NA
..
.
I
.
•
•
.
TOF
ORNL P
2052
.
:1
.
:
12.5
18 2.2
TPFTERET
1
.
11:25 |1.1.4 16
MICROCOPY RESOLUTION TEST CHART
NATIONAL BUREAU OF STANDARDS -1963
wa
OLNU-P 2082
conf-660313-1
GFSTI PRICES
4.C. 57.00; MN_.50
A POTASSIUM-STEAM BINARY VAPOR CYCLE FOR A
MOINEN-SALT REACTOR POWER PLANT*

MASTER
A. P. Fraas**
ABSTRACT
ut
1
.
..
..
.
-
,
instalare way
much monitor, w the cor der
om contato
.. !
...
-
+
--
--
A conceptual design for one embodiment of a binary vapor
cycle coupled to a raolted-salt reactor has been prepared to
determine whether such a plant is sufficiently attractive to
warrant further investigation. Its overall thermal efficiency
is estimated to be 54%, while its heat rejection to the con-
densers is about half of that for a modern steam plant. The
quantities of material required for the heat exchangers and
piping for both a coal-fired supercritical-pressure steam
plant and a nuclear-powered potassium vapor and supercritical-
steam plant are estimated and canpared along with the 8860-
ciated costs. The resulting cost and performance data indicate
that, the nuclear plant with a potassium-vapor and steam binary
cycle may give both lower capital charges and a much higher
overall efficiency than a coal-fired supercritical-pressure
steam plant.
mand wert. the her the United
LEGAL NOTICE
General
n
manum i
, ---rbereid. -- .
why internetten, maro, method, w more detaet dinner way wed during
ity
of
e
u het
tot, a m
ene, mensen met
to word met e
may wortor i de
tan, « imate mouw. my termin porwani wa Money o
imeaimorto y lo mencionada amb motor ting sama
nue tot
INTRODUCTION
Carnot demonstrated that the best way to increase the thermal effi-
ciency of a heat engine is to increase the peak temperature of the thermo-
dynamic cycle. Thus it appears that, because of material limitations
imposed by corrosion and 1088 in strength at high temperatures, conven-
tional steam power plants have about reached the upper limit of their
development toward higher thermal efficiencies. Figure 1 shows that,
*Research sponsored by the U. 8. Atomic Energy Commission under contract
with the Union Carbide Corporation.
.
**Oak Ridge National Laboratory, Oak Ridge, Tennessee.
N

