ML20235B312
| ML20235B312 | |
| Person / Time | |
|---|---|
| Site: | Monticello, Brunswick, 05000000 |
| Issue date: | 08/31/1987 |
| From: | Poloski J EG&G IDAHO, INC. |
| To: | |
| Shared Package | |
| ML20235A701 | List: |
| References | |
| NUDOCS 8709240060 | |
| Download: ML20235B312 (8) | |
Text
_ _ _ - _ - _ _ _ _ _ _ _ _ _.. _. _ _.
ATTACHMENT 2 t *'
- COMPARISON OF MONTICELLO AND BRUNSWICK RECIRCULATION PUMP TRIP J. P. Poloski l
August 1987 8709240060 870918 PDR ADOCK 05000324 P
1 On.
INTRODUCTION The data presented by CP&L to demonstrate the reliability of their RPT design compared to the Monticello design were examined.
Several items appear to have the potential to impact the results of their analysis. The items identified are:
l (1) The limited failure data chosen for estimating the failure probability of the field circuit breaker, model AKF-2-25.
l (2) The majority of AKF-2-25 breaker failures can be attributed to improper lubrication and/or misadjustment of the breaker internal parts.
DATA EVALUATION I
The reliability of any system progresses from the design stage to the operating phase.
In the design phase, the designers establish goals or an inherent reliability of the system.
During the operating phase, the system attains an achieved level of reliability. Typically the inherent reliability of a system is higher than the achieved reliability.
Factors such as maintainability, testability, and environment degrade the inherent reliability that is designed into the system. The inherent reliability of the Monticello and Brunswick designs are essentially the same except for the breaker and trip coil arrangement.
These two components are the dominant contributors to the overall reliability of RPT given the same ATWS trip logic. The Monticello design, assuming equal mean-time-to-failure (MTTF) for the field and 4kV circuit breakers, and same MTTF for the shunt trip coils, is the more inherent reliable design. This is attributed to the redundant shunt trip coils of the Monticello compared to the single shunt trip coil of the Brunswick design.
CP&L contends the breaker's MTTFs are different with the more reliable breaker being the 4kV.
The failure data supplied by CP&L is restricted to the Brunswick plants, Units 1 and 2, and the Pilgrim plant and only appears to concentrate on the field circuit breaker. 'To ensure that the data are representative of the overall population, a search of the Oak Ridge LER database was performed.
The search was restricted to the Recirculation 1
1
E. '.
System and culled out those records containing motor-generator and/or circuit breaker events. The result of the search yielded 111-records.
The records were reviewed for 4kV and field circuit breaker failures.
Listed below are1the results of the search and audit of the LERS and additional data provided by CP&L:
PLANT DATE ABSTRACT Oyster Creek-2-7-87 4160 breaker failed to trip Monticello 10-24-83 Trip coil for MG drive breaker failed Monticello 2-6-874 Drive Motor Breaker failed to open Field breaker failed to trip
' Vermont Yankee 10-17-81 MG field breaker failed to open Pilgrim 4-3-83 Field circuit breaker did not trip Pilgrim 6-29-86 Field circuit breaker failed to trio.
2-9-85 Field circuit breaker failed to trip 3-15-85 Main generator field breaker failed to trip Browns Ferry 3 5-7-81 Field creaker trip coil open Brunswick (a) 6-23-81 Field breaker failure 4-19-82 Field breaker failure 7-27-84 Field breaker failure Dresden 3 9-2-83 Shunt trip burned out I
3-1-82 Operator arm disconnected Note:
(a) Only identified by CP&L; no documentation to identify / evaluate the breaker failures.
Breaker failure probabilities were estimated based on the following assumptions:
(1) Two breakers per plant for the field circuit (2) Two breakers per plant for tne drive motors.
l 2
?....
1 (3)
Failures are reportable since 1980 thereby giving a 7) year period.
(4)
Since breakers are not required to be tested during power operations, the only demands are those experienced by unexpected transients.
An assumption of 6 breaker demands per breaker per year was utilized to estimate overall population of demands.
(5) Thirty-four plants comprise the population data set.
The demand population was estimated to be 3060 demands (34 plants x 7}
years x 2 breakers / plant x 6 demands / breaker / year). The following gross estimates for the 4160 and 480 V breakers were calculated:
I 1
480 V 3.6 E-3/ demand (11 failures /3060 demands) 4160 V 9.8 E-4/ demand
( 3 failures /3060 demands)
As mentioned earlier, the estimates calculated for the two breakers are to be considered gross indicators of the true failure parameter.
Subjective judgements pertaining to the population sizes, lack of detailed failure experience and all failures are not recorded in the LERs are some of the major factors contributing to the overall uncertainty of the estimates.
LER reporting requirements vary from plant to plant thereby making I
plant-to-plant comparisons difficult.
RPT availability estimates were calculated under various conditions to get a representative sample of the 480 volt and 4160 volt breaker designs of the Monticello and Brunswick plants. Tables 1 and 2 report the cutsets and associated, estimates. The range of RPT unavailability estimates account for different sources of failure data for the breakers and times for detection of open coil of the shunt trip devices.
The following range of estimates were calculated and extracted from Tables 1 and 2.
480 V.........(.000058,.0082) 4160 V.........(.00022,
.012) 3
Ib a.. ; o :
a
(
-TABLE 1.
