ML20010F639
| ML20010F639 | |
| Person / Time | |
|---|---|
| Issue date: | 08/18/1981 |
| From: | Novak T Office of Nuclear Reactor Regulation |
| To: | Harold Denton Office of Nuclear Reactor Regulation |
| Shared Package | |
| ML20010F640 | List: |
| References | |
| RTR-NUREG-0642, RTR-NUREG-642 NUDOCS 8109110050 | |
| Download: ML20010F639 (5) | |
Text
W y
j a
s g.
.D h:
I
'.-N kv
}
/1p-AUG 181981
\\ M lip MEMORANDUM FOR: Harold R. Denton, Diri+ctor Q
f f
VA Office of Nuclear Reactor Regulition h
LJ
[, ( <? l /g g y,,
8 l
THRU:
Darrell G. Eisenhut. Director 9
Division of Licensing FROM:
Thomas M. Novak, Assistant Directcr 4
for Operating Reactors, DL p
4 (4
SUBJECT:
THOUGHTS ON NUCLEAR PLANT CAPACITY FACTORS AND REA
.S SAFETY My May 19, 1981 memorandum on " Status of Forty Shutdown Fa':lities on May 13,1981" (Enclosure 1) was not intended to address 'ae more impor-tant question of how to assess operating plant safety performance. As you are well aware, NRC is expending considerable manpower on this question through the SALP program. Consequently, in order to avoid duplication of the SALP effort, these comments have been limited to the following closely related questions.
o What variables influence a nuclear power plant's capacity factor?
o Is there a relationship between capacity factor and safety performance?
First, however, it's necessary to address the improvement in the pl. ant operating status compared with the May data. As this memo is prepared, 56 of the 75 facilities counted on the August 10,19R1 Plant Status report are operating. Of the 19 shutdown unit: reported, 1 is awaiting a license and two of this number are the Three Mile Island units and D..sden 1 is it, chemical cleanup. Thu, the possible operating efficiency is 56 of 71 or 79%.
Looking.at the Enclosure 2 chart en Ct;mulative Unit Capacity Factor for the 64 BWRr and PWRs that have acctsulated significant operating experience, you can see the average capacity factor (cNely related to the operating effielency) is 62%. Therefore, at any 9ven time if 44 of the 71 plants licensed to oporate are "on the lin#, the nuclear power generation is at its historic.1 average norm.al ute.
Variablas thet Influence Nuclear Plant Capacity Factors o NSSS Type - Enclosure 2 shows that, based on a simple average, the CE units have the best performance followed by W, GE and
"'W.
Since the average for s11 is 62%, the variance 0109110050 810818 PDR MISC
+5% and -7% around this average.
PDR or,,u p Cont act:
E.L. Cor ner
. E7';54-
.m>
.l....
he,osumnowi%cume OFFICIAL RECORD COPY
- " *-3m a
p-----
9 enn y*,
wwg n;tq wm gjmmr
'X j,
g g'.
g g
g g
3 0 Maturation and Size - Enclosure 3 is a relevant, recent article from Technology Review. The authors found that the PWR apacity factors increased with maturation and decrease with si:'e (400 to 800 MWe vs. larger than 1000 MWe) and at higher I
coolant temperatures. They did not find discernible l
comparable effects on BWRs. They also reported that three of four unplanned outages are due to equipment failures and almost half of these failures were rooted in the main heat-removal systems (steam-generator /feedwater-condensate).
The article further cautions against presture standardization.
Similar results are presented in " Statistical Analysis of Power Plant Capacity Factors Through 1979" (NUREG/CR-1881, ASND81-0018), April 1981. The primary findings of this report for nuclear power plants are, (1) that both BilR and PWR cape city factors tend to increase with a unit's age, the increase being more pronour:ed for BWRs, (2) that there is heterogeneity among plants in that some tend to perform consistently above average, others consistently below; and (3) that there is no evidense of a size effect among BWRs while there is for PWRs.
In the latter case, a group of smaller PWRs (MGN from 450 to 600 MW) have averaged 10 to 15 percentage points higher than the remaininq larger PWRs (MGN fron 760 to 1216 MW).' Within these two subgroups, though, there does not appear to be an association of capacity factors with unit size.
It should be noted, however, that there are only one or two unit-years of data on several of the largest PWRs so tnat additional experience could easily change this conclusion.
This report's predicted capacity factor, at a 95% confidense level, is given in the following table.
Years 2-5 6-10 2-10_
BWRs 16%
65 i 15%
682 1 13%
73 14%
71 7 13%
58 22%
57 i 20%
The 95% statistical predicted capacity factor for a coal plant, for a 10 year period, is 56 i 17%. The authors consider the coal plant capacityyquite similar to the BWR and large PWR pre. fictions and conclude "there is no static-tical basis for claiming c camcity facter advantage for or.e type of plant (coal or nuclear) cver the other".
""^"4 DATEh oc,osu 3ie oo,soi.;scu e24o OFFICIAL RECORD COPY
"-3 2 *
- sg%
.y$
Q 3
Y
)
p
(
~3
, )
o Manageme'nt - Looking at Enclosure 2 again, indications of effects cf management my be seen.
For example, dual units like Dresden, Quad Cities D. C. Cook, Indian Point, Surry, Turkey Point, Zion and the Oconee units have capacity factors (CFs) within a 3% band of the s! ster tinits. Also, note the BWR/PWR combinations like:
(1) Commonwealth Edisons' Dresden 2 and 3 Quad Cities 1 and 2 and Zion 1 and 2 with Fs between E7 and 60; (2) Northern State.
Power's Monticello and Prairie Island 1 and 2 with CFs between 71 and 81%; and (3) Nor'theast Utilities with Haddam Neck at 82%, Millstone 1 and 2 at 63 and 625, respectively.
Could it be that NSP's conservative type of operation and NU's strong centralized engineering organization make a difference? On the negetive side are Duquesne Light's Beaver Valley with a CF of only 27% followed closely by Davis Besse, operated by Toledo Edison at 34% CF. The OR PMs for thse units attribute the problems to poor manage-ment and inadequate construction cuality control with Duquesne short on the management and Tcledo's problem being mainly QA.