.
RELESED
S
EMENT
&
$
SISACIS
IN NUCLEAR SCI...
AT
1
.
while the operating temperatures and thermal efficiencies of new steam
planto increased steadily throughout the first half of this contury, the
U
increase has cssentially stopped for the past decade (1)(2); some radical
new development will be required before it can resume.
.A
4.1.4
Ves,
...
...
..
.
could be used to reduce the pressure in the high temperature portion of
the cycle and thus ease stress problems (3), and a number of plants based
on a mercury vapor-steam cycle were built between 1917 and 1949. However,
corrosion and mass transfer problems have imposed an upper temperature
limit or 900 to 950°F on the mercury systems so that they have been un-
the
first
ies
r
to
en i
s
aar
l& higher by about 300°F.
There have been numerous efforts to find suitable high temperature
working fluide for binary vapor cycles. The best of the organic materials
are the i1phenyls (e.g., Dowtherm) but these give trouble with thermal
decomposition at temperatures above about 900°F. Sulphur, sulphur com-
pounds such as 80g, chlorides such as Albrz, and many other materials
have been considered (4), but corrosion considerations appear to rule out
all but the alkali metals. Extensive development work on alkali metal
systems for nuclear power plants during the past 15 years indicates that
the prospects are good for operating sodium and potassium in stainless
steel systems in the 1000 to 1500°F temperature range with relatively
iittle corrosioa (5)(6). The vapor pressure of sodium is a bit too low
in this runge to make it attractive as a working fluid, but the physical
properties of potassium at 1500°F are about the same as the corresponding
*-:-
r
-
*
-3-
values for water at 240°F (see Table 1). Many attempts have been made
to find other materials that would be better than potassium for high-
temperature vapor cycles, but only cesium and rubidium appear to have
the desired characteristics (4), and both of these are expensive and
difficult to procure.
A comparison of potassium with mercury, discloses that not only is
potassium far less corrosive than mercury, but it also wets stainless
steel surfaces extremely well (thus avoiding heat transfer problems that
have given trouble with mercury), and costs only about $50/fts as com-
pared to about $2000/ft3 for mercury: Further, the supply of mercury
18 80 limited that the amcunt required for a large modern plant would
represent a major capital investmeut, even if it did not drive up the
market price. The mercury inventory in a 400-Mw plant would be about
3 x 206 lb, or about 15% of the total world production in 1961. (7)
The health hazard represented by the toxicity of mercury vapor in air
18 a serious one - 80 much so that the mercury inventory represents an
atmosphere contamination hazard potential about as great as that a880-
ciated with the volatile P1881.on products in a reactor for a power piant
of the same capacity. That is, it (would take about as much air to dilute
the mercury vapor to standard tolerance levels as would be required for
similar dilution of the volatile fission, products accumulated in a nuclear
reactor in the course of a year of operation.
Potassium 18 not toxic, but it does burn in air. The ignition temper-
ature 18 sufficiently low that it sometimes ignites spontaneously at a tem-
perature just above the melting point. It reacts rapidly with water at low
......
.........
.
....
.........--------
*
**
*
***
n-*
-4
h
ha
ini
a nita
pressures, but experiments have shown (8) that the reaction rate is unifon
te mawi W
s
and not too rapid at temperaturoabove about 600°r. The potassium water
reaction itself 18 not explosive; however, if the hydroger evolved mixeu
with air, the resulting mixture may explode or even detonate. Thus, the
latter constitutes the principal hazard associated with the potassium-
water reaction, and, as in sodium-cooled reactors, the plant must be de-
signed so that the consequences of such a reaction will not be too serious.
de
.rrri diversidad
Worriotin
In the course of the initial ORNL work on space power plants in 1958
it became evident that a high cycle efficiency in a central station could
be obtained by coupling a auclear reactor to a potassium-vapor cycle from
which the waste heat would be rejected to a steam cycle. This arrangement
would achieve the advantages of the mercury-steam binary vapor cycle and
yet avoid both the relatively high pressures and the mass transfer problems
of a mercury system in the 1000 to 1600°F temperature range. It woulâ be
capable of an overall thermal efficiency of arowd sus, thus cutting not
only the fuel and capital costs but also the condenser cooling-water re.
quiremeni. The latter would be reduced to about half that of a conventional
ste en plant, a major advantage for many power plant locations either on
rivers or in arid areas where cooling towers are required.
An attempt was made early to 1959 to lay out a nuclear central station
utilizing a binary vapor cycle, but there were so many uncertainties that
.
.
'..
**
.
..
-
-
...-
further work was deferred. Subsequent experience in the molten- alt renc-
--
»*
tor program coupled with data from basic tests on boiling heat transfer, .
.
tiem
o
n
si
materials compatibility, and key components for potassium-vapor-cycle
space power plants provided sufficient background data to make possible
a
medi
•S
& preliminary design with enough detall to give a fairly good idea of
the sizes of the major pieces of equi prent and of the quantities of fuel,
working fluids, and expensive alloys required for the system. Using thie
information, a design study was carried out in 1963 to investigate the
probiens involved (9). During the past year experience with both the
moltep-salt reactor experiment and component tests for the boiling potas-
silm space power plant have brought ORIL's operating experience with red
not liquid systems to over 2,000,000 hr, giving further grounds for the
belief that a potassium vapor pover plant merits consideration. This
paper is intended to indicate what such a potassium vapor-steam dinary
vapor cycle pover plant night look like 1r designed for construction in
the '1980's, ito advantages, and something of the problems that would have
to be solved before ito design and construction could be undertaken.
PROPOSED SYSTEM
TO Illustrate the problems involved and the size of the major items
of equipment, it was decided to base the design on a molten-salt reactor
with a fuel outlet temperature of 1800°r. While this is well beyond the
present state of the art, it does not seem unreasonable for a plant to
be built in the 1980's. As shown in the flov sheet of fig. 2, the molten
salt from the reactor vould deliver hent through an intermediate heat
exchanger to an inert salt at 1700*7, which would, in turn, de circulated
to heat a potassium boiler that would deliver potassium vapor to. tur-
dipe at 1540°F. A 200° temperaturt drop vould be taken to each of the
tvo mult circuits. The potassium vapor leaving the turbine vould con-
dense at 1100*7, and the heat of concensation vould be transferred directly
-6.
to generate steam at 4000 poi and 1050°F with two reheats at 1050°F. Ibe
otructural material. 10 both ult circuit. would be a refractory alloy
something like the coiubium-1$ zirconium alloy that has been given much
attention in recent years, the potassium Systen vould employ stainless
teti throughout, except for the tubes in the potassium boiler, which
would be made of the refractory alloy.
. The first step in the design study was to make a thermodynamic analysis
of the cycle assuming the system flow sheet shown in Fig. 2. Since «pero
18 removed efficiently in the potasa.ww-vapor cycle, there 1e no incentive
to increase the peak steam cycle temperature wiher than to increase the
potassium-vapor density in the condenser and thus reduce the size of the
potassium turbine and condenser. Thus, the choice of the temperature level
at which heat 16 interchanged between the condensing potassium and the
Supercritical-prestare steun depend. on balancing the cost of small, varya
high-pressure steam equipment against large low-pressure potassium-vapor
equipment. Hough estimmtes indicated that near-vind costs would be obo
tained by condensing the potassium *t 1100°F and transferring the meant to
steak at a peuk temperature of 1050°F.
The temperature-entropy diagram for the binary vapor cycle corresponde
ing to the above conditions de presented in P18. 3, and the principal parti.
the ters are sumarized in Table 2. It was possible to employ the item flow
sheet for the supercritical unit Mo. 2 of the Daystore plact with esantially
no changes. (9,10) Murther, the overall thermal efficiency of the *****
portion or Ebe system could be obtained from that for the madyatove plant
by alloving for differences between the two ay#tum temming from such
factors « the stack losw. and the gover requirements for auxiliaries.
670
PLANT DESION
A rouctor thermal output of 1000 l was taken for the reference design.
several plant layouts were prepared of voich that shown in Mg. 4 appeared
to be the most promising. A wide variety of factors was considered in
evolving a series of layouts which end to that presented in Fig. 4. The
design precopts on which this was based are as follows :
d. The design should be such that any ccaponent of the syster may be
replaced with little disturbance of other parts.
2. There should be sufficient shielding between the reactor and the
Sual-to-inert-salt heat exchanger and sufficient neutron poison in the
latter (to abaord delayed neutroas) to reduce activation of the Inert salt
to a level that would permit limited contact maintenance of any component
in the inert salt system except the fuel-to-inert-salt heat exchanger.
3. Both the fuel pump rotor assemblies and the fuel-to-inert-salt
heat exchanger tube bundles should be designed so that they could be in-
serted into vertical wells with the seul flane in a shielded zone and
