report ORNL~TM;1855, Oak Ridge -
Natlonal Laboratory, June 30, 1967, pp. 21-22.

Oak Ridge Natlonal Laboratory, MSRP Quarterly’Progress Report _
Jan. 31, 1959, ORNL-2684, p. 5h. ,

L. G. Alexander et al., Molten Salt Converter Reactor Design Study

and Power Cost Estimates for a 1000 MW(e) Station, USAEC Report
GRNLTM—lOGO September 1965 -

B. W. Klnyon and G D. Whitman, Steam Generator-Superheater for

‘Molten Salt Power Reactor, ASME Publication 614WA-228 Nov. 26-
- Dec. 1, 1961 | ,

R.. S. Holcomb and M. E. Lackey, Performance Characterlstics of a
Short Reentry Tube Steam Generator at Low Steam Output, USAEC Report
ORNL-TM-3236, Oak Ridge National Laboratoiy (in press).

A. P. Fraas and M. N. Ozisik, Heat Exchanger Design, Wiley & Sons,
Inc., New York, 1965. |

~ B. 8. Shlralkar and P Grlffith The Deterioration in Heat Transfer

to Fluids at Supercritical Pressure and High Heat Fluxes, Rept. No.
70332-51, Dept. of Mechanical Engineering, Engineering’ Progect Lab,
Massachusetts Instltute of Technology (Mar. 1, 1968)

D. W. Lee and R. C. Spencer, Photomicrographic Studles of FUel

Sprays NACA Technlcal Report h5h 1933.

D. W. Lee, The Effect of Nozzle Design and Qperatlng Conditions on
the Atomization and Dlstrlbutlon of Fuel Sprays NACA Technical
Report h25, 1932. '

P. J. Birbara, J. E. 'éo#les'and'v C. A. Vaughen, Flash Boiler

Test II, Memo No., EPS-X-296, Massachusetts Institute of Technology,
'Englneerlng Eractlce School, Union Carbide Corp .5 Jan. 18, 1957.

Oak Ridge Natlonal Laboratory, MSRP Semiann. Progr. Rept. Jan. 31,
1964, USAEC Report 0RNL-3626, P- gh
 

14.
15.

16.
17.

18.

19.

- 20.

68

A. P. Fraas and M. N. Ozisik, Steam Generators for High-Temperature

' Gas-Cooled Reactors, USAEC Report ORNL-3208, Osk Ridge National

Laboratory, April 8, 1963. |

W. R. Chambers, A. P. Fraas and W. N. Ozisik, A Potassiun-Stean
Binary Vapor Cycle for Nuclear Power FPlants, USAEC Report 0RNL—3584
Oak Ridge Natlonal Laboratory, May 1964

N. L. Dickinson and C. P. Welch, Heat Transfer to Supercritical

Water, Trans. ASME, Vol 80, p- 746 1958

Westinghouse Turbine Operating Manual I. L. 1250- 32184, 300-350 Mm(e),

300 psi Hydraulic Governoring System, 1967.

M. M. Yarosh and P. Gnadt, Use of a Cavitating Pump for Control of a
Potassium Rankine Cycle System, paper presented at the Intersociety
Energy Conversion Engineering Conference,_Las Vegas, September 1970.

R. J. Bender, U,S. Welcomes Variable-Pressure Concept, p. 56, Power,
August 1970.

A. P. Fraas, Conceptual Design of & Molten Salt Reactor with Its
Intermediate Heat Exchangers and Fuel Pumps Integrated in a Common’
Pressure Vessel, USAEC Report ORNL-TM-2954, Oak Ridge National
Ieboratory (to be published).

O

r
 

 

b

&

C

89-90.

38.'

,'I'O.
4.

ho,

. 86.
87.
88.

al.
92.

 93-94,

96-98.
99-100.

Je

S.
M.
c.
E.
E.

Co

H.

R.

C.

d

W.
J.
F.

J -

S.
We

-J.

D.
A.
L.
Ww.
A,
W.

P,

H.
P.
Jd.
M.

69

INTERNAL DISTRIBUTION

L. Anderson

F.
E.

Bauman

Beall, Jr.

Bender

E.
S.
G.
Je
I.
B.
W.
W.
B.
L.
L.
R'
Jde.
P.

R..

E.
P.
D.
R.
G.
0.
N.
E.
W.
R.

._J.

I.

Bettis

Bettis
Bohlmann
Borkowskil
Bowers
Briggs
Collins
Cooke
Cottrell
Crowley
Culler
DiStefano
Ditto
Fatherly
Engel
Ferguson
Fraas
Fuller
Grimes
Grindell
Harms

Haubenrecih -

Helms
Hoffman
Kasten
Keyes, Jr.
Lundin

- 43,
b,
L6.

b7,
48,
kg,
50.
5.
52.
53.
L

.55
56-65.
66.

67.
68 69.
70,

1.

2.
3.
7)4'0
T5.
76.
TTe
8.
| T9e
80-81.
82.
83-8L.
85.

'EXTERNAL_DISTRIEUTION

David Elias, AIE-Washlngton

- R. Jones, AEC-Washington

Kermit Laughon, AEC-OSR
T. W. McIntosh, AEC-Washington
M. Shaw, AEC—Washington |

W. L. Smalley, AEC-ORO

 

~ ORNL-TM-2953

.R. N. Lyon
--H., G. MacPherson

R. E. MacPherson
H. E. McCoy

"H. C. McCurdy

H. A. Mclain
L. E. McNeese
J. R. McWherter
A. J. Miller

R. L. Moore

E. L. Nicholson
A. M. Perry

R. C. Robertson
M. W. Rosenthal
J. P. Sanders
A. W. Savolainen
Dunlap Scott

‘Mo Jo Skimler

I. Spiewak

D, A. Sundberg

R. E. Thoma

D. B. Trauger

A. M. Weinberg

J. R. Weir

M. E. Whatley

G. D. Whitman

L. V. Wilson

Central Research Library
Document Reference Section
Laboratory Records
Laboratory Records (LRD-BC)

Division of Technical Information Extension (DTIE)

Iaboratory and University Division, ORO

Director, Dlvision -of Reactor Licensing, AEC, Wash
Director, Dlvision of Reactor Standards, AEC, Wash,