ML20078R503
| ML20078R503 | |
| Person / Time | |
|---|---|
| Site: | Fort Saint Vrain  |
| Issue date: | 10/27/1983 |
| From: | Brey H PUBLIC SERVICE CO. OF COLORADO |
| To: | Jay Collins NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION IV) |
| References | |
| P-83348, NUDOCS 8311150173 | |
| Download: ML20078R503 (16) | |
Text
-
r pul) lie Service Comi):my f OcBmdk d
~
C I
2420 W. 26th Avenue, Suite 100D Denver, Colorado 80211 h
. l October 27, 1983 Fort St. Vrain Unit No. 1 P-83348 Mr. John T. Collins, Regional Administrator Nuclear Regulatory Commission m' M@((]
W '
Region IV Office of Inspection and Enforcement g
4 1983 611 Ryan Plaza Drive, Suite 1000 Arlington, Texas 76012 gi j
.i _hw DOCKET N0.: 50-267
SUBJECT:
Additional Information on H-451 Graphite
REFERENCE:
G.L. Madsen (NRC) Letter to 0.R. Lee dated August 2, 1983 (G-83282)
Dear Mr. Collins:
This letter submits the additional information on H-451 graphite requested by the referenced letter.
We trust that the information provided in the attachment to this letter satisfies your concerns.
If you have any further questions, please contact Mr. M.H. Holmes at (303) 571-8409.
Very truly yours, k
444t w H. L. Brey, Manager Nuclear Engineering Division HLB /JPL:pa Attachment QD B311150173 831027
)
PDR ADOCK 05000267 p
n-
.p.
Attachment to P-83348 Page 1 of 15 NRC QUESTIONS ON FORT ST VRAIN SEGMENT 9 RELOAD OF H-451 GRAPHITE QUESTION:
1.
The data and conclusions presented in GA-A16402 appear to differ in some areas from the information presented in the H-451 graphite generic licensing topical report GLP-5588, which NRC reviewed and approved in 1979.
GLP-5588 provided the Nsis for allowing the substitution of near-isotopic H-451 graphite fuel and reflector elements for the original reference needle-coke H-327 graphite elements in the FSV reactor.
In the new report, GA-A16492, it is indicated that creep rates in tension are higher than in compression, whereas in GLP-5588 it was stated that the same equation can be used for both tensile and compressive irradiation-induced creep. The recommended creep equation in GA-A16402 thus differs from that contained in GLP-5588. Moreover, in GA-A16402 it is indicated that the thermal expansivity, Young's Modulus, and Poisson's ratio are affected by the creep strain, whereas, there was no mention of such an effect in GLP-5588.
Discuss the safety significance of the revised design equations for irradiation-induced creep and other physical properties.
Explain how this new understanding of the effects of irradiation on. creep and other properties has been factored into the design of the H-451 graphite elements. Note that if the design equations and curves in GLP-5588 are no longer applicable to safety analyses, an amendment to that NRC-approved report should be submitted containing the corrected information.
Attachment to P-83348 Page 2 of 15
RESPONSE
Information on irradiation creep in graphite presented in Ref.1 included data from constant load creep experiments at 500*C conducted in the Petten reactor, which appeared to show creep rates in compression that were about 25% lower than creep rates in tension.
According to the generally accepted model for irradiation creep in graphite (Ref. 2), creep rates should be the same in tension and compression at low strains, but the compressive creep rate should fall below the tensile creep rate at compressive strains high enough for microcrack closure to occur.
Further data from the Petten experiments (Ref. 3), published since the review in Ref.
1, are consistent with this model.
The data in Ref. 3 show compressive creep strains within 15% of the tensile creep strains up to strain levels of about 0.5%.
At higher strain levels the compressive creep strains lag the tensile strains by a larger amount. The scatter of the data points is such that a 15% difference at low strains is not considered to be significant. Since there is no physical reason to expect a difference between compressive and tensile creep rates at low strain levels, and the data scatter is too large to establish a true difference, creep rates in compression are considered to be equal to those in tension for design stress analyses. Therefore, the average of the creep constants for compression and tension given in Ref.1 is used for stress analysis.
r-AttachmGnt to P-83348 Page 3 of 15 The irradiation creep strain for H-451 graphite design stress analyses is given by the expression:
Irradiation temperature less than or equal to 650*C.
