ML20058J360
| ML20058J360 | |
| Person / Time | |
|---|---|
| Issue date: | 06/29/1990 |
| From: | Amaev A, Shao L NRC, UNION OF SOVIET SOCIALIST REPUBLICS |
| To: | NRC |
| Shared Package | |
| ML20058J272 | List: |
| References | |
| JCCCNRS-WG-3, NUDOCS 9012020042 | |
| Download: ML20058J360 (11) | |
Text
,
o 1
I MEMORANDUM OF MEETING FOR WORKING GROUP 3 RADIATION EMBRITTLEMENT OF THE HOUSING AND 4
j SUPPORT STRUCTURES, AND ANNEALING OF THE HOUSING i
i i
The main topics discussed by the USSR and the U.S.
participants, plus'the important points emerging from those discussions were as follows:
TOPIC 1:
ANNEALING OF THE REACTOR VESSEL Scoce and Content of Discussions The discussions revealed details of annealing of Soviet VVER-440 reactor vessels and of determining vessel properties i
after annealing as a means for continued operations.
Other discussions concerned research on the parameters of annealing and of new proposed U.S. regulatory guidance for annealing, i
i U.S. -- Studies.in the U.S. of irradiation annealing (IA) and irradiation-anneal-reirradiation (IAR) have determined that submerged are weld materials (high copper and high nickel) reembrittle rapidly with fluence after 454*C or 399'C-168 hour annealing but retain a beneficial effect of IAR up to high fluence.
Notch ductility-and tensile tests independently show this behavior.
Residual embrittlement'(postanneal) appears to be a function of the specific material but not a function of the pre-anneal condition.
IAR behavior was found to be independent of the welding flux type.
IA tests of the Gundremmingen archive material showed full notch ductility recovery by 464*C annealing but not 399'O annealing.
Anomalously hign Jcta scatter was observed in fracture tests of the annealed conditicn.
USSR -- Nine reactor vessels have now been annealed, two cladded, and seven without.
A Bulgarian plant will be annealed in 1992, and the Czech Bohunice plant also in that time frame.
A remote macrohardness tester has been developed.
Boat samples measuring 5x60x70 mm are now being taken out of vessel walls to check pre-post properties.
Annealings are done in 24 days.
The USSR has developed a series of models for predicting reirradiation after annealing, which also predict U.S. data rather well.
Anreements and Conclusions l
It was agreed that annealing of the reactor vessel is an important topic for further discussion and research and that it would be beneficial ~for a U.S. delegation to witness the 1
9012O20042 900823 PDR REV9P NROUBUSR PDR i
%:~.
z ---
=.:..-
j 1
i l
annealing of a VVER plant during the next year.
Details of j-arrangements are to be worked out between the two sidt.s'.
The USSR has demonstrated the annealing of nine VVER-440 I
reactor vessels, including two clad vessels, and has developed remote computer controlled hardness testing equipment plus a process for removal of a 6x60x70 mm material sample for annealing verification.
The entire annealing procedure is conducted in 24 days.
TOPIC 2:
RESEARCH ON REACTOR VESSEL MATERIALS FROM REACTORS TAKEN OUT OF OPERATION Seone and content of Discussions The discussions revealed details of studies done on the Novovoronezh Unit 1 AES after 20 years of operation.and on the Shippingport Atomic Power Station neutron shield tank and'the J
Gundrammingen KRS-A react'or vessel after decommissioning.
U.S. -- Although the operating temperature is outside the scope of Regulatory Guide 1.g9, the characterization of material from the Shippingport neutron shield tank indicated that.the embrittlement of the steel is consistent with predictions based on NRC regulatory guidelines.
The Charpy-V notch impact energy tests revealed that the transition temperature shifts are lower than those expected based on results obtained from the High Flux Isotope Reactor (HFIR) surveillance samples.
The reasons for this difference are not well understood but may be related to the high thermal flux in the HFIR.
The shield tank weld metal is significantly tougher than.the plate material, and the weld specimen test results are strongly influenced by the locatio.1 of the specimen from within the tank wall.
Only specimens tak9n from the inner regions of the weld show embrittlement.
USSR -- Trepans were removed from the Novovoronezh vessel; a l
K shift from 0.6 CT specimens of 90'C was measured wk.ile a u
160*C shift (48-J) was measured with Charpy specimens from the same layer but from a different trepan with the same fluence.
