ML20012B897
| ML20012B897 | |
| Person / Time | |
|---|---|
| Issue date: | 03/31/1990 |
| From: | Su N NRC OFFICE FOR ANALYSIS & EVALUATION OF OPERATIONAL DATA (AEOD) |
| To: | |
| Shared Package | |
| ML20012B895 | List: |
| References | |
| TASK-AE, TASK-S902 AEOD-S902, NUDOCS 9003190134 | |
| Download: ML20012B897 (57) | |
Text
y-4
~
g.
1.
AE0D/S902 SPECIAL STUDY REPORT REVIEW OF THERMAL STRATIFICATION OPERATING EXPERIENCE March 1990 Prepared by:
Nelson T. Su Office for Analysis and Evaluation of Operational Data U.S. Nuclear Regulatory Comission Washington, D.C. 20555 i
i t
's IS 9003190134 900305 b
PDR ORG NEXDPDC
4-
e TABLE OF CONTENTS Page EXECUTIVE
SUMMARY
E-1 1
INTRODUCTION.......................................................
1-1 2 PHENOPENA AND CLASSIFICATION OF THERMAL STRATIFICATION.............
2-1 i
2.1 Phenomenological Description................................
2-1 2.2 Classification of Thermal Stratification....................
2-3 2.3 Descriptive Mode1...........................................
2-3 2.3.1 Global Thermal Stratification....................
2-4 2.3.2 Cyclic Thermal Stratification.....................
2-6 2.3.3 Thermal Striping..................................
2-6 2.4 Combined Effects............................................
2-7 3 OPERATING EXPERIENCE REVIEW........................................
3-1 3.1 Operating Events Caused by Thermal Stratification...........
3-1 3.1.1 Farley Unit 2 (PWR)...............................
3-1 3.1.2 Tihange Unit 1 (PWR), Belgium.....................
3-4 3.1.3 Genkai Unit 1 (PWR), Japan........................
3-7 3.1.4 Trojan Nuclear Plant (PWR).......................
3-0 3.1.5 Washington Nuclear Plant Unit 2 (BWR).............
3-10 3.1.6 Leibstadt(BWR), Switzerland......................
3-12 3.1.7 TVO I and II (BWR),
Finland.......................
3-15 3.1.8 Thermal Fatigue Cracking in Swedish BWR Reactors..
3-17 3.1.9 Craching in Feedwater System Piping (PWR).........
3-19 3.2 Operating Events with Potential Thermal Stratification......
3-20 3.3 Operational Data Searches...................................
3-21 1
w666mhmmm_. - - - _ _ - -
~
i
. o t
1 l
_ TABLE OF CONTENTS (Continued)
. Pag 4 SAFETY IMPLICATIONS, REGULATORY ACTIONS, AND INDUSTRY RESPONSES...............................................
4-1 4.1 Safety Implication of Thermal Stratification..............
4-1 4.2. Regulatory Actions........................................
4-l' 4.2.1 Information Notices and Bu11etins...............
4-1 4.2.2 Regulatory Review...............................
4-3 4.3 Industry Actions..........................................
4-5 4.3.1 Corrective Actions..............................
4-5 4.3.2 Long Term Actions...............................
4-6 5 FINDINGS AND CONCLUSIONS.........................................
5-1 6 REFERENCES.......................................................
6-1 APPENDICES APPENDIX A Criteria for Stable Thermal Stratification...........
A-1 APPENDIX B List of NRC Information Notices and Bulletins Related to Thermal Stratification............................
B-1 11
24 O~
TABLE OF CONTENTS (Continued)
LIST OF FIGURES
.P.agg 2.1 Turbul ent Penetration i n the RHR Line.........................
2-2 2.2 Thermal Stratification in the Pressurizer Surge Line...........
2-5 3.1 Farley Unit 2 - Location of Cracks in the High-Pressure Injection Line.................................................
3-2
- 3. 2 Farley Unit 2 - Downstream Temperature Profile'of Loop B.......
3-3 3.3 Tihange Unit 1 - Location of Cracks in the ECCS Injection Line.................................................
3-6 3.4 Genkai Unit 1 - Location of Cracks in the RHR Suction Line.....
3-8 3.5 WNP Location of Thermally Bowed Piping 3-11 3.6 Leibstadt, Switzerland - Arrangement of Feedwater Lines........
3-13
- 2. 7 Leibstadt - Temperature Profile of Feedwater Line..............
3-14 3.8 TVO II, Finland - Mixing Points of Feedwater, Shutdown Cooling System, and RWCU..:...........................
3-16 3.9 Swedish BWR Plants - Typical Arrangement of Feedwater Mixing Point..................................
4 3-18 3.10 Dresden Units 2 and 3 - Schematic Diagram of High-Pressure Coolant Injection System.........................
3-22 A.1 Limits of Stable Thermal Stratification with Leakage Rate = 1.0 gpm.................................................
A-3 iii j
i -
EXECUTIVE
SUMMARY
This report summarizes operating experience involving thermal stratification in piping systems of nuclear power plants which resulted in unisolable i
leaks from the primary system and unexpected pipe movement.
Thermal stratification refers to a condition in which two streams of fluid of different temperatures flow in separate layers without appreciable mixing.
Thermal stratification can be classified as global thermal stratification, cyclic thermal stratification, and thermal striping.
Global thermal strati-fication causes low cycle bending stresses and is responsible for unexpected l
piping movement.
Cyclic thermal stratification and thermal striping result j
in high cycle stresses which can potentially cause pipe cracks.
Thermal stratification was first identified in 1979 as the root cause for feedwater nozzle cracks which occurred at a number of pressurized-water reactor (PWR) plants.
No new industry-wide thermal stratification concern arose until an event at Farley Unit 2 in late 1987, resulted in an unisolable leak from the primary system.
Subsequently, reports from foreign reactors (i.e.,
Tihange Unit 1 in Belgium and Genkai Unit 1 in Japan) described similar events attributable to thermal stratification.
Data from installed instrumentation in some of these plants and in-depth investigation confirmed cyclic thermal strati-fication as the root,cause.
Since 1982, the licensee for the Trojan Nuclear Plar.t has observed unexpected
' movement of the pressurizer surge line causing contact of some pipe-whip restraints.
During the 1988 fuel outage, the licensee installed instrumentation to measure the actual temperature distribution and pipe movement.
Analysis of the data indicate thermal stratification routinely occurring in the surge line.
Additional events relating to thermal stratification include damage to pipe supports at the Washington Nuclear Plant Unit 2 (WNP-2) plant in 1984 and cracks on pipe joints in several foreign plants.
E-1
~
As a result of these events, the Office of Nuclear Reactor Regulation (NRR) issued Bulletins 88-08 and 88-11. These Bulletins require licensees to
'conduc. nondestructive examination (NDE) of lines having potential for thermal stratification and to analyze the structural integrity of the surge line.
The corrective actions taken by the utilities to date range from continuous monitoring programs to more agressive hardware changes. Although particular situations involving thermal stratification may have been prevented or the consequences may have been mitigated, none of the corrective actions could be expected to prevent or mitigate the consequences for all situations.
A mixed strategy of hardware and operational changes would probably yield the highest results in terms of eliminating the effects of thermal stratification.
Lessons learned and finds from the thermal stratification experience include the following:
(1) Thermal stratification has caused cracks, damaged supports, and contributed to thermal fatigue in high energy piping. Safety injection, RHR, feedwater and the pressurizer surge line piping systems have been affected. Some cracks have resulted in unisolable leaks of reactor coolant. Thermal stratification has occurred in pressurized-water reactors and boiling-water reactors.
(2) Operating experience indicates that many interfacing systems are potentially subjected to thermal stratification.
Review of operational events involving cyclic thermal stratification indicates that through wall cracks were generally found in the 90-degree elbow connecting the piping section containing stratified fluid and the piping section containing well-mixed fluid. Apparently, this location provides the minimum margins against thermal fatigue.
E-2
l e
(3) Searches of data bases from years 1980 to 1988 found over 2000 reports of pipe cracks and damage to pipe supports, snubbers, and pipe wipe restraints, in most of these cases, the failures were attributed to causes such as vibration induced fatigue or water hammer.
Some fraction of these failures was probably due to the effects of thermal stratification but not identified as such.
Detection of thermal stratification is difficult and usually only confirmed by measurements
)
from thermocouples placed around a suspected section of pipe.
(4) Typical monitoring programs to detect thermal stratification monitor only selected portions of lines which are considered to have a potential j
for thermal stratification.
