ML20053C049
| ML20053C049 | |
| Person / Time | |
|---|---|
| Site: | North Anna  |
| Issue date: | 05/26/1982 |
| From: | Shoemaker C NRC ATOMIC SAFETY & LICENSING APPEAL PANEL (ASLAP) |
| To: | |
| References | |
| ALAB-676, ISSUANCES-OL, NUDOCS 8206010447 | |
| Download: ML20053C049 (46) | |
Text
I,
. v, UNITED STATES OF AMERICA NUCLEAR REGULATORY COMMISSION ATOMIC SAFETY AND LICENSING APPEAL BOARD Administrative Judges:
Alan S.
Rosenthal, Chairman Dr. John H. Buck Dr. Lawrence R. Quarles
)
In the Matter of
)
)
Docket Nos. 50-338 OL VIRGINIA ELECTRIC AND POWER
)
50-339 OL COMPANY
)
)
(North Anna Nuclear Power Station, )
Units 1 and 2)
)
)
Messrs. Michael W.
Maupin, James N.
Christman and James M.
Rinaca, Richmond, Virginia, for the applicant, Virginia Electric and Power Company Messrs. Daniel T.
Swanson and Henry J. McGurren for the Nuclear Regulatory Commission staff DECISION May 26, 1982 (ALAB-676)
Several years ago, in the course of the review on our own initiative of two Licensing Board decisions in this operating license proceeding, 1! we took note, inter alia, of an unresolved generic safety issue of seemingly special relevance to the two-unit North Anna nuclear facility.
--1/
LBP-77-62. 6 NRC 1127 (1977) and LBP-78-10, 7 NRC 295 (1978).
bbb 8206 01o utn G S 'o ' g
5 2
ALAB-491, 8 NRC 245 (1978).
That issue pertained to the protection of safety-related components from missiles generated either inside or outside the facility. -2/ This subject had been dealt with in a series of Task Action Plans prepared by the NRC staff, d
and also had received the attention of the Advisory Committee on Reactor Safeguards.
Id. at 249.
The particular concern in the instance of North Anna stemmed from the fact that the orientation of its four Westinghouse turbines (two in each unit) was such that safety components might be in the path of a missile generated by a turbine failure inside the facility.
At that threshold stage, we directed the staff to supply us, in affidavit form, with "a full and detailed
_2/
In this context, a missile is a high-velocity fragment produced by the breakup of an object (such as a fence or barn during a tornado or, as will shortly be seen, a component within the facility).
_3/
In response to a Commission directive to develop "a program for resolution of generic issues and completion of technical projects," the Office of Nuclear Reactor Regulation developed a list of unresolved generic safety items and set up a series of tasks to study and resolve each item.
See NRC Program for the Resolution of Generic Issues Related to Nuclear Power Plants (NUREG-0410, January 1978).
For a discussion of the manner in which unresolved generic safety issues are to be dealt with in individual licensing proceedings, see Gulf States Utilities Co.
(River Bend Station, Units 1 and 2), ALAB-444, 6 NRC 760, 783 (1977).
't, 3
explanation of why it is acceptable to permit the North Anna units to operate in the face of" the unresolved missile issu'e.
Ibid.
Following receipt of both that explanation and further submissions by the parties and an amicus
- curiae, 1/ we entered an order in which we narrowed the inquiry to "the possibility of damage caused, not by objects originating outside the plant, but by pieces of the turbine breaking loose."
ALAB-529, 9 NRC 153, 154 (1979).
Additionally, the order called for an evidentiary hearing on the turbine missile matter.
The hearing was held in June 1979.
Although it had been our original intent to decide th~e issue together with another, but unrelated, safety question (pumphouse settle-ment) which also had been explored at that hearing, it did not prove possible to do so.
For, as noted in our decision with regard to pumphouse settlement, 5I "new information of
_4/
Of the parties before the Licensing Board, only the applicant and the NRC staff have participated to any extent in our examination of the missile issue.
Although, with our leave, the Union of Concerned Scientists filed an amicue curiae brief in connection with the staff's initial response to ALAB-491, that organization did not seek to involve itself in the proceeding thereafter.
_5/
ALAB-578, 11 NRC 189, 191 (1980).
\\v.
\\
t' 4
potential importance to the turbine missile issue has recently been brought to our attention, requiring us to withhold our disposition of that issue to await further 1
developments."
Over the past two years, those developments have been unfolding and, on a periodic basis, reported to us.
We are only now in a position to reach a satisfactory determination on the turbine missile issue.
We do so in this opinion.
5 I.
A.l.
There is no disagreement that the ascertainment of the annual probability (referred to as P4) that a safety-related component will be damaged as a result of a i
l turbine missile involves the determination and j
multiplication of three subsidiary annual probabilities: the l
probability (P ) that a turbine disc will break and a y
1' portion thereof will penetrate the turbine casing and become a missile; the probability (P ) that that missile will 2
strike a reactor structure containing a safety-related l
component; and the probability (P3) that the missile then will damage.the component.
As stated in equation form, P4=Py 2
3*
xP xP 6/
The reasons why we allowed operation of the North Anna
~~
units during the period that the turbine missile issue remained open are set forth in ALAB-589, 11 NRC 539 l
(1980).
l l
T k
/
y 0
5 In its initial response to ALAB-491, 1! the staff
-4 first assigned a value of 10 to Py; i.e.,
it assumed that, for'each turbine, there was one chance in ten thousand that during the course of a year a missile would be gen-erated.
8/
This figure was arrived at on the basis of
" historically observed turbine failures for the last 20 years or so."
EI In addition, the staff apparently assumed that the primary cause of the breaking of a turbine disc would be overspeed stress:
i.e.,
the stress on the disc produced by a rapid increase in the speed of the turbine rotor as a result of a dropping of the electric load.
This is seen from the staff's observation'that its reliance on historical turbine failure rates was conservative in that
~~7/
" Response to Atomic Safety and Licensing Appeal Board's Request for Information on the North Anna Units 1 and 2 Regarding Missiles," dated September 15, 1978 (hereinafter " Staff Response").
~~8/
See Enclosure 2 to Staff Response, entitled Task Action Plan for Task A-37, at p. 37/4.
~~9/
Staff Response, pp.
4-5.