Corrective measures have been taken by both utilities.
o Regulation - In order to get a feel for licensee!' evaluation of the safety impact of our (NRC) requirements, a single utility operating three units and constructing another unit was contacted. Northeast Utilities was asked to evaluate the s6fety significance of the NUREG-0737 requirements. They responded that in theira ssessment of the 63 items that apply to at least one of their facilities, 32 were judged to improve safety, 26 to have no effect on reactor safety, two that could possibly degrade safety and three itt 7s that the effect on scfety could not be determined, as yet. NU believes that of the 32 items that may improve safety, about 24 are cost Effective; that is for the money spent these improvements t.re worthwhile.
This, of course, depends a lot on implementation pressure.
If we insist on schedules that licensees cannot physically make, due to equipment design and acquirements and that regnire special shutdowns for installatic es, then itcms that could be cost effective become non-cost effective. is an interesting letter from NU to IE's V.
Stello summarizing r.rme af their concerns regarding our regulations. Fu alto expressed alarm in regards to the effect our regulations have had and could ha.ve on their capacity factor. They believe that th< CF for each Millstone unit has been reduced by about 7.5% per year for the last several years. They are apprehensive in regards to lost production from ot'r Appendix R Safe Shutdown, Environmental Qualifications and TMI items where requirements and imple-r-or,tatinn cr hedulcS_2 r_e_ar}itrat{ly_ent Aro stuch cnneerna:
l justified?
l omco
.I
. j..
psems7eno seu o2 m OFFICIAL RECORD COPY
' " " - * ~ "
up y,
i-
.a.
Relationship Between Capacity Factor and Safety Perfonnance Most of us c' the NRC feel that there is a direct relationship between how well a plant is run and nuclear safety. However, it does not necessarily follow that units wi-the highest capacity factor in Enclosure 3 are necessarily the safest nuclear power Plants. A high CF could be the result of conservative operations, good maintenance practices, well trained operators and luck or it could be a less than professional operation where production-at-any-cost is the aim.
For example, some facilities operated with main condenser tube leaks and/or poor secondary water chemistry control in the early days of operating their units.
Now the Piper must be paid in steam generator tube denting and support plate cracking. More condervative utilities have placed high priority on the protection of the stear: generators (and other major equipment) even when it cost CF in the beginning.
The incident record data shows that the highest nuclear safety is achieved when a unit is operating at a steady state power level with all equipment well maintained and in service. Most problems occur during startups, shutdowns, equipment failure caused and other transients and during outages.
Accordingly the wisdom of our requiring reactor shutdowns to correct potential safety concerns or to make NRC required modifications on arbitrary schedules may be questioned.
If we work toward implementations of NUREG-0642 (A Review of NRC Regulatory Processes and Functions) Technical Reconsnendation No. 9, a gradual improvement in the cuclear units' capacity factor would be seen and these units would have improved safety.
Recommendation No. 9 is: "ACRS believes that the fundamental safety goal of both NRC and the nuclear industry should be to achieve a degree of safety that is as good as reasonably achievable, taking into consideration appropriate technical, social and economic factors."
Original signed by:
Thomas M. Novak, Assistant Director for Operating Reactors Division of Licensing
Enclosures:
As stated t
See next page for concurrence and distribution
/
omcs >
ORB #3:DL ORB #3:DL ORB #3:DL A-
-L UK7euiz'e7 '$'Coine de'pY ~ (5Efais"*"
'TM sk" DUk' ". ut.. '.' "..' ' '. '.'. '. '. "..
su m m )
em) 8/12/81 8/13/R1 8/13/81 8/ I l'.81 8/.[h.. 81
. l.1
....3..
gc ronu aie ne soi ncu eue OFFICI AL R ECO'R D COPY '
-> =-m aa
t
. m, r%:
'N N ""~e f:
~
j 3
, Relationship Between Capacity Factor and Safety Performance Most of us at the NRC feel that thare is a direct relationship between how well a plant is run and n'iclear safety. However, it doesnnot necessarily follow tht units with the highest capacity factor in Enclosure 3 are necessarily the safett nuclear power Plants. A high CF could be the result of conservative operations, good maintenance practices, well trained operators and luck or it could be a less than professional operation where productior-at-any-cost is the aim.
For example, some facilitie; operated with main condenser tube leaks and/or poor secondary water chemistry control in the early days of operating their units.
Now the Piper must be paid in steam generator tube denting and support plate cracking. More conservative utili utilities have placed high priority on the protection of the steam generators (and other majer equipment) even when it cost CF in the beginning.
The incident record data shows that the highest nuclear safety is achieved when a unit is operating at a steady state power level with all equipment w?ll maintained and in service. Most problems occur during startups, shutdowns, equipment failure caused and other trancients and during outages.
Accordingly the wisdom of our requiring reactor shutdown to correct potential l
safety concerns or to make NRC required modifications on arbitrary schedules may be questioned.
If ne work toward implementations of NUREG-0642 (A Review of NRC Regulatory Processes and Functions) Technical Recommendation No. 9, a gradual improvement in the nuclearunits' capacity factor would be seen add these units would have improved safety. Recommendation No. 9 is:
"ACRS believes that the fundamental safety goal of both NRC and the nuclear industry should be to achieve a degree of safety that is as good as reasonably achievable, taking into consideration appropriate technical, social and economic factors."
As a bottom line to this effort, the staff should centinue to work at gaining a better understanding of the impact of perceived improvements on operating reactors but accept the possioility that certain changes may in fact detract from a continuing improvement in safety.
DISTRIBUTION:
Thomas M. Novak, Assistant Director for Central File Operating Reactors NRC PDR P.M. Kreutzer Division of Licensing L PDR E.G. Case NSIC Memo File TERA ORB #3-Rdg l
D. Eisenhut J. Heltemes R. A. Clark M. Conner
./
ORB #3.Dl' d
ORB #3:DL AD:0R:DL
'5""" " > N OR#3g,k[. N h.
cmcr y t
..!.h.k......
S
.m) 8/g/81 8//3 /81
/( 3/81 1
8/ /81
.. (..
- FJRC FORH 318 HOs tf 0} NRCM O240 OFFICIAL RECORD COPY
' " "-m m l
Enclosure i f',,
4D UNITED STATES
&85 NUCLEAR REGULATORY COMMISSION ier f.