Visual and dye penetrant examination would not detect cracks caused by thermal stratification until they are through wall since this form of thermal fatigue originates on the inner surface with crack depth propagating to the outside wall surface.
Fatigue cracks are difficult to dctect using even the more sophisticated ultrasonic testing and radiographic testing techniques.
(5) As plants age, those locations susceptible to the effects of thermal stratification are subjected to an increasing number of thermal cycles; and the likelihood of cracks increases accordingly.
Experience to date suggests that thermal stratification was not considered during the original design of current plants and would have to be a factor in any consideraticn of life extension.
(6)
Interfacing system leakage has a potential to cause thermal stratification in system piping and leakage through check valves, block valves, and isolation valves has been common.
Thus, an approach to minimize the potential for thermal stratification in many locations is to implement effective preventive and corrective maintenance of valves at system boundaries.
E-3
(7)
It was noted during the review that the leak-before-break approach has been accepted by the staff in specific instances in piping systems susceptible to thermal stratification, such as the pressurizer surge line. The fatigue and fracture mechanics analyses submitted to justify the application of leak-before-break rely on assumptions regarding the thermal cycles, thermal gradients, material properties, and crack geometry. The validity of the analysis over the long term is best supported by continued monitoring of operating experience related to the effects of thermal stratification.
(8) Existing NRC guidance and industry codes and standards may need to be revised to reflect existing operating experience, the results of ongoing NRC reviews of bulletin responses, and the current two-year Electric Power Research Institute study of thermal stratification.
Explicit acceptance criteria for analysis of systems subjected to thermal stratification may be needed.
I 1
l l
E-4
! '.i
=
1 I
)
f 1 INTRODUCTION s
Experience at domestic and foreign reactors indicates that thermal stratifi-cation has caused pipe cracks, damaged supports, and loosened flanges in the various high energy safety systems.
These safety systems include safety injection, RHR, and the pressurizer surge line in the primary coolant system.
Failuros of these systems could result in unisolable leaks of reactor coolant.
In many plants, the thermal loads resulting from thermal stratification were not considered during the design of these systems.
[
Problems caused by thermal stratification in nuclear power plants were recognized as early as 1979.
Thermal stratification is a condition in which two streams of fluid of different temperatures flow in separate layers without appreciable mixing.
The two streams remain stratified if the flow parameters meet certain physical conditions.
Originally, the phenomenon of thermal stratification and the associated thermal stresses were found in the feedwater system piping.
Recent operating experience, however, indicates that other high-energy piping systems are also affected.
As a result, the NRC took various regulatory actions in response to these problems.
In this report, the Office for Analysis and Evaluation of Operational Data (AE00) presents the results of its review of operating experience related to thermal stratification and the associated corrective actions that utilities have taken.
a 1-1
2 PHENOMENON AND CLASSIFICATION OF THERMAL STRATIFICATION 2.1 Phenomenological Description Thermal stratification is a condition in which two streams of fluid of different temperatures flow in separate layers without appreciable mixing.
I When the flow rate is low, turbulence is low and the potential for mixing is minimal.
The lighter hot fluid stays above the heavier cold fluid.
l The potential for stratification increases as the temperature difference between the hot and cold fluids increases.
Increasing fluid temperature difference increases the effect of density variation and the buoyancy force.
Fluid velocity has a negative effect on the potential for thermal stratification.
As the fluid velocity increases, the potential for turbulent flow and mixing increases.
The Richardson number discussed in Appendix A is a measure of these effects.
The Richardson number does not address the effects of piping configuration such as the pipe length and slope or the mixing effects of elbows and tees.
Nonetheless, the Richardson number is used as a screening criterion for assessing the potential for thermal stratification (Ref.1-4).
The effects of piping geometry are difficult to determine analytically and, consequently, efforts have been made to determine the effects experimentally.
For example, tests (Ref. 1) have been conducted to model flow conditions where the RHR line is connected to the reactor coolant system (RCS) hot leg (see Figure 2.1).
These tests sought to determine the penetration distance of turbulent flow from the hot leg into the RHR line.
The turbulent flow penetration distances is one of the parameters needed to identify piping that may be susceptible to thermal stratification and the corollary, those configurations which may be excluded from further consideration.
t 2-1
l I
HOT LEG i
r j
P l
d 1
i TURBULENT PENETRATION i
p s
i
(
(
>4 8
l L
Figure 2.1 Turbulent Penetration in the RHR Line 22
(,
,e The occurrence of pipe cracks or damaged hangers may indicate a thermal strati-fication condition.
However, confirmation of thermal stratification in piping i.
systems is best achieved by measurements of temperature circumferentially around the suspected section of pipe.
2.2 Classification of Thermal Stratification For purposes of this report and in agreement with current industry usage, the thermal stratification phenomena can be classified as: (1) global thermal stratification, (2) cyclic thermal stratification, and (3) thermal striping.
.These phenomena can occur in piping which communicates fluids with large temperature differences and low ilow rates.
The classifications provide a framework for discussion of a number of different problems that arise from thermal stratification.
These problems depend to some degree on a variety of factors including: (1) magnitude of th'e temperature changes, (2) number of thermal cycles, (3) frequency of thermal oscillations, (4) material characteristics of the piping, and (5) external restraints on pipe movement.
The consequences of global thermal stratification include: (1) macroscopic movements of the pipes and resulting hanger damage, (2) generation of stress in the pipes, which might not have been considered in the design of the piping system, and (3) low cycle fatigue.
The consequences of cyclic thermal stratification and thermal striping include pipe cracks resulting from high cycl'e fatigue.
- 2. 3 Descriptive Model The thermal-hydraulic conditions in the pressurizer surge line during various plant operations provide a model for development of a description of the different classifications of the thermal stratification phenomena.
In the case of the pressurizer surge line, all three classifications can occur at the same time, simply due to plant conditions which arise during normal startup and shutdown, i
i The following subsections illustrate this occurrence within the context of the three classifications.
I 2-3
i 2.3.1 Global Thermal Stratification Global thermal stratification refers to a flow condition in which the separating thermal layers of fluid are stable and vary slowly with time.
This classification is loosely termed " thermal stratification" by the industry.
However, to differentiate this from other types of thermal stratification, the term " global thermal stratification" is used in this report because it causes global thermal stresses on piping.
The concern over global thermal stratifi-cation is that it is a mechanism for pipe movement and potential dagage to pipe supports, hangers, and pipe whip restraints.
In the hot upper region of the pipe, compressive stresses develop as a result of constrained expansion (Ref. 5).
Since the pipe is flexible, the thermal moment gives rise to a bending stress which is superimposed on the membrane stress.
These low-cycle axial and tangential stresses due to strat-ification depend on the interface level and temperature difference.
During startup of a PWR, for example, the pressurizer is heated to form a
" bubble" while the remainder of the RCS is relatively cold.
As the bubble forms, hot water is displaced out of the pressurizer into the surge line.
This stream of hot water does not mix with the colder water already in the surge line.
A relatively stable condition of thermal stratification develops.
Figure 2.2 illustrates the separation of the hot and cold fluids in the pressurizer as a bubble forms.
In this example, corresponding to the classification of global thermal stratifi-cation, the pressurizer is initially heated to a temperature of 425'F to establish system pressure corresponding to the net positive suction head NPSH requirements of the reactor coolant pumps.
Since the rest of the RCS remains at-approximately 120'F, the peak temperature difference is about 305*F.
The interface between cold and hot fluid in the pressurizer surge line persists
_throughout the startup, although the temperature difference reduces as the RCS is heated by pump heat.
At normal operating conditions, the temperatures of the pressurizer and the hot leg are approximately 650 F and 600 F., respectively, i
j 2-4
_y a-
4 TEWPERATURE STRATIFICATION IN HORIZONTAL RUN
\\
PRES 80RIZER J\\
N x
slxxxxxxxx\\\\\\\\\\
CB,Gy,/yy/y//mv v n ~ ~
~u O
/
Th 7p a PRESSURIZER ftWP.
Th a NOT LEO TEWP.
t TYPICAL AXIAL TEMPERATURE PROrlLE WITH STRATIFICATION AT SECTION A-A 7
N 7
k>
sp/
/
STEADY-STATE pt9 vest evTsvest Figure 2.2 Thermal Stratification in the Pressurizer Surge Line 2-5 I
L, I'
2.3.2 Cyclic Thermal Stratification l
Consider again the stratified condit; ion in the pressurizer surge line as shown in Figure 2.2.
