ML20029D406
| ML20029D406 | |
| Person / Time | |
|---|---|
| Site: | Pilgrim |
| Issue date: | 04/18/1994 |
| From: | Taylor J NRC OFFICE OF THE EXECUTIVE DIRECTOR FOR OPERATIONS (EDO) |
| To: | Studds G HOUSE OF REP. |
| Shared Package | |
| ML20029C549 | List: |
| References | |
| CCS, NUDOCS 9405050321 | |
| Download: ML20029D406 (10) | |
Text
V7 ')
l (9
i'S 3
UNITED STATES
($g$
^!
NUCLEAR REGULATORY COMMISSION
[
WASHINGTON, D.C. 20555-0001 April 18, 1994 The Honorable Gerry E. Studds Member, United States House of Representatives 1212 Hancock Street Quincy, MA 02169
Dear Congressman Studds:
I am replying to your letters of February 9 and March 1,1994, in which you asked the U.S. Nuclear Regulatory Commission (NRC) to respond to concerns raised by your constituents regarding the main turbine at the Pilgrim Nuclear Power Station. The letter from Robert A. Pollard, Union of Concerned Scientists, enclosed with your February 9 letter, included a list of specific questions to be addressed. The responses to that set of questions are given in the enclosure.
In his letter, Mr. Pollard also expressed concerns about NRC's overall treatment of this issue relative to satisfying regulatory requirements and guidance.
In this regard it is important to understand the basis, with respect to turbine integrity, upon which the Pilgrim plant was granted an o)erating license. Turbines are not nuclear safety related equipment; t1erefore, NRC's guidance on maintenance of turbine integrity has been limited.
Pilgrim was licensed based on evaluations demonstrating that failure of the turbine would have no adverse effect on nuclear safety systems because those systems are adequately protected by physical barriers capable of withstanding the impact of the worst hypothetical turbine disk failure.
Based on Pilgrim's present analysis, the NRC concludes that the Pilgrim plant turbine does not pose a threat to plant safety-related equipment required to mitigate an accident. A more detailed discussion of this subject is contained in our Response No. 1 in the enclosure.
Finally, with regard to Ms. Ott's letter, which you enclosed with your letter of March 1, 1994, there is no regulatory requirement pertaining to turbine disk failure probability.
The NRC expects licensees to satisfy the applicable guidance documents that exist at the time of plant construction (e.g., NUREG reports and regulatory guides), or to propose alternatives that would be evaluated by the NRC staff as was the case with Pilgrim.
The internal NRC memorandum that Ms. Ott referenced in her letter to you incorrectly referred to " guidance documents" as being NRC requirements.
This incorrect use of the o$o00j93 m nas e
~
m J
\\
RESPONSE TO RE0 VESTS FROM CONGRESSMAN STUDDS RELATING TO TURBINE DISK CRACKING IN PILGRIM NUCLEAR POWER STATION. UNIT 1 In April 1993, General Electric (GE) inspected the rotor in low pressure turbine "A" (LPA) at Pilgrim Unit 1 and found flaw indications in disk numbers 4, 5, 6, and 7.
Based on its analysis, GE recommended that Boston Edison Company (the licensee) either remove the seventh stage disk on the generator side (7GA) or warm the LPA rotor before starting.the plant. The licensee later retained Structural Integrity Associates, Inc. (SIA) to evaluate the flaw indications in disk 7GA.
In June 1993, the NRC staff reviewed the GE and SIA analyses and assessed the failure margins of disk 7GA.
On August 4,1993, the NRC provided its evaluation as part of a response to an inquiry from U.S. Congressman Studds, U.S. Senator Kerry, and U.S. Senator Kennedy regarding the cracks in the Pilgrim turbine.
On February 9 and March 1, 1994, Congressman Studds requested the NRC to respond to questions raised by his constituents regarding the cracks in the LPA turbine in the Pilgrim Nuclear Power Plant. The following are responses to Congressman Studds' request:
Reouest No. 1:
Identify all existing regulations and regulatory guidance (e.g., regulatory guides and policies) that are relevant to ensuring that turbine missiles do not pose an unacceptable risk to public health and safety.
If any of these are not applicable to the Pilgrim plant, explain why.
