ML19309A770
| ML19309A770 | |
| Person / Time | |
|---|---|
| Site: | North Anna  |
| Issue date: | 03/24/1980 |
| From: | Hazelton W Office of Nuclear Reactor Regulation |
| To: | |
| Shared Package | |
| ML19309A769 | List: |
| References | |
| NUDOCS 8004010313 | |
| Download: ML19309A770 (17) | |
Text
.
03/24/80 (s /
UNITED STATES OF AMERICA NUCLEAR REGULATORY COMMISSION BEFORE THE ATOMIC SAFETY AND LICENSING APPEAL BOARD In the Matter of
)
)
VIRGINIA ELECTRIC AND POWER COMPANY
)
Docket Nos. 50-338 OL
)
50-339 OL (North Anna Nuclear Power Station,
)
Units 1 and 2)
)
DERIVATION OF TURBINE DISC CRACK CROWTH RATES FOR NORTH ANNA UNIT 1 On March 3, 1980, the Appeal Board issued a " Memorandum and Order" in which it directed the Staff, among other things, to furnish it a more complete explanation of how the Staff constructed and would justify the crack growth rate curves employed in the previously submitted Staff analysis. This docu-ment responds to that Memorandum and Order.
1.
Staff Methodology In Attachment 1 to this document, I have set forth the Staff's approach to calculating turbine disc crack growth rates for operating nuclear power plants.
In Attachment 1, I explain the background end methodology for developing methods to predict the largest probable size of turbine disc cracks, and for determining the critical size of cracks that could cause failure.
I also discuss the conservatisms used by the Staff in developing crack growth rates and critical crack sizes that justify their use.
Predicted maximum crack sizes are derived from curves based upon actual service experience, and depend on temperature and minimum specified yield strength of the material.
In Attachment 1, at page 3, I set forth an equation for calculating the critical crack size for turbine disc bores.
This equation varies somewhat from the I
equation used by the Staff in its earlier testimony before the Appeal Board in this proceeding.1/ The essential difference between the two equations is that my equation takes into consideration a function related to the shape of the assumed crack, whereas the earlier testimony did not.
2.
Results of Calculations In Attachment 2, I have included Tables 1-3 and Figure 1, which represent the results of the Staff's review of the Westinghouse turbine crack data and Staff crack growth rate calculations for the North Anna Unit 1 turbine. These Tables and Figure are revisions of Tables 1-3 and Figure 1 which were attached to the
" Staff's Bases For The Continued Operation of North Anna Unit 1," which was submitted as Attachment 2 to a letter to the Appeal Board from Daniel T.
Swanson dated February 19, 1980.
The revisions to Tables 1-3 and Figure 1 were made to reflect more current and accurate data.
As a result, the crack growth rate curve for high strength discs has been revised.
In addition, calculations are also done considering that the crack will grow to a shape more nearly like that actually seen in service, where the length is no more than twice the depth. This will provide a "most likely" value of critical crack depth, but is not felt to be conservative enough for regulatory purposes.
Our calculation of cracks for the most critical first stage disc remains at a: px mimately 19% of the critical crack depth at design overspeed and this will ou.y increase to 31% of critical size by December 1, 1980, the time of the next refueling outage. / We now calculate that the most critical second stage 2
A! "NRC Staff Testimony Regarding Turbine Missiles" by K. Campe, et al.,
dated April 27, 1979, pp. 28, 29.
! The two most critical states in the North Anna turbines are the first and second stages.
The later stages run at such low temperatures that stress corrosion will proceed at a very slow rate, if at all.
.-.. =.
-..=.
disc could have a crack 29% of the critical depth at design overspeed, which could grow to 49% of the critical crack depth by December 1,1980. This 4
compares with our earlier reported estimates of 27% and 45% for the current and projected crack sizes, respectively. As stated previcusly, the current staff position is that when the postulated crack is less that 50% of the j
critical crack size, there is sufficient margin to account for any uncertainties in the calculations and therefore no necessity to become con-cerned about turbine operation.
The results do not alter the Staff's conclusion that there is reasonable assurance that North Anna Unit I can continue operation until December 1, 1
1980 with a very low probability of turbine disc rupture.