It elsewhere appeared that the source of the data was Bush, Probability of Damage to Nuclear Components Due to Turbine Failure, 14 Nuclear Safety 187-201 (1973).
s
_g D
6
" modern turbines have substantial design improvements in terms of materials and overspeed protection systems." 10/
!!aving settled upon a P value, the staff then turned y
to P and P It f und the probability that a generated 2
3 missile would. strike a vital reactor structure (P ) to be 2
-1 2 x 10 (two chances in 10).
It further conservatively aasumed that P3 = 1; i.e.,
that any such strike perforce 5.ould cause unacceptable damage within that structure.
Thus, the staff's ultimate result was that the upper limit probability of turbine missile damage to a safety-related
-5 l
component was 2 x 10 per turbine year -- that is, 10 x (2 x 10-1) x 1, 11/
-4 We were told by the staff that, notwithstanding what it deemed to be the conservative assumptions built into that computation, it nevertheless had required the applicant "to 10/
Staff Response, p.
5.
~~11/
Ibid.
See also, North Anna Safety Evaluation Report, Supp.
2, Section 10.7.
It should be noted that the Staff Response did not explicitly set forth a value for P
This value was readily determinable, however, g2v.en the assigned values for P and P and the y
3 calculated value for P 4 i
l l
.r
a m
4 A
2
%-+--
7 adhere to certain measures designed to further reduce missile risks." 12/
Our attention was specifically directed to'the turbine valve inspection requirements and the maintenance and testing procedures set forth in its Safety Evaluation Report for the facility. 13/
In the totality of these circumstances, the staff saw no reason to halt the operation of the facility. 11#
2.
The evidence presented by the staff and applicant at the June 1979 hearing similarly was rooted in the premise that any turbine failure likely would be brought about by brittle or ductile fracture 1EI stemming from an overspeeding-i 12/
Staff Response, p.
5.
13/
Id. at pp.
5-6.
14/
Id. at p.
6.
~~15/
Such fractures are the product of mechanical forces being brought to bear upon the object.
Susceptibility 4
to a breakup of that nature under a particular stress is influenced by, inter alia, the toughness and malleability of the material.
t
8 of the turbine due to loss of electric load. 16/
For its part, the staff essentially adhered to its prior analysis 1
and the conclusion reached therefrom that P could 4
-5 conservatively be assigned an upper limit of 2 x 10 per turbine year. 12 The applicant's testimony stressed that the historical data on turbine disc failures (upon which the staff had based its P value of 10~4) had been derived from old steam y
turbines and that the turbines now in use have a lower 16/
With respect to the possibility of failure induced by
-~
stress corrosion (i.e. cracking caused by a combination of relatively high stress and corrosive environments),
it was assumed that any corrosive impurities in the steam would be deposited in the low pressure areas of the turbine where steam condensation occurs and that the routine water chemistry monitoring performed at North Anna would prevent serious problems.
See Staff Testimony, foll. App. Tr. 580, at 24-25, 37-38; see also App. Tr. 598-600.
17/
Staff Testimony, foll. App. Tr. 580,at_30.
The staff added, however, that it deemed 7.3 x 10 to be a more realistic estimate. Id. at 55.
4 9
probability of brittle or ductile fracture because of such considerations as improved materials and quality assurance and better overspeed control. 18/
In this connection, we were furnished with a description of the North Anna turbines, their overspeed detection and control devices and the extensive inspection system employed to ensure the reliability of those devices, 1E With these factors in mind, and utilizing a fault tree analytic method, 20/ the applicant arrived at the conclusion that the annual per turbine probability of generating a destructive overspeed 18/
VEPCO Testimony on Probability of Generating Turbine Missiles and Turbine Overspeed Protection System, foll.
App. Tr. 19, at 9.
See also App. Tr. 478-82.
~~19/ VEPCO Testimony, fn. 18, supra, at 2-5; VEPCO Testimony on '_ and P3 and Turbine Inspection, foll. App. Tr. 19, at 6-8.
This testimony is reproduced in the Appendix to this opinion.
20/
It turned out that the witness who testified at the hearing on the fault tree analysis was unable to supply the foundation data employed in that analys:s:
i.e.,
the individual values of the various components
(" root events') of the tree.
These data were subsequently furnished under protective order.
See ALAB-555, 10 NRC 23 (1979).
(
e 4
-'d 10 i
-6, 31/
Insofar as an overspeed missile would be 1.7 x 10 within design limits was concerned, the probability of a missile resulting from such an event was estimated to be
-10, 32/
1.05 x 10 B.1.
In December 1979, while we still had the turbine missile matter under advisement, the staff informed us that cracks had been found in "several" low pressure turbine 21/
VEPCO Testimony, fn. 18, supra, at 6-8.
" Destructive overspeed" was defined as "the lowest calculated speed at which a low-pressure rotor disc will burst, based on the average tangential stress being equal to maximum ultimate tensile strength of the disc material, assuming no flaws or cracks in the disc."
Id. at 6.
22/
Id. at 8.
The applicant explained that " [a] design overspeed of 120 percent of rated speed is based on the precept that, should the turbine speed governing system be incapacitated so that on loss of full load the turbine is tripped by the overspeed trip mechanism, the calculated speed attained will not excced 120 percent of rated speed.
Turbine rotors are designed for this condition with appropriate margin and tested to 120 percent of rated speed."
Id. at 6.
The applicant's testimony also addressed the mode of calculation of P2 and P3.
See VEPCO Testimony on P2 and P3 and Turbine Inspection, foll. App. Tr. 19.
See also, Final Safety Analysis Report, Table 10.2-4, Figures 10.2-4 and 10.2-5; Amendment 36 to the FSAR, Figures 10.2-1 through 10.2-3.
For present purposes, it is unnecessary to discuss that evidence in detail.
11 discs of Westinghouse manufacture. 21/
The report indicated not only that disc cracking was much more widespread than previously assumed, but also that stress corrosion appeared to be involved.
Further, the staff stated that Westing-house was in the process of recalculating the energies of possible turbine missiles.
The assigned reason was that recent tests had indicated that non-symmetric fragments of a fractured turbine disc might, as a result of their impact with other internal turbine parts, achieve higher energies as missiles than had been previously estimated. 25/
It was this information that prompted our decision to j.
hold the turbine missile issue in abeyance to await additional developments.
See pp.
3-4, supra.
