EXPERIMENTS WITH CARAMEL FUEL IN OSIRIS REACTOR DESCRIBED

• Document Type: CREST
• Collection: General CIA Records
• Document Number (FOIA) /ESDN (CREST): CIA-RDP08R00805R000100260005-1
• Release Decision: RIFPUB
• Original Classification: K
• Document Page Count: 28
• Document Creation Date: December 27, 2016
• Document Release Date: January 7, 2013
• Sequence Number: 5
• Publication Date: April 27, 1982
• Content Type: MISC

Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
Ah- NO
1
EXPERIMENTS WITH CARAMEL FUEL IN OSIRIS REACTOR DESCRIBED
Paris REVUE GENERALE NUCLEAIRE in French No 6 Nov-Dec 81 pp
567-576
[Article by Jean-Marie Cerles, Chief of Osiris reactor opera-
tions, CEA; and Ghislain de Contenson, engineer in the Tech-
nology Department, CEA]
[Text] After a presentation of the Caramel fuel to be used in
the Osiris reactor, the authors describe the first experimental
phases of the use of this fuel in the reactor. They then
analyze the behavior of the fuel and its consequences for opera-
tional purposes.
Starting in the 1960s and until the present time, the many
research reactors all over the world have used highly enriched
fuel. The choice of this type of fuel, which offers some
technical advantages, was made possible because the United
States made the necessary fuel available. The technical ad-
vantages of this fuel are quite obvious: the very low amount
of highly enriched uranium required can be placed in a low
density matrix. In the present case, a uranium and aluminum
alloy is used to produce fuel elements consisting of thin
plates. These fuel elements are perfectly suited for producing
the high specific powers associated with the high neutron flows
needed for these reactors.
This dissemination of large amonts of highly enriched uranium
led to a growing awareness of the associated dangers of pro-
liferation, and during the 1970s, a movement developed, tending
to restrict the supply of highly enriched uranium. Then there
arose the problem of converting the highly enriched fuel used
in research reactors into a slightly enriched fuel, with the
upper limit for enrichment set at 20 percent U 235.
While research and development programs for the production
and qualification of fuels that could be used for slightly
26
Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
 Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
enriched fuels were being set up in a number of countries, such
as the United States, Germany, and France (CERCA [Company for the
Study and Manufacture of Atomic Fuel]), in 1977 the CEA [Atomic
Energy Commission] decided to conduct an experiment using slight-
ly enriched fuel with the Osiris reactor at Saclay. The tech-
nology for such an experiment, developed at the CEA, was already
available. This fuel, known by its generic name of Caramel,
consists of uranium oxide strips, whose thickness can vary,
encased in zircaloy and assembled to form plates. By using
this fuel, the problem of increasing the density of the uranium
matrix necessary for the reduction of enrichment was resolved:
9 g U/cm3, compared with 1.6 g U/cm3 in the U-Al alloy.
The Osiris reactor has been using this new fuel for nearly 2
years now (it began in January 1980). This experiment has
provided a very complete and strict test; its duration has now
been long enough to produce meaningful results in the following
areas: operating conditions, consequences for testing, behavior
and monitoring of the fuel.
Figure 1: Elements of a Caramel plate.
(1) _ Couvercle
Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
 Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
Key for Figure 1:
1. Lid
2. Tip
3. Grid
4. Caramels
5. Side strips
I The Caramel Fuel
1.1. Presentation
One result of the various programs conducted in France for the
design, manufacture, and development of nuclear fuels has been
the CEA's development of a uranium oxide fuel (called the Cara-
mel fuel) whose characteristics and performance are essentially
quite different from those of the conventional rod or pencil
type of fuel. The Caramel fuel is distinguished from the con-
ventional type of nuclear fuel by the following features:
a. Its plane geometry;
b. The lack of any free volume;
c. Good contact between the oxide and the protective
shield (absence of any clearance);
d. The splitting of the fuel into a large number of
compartments, all tightly sealed in relation to each
other.
The basic element of the Caramel fuel is in the shape of a
plate, formed by two zircalov sheets which act as lids. Be-
tween these sheets are the strips of UO2, which are themselves
arranged in the cavities of a grid. The plane geometry is an
effective way to convert the disadvantage of:the poor conduc-
tivity of uranium oxide into an advantage, by combining a high
specific power with a low fuel temperature.
Because of their use in water-cc5oled reactors, the components
of the Caramel fuel--the UO and the zircaloy--are very well
known in terms of their physical properties, their behavior
under irradiation, their manufacture, etc.
The special features of the Caramel fuel do require particular
operating conditions. Because of the lack of free volume (ex-
cept for the open porosity of the fuel), the temperature
28
Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
 Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
at which fission gases begin to be released must never be
exceeded. For this reason, it is essential to have excellent
contact between the oxide and its covering from the very start;
this contact should last for the entire life of the fuel ele-
ment. The manufacturing methods described later can ensure the
good quality of this contact.
(1)
(2)
(3)
(4)
FAB CAlq
GRILLES '
(5)
(6)
UF6 EFF~HI
ta+vtT
4
PDJDPd UO 2
PLAQUETIES CARAMEL
(9)
PLAQUES
(7)
TOIES ZRY
(10 ) ,.,?..TACX
(11)
PLAQUES Bolts
cREMAILLERES (13) rn?r
(12)1 o E ?? na.~,~ (14)
(16) t T~ 1sT,.,,,.
3
(17) tTV?,~
r
Figure 2: Manufacturing schematic for
Key:
1.