Material was removed from the outer surface of the vessel and I
heat treated at 660'C-2 hrs to recover unirradiated properties.
Microhardness measurements showed no diffusion of Cr and Ni from cladding into base metal after 20 years operation at Novovoronezh.
Nikolaev has developed a correlation for predicting annealing recovery as a function of annealing and irradiation temperature.
Anreements and Conclusions Both sides are satisfied that "unirradiated" material properties can be derived by proper annealing of low fluence 2
e l
G e =D + * ** O e* gem eum,
,e g
--_____-_...-,,_.--,,.,,----,---,,,-.,v,v.-,
- ~ -
l I
\\
l.
vessel materials.
The USSR is satisfied with results from Novovertneth studies, whils the U.S.
is still studying data from l
the Shippingport NST pred Gundrammingen forg'.ng materials.
4 TOPIC F.:
RADIATION EMSRITTLEMENT OF MATf. RIALS OF THE VESSEL IN MODEL TYPE VVER-1000 Scoca and Content of Discussions i
Information was presented concerning embrittlement characteristics of VVER-1000 vessel materials following exposure in both power and test reactors.
USSR -- The 16x2MHWA material for the VVER-1000 reactor j
pressure vessels has considerably higher Ni content than the VVER-440 composition, and the VVER '4000s operate at 230'C vs.
270'C for the 440s.
At contents greater than 1.3% Ni, a distinct increase in embrittlement occurs at high fluence.
Ni becomes a much more radiation sensitive elenant, probably in cooperation with Cu.
Finer grain size lowers the initial transition temperatures, but fine and coarse grained steel show about the same radiation increases.
It is now suspected, but not yet-proven, that P does not have as significant an effect in the improved higher purity steel.
Phosphorus content in the first generation of VVER-440 steels tends to be much higher.
A relation has been developed to predict VVER-1000 material re.diation embrittlement, based on test reactor experiments.
]
l Anreements and Conclusion's With the higher' operat'ing temperature of VVER-1000s vs. tha
-440s (290 vs. 270'C) and Ni addition of as much as 1.8%, the J
VVER-1000 composition 15x2HMWA steel shows much more similarity to embrittlement of U.S. RPV steels in that Cu and Ni interactions are the most dominant contributor, rather than P content.
TOPIC 4:
RADIATION EMBRITTLEMENT OF REACTOR VESSEL MATERIALS AND APPLICATION TO OPERATING REACTORS Scope and Content of Discussions Information was presented concerning irraciation effects on crack street and vessel cladding. on experimento1 fluence rate studies, and the application of VVER-440 materiais' properties in the irradiated condition to operating plants.
U.S. -- The irradiation-induced temperature shift of the crack initiation toughness (N,) curves for high-copper submerged-arc welds is somewhat greater than the temperature 3
~
~_.
i i
shif t indicated by the Charpy impact test results ai, the di Joule
. energy level.
However, preliminary results indicate that the irradiation-induced temperature shift of the crack arrest toughness (q,) curve for high-copper submerged-arc. welds is the i
same as the temperature shift indicated by the Charpy impact test j
results at the 41 Joule energy level.
l Irradiation at 244'C reduced the ductile crack initiation and ductile tearing resistance of type 304 toughness (Ja)l weld overlay. cladding to levels comparable with' i
e stainless stee those f rom a high-copper, low Charpy upper-shelf energy l
submerged-arc weld.
J Exploratory stuaien indicate that the amount of change in i
j mechanical properties by a given neutron fluence is dependent on fluence rate in some but not all cases.
The depencency appears j
to be a function of fluence level (low vs. high) and material composition or material type (plate vs. weld), or botn composition and type.
Charpy-V and tensile test results are in j
agreement on the direction of too fluence rate effect; a general i
agreement of.Charpy-V and J-R curve test assessments of fluence i
4 i
rate was also found in exploratory tests.
Recent fluence rate investigatione of Gundrammingen archive and vessel trepan materials have not fully explained prior anomalous test results from these materials.
USSR -- Values hhvo been developed for the coefficients of the embrittlement prediction equation.
There cannot be one l
single analysis to cover'all ranges of composition or fluxes.
At i
both low and high fluxes at 270*C for VVER-440 steels, F is deminant, but there'is a strong P and Cu interaction.
This j
interaction is dominant except when both P and Cu are low.
A remotely operated, automated and computer controlled ball-indentation hardness tester has been developed to verify t
l embrittlement recovery of annealed vessels.