In February 23/
December 12, 1979 Memorandum from M.L. Boyle to S.S.
~~
Pawlicki, entitled " Notification of ASLAB of l
Westinghouse Turbine Disk Cracking," at p.
1.
This information was transmitted to this Board under the standard Board Notification procedure instituted by the staff several years ago.
24/
Ibid.
25/
Id. at p.
2.
i
~
12 1980, we were apprised of meetings during the prior two months which involved the staff, Westinghouse, an ACRS sub-committee and utility representatives. 26/
Westinghouse had supplied the staff with information which dealt specifically with cracking in its 1500/1800 RPM turbines.
According to the staff, that information revealed that:
Since the initial observation of cracks in the disc keyway, Westinghouse has been training teams to inspect the turbines of all units, both nuclear and fossil-fueled, that are suspect.
To date, 19 rotors have been inspected and 10 have been observed to have cracked discs.
Of the 183 discs inspected 14 have been cracked.
Cracks up to 0.4-inch in depth have been observed in discs with as little as 40 months service.
In addition, two discs in a turbine with 78 months service were observed to have bore cracks of up to 1.2-inches depth and 2.5-inches length but no keyway cracks.
Other cracks have been observed in the face of discs be-hind a spacer and in the lip of the spacer groove.
Analytical examinations reveal the presence of chlorides and hydroxide in the keyway cracks, but the initiation of stress corrosion has not been correlated quantitatively with secondary water chemistry.
The cracks in the discs bores do not appear to be caused by stress corrosion.
Westing-house has developed ultrasonic techniques that do not require removal of the discs, however, an inspection requires approximately 14 weeks for one end or approximately 19 weeks for both ends when performed at the Westinghouse facility at Charlotte, North l
Carolina.
It is hoped that inspections can be 1
26/
February 5, 1980 letter from Daniel T.
Swanson to this Board.
I
13 performed in the field in 1980. 27/
In light of these disclosures, we asked the parties to address further in written submissions the matter of crack initiation and growth.
The responses to that request pointed do data -- derived partially from the widespread turbine disc cracking experienced in Great Britain in the early 1970s -- which were said by the staff to establish that, once initiated, cracks grow at a rate influenced by
" disc material and heat treatment, keyway and bore design, temperature of operation, and to some degree steam chemistry." 28/
The staff also took note that, by the employment of standard linear elastic fracture mechanics theory,'it is possible to determine at what length the crack
--27/
January 9, 1980 memorandum from William J.
Ross to A.
7 Schwencer, entitled " Summary of Meeting with Westinghouse Related to Cracking in Turbine Rotors," at pp.
1-2.
(This memorandum accompanied Mr, Swanson's letter, fn. 26, supra).
28/
See Attachment 1 to "NRC Staff Response to Appeal Board's Memorandum and Order of March 3, 1980," dated March 24, 1980, at pp.
1-2.
Although not there specifically stated, it appears that, following initial development, cracks tend to grow linearly with time (i.e., at a constant rate).
See fn. 41, infra.
i l
l l
t
O s
0 14 would become critical; i.e., might cause disc failure. 2g/
2.
During the remainder of 1980 and the first half of 1981, the applicant (through its contractor Stone and Webster), Westinghouse and the staff continued their analyses of the cause of turbine disc cracking and the rate of growth of initiated cracks.
These analyses proceeded against the background of two intervening events: (1) the discovery of cracks in a turbine in the Farley facility which indicated a crack growth rate larger than that previously encountered; 30/ and (2) an ultrasonic
--29/
Id. at p.
4.
More particularly, the following formula is employed to determine critical crack size (Acrit}*
2 A
0 ic
=
crit
(
1.21n o
Q is a complex function related to the shape of the assumed crack and the ratio of the applied stress to the yield strength of the material; o is the nominal stress at the bore; and K is the fracture toughness g6alue is usually obtained of the material.
The K from the empirical relakfonship developed with the use of Charpy V impact test results.
See Barsom and Rolfe, Correlations Between Kic and Charpy V - Notch Test Results in the Transition - Temperature Range, American l
Society for Testing and Materials, Special Technical l
Publications, 466, 1970, pp. 281-302.
t
~~30/
See December 23, 1980 letter from Thomas M. Novak to this Board.
l
15 inspection of the North Anna 1 turbines during a routine shutdown in early 1981, which located two cracks in the i
keyway of the rotor of one of them.
The latter cracks were, respectively, 0.360 inches and 0.2 inches in depth and 1.0 i
inch and 0.5 inches in length. 31/
A later examination of one of them showed that "[t]he major portion [of cracking was] * *
- developed by intergranular stress corrosion." 32/
Confronted with this information, the applicant decided to replace the rotor with a compatible one from the disabled Unit 2 of the Three Mile Island facility (an ultrasonic inspection of the latter had disclosed that it was free of defects).
This replacement was accomplished in late February 1981.
In an unpublished memorandum and order entered on February 23, ILJ1, we approved the resumption
~~31/
See January 15, 1981 Preliminary Notification of Event or Unusual Occurrence (PNO-II-81-05), attached to January 21, 1981 letter from Thomas M.
Novak to this Board.
-~32/
VEPCO Exhibit V-22, entitled " Investigation of Keyway Cracking in LP Turbine Discs, Interim Data Summary" (October, 1981), at p.
5-1.
This investigation was conducted by the Southwest Research Institute of San Antonio, Texas, i
16 of Unit 1 operation at such time as the staff was satis-fied that the rotor had been properly installed. SSI The timing of the next inspection was left open. 34/
3.
In May and June 1981, Westinghouse supplied the applicant with three reports on the subject of turbine disc cracking and inspection.
Copies of these documents were thereafter sent to the staff and this Board as VEPCO 33/
As the February 23 order reflects (at p.
3), this approval was founded upon not only the documentary submissions to us but also, additional information which we obtained during our visit to the North Anna Unit 1 turbine building on February 19, 1981.
Unit 1 actually resumed operation on April 6, 1981.
--34/
In that connection, we explained that, during the course of our site visit, the staff had stated that it was reconsidering its analytical model pertaining to crack growth in light of the then recent cracking experience at the Parley facility.
February 23 order, at p.
3.
It should be noted that, throughout this period, our concern was focused upon Unit 1.