Response
The following NRC regulations and regulatory guidance are related to the subject of turbine integrity:
A)
General Design Criterion 4 (GDC 4) of Appendix A to Title 10 of the.co_d_g of Federal Reculations (10 CFR Part 50)
B)
Standard Review Plan (NUREG-0800) Sections 3.5.1.3 and 10.2.3 D)
NUREG-1048, " Safety Evaluation Report Related to the Operation of Hope Creek Generating Station," Supplement No. 6, U.S. Nuclear Regulatory Commission July 1986.
The Pilgrim Nuclear Power Plant was licensed before the issuance of any of the above regulations or regulatory guidance. However, to satisfy the staff criteria that were in place at the time the plant was licensed (1971), the licensee performed a turbine missile analysis in 1971, using the worst-case assumptions, to show that the plant design provides )rotection from turbine missiles. The analysis assumed a failure of the tur)ine disk at 110 percent of rated speed, with the disk separating into three 120-degree segments. This postulated failure mode was determined to yield the most severe impact on e
. structures (i.e., a missile with the highest impact energy). The licensee's analysis was consistent with the later requirements of GDC 4 and demonstrated that the structural barriers in the facility (i.e., walls and floors) are capable of protecting safety systems from turbine missiles. This analysis is discussed la Section 11.2.4 of the Pilgrim Final Safety Analysis Report (Re f. 1).
Reauest No. 2:
Explain the bases for allowing the Pilgrim plant to operate with a probability of turbine disk failure greater than 1 in 100,000 per year until the next scheduled cutage, but not after that outage.
Response
The licensea, on his own initiative, decided to replace the turbine during the last refueling outage. The delivery date of the replacement turbine coincides with the next outage. While staff has no objection to the licensee's decision to replace the turbine, the staff did not require the replacement. This is because the plant, with its existing turbina, continues to meet the licensing basis for the plant and does not pose an unacceptable risk to public health 4
and safety. Moreover, allowing operation of the Pilgrim turbine with a failure probability in the range of 1.0E-5 to 1.0E-4 per year is consistent i
with the current NRC turbine system reliability criteria. Criterion B of that guidance states that for an unfavorable turbine orientation with a failure probability in the range of 1.0E-5 to 1.0E-4 per year, the turbine may be kept in service until the next scheduled outage, at which time action should be 1
taken to reduce the turbine missile generation probability to meet the criterion of 1.0E-5 per year before returning the turbine to service.
However, these criteria were not in force when the Pilgrim Plant was licensed and, therefore, do not apply to Pilgrim. The NRC turbine system reliability criteria are contained in Reference 2.-
Reauest No. 3:
For each parameter used in the turbine failure analysis described in the NRC's August 4,1993, response, explain the basis for the value used and state whether that represents the worst case assumption. The parameters to be discussed include turbine speed, disk temperature, crack size, crack growth rate and critical flaw (cracx) size.
Response
The following parameters were assessed in the staff's review of the Pilgrim turbine analyses:
i
. A) Turbine Speed j
The GE and SIA analyses were both based on the rated speed of the turbine, j
which is 1800 rpm. The NRC staff's evaluation, which focused on the differences between the GE and SIA analyses, also assumed a speed of 1800 rpm. A worst case assumption would have been to assume a speed consistent with the emergency overspeed set point, which is between 110 and 111 percent of rated speed. This condition was evaluated for the licensee by SIA in an analysis dated June 4,1993. This evaluation indicated that the stresses would increase by about 7 percent at 110 percent of rated speed.
For normal operating temperature (i.e., 77.8 *C [172 'F]), this equates to a factor of safety for stress between 1.57 and 1.77 at the end of the operating cycle, depending on the assumed crack growth rate. On the basis 1
ofoperatingdata,thestaffused35.6 C [96 *F] instead of the 18.3 C
[75 F] used in the staff's June 1993 (See Item 8, below) evaluation. The i
factor of safety for stress during an overspeed event occurring at this temperature would be between 1.20 and 1.37, depending on the assumed crack growth rate.
Finally, in its probabilistic analysis (see summary paragraph at the end of this response), GE considered tne possibility of abnormal events including turbine speeds up to 195 percent of rated speed.
l B) Disk Temperature The GE and SIA analyses assumed a startup temperature of 18.3 "C [75 'F]
and a normal operating temperature of 77.8 C [172 *F].