2.
Enriched UF6
Conversion
PIEDS .VERRLIUS
Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
 Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
3. U02 powder
4. Pressing
5. Caramel strips
6. Manufacture of grids
7. Zircaloy sheets
8. Assembly/soldering
9. Plates
10. "Shelling" [form of soldering]
11. Soldered plates
12. Racks
13. Enrichment control
14. Feet, bolts
15. Assemblies
16. Final controls
17. Delivery
The splitting of the fuel into a large number of tight compart-
ments is a major safety factor. Because of its dispersion, any
possible loss of seal only puts a small amount of fissile ma-
terial in contact with the coolant, which limits any possible
contamination of the primary circuit.
The operating temperature, well below the temperature at which
fission gases begin to be released, is another safety factor.
The first studies done at the CEA were oriented toward such
applications as electricity or heat-generating reactors and
naval propulsion reactors, and these used very thick Caramel
(4 mm), because the specific powers required were limited. The
thin Caramel, which was developed later, has been used in the
Osiris research reactor, for which very high specific powers
are needed (the average value is approximately 1,640 W/cm3, and
the maximum value is 4,300 W/cm3).
Table 1 lists the main features of the reactors which are
operational or planned (such as the Thermos reactor) with
Caramel plate fuels.
Description of the Osiris Fuel
The basic module of the Osiris fuel element of the Caramel
type is a plate 700 mm long, 80 mm wide, and 2.25 mm thick.
Figure 1 lists its various components. The U02 fuel is pressed
into a parallelopiped with a square section of 17.1 x 17.1 mm
and 1.45 mm thick. These strips are arranged in the cavities
of a grid and encased between two sheets of zircaloy. This
unit, fitted with side and end pieces made of zircaloy, undergoes
30
Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
 Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
a series of soldering procedures in order to ensure the perfect
seal of each U02 strip in relation to the outside and in rela-
tion to its nearby strips. These manufacturing procedures are
described later.
Puissance
Epaisseur
Temperature
Pression
Puissance
sp6cifique
Puissance
sp3cifique
RBacteur ( 1)
(MW)
CARAMEL
gaine
rOfrig?rant
6 )
mo
enne
7 )
maximale
(2)
(mm)
(4) (-c)
bars )
(51
y
(w/cm3)
(w/cm$)
Pile piscine (OSIRIS)( 8 )
70
1,45
140
3
1 640
4300
Calogbne (THERMOS] 9 ).
100
2,25
160
11.
275
1 070
ISIS
0,7
1,45
95
1,5
16
43
Table 1: Principal Features of Operational or Planned
Reactors Using Caramel Fuel
Key:
1. Reactor
2. Power (MW)
3. Caramel thickness (in mm)
4. Temperature of casing (?C)
5. Coolant pressure (bars)
6. Average specific power (W/cm3)
7. Maximum specific power (W/cm3)
8. Swimming pool reactor (Osiris)
9. Heat-generating reactor (Thermos)
Clusters are formed of 14 or 17 parallel plates, which are
kept joined by slotted side pieces. They are fitted with a
base or foot to provide a water supply, and a head for handling.
Table 2 lists the main features of an Osiris fuel element.
These features are shown in comparison with those of a highly
enriched uranium fuel element, whose structures are made of
aluminum. 4
The essential differences consist of:
a. The metallurgical nature of the fuel: U02 instead of
U-Al alloy;
b. The nature of the enclosure and structures: zircaloy
instead of aluminum;
Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
 Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
c. The thickness of the plates; 2,25 mm instead of 1.27 mm;
d. The total amount of uranium, which is considerably in-
creased.
Type d'616ment combustible (1)
( 2 ombre
de plaques
Epaisseur
d'une plaque
3) (mm)
Epaisseur du
canal d'eau
(4)(mm)
Charge
en uranium
5) (kg)
Charge
n U 235
(6~(kg)
Element a uranium tres enrichi U-AI ( 7 )
Element standard (8)
24
1,27
2,1
0,420
0,390
Element de commande (9)
20
1,27
2,1
0,282
0,262
Element CARAMEL
(10)
Element standard
17
2.25
2,6
8.6
0,602
Element de commande
14
2,25
2,8
5,3
0,371
Table 2: Comparison of Highly Enriched Uranium Fuel Element
and Caramel Fuel Element for the Osiris Reactor
Key:
1. Type of fuel element
2. Number of plates
3. Thickness of plate (mm)
4. Thickness of water si pp1r_ductp Cmm)
5. Amount of uranium (kg)
6. Amount of U 235 (kg)
7. Highly enriched uranium U-Al element
8. Standard element
9. Control element
10. Caramel element
The amount of fissile uranium (U 235) also had to be increased
in order to compensate for absorptions caused by the U 238.
1.3. Manufacturing and Quality Control
1.3.1. Manufacture of Strips
The manufacturing of the Caramel fuel is done in facilities
belonging to the CEA or to its subsidiaries. Some components,
such as the zircaloy sheets, and cast or machined parts, are
provided by industry. The manufacturing schematic is given
in Figure 2.
32
Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
 Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
The following manufacturing phases can be distinguished.
Production of oxide strips (Caramel). The oxide strips are made
by pressing UO2 powder, using moisture. The process used is
called the DCN (Double Normal Cycle). The pressing is done
under hydrogen, at a temperature close to 1,600?C. The density
of the pressed Caramel is equal to 94 percent of the theoretical
density.
We should mention that the production of Caramel fuels benefits
from the experience acquired in this field with PWR [Pressurized
Water Reactors) fuels.