Unit 2 did not commence operation until mid-1980 and we were satisfied from the evidence at hand that the turbine disc cracking phenomenon is time-related.
See ALAB-589, 11 NRC 539, 540 (1980).
See also November 18, 1980 order (unpublished) at fn.
1.
In an April 3, 1981 order (unpublished), however, we recorded our exoectation that the Unit 2 turbines would undergo an ultrasonic inspection at the first refueling shutdown of that unit.
i I
l l
i i
17 25! and V-3, 22 with the notation Exhibits V-1, 22! V-2 that they were considered by Westinghouse to include proprietary information.
In October 1981, the applicant filed, as its Exhibit V 4, a Stone and Webster report concerned with the probability of turbine missile damage at North Anna. $ ! Portions of this document likewise were asserted to contain proprietary information.
Because of our agreement to treat as confidential disclosures in those reports which might reasonably be taken
---35/
Criteria for Low Pressure Nuclear Turbine Disc Inspection, June 1981.
I 36/
Missile Energy Analysis Methods for Nuclear Steam Turbines, May 1981.
--37/
Procedures for Estimating the Probability of Steam Turbine Disc Rupture from Stress Corrosion Cracking, May 1981.
--38/
Summary Report: Turbine Missile Damage Probability Analysis for North Anna Units 1 and 2.
n_ _
18 l
to be proprietary in character, we confine ourselves here to a summary of their content with no direct reference l
to any such disclosure. SEI a.
Exhibit V-1.
This report addresses the matter of the frequency of ultrasonic inspections required to prevent a turbine disc breakup as a result of cracking. 40/
The inquiry involves consideration of the causes of cracking; the growth rate of a crack once initiated; and critical crack size (which is influenced by the shape of the particular crack).
With regard to the last-mentioned
--39/
Westinghouse's first claim of entitlement to a protective order was advanced in connection with the data in support of its fault tree analysis.
See fn.
20, supra.
Ali hough finding that the affidavit submitted wiu, that claim did not satisfy the requirements laid down in Kansas Gas and Electric Co.
(Wolf Creek Duclear Generating Station, Unit No. 1),
ALAB-327, 3 NRC 408, 416-18 (1976), we nonetheless decided to provide the requested protection.
This was because no party had interposed an objection and we were desirous of obtaining the information in question "without untoward further delay."
See ALAB-555, 10 NRC 23, 27-28 (1979).
That protection was then automatically extended to all further Westinghouse submissions asserted to contain proprietary i
information.
As noted in ALAB-555, however, the course
~
which has been followed in this instance is not to be taken as a precedent.
Id. at 28.
--40/
As previously noted, Westinghouse has devised techniques for conducting such inspections of low pressure turbine discs without turbine disassembly.
See p.
12, supra.
See also, VEPCO Exhibit V-7, entitled " North Anna Unit #1 Low Pressure Disc Ultra-1 sonic Inspection Report" (January 16, 1981).
l t
~ _. _.
e 19 factor, Westinghouse resorts to the linear clastic fracture mechanics formula which is discussed above.
See fn. 29, suprh.
Furthermore, according to the report, there are many variables affecting the rate of growth of a crack following its short initiation period.d1/
Those variables include environmental conditions (i.e., whether corrosive elements are present), the location of the crack and temperature.
See p.
13, supra.
Because the turbine environment may vary widely from facility to facility, Westinghouse relies upon available experimental data on crack development to establish ranges for the rate of crack growth.
In essence, Westinghouse arrives at its recommended l
inspection intervals by employment of a crack growth rate model developed from the accumulated data pertaining to crack initiation, crack shapes and rates of growth.
According to the report, the ultimate aim is to have the inspections frequent enough to insure that no crack reaches 3/4 of its calculated critical size without detection.
In this connection, the report noted that several conservative assumptions undergirded the critical size calculation, with
--41/
The report indicates that, during that initiation period, the crack may grow rapidly; experimental test results show, however, that there is then a slower, constant growth rate until critical crack size is reached.
i
20 the result that the actual critical size would be substantially greater.
b.
Exhibit V-2.
This comprehensive treatise (approximately 300 pages in length) is concerned with turbine missile energy and is divided into three parts.
Part A discusses "the kinetic energy of external missiles that hypothetir:11y could result if a turbine rotor were to rupture at normal speed, design overspeed, or destructive overspeed" for units in nuclear power facilities.
It considers the methods of " calculating the kinetic energy of postulated rotor fragments before any interaction with the turbine stationary parts."
Exh. V-2,
- p. A4.
Part B "is concerned with the determination of whether or not a disc burst will result in missiles being ejected from the turbine casing, and if so, the external kinetic energy properties of those missiles."
It includes the results of analytical procedures, and of laboratory tests which confirmed some portions of the analyses.
Exh. V-2, p.
B5.
The basic analysis follows the Hagg-Sankey method and is applied specifically to turbines and their casings.
21 Id. at B13. 42/ It is not a probabilistic study of missile penetration but, rather, essentially a deterministic analysis of the consequences of various size disc fragments striking the casing.
This initial strike might be followed by repeated impacts from the same fragment or other fragments with additional resultant damage.
Part C describes the methods and procedures for evaluating the effects of disc fragments hitting the blade rings and outer cylinder of the turbine.
The effects of fragments hitting four representative locations around the outer cylinder are considered.
c.
Exhibit V-3.
This document focuses upon turbine l
disc rupture occasioned by stress corrosion cracking.
It uses data from Westinghouse turbine experience to determine (1) the location of cracks (i.e.,
in the bore itself or in one or more keyways or both); and (2) the range of crack growth rate (s) for various crack locations.
As noted in Exhibit V-1, this study supports the theory that, whether located in the bore or in a keyway, the cracks are the product of stress corrosion, since no cracking has been observed in dry steam areas.
~~42/
See Hagg and Sankey, The Containment of Disc Burst Fragments by Cylindrical Shells, American Society of Mechanical Engineers Paper No. 73-WA/PWR-2, August 1, 1973.
i e
22 The report concludes with three figures (9-11) which plot the probability of missile generation (P ) against the y
inspection intervals in years.
The figures cover three turbines of differing materials.
In general, they reflect a
-5
-6 P
value of 10 to 10 for 1 to 2 year inspection y
-3 intervals for two of the turbines, but a value of 10 to
~4
-2 10 for the third.