- cr " h[1-exp(-2.5x10-20,)]
-21 de E*fe E(e,T)
+ 1.6x10 o
o o
Irradiation temperature greater than 650'C:
[1-exp(-2.5x10-20,))
=
c cr
- oE*[1.6x10-21+5x10-28 (T-650)2.6]
e de where:
E o E(e,T) g
= total uniaxial creep strain (cm/cm) c cr
= stress (MPa) o E
= Young's modulus when load is applied or changed (MPa) e
= fast neutron fluence, n/cmz (Egreater~ than 0.18 MeV, HTGR spectrum) 21 E*
= Young's modulus at a fluence of 1.4 x 10 n/cm2 (MPa)
E
= unirradiated Young's modulus (MPa) g E(e,T) = Young's modulus after a neutron fluence e at temperature T (MPa)
T
= irradiation temperature ( C)
The first line of the equations represents the transient part of the creep strain, which is slowly recovered when the stress is removed during irradiation. The second line is the steady state part
Attachm:nt to P-83348 I
Page 4 of 15 of the creep strain, which is not recoverable.
7he equations are
-valid for temperatures between 300 and 1100*C, t;uences up to 1022 e
n/cm, and creep strains up to 0.5%.
With regard to the effects of creep strain upon thermal expansivity, Youngs' modulus, and Poisson's ratio, as noted above the creep strain equations are valid for creep strain up to 0.5%, which encompasses the range of creep strains expected for Fort St. Vrain fuel elements. Based upon the information presented in Ref.1, at a creep strain of 0.5% the following ratios of these parameters to their values for unstressed specimens are obtained:
Parameter 0.5% Creep Strain /Zero Creep Strain Thermal Expansivity 0.91 Young's Modulus 1.02 Poisson's Ratio 1.15 These variations are less than or equal to the uncertainty in each of these parameters. Accordingly, while the stronger effects of fast neutron fluence and temperature upon these parameters are taken into account, the secondary effect of creep strain is neglected in graphite design stress analyses.
The new equations for irradiation-induced creep were not factored into the design of the H-451 fuel elements to be used in Reload
Attachment to I
P-83348 Page 5 of 15 Segment 9.
The information in Ref. I was published in August 1981, and the more recent Petten data were published in September 1982; the fuel loadings for Segment 9 were established in December 1980. Thus, the new creep equations were not available in time to influence the design of Segment 9.
To assess the significance of the new creep equations, however, an evaluation of fuel element stress and strain has been conducted with both the old and new equations. The evaluation was conducted with the GBEAM code, which is discussed in the response to Question 3.
Details and results of this comparative analysis are described in the response to Question 3.
In addition to the new H-451 graphite irradiation-induced creep equations presented here, changes have been made in the material property data base for H-327 graphite since the issuance of GLP-5588 First, the creep formulation for H-327 graphite has been changed to be analogous to that for H-451. The new H-327 formulation is based upon the same physical considerations which led to modification of the H-451 creep formulation.
Second, a general review of H-327 i
irradiation-induced dimensional change data was conducted which incorporated new information from capsule irradiations in the Petton reactor. As a result, new dimensional change curves were generated.
The revised dimensional change curves are generally consistent with post-irradiation measurements on more than 100 fuel and reflector elements at Fort St. Vrain (Refs. 4 and 5).
f u
Attachm:nt to P-83348 Page 6 of 15 Amendment pages to GLP-5588, which was a GA topical report, will be submitted to NRC by GA Technologies under separate cover.
The amendment will contain the H-451 irradiation-induced creep equations presented in this response and a summary of the changes which have been made in the H-327 material property data base.
QUESTION:
2.