Deposit of Antidiffusion Barrier
Each Caramel is covered with a layer of chrome, deposited by
cathode pulverization. This chrome acts as a barrier, prevent-
ing the diffusion of oxygen in the zircaloy, which could take
place during the process of soldering by diffusion, which is
also called "shelling," and is described later.
1.3.2. Manufacture of Plates
The following elements are combined to form a plate (Figure 1):
a. oxide strips;
b. zircaloy grid, made by soldering zircaloy wires;
c. a nickel tip, which has an effect on the neutrons;
d. side pieces made of zircaloy;
e. end pieces, also made of zircaloy; and
f. zircaloy sheets used as enclosures.
The Caramel are placed in the grid cavities. The entire unit--
the grid with its Caramel fuel, the side pieces, the end pieces,
and the tip--is placed between tCvo sheets of zircaloy, and sol-
dered to form a tight seal. The plate closing procedures are
done in the following order:
a. After resistance soldering of the various parts, the sides
are soldered to the roller, using a resistance solderer;
Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
 Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
b. The end pieces are then soldered by electron bombardment,
which also places the unit in a vacuum;
C. After the plate is closed in this way, it is transferred
to a diffusion soldering enclosure, where it undergoes
a high temperature (,,%-800?C) and high pressure (= l,000
bars) treatment for 4 hours; this procedure is also
.called "shelling." This process ensures that all the
zircaloy components, especially the grids and sheets,
are properly soldered, separating each strip of UO2 and
ensuring a good contact between the oxide and the en-
closure sheet.
d. This shelling process is then followed by a control
procedure, done in a vacuum at 700?C.
1.3.3. Quality Control
All the materials and components must meet the stringent
specifications: specific control procedures have been estab-
lished, and a quality control organization has been set up.
These controls are done at all phases of production, and cover
in particular:
a.
b.
Measurement of components and assembly
ducts) ;
Enrichment of each fuel strip;
(measurement-of
c.
The quality of soldering between metal parts;
d.
The quality of the contact. between the oxide and the
enclosure.
These controls are done during the manufacturing process in
the following manner.
1.3.3.1. Controls during the Pfbduction of Basic Elements
UO2 strips: measurement (length, width, thickness, density,
appearance, control of chrome deposit).
The tolerable surface density defects have been defined on the
basis of correlations between the dimensions of the defect and
the residual thickness of the sheet after diffusion soldering.
Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
 Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
Zircaloy parts: measurement. Surface condition after abrasion.
1.3.3.2. Controls of finished plates
a. appearance;
b. measurement;
c. control of absence of entrapped gases, by thermal
treatment at 700?C in a vacuum, with the plates free;
d. control of quality of soldering by micrography and
corrosion testing;
e. enrichment controls.
The enrichment control of the Caramel plates consists of
determining the surface content of U 235 by neutron measure-
ment. This control offers a number of advantages:
a. overall control of the various manufacturing tolerances
of the Caramel elements;
1. enrichment;
2. uranium titer in the UO2;
3. density of the pressed UO 2;
4. thickness of the Caramel;
possibility of a systematic control of all the strips
which combine to make up a fuel element;
c. speed of control in relation to conventional methods;
d. exact measurement of the combined manufacturing vari-
ances;
e. control of a characteristice(surface content of U 235)
that can be used directly to calculate the reactor's
performances; and
f. control of the quality of the oxide-enclosure contact.
The quality of the oxide-enclosure contact appears to be an
important parameter in the design of the Caramel fuel. A
method of thermal analysis by infrared thermovision has been
developed.
Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
 Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
1 ) NUMLERO PLAQUE : OS 2645 S T
235U/CY.2 (E - 7.00
(2)CHARGE NOMINALE 90.5 MG
235U/CM2 (. - 0.5)
(3 )ma q MOY.l ESUREE : 91.E MG
Z)
4 )ECART CORP.ESPONDANT : 1.4 Z
ECART I
A
B
2
100
200
300
400
7
--
--
0.4
1.0
1.4
!
1.1 1
2
1.9
1.0
0.8
!
2.8 !
3
1.9
- 0.1
1.9
I
0.5
4
- 0.3
- 0.2
1.0
1
1.4
5
1.9
1.2
2.0
1.3
6
3.0
0.7 !
0.6
1.5
7 !
1.6
1.2
1.6
3.3
6
1.I
2.4
1.6
1.2
9
2.7
2.2
1.1
1.5
10
0.9
2.0
0.0
!
1.1
11
2.4
0.0
0.6
1.6
12
2.1
1.4
2.2
1.2
13
2.7
0.9
1.3
0.6
14
2.1
0.7 !
.I
1.2
IS
1.3
1.3
2.0
!
- 0.2
I6
1.3
1.6
2.1
2.5
17
2.5
2.1
2.3
1.6 1
18
2.4
0.7
2.6
1.4 1
19
2.5
1.1
1.7
2.0
20
0.6
1.1 !
0.8
I
0.5 !
21
1.2
0.7
2.1
1.5
22
2.1
1.2 !
0.9
!
2.1
23
2.4
1.5 !
2.6
1.8 I
24
0.4
1.5
1.1
0.7 !
25
2.2
1.7 1
0.4
2.1
26
1.5
1.7
1.3
1.3 I
27
2.9
1.6 !
0.9
2.7 !
28 !
2.4
0.8
0.7
1.8
29
1.0
1.0
1.1
1.5
30
- 0.1
U.9 !
1.5
!