Correlations have l
been developed with ultimate tensile stress.
Brinell hardness was measured using a 3 mm ball for irradiated material at about 234 versus about 200 for annealed; this converted to about 69%
recovery.
No machining of the unclad vessel surface was necessary.
Such measurements will not be done on cladded vessels.
The machine has been used on the Greifswald Unit 2 reactor.
Studies are under way to convert the measurements to stress-strain curves.
Anreements and Conclusions it was agreed that the effects of irradiation on vessel cladding are important and that more work needs to be done.
The study of materials from decommissioned materials is considered by both sides to be very important.
Research should be continued in 4
.,,_en-m.
-~-m---,r
, - -, - ~
..---,r-~..-,,-.,,.~,-_,---.-..-.,....m w.
~--
a.-,
T i
I I
as many areas as possible, especially fracture mechanics, flux rate, structural analysis, etc.
4 i
TOPIC 5:
MECMANISMS OF RADIATION DAMAGE OF VESSEL MATERIALS i
Scone and Content of Discussions Some issues concerning the modeling process for radiation j
damage mechanisms, the influence of microstructure and fast i
neutron flux density, as well as other recent results and j
generalizations of mechanisms were presented.
U.S. -- A number of key mechanisms which. control embrittlement have been at least semieuantitatively determined.
l These include hardening due to formation and growth of i
precipitates, temperature shifts of the fracture toughness due to hardening, and radiation-enhanced growth of copper-rich precipitates.
Other key mechanisms have been qualitatively i
identified but are not as well understood.
These include various l
1 features associated with the material matrix and certain j
independent as well as synergistic effects of nickel, effects due to neutron flux and neutron energy spectrum, effects of annealing and reirradiation, and relationships between mechanical properties and micromechanics.
y For more reliable predictions of embrittlement, specific l
activities appear to be needed.
These include well-designed and well-controlled experiments which will allow for separation of l
variables r,d identification of' specific mechanisms, use of l
l state-of-the-art analytical equipment to characterize the l
microstructural features, and development of large-scale computer i
simulation models which incorporate the known kinetic and thermodynamic theories.
j USSR -- Modeling processes are being developed to predict i
radiation embrittlement; the model used for copper precipitation j
is that of Odette.
No specific. Soviet experiments have beer done to confirm the model for VVER-440 or -1000 steels.
Good agreement is found with American results on the modal.
An issue for the future is if the model will also work for annealing and reirradiation predictions.
l An extensive TEM and SE fractography study has.been done on VVER-440 steels from Novovoronezh and test reactor irradiation.
I A case was shown for ductile tearing of a fine grain steel at l
temperatures below the critical temperature.
It is reported that no temper embrittlement occurs for~ annealing at 500'C for normal j
annealing times, and perhaps so for multiple anneals.
e h
5
.m
-x
- s,_
a
)
h Apreements and conclusions i
The U.S. and the USSR have obtained good agreement using the Odette semi-empirical model for prediction of radiation l
embrittlement of U.S. steels; studies with USSR steels are yet to be Mone.
There is agreement that Copper precipitation is the dominant mechanism for radiation embrittlement of U.S. steels, and that the model needs to be expanded for prediction of annealing and reirradiation behavior.
i i
TOPIC 6:
THERMODYNAMICS OF THERMAL SHOCK, ELASTIC-PLASTIC 4
FRACTURE MECHANICS AND VALUES OF FAILURE PROSASILITY OF THE VESSEL Scoon and Content of Discussions i
Information was presented and di wucoed concerning i
t thermodynamics and scenarios for e.;cidents under PTS conditions, l
analysis methodologies and risk assessments for PTS, and evaluation of the resistance to brittle fracture of vessels under PTS conditiona.
+
U.S. -
'.'n the U.S.,
PTS transients are a potential threat 4
to PWR vessel integrity and therefore must be considered in an
{
evaluation o' vessel life expectancy.
An integrated i
probabilist.c evaluation of the PTS issue, similar to the j
d NRC-sponsored IPTS studies conducted in 1985 for three specific j
PWRs (Oconee, Calvert Cliffs, and HB Robinson), appears to be a i
reasonable approach.
l However, the results of the IPTS studies indicate that the i
screening criteria included in the NRC PTS Pule may not be i
appropriate for all PWR vessels in the U.S.
This is because i
small dif ferences in plant design and operat.ing procedures can make a big difference in the calculated freauency of vessel failure, and because only three PWRs were r,ubjected to a l
specific-plant analysis.