contact maintenance activities carried out at the flange, c.8., seal veld.
ing and inspection of the flanges.
14. Differential thermal expansion in the piping connecting the hot
components should be absorbed either in simple bending of the piping or in
toroidal expansion Joints. Stainless steel bellove should be used only in
low-pressure, low-temperature zones, such as those around penetrations in
the containment envelope.
9. All rour fluid system (tuel, inert walt, potassium, and water)
should be desired to provide sufficiunt theran convection to remove
2
ST
.....
.
M
ind i, men NM.
arterheat in the event of complete failure in the power supplied to all
the pumps.
6. The inventory of the fusi, ert salt, potassium, and refractory
alloy structural material should be kept to a minimum to reduce both the
costs and the hazards.
The hcat exchanger sizes were based on 3/8 in. OD refractory alloy
tubing with a 0.040-10.-thick wall. Experience indicates that fouling
of these small-diameter tubeo is not a problem in systems wnere fluid
cleanliness requirements are as stringent as they are for the molten-salt
and potassium systems. It should be mentioned that the bore of the super-
heater tubes in the Philo plant on the Ohio Power Company, a supercritical
pressure stean plant, 1. 0.40 in. (11)
Reactor Core and Heat Exchanger Assembly
As shown in Fig. 5, the reactor core and the fuel-to-inert-cult heat
exchanger assembly are mounted in a 17-ft-diam, 26-ft-high, cylindrical
shell. The core 16 10 st 10 diaweter and 10 ft in neight and 18 mounted
in the lover portion of the shell. It is surrounded with 3 ft of graphite
reflector with the 3-ft-long blocks laid with their axer normal to the core-
rerlector interface so that the heat generated in the graphite reflector can
de removed by fuel Kloving over either the inboard or outboard ends of the
graphite blocks.
Ova :-* - .....
intensitas medicina.
Pel flavt axially upward through the core and then up through a cea-
trul columns approximately 7 At in diameter to the top of the heat exchanger
region. Duel slovs dovmvard and then back upvard over a burtle between
Uutuber in the heat exchanger. It then returna downward in the annular
er en
lemet
b
pro
..
vitara
i tri i
l i do in
the me
mory
sti...nic.
is...
main
region between the heat exchangers and the outer shell through six axial-
flow pumps flanga-mounted in wells above the top of the fuel expansion tank.
The pumps are driven by long overhung shafts with outboard bearings lubri-
cated by the fuel salt. (A pump with a molten-salt bearing of this type
has been operated satisfactorily for 12,000 hr at ORNI.) The heat exchunger
consists of six tube bundles, each about 4 ft in diameter and 6 ft in length.
These units would be larger than the MSRE unit shown in Fig. 6, (12) but
they would be essentially similar in basic design.
The control rods are mounted in the central region so that they operate
through the fuel-channel riser from the core to the heat exchanger inlet.
The control rod drives, the pumps, and the heat exchangers are mounted with
their flanges in a relatively cool zone above a 4-ft-thick shielded region
over the heat exchanger and fuel expansion tank. The layout is such that
any pump or heat exchanger may be removed by vertical withdrawal from the
top without disturbing any other unit. The weight of each unit is such
that it more than counterbalances the pressure forces acting at the flange,
80 only a seal weld 18 required.
A neutron poison is placed between the top of the graphite reflector
over the core and the heat exchanger to reduce activation of the inert
salt in the secondary fluid circuit. The cooled fuel stream returning to
the core around the outer perimeter of the reflector exerts a modest net
radial pressure inward on the graphite blocks, thus tending to force them
together and minimize the clearances between them.
Potr8sium Boiler, Turbine, and Condenser Units
The potassium-vapor turbines look much like the low-pressure steam
turbines for a convential steam power plant. With an 1860-rpm shaft
.
-10-
speed and two double-flow turbines (to keep the size of each unit dova)
the rotor diameter at the outlet end 18 about 10 st to give 80 mW not
electrical output per turbine. From five to seven stages are required,
with the turbine rotor inlet around 6 ft in diameter.
The very large volumetric flow rate of the potassium vapor, both out
of the boiler to the turbine and out of the turbine into the condenser,
made it desirable to incorporate both the potassium boiler and the potas
sium condenser in units direct.ly beneath the turbine in much the same
fashion as is standard practice with conventional steam turbine condensors.
To minimize the temperature gradient in the vicinity of the turbine shaft
seals and bearings and to keep the temperature of the turbine casing down
to 1100°F where its strength would act be much lower than at room tempera-
ture, the condenser was divideú into two sections, with one at either end
of the double-f.low turbine. The boiler region was placed between these
tuo condenser regions. The general layout is shown in Fig. 7.
As in the heat exchangers incorporated in the reactor pressure ves.
sel, it seemed best to design so that both the potassium boilor and the
-
potassium condenser would consist of a multiplicity of tube bundles, each
-
-
--
-
-
-
-
-
of which could be installed by moving it into position along a vertical
axis and securing it in place with a seal weld. Fortunately, the pressure
-
on the potassium-vapor side is nominal in both cases, i.e., about 15 psi
alove atmospheric in the potassium boiler and about 12 281 below atmos.
pheric in the potassium condenser. Thus, the pressure loads on the potas.
sium side are small. In fact, the pressure force 18 of about the same
order as the weight of the tube bundle.
-ll.
A pair of inert-salt pipes delivers salt from each of the fuel-to-
salt heat exchanger tube bundles to one of a set of six tube bundles mounted
in the potassium boiler casing below each turbine. Each of the potassium
boiler tube bundles makes use of about 1000 U-shaped tubes. The U's are
about 5 ft long and Inverted so that each tube bundle can be withdrawn
vertically downward from a flanged joint at the bottom of the boiler casing.
An advantage of this potassium boiler configuration is that the physical
properties of the fluids (13, 14, 15) permit high heat fluxes so that the re-
quired surface area is small and the potassium inventory can be reduced to
a relatively low value for such a large plant, 1.e., about 800 ft3, or
35,000 lb.
'The supercritical-pressure boiler and the reheaters are made up of
U-shaped tubes arranged in cylindrical bundles and mounted in a casing
under the turbine to serve as the potassium condenser. The feedwater (or
reheat steam) enters one leg of a U and flows vertically upward and then
returns downward to the tube-bundle outlet. It is to be noted that there
is no sharply defined phase change in a supercritical-steam boiler - only
a steep gradient in the fluid density. This occurs mainly in the first
portion of the heated length of the boiler tube.
The most difficult set of tube bundles to design was that for the
supercritical-water boiler because the high pressures on the water side
made it necessary to employ very thick-walled header drums and tubes.
While the temperature in the potassium boiler is high, the pressures there
are low 80 that the stress prcolems are much less severe. The close spac-
ing required of the tubes in the potassium condenser region in order to
py
-12-
.
keep the unit sufficiently compact to fit the turbine led to the use of
bifurcated tubes to reduce the number of penetrations in the superheated.
steam header drum anit thus reduce the stress concentration problem there.
The reheater tube bundles have the same tube length but only about
one-fourth of the tube matrix volume of the supercritical-boiler tube
nt artisanaladislacioni
bundles and are located at the in-board ends of the superheater header
e
drums, as indicated in the reduced-scale plan view of Fig. 7. This arrange-
viraram mkomuhitninamaanishm
ment was chosen to facilitate access to the connecting piping and the clo-
sure welds for the various units. The steam pipes out of the superheater
header drums are 10 in. in diameter and are run horizontally just below the
closure flange level to reach a region well clear of the congested space
beneath the turbine. The much larger pipes for the reheater tube bundles
are stiffer and pose more difficult differential thermal expansion problems,
and hence they are run vertically downward from the tube bundles in the
-
-
-
-
-
-
central region. With this arrangement, any one of the tube bundles can be
-
-
.
-
-
-
--
-
-
approached with a specially built machine designed for installing and re- '
moving tube bundles with a hydraulic lift. After severing the connecting
piping, attachment points on the tube bundles would be connected to the
elevator mechanism on the machine, 'and the closure weld would be cut with
appropriate tooling. The tube bundle would then be lowered and removed
from the area, and a new bundle would be installed.
In the layout shown in Fig. 4c the columbium pipes carrying the inert
salt from the reactor and intermediate heat exchanger assembly up to the
potassium-boiler tube bundles are enclosed within a large steel casing 80.
that they can be kept in either a vacuum or an inert gas atmosphere to
...
-
-
-
---
--
-
-
-13-
avoid difficulties with oxidation. A bellows joint 18 required between
the casing around these pipes and the stainless steel shell or the potas-
sium boiler. A removable sleeve around these pipes is required to provide
space for insertion of the boiler tube bundles. The use of horizontal
pipes passing outward from the superheater outlet drums for the super-
critical-water system leaves the floor free for access to these removable
sleeves.