As plant heatup continues, reactor coolant pumps are started and charging and letdown flows are established to maintain level control in the pressurizer.
These evolutions lead to inflows and outflows from the pressurizer.
Because the flow rates i,n the surge line,are small compared to the diameter L
of the pipe, the interface between the hot and cold water is not disturbed sufficiently to cause significant mixing.
However, the physical location of the boundary between hot and cold fluid is shifted back and forth, depending on.the flow conditions.
Thus, some locations on the surge line are exposed alternately to hot and cold water.
This situation is called cyclic thermal stratification.
Plant operational data indicate that the period of cyclic thermal stratification temperature changes is generally on the order of a minute or more.
Instrumentation used to monitor the temperature changes of the surge line confirms the cycling of metal temperature in response to the fluid conditions inside the pipe.
These cycles continue during all plant operations.
Cor-responding to the decreasing temperature differences between the pressurizer and the hot leg, the magnitude of the cyclic temperature changes also decreases as I
the plant heats up to normal operating conditions.
2.3.3 Thermal Striping l
Thermal striping is the term applied to the temperature fluctuations experienced by the inside' wall of the pipe as a result of wave-like fluctuations of the boundary between the hot and cold fluids.
These fluctuations are generated by flow perturbations and propagated due to thermo-hydraulic instabilities at the boundary.
These fluctuations cause the inside of the pipe to be exposed alternately to the hot and cold fluid.
2-6
U x
I
- Experimental data (Ref. 7) indicate that the period of the temperature changes l
associated with thermal striping is generally on the order of 10 seconds or i
i less.
The effects of thermal striping on the pipe are similar to those of cyclic thermal stratification, except that the frequency of the temperature
{
changes is greater.
t The metal temperature responds to the fluid temperature fluctuation, but the temperature profile of the pipe interior changes in relation to convective and conductive heat, transfer at the fluid to pipe interface.
Thermal striping i
causes localized stress on the pipe metal and is considered a mechanism for crack initiation due to fatigue (Ref. 2).
Since the feedwater line cracking problems were identified in the PWR plants (NRC Bulletin 79-13), various research programs have been conducted in an attempt to resolve this issue (Ref. 5, 7-14).
2.4 Combined Effects In many actual situations, all three classifications of thermal stratification occur.- Relatively constant flow leads to global thermal stratification; varia-tions in the flow cause the boundary to cycle back and forth; and the fluctuations
+
due to thermal striping are superimposed on the boundary. Where thermal stratification can exist, the combined effects of these phenomena should be considered, c-1 i
t 2-7
F,p 3 OPERATING EXPERIENCE REVIEW 3.1 Operating Events Caused by Thermal Stratification 3.1.1 Farley Unit 2 (PWR)
On December 8,1987, Farley Unit 2 was operating at 33 percent power when an unidentified leak occurred inside the containment.
The following day, licensee personel entered the containment, with the plant in hot standby condition, to locate the leak and determined that the leak could not be isolated.
The reactor was shut down so that repairs could be made (Ref. 15).
The leak was located in the RCS loop B cold-leg safety injection line between a check valve and the RCS loop.
Ultrasonic testing found a throughwall crack on the six-inch safety injection line.
The crack was located at a weld connecting an elbow and a horizontal spool, as shown in Figure 3.1.
The defective section of piping was removed and sent to a hot-cell facility for failure analysis.
The results of the analysis indicated that the crack was caused by cyclic thermal stresses attributed to thermal stratification in the safetyinjectionline.
Following the December 8, 1987, event, the licensee conducted an investigation to determine the root cause of the cracks in the safety injection line (Ref.
I 16).
The investigation included temperature measurements, vibration measure-ments, and metallurgical evaluations.
The licensee installed temperature-measuring instruments on the high pressure injection lines.
The data confirmed that the fluid in the B loop, where the crack was located, exhibited I
high thermal stratification and the temperatures fluctuated significantly I
while the other loops showed very little thermal stratification.
Figure 3.2 shows the temperatures measured near the weld joint of the B loop which indicated the maximum temperature difference between the top (hot) and bottom 3-1
~ -.
4
. =
4; =
c
=
3" w
==- 2*
COLD LEG
,,6*
_RHR A. _
CHARGING LOOP A
~
PUMP r~g 3
h, o
6* \\ _
l l8 CHARGING CO..D LEG f
o L
LOOPB
\\
~
" ' ~
n l l PUMP
%.)
o
"_RHR Y20' C_
CHARGING COLD LEG
/.
8*'
~
Y
~
PUMP LOOP C
)l Y
3* b I
SEE
+
+
1
( DETAIL "A"/
.om w :::
- u LEAK PATH SCHEMATIC DIAGRAM OF SAFETY INJECTION L
_n n
)
L.
l
_l I
i CRACK 4
r RCS COLD LEG I
l LOOP B l
DETAIL "A" 1
y l.
Figure 3.1 Farley Unit 2 - Location of Cracks in the High Pressure Injection Line 3-2
q ui D?
ol s%
)
FARLEY SIS CL TEMPERATURE PROFILE 04 e
sotton i
- TO
My g g V S h N e [ % [ g *k
/ V p &
m v y v w d
o w n
w m
"o TM h _ A, d.
_n a A
J 1A A
_r dg c
'QTJ\\f )')
\\/
]\\ /' V A
' I
V'. _.
h 6
a y
~%fM Y
J w
~
o 5
I m
" c' LO )P h DOMSTI EMI '80 [T0t1 82
- l d
k l
[
"a k
/U A A P 4 M\\f y
~5 j h I
~
f-g u
r u
i 130 c
f 1
2
, oa.
10.7 10 8 11.1 11.3 11.3 11.7 11.8 12.1 12.3 12.3 12.7
't g (psou 4i) i
(
r-t (cold) of the pipe was approximately 215'F.
This temperature difference was caused by a small leak of about one gallon per minute (ppm) flowing through a closed bypass valve of cold charging water line.
Temperatures also fluctuated particularly at the bottom of the pipe with a cycling period between 2 and 20 minutes.
This cyclic thermal stratification subjected the pipe to the stress discusseo in Section 2.3.2.
The metallurgical evaluations further confirmed the root cause.
Results of the evaluations indicated that cracking was initiated on the inside diameter (10) surface and progressed radially outward to the outside diameter (00).
Crackiag also appeared at several locations over a broader region of the 10 surface, rather than from a single isolated initiation site.
Based on these observations, the licensee concluded that the cracks were caused by the high cycle fatigue mechanism.
The licensee presented the results of the investigation to the NRC staff (Ref, 6).
The licensee also proposed further actions to deal with this issue.
These actions included additional monitoring and more analyses to be performed by a NSSS vendor.
As a consequence of this event, NRC issued Bulletin 88-08, " Thermal Stresses in Piping Connected to Reactor Coolant Systems." This bulletin required licensees to identify and inspect unisolable sections of piping connected to the primary system which might be susceptible to thermal stratification and to provide assurance that any thermal stresses which resulted would not cause fatigue failure.
3.1.2 Tihange Unit 1 (PWR), Belgium On June 18, 1988, while Tihange Unit 1 was operating at full power, a primary coolant leak to the containment occurred.
The leak was unisolable and its origin could not be identified from the control room.
The plant was brought to hot shutdown so that the leak could be located.
A visual inspection inside the primary containment showed that the leak was located where an emergency core cooling systems (ECCS) injection line connected to the RCS hot leg.
The plant was then brought to cold shutdown.
1 l
3-4 L
~
' Examination of the injection line revea'ed several cracks.
The location of the cracks is shown in Figure 3.3.
One crack was in the base metal of the elbow wall; it was 3.5 inches long on the inside surface of the elbow and 1.6 inches long on the outside surface.
Another crack was found in the spool connecting the elbow to the nozzle in the RCS hot leg.
Two small cracks also existed near the weld connecting the elbow to the check valve.
The cracks were caused by thermal cycling due to leakage of cold water through an isolation valve (not shown in Figure 3.3).
The le'akage allowed cold water to flow into the pipe section that was connected to the RCS.
This created stratified liquids with significant temperature differences in that section of the pipe.
On the basis of the similarity of the failure mode, the staff believes i
that the Tihange temperature profile should be very close to that reported for Farley Unit 2 as discussed above.