~JE WASWNGTON, 3. C. 20555 l -
h
[
MAY l9I381
/
MEMORANDUM FOR:
Harold R. Denton, Director
}
Office of Nuclear Reactor egulation W
SITT-~
Darrell G. Eisenhut
'ector r
Division of Lic FROM:
Thomas M. Nov, Assistant Director fer Operating Reactors. GL
SUBJECT:
STATUS OF FORTY SHUIDOWN FACILITIES ON MAY 13, 1981 In response to your inouiry of May 13, 1981, "Over 40 facilities down:
What's. happening?", the following summary information is provided.
Grouc 1 - Sorino Refuelino Outaces As you are well aware, many licensees deliberately plan refueling cutages l
in the spring (and the fall, also) as this is the period of lowest power i
consumption. The facilities in refueling outages with no other major problems are:
Arkansas 2 Bre.sns Ferry 1 Cocper D. C. Cock 2 (now starting up)
Duane Arnold 4
G nna Kewanee
aine fankee Nine Mile Point i North Anna 2 Point Beach 2 Troj an Grouc 2 - 3hutdown for Major Problems i
The following facilities have experienced major.rechanical failures or other problems as noted.
Three Mile Island 2 is included in this group for completeness.
4 Srowns Ferry 2 trip recovery (has returned to pte.er)
Bruns 'ick 1 - miscellaneous repairs Brunswick - RHR heat exchanger repair C;vis Sesse 1 - (has returned to. power)
Farley 2 - stattup physics testing s
- /
p /!/ m'y fv 'Lt f4h b
f}/,,'
o p u o'\\ f ' _
o y '-
a nf
s s
1 Indian Point 2 - major repairs (in the process of startup)
Millstone 1 - damaged turbine Peach Bottom 2 - recirc. ouma motor bearings-Rancho Seco-repairs (has returned to power)
San Onofre - steam generator repairs Sequoyah 1 - steam generator repairs Surry 1 - steam generator replacement Three Mile Island 2 - Accident Turkey Point 3.- Failed main generator Grouc 3 - Shutdown in A;.cordanc& with or to meet NRC Recuirements This group of facilities is. separated 'rtm Group 1 only because.be major repairs stem from some NRC requirements to ensure or improve raactor safety.
Dresden 1 - Ordered shutdown /in hearing Hatch 1 - Mark 1 containment modifications Millstone 2 - Inoperable mechanical snubbers / Bulletin 81-01 Oyster Creek - SWR scram system and environmental qualifications /
Bulletin 79-01 Peach Bottom 3 - Mark 1 containe;nt modifications Yunkee Rowe - Modifications for AFWS upgrade, TMI rcquirements and SEP Grouc 4 - Shutdcwn Avaiting ::RC Licensing Action This grouping is necessary since these f acilities were counted in the 40 shutdown units.
Diablo Canyon 1 - 12 months-delay LaSalle - no delay McGuire - 6 mcnths delay Salem 2 - ready sir.ce May 1979 Susquehanna - 12 months delay Summer - 5 months delay Three Mile Island 1 - hearing Zimmer - no delay These four groups categorize the reascas for the 40 shutdown fac;1ities as shown on the enclosed IE Plant Status Report of May 13, 1981.
In addition, the folicwing facilities are limited or are operating significantly below their licensed power level.
Big Rock Point - 88% due to thermal limits on fuel Davis Besse 1 - three RCP operation limits power to 175%
Fort Calhoun. 45% conserving fuel (low area power consumption) i e
e
_3_
-Fort St. Vrain - 70% license limit Indian Point 3 - e7% due to failed main transformer Lacross - 94% to extend core life l
Pilgrim 1 - 90% limit due to concenser delta temoerature Point Beach 1 - 80% by licensee ~o reduce steam generator tubes degradat;cn hW w'
s homas M. Novak, Assit tr..it Director for Operating Recctors Division of Licensing
Enclosure:
As stated t,yco An-l M;E h f k g;, w n y.
en
% ~ l> rad w M
& c wn ou dp /-
i m e.u:,
w
~w
/
/
l i
e r
l j
~
1 i
'\\
l I
e 9
~
_.,s
cN, t
- 4 1
s t_,_,
L i
/.,*>.,
'JMTED STATES j
NUCLEAR RdGULATORY COMMISSION I
1
'. \\V,,s!
WASHINGTON. D. C. 20555 l
=.,
l OFFICE OF THE -
1 DIRECTOR OF NUCLEAR REACTOR REGULATION Tom Novak -
Over 40 facilities down! What's happening?
l a
Harold Denton May 13, Ic81 a
e.
8*,
,p _
=..,.. g..;'-sm; =--
gy m
g:
... ~..-...
~ '...
- ~..'.