In view of the discovery of cracked H-327 graphite elements at the last refueling and the planned insertion of H-451 graphite elements in Segment 9, we believe some surveillance, consisting of interim visual exaninations equivalent to those recomended for "first core loadings" (see attachment to Speis' January 3, 1979 letter), is necessary for the new reference fuel elements. This will provide confirmation that the H-451 graphite elements are performing satisfactorily as predicted and are not encountering the same or worse cracking behavior than the original H-327 reference mate' rial.
Accordingly, please propose a surveillance program that will provide such confirmation.
RESPONSE
In addition to the existing Fuel Test Element Surveillance Program, Public Service Company of Colorado proposes the following surveillance program for Segment 9 H-451 graphite to provide for interim visual examinations equivalent to those recomended for
T Attachment to P-83348 Page 7 of 15 "first core loadings" and post irradiation examination.
Such a surveillance program will provide confirmation that H-451 graphite elements are performing satisfactorily relative to cracking behavior.
H-451 Graphite Surveillance Program Fuel Test Elements:
During the first reload, eight Fuel Test Elements (FTEs) made of H-451 graphite were inserted in the core to demonstrate, among other things, acceptable performance and safety of H-451 graphite.
These elements are all located in the third fuel layer and are exposed to inter-region gap flows, (see response to Question 3) as was the first cracked H-327 block, so the FTEs are in suitable locations to monitor cracking behavior of the type encountered with H-327 fuel elements.
Furthermore, since the FTEs will have been exposed to almost 500 effective full power days (EFPD) of operation prior to the upcoming third reload, the FTEs " lead" the standard H-451 fuel elements relative to irradiation and thermal effects by a substantial margin.
The first Fuel Test element removed from the core (FTE-1) was l
l nondestructively examined in April 1982 a f ter 189 EFPD of 21 exposure (about 0.65 x 10 n/cm2) under a DOE-funded program.
The examination included 1) a visual inspection, 2) dimensional measurements, and 3) gamma dose rate and neutron count l
l l
l f
L
Attachmsnt to P-83348 i
Page 8 of 15 measurements.
The structural perfomance and dimensional stability of the H-451 graphite element were excellent.
As each of the remaining Fue', Test Elements is removed from the reactor core, it will be visually examined for corrosion, cracks, scratches and other abnormalities, as a minimum.
Interim Visual Examination:
During Refuelings 4 through 8,
inclusive, PSC will perform in-core visual exhmination of at least twelve (12)- Segment 9 H-451 fuel element surfaces during each refueling using the Reactor Viewing Device.
The surfaces to be examined during each reload will be those of fuel elements in regions 3,13 and 18 that are adjacent to regions being refueled and that provide a normal (right angle) viewing angle to the Reactor Viewing Device.
Since these surfaces are inter-region surfaces that will have been exposed to inter-region gap flow effects, as were the two cracked fuel elements (see response to Question 3), this examination is appropriate.
The inspections will be videotaped for record purposes, if practical.
The Reactor Viewing Device (RVD) is a miniature, radiation tolerant TV camera.
It is capable of providing 650 vertical lines of resolution and 400 horizontal lines of resolution and discerning eight and nine shades of grey vertically and horizontally, respectively.
Original checkout tests of the RVD demonstrated its ability to detect hairline (1/64 inch) cracks.
r
Attachment to P-83348 Page 9 of 15 The level of performac.ce achieved during surveillance inspections should be adequate to confirm that the H-451 graphite elements are not experiencing cracking that could lead to loss of cooling configuration or interfere with handling of the fuel elements.
Post-Irradiation Examination:
Five Segment 9 fuel elements will be precharacterized prior to the forthcoming refueling.
When these elements are
- removed, during Reload 9,
the five precharacterized fuel elements will be non-destructively examined in the hot service facility.
This on-site examination will include:
1.
Visual Examination 2.
Measurements to Determine Graphite Dimensional Changes Data evaluation for the interim Segment 9 visual examinations, the Segment 9 P.I.E., and the Fuel Test Element examinations will be provided to NRC as they become available.
QUESTION:
3.
In GLP-5588 it was stated that the H-451 fuel elements would have higher calculated stress levels than H-327 graphite elements, but that design stress margins would be improved by the use of H-451 elements because the higher strength of H-451 more than compensated for the increased stresses.