A major advantage of the layout of Fig. 7 is that gravitational forces
aid in scavenging condensate from the condenser and also help to recirculate
the liquid potassium in the potassium boiler. It should be pointed out that
the potassium boiler 18 designed for recirculation with a boiler exit quality
of about 10%. A vapor separator in the vapor region above the boiler 18 de-
signed to remove the liquid potassium and return it to the base of the boiler.
The vapor quality supplied to the potassium turbine should be in excess of
99%. Experience with natural-thermal-convection boiling-potassium systems
in the course of the space program indicates that this configuration can be
built to work as effectively with potassium as it would with water, and it
should give much the same performance characteristics.
Differential expansion in the potassium-vapor turbine casings poses a
major problem, particularly since the vapor inlet region operates at about
1540°F, while the condenser region runs at 1100°F.
The most promising
approach seemed to be to make the portion of the turbine casing that sup-
ports the stator integral with the potassium boiler and couple this into
the bearing housing to give good concentricity in the stator region. The
potassium condenser casing was then arranged outside this and connected
rii
ܗܪܙܫ
to it rigidly only at the bearing housing. While this layout presents a
number of difficult detailed design problems, it appears that one of the
most important, i.e., differential thermal expansion, can be solved with
flexible omega or bellows joints placed between the boiler and the con-
denser casings on the one hand and the turbine casing on the other. With
the oil-lubricated bearings envisioned, the thermal gradient between the
hot portion of the rotor and the 200°F bearings must be kept to a reason-
able value. A buffer zone about two shaft diameters long will probabli
be required io give an acceptable thermal gradient in the shaft. This
zone can be used for the shaft seal. A reentrant section in the housing
will be required to accommodate the thermal gradient between the bearing
and the large-diameter high-temperature portion of the casing.
The shaft seal configuration of Fib. 7 was chosen, in part, because
it seemed to give a good sound design and, in part, because two small potas-
Bium-vapor turbines have been built using essentially similar seals. (One
of these 18 at the De Gas Turbine Division at Evendale, Ohio.) In both of
these, the turbine wheel 18 about a foot in diameter and 18 separated from
oil-lubricated bearings by a close-clearance seal with argon supplied to
the region between the potassium ve.por and the oil. Two bleeds from the
labyrinth are used in each case, one to carry off a mixture of argon and
potassium vapor and the other to carry off a mixture of argon and oil.
Auxiliary equipment 18 being employed in these systems to separate the
potassium from the argon and the oil from the argon, hence the feasibility
of conducting these two operations should also be determined in the course
of the test programs scheduled for the coming year.
---
- ------................
...--------
-15-
While not shown in Fig. 7, two 100-hp potassium feed pumps are located
inside the shell of the condenser for each turbine unit. The l-ft-diam
single-stage centrifugal impeller of each pump is coupled directly to a
2-ft-diam single-stage potassium-vapor turbine wheel. A pump of this sort
with a 6-in.-diam turbine wheel and potassium-lubricated bearings has been
run at ORNL. (A set of parts 18 shown in Fig. 8) These units should be
very reliable, since they would be independent of failures in such compo-
nents as electrical switchgear. Tests have shown that such a turbine-
driven pump unit can be "bootstrapped" by simply applying heat to the
boiler.
Regenerative feed heaters are placed beneath the potassium turbine
between the boiler and the condenser. These units are designed to employ
vapor bled off between stages of the potassium turbine to preheat the po-
tassium feed to a temperature near the boiling point before supplying it
to the potassium boiler. These make it possible to avoid thermal stresses
in the boiler that would be associated with the admission of liquid feed
at a temperature of 440°F below the boiler temperature, and reduce the
size of the potassium boiler from 4 to 8%, depending on the details of
the bleed system used for the regenerative feed heater and the proportions
employed.
RELATIVE COSTS
While tre many novel components in the proposed plant make it out of
the question to prepare a definitive cost estimate, some notion of its cost
relative to that of a conventional coal-fired plant can be gained from
--
DS
..
.------
-
.-
.
...
comparison of the quantities of material required and their unit costs
given in Table 3. In the first place, the stoon syotum should cost the
same as that for a coal-fired plant, except for the steam generator and
the piping connecting it to the turbine. The connecting piping for the
325 Mw Badystone Supercritical Unit No. 2 totals approximately 5400 st
kod weights approximately 1,000,000 lb. In the plant layout of Fig. 4,
the corresponding piping has a total length of approximately 2000 ft
and hence would cost ecout 40% as much. Although detailed data for the
steel in the boiler tubes for the Eddystone plant vere not available,
data were available for a somewhat similar boiler, 1.e., unit No. 5 of
the TVA Colbert plant. This unit has a 2400-p81 straight-through boiler
designed to deliver 500 Mw of gross electrical output. The weight of
the tubing in this boiler is approximately 5,000,000 lb, while the con-
uecting piping weighis approximately 1,600,000 lb. The latter value 16
consistent with the Badystone Supercritical Unit No. 2 ir allowance is
made for the difference in capacity. The stainless steel tubing in the
potassium-condenser, water-boiler, superheater, and reheater units in
the proposed plant has a total weight of approximately 400,000 lb, 1.8.,
about 10% of the weight of the corresponding tubing required for Colbert
unit No. 5. Because of the higher pressure, the tubing weight for Eddy-
stone i nit No. 2 would probably be considerably greater - a 50% increuse
in weight was assumed in mepering Table 3. This drastic reduction in
tube weight required for the proposed binary vapor cycle stems partly
from the higher heat transfer coefficients and higher heat fluxes avail-
able in the potassium condenser, partly from the smaller diameter tubing
.
-17-
employed, and partly frow the fact that only one side of the tubes in the
mediant-heating done in a coul-fired boiler te weed for hent transfer pur.
poses. me dikeest factor 16, of course, the uniform temperature on the
potassium side, which means that there is no danger of a hot spot ronin
and leading to a durn-out condition, and hence the heat exchanger can de
designed for high and fairly uniform heat flux over the entire surface
roa.
mis high heat flux, coupled with the absence of a flov distribution
problem, makes it possible to employ an extremely compact heat transfer
matrix so that the total volume required for the potassium condanser is
much less than that required for the condenser of a conventional steam
turbine of the same output. It should be noted, in particular, that one
of the principal barriers tu heat flow in a steam condenser tube 16 the
relatively low-thermal-conditivity film or condensate. The high thermal
conductivity of the liquid potassium essentially eliminates this barrier
to host transfer. However, the biggest factor in reducing the #126 of
the potassium condenser in the proposed design, as compared with a con-
ventional steam condenser, is the 150°7 san temperature difference be-
tween the condensing vapor and the fiuid being heated, 1.6., over five
times the value ordinarily used in conventional steam plants.
The columbi teine tuo ing required for both the suel-to-inert-salt neat
exchanger and the inert-halt-bc-potassium boiler has a total weight souse-
what lo.. than 100,000 id. CRNL betallurgists estimate that by 1980 the
cost per foot of this tubing will be roughly 20 times that of stainless
steel tuding for the same tube diameter and well thickness. Even on this
besla, Hovever, the cost of the columbi won tubing for the proposed plant
enould me about the new the cost of the steel tubing for a conven.
tioaud coul-fired plant such as unit No. 5 of the TVA Colbert plant.
The columbiu piping betwen the nel-to-inert-alt ment exchanger
and the salt-to-potassium boiler would walion about 30,000 lb. thi. vould
be equivalent in cont to approximately 600,000 id or stainl... steel
piping. Since the wall thickness of this piping need be only 1/6 to 1/2 in.,
Ito cost coupled with the cost of the steam piping in the proposed plant
would still be less than that of the corresponding stean piping for a coni.
fired plant or equivalent capacity.
The cost of welding the tube joints depends on the uterial and #ita
of the ture, the accessibility of the joint, and whether the joint can be
fabricated in the shop or must be welded in the field. A large fraction
of the joints in a coul-rired boiler must be rield-welded whereus all of
the tube-to-hender scints in the heat exchangers for the proposed potas.
Si un vapor plant would be made in the shop. Since the greater part of
the cost of cuch joints lies in the X-ray inspection and the rounder of
Joints that can be included in a #ingle X-my phoco, it was found that
the cost of shop welding the easy-to-ge:-at un
diamter columbium tube.to.
4
Vi-4.
munder joints in their compact arrays actually appears to be las than
the average cost of the shop and rield-velded joints in the large, open
· tube arrays of a conventional steam boiler.
An important area in which the cost of the binary-vapor-cycle plant
proposed should be much less than that for a coowntional coul-fired steam
plant is in the size of the building. The surnace for a conventional coule