The licensee took the following corrective actions:
(1) All affected parts were replaced or refurbished; (2) Temperature-measuring instruments were installed in different sections of the pipe; (3) -Permanent pressure-monitoring instruments were installed to monitor leakage in the ECCS lines by verifying that pressure upstream from the check valves was lower than the RCS pressure; (4) Operating procedures were modified to mitigate the effect of injection of
-cold water into the primary system; (5) An investigation was conducted to determine if a similar problem could occur in other systems connected to the RCS; and (6) A long term solution is being developed which will be implemented during the next refueling outage.
3-5
e.
.o 4
J 1
a. re.
1 tx verve f
C tcL 9 2
./ -
w v
/
/
v
. -.1 4.
.-r..
l 7
t e, w e, c i.........
. m.. u..
1 l
j
- d I '
- T* * ".* A A
l GN90 E VALVf
/s
/ N' %
(f asn we
,. /
j
/
s
......i
=
1 1
(
a=
i
.......i I
I Jis ro.e D.b A
2
.a L
. i ~.n 3
(
l y
1 suar ~ e.a
\\
,g 1
.,t\\
w im a=s..
Figure 3.3 Tihange Unit 1 - Location of Cracks in High Pressure Injection Line 3-6
O
['
3.1.3 Genkai Unit 1 (PWR), Japan On June 6, 1988, while Genkai Unit I was operating at 100 percent power, a primary coolant leak to the containment occurred.
The reactor was shut down for inspection.
The visual inspection revealed that the leakage was located at a weld joint of an elbow located in a horizontal pipe section upstream j
of the A-RHR first isolation valve (Figure 3.4).
The pipe section was directly connected to the RCS loop A hot leg.
The leak was'unisolable.
The crack was found in the weld metal of the elbow and was about 0.06 inches long and 0.008 inches wide.
Plant personnel conducted an investigation and concluded that the crack had been caused by the high cycle fatigue that resulted.from cyclic thermal stratification in the horizontal pipe section.
The thermal stratification was caused by leakage through a gland of the PHR first isolation valve.
The thermal expansion and contraction of the valve disk caused the valve seat gap to open and close periodically.
Flowing hot fluid caused thermal expan-sion of the valve internals, which stopped the leak.
When the leak stopped, valve internals contracted and reinitiated the leak.
When the valve was leaking, the high-temperature fluid flowed through the pipe section.
When the valve stopped leaking, the fluid in the pipe cooled.
Therefore, cyclic thermal stratification was created and subjected the pipe to the thermal stresses as discussed in Section 2.3.2.
The corrective actions included:
(1) Replacement of all affected parts; and (2)
Installation of temperature monitoring instruments.
3-7
I o
l From Reactor Yessel i
I sl N.._..M
\\-
RC$ Hot leg A i
/
./
To $ team Generator 8" Diameter ($chedule 140)
T RHR ! solation Yalve I
I Pipe Leak
,, q Location N,
y g
Leak.0ff Pipe i
s N,
i lToReactorcoolant 3,
Crain Tank h
track
's Location f
v k To RHR Pump l
Figure 3.4 Genkai Unit 1 - Location of Cracks in the RHR Suction Line 3-8 r
y 3.1.4 Trojan Nuclear Plant (PWR)
When the Trojan Nuclear Plant was shut down for refueling outage in April 1988, the licensee found that the gaps associated with pipe whip restraints of the pressurizer surge line had changed and the piping was in contact with one of the restraints.
These anomalies contributed to the forced, unexpected displace-ment of the pressurizer surge line, which had been observed since 1982, when the licensee began to monitor the line.
Although the licensee had repeatedly adjusted shims and gap sizes on the basis of analysis of various postulated conditions, the problem had not been resolved.
The licensee's most recent
{
investigation indicated that the movement of the pressurizer surge line was caused by thermal stratification of water within the line (Ref. 16).
The licensee performed a piping stress analysis modeling stratified flow (Figure 2.2) in the pressurizer surge line for the piping configuration in Trojan with a bounding differential temperature of 300 F.
Results of the analysis indicated that the surge line will deflect more when stratifi-cation is considered than when no stratification is assumed.
This deflection varies with increasing differential temperature between the top and bottom of the pipe.
The analysis also indicated that the surge line under stratified flow conditions would deflect downward, contact pipe whip restraints, and undergo plastic deformation which would result in the cold set of the pipe above its original location.
This agrees with the observed vertical set of the surge line.
The licensee presented the results of investigation and proposed corrective actions which include:
(1) Perform inspections and nondestructive examination (NDE) of the line; (2) Conduct piping integrity evaluation; and (3) Initiate a monitoring program to establish the actual temperature distribution and line movements.
3-9
p e
L l -
-As a consequence of this event, the NRC issued Bulletin 88-11 " Pressurizer Surge Line Thermal Stratification." That bulletin required licensees to
' conduct a visual' inspection of the pressurizer surge line and to determine that the pressurizer surge line meets the applicable design codes, including loads t
generated by thermal stratification.
L 3.1.5 Washington Nuclear Plant Unit 2 (BWR) p On August 22, 1984, Washington Nuclear Plant Unit 2 (WNP-2) experienced a thermal
-transient that damaged a portion of the feedwater system (Figure 3.5).
The event occurred following an outage of about five days when plant personnel began i
to admit feedwater slowly to the reactor vessel while the reactor was at about 1 percent power.
About 15 minutes.after flow was initiated, licensee personnel heard a dull " thud" in the plant.
The licensee found that several feedwater piping hangers and snubbers had been damaged and a flange had been loosened, allowing a small leak of feedwater (Ref. 17).
The licensee determined that the failures were caused by thermal deflection of the feedwater pipe, induced by stratified flow.
This stratified flow was a result of slow admission of cold feedwater to a pipe filled with high temperature water heated by RWCU (reactor water clean-up) backflow.
Following the event, the licensee installed instrumentation in the feedwater line to detect and record pipe movement and temperature differences between the top and bottom of the pipe.
The licensee reported that the cool water stratified at the bottom in the long horizontal runs of the feedwater pipe resulting in a peak top-to-bottom temperature difference of 200*F.
The feedwater line was observed to move downward, causing the hanger failure.
To prevent future damage to pipe supports, the licensee redesigned the pipe
. support system to accommodate large pipe displacements.
On the basis of a bounding case of thermal loads, the licensee replaced certain rigid restraints with spring supports and snubbers capable of large displacement.
This resulted in a more flexible support system.
3-10 i
_n
co-sg -
.v.
L'
\\
]
~v l
REACTOR PRESSURE THERMALLY VESSEL BOWED PIPING i
,x l
}
,,. A-h>
~l l.
b#
~.'
s RWCU RETURN
(_),
,?>.
kA I
3'/
g
/
P' s s
) ',
v s
f n
.s
(
l CONTAINMENT
(&
VESSEL YC4 4
4
- %m
- 0 4
l 4
I Figure 3.5 WNP Location of Thermally Bowed Piping 3-11
4 Notrecurrence has been reported since the redesigned piping support system waslinstalled in 1984.
As.a consequence of this event,.the NRC issued Information Notice 84-87, " Piping Thermal Deflection Induced by Stratified Flow."
3.1.6 Leibstadt (BWR), Switzerland
- In October 1984, during startup testing, repeated leaks were observed at the-flanges of the ventu.'i flow meter in the feedwater line (Figure 3.6).
An 4
investigation determined that the leaks were caused by thermal stresses from temperature stratification in the feedwater line, i
i The feedwater line was instrumented to dete mine the causes of thermal stratifi-~
I i
cation.
Two possible conditions which may have caused thermal stratification
{
were identified.
First, hot water from the reactor cleanup (RWCU) system could leak into the feedwater line, which could be at ambient temperature.
- Second, j
water could backflow through the feedwater sparger to the feedwater nozzles.
\\
Additional inctrumentation was' installed to measure the_ temperature difference
_ profiles of these two potential conditions.
These measurements indicated that the-maximum-temperature difference for backflow of hot RWCU into the cold _ feed-water system was 356 F, while the maximum temperature difference for backflow l
of~ hot reactor water to the feedwater nozzle was 230*F (Figure 3.7).
-l 4
To prevent recurrence, the licensee took the following corrective actions:
-(1) ~ Gate valves were replaced with check valves to-prevent backflow of RWCU system water into the feedwater line; i
-(2) Operation of the RWCU system was changed so that the temperature of the.
returning RWCU water approximately equals feedwater temperature; and d
3-12 s
4
..y.
t
'..L 3;
,-h 5
v VENTURI.
~$
Z l
X
~
- FEEDWATER I
'~'~~~.~'~'~' 1 1i VEi4 TURI.