The NRC staff assumed the same values in its June 1993 evaluation.
However, on the basis of operating data the actual temperature of disk 7GA during the startup is g
higher than 18.3 C [75 "F] because the LPA turbine is prewar.ned by the gland seal steam to at least 35.6 C [96 *F] for about 10 hours1.157407e-4 days <br />0.00278 hours <br />1.653439e-5 weeks <br />3.805e-6 months <br /> before rolling the turbine. This difference in temperatures is significant because of its effect on the critical stress intensity factor, K The K value is 19 percent higher at 35.6 *C [96 *F] than it is at 18.3,G.C[75*F)e Recalculating the factors of safety presented in the June 1993 evaluation using 35.6 C [96 *F) as the startup temperature results in factors of safety between 1.28 and 1.48 on stress and between 1.44 and 1.84 on flaw l
size, depending on the assumed crack growth rate.
C) Initial Crack Size During the April 1993 turbine inspection, GE was unable to reliably determine the depth of the crack located under the web of the disk.
In its analysis GE assumed a flaw depth of 6.35 mm [0.25 in.].
This value corresponded to the average depth of flaws measured in turbine disks with similar indications for which ultrasonic testing could not determine the flaw depth. These measurements were made after the disks'were removed from service. The maximum flaw depth found in these disks was 17 mm [0.67 in.].
SIA performed an analysis using a flaw depth of 6.35 mm [0.25 in.]
)
extending the entire length of the keyway in the bore of the disk. SIA i
. also performed a more sophisticated analysis using a flaw depth corresponding to the maximum value of 17 mm [0.67 in.] with a length of 85 mm [3.35 in.].
The assumption of a finite length flaw is consistent with the cracking that occurs in turbine disks and is the most accurate of the analytic models used. The subject of analytic models is discussed in greater detail in Item E, below. The SIA analysis showed that with a crack 17 m [0.67 in.] deep by 85 mm [3.35 in.] long and a maximum growth rate of 1.52 mm [0.06 in.] per year, the critical crack depth was 25.9 mm [1.02 in.] and the time to failure for disk 7GA would be in excess of 5 years.
D) Crack Growth Rate GE used a mean plus two standard deviation value of 1.524 mm [0.06 in.] per year for the crack growth rate based on a generic compilation of data from GE, Westinghouse, and Central Electric Generating Board of Great Britain (CEGB). SIA used 1.524 mm [0.06 in.] and 0.508 mm [0.02 in.] per year in its analysis. The 0.508 mm (0.02 in.] value was the mean value from the above data. The NRC staff performed calculations using both growth rates.
Based on the generic data, the probability of achieving a growth rate of 1.52 mm (0.06 in.] per year would be about 2E-2.
In addition, the measured crack growth rate based on results of inspections of the Pilgrim LPA turbine is about 0.508 mm [0.02 in.] per year.
E) Analytic Model GE, SIA, and NRC performed calculations that yielded critical crack depths of 8.64 mm [0.34 in.], 13.72 mm (0.54 in.] and 11.43 mm (0.45 in.],
respectively. GE's calculation was cased on a model having a crack in a semi-infinite plate with uniform stress.
SIA modeled the problem as a hole with a crack in a finite plate with attenuated stresses.
The NRC staff's calculations were based on a thick-walled cylinder model with attenuated stresses.
GE's assumption of a constant stress field is conservative; however, the attenuated stress profile more accurately represents the actual stress distribution.
In addition, the NRC staff considers the thick-walled cylinder model a more appropriate representation of the shrunk-on disk than either the simple plate model or the hole-in-a-plate model. The NRC model resulted in greater factors of safety than the GE
)
model but less than the SIA model. All of the above analyses are conservative in that the analytic models assume the crack is infinitely long; that is, the crack exists over the entire length of the keyway in the i
disk. SIA also performed a more accurate analysis in which the analytic model was for a finite length crack. This finite length flaw model 4
represents more accurately the cracking that occurs in turbine disks than the infinite length flaw model.