-19-
rired plant commonly runs 120 st nigth and requires substantiad amounts
or space, both above and donanth it, and a large floor art. The mving
in the steel required for the building, cond storage bine, wh hoppers,
etc., should much more than compensate for the steel required in the con-
tainment shall around the reactor, and true the overall building cost
should be less for the nuclear plant.
In s om ry, while it is out of the question to prepare . definite
cost comparison at this stage, it appears possible that the capital cust
of potassium-stean dinary-vapor-cycle plant night actually be less than
that of a conventional stear plant.
TEASDILITY PROBLDAS
me principal busic feasibility questions for the system are associ-
*ted with the corrosion of the refractory alloy by both the fuel salt and
the inart alt in the intermediate circuit and the compatibility of the
refractory alloy with boiling potassium in a stainless steel system where
***' transfer night be a problem. In view of the low fuel costs that it
appears possible to obtain with molten-sult reactors, and the promising re-
sults of this study, it appears that the writs of the power plant proposed
are sufficient to justify some research to evaluate its feasibility.
The basic concept is also applicadle to other reactor types, 6.8., a
sodium- or potansiw-cooled funt reactor, and these reise other feasibility
questions. A boiling-potassium reactor 1. a particularly interesting possi.
bility, since it would make it possible to build the entire system of stain-
less steel and to eliminate tbe intermediate heat exchangers if . substantial
.
. wkunt of radioactivity could be tolarated in the potassium turbineswhile,
...1*..
1..'.WN
A
-
in principle, the cycle could we applied to foss1l-fueled plant, air-side
corrosion vould probably while on this impractionblo, particularly since the
poor heat transfer coefficient on the air
lae would make it necessary to
wse a large quantity of expensive high-alloy materials in the potasium
boiler.
In addition to the above questions there are a host of problems uso.
ciated with the detail design of components. Many of these including thos,
concerned with the heat exchangers and pumpe have been wat and otiofac. ·
tory solutions have been found for smaller components operating under:
somewhat less severe conditions. (16,17) Others, suck u thoja mentioned
In the section on the turbine, vill demand a high order of engineering
completance if the development time and costs are to be wept vitain reason-
able bounds.
.... ewaneo..
.....
a
....
dit men det har en stor hoe ho,77
.
A
a n
muw
wenseny
......
.21.
REL ERENCES
1.
The expanding Market for Electric Power, C. W. E. Clarke, and W. P. Gavit,
Mechanical Engineering, p. 483, 1954.
2. Electric Power Generation - Past, Present, and Future, J. B. McClure
and A. O Mellor, Mechanical Engineering, 1956, p. 521.
3.
The Promet Mercury-Vapor Process, W. L. R. Examet, mansactions ASME,
Vol. No. 46, 1924, p. 253.
Properties of Inorganic Energy Conversion and Heat Transfer Fluids
for Space Applications, W. D. Weatherford, et al., WADD Technical
Report 61-96, March 1961.
A Symposium on Materials for Sodium-Cooled Reactors, 1963 Winter Meeting
of the American Nuclear Society; Vol. 9, Nuclear Metallurgy, JMD Special
Report Series No. 12.
6.
Type 316 ss, Inconel, and Haynes Alloy No. 25, Natural Circulation
Boiling Corrosion Test Loops, D. 8. Jansen and E. E. Hoffman, USAEC
Report CRML-3790, June 1965, Oak Ridge National Laboratory.
7. 1961 Minerals Yearbook, Vol. I, Metals and Minerals (except Fuels),
1.8. Department of Interior.
8.
The Reaction of Kak and 820, E. C. King, Technical Report XII, Min
Safety Appliances Co., Contract N9 onr-85801, January 1952.
A Potassium-8tcan Binary Vapor Cycle for Nuclear Power Plants, W. R.
Chambers, A. P. Fraas, and M. N. Ozisik, USAEC Report ORNL-3584,
May 1964, Oak Ridge National Laboratory.
Engineering the Eddystone Plant for 5000 lb 1200-deg Steam, J. A.
Harlow, Transactions ASME, Vol. No. 79, 1957, p. 1410.
First Commercial Supercritical Pressure Steam-Electric Generating
Vnit for the Philo Plant, 8. X. Fiala, Transactions ASME, Vol. No. 79,
1957, p. 389.
12.
12. Molton-Salt Reactor Program Semiannaal Progress Report for Period
Ending July 31, 1964, R. B. Briggs, USABC Report ORNL 3708, p. 141,
Oak Ridge National laboratory.
13.
Forced Convection Saturation Boiling of Potassium at Near Atmospheric
Pressure, 1. W. Horrman and A. I. Krakoviak, Proceedings of the 1962
Him towperature Liquid Metal Heat Transfer Technology Meeting, USAEC
Raport BUIL-756, pp. 182-203, Brookhaven National Laboratory.
22.
web
14. Molten Salt Reactor Program Progress Report for Period from March 1 to
August 31, 1961, USAEC Report ORNL-3215, pp. 132-134, Oak Ridge National
Laboratory.
ze
15.
A Physical Property Summary for ANP Fluoride Mixtures, s. I. Cohen, et al.,
USAEC Report ORNL-2150, August 1956, Oak Ridge National Laboratory.
e
16. Design Precepts for High Temperature Heat Exchangers, A. P. Fraas, Nuclear
Science and Engineering, Vol. 8, 1960, p. 22.
--
Sw
17.
Reliability as & Criterion in the Disign of Space Power Plants, A. P.
Fraas, paper presented at the Joint SAE-ASME Meeting in New York, April,
1964.
-
.
mm..
.
.
-23-
Table 1. Comparison of Physical Properties of Potassium
and Water for Condensing Conditions (Lata from Ref. 4,
Potassium
Water
Liquid
Vapor
Liquid
Vapor
Temperature, °F
Pressure, psia
Specific volume,
1040.0
1.50
0.02269
1040.0
1.50
267.15
115.6
1.50
0.01619
115.6'
1.50
228.65
1170.4
1111.8
283.0
887.4
83.56
1028.14
0.1823
0.1266
0.998
0.43
Enthalpy, Btu/10
Heat of vaporization,
Btu/lb. °F
Specific heat,
Btu/16. °F
Viscosity, lb/ft.hr
Thermal conductivity,
Btu/hr.ft•°F
Prandti No., COM/"
Surface tension, 1b/ft
0.37
21.
0
0.0189
.00363
1.42
0.371
0.029
0.012
0
0.659
1.04
0.00321
0.0041
3.82
0.00469
s
-24.
Table 2. Temperature and Pressure Conditions for the
Steam and Potassium Cycles Shown in Fig. 3,
and Estimate of Cycle Efficiencies
Location
Pressure
(pola)
Temperature
Enthalpy
(Btu/10)
Steam cycle
READERSchumanuallNASTAGNE
amatira
4000
1133
1043
263
251
0.50, vacuum
0.50, vacuum
4300
29.0
26.0
2.4
2.14
30.0
1050
786
1050
-705
1050
115
115
558.4
1540
1500
1100
1100
1500
1446.1
1374.0
1535.0
1372.7
1552.9
1111.6
82.9
560.1
1210
1195
1068
Potassium cycle
10
Extraction
22
12
290
-
-
368
----
-
- - -
-
-
-
(H10 - HEX) + (1 - F)(Ex - ,)
H10 - H13
-
Efficiency of potassium cycle =
...
(1210 - 1195) + (1 - 0.0865)(1195 - 1068)
1210 - 368
= 15.6%
Efficiency of steam cycle for coal plant = 40.7%
Efficiency of steam cycle for nuclear plant = (no stack, etc. losses) = 46.2%
Efficiency of combined plant = 0.156 + (1 - 0.156) 0.462 = 54.6%
-25.
.
Table 3. Estimated Boiler Tubing and Piping Requirements
for Both the Proposed Plant and a Coal-Fired Plant
of the Same Net Output
Total
Weight
(1b)
Cost in
Thousands
of Dollars
$
Proposed Plant of 550 Mwe
Steam system
Piping
400,000
Boiler, superheater, and 430,000
reheater tubing
Tube-to-header joints
(134,000 joints)
Total for steam system 830,000
600
700
4,700
$ 6,000
40,000
90,000
800
9,000
600
1 Salt system
Piping
Tubinga
Tube-to-leader joints
(19,000 joints)
Reactor and heat exchanger
pressure vessel
Total for salt system
TOTAL
40,000
800
170,000
1,000,000
$11,200
$17,200
1,700
8,000
Coal-Fired Plant of 550 Mwe
Piping
1,700,000
Boiler, superheater, and
8,000,000
reheater tubing
Tube-to-header joints
(55,000 joints)
TOTAL
9,700,000
2,500
$12,200
Stubing for both the fuel-to-salt and the salt-to-
potassium heat exchangers.
! ....*.*
ORNL - DWG 64-1234