The P probability increases to 10 to y
-3 10 for 4 year inspection intervals, d.
Exhibit V-4.
The purpose of this Stone and Webster report, directed specifically to the North Anna facility, was to " compute the annual probability P of 4
significant damage to plant structures, systems, and components resulting from postulated turbine failure."
Exh.
V-4, p.
1.
As earlier noted (p.
4, supra), P is the 4
product of three factors, Py,P2 3'
and P The report assumes, as had the initial staff presentation to us, AS a 10 value for P It then
-4 1
independently calculates values for P and P in rder to 2
3 arrive at the P value.
The final summation of the 4
results of the analysis is contained in the following 43/
See p.
5, supra.
i 1.
=-
. ~
l 1
23 j
table (Table C-19 of Appendix C to Exh. V-4): -44/
i
[
r i
i i
f i
i i
i.
i i
1 i
i I
i
-44/
As used in the table, the term " unit" refers to a single turbine and the term " operating period" to the length of time between fuel loadings that the reactor was in actual operation.
r i
1 4
4 t
,an
,,-->.--,--n,-
--,--e-,wn-,,
,y----
e
--,---c.
--m---m,,-
24 TABLE C-19 TOTAL DAMAGE PROBABILITY (P )
4 UNIT 1 P;/ UNIT / OPERATING PERIOD l
OPERATING TIME (YRS.)
CRITERIA A CRITERIA B Ps = 0.05
-6 1
1.019x10 4.812x10 5.081x10
-5
-6
-6 2
2.265x10 1.053x10 2.133x10
-4
-6
-6 3
1.525x10 2.110x10 9.629x10 UNIT 2 P / UNIT / OPERATING PERIOD 4
OPERATING TIME (YRS.)
CRITERIA A(
CRITERIA B(
Ps = 0.05 1
8.676x10-4.795x10-4.988x10-
-5
-6
-6 2
1.696x10 1.381x10 2.159x10
-4
-6
-6 3
1.136x10 3.447x10 8.955x10
( I CRITERIA A Present conservative NRC approach equates the initiation of scabbing within a safety related cubicle with a damage probability of 1.0
(
CRITERIA B Slightly unconservative approach which neglects scabbing damage if missile perforation is prevented (3)
P
= 0.05 Assumes a realistic probability of s
5% for ensuing safety related damage if a missile strike results in scabbing without perforation.45/
g/
This table is not claimed to contain proprietary in forination.
" Scabbing" refers to the creation of a secondary missile (such as a piece of concrete) within a vital structure as a result of the turbine missile striking (but very likely not penetrating) a reinforced wall of that structure.
25 4.
As is seen from the foregoing, the various 3
analyses put before us by the applicant approach the problem i
at hand from two different standpoints.
Exhibits V-2, V-3 and V-4 all are addressed to the P assessment; i.e.,
to the 4
annual probability that a safety-related component would be damaged by a turbine missile.
In contrast, Exhibit V-1 is not principally concerned 2
with the P4 probability.
Rather, its focus is upon the avoidance of a stress corrosion-induced disc fracture through the mechanism of routine ultrasonic inspections at specified intervals:
i.e., upon preventing the occurrence of the event that might possibly give rise to a turbine missile and, ultimately, to unacceptable damage.
- Moreover, Westinghouse's methodology for fixing those inspection intervals has a deterministic foundation: it utilizes actual data from which it is possible to ascertain with reasonable precisivi the rate of growth of a crack in a particular turbine disc.
Being a generic study, Exhibit V-1 does not focus upon the North Anna units specifically.
In a memorandum E
to this Board, however, applicant's counsel represented that the Exhibit V-1 analysis calls for inspections every 43.3 operating months (in the case of Unit 1) and
~
26 1
overy 39.9 operating months (in the case of Unit 2). d5#
This conclusion rested upon two premises.
The first was that, unlike the No. 2 discs, the No. 1 discs in each turbine in each unit are " contained";
i.e.,
should one of.those discs fracture as a result of stress corrosion, no fragment would penetrate the protective turbine casing and thereby become a missile capable of damaging to a safety-related component.
In justificatian of this premise, counsel pointed to Westinghouse's calculations founded upon the application of its methodology for determining fragment and missile energy (VEPCO Exhibit V-2, supra) -- the results of which calculations are found in VEPCO Exhibit V-13. All The second premise was that the inspection intervals should be based upon the uncontained No. 2 disc in each unit that was most " critical":
i.e.,
had the greatest vulnerability to the production of a turbine missile should 46/
Memorandum of VEPCO's counsel on North Anna 1 and 2 Turbine Missile Analysis, dated October 21, 1981, at pp.
2, 12, as revised by enclosures to January 21, 1982 letter from James N. Christman to the Chairman of this Board.
These intervals are premised on no cracks having been discovered at the prior inspection.
If a crack had been found, its calculated growth rate would determine when the next inspection would be necessary.
47/
Id. at p.
10.
k
I o
27 cracking occur.
From an application of the Westinghouse analyses, this appeared to be the No. 2 disc at the gener-ator end of Turbine No. 1 (Unit 1) and the same disc of Turbine No. 2 (Unit 2). AE Counsel went on to note, however, that the applicant might'nonetheless elect to inspect the discs at shorter j
intervals; more particularly, after approximately 33 operating months (which represents two refueling cycles).
We were told that there was an economic rather than a safety reason for following that course.
Specifically, even though a No. I disc fracture would not (according to Westinghouse) result in a turbine missile and therefore would have no safety implications, the fragment (s) likely would seriously damage other internal parts of the turbine at substantial financial cost,to the applicant.
Accordingly, it might be in the applicant's pecuniary self-interest to establish its inspection schedules without regard to whether a particular disc was or was not contained.
In this connection, the Westinghouse analyses reflected that, to prevent the possible fracture of a No. I disc, the inspection intervals for Units 1 and 2 should be 34.9 and 32.5 operating months, 48/
VEPCO Exhibit V-6, Table 1.
~
s
28 respectively. 49/
5.
On January 22, 1982, the staff submitted the written testimony of two witnesses which addressed,' inter alia, the staff's criteria for turbine disc inspections. 50/
These witnesses, both metallurgical engineers, stated that
/
they had reviewed applicant's Exhibit V-1 and had concluded that "the inspection schedules derived by its use are consistent with [the staff's] past criteria and current understanding of the cracking problem."