.e
~~
5 e
/
O iJ l IL 4f vu ff,
]O p7 I
loo F J
i t
{
+ -
~
e kNO
__m_____'___________
REASONORCOMMENT(fc I
PLANT NAME PLANT STATUS 7
==% POWER __DOWN__-
____C
$\\,
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ = = _ - -
BEAVER VALLEY 1 l 100 l
l i 9 lbMMa CALVERT CLIFFS 1 100 l
CALVERT CLIFFS -
100
-,' '[
d FITZPATRICK
- 100 l
l GINNA l
l X
- REFUELING
(
[ h. k bOi A \\ t RESTART DATE-6/8 l
HADDAM NECK 100
- INDIAN POINT
- X
- HEATUP PLANT FOR PRE-PHYSICS TESTING
{})
INDIAN POINT 3 67 1 TRANSFORMER AVAILABLE MAINE YANKEE X
REFUELING
^
RESTART DATE-7/3 MILLSTONE 1 X
- TURBINE REPAIR RESTART DATE-6/7 MILLSTONE 2 X
l REPAIRS IN CONTAINMENT i
l l
REGTART DATE-O DATE NINE MILE POINT 1 l
X l REFUELING l
l l
RESTART DATE-6/8 f
OYSTER CREEK l
l X
MAINTENANCE RESTART DATE-5/26 PEACH BOTTOM 2 l
l X
REPAIR RECIRC. PUMP SEAL UNIT RESTART DATE-5/20 PEACH BOTTOM 3 l
l X
- REFUELING l
l RESTART DATE-S/21 l
PILGRIM 1 l
90 i
! CONDENSER DELTA T j
SALEM 1 99 l
SALEM 2 X
- AWAITING LICENSE SUSOUEHANNA X
- AWAITING LICENSE THREE MILE ISLAND 1 X
- NRC OPDER THREE MILE ISLAND 2 X
NRC ORDER
- VERMONT YANKEE 40 l
l COMING UP FROM TRIP YAK CE ROWE X
REFUELING l
l l -
RESTART DATE-7/2 REGION II PLANT STATUS-5/13/S1
___________ _____-______- ____-____________ _ _ =
=-_
__-- _% POWER __DOWN_______________________________
BROWNS FERRY 1 l
l X
l REFUELING 7
RESTART DATE-8/2 BROWNS FERRY 2 X
RECOVERING FROM REACTOR TRIP l
l l
RESTART DATE-O DATE BROWNS FERRY 3 89
- COMING BACK UP j
BRUNSWICK 1 l
l MAINTENANCE OUTAGE BRUNSWICV
- X l REPAIRS I
RESTART DATE-O DATE l
CRYSTAL RIVER 3 100
't FARLEY 1 99 l
l FARLEY 2 X
! HEATUP FOR PHYU l
RESTART DATE-5/15 X
! REFUELING 4
HATCH 1 RESTART DATE-6/2 HATCH 2 l
98 MCGUIRE l
X
! MAINTENANCE l
l l
RESTART DATE-6/6
I NORTH' ANNA 1 l 100
. NORTH ANNA 2 X
- SCHEDULED CiUTAGE RESTART DATE-5/?3 OCONEE 1
- 100 OCONEE 2 l 100 OCONEE 3 l 100 ROBINSON
- 96 SEQUOYAH 1 X
SG REPAIRS, RESTART DATE-5/15 ST. LUCIE 1 l 100 l
l SUMMER X
- AWAITING LICENSE SURRY 1 X
- SG REPAIRS l
l RESTART DATE-S/81 SURRY '
100 l
l TURKE/ POINT 3 X
COLD SH"UTDOWN FOR REPAIRS RESTART DATE-7/1 TURKEY POINT 4
- 100 l
REGION III PLANT STATUS-5/13/81
_ _ _ _ _ _ = - - - - - - - -
PLANT NAME PLAFT STATUS REASON OR COMMENT j-
________________________% PCMER__DOWN__=__---
-=
BIG ROCK POINT l
88
- THERMAL LIMITS ON FUEL COOK.1
- 100 i
- COOK 2 X
- MODE 4 LEAKING VALVE DAVIS-BESSE 1 l
l X
- MODE 2 HAVE ESTABLISHED CRITICALITY DRESDEN 1 l
l X
CHEMICAL CLEANING RESTART DATE-INDEF DRESDEN
- 50
- COMING UP TO FULL POWER DRESDEN 3 l
90 l
l COMING BACK UP DUCNE ARNOLD X
l REFUELING l
l l
RESTART DATE-5/18 REWAUMEE X
l REFUELING RESTART DATE-6/1 LACROSSE 94
, EXTENDING CORE LIFE LASALLE l
X
- AWAITING LICENSE
- MONTICELLO l
55 COM'_NO UP PALISADES 99 l
l POINT BEACH 1 80 REDUCE TEMPERATURE STRESS POINT BEACH 2 X
- REFUELING l
l RESTART DATE-5/24 PRAIRIE ISLAND 1 l 100 l
l PRAIRIE ISLAND 2 l 100 QUAD CITIES 1 98 l
}
QUAD CITIES 2 94 ZIMMER l
X W4AITING LICENSE ZION 1 90 l
l COMING UP TO FULL POWER ZION 2 90 l
l COMING UP TO FULL POWER e
/
a
-I' REGION IV PLANT STATUS-5/13/61'
_=_
_ _ _ = - -
_ __====
"LANT NAME PLANT STATUS REASON CR COMMENT
____% POWER __DOWN___
l REFUELING RESTART DATE-6/30 COOPER l
l X
- REFUELING RESTART DATE-5/20 FORT CALHOUN 45 FUEL CONSERVATION FORT ST. VRAIN 70 NRC LIMIT REGION V PLANT STATUS-5/13/S1
=
_-==
PLANT NAME PLANT STATUS REASON CR COMMENT F OWER__DOWN_ = = = =
DIABLO CANYON l
1 AWAITING LICENSE RANCHO SECO X
- SHUTDOWN FOR REPAIRS l
l l
RESTART'DATE-5/15 SAN ONOFRE X
COLD SHUTDOWN; SG REPAIRS l
l l
RESTART DATE-5/31 TROJAN X
l REFUELING RESTART DATE-6/29
- REASON OR COMMENT HAS CHANGED IN PAST 24 HOURS
]
I O - hEACTOR SCRAM W'ITHIN F AST 24 HOURS NOTE - REACTOR STATUS DATA COLLECTED BETWEEN i
6 A.
M.
AND S A.
M.
EACH DAY EXCEPT HOLIDAYS AND WEEKENDS SIGNIFICANT EVENT REPORT SIGNIFICANT E /ENT REPORTING PERIOD ~ FROM 0S00 5/12/S1 TO 0S00 5/13/S1 FACILITY DAVIS-BESSE REPORT TIME 1546 EVENT REACTOR TRIP RCP TRIPPED. INADVERTENTLY INSTALLED TRIPPEE 3
RELAY. NO ESFAS EVINT TIME 1457 CATEGOR) OF EVENT 10 CFR PART 50.72 REPORTABLE EVENT PEGION III ACTION TAKEN BY HQTRS DUTY OFFICER NOTIEFIED REGION III (MC GREGOR) og 4
4 0
^
Enclosure.2 d
CUMULATIVL UillT CAPACITY FACTOR GE BWRs Cl W PWRs CF CE PWRs CF Browns Ferry 1 51 Beaver Valley 27 Arkansas 2 68 2
50 Cook 1 68 Calvert Clifts 1 70 3
66 2
68 2
78 Brunswick 1 56 -
Farley 1 55 Ft. Calhoun 65 2
45 Ginna 69 Maine Yankee 68 Cooper 63 Iladdam lleck 82 Millstone 2 62 Dresden 2 57 Indian Point 2 56 Palisades 47 5
3 57 3.