I 1
-n-
~
O 4
ATTACHMENT I
l
\\
l I
STAFF CALCULATIONAL METHODS FOR TURBINE DISC CRACKS Since November lf 1979, the Engineering Branch of D0R has been deeply involved in developing and evaluating methods to predict the possible size of turbine disc cracks, and to determine the critical size of cracks that could cause failure. We have relied heavily on the experience of the British, developed as a result of widespread turbine disc cracking they experienced in the early 1970's. We also have used much useful information supplied to us by Westing-house and General Electric. The basic concepts and methodology are well known and have been in common use for some time.
The major problems encountered are in the quantification of the relationship between the many materials parameters used.
Crack Growth Rate Predicting crack growth rates, especially for stress corrosion or environ-mentally enhanced fatigue crack growth, is impossible without experimental i
data that are obtained under conditions expected in service. At the present time, it is not known what the exact conditions are at turbine disc bores and keyways that cause cracking.
It is known that caustic and some acids will l
cause cracking of turbine disc steels, but laboratory and field tests also have shown that under the right conditions, cracks can be initiated and propagated by pure steam or high temperature water.
It is also known from laboratory tests that under some conditions cracks need a significant period of incubation to initiate, whereas under other conditions cracks will start to grow as soon as service conditions are applied. These situations make the job of accurately predicting actual crack growth rates and crack sizes in service an impossible task. We have not tried to do this.
~
_2_
What we do try to do is to predict what the worst case is likely to be.
If enough data are available representing the total spectrum of relevant con-ditions, the worst cases can be considered an upper bound to probable future crack growth rates. Of course wc cannot be sure that the data include worst possible cases, but if sufficient conservatisms are placed on the use of the crack growth determinations, the method can provide reasonable assurance that inspections will be performed before cracks grow to unacceptable depths.
The British did a great deal of work correlating their cracking experience with relevant parameters. The problem of making technical sense out of the vide variation in crack. sizes actually found in discs was a formidable one.
They found that disc material and heat treatment, keyway and bore designs, temperature of operation, and to some degree steam chemistry, were major parameters.
It should be emphasized that they never could predict whether c given disc would be cracked or not; or if it were, how deep the cracks would be.
The most useful informatica from their extensive studies is that the crack growth rate appears to be dependent on temperature, disc material strength level, and stress level. Accordingly, Westinghouse, General Electric, and we have been plotting curves based on depths of cracks found to date and total operating hours to discovery of cr.ck, as a function of operating temperature and material yield strength.
The b asic premise is that in accordance with the usual metallurgical and chemica; reaction rate laws, the higher the temperature, the faster the rate.
'he relationship appears to follow the classic Arrhenius law that the logtrithm of the rate is a linear function of the reciprocal of the absolute temperature. Further, there appears to be a l
clear distinction on the basis of minimum yield strength specified for the l
t
particular disc.
This is probably because of two effects:
the higher the yield strength the more susceptible the steel is to cracking, and the higher the stresses are (the design stress is the basis for the specified minimum yield strength of the material).
The curves :-a use are shown on the attached Fig. 1 (for convenience we use actual temperature, not the reciprocal; this causes the lines to be slightly curved).
There are three curves shown, for different yield strength materials.
The slope of the curves is based partially on British data, which generally are at lower temperatures than our cases.
This Figure also includes the data points from inspections performed on domestic Westinghouse turbines. Note that there is a great deal of scatter, but that the curves shown are the upper bound (or " worst case") of all the crack growth rates observed for each strength level.
Since these curves were developed, two additional cases of rather severe cracking have been discovered.
They are shown on the Figure as points for Yankee Rowe and Cooper Station.
Note that neither falls above the curve appropriate for the disc material.
These curves are subject to continual revision; the latest revision to the curve for high yield strength material is shown as a dashed line. This slight revision was found to be necessary when more accurate temperature and crack depth data became available for the Indian Point and Beaver Valley cracking incidents.
Although we do not expect that these curves will change significantly, we intend to continue our efforts to understand the basic principles involved in the cause of the cracking and crack depth predictions. We will modify our criteria whenever additional information indicates that it is prudent to do so.