1.6 1
31
1.2
2.6 1
- 0.2
1
0.1
32
1.3
1.7 !
1.2
0.8 !
33
2.1 -
1.3
0.9
1.0
2.0
2.4 .
ECART 1111( ) - 0.3 (MV- 3.3 : CHka/NOM1NA1E
CHARGE 235U/CM2 PLAQUF. HOFk)GENE (91)
Figure 3: Caramel lates: Example of control of content of
U 235/cm.
Key:
1.
Plate number: OS 2645 S T
2.
Nominal content
3.
Average content measured
4.
Corresponding variation
5.
Percentage of variation
6.
Minimum
variation
7.
Maximum
variation
8
Nominal
content
.
9.?
235 U/cm2 content; homogenous plate
36
Declassified and Approved For Release 2013/01/07: CIA-RDP08ROO805R000100260005-1
 Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
However, this method is not being used now, for the test done
at 700?C in a vacuum can be used to test the plates under con-
ditions which are more stringent than the conditions of opera-
tion in a pool reactor. Until the present, this test has been
found to be highly satisfactory.
1.3.3.3. Control during the final assembly
a. overall measurement of the assembly (by placement in
a standard measurement gauge);
b. measurement of all ducts;
c. control of superficial pollution.
The controls and final acceptance are handled by an organization
independent of the fuel manufacturer. Moreover, a system of
quality assurance has been set up and works with studies, de-
velopment, and testing.
A descriptive file is drawn up for each fuel assembly. These
records contain all the data for the assembly:
a. data concerning the various componenets: source, method
of production, mechanical characteristics, measurements
and weight, chemical analyses;
b. measurement results for each plate;
c. observations after a visual examination of each plate;
d. results of the enrichment control, done for each strip;
e. results of the measurement of all water ducts (a recording
is made for each row of strips).
Figure 3 gives some examples of the data compiled for each fuel
assembly, and contained in this descriptive file.
1.3.4. Manufacturing Experiences
The experience acquired during manufacturing is based on the
development of a large number of assemblies, in addition to
the production of units for experimental irradiations. Table
3 gives a summary of this experience.
Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
 Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
(6)
(7)
PAT (')
CAP (")
OSIRIS
2 Nombre
d assemblages
(3) Nombre
de plaques
576
44
3400
U02
Nombre i (5) (Poidskg)
(4 de plaquettes U02
101 952 1 497
13 72B 192
462 400 2019
Table 3: Assemblies produced
1. Reactor
2. Number of assemblies
3. Number of plates
4. Number of U02 strips
5. U02 weight (kg)
6. Ground Prototype: PWR type reactor used as a prototype
for naval propulsion, located at Cadarache.
7. Advanced Boiler Prototype: PWR type reactor with a
power of 100 thermal 1,1W, also located at Cadarache.
At present, the production capacity is 200 assemblies of the
Osiris type a year.
1.4. Qualification of the Caramel Fuel
1.4.1. Various Qualification Programs for the Caramel Fuel
A broad testing and qualification program for the Caramel fuel
has been undertaken and executed within a broad range of speci-
fic powers and combustion ratios. Its purpose was to determine
first of all the technological limits, then the safe and reli-
able areas of operation of this type of fuel. This program
includes both parametric tests in irradiation test circuits in
test models with a limited number of Caramel elements, and
irradiations of experimental assemblies doneecinrthe Osiris
reactor and in the CAP and CAT p yPe
Following are the major outlines of this program, as well as
.the principal results that have been obtained.
1.4.1.1. The EL 3 Program
A first exploratory program designed to determine the technolo-
gical limits of the fuel was begun at Saclay in the EL 3 reactor
between 1965 and 1970. It worked with 17 irradiation test units
Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
 Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
(including five or nine Caramel each) which were irradiated with
variable combustion ratios and specific powers.
The range of specific power explored extends from 1,000 to 3,000
W/cm3. The enclosure temperature: 280?C to 340?C. The results
of these tests showed that:
a. It is possible to achieve combustion ratios of 30,000 MWj/t
with specific powers as high as 3,000 W/cm3 without dama-
ging the fuel. Beyond 30,000 MWj/t, if this specific power
is maintained, there is a danger of causing the strips to
swell, thus releasing gases. This would not necessarily
cause the enclosures to break.
b. On the other hand, if the specific power is decreased
at this stage of irradiation, it is possible to reach
combustion ratios up to 50,000 MWj/t without any notable
change in the structure of the UO2 strips.
1.4.1.2. Siloe Program
Two irradiations without external pressure were conducted in
the swimming pool reactor, Siloe, at Grenoble, using two small
units with a few test models each.
The irradiation conditions were as follows:
Specific power 1,060 W/cm3 1,500 W/cm3
Duration 245 days and 202 days
Combustion ratio 18,300 MWj/t 25,100 MWj/t
Temperature 100?C 100?C
The results obtained show a very good behavior of the assemblies
in the reactor and a very good appearance after irradiation.
1.4.1.3. Program of Experimental Irradiations in the Osiris
Reactor
These are still experimental irradiations of test units, each
with several Caramel, and should not be confused with the irra
diations of Osiris elements done in actual size. The irradia-
tions were done in an NaK container at a temperature of 300?C
with external pressurization of 140 bars.
Historically, two series of irradiations were done:
a. In 1973-1974, with Caramel thicknesses of 3 and 4 mm;
Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
 Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
b. In 1975-1978, with just a thickness of 4 mm; the objective
was to determine the technological operating limits.