In this evaluation the crack was assumed to be 17 mm [0.67 in.] deep and 85 m (3.35 in.] long. The results of the analysis showed that the time to failure for disk 7GA would be in excess of 5 years based on a critical crack size of 25.9 mm [1.02 in.] and a maximum growth rate of 1.52 mm [0.06 in.] per year.
i
. F) Critical Stress Intensity (K )ic The critical stress intensity (K ) is a material property that provides a 3c measure.of the resistance of the disk material to fracture as a function of Like many material properties, the K,in the GE and SIAfactor shows a range temperature.
of values at a given temperature. The values used analyses were lower bound values that correspond to values approximately two standard deviations below the mean. This is a conservative assumption.
The probability of having a critical stress intensity factor this low or lower is about 0.02.
While a comprehensive probabilistic assessment that treats each of the above parameters as a random variable provides the most effective method for assessing the overall conservatism of the analysis, such an assessment was not done since it was deemed not necessary to demonstrate that a safety problem did not exist. This is because, considering the probabilities of the assumed crack growth rate and critical stress intensity factor alone yielded a failure probability of less that 4 0E-4 per year (i.e., 0.02 X 0.02 = 0.0004), and considering the probability distributions of the other variables in the analysis, the staff concluded that the overall probability of failure would be less than 1.0E-4 per year.
Subsequent review of the GE analysis (see response to Request No. 6) indicated an overall probability of failure of 1.1E-5 per year as of April 1995.
Reauest No. 4:
If the analysis described in the August 4,1993, response did not include consideration of the stresses on the cracked disks that could be caused by the type of vibrations discussed in Information Notice (IN) 94-01, perform such an analysis and report the results.
Explain the basis for any assumptions used in the analysis.
Response
The analysis described in the August 4,1993, response did not include stresses on the cracked disks that could be caused by the type of vibrations discussed in IN 94-01. As discussed in the response to Request No. 1 (above),
the licensee performed a turbine missile analysis for Pilgrim and showed that safety-related equipment would be protected from the worst-case turbine missile hazard. Therefore, given the licensing basis of the plant with respect to turbine failure, the requested analysis is not necessary to assure public health and safety.
Reauest No. 5:
If.the analysis described in the August 4,1993, response did not consider the stresses on the cracked disks that could result from a blade failure, perform such an analysis and report the results.
Explain the bases for any assumptions used in the analysis.
j
o
. Response:
The GE and SIA analyses and the NRC assessment did not evaluate the stresses on the cracked disks that could result from a blade failure. However, as discussed in the response to Request No. 4, the requested analysis is not necessary to assure public health and safety.
Reauest No. 6:
The NRC stated that it " intends to perform a confirmatory review of the GE analysis and its methodology." Provide the results from that review or explain why it was not performed.
Response
Within 1 day of completing its review of the SIA analysis, the staff confirmed that the GE analysis (Ref. 3) had been performed consistent with a methodology (Ref. 2) that was approved by the NRC in the mid-1980s. The staff approved the GE methodology as documented in NUREG-1048, " Safety Evaluation Report Related to the Operation of Hope Creek Generating Station," Supplement No. 6, July 1986 (Ref. 2).
In further assessing the cracks in the Pilgrim turbine, the staff has reviewed the actual analysis performed by GE. GE calculated a missile generation probability of 5.lE-6 per year as of April 1993 and a missile generation probability of 1.1E-5 per year by April' 1995 for the LPA turbine.
It should be noted that consistent with the NRC-approved methodology for calculating turbine disk reliability and established NRC guidance for acceptable reliability levels, the GE probabilistic analysis supports operation of the Pilgrim turbine until the end of the current operating cycle.
References
- 1. Pilgrim Nuclear Power Station, Unit 1, Updated Final Safety Analysis Report.
- 2. NUREG-1048, " Safety Evaluation Report Related to the Operation of Hope Creek Generating Station," Supplement No. 6, U.S. Nuclear Regulatory Commission July 1986.
- 3. " Probability of hissile Generation in General Electric Nuclear Turbines,"
General Electric Company Large Steam Turbine-Generator Department, January 1984 (proprietary).
April 18, 1994 1
The Honorable Gerry E. Studds term is unfortunate and to some degree is responsible for the confusion that currently exists.
I hope that this explanation anu the additional information provided in the enclosure help to resolve the concerns that you and your constituents have expressed.