4_
Critical Crack Size The calculational method used to determine the critical size of crack that could cause disc failure is based on linear clastic fracture mechanics (LEFM) con-cepts. The formulation used has been in use for many years, is generally well known, and has been verified by actual disc burst tests by several companies, including Westinghouse and General Electric. The important material parameter is the critical stress instensity factor K which is obtained by direct measurement using standard specimens and test methods, or by standard correlation methods developed to convert results from the simpler Charpy V impact test to K values. The specific LEFM formulation for critical crack IC size (Acrit) calculation for cracks in disc bores is:
!K A
=
crit 1.21 H Where:
o Is the nominal stress at the bore Q Is a compicx function related to the shape of the assumed crack If the crack is in a keyway, the depth of the keyway is considered part of the crack. Q varies with the shape of the crack and the ratio of the applied stress to the yield strength of the material.
In the case of turbine discs, where the stress is approximately 2/3 of its yield strength Q varies from 1.0 for a crack 10 times as long as it is deep to 2.3 for a crack twice as long as it is deep. Because Q affects the critical crack size in a linear manner, it is important to know the probable shape if the calculation is to be meaning-ful. Westinghouse has calculated critical crack sizes using a postulated shape of 10 to 1.
This new appears to be very conservative, as it results in calcu-lating critical crack depths less by a factor of 2.3 (the ratio of the Qs) than if a shape of 2 to 1 is assumed, for example.
Now that many cracks have been found, a more realistic crack shape can be used for calculations.
Most cracks, and particularly the deeper ones, appear to be no more than twice as long as they are deep. This is why we chose to use a 4 to 1 shape. This is still conservative, but is considerably more realistic than the 10 to 1 l
assumption.
Note that if we wanted to actually predict cr'cical crack sizes, I
a shape between 1 to 1 and 2 to 1 (in line with actual observations of crack shapes) would be used. As an example of the difficulty caused by using over conservative shape factors, the critical size keyway crack for design over-speed of the first stage disc in Cooper Station was calcriated to be 0.95 inches in depth.
The recent inspection at Cooper found that a keyway crack about 2.9 inches deep was present although it had not caused a failure. This is because the critical depth of a realistic 2 to 1 shape keyway crack would be about 3.4 inches for normal operating speed, rather than the conservative but less realistic estimate of 0.95 inches.
The determination of a K value r each disc is done in a very simple manner.
IC Actual K measurements are not practical, so standard correlations with the IC Charpy V impact test results are used. The correlation we use was developed some years ago by Rolfe and Novak of the U.S. Steel Research Laboratory.
It is an empirical expression that has been shown to be fairly accurate for steels of the type used for turbine discs.
Parameters of interest are the yield strength of the steel and the absorbed fracture energy measured by standard Charpy V impact tests:
KIC
" **"E
~
eld at eng Sl As each disc has different strength and Charpy properties, the calculation is performed for each specific disc of interest.
4.
Summary 1.
Predicted maximum crack sizes are derived from curves based on actual service experience, and depend on temperature and minimum specified yield strength of the material.
2.
Critical crack depths are calculated conservatively for regulatory purposes using a 4 to 1 crack shape.
If it is desired to predict critical crack sizes more accurately, a 2 to 1 crack shape should be assumed, because this shape is similar to that of deep cracks found in service.
i i
l 1
i
4 S
ATTACHMENT 2
1 e
D
TABLE 1 Predicted Maximum Crack Depths Disc Temp.
Strength
<a te Depth 3/15/80 Depth 12/1/80 Level in/mo (14.2 mo)
(23.7 mo) 1 329 Med.
.028
.40
.66 2
267 High
.026
.37
.62 3
213 Med.
.004
.06
.09 4
182 Med.
.002
.03
.05 TABLE 2 Critical Crack Depths Disc Location K c /Tn.
i Stress Critical Keyway Crack Depths *
(K51 (KsI) a/2c =.1 a/2c =.25 a/2c =.5 (W Basis)
(NRC Basis) (Most likely) 1 LP-1 Gov 200 76 1,43 2.1 3.76 2
LP-1 Gov 187 87
.84 1.26 2.4 3
LP-2 Gov 228 72.3 2.2 3.1 5.6 4
LP-2 Gov 178 70 1.33 1.9 3.5 c
Determined by calculating Critical Depth for a Bore crack, then subtracting the depth of the keyway. This considers that the keyway is an extension of the crack.