They added that
"[a]dherence to these inspection schedules will' provide an
~
acceptably high degree of assurance that discs will be inspected before cracks can grow to a size that could cause f
49/
October 21, 1981 memorandum, at pp. 12-14, as revised
~~
on January 21, 1982.
Thus, as apparent, it has been determined that cracks in the No. 1 discs would have a higher growth rate than those in the No. 2 discs.
50/
NRC Staff Testimony of Harren S. Hazelton and Clifford D.
Sellers Regarding Turbine Disc Cracking.
s 29 disc failure." 51 Because this testimony did not make direct mention of the October 21 memorandum of applicant's counsel, we sought clarification from the staff on whether its witnesses were endorsing the inspection intervals which counsel had represented to be the product of the Exhibit V-1 analysis.
In its response of May 19, 1982, the staff informed us that it had not us yet completed its review o'f the Westinghouse analyses and calculations underlying the conclusion that the No. I discs are contained and consequently need not be factored into the inspection interval determination.
That being so, the staff is of the opinion that the 34.9 and 32.5 operating month intervals (see pp. 27-28, supra) should obtain at this juncture. 12/
--51/
Id.,
at 18.
It might be noted, however, that the staff's criteria include the standard that "a new disc, or a disc found free of cracks by inspection, can operate until the time calculated for a new crack to grow to one half of critical depth.'"
Ibid.
As earlier seen, p.
19, supra, the ilestinghouse standard appears to be 3/4 rather than one-half of critical size.
--52/
Supplemental Testimony of Warren S. Hazelton and Clifford D.
Sellers Regarding Turbine Inspection Schedules for North Anna 1 and 2, at 3-4.
The witnesses added, however, that the staff would likely agree to an extension of the inspection interval of up to 10% to accomodate refueling schedules for North Anna l
1 or 2.
Id. at 3.
This is because there are "such wide margins of safety incorporated into the inspection intervals."
i l
I
30 Given th.'s staff position, the applicant (through its Vice President for Nuclear Operations) has committed itself in writing to conduct the inspections in accordance with the 34.9 and 32.5 operating month schedule unless and until the staff approves a modification of that schedule. 53/
Such approval would presumably be forthcoming when and if the staff becomes satisfied that, in fact, the No. I discs are contained and, thus, the fracture of one of them would not threaten to create a turbine missile.54/
II.
On the basis of the record before us, as summarized in material part above, we reach the following conclusions.
1.
When we commenced our turbine missile inquiry in 1979, brittle fracture 55/ was apparently thought to be the principal cause of turbine disc cracking.
But such is not the case.
Rather, it is now recognized that the more 53/
May 18, 1992 letter from R.H.
Leasburg to Harold R.
Denton.
~~54/
A limited schedule alteration might also be allowed to l
synchronize the inspection with a routine refueling i
outage.
See fn. 52, supra.
l
~~55/
In the ensuing discussion, our use of the term " brittle l
fracture" includes ductile fracture as well.
l
31
- erious concern is stress corrosion.
Unlike brittle fracture, which generally is the product of turbine ov'erspeed, stress corrosion cracking can occasion a disc failure at normal turbine operating speeds, as well as under l
l startup stress.
In this connection, the evidence suggests that the very emphasis on improved material to enhance brittle fracture toughness may have also increased the susceptibility of the discs to stress corrosion. 56/
2.
So long as the applicant maintains its existing l
and extensive inspection system with regard to the l
reliability of the overspeed detection and control de-vices, 57/ there will be no undue risk to the public health and safety stemming from a turbine missile generated by a brittle fracture.
Stated otherwise, the totality of the evidence indicates, and we so find, that, if the likelihood of a destructive overspeed is minimal, the annual probability of such a raissile being created (P1) and ultimately causing damage to a safety-related component (P )
4
--56/
See Testimony of Hazelton and Sellers, fn. 50, supra, at 12-13, 57/
See p.
9, supra, and the Appendix to this opinion.
32 falls well within acceptable limits.58/
It is nonetheless obvious, however, that continued attention should be given to proper disc metallurgy, which will further serve the end of preventing brittle fracture-induced disc failures and, thus, the possible creation of turbine missiles.
3.
Although the root cause of brittle fracture is known, the same cannot be said with respect to stress corrosion cracking.
There does appear to be general agreement that such cracking is not associated with dry steam and hence it can be expected to occur only on low pressure discs where steam condenses in the rotor areas.
l But no specific contaminant in the water has been correlated with the initiation of cracks either in the rotor itself or
--58/
In this regard, two considerations should be kept in mind _yith respect to the staff's initial assignment of a 10 value to P based upon historical turbine 3
failure data 5 (which, in turn, had led to the assignment of a 2 x 10 value for P4).
See pp. 5-6, supra.
The first is that the historical data to which the staff resorted undoubtedly had included all turbine failures
-- whether occasioned by brittle fracture or, instead, i
stress corrosion.
As we have reiterated, however, the latter is now understood to be the major cause of disc cracking.
Secondly, the likelihood of brittle fracture i
was much greater then than it is now because of l
improvement in materials and quality assurance, as well as better overspeed control.
See pp.
8-9, supra.
In short, the historical data have little meaning insofar as the present probability of a brittle fracture-induced turbine failure is concerned.
l l
l
e 33 i
in the keyways. 59/
Moreover, stress corrosion may be encountered in a disc (or discs) of one turbine but not in thbse in another turbine in the same facility. 60/
Still further, while initiated cracks expand at an essentially constant rate, that rate may vary from rotor to rotor.
In this connection, it is not certain whether the maximum crack growth rate has yet been experienced; nor is it known what conditions might bring about still greater growth rates.
In these circumstances, with respect to stress corrosion-induced cracking we have chosen to eschew reliance 59/
See Attachment 1 to March 24, 1980 "NRC Staff Response to Appeal Board's Memorandum and Order of March 3, 1980," at p.
1.
The attachment was prepared by Mr.
Hazelton, who attested to its content by affidavit.
-~60/
See Testimony of Hazelton and Sellers, fn. 50, supra, at 14.
The witnesses referred specifically to the experience at the Yankee Rose facility:
"The two number one discs were of identical design, were manufactured accord;ag to the same procedures, were made from the same steel ingot, and of course were subject to as similar temperature and environmental conditions as we could possibly hope to have.