56 St. Lucie 76
.Duane Arnold 55 Kewannec 76 3
~
.8 CE-PWR Ave.
67 Fitzpatrick 61 North Anna 1 66 llatch -1 61 2
91 4
2 54 Point Beach 1 73 4
Millstone 1 63 2
80 B&W PWRs CF Monticello 74 ?
Prairie Is. 1 71 Nine Mile Point 62 2
81 Arkansas 1 58 i
Oyster Creek 68 Robinson 2 70 Crystal River 54 1
P.jach Bottom 2 64 Salem 1 45 Davis Besse 34 3
68 San Onofre 68 Oconce 1 61 Pilgrim 1 56 Surry 1 53 2
61 Quad Cities 1 60 2
52 3
63 2
60 Troj an 52 Rancho Seco 56 Vermont Yankee 68 Turkey Pt. 3 71 Three Mile is. 1 53 j
4 68 22 BWR Ave 60
-Yankee R_ owe 70 0 B&W-PWR Ave.
55 I
Zion 1 57 2
59 64 BWR&PWR Ave.
62 26 W-PWR Ave.
65 52 PWR Ave.
63
t l
I W.aat Maxes Xuc_ ear Power P: ants 3reak Down?
l by Steve Thomas and John Surrey l
The reliability of today's nuclear reactors l
generally needs improvement, so standardiza. ion of current designs is premature.
I
] t.suc debates on nuclear power have centered on Previous independent analyses of nuclear plant per-l matters of safety and proliferation and the wider formance have general y focused on a particular ceun-issue of cent.alized decision making. But the reliabili-
_try (t!sually the United States) and year and failed to ty of nuclear plants ccmpared with expectations and examine the causes of reactor unavailability. With the effect of reliability on energy costs have largely these shortcomings in mind, we undertook a study of escaped public attention. The reliability issue will all commercial nuclear p! ants in the non-Communist inevitably come to greater prominence as more nucle-countries by 1978, including 17 countries with a total i
ar plants enter service and their performance and rated capacity of 91.2 j;igawatts. Apart from readily costs are reflected in the price of electricity.
identifiable prototypes, no plant is excluded even if its Nuclear plants have high capital costs and low performance has been " unrepresentative."
operating costs relative to fossil-fuel plants, so they Performance of nuclear power plants is usually are operated at their full authorized rating whenever expressed as a plant capacity factor, obtained by i
possible. Indeed, n'uclear-investment decisions involv-dividing actual output by maximum design output. To ing huge, long-term capital expenditures are made.. ensure that the capacity factor is a valid measure of with little knowledge of how current designs are life-recctor availability and performance, we examined l> to perform over the long term. When reactors are the record of all commercial plants operating from not available for service, or if they must operate well 1975 to 1977. We found that only two units had oper-below their full rating, the cost of replacement power ated, briefly, at reduced output even when available from other types of generating plants or utilities is for service at their full authorized rating. All other usually very high'. In view of the grov irrg number of output stiortfalls were due to plant-related factors reactors worldwide and the changing regulatcry cli-such,as refueling, planned repairs and maintenance, mate, plant performance must<be continually moni-and equipment failures.
tored to improve reliability and reduce investment Of the four types of commercia. nuclear reactors, risks.
by far the most numerous are the pressurized-water j
l May/ June 1981 Techneiogy Fevien 57
+
~
j
-e.
~
T V per%rr%.mteef 42 te! g-astst rcrctus had cp.rtting 80tir this } ett. The
.. riuclear pom r pt..,ts en 17 the ;~c'sst rectrg. Early track treerd cf prsssuri:ed.
' c;untrcs from 157s 131s79, dwgns 6f g!s-csclaf ructor3 witir r;actirs was d>itt a The c;pacdf seroris th) wers 3bancton;4 in 1?5s hrary birw by th3 cccid;nt ct I
Ovcilst'e caparJty timded by becaus2 of high c!nstruction Thrse El3 Istr.ad in Apfd design capacity. Pressurized certs and relative 1579. All nuclear power plants heavy water reactors had the inefficiency; an advanced sited in J.gan suffered larges; capacity factors, while gas cooled design may be prolonged outages.-
T J g.3,,e m eny
?- pressuraed program of PWR constructior. <ince 1973.
i >: tor a..oent or.ctuai nea<y wa'er The fourth type of reactor, the pressurized heavy-
!$@NI[urYeb7 viri 8 water reactor (MWR), uses heavy water as the coolant to x: ant;
, - [l,5 5^**
and unenriched uranium for fuel.Of the ten commer-g reacters
-l cial PHWRs currently in service, nine a'e of the Cana-
. '. AcaseoFed dian CAxou design; the other is a similar German
-v reaces esign. PHWRs have two distinct advantages: the use
==:3cjnyter of heavy water allows more eilicient absorption of the 7,
energy released,m the raactor, greater fuel " burn,
up," and therefore fuel economy; and they can be 63 f%
' t.
- / A 6 -- " h..
refueled without being shut down.The great majority 1
of reactors currently in service and under construc-
~~
S3 N' +- [
4 tion are light-water reactors ordered between 1967
~3 and 1974. An initial surge in orders in the United -
20 States frun 1967 to 1970 was followed by a second surge from 197^ to 1974 in the United States, West-
'I ern Europe, and Japan. Since then, forecasts of reduced electrical demand, escalation of construction a
costs, and public resistance have led to a near-morato-rium in ders ad many C3nceNadons.
- s75 ters 1977
-979 1979.
Year.
Batting Aserages reactor (PWR) and the boiling-water reactor (BWR). In l
the PWR, Cooling water is kept in the liquid state With average capacity factors of a, least 70 percent, j
under preuurc. which necessitates the use of large, the !cvel on which the utilities bred their planning.
l thick. s:ce! pressure s essels:in the B% R, cooling water PHWRs were consistently the top of the Icague from is allowed to boil. Both types use enriched uranium 1975 to 1979. GCRs have also been reliable but their fuel ar.d must be shut dcwn for three to six weeks per average output decreased when the capacities at year to allow spent fuel rods to be rep! aced. Both which the British units were rated were reduced,to l
reactor ty pes were originally developed in the United slow the corrosion of mild steel parts by the carbofr-States-the PW R by Westinghouse and the swr by dioxide Coolant.