TABLE 3 Per Cent of Critical Crack Depth for Maximum Predicted Cracks Disc 2/15/80 Per cent of Crit.
Per cent of Crit.
Per cent of Crit.
Predicted a/2C =.1 Crack a/2C =.25 Crack a/2C =.5 crack Max (in)
(W Basis)
(NRC Basis)
(Most likely) 1 LP-1 Gov.
.40 28%
19%
11%
2 LP-1 Gov.
.37 44%
29%
15%
+
12/1/80 1 LP-1 Gov.
.66 46%
31%
18%
2 LP-1 Gov.
.62 74%
49%
26%
/
4 TURBINE CRACK RATE AS FUNCTION OF TEMP. & MATERIALS A - HIGH YIELD STRENGTH d
hg*pl
.3 /
0, Mg e-MEDIUM YIELD STRENGTH W e - LOW YIELD STRENGTH (P 5 0.I c'
a
- g..
. a.., ;.
=. _. =...
1.-
p
.{ ; i E@g -i
- j 4pji
- e-
- i-1.:-
--J
.ya Md.i
- . i.
- -
}8
- _j j 20 N
s
. : -(f gy.
g
-:- N 3pl cyae5
. igw mi e a
.a ggg
' ur, ja nu i!
g 9.g6 7-..
4,
.. wr gc m,
.g N.....
g p M f- - j
.J
,m.__-
...~_
_.g
.w
.N-k_ _.
Lce:
w r
s i: E -@ ' -
Q MAS
. Mif li
- E - -. gd
- =-E IE ji:-
-+ iEi! ' IIMQ u-r@..
4H= j5:
N7g x _.
- A
- c,
._u
~
.u r:
=
==- -_ m M.r-
- Ti r
- =
x tm n.r' Mr r L ;--
- _
- -;. ay...;
.. (.
.L.._ rr-T x
X-O L-
=. - -
-_+ -
t-z p.
... r-... _..
. - _t-.. r t __.. %..
- - ~ ~
- A -
^
ri7-__
..-,2_+_
o
. I _...
. +. -..
e _ __
p.t..
- E.
-_. 1.
y
..I a ;.,... " ~..
..,LQ-_ N._.
g t r.
-* n. _ i N- % %.
^
L *-
-r m
s g%
- g..
. 7-- _g
_ -fg.
g._
h h._i
{a{,..p ptg K..D_
'1
.u
..-2 5
y 4y
.c i
2.__-
7 l' I F__..
t l__.t 3
.Tk.. lTlt i i-t-N' I
i U
- -"-NY i:b-
_3._.-
g-a
. q.
o i..
z N.N ~
-~1 T
0.01 ~T Ti d b i
'N eh
- i. : 'N%
4.i :
q ll %
?
\\
-. 1, J
- -i-e
- - h
. g:
-;r.
r._ ; -
y_
- m; ' h t:-
=
i.
u-e.
N 1
.Jre.u un =2...-r
- =
9 3
F.
[i -jli-'-[g
.h_~ Kd
- -j
- si i, ; j 2_
- 3E 'M N
g j: '-
'H[j 22, 12 -i --i
-H;
]
1:
- N=
g%
f-
.:T
- -i T r_. = L
__.x_.-
= :.
- d. :d fin 4 w6 ktI=-
c :.:
y
=i M-t ~.--1*'
z
^;'-
~
L ---K
==
==-
y 2.:. r: :_.zn
=
- -'..-~,
.r e
yg
_ Ayp -m._,
_ Q _- 3. g, y =..
c._ ;
. _ _ n.
4
.t---
.- 3 _
-.13.
&., g
, b-
~. _
"%__ W U
2 i' ' '
^
. _3 m
1%
p
~ _ -
y 19g =0'. 0 '_:
- r
. _ r. d__._
}_:.
- a.....
r.<g.
+ :tixp {8.75 1
'8
- ~} 3 w /[
._;m n:WF 85 m Frxrar-WM ty a c.t-- ps 7 =p.p.;4g-gi :
~
- W,-.: :-.r"="
~ - ~ - - 24 3. t ;._ -
L ::. {tr- =.:-_' '
r=
R rl -
~r: =
2:.
_. t :
2 --.4-.
._v.
- p. w.v.is!
. _.un f,y
.. n u r
. L, ma. Otti..