t
wai
e I
m
THERMAL EFFICIENCY (%)
til ti
step ahensyn
e into
m
4000
reiniguiente morti
co
المنهننمنملن علمهم
INLET PRESSURE (psia)
مع
1000,
1000
INLET TEMPERATURE (OF)
.
-.
.-
-
.
-
.
400
1910 1920 1930 1940 1950 1960
1970
- YEAR OF PLANT CONSTRUCTION
Fig. 1. Thermal Efficiency and Turbine Inlet Temperature and Pres-
sure for the United States Steam Power Plants as a function of the Year
of Construction. (Data replotted from refs. 1 & 2)

Oral-Dus tetas
· S, FUEL-TO-SALT HEAT EXCHANGER
S-2, SALT-TO-POTASSIUM BOILER
B, SUPERCRITICAL-PRESSURE STEAM GENERATOR
RV-1, REHEATER I
Rito2, REHEATER 2
FUEL NO. 133: LiF-Befz-UF, -Thf. (67-18.5-2.5-14 mole 2)
SALT NO. 44: NoF-KF-LiF-UF (409-43.5-44.5-11 mole 2)
.
GENERATOR
1540°F
POTASSIUM TURBINE
GENERATOR
VERY-HIGH-
PRESSURE
TURBINE
HIGH-
PRESSURE
TURBINE
LOW- ANO
INTERMEDIATE-
PRESSURE TURSINE
GENERATOR
S-2
POTASSIUM-TO-
STEAM BOILER
RH-2
RH-1
STEAM
CONDENSER
12 psia
VACUUM
1510°F,
26 psic
1050°F
1043 psio
1500 F
29 psio
1050°F
251 psia
1050°F,
40C.O psio
i 1100°F,
2.4 psia
|
POTASSHIA
FEED HEATER
| 786 F,
| 1833 psia
70sof,
263 paio
1600°F
1700°F
EXTRACTIONS
115°F
CORE
558°F, 4100 psic
17000 F
1800°F
FEED WATER HEATERS
LAOLTEN-SALT REACTOR
.
.
.
...
.
.
Fig. 2. Flow Diagram for Binary-Vapor-Cycle Power Plant.
Ontho - OWO 84-1336
0081