In fact, results of the IPTS study might be changed significantly by considering new information, including a new radiation-damage trend cu'*ve (Regulatory Guide 1.g9, Revision 2), shallow-flaw effects, the effect of stratified flow on the potential for propagation of l
circumferential flaws, and cladding effects.
USSR -- The review by the U.S. experts of the USSR analysis f
of the HB Robinson HYPO" vessel under PTS loading was discussed.
Areas of similarity and areas of difference were identified and l
discussed.
l The USSR is now planning to build vessels capable of
(
operation up to 60 years without an enneal.
Steel of improved quality will be used.
It is being done on an industrial scale.
l t
i l
f t
_. -, -.- -. - _ ~
.. -. - -. -... ~. -...
~..
(
l l
An entire new analysis is under way, including fatigue crack J
grow'h, fatigue, nozzle material serviceability, and fracture
' toughness.
The methodology for vessel strength under PTS was discussed; many similarities to the U.S. are apparent.
Model i
vessel tests for thermal shock are being conducted.
A series of j
screening criteria were developed for determining annealing times for VVER-440 reactors.
The criteria are related to temperature
- of injection water and type of transient.
The highest criterion for application to the'model 230 VVER-440 reactors having i
significant system' modifications is for 193'C (379'F).
i j
Anreements and Conclusions USSR considers degree of mixing and stratification i
associated with injection of emergency core coolant when performing a fracture analysis.
They have developed their own a
l analytical mixing models and have conducted a series of thermal hydraulic experiments to investigate the validity of the models.
L The experiments indicate that for their particular case (geometry, etc.) circumferential and radial variations in coolant temperature at the core bottom elevation are " negligible."
At higher elevations they are not negligible in terms of calculating the nominal axial stress component in the vessel wall associated i
i with the circumferential variations in coolant temperatures.
This strees is important in the calcu'" tion of K for circumferential flaws.
The USSR mixing experiments have included one-seventh-scale low-pressure, ons-seventh-scale high-pressure and full-scale (real reactor) fluid-flow configurations.
Numerous thermocouples 4
were used for measuring coolant and surface and each incluaed a typ'. cal six-nozzle arrangement.
For the full-scale tests, both high-pressure and low-pressure injections were considered.
Theofanous indicated that he was preparing a report on mixing that thus far includes "all" of the significant available mixing-experiment data, with the exception of the USSR data.
He invited the USSR to provide their data to him for inclusion in his report.
(Theofanous is particularly interested in the full-scale-facility data.)
The Soviet side will provide a short description of the full-scale tests.
PTS is evaluated using a deterministic approach with
" worst-case" transients.
Event trees are used to define transient but not all branches are considered.
As a result, the defined transients are not as severe as the most severe U.S.
transients, but the USSR determinist.ic approach tends to be more censervative because the most severe U.S. transients tend to have relatively low frequencies of occi.rrence.
Flow stagnation assumptior. is considered to be a significant feature in the VVER's for the case of steam-line breaks.
T l
t
-m.
,.,..-_-,.r,..
.,,..1**.*"_"L**
m
.,c..*..m.,..
.,,,,y,,,
.r
..,,~__.._,,.m,_
__,_m..,
1 i
l i
?
1 1
A comparison of U.S. and USSR probabilistic l
fracture-mechanics analyses is attached, j
l
(
l l
PROPOSED ACTIVITIES FOR COOPERATION, WORKING GROUP 3 j
l 1.
Irradiation of U.S. and USSR Steels.
Test specimens, including Cha"py-V, tensile and 0.6 TCT, will be provided of U.S.
pressure vessel steels and of USSR VVER-440 and -1000 steels for i
irradiation in the other's test and power reactors.
Testing will be done in the country of irradiation.
Results will be compared.
2.
Surveillance. claddina and Fracture Touchness Data.
Data summaries and data bases will tMe exchanged on reactor vessel surveillance data, cladding data, and on fracture toughness data.
3.
Study of Decommissioned Reactor Materials.
USSR will j
provide small samples of vessel material irradiated in the i
Novovoronezh and Armenian VVER-440s for microscopic study by Odette at UCSB and by MEA for micro-mechanical properties.-
J 1
4 Rwactor Vessel Stress Analysis' Validation.
The USSR will i
provide detailed measuremer.cw wf stress, strain, and temperature i
for a VVER-1000 reactor vessel, as well as detailed dimensions I
and appropriate mechanical properties; the USNRC will perform a stress analysis of the VVER-1000 vessel and validate the
{
calculation by th6 measurements.
i 5.