YL RHR l-1
~
%v W
s FEEDWATER.
s a
i u,
P Figure 3.6 Leidbstadt, Switzerland - Arrangement of Feedwater Lines 3-13 i
f
W I (.
"., ; q :.
=.
500'-
INCREASE IN FEEDWATER FLOW 400 3(
-v h
g 2
y m
275'F g
LINE ATMM = M'F 2
m t
~
BOTTOM OF 68'F FEEDWATER LINE if I
I i
I l
i i
i
., li 2
3 4
5-TIME (HOUR)
Figure 3.7 I
-Leidbstadt - Temperature Profile of Feedwater Line 3-14
5~
- (3)-.A special mixing device was installed in the RWCU system to minimize thermal stratification where the hot water from the regenerative heat exchanger mixes with cold bypass water.
3.1.7 TVO I'and II (BWR), Finland During the annual maintenance inspection conducted at the Teollisuuden Volma I
Osakeyhtio (TVO) I and II plants in July 1983, cracks were discovered at the
=
mixing point between the shutdown cooling system and the feedwater lines.
At the TVO I and II units, the RWCU return is connected to the shutdown cooling system before the shutdown cooling system is connected to the feedwate ' system (Figure 3.8).
During normal operation, the feedwater temperature is about 356"F and RWCU water temperature is about 527*F.
During startup and shutdown
~
the feedwater temperature could be as low as the ambient temperature of 68 F.
During startup and shutdown, when hot RWCU water is injected into the cold feedwater line, the temperature difference between these systems reaches its maximum, and thermal stratification becomes significant.
Stresses resulting from this thermal stratification were identified as the cause of cracks at the mixing points.
The corrective actions included:
(1) Additional inspections; (2) Replacement of the affected parts; and E
- (3) Installation of a new type of mixer.
The -long-term effectiveness of these corrective actions has not been determined.
=
3-15
=
- s.. n 9
4:
RHR SHUTDOWN COOLING AND RWCU
'c RETURN s
s 5
s
\\'
y s
s
\\
N
\\
g N
I g
N N
i d
D
\\\\NNNNNxxxNN' xxxxxNNA's\\\\\\\\\\ -
l,
,.K l
- Kd l
- KH l
- K4 Ns FEEDWEATER :
- .gg
'Hd. ~
l l
l W
j i
-l I-l Figure 3.8
' TVO II, Finland - Mixing Points of Feedwater, Shutdown Cooling and RWCU 3-16
____m= _ _ _ _ _ _ _ _. _ _. _. _. _
I___.__
._ - ~. _ _
__~-_ _ _ _ _ _
._____m_.m_.m__m.__
_______m____________.
4 a
3.1.8 Thermal Fatigue Cracking in Swedish BWR Reactors During inspection of the Oskarshamn 2 nuclear power plant in the summer of_1980, cracks were discovered at a branch pipe connection between the feedwater system and the shutdown cooling system (Ref. 18).
This led to the examination of
^ similar branch pipe connections in other Swedish BWR plants.
Cracks were found in similar locations in Barseback 1, Ringhals 1, and Forsmark 1.
In 1983, t
cracks were also discovered at Oskarshamn 1 and then'again at Oskarshamn 2.
Figure 3.9 shows a typical arrangement of the affected pipe connection.
The temperatures occurring in these pipe connections are very similar to the temperatures in TV0-II, the Finnish plant, as discussed in Section 3.1.7.
Several corrective actions were taken to prevent recurrence:
(1) Cracked thermal sleeves were replaced; (2) The affected piping was repaired; (3) The Swedish Nuclear Power Inspectorate initiated a developmental program and requested concepts for improved design; (4) Temperature measuring instruments were permanently installed in some' branch piping connections; and (5) New thermal mixers were developed.
3 The mixers, intended to create strong turbulence, were. installed in all Swedish BWR plants to mitigate the thermal stratification problem.
However, during the inspections of the Forsmark 1 and 2 plants in the spring and summer out-ages of 1983, new thermal fatigue cracks were discovered in the T-joints.
3-17
p._, - -
n i.
e
~. '..
g:
t f-(
8 s kg,*is a
gf' L:3 : u
/ 18 ?.,
- 3 RR
/j mvm GEN
/
oeu
/
'y
^ Y A\\\\\\D D..
?
f Auxiliary l
Q eed-Water System l l
30'C (Intermittant) l Operating Temperature = f II'W##
\\
\\
^
Crach Le,. + 4 g s
/h 4,
t M
,s
, e h
h 1
N is % w h%wwwN I
'8 i
f'-
l Crack Locaticn l Thermal Sleeve l reed-Water S stem l
Operating Temperat.
l IS0* C
=
' Material : Stainless Steel XS Crtii 1E9 with a maxm:.
carten centent of 0.051 - Corresperds to
.ASO4 A304 1
l M b D h\\ N M D \\\\\\ D D D D \\
^
Figure 3-9 Swedish BWR Plant - Typical Arrangement of Feedwater Mixing 3-18
c.:
- This discovery led the Swedish Nuclear Power Inspectorate to conclude _ that the
- new mixers would not solve the problem and to decide that the pipe material i
would be the next factor for investigation.
In the interim, the'Swedish Nuclear Power Inspectorate imposed requirements for more frequent inspections-and established acceptance criteria for continued operation if cracks were detected.
3.1,9 Cracking in Feedwater System Piping (PWR)
IE Bulletin 79-13 described operating events involving cracking in feedwater system piping including. events at 15 Westinghouse and Combustion Engineering-designed plants.
These plants reported a crack in the feedwater elbows
. adjacent-to welds in elbows of steam generator nozzles.
While conducting.an extensive metallurgical investigation, Westinghouse concluded that the cracking had been caused by corrosion fatigue.
Subsequently,
- instrumentation was installed in-several plants to obtain data to determine the potential forcing function contributing to cracking.
The data suggested that stratified flow conditions existed in the feedpipe weld region during zero and low power operations.
This thermal stratification tends to support the fatigue aspect of the postulated failure mechanism.
As a result of the findings,Bulletin 79-13 requested all PWR-licensees to:
(1)' Perform radiographic and ultrasonic examination of all feedwater nozzle-to pipe welds and of adjacent pipe and nozzle areas; (2) Repair or replace the cracked piping identified from the examination; (3) Perform another examination at the next refueling outage; or (4) Change operational procedures such that the feedwater level within the steam generator is maintained essentially constant and no intermittent cold auxiliary feedwater injection in utilized during startup, hot standby, or cold shutdown operations.
3-19
The licensees have committed to the above requirements.
3.2 Operating Events with Potential Thermal Stratification
'The thermal stratification potential exists in piping where two systems at
'different temperatures and pressures are connected.
All ECCS, RHR system, charging and' letdown, reactor core isolation cooling (RCIC), and other systems are connected to the primary system.
Check va'1ves or block valves are used to isolate these systems from the primary system.
However, operating experience indicates that, in many instances, leakage through these isolation valves occurs.
These systems may experience thermal stratification and associated thermal stresses-in the pipes.
For example, on January 20, 1989, at Arkansas Nuclear One, Unit 1, following a reactor trip, the operators started one high pressure injection (HPI) pump to maintain the pressurizer level (Ref. 19).
This operation created a pressure differential across a check valve that prevents'RCS coolant backflow.
Because this check valve did not seat properly, some RCS coolant at high temperature leaked through the check valve into the other HPI system piping where the fluid is-colder.
The licensee became aware of RCS flow beyond the primary system when a tape attached to the HPI piping began to melt, smolder, and smoke, activating a smoke detector. 'The NRC issued Information Notice No. 89-36 to alert the licensees regarding HPI system piping exposed to fluids in excess of design temperature.
However, should the leakage be small, the HPI system piping would potentially be subjected to thermal stratification and associated thermal fatigue.
The recent event at.the Dresden Nuclear Plant provides another example of
-potential thermal stratification.
The licensee discovered damage to-piping supports on Unit 2 on October 28, 1989 and on Unit 3 high pressure coolant injection system (HPCI) on October 31, 1989 (Ref. 20).
Prior to the event, the licensee had observed a temperature gradient along the HPCI discharge line since February 1989.
The temperature gradient was determined to be caused 3-20
x 1
by backflow of reactor coolant through the outboard discharge valve'(see Figure 3-10).
With the temperature gradient in the line, the licensee conducted surveillance tests'of the Unit 2 HPCI system on October 5 and 13, 1989.