The test units in the first series exceeded 40,000 MWj/t with
a thickness of 3 mm and 37,500 MWj/t with a thickness of 4 mm.
The examinations conducted revealed the absence of any release
of gases and the very good behavior of the separators.
The results cover in particular:
a. The thermal conductance of the uranium oxide as a function
of the combustion ratio;
b. The determination of the extreme limit for the normal
operational temperature ;
c. The extreme temperature of the separators;
d. The operating limits for the two thicknesses of 3 and 4
mm during operation.
Figure 4 gives an example of the operational limit for a Caramel
4 mm thick up to a mass combustion of 50,000 MWj/t U. If we
know the extreme temperature for the fuel operation, it is
possible to transpose these results to other Caramel configura-
tions. So, in the case of the Osiris reactor, the extreme
specific p o,.?aers would be above 5,000 W/cm
Other test irradiations were done in a loop on the Osiris
reactor. Two irradiations were done as part of the Irene pro-
gram, using a pressurized water loop which reproduced the real
coolant conditions. One of them caused a leak detection signal
due to the instrumentation (passage of thermocouples). This was
used for safety studies (Ir-06). The second concerned the study
of deposits of corrosion products. Moreover, power cycling
tests were done, still using the 4 mm-thick Caramel, before
the shutdown of the Osiris reactor in July 1978, on test models
which reached a combustion ratio of 30,000 MWj/t. These cyclings
were done under the following conditions:
a. high level 1,250 W/cm3
b. low level 375 W/cm
-1
c. speed of rise and descent 400 W.cm 3/mm
d. number of cycles 3,634
40
Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
 Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
Puissance specifique W/cm3
irradiation MWj/t
Figure 4: Operational limit for a Caramel 4 mm thick
'Key: 3
1. Specific power W/cm
1.4.2. Carine Experiment: Study of Fuel Shield Rupture in
Osiris
In order to evaluate the safety in operation of such a fuel in
relation to the risks of a release of fission products or fis
sile matter in the primary circuit of the slightly pressurized
reactor, a test of a shield rupture was done under conditions
representative of the Osiris operation. This test was conduc-
ted in the independent test loop of the EL 3 reactor. This
loop, located in the heavy water tank, is a heavy water cooling
circuit distinct from the reactor's circuit. Thus, any possible
pollution of heavy water is limited to this circuit.
1.4.2.1. Test Conditions
The fuel element, which was produced according to the same
process as the standard fuel element, contained 32 strips of
UO2 enriched to 7 percent. The shield de5ect was a circular
hose with a surface of approximately 1 mm . The specific power
during irradiation was raised to 3,050 W/cm3. Cooling was pro-
vided by the circulation of heavy water at a flow rate of 10
m/s. These conditions are quite similar to those in Osiris.
Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
 Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
The fuel plate was placed in an aluminum case channeling the
water to both sides of the fuel plate in order to provide ade-
quate cooling.
The power released in the fuel was obtained by thermal records.
The rise in the cooling water temperature was measured by a
triple probe placed at the intake of the assembly in the reactor
and by three triple thermocouples placed at the outlet. The flow
rate was measured by a meter placed on the circuit.
Detection of the shield rupture was done by two parallel systems.
One used a BF 3 counter which was normally in use on the loop,
while the other was-equipped with a 3He counter similar to the
one installed in Osiris.
Changes in the DND (Delayed Neutron Detection) signal as a
function of time are shown in Figure 5. We notice the presence
of three pseudo-plateaus corresponding to increasing levels of
activity. Their duration is about 5 hours for the first, 30
hours for the second, and 3 hours for the third. During the
duration of the first two pseudo-plateaus, there were bursts of
activity reaching peaks whose amplitude was close to the average
signal.
Taux de comptage ( 1 )
2 Pics
44 4~ 72
1 7 3
96 ieiul
4 jours
Figure 5: Changes in the Counting Rate of Neutrons Delayed
during Irradiation
42
Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
 Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
Key for Figure 5:
1.
Counting
rate
2.
Peaks
3.
Doubling
time
4.
Hours/days
Then there was an abrupt signal acceleration following an
exponential curve.
The time required for the signal to double--30 hours during the
first phase--was only 3 hours during the second phase.
After the third pseudo-plateau, there were recorded peaks of
activity, with the average signal first-changing rapidly and
then stabilizing for 1 hour 40 minutes.
The delayed neutron detection equipment used was representative
of the device installed in the Osiris reactor, which permits a
determination of the time necessary for reaching the detection
threshold of a shield rupture. The results indicate that the
detection threshold is exceeded after the third pseudo-plateau.
Despite the continuation of irradiation in these conditions for
about 2 hours, we found a very low level of radionucleid activi-
ty indicating a fuel leak. Moreover, the examinations made of
the test fuel element after irradiation showed an absence of
any significant change in the opening initially made and in the
underlying fuel.
This test therefore showed the excellent behavior of the Caramel
fuel in the event of a shield rupture, and also the possibility
of detecting such a rupture in Osiris before any significant
contamination of the reactor circuits could occur.
1.4.3. Irradiation of Three Complete Fuel Elements in Osiris
Before the start of service of the Osiris core, three precursor
elements were irradiated in the old core. These irradiations
were done during the first 6 months of 1978.
The first, element put in the reactor was removed during the
shutdown of the reactor for maintenance in July 1978, at an
average irradiation level of about 18,000.MWj/t U. A destruc-
tive examination of this element was made at the Irradiated Fuels
Research Laboratory at Saclay before the fueling of Osiris.
Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
 Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
The other two elements were put in the reactor during the core
fueling with Caramel fuel, for durations of one and two cycles,
respectively, in order to actually have precursors, up to an
average irradiation level of 18,000 to 20,000 MWj/t U.
II Changing the Fuel in? Osiris
2.1. Refueling Schedule
The decision to conduct an operational experiment using slightly
enriched fuel in Osiris was made in early 1977. The fuel chosen
was the only one that would be available in the near future, the
Caramel fuel. For the requirements of this operation, the thick-
ness of the strips was reduced to 1.45.mm, as we said earlier.
The operation itself, with its main features: studies, work on
the reactor, fuel testing, experimental studies, safety studies
and reports, fueling and tests, took place following this sche-
dule:
a. General studies, development of the fuel: 1977;
b. Testing of fuel elements in Osiris: early 1978;
C. Work on the Osiris reactor for the Caramel transformation:
from autumn 1978 to the summer of 1979;
d. Isis safety report: January 1978;
e. Fueling of Isis and start of experimental studies on the
core: October 1978;
f. Thermal studies on the loop in Grenoble: second half of
1978 to early 1979;
g. Osiris safety report: November 1978;
h, Fueling of the Osiris core: September 1979;
r
i. Tests of power increase: December 1979 to February 1980.
Note: The reactor was stopped at the end of July 1978 for the
usual summer shutdown and for some major work, involving re-
pairs of linings of'the reactor (ducts, deactivation tanks),
combined with work required for the change of fuel.
44
Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
 Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
2.2. Adaptation of the Facility
The Osiris reactor was chosen for this experiment, involving
the use of a slightly enriched fuel, because of its high per-
formances. The fuel is thus placed in severe operating con-
ditions which exceed the needs of all research reactors.
It is these conditions, combined with the fact that the Osiris
reactor had no margin of U-Al fuel, beyond the margins required
for safety, which explain to a great extent the need for cer-
tain modifications in the core's primary circuit. We should
recall that the reactor had been originally designed for opera-
tion at 50 MW; the power of 70 DM was not reached until 1968,
after a slight modification of the pumps.
The reduction in the number of plates per element (17 instead
of 24) therefore had to be compensated for by an increase in
the primary flow and thus by a change in the pumps and their
motors.
As the primary circuit had four groups of exchangers and pumps
operating in parallel in the U-Al core, each pump was associ-
ated with an exchanger. Three of these units were used simul-
taneously.
Because of the impossibility of increasing the flow in each of
the exchangers simultaneously with the change in the pumps, a
modification in the principal circuit allowed the four exchangers
to be supplied in parallel from each of the four pumps. A main
collecter now connects the outlet of the pumps to the intake of
the exchangers. Only three of the pumps are used at the same
time; the fourth remainson standby. However, the flow goes into
the four exchangers.
The other adaptations made for operation with the Caramel fuel
were very minor and were limited to a few reinforcements of the
dry fuel storage facility.
The modification of the DRG [Shield Rupture Detection System]
had already been considered for,.the U-Al fuel. In the former
situation, the DRG only gave an alarm signal and only used a
detection system for delayed neutrons. The modification made
consisted of introducing the DRG in the sequence of protective
actions for the signal caused by delayed neutrons (action at
2/3) and adding a measurement of the gamma signal.
Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
 Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
2.3. Experimental Studies
In addition to the irradiations and tests of the fuel, a major
experimental program was used to support the core transforma-
tion:
a. A study of the core with the Isis reactor, starting in
October 1978, a year before the fueling of Osiris;
b. A study of the hydraulic behavior, loss of load and
vibrations, and of the fuel element in a test loop at
Saclay;
c. A study of the thermal behavior of the duct at
Grenoble.
2.3.1. Experimental Study of the Core
This study, done on the Isis, was designed to verify the data
for the core, which are highly important for safety, and to
determine the neutron flows and gamma heating in the places re-
served for the experiments. This was essentially a complete
qualification of the method of calculation used.
In the first phase, the program consisted of:
a. Making a subcritical approach by gradual fueling of the
core, and determining the minimal critical size, then the
available reactivity of the core which was to be opera-
tional in Osiris;
b. Measuring the efficiency of each of the six control rods;
c. Determining the distribution of power inside each of the
fuel elements of the core in order to know, in every situ-
ation, the maximum power to be removed by a water duct.
To do this, the Isis core was identical to the first Osiris
core. The flow measurements in 'the experimental sites which
were not essential for the safety study of the core and for the
compilation of the corresponding records were done in a later
phase.
These measurements enabled us to verify some major parameters
for irradiations: the fast and thermal neutron flows, and gamma
heating.
46
Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
 Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
Flux thermique
(2)
Coeur U-Al
Cceur oxyde
maximum
(1 )1014 n/cm2/s
cycle D 62
Premier
Deuxieme
cycle
cycle
Moyenne des
A 0/0
A 0/0
6 (resp. 4)
emplacemen11
6)
centraux
2,24
1,46-35%
1.33 - 41 0/0
Moyenne des
emplacemenjs
(9)
7
face Nord `` ))
2,42
1,86 - 23 0/0
1.81 -25%
Moyenne des
emplacements (10)
face Ouest
1,73
1,64- 50/0
1,53-12%
Moyenne des
emplacements (10)
face Est (9)
1,36
1,55 + 14 0/0
1.55 + 14 0/o
Table 4: Results of Experimental Core Study
Key:
1. Maximum thermal flow
2. U-Al core: cycle D 62
3. Oxide core
4. First cycle
5. Second cycle
6. Average of six central locations
7. Average of locations on north side
8. Average of locations on western side
9. Average of locations on eastern side
These measurements confirmed the calculations, and gave us a
precise estimate of the changes to be expected under test
conditions. These data have also been confirmed by the ex-
perience acquired during the operation of Osiris.
