ML20058J534
| ML20058J534 | |
| Person / Time | |
|---|---|
| Issue date: | 07/09/1990 |
| From: | Serpan C NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES) |
| To: | Shao L NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES) |
| Shared Package | |
| ML20058J272 | List: |
| References | |
| JCCCNRS-WG-3, NUDOCS 9012020104 | |
| Download: ML20058J534 (10) | |
Text
_ _,... _
_... _ _ _ _ _ _ ~ -
r.
>D 780
/
UNITED STATES i 7 NUCLEAR REGULATORY COMMISSION c
{
W ASHINGTON, D. C. 300M I
-l JUL 0 e 1990 l
NEMORANDUM FOR:
L. C. Shao Director DivisionofEngineering i
Office of Nuclear Regulatory Research.
l FROM:
C. Z. Serpen Jr., Chief i
Materials Engineering Branch Division of Engineering, RES
SUBJECT:
REPORT OF TRAVEL TO MOSCOW i
During the week of June 25-29, 1990, I participated in meetings of Working Group 3 of the JCCCNRS exchange between NRC and the USSR. A copy of the signed-Memorandum of the Meeting, along with sumary nGtes taken during the meeting are enclosed.
Some of the more interesting items learned included the following:
Nine VVER-440 reactor vessels will have been annealed by the end of June 1990.
Advances in annealing validat'on include the development of a remotely operated and computer. controlled macro-hardness tests for use on unclad vessels, and development of a process to remove a 5x60x70-sn piece of vessel material for direct property measure-ments.
The Soviets aave reduced the phosphorois content in the "second generation" of VVER-440 vessel ster.*is, as well as in the 15x2HM A composition fo* the yyrR-1000 '.essels. The higher nickel content (61.8%) now makes Ni a much more radiation sensitive element, probably through interaction with Cu, as is typical for US pressure vessel steels.
Careful evaluation of the US and USSR calculation of probabalistic fracture mechanics of vessel failures showed that the differences could be explained by differences in the analytical models. For example, we use Monte Carlo calculations while the Soviets use a sonewhat simplified closed form equation 9012O20104 900823 PDR REV9P NRG R-f n~
--=..
,-.e
Items proposed for future cooperation should be quite beneficial to NRC, but will not be cost-free.
Forexample;(1)weproposedtoexchangeandirradiate each other's RPV steels; this way, we gain important information on the effect of P at lower Cu content. Most importantly, EPRI, through their meeting participant, Tim Griesbach, have offered to irradiate Soviet steels in three different power reactor surveillance capsules.
(2)Weproposedtoinstrument a VVER-440 reactor vessel that is to be annealed to validate a US temperature and stress calculation of the vessel. Here again, we believe that EPRI and its collaborators (DOE, industry) will contribute heavily to this effort.
(3) We propose to make a detailed stress analysis of a VVER-1000 reactor vessel for validation by Soviet stress, strain and temperature measurements.
The working and personal relations between US and Soviet personnel continues to get better, so that cooperation is expected to continually improve and be more beneficial, original eignea bt
- c. t. Sorpan. Jr.
C. Z. Serpan, Jr.
DISTRIBUTION:
EBeckjord TSpeis JCortez LShao RBosnak MEB Staff KBurke, IP JTaylor, EDO JRichardson, NRR CCheng, NRR PWu, DOE RNanstad, ORNL RCheverton, ORNL JHawthorne, MEA TGriesbach, EPRI i
2
t
\\
June 25, 1990 Working Group 3 i
Prof. Amir Ameev opened the meeting with greetings and a short summary of the status of embrittlement.
Shao asked for an extra session on reactor failure i
probability including US and USSR experts, and for discussion of mutual reactor issues. Also, he esked that we be able to witness an annealing of a Soviet reactor.
1.
Annealing of the reactor vessel.
1.1 A. Amaev -- Scientific-engineering and working plans for conducting annealing of VVER-440 vessels and experience.
Developed remote control density (?) meter to measure hardness.
Development of NDE proceoures for before and after anneal.
Nine vessels annealed, two with and 7 without cladding.
Also in 1992 a plant in Bulgaria.
The Czech Bohunice plant I
may be annealed in 1993, maybe earlier, with either Czech or Soviet equipment.
f Soviets plan on 7000 operating hours per year; thus 30 year plant goes for 30 x 7,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br />.
Revealed that templates (specimens) can be taken out of vessel walls to check annealing process.
5x60x70 mm.
First in GDR, I
now in USSR.
Annealing goes for 150 hours0.00174 days <br />0.0417 hours <br />2.480159e-4 weeks <br />5.7075e-5 months <br />, plu: heatup and cooldown.
Annealing conducted in 24 days. NDE before and after.
Determination of radiation embrittlement of vessel i
1.2 Sokolov materials after annealing, as basis for operations.
l No original surveillance programs, but later programs used similar, not archive, materials.
18 Af ter two annealing cycles at 460*C, > 9x10 get appreciable benefit of improvement not just recovery or reirradiation.