SEE TABLE 2 FOR CYCLE CONDITIONS
009)
-
-
BOILER
EXTRACTION
1400
REGENERATIVE
FEED HEATER
POTASSIUM CYCLE .
TURBINE
1200
CONDENSER
TEMPERATURE (OF)
TURBINE
REHEATER 1
TURBINE
REHEATER 2
SUPERHEATER
009
TURBINE
400
REGENERATIVE
FEED HEATER
TO FEED WATER HEATERS
STEAM CYCLE
200
CONDENSER
Lai
S, ENTROPY
ľ Fig. 3. Temperature-Entropy Diagram for the steam and Potassiumi
Cycles.

ORNL-OWS 14-12
VERY-HChe. messine
ANO HIGH Messut NIES
CONTAINMENT VESSEL
OPERATING FLOOR LEVA
GEETH
(110
ARCON
ATMOSPHERE
POTASSIUM TURONES
WO HEAT EXCHANGERS
WATER-
INERT-SAU PIPING OF
REFRACTORY METAL ALLOY
The three
DITIO
uit
!!!!!!!!!!!i!!!
REACTOR
MASCHEXT LEVEL
100000
ED
Fig. 4. (a) Typical Elevation 'mrough 550 Mw, net electrical Output
Binary-Vapor-Cycle Power Plant.
one-tw
.100