A summary of the results is given in Table 4 (above).
Rapid Neutron Flow (E > 1 MVe)
The calculations had shown that the rapid flow was affected in
an inverse ratio to the size of the core. This result was veri-
fied with Isis and with Osiris. In fact, we found that the
Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
 Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
average loss in rapid flow was a factor strictly and inversely
proportional to the increase in the core size, when the number
of elements was raised from 39 to 44, for hydraulic reasons
peculiar to Osiris.
The very significant absorption of the new Caramel fuel, caused
by its high uranium content, is expressed, for the same power,
by a significant decline in thermal flow in the system (+,--40
percent for these experiments). However, the spectrum quality
is better (approximately 40 percent gain in the rapid flow/
thermal flow ratio).
The decline in thermal flow in the core has no unfavorable
consequences for the experimental program, as the system is
used solely for its rapid flow:;characteristics.,...
The thermal flow on the periphery is reduced on the average by
8 percent, but here we must also bear in mind the increase in
the core size and the number of sites that can be used on the
external grids.
1"' Heating
Y heating decreases in the system, and this is an advantage'
for the experiments:
a. U-Al core 10 to 15 W/g
b. Caramel core 4 to 8 W/g
2.3.2. Thermo-Hydraulic Study
The measurements were designed to:
a. Determine the loss of load of the fuel element, bearing in
mind the surface irregularities of the plate, caused by the
special Caramel technology." This study was handled by the
Reactor Service at Saclay, using the existing hydraulic
loop for the tests of Osiris elements and irradiations;
b. And to ascertain the absence of vibrations in the standard
fuel element and in the control element.
48
Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
 Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
These measurements were made along with the determination of
the loss of load in order to verify the lifting force of the
elements and of the control rods (resulting from the upward
circulation of the primary fluid in the reactor-core).
A second series of tests conducted at Grenoble by the Department
of Energy Transfer and Conversion concentrated on a'thermal
study of the duct. The results were analyzed jointly by the
Department of Energy Transfer and Conversion and by the Reactor
Study and Applied Mathematics Service.
All the results have shown that the Caramel element behaved much
like the old U-Al element. The loss of load caused by friction
is increased by about 5 percent but, considering this increase
in the friction coefficient, the thermal behavior of the duct
is slightly modified by the Caramel structure. The test results
were analyzed under the code name of FLICA.
2.3.3. Safety Records
Because of the importance of the modifications, a detailed
examination of their implications in terms of safety was con-
ducted. In the very first phase, before the work, a complement
to the Safety Report was issued and, after the fueling, a com-
plete reedition of the Safety Report was released which covered,
in addition to this safety study and the modifications made,
the results of the start of service tests.
This reexamination of safety covered all the documents. However,
the major studies and the most important modifications concerned
the following areas:
a. Core physics: loading of the fuel, configuration,
efficiency of the rods, power distribution, replacement
of the fuel;
b. Primary cooling circuit of the core: new configuration
of the circuit, pumps, exchangers, flows, loss of load,
vibrations;
c. Thermohydraulics of the core at its permanent operational
status, in transitional phases, and in natural con-
vection;
d. monitoring of the first barrier: new DRG system; filtra-
tion circuit; automatic insulation of purification
system;
Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
 Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
e. Study of accidents and their radiological consequences
on the environment.
III Behavior of the Fuel: Consequences for Operation
3.1. Record of Operation'
The Osiris reactor was loaded with Caramel fuel in October 1979.
The power increase test cycle took place in January/February
1980. Since then, and until July 1981, in addition to this
test cycle, there have been 12 cycles of about 4 weeks each,
operating at 70 MW, for a total of 21,600 MWj (see Table 5).
Cycles
Charge en uranium 235 Energie
en debut de cycle fournie
(1) (g) (2) (MWi)
Duree du cycle
(3)en JEPP 70 MW
Irradiation moyenne
des elements sortis
(4) (MWj/t U)
E O
20 733
1 542
22
4 970
E 1
20 938
1 546
22,1
9 460
E2
21 586
1 813
25,9
16 460
E3
21 397
1 864
26,6
18 670
E4
21?.164
1 670
23,9
20 090
E 5
21 555
1 622
23,2
22 230
E 6
21 817
1 650
23,6
24 370
E 7
21 875
1 666
23,8
23 850
F 1
21 482
1 732
24,7
24 300
F 2
21,027
1! 774
25,3
24 860
F 3
21 408
1 938
27,7
24 960
F 4
20 908
1 860
26,6
28 360
F 5
20 480
994
13,7
27 455
Table 5: Osiris Reactor: Cycle Characteristics.
Key:
1. Content of uranium 235 at start of cycle (g)
2. Energy supplied (MWj)
3. Duration of cycle in JEPP 70 MW
4. Average irradiation of elements removed (MWj/t U)
For each of these cycles, Table 5 gives the uranium 235 content
at the start of the cycle, the energy supplied, the duration of
the cycle, and the average irradiation of the elements removed.
For the first core, the average enrichment had intentionally
been reduced to about 6 percent, with fuels enriched to 4.75
percent, 5.62 percent, and 7 percent. The fuel now used is
enriched to 7 percent. Towards the end of 1982, this enrich-
ment will be increased to 7.5 percent.