]
They have developed a series of nodels to predict shif t of j
critical temperature, including a conservative model. They j
seem to wind up with a typical residual 20*C shift.
j 3
N I
Soviets refer to high copper as 0.15-0.20, whereas they assume US "high" is.35 or more.
They seem to be able to predict US embrittlement data using their schemes, but must use different models for different
)
copper contents.
Still, they are conservative.
l 1.3 Hawthorne - Embrittlenent and annealing status.
)
Hawthorne showed that the US sees a trend of lower reembrittlement rate at higher annealing temperatures, whereas USSR did show some convergence for 340 and 440' anneals. JRH pointed out that weld flux does not affect recovery, but material composition does..
JRH pointed out that K at 100 MParih is greater than C 41-J jc y
for a number of materials, but maybe about equal for C 68-J l
y But Sokolov noted that even at lower temperaturc " wet" anneals-they attained a sort of saturation of reembrittlement upper limit.
2.
Research on vessel materials from operating plants.
2.1 Research on Novovoronzh AES No. 1 - Vessel material fracture.
Nikolaev - Used vessel material, annealed 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> at 650'F for unirradiated material. Measured 0.5 ct fracture toughness of shift of 90*C, but C shift at 48J of 160'C; however, these were y
two different trepans, nonetheless at about the same fluence.
So, it's a problem, and they would like to come back to this.
2.2 Nikolaev - Structure of the state of the vessel.
Discussion of the efficiency factor of annealing.
2.4 Krasikov - Microanalytical research of materials under cladding, at the HAZ, after long thermal exposure in the reactor and under annealing.
The purpose was to determine any diffusion of doping elements from cladding (especially Ni) into the base metal. Tests done by micro-hardness measurements. The results show virtually no diffusion of Cr and Ni into base metal after 20 years, even after 2 hour2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> anneal at 650'C 2.5 Serpan - Shippingport NST
')
4
3.1 Radiation Embrittlement of VVER-1000 materials.
Kryukov - Materials for VVER-1000's up in nickel content over VVER-440, and 1000's operate at 290*C rather than 270'C for 440's. They see indications of effect of Ni in the 15x2HM A (for1000's).
In fact, for Ni content > 1.3%, embrittlement i
increases rapidly.
Embrittlement performance is improved by finer grain size, but only because small size grain has lower initial T.T.
But now, they believe VVER-1000 material is sensitive to Cu, and at higher levels, is significantly additive with Ni, They also believe, but can't yet prove that P does not have a significant effect.
Note that P affects thermal aging, and they don't have data to prove from long-term exposure.
3.2 Vishkarey - VVER-1000 materials in research reactors.
13 I
Flux was 3 x 10 Showed a significant decrease of embrittle-ment rate for increase of T from 250*C to 290'C, and have developed relations to predict the behavior.
June 26, 1990 4.
Radiation Embrittlement of reactor vessel materials, and application to j
operating reactors.
4.1 Nanstad - Radiation effects on arrest on clad and toughness shift of cladding. Crack arrest toughness shifts roughly the same as Charpy, for.23 and.31 Cu welds, but over and under prediction not consistent. The curve shape is not seen to cha',ge, as of yet.
4.2 Hawthorne - Experimental fluence rate studies.
Some effect-of rate seen, but is material depent'ent; that is, some-times plate affected, sometimes weld. GundrerAingen showed lower shelf for weak L-C direction than test reactar irradiation.
It possibly can be related to inhomogeneity of forgings. Hawthorne-showed no effect of 40:1 thermal: fast flux on Gundremmingen steel; Kryukov observed they had shown similar results last year (for.
hightemperatureirradiations).
5
i.
1 l
l,.
l 4.3 Levit - Surveillance data analysis for critical temperature.
They have much data, but do not have as good an evaluation and analysis as they would like.
By studying only power reactor data they reduced variances caused by different temperatures and neutron spectra. By restricting analysis to these data they hoped to
)
F1/3 at fluxes 1
improve correlations for their relationd T 'AF k
of 4x10 and 1012(>.5MeV).
In the study P ranged II between 0.01 to 0.055, and Cu from 0.04 to 0.21%. At both lower and higher fluxes there is contribution from both P and Cu, for 270'C irradiation in VVER-440's. Although P embrittlement seems to saturate faster at low fluxes, P embrittlement seems still to be dominant, with Cu contribution rather nominal. When P content l
1s low, the main contribution is from P and Cu interaction but when P and Cu ere both low, the most embrittlement is caused by mechanisms other than those normally expected.
I 4.4 Bakirov - Means for measuring and analysis of embrittlement of vessel materials from annu l.
A video film shown of doing the post-anneal hardness measurement of the NORD-2 reactor. They used a 3-nn ball indenter to make measurements.
The device fixes magnetically to the vessel, with movable frame to accommodate difficult diameters. More than 15 poir.ts (measurements) are made.
Hardness was determined as approximately 235 before and about 200 HB after anneal.
(Brinell hardness) Recovery is about 70%.
Considerable backup work done to relate indentations to stress strain behavior.
However, one Soviet suggested that the changes are too small to be taken without much reservation.