- However, the stress models, FESIC,
Attachment to P-83348 Page 10 of 15 SAFE /GRAPHIT and the like, that have been used for HTGR graphite ~
stress analysis, have never received NRC review, and the finding of i
cracked blocks in FSV, where none had been predicted, calls into question the reliability of these modeling techniques. Therefore, please discuss any improvements that have been nade in the graphite stress models or analytical input currently in use for the FSV graphite blocks.
Demonstrate how the new H-451 blocks will retain adequate margin for structural integrity. What exists in the form of confirmation of_ the adequacy of the analytical models now in use?
RESPONSE
i The stress models, FESIC and SAFE GRAPHITE, used in the evaluation of H-451 graphite in GLP-5588 were used in the original Fort St. Vrain FSAR fuel element stress analyses. These analyses were accepted by NRC (AEC) as suitable to form part of the licensing.
basis for Fort St. Vrain. Use of these methods was endorsed by the NRC Project Manager for Fort St. 'Vrain at the time when GLP-5588 was being prepared by GA and has been continued in reload safety analyses in a manner consistent with the recommendations of NRC Branch Technical Position D0R-1, " Guidance for Reload Submittals."
Newer methods for HTGR fuel element stress analyses -have been developed since the submittal of the FSV FSAR.
It should be noted, j
- however, that these methods have never received NRC review, nor have they been validated sufficiently for NRC review.
Furthermore, they i
1 ye. U,,.
---..,,y
.m y
,y
,,~-w,,,,,,..y,..,,yn,,.--
,.i.,-
e y
Attachment to P-83348 Page 11 of 15 l
have never been used in analyses reviewed and approved by the NRC.
Accordingly, the following description of these methods and their use is provided for information only.
Fuel element stress analyses for the large HTGR are performed primarily with the codes GBEAM and TWOD. GBEAM is a one dimensional structural analysis code using the classical elementary beam theory assumptions.
It performs time-dependent axial stress analysis of irradiated graphite components. A standard linear viscoelastic solid is used to model the material response of graphite under irradiation.
TW0D is a two-dimensional finite element stress analysis code used for the detailed stress and deformation analysis of the graphite fuel elements.
The code uses a generalized plane strain assumption and a material model similar to that of GBEAM.
Due to the use of higher ordered finite elements (e.g.,
8 node quadrilaterals) accurate modeling of the multi-holed fuel elements is achieved for both the axial and three in-plane stress components (x, y, and shear).
With regard to confirmation of the adequacy of these analytical models, they have been subjected to traditional verification methods such as checking against hand calculations and against theoretical formul ations of simple problems.
The results of these code verification exercises have been satisfactory.
In addition, these rodels are being used at GA in the DOE-funded evaluation of the operating history of the FSV Segment 2 fuel
Attachm:nt to P-83348 Page 12 of 15 elements with cracked webs. DOE has funded an evaluation of these elements with newer methods to assess potential implications for the large HTGR.
These analyses, in which the latest H-327 graphite material properties are being used, have not yet progressed to a point where the cause of the cracks can be definitively determined.
The analyses have, however, indicated a probable cause of the cracks.
The following is a discussion of the preliminary results of these i
analyses.
The peak operating in-plane stress of Segment 2 fuel element S/N 1-2415 was calculated to occur in the area where the crack.was observed.
The dominant effects which lead to high stress at this location are:
- 1) inter-region gap flows, which cause the edge of the fuel element on the outside of the region to be relatively cold and in tension; 2) a skewed power and flux distribution in this fuel
- element, in particular during all of Cycle 2 operatica, when the region control rods were fully or partially inserted. High power and flux (thus fluence) are experienced in the same area where the edge of the fuel element is relatively cold. The high power causes the graphite just inside the ed,e to be hot, thus increasing the thermal j
gradient in this area, and 3) a product of (1) and (2), wherein the l
combination of high fluence and low temperature lead to greater irradiation-induced contraction in the fuel element edge where the crack occurred, increasing tensile stresses.
l l
f 1
L
Attachment to P-83348 o
~
Page 13 of 15 The peak stress calculated at the crack location was about a factor of 1.5 below the graphite strength for nominal, time-averaged operating conditions; hewever, sensitivity calculations perfomed for this block show operating stress may be increased 50% - 100% due to possible variations in operating conditions.