Sn
Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
 Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
Each element remains in the core for five to six cycles. During
each cycle, there is a partial replacement of the fuel and a re-
arrangement of the remaining elements.
3.2. Conditions of Operation of the Fuel and Statistical
Record
The conditions for the use of the fuel are quite stringent.
In particular, the average and maximum specific powers are well
beyond the values found for pressurized water reactors or for
those possible for almost all the research reactors existing
anywhere in the world:
a.
Average specific power:
1,640 W/cm3 6f U02
b.
Maximum specific power:
4,300 W/cm3 of U02
In order to be certain of the good behavior of the fuel in the
reactor, a program of systematic nondestructive examinations
was undertaken. This entails monitoring the water ducts; this
is done by a stress gauge and a comparison with the measurements
made after manufacture. It covers all the assemblies unloaded
after the first six cycles of operation and half of the assem-
blies removed after the next six cycles; this amounts to about
90 assemblies.
This program, which is now underway, will be completed by
destructive examinations of one of the most heavily irradiated
elements, having reached a mass combustion of 30,000 MWj/t U.
All the measurements done will help us to make an overall
evaluation, statistically representative of the evolution of the
characteristics and of the behavior of the Caramel Osiris
fuel element under irradiation.
Since the start of service in Osiris, the average irradiation
authorized has risen from 20,000 MWj/t U, the value designated
as the first objective, to 25,000 MWj/t U in November 1980, and
to 30,000 MWj/t U in February 1981. These authorizations are
based on the good behavior of the fuel, the systematic non-
destructive examinations made of the irradiated elements removed
from the reactor, and destructive examinations of some elements.
For the moment, the average irradiation level of 30,000 liwj/t
U has, not been reached in all of the elements removed. This
level will be reached when the replacement. elements are enriched
to a level of 7.5 percent.
Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
 Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
At present, 63 elements have exceeded 20,000 MWj/t U; of these,
26 are over 25,000 MWj/t U, and eight have reached an average
irradiation level between 28,000 and 30,000 MWj/t U.
We should add that, for a-fuel element whose average irradiation
is close to 30,000 MWj/t U, the maximum irradiation is between
40,000 to 43,000 MWj/t U.
The behavior of the fuel has been good and no major incident has
caused any serious problems with contamination of the circuits.
The only noteworthy incidents seem caused by problems arising
from the manufacturing process used for the Caramel in this
configuration and for the reactor's conditions of use.
During the power increase test cycle, three elements had to be
removed even before the nominal power level was reached, because
of abnormal rates in the DRG. Later, this sort of incident did
not recur. An examination of one of the elements in hot cells
at the Irradiated Fuels Research Laboratory at Saclay showed
that this was caused by an initial manufacturing defect. After
the first series of approximately 100 elements, the plate con-
trols were improved and the satisfactory behavior of the fuel
made it possible to increase the irradiation levels, as we in-
dicated earlier.
Some bursts of Y,activity in the DRG at the end of 1980 and in
early 1981, which did not release iodes or disrupt the delayed
neutron signal, were probably a sign of micro-leaks in some of
the Caramel elements. Systematic examinations should enable us
to determine the origin, either a shield defect or swelling.
3.4. Consequences for operation
As the few initial incidents mentioned above never caused any
real circuit contamination problems, there have been no severe
radiological consequences affectfing the personnel. On the con-
trary, the individual dosimeter readings show a general decline
in the doses received by the pperating personnel, in relation
to the period when the highly enriched U-Al fuel was used. This
is caused essentially by the replacement of the aluminum shield
by a zircaloy shield, which has meant a very large decline in
the 24Na activity of the water in the primary circuits and
capacitors.
52
Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
 Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1
The performance losses in neutron flows found in irradiations
confirm the experimental studies on Isis and the preliminary
calculations. The consequences are quite limited for the test-
ing program. The reduction of the fast neutron flow by about
9 percent has no disadvantage other than lengthening the dura-
tion of irradiations. The notable loss in thermal neutron flow
inside the core causes no problems, as the corresponding sites
are used for rapid flow irradiations. Externally, the average
8 percent loss in thermal neutron flow is compensated by the
adjusted position of the irradiations and the increase in the
number of high-flow sites.
This demonstration of the use of a slightly enriched fuel in
a high-performance research reactor is a technical experiment
of interest from all points of view.
Generally speaking, it seems that a research reactor, even
one with performances as high as those of Osiris, does not
necessarily need a highly enriched fuel. Of course, this re-
duction in enrichment is accompanied by a certain modification
in the characteristics of the core in terms of the neutron
flows. These modifications have not significantly hindered
most of the irradiations.
In the particular case of the oxide plate fuel, known as Caramel,
its behavior is quite satisfactory. The few defects found have
been minor and the reactor can be used with a completely normal
regularity. The change, from a radiological point of view,
has been of benefit for the personnel.
This demonstration experiment successfully conducted with
Osiris gives France an alternative solution for the highly
enriched fuel used in its multipurpose research reactors. It
strengthens France's position for exporting research reactors,
as international trade in highly enriched uranium may in the
future be tightly restricted.
It is certain that, for most of the existing reactors and
announced projects, the irradiation conditions for this fuel
would be much less stringent than those that have been tested
at Saclay for almost 2 years.
7679
CSO: 5100/2109 F,ND
Declassified and Approved For Release 2013/01/07: CIA-RDP08R00805R000100260005-1 --