BRUNSWICK CUTSETS BRUNSWICK &
BASE NUREG-4126 PILGRIM DATA LER MILLSTONE BKR-A 1.25'E-3 1.1 E-4 9.6 E-5 9.8 E-4 6.2 E-3 BKR-B 1.25 E-3 1.1 E-4 9.6 E-5~
9.8 E-4.6.2 E-3 TC-A-FTO
.1 E-4 1.4-E-5 1 E TC-A-OPEN 1.6 E-6 1.6 E-6 1.2 E-3 TC-B-FT0 1 E-4 1.4 E-5 1 E-4 TC-B-0 PEN 1.6.E 1.6 E-6 1.2 E-3 LEVEL-A
- LEVEL-B 2.6 E-10 2.6 E-10 LEVEL-A
- LEVEL-D 2.6 E-10
.2.6 E LEVEL-C
- LEVEL-B 2.6~E-10 2.6 E 10 LEVEL-C
- LEVEL-D 2.6 E-10 2.6 E-10 TOTAL.
2.5 E-3 4.22 E-4 2.2 E-4 4.6 E-3 1.2 E-2 L
4
/
TABLE 2.
MONTICELLO CUTSETS PILGRIM &
BASE NUREG 4126 BRUNSWICK-
'LER INPO. MILLSTONE.
BKR-A~
1.25 E-3 2.9E-5 4.1 E 3.6E 1.2 E-3 2.95 E-3 BKR-B 1.25 E-3 2.9 E-5 4.1 E-3 3.6 E-3 1.2 E-3 2.95 E-3
'TC-Al-FTO
- TC-A2-FTO 1 E-8 3.7 E-7 3.7 E-7 TC-Al-FTO
- TC-A2-OPEN 1.6 E-10 9.8 E-10 7.3 E-7 TC-B1-FTO
- TC-B2-FT0 1 E-8 3.7 E-7 TC-B1-0 PEN
- TC-B2-FT0 1.6 E-10 9.8 E-10 TC-B1-FT0
- TC-B2-0 PEN 1.6 E-10 9.8 E-10 TC-Al-OPEN
- TC-A2-FTO 1.6 E-10 9.8 E-10
.TC-Al-OPEN'* TC-A2-0 PEN 1.6 E-10 9.8 E-10 LEVEL-A
- LEVEL-B 2.6 E-10 LEVEL-A
- LEVEL-D 2.6 E-10 LEVEL-C
- LEVEL-B 2.6 E-10 LEVEL-C
- LEVEL-D 2.6 E-10 TOTAL 2.5 E-3 5.8 E-5 8.2 E-3 7.2 E-3 2.4 E-3 5.9 E-3 L
E u
5
c;..
The range of estimates indicate that the RPT unavailability associated
.with the 480 V generator field breaker design to be lower than those associated with the 4160 V motor drive breaker design.
The lower-valued estimates for both the 480 V and 4160 V breakers represent the data extracted from NUREG/CR-4126.
As pointed by CP&L, the low voltage data is not representative of the metal clad breaker, AKF-2-25.
The lower estimate can be interpreted as a potential estimate of RPT unavailability if the metal clad breaker experienced the same performance as the low voltage molded case breakers.
A review of the failure data pertaining to the AKF-2-25 failures identifies binding and/or improper lubrication as the mechanism for failure of the breaker to open on demand.
Since these failures are risk significance with respect to ATWS, an improvement in the preventative maintenance of these breakers is essential.
General Electric Service Information Letter, SIL 448, is intended to correct these deficiencies in the AKF-2-25 breakers.
New lubricants and more frequent preventative maintenance by trained personnel have been identified by GE to remedy the failures associated with binding and/or lubrication of the mechanical trip device of AKF-2-25 circuit breaker. The effectiveness of these recommendations will take time to demonstrate. The new lubricants and preventative maintenance policy should potentially eliminate the binding of AKF-2-25 breaker thereby improving the reliability of the breaker to that of the 4160 design.
SUMMARY
The recirculation pump trip of the Monticello design, from a design standpoint, has the potential to be more reliable from a safety aspect than the Brunswick.
The primary reason for this being the lesser number of single failures in the Monticello design. The dominant single failure in both designs are those associated with the 480 volt field circuit breaker of the Monticello and the 4160 volt drive-motor breaker of the Brunswick.
Due to the poor quality of the failure information reported in
]
the LERs and by CP&L, a plant-to-plant comparison was difficult to analyze.
Instead, a range of estimates for RPT unavailability were calculated to show the potential spread or uncertainty associated with the l
6 j
o...
i l
data. These estimates are not to be intended as exact or true estimates of RPT unavailability.
The ranges are to be interpreted as gross indicators.
As indicated by the ranges, the Montic211o design appears to be potentially better with respect to unavailability of the PRT function than the Brunswick design.
Incorporation of the improved breaker lubricants and preventative maintenance into the AKF-2-25 breaker operations as prescribed by GE should potentially eliminate binding of the breaker.
However a program should be established to ascertain the effectiveness of these recommendations.
Other alternatives, such as molded-case or in-line series breakers, should be examined in the event the remedies identified by GE do not improve the AKF-2-25 breaker performance.
4 l
j t
7
. _ _ _ _ _.._______-___- ______