Following these tests, the licensee discovered damage to the piping supports which was possibly caused by water hammer.
Thermal stratification could have caused similar damage.
The root' cause of the event is being investigated.
Similar flow conditions had occurred in the HPCI discharge line at Brunswick on March 11, 1987 (Ref.'21), however, no damage was reported.
In summary,' interfacing system leakage has a potential _ to cause thermal stratification in the system piping.
Leakage through check valves, block valves, and isolation valves occurs in many operational events.
3.3 Operational Data Searches Searches of the sequence coding and search system (SCSS) for events involving pipe cracks revealed 1784 related licensee event reports (LERs) from 1980 to March 1989.
Among these events, 136 LERs were related to.the chemical and volume control system (CVCS), 36 LERs to the RHR system, and 17 LERs to the
. pressurizer in PWRs.
These three systems (i.e., CVCS, RHR, and pressurizer) are. systems which have experienced thermal stratification.
Most of the cracks were found in weld joints.
Usually, the crack was attributed to vibration.
'A review of the abstracts.of these 189 LERs did not identify thermal strati-fication as a root cause nor could thermal stratification be inferred from the description of the event.
Another search of the SCSS database looked for all LERs reporting damage to snubbers, supports, pipe hangers, and pipe whip restraints.
The search yielded 52.LERs.
Review of the abstracts of these LERs revealed that thermal stratification was not reported as the cause of the damage.
3-21
- p. ~;,
_ a ;-
n,
'a f
Q c
HPCI V
BOOSTER HPCI J k PUMP DISCHARGE PUMP TO TORUS
?TO CST u
)
>4 FEED TER UNE GLAND SEAL e
CONDENSER LEAKAGE z
Cr C OUNGWATER PUMP r,
r, U
O FROM TO CONDENSATE -
CONDENSATE STORAGE STORAGE TANK TANK Figure 3.10 Dresden Units 2 and 3 - Schematic Diagram -of High-Pressure Coolant Injection System 3-22 1
7, n y,
The staff'also searched *the nuclear plant reliability data system (NPRDS) for events related to cracks of pipe or fittings or damage to pipe supports for the RCS, the RHR system,' the low pressure (LP) safety injection system, the CVCS, and the high pressure (HP) injection system.
The searches revealed 71 events for the' RCS 18 events for the RHR or LP safety injection system, 90 for the CVCS, and 26 for the HP injection system.
None of these events describe a icause which could be related to thermal stratification.
Awareness of operating experience involving thermal stratification is growing due to industry follow-up of the events described above.
Some investigations of pipe cracks and hanger failures have included installation of instrumen-tation which has confirmed the occurrence of thermal' stratification.
No doubt, some of the large number of cracks identified in the past as due to fatigue caused by vibration or other causes were actually caused by phenomena related to thermal stratification.
g..
3-23 3
- .g r
s
.h I
4' SAFETY IMPLICATIONS, REGULATORY ACTIONS AND INDUSTRY RESPONSE 4.1 Safety Implication of Thermal Stratification
' Experience at domestic and foreign reactors indicates that thermal stratification has caused cracks, damaged supports, and loosened flanges in high-energy piping systems.
These systems include safety injection, RHR, and the pressurizer surge line in the primary coolant system.. Failures of-these piping systems could result in unisolable leaks of reactor coolant.
In many plants, the thermal loads resulting from thermal stratification were not considered during the design of
.these systems.
s-
.As plants age, those locations susceptible to the effects of thermal stratification could'be subject to an increasing number of thermal cycles and the likelihood of cracks could increase.
Also, plans for life extension of current plants could be affected.
Continued-attention to operating experience will help identify areas of concern and potential solutions.
4.2 Regulatory Actions
- 4. 2'.1 Information Notices and Bulletins As early as 1979, the NRC recognized the direct consequences of thermal
-stratification. In response to the reported cracks on the feedwater nozzle-to pipe welds of.PWR' plants, the NRC issued IE Bulletin 79-13.
In
' December 1984, Information Notice 84-87 was' issued to alert licensees 'to the WNP-2 incident involving piping support damage due to thermal _stratifi-cation in a-feedwater line.
l
-The Farley Unit 2 event on December 9,1987, prompted the NRC to issue
~ Bulletin 88-08, " Thermal Stresses in Piping Connected' to Reactor Coolant l'
l 4-1
c; H
i Systems."' Supplements 1 and 2 to Bulletin 88-08 were issued subsequently to provide additional information and to alert licensees to the Tihange Unit 1 event.
Supplement 3 to Bulletin 88-08 was issued to discuss the Genkai event.
Bulletin 88-08 requests licensees to:
(1) Review systems connected to the RCS to determine whether unisolable sections of piping are subjected to stresses from thermal stratification; (2) Examine nondestructively the unisolable section of piping as identified in item 1; and (3) Plan and implement a program to provide continuing assurance that the unisolable sections will not be subjected to combined cyclic and static thermal and other stresses that could cause fatigue failure during the remaining life of the unit.
NRR has received responses to Bulletin 88-08 from all licensees, These responses provide the results of licensee review of systems-connected to the.RCS to identify unisolable sections of piping.
The AE00 staff has sampled responses from each type of NSSS. 'The PWR licensees indicated that the auxiliary pressurizer spray line, normal charging line, and alternate charging line are potentially subject to stress from thermal stratification.
There are unisolabic sections of these lines.
Some licensees, however, simply responded that there are no such unisolable sections of piping in their systems, offering no detailed discussion.
All of these responses
' are being reviewed by NRR and its consultants.
The staff is in the process
~ of developing acceptance criteria and an action plan to address the responses to the Bulletins.
As a' result'of the Trojan experience of unexpected movement of the pressurizer surge lines, the NRC issued Bulletin 88-11 " Pressurizer Surge Line Thermal Stratification," on December 20, 1988.
The bulletin required licensees to:
I 4-2
s (1) Conduct aLvisual inspection of the pressurizer surge line; and (2)' Determine whether the pressurizer surge line meets the applicable design code and other final safety analysis report (FSAR) and regulatory commitments for the licensed life of the plant, considering the phenomenon of thermal stratification and thermal striping in-the fatigue and stress evaluation.
- The appli, cable code cited,above is the ASME Code,Section III.
Although the code does not specifically address thermal stratification conditions, the code does require design consideration of fatigue which would include the concern of low and high cycle thermal fatigue resulting from thermal stratification.
Appendix B lists-NRC bulletins and information notices that deal with the issue of thermal: stratification.
4.2.2 Regulatory Review In response to'the recent events, NRR has held meetings with owners groups, NSSS~ vendors, and individual utilities.
The NRR staff is actively reviewing the. licensees' responses to Bulletins 88-08 and 88-11, in particular, responses from the so-called near term operating licenses (NT0Ls), which had applied the-leak-before-break (LBB) approach to justify design of the piping system without pipe whip restraints.
These licensees have submitted detailed analyses which include the dynamic effects of low and high cycle thermal ' fatigue resulting from.
thermal stratification.
These analyses are intended to validate continued application of the LBB concept for the 40 year plant life (Ref. 22.).
The staff has accepted the design of the pressurizer surge line for the South Texas plant
-(Ref. 23).
The concept of LBB allows piping designs which eliminate certain pipe whip restraints and jet impingement barriers intended to prevent damage to safety 4-3
i equipment due to catastrophic failure of high energy pipes.
The rationale for L~
the relaxation of the requirement for pipe whip restraints is that leaks will develop and be detected prior to catastrophic break in the high energy line, and will allow time for safe shutdown of the plant.
The regulatory basis for the application of the LBB concept is General Design I
Criterion (GDC) 4 of Appendix A to 10 CFR 50 which states that " dynamic effects associated with postulated pipe ruptures in nuclear power units may be excluded from the design basis when analyses reviewed and approved by the Commission 1
demonstrate that the probability of fluid system piping rupture is extremely low-under conditions consistent with the design basis-for the piping."
s 1
Detailed guidance for development of analysis in support of the LBB approach j
'isprovidedinNUREG-1061(Ref.24)andtheproposedStandardReviewPlan(SRP)
Section3.6.3(Ref.25). NUREG-1061 guidance includes a provision which restricts application of the LBB approach in "high energy fluid system piping, or. portions thereof, that operating experience has indicated particular
- susceptibility to failure:from the effects of corrosion (e.g., intergranular stress corrosion cracking), water hammer, or low and high cycle (i.e., thermal,
'l
-1 mechanical) fatigue." Likewise, the SRP includes a provision that "there is t
adequate mixing of high and low temperature fluids so that there is no potential for significant cyclic thermal stresses."
q Prior to the recognition:of the existence of stratified conditions in the surge
.line, the original design of the South Texas surge line assumed well-mixed fluid, such that the effects of low and high cycle thermal fatigue could be excluded.
l With the need to consider thermal stratification, the NRR acceptance of the j
I South' Texas surge line analysis for LBB continues to be based on NUREG-1061 and SRP 3.3.6.