.,I,i
-r---
M, m.
- 1*--I crT1 ' M- +..'___r,__" _. - -' "ii.
.. - _ _ + -
_-r-..
g
...q:'g3gj
- n..:. :----_
. _ r--
n.-7.
r
- --.'..1__
w..__..
m a.
_g.
1 x
,.. w.... j..
3s.
7..
..p.$-.-.Av..
. _6_.
.)y4
,__.7.F.,.,.,... - _ _.. _+
-- c.
r
.4 2
7 -
y 4
r
. +
x
.. J__ 4:
_._,_.....-TgJ
_._ M,,.f
.p..
I
- jj._.
y___
T...._
ll>
...a. __
qi,,
_7__.. 4 i;
i i
_. _2.
i 0.001 -ti, i
i l
l 340 320 300 280 260 A40 220 200 360 3_80 TEMPERATURE. OF
UNITED STATES OF AMERICA NUCLEAR REGULATORY COMMISSION BEFORE THE ATOMIC SAFETY AND LICENSING APPEAL BOARD In the Matter of
)
)
VIRGINIA ELECTRIC AND POWER COMPANY
)
Docket Nos. 50-338 OL
)
50-339 OL (North Anna Nuclear Power Station,
)
Units 1 and 2)
)
CERTIFICATE OF SERVICE I hereby certify that copies of "NRC STAFF RESPONSE TO APPEAL BOARD'S MEM0-RANDUM AND ORDER OF MARCH 3, 1980" in the above-captioned proceeding have been served on the following by deposit in the United States mail, first class, or, as indicated by an asterisk, through deposit in the Nuclear Regulatory Commission's internal mail system, this 24th day of March, 1980:
Alan S. Rosenthal, Esq., Chairman
- Michael W. Maupin, Esq.
Atomic Safety and Licensing Appeal Hunton & Williams Board P.O. Box 1535 U.S. Nuclear Regulatory Commission Richmond, VA 23212 Washington, DC 20555 Dr. Kenneth A. McCollom Dr. John Buck
- Assistant Dean Atomic Safety and Licensing Appeal College of Engineering Board Oklahoma State University U.S. Nuclear Regulatory Commission Stillwater, OK 74074 Washington, DC 20555 Mr. James M. Torson Michael C. Farrar, Esq.*
501 Leroy Atomic Safety and Licensing Appeal Socorro, NM 87801 Board U.S. Nuclear Regulatory Commission Anthony Gambardella, Esq.
Washington, DC 20555 Office of the Attorney General 11 South 12th Street - Room 308 Dr. Paul W. Purdom, Director Richmond, VA 23219 Environmental Studies Institute Drexel University Mrs. June Allen 32nd & Chestnut Streets 412 Owens Drive Philadelphia, PA 19104 Huntsville, AL 35801 i
Mr. R. B. Briggs 110 Evans Lane Oak Ridge, TN 37830 I
l l
l
Mr. Dean P. Agee Senator Allen R. Carter, Chairman Executive Secretary Joint Legislative Committee on Board of Supervisors Energy Louisa Courthouse Post Office Box 142 Louisa, VA 23090 Suite 513 Senate Gressette Building William H. Rodgers, Jr., Esq.
Columbia, SC 29202 Georgetown University Law Center 600 New Jersey Avenue, N.W.
Atomic Safety and Licensing Board Washington, DC 20001 Panel
- U.S. Nuclear Regulatory Commission Mr. William Warren Washington, DC 20555 722 St. Christopher Road Richmond, VA 23226 Atomic Safety and Licensing Appeal Panel (5)*
Richard M. Foster, Esq.
U.S. Nuclear Regulatory Commission Musick, Williamson, Schwartz Washington, DC 20555 Leavenworth & Cope, P.C.
P.O. Box 4579 Docketing and Service Section (7)*
Boulder, CO 80306 Office of the Secretary U.S. Nuclear Regulatory Commission Mrs. Margaret Dietrich Washington, DC 20555 Route 2, Box 568 Gordonsville, VA 22042 Ellyn R. Weiss, Esq.
Sheldon, Harmon, Roisman
& Weiss 1725 I Street, N.W.
Suite 506 Washington, DC 20006
/ &E m
Daniel T. Swanson Counsel for NRC Staff
.