Sincerely, Original signed by James ht Taylor James M. Taylor Executive Director Peradons cc:
Sen. John F. Kerry
Enclosure:
Response to Requests DISTRIBUTION:
Docket File (50-293) (w/ incoming)
W. Butler NRC & Local PDRs (w/ incoming)
OGC EDO #0009775 OPA J. Taylor OCA J. Milhoan SECY # CRC-94-0121 H. Thompson NRR Mail Room (EDO 9775 & 9858 w/ incoming)
J. B1aha N. 01 son W. Russell /F. Miraglia C. Norsworthy L. Reyes, Acting C. Nagel PDl-3 Reading R. Eaton w/ incoming S. Varga S. Little t
J. Calvo J. Linville, RI J. Strosnider
- Previously concurred
/
_ffEf[
OFFICE PQIN:EM PDI-30(k
- EMCB
- DE:D PD NAME REkt[n:dt SLiftle JStrosnider BSheron WBbtker j h194 Y //3/94 03/25/94 0_3/28/94 N/h/94 DATE 0FFICE
- TECH ED
- ADRI
- DRPE:D
- ADP (A)
NRR:D a y EDO(
NAME Mejac JCalvo l SVarga LReyes WRuNel5 JTpflop DATE 03/21/94 03/28/94 03/28/94 03/29/94 t/0/94
//j /94 0FFICIAL RECORD COPY pg g c,
bf__,,
yl:~
zA,i hK DOCUMENT NAME: A:\\PIGT9775 y
hs og o
c; ND
. ?r c.. q gig f;
C?T" M h
hy{g h&(h F#q {:l.
- A'!!y EDO Principal Correspondence Control FROM:
DUE: 03 94 EDO CONTROL: 0009775 DOC DT: 02/09/94-FINAL REPLY:
'R3p. Gerry E. Studds TO:
Dennis Rathbun FOR SIGNATURE OF :
- GRN CRC NO: 94-0121 Executive Director DESC:
ROUTING:
ENCLSOES LETTER FROM ROBERT POLLARD, UNION OF Taylor CONCERNED SCIENTISTS RE PILGRIM NUCLEAR POWER Milhoan PLANT Thompson.
Blaha TTMartin, RI Lieberman, OE DATE: 02/15/94 Cyr, OGC
-ASSIGNED TO:
CONTACT:
NRR Murley SPECIAL INSTRUCTIONS OR REMARKS:
February 16, 1994 NRR ACTION:
DRPE:VARGA' NRR ROUTING:
TEM /FJM i
' '? ; jnN 6(f 9
ONE l O
!UI 5 ?I'-[
f AI NRR MAIL ROOM l
bY ~7/JdM:-. M M _ I_Y'/___ _, {
(
/-
'i 7
j _ _ _. _ _. _. _ _ _,.
_ _. :a o
i
+
47
- 2 s
,.n E
n,,,.,.
.u
- l. '
4 OFFICE OF THE-SECRETARY CORRESPONDENCE CONTROL TICKET
<,i;;vp
[j.
E' PAPER NUMBER:
CRC-94-0121 LOGGING DATE: Feb 15 94:
IACTION OFFICE:
. EDO AUTHOR:
GERRY STUDDS AFFILIATION:
U.S.. HOUSE.0F REPRESENTATIVES ADDRESSEE:
RATHBUN-i
-LETTER DATE:
Feb 9 94 FILE CODES IDR-5 PILGRIM:
p
SUBJECT:
OPERATION OF THE PILGRIM NUCLEAR POWER PLANT
. ACTION:
Direct Reply DISTRIBUTION:
OCA TO ACK, DSB SPECIAL-HANDLING: NONE CONSTITUENT:
MARY OTT NOTES:
DATE DUE:
. Mar 2 94 4
SIGNATURE:
-DATE-SIGNED:-
l
- AFFILIATION:
' l ci
.i ii
- )
ij l
,. o c
(-
Lt:
s EDO --- 009775 Y
____.________.__a.2.n_J.__._.-_--.--_m. - _ _ _ _ _ _ _ _. - - _ _ _ _ _ _ _ _ _ - _ _ _ _ _. _. _
m___
-