However, the provision that the fluid is well mixed is not met but is addressed by explicit calculation of the stresses which arise from fluids which are not well mixed.
These additional stresses are then applied in the l
)
mechanistic fracture mechanics analysis described in NUREG-1061 and SRP 3.3.6 as'the basis for LBB and the fatigue analysis required in Section III of the ASME code.
44
r7p..
l,
?
c4 a m.
4.3 Industry Actions 4.3.1 Corrective-Actions
,o Plants which previously identif.ied thermal stratification as a concern b
have taken a variety of corrective actions to deal with the thermal r
stratification problems.
The following summarizes these actions.
(1) Monitoring Program - Through various monitoring programs, plant operators I
attempt to (a) measure the actual temperatureLdistribution and line movements in the piping system involved, and (b) detect leakage in the line (Genkai), (c) limit temperature differences of fluid in the surge line, or (d) monitor the ECCS line pressure (Tihange) to verify that the pressure in the affected section is always lower than the RCS pressure..
Monitoring programs are only a partial solution since they do not monitor all: lines which have a potential for thermal stratification.
Licensees should recognize the'importance of eliminating or detecting small leaks; however, none of the affected plants explicitly proposed the initiation of a preventive maintenance program on primary system isolation valves as-a corrective action.
+
.(2) Reliance on Inservice Inspection - Cracks caused by thermal stratification initiate on the inside surface of the pipe and propagate to the outside.
Visual and dye penetrant examination will not detect these cracks until they are through wall.
These fatigue crack.s are difficult to detect using even the more sophisticated ultrasonic testing (UT) and radiographic b
_ testing techniques until they have progressed to near through wall.
I
\\
4-5
b
'(3) Hardware Modification - Examples of hardware modifications include:
-(a) replacement of gate valve's with check valves (Leibstadt) and (b). development of an improved mixing device (Swedish plants).
Hardware
~ modification did provide a partial solution to the problem as the Swedish BWR operating experience indicates.
-(4) Operation Procedural Changes - Some utilities modified operating procedures to reduce temperature differences at the point where fluid from two systems meet or reduce the time that the unit operates 'in a configuration susceptible to thermal stratificatio'n, r
This measure, accomplished by heating or cooling one of the fluids, is feasible for certain applications.
(5) Changes to Piping Supports - Piping supports have been changed to allow more movement of the affected piping as an effective way to deal'with bending stresses resulting from global thermal stratification.
In some cases, to meet seismic criteria, o.lder plants installed more E
supports than are currently needed.
The new seismic criteria allow more damping and, as a result fewer supports may be required.
For example, on the basis of stress analysis, Trojan and Beaver Valley Uriit 2 I
.have-removed some piping supports to accommodate thermal stresses.
4.3.2 Long Term Actions Owners groups for PWR and BWR reactors recently initiated a program called
. SCATS. (striping, cycling, and thermal stratification) to investigate thermal-s stratification as a generic-issue (Ref. 26).
We understand that the industry also requested EPRI to conduct a study of thermal stratification issues.
An action plan has been developed, and it recommends a two year effort to resolve
.the-issue.
4-6
.n
s a
5; FINDS AND CONCLUSIONS The following paragraphs summarize the finds and lessons learned from the thermal stratification review.
(1) ' Thermal stratification has caused cracks, damaged supports, and contributed to thermal fatigue in high energy piping. Safety injection, RHR, feedwater and the pressurizer surge line piping systems have been
.affected.
Some cracks have resulted in unisolable leaks of reactor
'hermal stratification has occurred in pressurized-water coolant.
T reactors and boiling-water reactors.
-(2) Operating experience indicates that many interfacing systems are potentially subjected to thermal stratification.
Review of operational events involving cyclic thermal stratification indicates that through-wall' cracks were generally found in the 90-degree elbow
' connecting the piping section containing stratified fluid and the piping section containing.well-mixed fluid. Apparently, this location provides the minimum margins against thermal fatigue.
(3) Searches of data bases from years 1980 to 1988 found over 2000 reports of pipe cracks and damage to pipe supports, snubbers, and pipe wipe restraints. In most of these cases, the failures were attributed to
- causes such as vibre. tion induced fatigue or. water hammer.
Some fraction-of these failures was probably due to the effects of thermal stratification but not identified as such. Detection of thermal stratification is difficult and usually only confirmed by measurements e
from thermocouples placed around a suspected section of pipe, i
5-1 F
7
- w
~m:.(4) Nypical monitoring, programs to detect thermal stratification monitor only selected portions of lines which are considered to have a potential for thermal-stratification.
Visual and dye penetrant examination would not detect cracks caused by thermal stratification until they are through wall since this form of thermal fatigue originates on the inner surface with crack depth
-propagating to the outside wall surface.
Fatigue cracks are difficult to detect using even the more sophisticated ultrasonic testing and radiographic testing techniques.
(5) As-plants age, those locations susceptible to the effects of thermal stratification are subjected to an increasing number of thermal cycles; and the likelihood ~of' cracks increases accordingly.
Experience to'date suggests that thermal stratification was not considered during the original design of current plants and would have to be a factor in any.
consideration of life extension.
(6)
Interfacing system leakage has a potential to cause thermal stratification in system. piping and leakage through check valves, block valves, and isolation valves has been common. Thus, an approach to-minimize the potential for thermal stratification in many locations is to implement effective preventive and corrective maintenance of valves at system boundaries.
.(7)
It was noted during the -review that the leak-before-break approach has been accepted by the staff in specific instances in piping systems susceptible to thermal stratification, such as the pressurizer surge line.
The fatigue and fracture nechanics analyses: submitted to justify the application of. leak-before-break rely on assumptions regarding the-thermal cycles, thermal gradients, material properties, and crack geometry. The validity of the analysis over the long term is best supported by continued monitoring of operating experience related to the effects of thermal' stratification.
5-2
z. :-
1 (8)L Existing-NRC-guidance and industry codes and standards may need to be-l
' revised to_ reflect existing operating experience, the results of ongoing-NRC reviews of bulletin responses, and the current two-year Electric Power.Research Institute study of thermal stratification.
Explicit acceptance criteria for analysis of systems subjected to thermal stratif,1 cation may be needed.'
+.
5-3
e,-
b v,
., REFERENCES 1.
WCAP-12171, " Evaluation'of Thermal Stratification of the Surge Line and RHR Line for Comanche Peak Unit 1," February 1989.
2.
BAW-2085, " Submittal in Response to Nuclear Regulatory Commission Bulletin 88-11, " Pressurizer Surge Line Thermal Stratification," May 1989.
3.
WCAP-12248, " Evaluation of Thermal Stratification for the Comanche Peak Unit 1 Pressurizer Surge'Line," April 1989.
4.
WCAP-12087, Rev.' 1, " Evaluation of Thermal Stratification for the South Texas Units 1 and 2 Pressurizer Surge Line," January 1989.
5.
Wolf, L. et al..." Thermal Stratification Tests in Horizontal Feedwater Pipelines," NUREG/CP-0091, Vol.
5., pp. 437-464.
6.
Reeves, E., " Summary of Meeting Held on January 15, 1988, Between NRC and APCo Representatives to Discuss Generic Implications of a Cracked Six-inch Safety Injection Pipe at Farley Unit 2," February 8,1988.
7.
Haefner, W. and Wolf, L., " Derivation of Mixing Parameters from the HDR --
Thermal Mixing Experiments," Battelle Institute e.v., Frankfurt an Main (Germany, F.R.) International Journal of Pressure Vessels Piping, Vol. 33.1. 198.
pp. 41-97.
8.
Chen, F.F. et al., ? Turbulence Modeling of Thermal and Fluid Mixing in PWRs During High Pressure Coolant Injection Using COMMIX-1B, Numerical Methodscin Nuclear Engineering," Part 1, INIS-mf-11168, August 1983, pp. 226-247.
~ 9.. ' Miksch, M., Len, E., and Loebberg, R., " Loading Conditions in Horizontal Feedwater Pipes of LWRs Influenced by Thermal Shock and Thermal Strati-fication Effects," Nuclear Engineering Design (Netherlands), Vol. 84:2, January 1985, pp. 179-196.