In part, these variations were smeared in the time-averaging process used to derive a single set of data for each calculational interval.
Also, it was shown that stress is sens* tive to the gap flow rate.
Since the gap is not mechanically fi).ed along the entire column length, the gap width is expected to vary considerably throughout the core.
The stress analysis performed to date does not conclusively predict the crack that was observed, but it has indicated the importance of certain operating conditions to stress formation.
Based on the results obtained thus far, it is not inconsistent that a crack did occur at this location.
Further analysis and evaluation, planned for FY-84, will look more closely at the actual operating history and the sensitivity to variables used in the analysis to establish the most likely scenario for causing the observed crack.
With regard to the structural adequacy of H-451 graphite, H-451 graphite is superior to H-327 graphite for use in HTGR fuel element design due to its higher strength and its better dimensional stability, and because it does not experience net expansion at peak design fluences. Therefore, although the cause of the cracks in the
Attachm:nt to P-83348 Page 14 of 15 FSV Segment 2 fuel elements has not yet been definitively determined, use of H-451 in place of H-327 graphite will be an improvement.
To illustrate this. point, the preliminary TWOD analyses of the Segment 2 fuel element discussed above were repeated; however, the !!-
- 327 material properties were replaced with H-451 properties, including the creep formulation presented in the response to Question 1.
Using this approach, one-can assess the relative performance of these graphite types for a hypothetical situation in which one assumes that the Segment 2 fuel element was made from H-451 graphite instead of H-327.
The results _ of these preliminary TWOD analyses of the Segment 2 element indicate that, for the nominal operating conditions of the Segment 2
fuel
- element, the calculated in-plane and axial stress / strength ratios with H-327 graphite are about 8% higher than those obtained with H-451 graphite properties.
It is concluded, therefore, that for the conditions experienced by this fuel
- element, the probability of cracking would have been somewhat lower had the element been made from H-451 graphite.
In addition, in response to Question 1, a GBEAM analysis of the Segment 2 fuel element was conducted assuming H-451 graphite properties.
Both the H-451 irradiation-induced creep formulation in the response to Question 1 and that from GLP-5588 were used.
This
Attachment to i
P-83348 Page 15 of 15 analysis enables a side-by-side comparison of the effects of the new creep formulation to be conducted.
The resul ts of the GBEAM evaluation, which considered both operating and shutdown
- stresses, indicate that maximum stress / strength ratios calculated with the new creep formulation were increased by 15% above those obtained with the old formulation when the actual operating history of the element is simulated. However, as noted in the comparative analysis of H-451 and H-327 graphite, the stress / strength ratios obtained for H-451 graphite are still lower than those obtained for H-327 graphite.
It is concluded, therefore, that the new creep formulation has no adverse impact upon fuel element performance.
References 1.
R.
J.
Price, " Irradiation-Induced Creep in Graphite: A Review,"
GA-A16402, August, 1981.
2.
B.
T. Kelly and A.J.E. Foreman, "The Theor 12, 151 (1974)y of Irradiation Creep in Reactor Graphite," Carbon 3.
M.R.
- Cundy, G.
Kleist, and D. Mindermann, " Irradiation Creep Perfomance of Graphite up to High Neutron Fluence," Carbon
'82 (Proc. 6th London International Carbon and Graphite Conf.), p.
353 (Society of Chemical Industry, London,1982).
4.
C.
M.
Miller and J. J. Saurwein, " Nondestructive Examination of 51 Fuel and Reflector Elements from Fort 4t. Vrain Core Segment 1,"
GA-A16000, December, 1980, PSC letter to NRC P-81254, November 16, 1981.
5.
J.
J.
Saurwein, " Nondestructive Examination of 54 Fuel and Reflector Elements from Fort St.
Vrain Core Segment 2,"
GA-A16829, October, 1982, PSC letter to NRC P-83196, June 2, 1983.
<