MUSTON
anaanING FUEL PUMPS
COXTOL 100 CAIVO

A
CONTAINMENT VESSEL
0
5
10
15
20
25
-----------
..
.
....
.
.
.menu.. no one ....
0
POTASSIUM TURBINES
HEAT EXCHANGERS
100
WO
rduto
0
e
_
0
read their
p
ahendamine
-
O
l
u o
S
Fig. 4. (b) Plan View of Reactor Area.
med
-.
-
-
:
Ornit-ons Women
mous
• Tunas MMO
NA EXOURS
GENERATOR
uu
050 S 20.3
foot -
BORETON

a
GENERATOR
VINTERMEDIATE-
VPRESSURE TURBO
LOW-PRESSURE TURBINE
Verthenomen
ANO SPREME
rumemes
W
Fig. 4. (c) Plan View of Primary Piping. The net electrical output
of the generator driven by the two potassium vapor turbines is 160 MW.
while the total output of the two generators driven by the steam turbines
would be 400 MW.

.

:
ma

.
.
*
.
.
.
TV
.in
11
*
-
.
12
.
.
TE;
>
2
HOUD
.
***
YRE
*
tor':.
leveransen EXPANSION TANK
.
..
.
.IN
·
m
.
.
-
MEAY
EXCHANGER 181 |
overensereno AXIAL now
!
.
-
.
..
.
(N 000 ROUND
I.IS
.
come
- RETURN now
MUS
Hili
NOM
QUEMTION
Fig. 5. Molten-Salt Reactor for Bloary-Vapor-Cycle Power Plant,

.
...
........
: mit
.
12
2
unde
. . omnis ..
.-
e
.: White
.
ace
.
Fig. 6. The fuel-to-Inert Salt Heat Exchanger of the MSRE as it
Looked when the Tube-to-Header Joints were Being Welded.

S.
..
anmowe -
ARGON-BUFFERED
SEAL
TUROINE
B. POTASSIUM TO STEAM BOILER
S2, SALT TO PATASSIUM BOILER
RH-1, REHEATERS
RH-2, REHEATER 2
er
o] EHER CO]
up
w
en
@] EBER @
BEAR
som er
en
Woche
PLAN VIEW OF TUBE
BUNOLE LOCATION DRAWN
TO REDUCED SCALE
BLEED UNES TO
FEED HEATER
- BAFFLE
VAPOR SEPARATOR
norama
i
2.
O
widu
Lund
INERT SALT IN
STEAM OUT
STEAM IN
STEAM OUT
/
WATER IN
INERT SALT OUT
SIDE ELEVATION
feet
END ELEVATION
Fig. 7. Potassium Turbine and Heat Exchanger Unit for Binary-Vapor-
Cycle Power Plant.
........
..
.
...

1
1
.
**
.
11.,
T
ANS
EL
AV
. tiirus who wetu wa
.
V
.
*

t
2-

S4
RA
LA
VE
-
.."-
--
vi
-
--
---
-
.
.
.
+
INCHES
LALELLA
DE LA
110 111
14.3
22
Fig. 8. Paris of a Disassembled Free Turbine-Driven Feed Pump for a
Potassium Vapor Cycle System. .
END



e.
DATE FILMED
5 / 24 / 66







i
.
.