Yet one suffered hundreds of significant stress corrosion cracks, and the other had none."
It is possible, of course, that machining errors may cause misfitting between the rotor bore and the shaft onto which it is shrunk.
A bore slightly too small or a shaft too big could cause much larger stresses in some rotors than in others.
This matter was not addressed by the parties.
i l
l
34 upon the results of the Westinghouse and the Stone and Webster probability studies which have been presented to us.
While there may be no serious flaws in the analyses and calculations which underlie those results, the abiding uncertainties make such reliance imprudent.
We are reinforced in this view by what the staff told us in March 1980:
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 cause cracking of turbine disc steels, but laboratory and field tests also have shown that under the right conditions, cracks can be initiated and prcpagated 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.
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 conditions, the worst cases can be considered an upper bound to probable future crack growth rates.
Of course we
{
cannot be sure that the data include worst possible i
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. 61/
These observations seem as valid today as they were two years ago.
61/
See Attachment 1 to March 24, 1980 staff filing (fn.
28, supra?, at pp.
1-2.
4 b
35 Rather, Westinghouse's deterministic approach detailed in Exhibit V-1 commends itself to us.
As previously seen, it is possible, utilizing empirical data, to determine with reasonable certainty the length of time that will elapse before an initiated stress corrosion-induced crack might reach critical size.
The Westinghouse methodology employed in making that determination has the staff's endorsement, and our own examination of it gives us no reason to disapprove it.
Insofar as the precise ultrasonic inspection schedules for the two North Anna units are concerned, we encounter no difficulty in accepting the calculations for the No. 1 and the No. 2 discs which have been derived by the parties from the application of the Westinghouse methodology.
What remains is the question of which discs should control those schedules in the interest of providing reasonable assurance that any stress corrosion-induced crack would not bring about the generation of a turbine missile; 62/ that, in turn, hinges upon whether the No. 1 discs are contained.
We might, of course, hold this proceeding in still further abeyance to await the completion of the staff's review of the Westinghouse analyses and calculations which produced an
~~62/
We are not here concerned with what inspection intervals might be advisable from a purely economic standpoint; that is for the applicant to determine.
l l
T 4
36 affirmative answer on the containment matter.
- Given, however, the applicant's willingness to commit itself to the inspection intervals founded upon the No. 1 discs until such time as it receives the staff approval for a schedule alteration, we see no compelling necessity to prolong our already extended involvement in the turbine missile inquiry. 63/ Among other considerations, it is most likely that, the outcome of this specific staff review to one side, the continuing investigation of the causes and consequences of stress corrosion-induced disc cracking will bring to light new information bearing upon appropriate inspection schedules.
Thus, nothing we might prescribe in that regard could be expected to have validity in perpetuity.
This being so, it is best to decide the matter now on the basis 63/
It should be noted in passing that it might not make any practical difference whether the inspection schedules are geared to the No. 1 or,. instead, the No.
2 discs.
The end of the 43.3 and 39.9 operating month inspection intervals required (on the basis of currently available information respecting crack growth rates) for the No. 2 discs will be reached during the third fuel cycle (each fuel cycle being of approximately 16 to 17 operating months in duration).
This being so, in all events, the applicant likely would. conduct the inspections at alternate refueling outages to avoid having to shut down in the middle of a
-fuel cycle for that purpose.
Additionally, there is the already discussed consideration of basing the inspections on the No. 1 discs to minimize the possibility of economic loss. occasioned by a fracture of one of those discs.
O 37 of the information currently at hand, leaving it to the staff to deal with fresh developments as they occur in the fulfillment of its role as the ongoing monitor of nuclear facility operation.
For the reasons heretofore stated, we find reasonable assurance that the full-power operation of the North Anna Units 1 and 2 steam turbines will not pose an undue risk to the public health and safety provided that:
1.
The inspection procedures pertaining to overspeed detection and control are' maintained.
Any modification of the existing procedures are to be subject to prior staff approval.
2.
The turbine discs of Units 1 and 2 are subjected to ultrasonic inspection at intervals of no greater than 34.9 operating months (Unit 1) and 32.5 operating months (Unit 2) unless and until the NRC staff authorizes an increase in those intervals. 64/
64/
Although not likewise a requirement, we strongly
~~
suggest that the turbine vendor be urged to continue its disc redesign effort now underway.
The proposed design modification shown to us during our visit to the i
North Anna facility last year appeared to be a step in the right direction.
~
o
~
38 The staff is to prescribe the manner in which these provisos are to be memorialized.
In this connection, we note simply our belief that it should not prove necessary to include them in the facility's technical specification.65/
--65/
To be sure, certain existing inspection procedures pertaining to overspeed detection and control devices are now included in the technical specifications for at least Unit 1.
See pp. 45-46, infra.
All that we mean to suggest is that there are adequate means available for insuring observance of requirements of the type hereinvolved short of conversion into technical specifications.
The May 19, 1982 letter from staff counsel which accompanied the supplemental testimony of Messrs. Hazelton and Sellers (see fn. 52, supra) bears this out.
Mr. Swanson noted that the applicant will be revising the FSAR for the two units to reflect the turbine inspection intervals to which it has committed itself.
He went on to observe that, as a result, the inspection schedules will be subject to the restric-tions of 10 CFR 50.59.
For this reason, the staff "does not intend to impose [those schedules] in the form of a license condition or technical specification in the North Anna Units 1 and 2 operating licenses."
m
39 It is so ORDERED. N!
FOR THE APPEAL BOARD b-h mbd C. J n Shoemaker Secre ary to the Appeal Board l
-66/
This decision terminates our review of all issues in this proceeding other than that related to radon emissions in the mining and milling of uranium fuel.
The radon issue will be addressed in this proceeding (in which it is not in contest) following its i
resolution in other proceedings in which it is i
contested.
i r
i
40 APPENDIX 1.
Excerpt from VEPCO Testimony on Probability of Generating Turbine Missiles and Turbine Overspeed Protection System, Introduced into Evidence Following App. Tr. 19, at pp. 2-5 II.
DESCRIPTION OF NORTH ANNA TURBINES The turbine-generators for North Anna 1 and 2 were designed and manufactured by the Westinghouse Electric Corporation, which has supplied turbine generators to industry for over 75 years.