Vessel Failure Proba'bility.
The U.S. and USSR will continue l
analyses of reactor vessel failure prcbability using the
{
HB Robinson HYPO benchmark as the pr. nary item for discussion and comparison.
The U.S. will consider the possibility of conducting probability failure analyses of HB Robinson (real case), NORD-3 I
and new generation VVER-1000.
6.
In Situ Vessel Annealina, s.
The U.S. can witness the conduct of annealing of a Soviet VVER-440 reactor vessel at the earliest available time.
b.
The U.S. Will cooperate to help provide appropriate instrumentation for stress, strain, and temperature measurements during a future anneal of a Soviet VVER-440 reactor vessel (preferably a.,
above) and will perform an independent thermal and stress analysis of the plant.
c.
The USNRC and the USSR will both
?opare draft regulatory guidance for the conduct of annealing of reactor vessels.
These drafts will be exchanged,-reviewed, and critiqued ol,_
by the other side prior to formalization by both countries for l
their own regulation.
8
7.
Besator vessel Intearity.
Data and existing reports will be exchanged on experimental studies of vessel integrity, especially l
for a riange of loading conditions.
Based on evaluation of the 1
data by both sides, future cooperation or exchange in this area l
may be proposed.
The U.S. will continue to seek to gain the i
entry of the USSR into the Fracture Analysis Group of the
{
1 OECD-CBNI Principal Working Group No. 3 to further this cooperation.
S.
Inhomoneneity Data.
The U.S. and the USSR will exchange 1
available data on inhomogeneity of plate and forging material, especially as this is caused'by manufacturing processes.
l g.
Thermal Mixino Models.
Blind test evaluations will be performed of thermal mixing models.
The U.S. will provide the i
Perdue. CREARE 1/2-scale model and HDR benchmarks for USSR analysis, and the USSR wil1 provide their full-scale VVER thermal mixing test for U.S. analysis.
Results will be exchanged at the i
next annual meeting.
t i
t i
i l
i
?
i i
l l
r l
i i
o L
g
\\
l
i Working Group 3 Memorandum of Meeting signed in Moscow, USSR.
1 i
U.S. Co-Leader USSR Co-Leader Oh 19,(770 Ag4$,(pf Lawrence C.' shao
(/
Amir D. Amaev f
d 4
5 i
i t
f i
t
?
5 l
i i
3 e
.. ~...
-,m5+
g-w_,
->#-.--y
.,--.y.
e-*-
-w--
w-ew-w mw,w--wei-V v'-+s-r-s-n--
+
j 2
l l
WG-3 ATTACHMENT j
I
\\
l Ggmnarison of U.S. and USSR Probabilistic Fracture-Mechanica i
l Analyses
{
The USSR performed probabilistic fracture-mechanics i
calculations for 21 of the HS Robinson transients included in the U.S. Integrated Pressurized Thermal Shock (IPTS)' Program.
Ratios of U.S. to USSR calculated vales of P(F/E) ranged from 0.1 to 45, with most being in the range of 2 to 5.
Specific aspects of analysis that probably account for the j
differences are as follows:
a)
Calculation of stress-intensity factor.
U.S. used a detailed finite-element analysis, while USSR used ASME approximats approach.
b)
RTm, was simuleted and used as an independent variable by the U.S. While the USSR simulated Fo, Cu, and RTag.
J c)
Soth two-and thr0e-dimensional flaws were considered I
by the U.S.
It appsars that the USSR used three-dimensional flaws but with different aspect
}
- ratios, f5 d)
The particular comparison made by the U.S. used a ffaw I
l density as deterministic value equal to 1 flaw /#, but the USSR used this densit I
i the mean equal to 1 flaw /#y as probabilistic value with e
e)
The calculation of P(F/E) was performed using Monte l
Carlo techniques, while the USSR used a closed-form approach.
f)
Initially there was s'significant difference in the calculated wall temperatures.
However, the USSR improved their thermal model and now obtain temperature essentially the same as those calculated by the U.S.
Sased on sensitivity studies conducted by the U.S. in connection with the original IPTS studies, it is concluded that the differences in the U.S. and the USSR values of P(F/E) can be explained by the differences in the analytical models.
1 1
4
...~.
~,.,.e..