For clad vessels, they are developing ways to remove the clad, but don't know how to replace. They believe they might get a secondary correlation using the clad.
Amaev stated that it's more important to keep cladding intact than L
to make hardness measurements.
t-1 6
l
5.
Mechanisms of.adiation damage.
5.1 Kevorkan - Medeling processes of vessel embrittlement.
1 The one model they use for copper precipitation is that of Odette.
But no Soviet experiments have been done to confirm the model for VVER-440 or 1000 steels. They find good agreement with American colleagues with this model, and now wonder if it will also work for annealing and reirradiation.
5.2 Griesbach - Radiation damage mechanisms, i
Stated that R.G. 1.99 relationships appear to reasonably predict Charpy (RTNDT)shiftsforUSreactors.
5.2 (Cont.) Nanstad.. Radiation damage mechanisms.
4 iune27.1990 t
5.3 Gurovitch - Electron Microscope Study of embrittlement of reactor vessel steels.
Studies done on steels of composition P-0.010-0.033 Cu 0.06-131, Fluence 1.1 - 50x1019.)E 0.5 MeV.
20 2
No defects observed until fluence reaches 1x10 n/cn E> 0.5 MeV.
Overall study has microstructures and defect densities, etc. of steels having relatively high P but lower copper, which could be of high use to NRC.
Interesting details of disappearance of all defects at 500'C anneal, and X10 reduction of intra-grain inclusions.
They have fine-grained steels that show ductile tearing at temperatures below the critical temperature.
They only get tempera-ture embrittlement after 700-1500 hours of annealing at 500*C.
5.3 Nikolayev - Fracture of Novoyoronezh. reactor vessel material.
Shift of Novo, base is 80*C, weld 160*C, Corresponding to iP = 0.013 n 0.033, respectively, based on trepans. Annealing s
for 150 hours0.00174 days <br />0.0417 hours <br />2.480159e-4 weeks <br />5.7075e-5 months <br /> at 470'C or two hours at 650'C eliminates all dis-location loops, while fine ppts grow nearly' double in size, for both base and weld metal.
5.4 Kapinos - Small angle X-ray scattering work on irradiated iron.
L Very interesting, good work.
L 7
~,.
-,+
o 6.4 Kar20v - Methodology for PWR PV Strength under PTS.
T.S.
work was started in 1973. They have tried to include the effect of warm prestress to help preclude fracture, j
I 1.
Initiation of brittle rupture only occurs on increasing loads 2.
Cracks even if initiated should arrest 3.
Actual crack size must be considered.
6.5 Maksemov - Embrittlement effect on VVER-440 vessels, f
ThecoefficientofembrittlementisA270a800(5P+0.07Cu).
l p
Analysis presented on annealing parameters of NORD plants.
6.6 Tutnov-Prediction of rupture, i
Verification of calculation, and PRA analysis.
Verification of computational programst studies began 20 years ago, including LEFM, EPFM, etc.
I June 28.19';0 6/ Thermal Model for PTS for probability of failure calculations. A large series of experiments in models and in full scale reactors l
to measure ' flaws, temperatures, mixing and distribution.
Theofanis offered to include all Soviet heat transfer and thermo-hydraulic experiments in the NUREG report he is preparing.
j 6.8 Cheverton - He concludes that PTS represents a threat'to PWRs so we have developed a rule to handle it. The probabilistic approach is appropriate, and NRC has implemented it in R.G. 1.154.- The studies suggest that the screening criterion may not be appropriate l
for all vessels in the US. The IPTS results may be changed I
appreciably by new information and studies to be undertaken.
i Comparison of US and USSR calculations of HBR-HYPO.
It is assumed to have an RT of 270'F, for 21 transients.
Results:
sone NDT 11 were within factor 2-5, while the extremes ranged from 0.1 to 45.
Finally the range of conditional probebilities was from j
10-8 to 10-3 8
L
t i
[
Karzov - Regarding their IT/ PTS tests; Test vessels were heat treated to sinulate *,he highest levels of embrittlement. Stated j
(agreed)thatthermaltransientsarenotdominant.
His lab has I
developed all the steels and their production technology. Stated that volume of flaws in welds is a factor of 10 higher than-in j
base metal.
(But Cheverton countered that surface flaws are of moreoverallconcern.)
j US/ USSR IPTS Analysis Differences Ky calculation US Finite Elenent (Influence Coefficients)
USSR ASME Code Curves US probably somewhat more conservative.
independent Variable i
US RTNDT l
USSR Fo(Cu,RTNDT, RTHDTsimulated) l ISSR probably more accurate here Type Flaw US 2D (a/w< 0.2);
3D(b=2m,a/w20.2)
USSR i
I Flaw Density US 1 flaw /m3 (not sinolated) for best estinate crack) l 3
USSR Simulated, averagee flaw /m,
t P(F/E) Calculation I
US Monte Carlo USSR closed form Number of Regions l
US Welds, Plates i
USSR I
)
i I
9
h' t
i l
i I
i i
i i
Working Group 4 Trip Report i
I i
l' l
1' l