Each turbine at North Anna is a conventional 1800 rpm tandem-compound unit, consisting of one double-flow high-pressure cy-linder and two double-flow low-pressure cylinders.
Each tur-bine is provided with four moisture separator reheaters.
Tur-bine extraction connections supply steam to six stages of feed-water heaters.
Each high-pressure steam pathway to the high-pressure cy-linder contains a throttle valve and a governor valve.
A reheat stop valve and an interceptor valve are provided in each cross-over pipe between each moisture separator and each low-pressure turbine cylinder.
The turbine control system is of the electro-hydraulic type, ensuring rapid speed of response and control of turbine operation.
O 41 The protective devices for the turbine include a low bearing oil pressure trip, a solenoid trip, overspeed trips, a thrust bearing trip, and a low vacuum trip.
The solenoid trip will be actuated by malfunctions of the Steam and Power Conversion System, such as a reactor trip, generator trip, or loss of electro-hydraulic governor power.
The control system includes an overspeed protection con-troller, which acts to limit turbine speed in case of a load separation.
The controller operates to close the turbine governor valves and the interceptor valves until the overspeed condition is corrected.
Nonreturn valves are installed in the turbine extraction steam lines to minimize turbine overspeed following a trip.
The North Anna turbines are equipped with an overspeed protection system consisting of an overspeed pro-l tection controller ("OPC"), a mechanical overspeed trip, and an electrical overspeed trip designed to operate as follows:
A.
OPC--Anticipator (30% load or greater)
When the turbine is operating at 30 percent load or greater (as measured by a pressure corresponding to steam flow) and a load separation occurs opening the generator circuit breaker, anticipator logic in the system opens two redundant solenoid valves in the hydraulic oil system, causing the turbine governor valves and interceptor valves to close in anticipation of over-speed.
l w
s s
-42 B.
OPC--Auxiliary Speed Channel (103% overspeed)
An auxiliary speed channel that shares none of the com-ponents of the overspeed systems described below (except the power supplies that are redundant) receives frequency pulses generated by a separate reluctance pickup and converts them to a proportional analog signal for control of overspeed.
If the turbine overspeed exceeds 103 percent of normal rated speed, this system will open the two redundant solenoid valves in the hydraulic oil system mentioned above and close the tur-bine governor and interceptor valves (if they have not already been closed by the anticipator).
C.
Mechanical Overspeed Trip (110-111% overspeed)
At about 111 percent of rated speed the mechanical over-speed trip system will cause the flow of steam into the turbine to cease.
This mechanism consists of a trip weight that is carried in a transverse hole in the rotor extension shaft with its center of gravity offset from the axis of rotation, so that centrifugal force tends to move it outward at all times.
The trip weight is held in position by a compression spring.
If the speed of the turbine increases to a speed above the setpoint, the centrifugal force overcomes the compression of the spring, and the weight moves outward and strikes the trip trigger, which initiates a sequence of events causing all valves i
i i
43 capable of admitting steam to the tur'bine (that is, the throttle, governor, reheat stop, and int rceptor valves and extraction nonreturn valves) to close.
D.
Electrical Overspeed Trip (110-111% overspeed)
Another method of tripping the turbine on overspeed is provided by the primary speed channel, which receives a con-tinuous turbine speed signal from a variable reluctance trans-ducer mounted at the turbine shaft.
(This primary speed channel is the one used by the normal governing system but is separate from the auxiliary speed channel mentioned above.)
The trans-ducer output is converted to a precise d-c signal with a level proportional to turbine speed.
At 111 percent of turbine speed, this system operates the solenoid trip located on the emergency trip block.
Actuation of the solenoid trip will initiate a se-quence of events causing all valves capable of admitting steam to the turbine to close.
2.
Excerpt from VEPCO Testimony on P2 and P3 and Turbine Inspection, Introduced into Evidence Following App.
Tr. 19, at pp. 6-8 IV.
INSPECTION AND TESTING Detailed procedures for testing the cverspeed trip system j
have been developed for North Anna Units 1 and 2, as stated in the FSAR:
s s
44 1.
A thorough check of the throttle and governor valve stem freedom will be made once each week.
2.
A thorough check of the reheat stop and interceptor valve stem freedom will be made once each week.
3.
Motor-driven oil pumps and controls will be tested once each month.
During normal operation this procedure in-volves testing the bearing oil pump pressure switch by reducing the pressure through the bleed-off valve to a point where the switch makes contact, completing the circuit to the AC pump motor.
The emergency oil pump pressure switch can be tested by continuing to reduce the pressure to the point where this switch makes contact and operates the emergency oil pump.
The actual pressure at which each switch operates is compared to the pre-scribed setting.
4.
The following oil trip test devices located at the governor-end pedestal will be tested prior to each turbine startup (if they have not already been tested the previous week) :
a.
Overspeed trip oil test device b.
Low vacuum trip c.
Low bearing oil pressure trip d.
Thrust bearing oil trip 5.
The overspeed trip will be tested by overspeeding the turbine-generator unit during each refueling.
4 45 The Technical Specifications for North Anna Unit 1 include certain inspection requirements:
4.7.1.7 The structural integrity of the steam turbine assembly shall be demonstrated; a.
At least once per 40 months, during shutdown, by a visual and surface inspection of the steam turbine assembly at all accessible lo-cations, and b.
At least once per 10 years, during shutdown,-
i by disassembly of the turbine and performing a visual, surface and volumetric inspection of all normally inaccessible parts.
4.7.1.8.2 The above required turbine overspeed protection system shall be demonstrated OPERABLE; a.
At least once per 7 days by cycling each of the following valves through one complete cycle.
1.
4 Turbine Throttle valves 2.
4 Turbine Governor valves 3.
4 Turbine Reheat Stop valves 4.
4 Turbine Reheat Intercept valves b.
At least once per 31 days by direct observation of the movement of each of the above valves through one complete cycle.
c.
At least once per 18 months, by performance of CHANNEL CALIBRATION on the turbine over-speed protection instruments.
s s -,
J i
i 46 i
d.
At least once per 40 months, by disassembly of at least one of each of the above valves l
and performing a visual and surface inspection 1
of all valve seats, discs-and stems and verify-ing no unacceptable flaws or corrosion.
)
i i
t i
F W
t l
l
{
t t
i
~
l 1
i T
i i
-