General Electric. They are now used throughout the Average PWR performance has fluctuated but has world, notably in West Germany, Japan, France, and generally exceeded average BWR performance, which Sweden, as a result of technical licensing that enabled has remained at the bottom of the league--except in foreign suppliers to import.end even adopt, American 1979, the year of the accident at the PWR at Three technology.
Mile Island.
Another type of reactor, the gas-cooled reactor PWRs arad BWRs generally have greater planned out-(GCR), was developed in Britain and France in the ages and therefore lower capacities than GCRs and 1950s, oni,y to be abandoned in the late 1960s when PHWRs, partly because they must be shut down for
~
construction costs grew too large. GCRs use carbon-refueling. Pad partly because the low incidence of dioxide gas as the coolant (with some advantage over. equipment failures among GCRs necessitates ! css I
water in intrinsic safety) and unenriched uranitirn planned outage-for repair. And in the United States, fuel, and refueling does not require that they be shut requirements of the Nuclear Regulatory Commission down. In 1965 Britain sted for a more elaborate for backfitting with new components have also caused design, the advanced gas-coole- ; ? actor (AGR), which shutdowns and affe..ed capacity factors. These gener-promised greater efficiency and lower costs. Huwev-ic averages conceal wide variations among individual cr, construction difficulties have set baa the whole ' plants, however. Indeed, such variability is the domi-program to such an extent that the first aGR ordered nant feature of nuclear plant performance, especially (Dungeness B, some 16 years ago) is not yet operat-among light-water reactors, as reflected in the fre-ing. The French abandoned construction of tt :ir very quency and duration of unplanned outages.
cestly Gcas ir,1969 and have pursued an aggressive We tried to explain the variability in the perfor-f.tay/J me 1981 Ter.hn0logy Reuen 59
- ~ - ~,
q-
~~ - :
- a Very few outages resulted from failures of " nuclear core" components, probably because of exacting engineering standards.
w== - -
_. _ _ _ _ = = = - =. =. = = =. =. = -
= -
= = = - - -
_ _ ~ ~
W" - ~=- +==.==::=U. =..: ___.._ = =.. - - -- =. L e-
'--'~~'S'=-
T--___
_5=.=.---
-_-..~_.:===--=--
- -,.. _. _ _ - = - -
m: -r.
E
- _._._._.=g_T_....-
- -- __- = =-
%=_---
=-_-___.;_.-_:-_=_=_u._.._
. _ =. =, _ _. _
=_=-_
==,
. - =
. u =_ _-. _z d' -%-X. ~nf_==". 0. _ - ~_
E
-. _==..__.:-.=-
7 =-
=-F _ ;=%C"
~
- =-=-=--,__
. _ = = -
=-
=55' 2:F= in'~. =~ ';--.. =c_- ~E.-
b__
b'i
__ Dh _ _..g.. _ Z-k =__ # c ___~- 'I-E m_-
_~~ ---
~
r_ -. f r--
=n-----
_.===--.
u :
=
W2^. i=K- -M ~-T.
= ;;~_=:?=12-i4.5^ ~==~=_-M, l,!;-ni 1_-%t _^?=^q=_5_ g ' 2_.;;; _a2
~
~
^
_-=,=:-__.-=.---
~~
- - _ = _ _ = _ - -
_ _.q_
_1-----.-
- - - - - =. - - -
.f;;I=t= =- - m. ; _:i =.=-
. -. _-;; e * =_========- a_ _t-g== _ _ g_-== =-m C ~ ~ _ ~~ =~_===_. ~--~~= _'
jf_^~
4
- F,W_f.cf;;
f____
~ ~ = - - ~ ~
~
g._. _. _.. __ ;- - -
' =E=- -
- 7
..2 = -
=
~_='."'*.15:v_-6. _b2-' -
--~E-N ~. ~. :;__== m _ _ _ _
j.
i
.=i.-=.-.=.__.__--_===.._______-- - - -_ _ _ __
cr _ _--
==
=i
-.. =.
. _ _ = _ = = =
+- :-
t
-r y
5*
[
q W. %r;
% g,.%; n %)-~~1 (.
G;p.fikr,9 L
p.y Q
3 V'br dl( > d.!
~"
~
i ~ II l
I5
[.
N 3
- 0. - %
m
.. g lg p.s.re
- .h
._I i[
. _ _. = - - -
_ ~
(MQET 3%.
I M
IT M
{
=~==I'==M=*=W
=~==L=~==X==;==I= :
)
='
x
}
M.
' Aw' C d 4,
h h,
w
.m c-.,_
.m 3
t marce cf nuclear power p5nts by means cf regression not explain the variation in capacity factors, we analysis, fe!!owing the lead of previous authors. Plant adopted a more empirical approach to test some com-i l
capacity factor was chosen as the dependent variable mon hypotheses about nuclear plant performance.
l' and age, size, coolant pressure, and temperature as The first hypothesis is familiar enough: it hcids the independent variables. We then performed sepa-that performance typically improves as start-up prob-I rate regressions for PWR5 and BWRs using the average lems are overcome and plant operators learn the finer.
capacity factor ofindividual plants for the two neriods points of maintaining the reactor. After the phnt has 1975 to 1976 and 1977 to 1978 (twc-year averages operated for two years, equipment failures will be were used to iron out performance fluctuations owing i,nfrequent until components oegin to wear out much t
to refueling shtit' downs).
e later. This hypothesis was based on the performance l
But this analytical technique could explain only,7,.-
record of relatively small plants; unfortunately, to-t to 21 percent of overall rariability. Furthermore, day's large rea'cttrs of 900 to 1,200 megawatts have b;
there was no consistency in the relative effects of the not matched the performance of smaller plants at explanatory variables between the two periods. The comparable ages. (On-line -xperience can suggest rnain diffictity with this approach is that age, size, when p!anned shutdowns to refuel BWRs and PWRs ca't and coolant conditi.ons are not irdependent.
permit the repair of defective equipment, possibly to i
prevent an unplanned outage.)