6-1 1
i i
- , y....
10.
Kasza, K.E.'and Kuzay, T.M., "The Influence of an Elbow on Horizontal Pipe p
Flow Thermal-Stratification," Transaction American Nuclear Society, Vol. 43, 1982,;pp. 780-857.
- 11.. Hu, M.H., Houtman, J. L. and White, D.H., " Flow Model Test for the i
Investigation of Feedwater Line Cracking for PWR Steam Generators,"
L ASME Vol. 81-PVP-4, 1981, pp. 1-10.
~12.
Fujimoto T. et al., "(xperimental Study of Striping at the Interface of Thermal Stratification," J. Heat Transfer (United States) Vol. 15, 1981, pp.73-150.
13.
Lyon, W.F., Hassan, Y A. and Peddicord, R.L., " Simulation of Thermal Mixing in Main Stream Generator Feedings," Transaction Am. Nuclear Society Vol. 55, 1987, pp. 737-809, 14.
Black,.P. S.; et al., " Tests of the TRAC Code Against Known Analytical Solutions for Stratified Flow," Nuclear Engineering Design (Netherlands)
Vol. 108: 1/2, June 1988, pp. 121-133, p
15.
LER 87-010, Alabama Power Company, Docket No. 50-364, dated January 6, 1988.
16.
Portland General Electric Company, Trojan Nuclear Plant, Docket No. 50-344,
" Pipe Whip Restraint Gaps Pressurizer Surge Line Movement," May 1988.
17.
Washington Public Power Supply System, WNP-2, Docket No. 50-397, " Design Engineering Report, WNP-2 Feedwater Piping Thermal Deflection Events,"
l October 8, 1984.
.18.
Swedish Nuclear Power Inspectorate, " Thermal Fatigue Cracking in BWRs,"
1984.
19.~ -LER 89-002, Arkansas Power and Light Co., Docket No. 50-313, dated March 31, 1989.
6-2
g.
p.. n.,1
--;o e
.y....t.
M -Y
[20.
U.S.-l Nuclear Regulatory Commission, Inspection' Reports, Nos.'50-237/-
[
- 89023'(DRP) and 50-249/89022 (DRP), dated November 16, 1989.
g E
- 21. - LER 87-004, Carolina Power and Light' Co., Docket No. 50-324,: April 10, I
f"
' 1987;--
=:
);
~f
--22.
Memorandum Hou, S.N.=to Marsh,.L.B, " Status of.Bulletin 88-11 Review-.
b Regarding POVR Surge Line Thermal Stratification," August.2, 1989.
23.
NUREG-0781 " Safety-Evaluation Report Related to the Operation of South F
- Texas Project, Unit 2," March 1989.
t s
24."NUREG-1061,Vol.3;"ReportoftheU.S. Nuclear.RegulatoryCommission-Piping Review Committee - Evaluation of Potential for Pipe Breaks,"
f November'1984.
25.
Standard Review Plan, Section 3.6.3, " Leak-Before-Break Evaluation Procedures," Federal Register Vol. 52, No. 167, August 28, 1987.
26? Handout-from meeting with representatives from Westinghouse and Comanche Peak licensee on July 27, 1989.
]n c
-i
- I
-}
t 9'}
[:
O U
,,1.i,.
, Appendix A Criteria for Stable Thermal Stratification 1
Therm'al. stratification means nonuniformity _in fluid _ temperature in a direction transverse to the flow path.
For simplicity, two streams that coexist in a pipe with hot stream of temperature T and cold stream _of temperature T are assumed.
h c
Normally, the hot stream stays on the top portion of the pipe and the cold stream stays on the bottom.
Given this definition, stable thermal stratification means that a finite bound-ary exists between the hot and cold streams.
The stability of stratified flows depends on the stratification equation:
g dp p dy R$=
gg12 (A-1)
,dyj g known as the Richardson number (Ref. A-1).
Here, g denotes the acceleration due to gravity, p the density, U the fluid velocity, and the positive direction of-y _is measured vertically upwards.
The subscript w. refers to the value of the velocity gradient at the wall.
Where R$ = 0, the fluid is homogeneous.
Equation A-1 can be expressed in terms of the parameters, which will be more practical for thisLproblem, as follows:
5
$ = 16 S AT d (A-2)
R Q2 L
where p=volumeexpansivity-(1/F)
AT.= temperature difference ( F) d = inside diameter of. pipe (in.)
Q = flow rate (gpm)
A-1
<r
.f -
L
.c H e' j
p For the pressure of-interest (s2500 psi), the volume expansivity,is about 5 x 10 4.c With this value, equation A-2 becomes:
8
'o g-AT d5 (A-3) j E 8 x 10 3
-._, _._._.. _. Q2 L:
Using the Richardson number,,the limit of stability has been determined:
)
- (Ref. A-2) as
v b
R9=1 (A-4)
'with R$ > 1 denoting stable stratification and R < 1 unstable stratification.-
For illustration, assuming flow rate = 1.0 gpm (Technical Specification leakage limit), one can calculate the limit of stability for varied pipe sizes versus temperature differential-between the hot and cold streams.
The results'of the calculation were plotted in Figure A.1.
Figure A.1 provides general criteria that determine the threshold of potential therma 1' stratification.
For example, a potential thermal stratification will i
~
Lexist in a pipe with an inside ' diameter of two inches if the-temperature differ-ence between the hot and cold streams is larger than 250 F.
Applying' Equation A-3, the mixing flow rate for the. pressurizer surge line was y
calculated to be approximately 700 gpm for a typical 14-inch schedule 140 pipe-with a' temperature difference of 310 F.
Since the insurge or outsurge flow rate during plant-heat-up,.cooldown, or-normal plant operation is only a few
'gpm,. thermal stratification always exists in the pressurizer: surge line as' indicated by the related criterion.
'1 A-2
- ~
3
- ,y-t 1
4' ;[p h
i*..-
3 10 ' --
1 I
I I
I I
i 8
6 4
f 2
2 c
10 L
g w
oa
- 6 STABLE THERMAL
'E STRATIFICATION
$^
4
~
.S THERMAL O
MIXING
.p
-w a:
g
(
- 2 a
E IV 10 8
6 4
g
/
L 2
i I
1 1
I i
1 2
3 4
5 6
7
- e' PIPE INSIDE DIAMETER (INCH)-
t Figure A.1 3
Limits of. Stable Thermal Stratification with Leakage Rate = 1.0 gpm A-3
+,
f*
-.. ' i.,
b i:.
REFERENCE A
{
A-1 Schlichting, H., " Boundary La:/er Theory." McGraw-Hill Company, 6th Edi tion,1968.
A-2 Turbner, J.S., " Buoyancy Effects ir, Fluids," Cambridge Unive'.sity Press, 1979.
{
P L-W L.
=
a ll7 JL-5_
M A-4
=
E.
w,
+
c i, *,,
r.
l APPENDIX B'
(
I list of NRC Information Notices
.q and Bulletins Related to Thermal Stratification
.i
, - Information Notice or Date of F-
- Bulletin Sul, ject
- Issuance y
Bulletin 79-13 Cracking in Feedwater 10/16/79 1
[
System Piping
[
IN'84-87
. Piping Thermal Deflection 12/3/84
(
Induced by Stratified Flow
~
l IN 88-01 Safety Injection Pipe 1/27/88 l
Failure l'
i Bulletin 88-08 Thermal Stresses in Piping 6/22/88.
q Connected to Reactor Coolant g
F Systems
- Bu11etin.88-08 Thermal Stresses in Piping 6/24/88-
~
- Supp. 1 Connected to Reactor Coolant j
. Systems J
. Thermal Stresses in Piping 8/4/88-4 i'
Supp. 2 Connected to Reactor Coolant Systems l
a' IN 88-80' Unexpected Piping Movement 10/7/88 i
Attributed to Thermal i'
Stratification u
l,
?
B-1 j
b
t i i
y
- f
,t,.
k i
^
t_..
.: a a.
<OJOO 1
APPEl4 DIX B (continued)
Information Notice or Date of Bulletin Subject Issuance
.Bulletin 88-11 Pressurizer Surge Line 12/20/88 Thermal Stratification
-Bulletin 88-08 Thermal Stresses in Piping 4/11/89 Supp. 3 Connected to Reactor Coolant Systems
(
l l
1 i
i B-2 J