1 Maturation, Size, and Temperature Effects To try to understand the effect of maturation, we grouped reactors of similar characteristics and age.
Because regression analysis (even with the i 9 ion Using capacity factors for every year that cach reactor -
of additional independent variables) apparently could had been in service, we were able to meter maturation so Tecnnc':;f Rev:ew uavane ast '
. n
,.e-
,s c t-
..s
I in a group of technically similar reactors and compare Variations in Manufacturers and Nations the performance of reactors of similar age but differ-ent technical characteristics.
The variation in performance among reactors may There was clear evidence of maturation for all sizes partly result from international dirTerences in quality of PWRs-but not prior to f7ur to five years of opere of maintenance and operator training and different tion, compared with the two :xpected. Units of 400 to levels of quality control among manufacturers. Of 51 800 megawatts have consistently achieved an average PWRs operating on January 1, 1978, 39 were in the capacity factor of over 70 percent from their fourth
' United States, 7 in Japan, and 5 in Germany. Thus.
year on. Data for units larger than 1,000 megawatts the average worldwide performance of PWRs. is are limited but suggest that their performance H, so weighted toward U.S. performance, which was better far been inferior to that of the 400-to-800 m-w att than Japan's but worse than Germany's.
group of comparable age. The smallest and generally Among P'VR manufacturers, Westinghouse as mar-oldest PWRs were usually mediocre performers, al-ket leader R. the average, Babcock and Wilcox units though two relatively new small PwRS have done well.
performed far below average even before the accident Furthermore, PWRs with lower coolant temperatures at Three Mile Island, and Siemens/KWU (West Ger-tended to achieve larger capacity factors than those man) units have worked much better than Westing-with' higher coolant temperatures. (Except among the house units in all size ranges.
oldest PWRS, coolant pressures have not varied much No country apar t from Switzerland, whkh has only and have not significantly affected PWR peria mance.)
one unit, has had satisfactory performance from its Evidently, capacity factor decreases as reactor size BWRs. Several of the four units in West Germany anc increases and coolant temperature rises, but it is sta-the seven in Japan were out of service for more than a tistically impossible to separate the two effects.
year. All German BWRs will require extensive modifi-Similar analyses on BWRs revealed no discernible cation, and two units (Gundremmingen and Lingeni maturation, size, or coolant effects. We noted that the wi'l be decommissioned after only ten pa.s of ser-performance of BWRs of all ages and sizes were less vice. The SWRs made by AEG/KWU (West Germa-satisfactory and predictable than that of other types of ny) have the poorest operating record; those of Gen-reactors.
eral Electric-again, the market leader-set the as er-By centrast, nine of th: ten PHWRs achieved 70 age; and these made by Asea Atom (Sweden) are the perent capacity factors from their first year, and the best of an unsatisfactory bunch.
(
performance of larger units is similar to that of small -.
er oncs. Howeser, the size iange, 320 to 791 mega.
Vinta;e Effects watts, is smaller than for light-water reactors. The mediocre performance of the tenth PHWR (Rajasthan, Plants of the same age, technical characteristics, and in India) is ascribed by the utility to " grid distur-rrake constitute a given vintage. Pe-formance is gen-bance"-factors unconnected with the plant itself.
erally expected to improve in successive vintages with The record of GCR performance is heavily influ-design modifications and more eflicient operation. In enced by the Britir' reactors, which account for 20 of practice, it is difficult to test this hypothesis because the 26 operating in 1978. Although the largest is few nuclear plants have sufficiently similac technica2 rated at only 676'mej;awatts, ocR performance tends specifications, and even those that are similar can to decline with increising side and as coolant condi-have significantly different major components.
tions be: cme more severe. Only the smallest ocRs-Some indication of the importance of vintage rated at less than 200 megawatts of capacity-have effects can bTgauged by ecmparing the performance average capacity factors exceeding 70 percent. And of duplicate units on'the same sites. We found tha:
I despite the deratings to mitigate corrosion, recent eight of ten duplicate units on nine sites in the Unitec events suggest that GCRs are wearing out. Potentially States performed appreciably oetter than their prede-l serious cracking at several Magnox pJants will necessi-cessors. This improvement probably stems from the tate lengthy outages; it is not yet known if similar learning achieved during construction, site manage-cracks are present at other Magnox plants. Because ment, and early operation and suggests the value or they were designed for a 20-year life and most ae maintaining a close relationship among manufacturer.
approaching that age, repair costs could exceed the architect-enginect, and utility.
value of retaining the afTccted plants in service.
i vape issi wocu ne,.e.
e-
.A
. ~. - - -
- c. -
eq were in " conventional," or nonnuclear, subsystems-Equipment F ikres the main heat-removal system, steam generator, feed.
To determine the most common causes of unsche-water-condenser system, and turbine generator.
duled outars, we used utility reports of the duration These subsystems must withstand mechanical stress and catise of plant outages for the three years 1975 to and chemical corrosion oflarge flows of heat, steam, 1977, published by the International Atomic Energy and water and the vibration and other stresses of Agency. Unfortunately, utilities report only equip-heavy rotating machinery. In contrast, very few out-ment failures that necessitate immediate shutdown.
ages resulted from failures of" nuclear core" compo-
'Other less sericus or potential equipment problems nents, proba.bly because of exacting engineering stan.
that are recti 6ed or for which repairs can be post-dards.
Half the outages from equipment failure in PHWRs, poned until the next planned refueling shutdown a,re and a smaller but signiScant fraction in PWRs and not listed. Nevertheless, useful insights into the deter-BWRs were rooted in main heat-removal systems.
minants of plant outage can be gained by analyzing These failures were more prevalent among newer and the reported equipment failures.
Most lost hours occurred in rare and unpredictable larger plants. Almost half the equipment-failure out-prolonged outages. Thus, the outages for each main ages among PWRs were caused by faults in the steam-subsystem were not significantiv correlated with age, generator and feedwater-condens:r systems. A large size, and steam conditions.
proportion of the steam-gener3 tor problems caused For all three current types of nuclear plant, equip-long outages, some lasting more than a year.
ment Tailure accounted for three of four unplanned Turbine generators caused nearly one-fifth of outages. In turn, three of four equipment failures equipment failures in all types of power plants.
H
/
l A program ci cent.nuing educa-The Advanced Study Program ccm-For Wer ate,a-e an
/
/
f j
aer' am 'um. maa j/
l thn fo* exper.enced engineers.
Oines classrecm study. sem.nars.
]CgQ f
f; scient:sts. technical managers.
gu.ded independent study, and g
i and educators.
research.
Director. Ac.anced /
4
,/
Stuey Program
/
/
MIT Room 9-4357 A unique opportunity to enhance The Program is d:vided into seg-
/
/
f p.ofessional capabilities thio 0gh.' ments which coincide with MIT Cambnc;e. 7
' I9
/
/
j increased technical competence.
summer. f all. and,sp. ig terms.
c l
j and to brcaden understanding of FeMows typ; cal!y',stan at the 7
/
emerging technologies.
bepinning of the f all term. or j
./
/
/
)
a with spec:st review subjects 7
j
,e f
,e Anintensive experience.indi-cffe' red during 'he surnener.
7
/
,/
/-
e U
/
vidually tailored to the back-7
/' j/
/
ground of each participant.
The Center for Advanced j
/,/
Engineering Stucyis new
[!
/ g g/ pf c(/ d /.+gj g
A,'/
j Resources throughcut MIT are accepting esplicafens c
c ?'f
'oOe 4/
c f
availabte to the FXows of the for the Fa!!1981 o
Center.
Program.
7 4
y
...yt.u ssi i
e2 Tecnnoiocy %.-
. ~. -
.o w w ; rs,
_e.
-e
'L:mli w w of miaue are bt ner undmtdl,imi dfettise reniedies are devek.pell.
stantlardization muid wd! >crve to reproduce faults along with good design elements.
Among light-water reactor;, the chief problems with means that inv'estment is risky. The risk is highest for turbine generators were crosion, corrosion, and fa-small utilities, where a. single nuclear plant represents tigue of the blades in the low-pressure section of the a sizeable fraction of system capacity; and all utilities slow-speed (1,800 rpm) saturated-steam turbine, to-must consider the high cost of replacement power gether with unspccified generator faults.
when nuclear plants are shut down.
The size of PWRs and BWRs seemed to play at least a Economic uncertainty. From an overall economic statistical role in the frequency of equipment failure.
viewpoint, the fact that nuclear core components are Among large PWRs, equipment failurcs in the main reliable is irrelevant; an outage of given duration is neat removal system and feedwater-condenser system serious in terms of replacement power costs no matter w ere appreciably higher than among smaller plants, as what its cause. Good performance can be achieved were output losses from testing, regulatory limita-only if all coraponents that cannot be replaced while tions, and operator error. But steam-generat~or fail-the plant is in operation are equally reliable, including ures were appreciably lower among newer large PWRs, valves, pumps, f ad pipework. An on-site inventory of suggesting that tube Icaks generally occur after an spare components can reduce the risk of prolonged incubation period. Operator error was the biggest outages, nd sustained research and development to cause of outage among large BWRs. owing to the shut-impro',e the reliability of " conventional" subsystems down of the Brown's Ferry units following the notori-of nuJear power plants may also improve the reliabil-ous cable tray fire With that exception, there were no ity of similar srbsystems in fossil-fuel plants.
important differences in the causes of unscheduled Bigger is not better. Larger nuclear units, especial-outages between larger and smaller BWRs.
ly PWRs of over 1,000 megawatts, will continue to,
All Japanese nuc! car power plants have suffered represent a greater investment risk'than medium-prolonged outages, many exceeding six months. Ja-sized PWRs of, say,400 to 800 megawatts, unless reli-pan's PWRs had persistent steam-generating problems, ability improves dramaticall) Wer the next few and the EwRs have experienCcd faults in Control-rod years.
drive mechanisms and Icaks from corrosion and Prcb! cms with suppliers and national standards.
cracking in the pipework.
Rep!ication of a gis en design may theoreticelly lead to better product re!! ability and lower costs, but the suc-Premature Standardization cess of West German PWRs and the dismal record cf both PWRs and Bb Rs in.Iapan Carry another important Recent!) there has been much interest among utilities _. message. To reduce recurring equipment failures'that and phnt suppliers in total plant standardization.
cause heavy output losses, manufacturers and engi-Standardization could well reduce liceming delays, neers must ensure that the quality of components.
construction costs, and equipment failures as manu-structural materials, and site work is uniformly supe-facturers and utilities incorporate experience into rior. If utilities are to anticipate and prevent failures.
plant design. But there remains the nagging problem they must maintain experienced teams of operators, of esaluating the causes of i aredictable perfor-technicians, and maintenance personnel and encour-mance. Vastly different implications for designers age or even institutionalize continuing contact be-would issue from analyses that pinpoir ted generic tween plant engineers and the original design team.
design.and materials defcets, substandard site work, No single factor cu guar, tee success; continuing, inadequate' quality contol, and igpropriate treat-meticulous attention m /s* be devoted to every facet of 1
meet of cooling water. Until'these causes of outage nuclear plant design, construction, and operation.
are better understood and effective remedies are developed, standardization could well serve to repro-4 John S ny has led the Energy Programme at the Science Policy duce faults along with Eood desi n iemen'ts. There-7 8
Rescare Umt at the Unnersit of sussex in Eng!Ind since 1969. After i
fore, we favor continuation of incremental ordering graeusting frem the t.ondon school or tcenomics in i956. he ortce as a rather than a long-term program of reactor standari-so'ernment econoraic advisor. sie,e Thomas joined the Energy P o-zation based on today's umts. We basEthis Conclusion 19M and then spent tix years in operational research m, o ga mei M. He gra uaw in cheminry from Brin industry and at on several of our fmdm, gs:
sussex t;niversity.
Unpredictability. Excluding PHWRs, which have ne E cgy Pregramme is funded by the British Research Councils, the Dc;.rrnent f Energy, and the British fuel industries. This stud) is consistently Performed well, the main feature of part of so enseing research program on the technotes.:ai, eeenomic, c:
nuclear plant perform ince has been variability, which social anects er nue: car po.er.
May/J.e iSat Technsto;;y Fen
- 63
..; 2.a
~
~
n
~
- a. s
.-