ML20215B540
| ML20215B540 | |
| Person / Time | |
|---|---|
| Site: | Surry  |
| Issue date: | 01/28/1987 |
| From: | Hauber R NRC |
| To: | Kanda S JAPAN |
| Shared Package | |
| ML20215B065 | List: |
| References | |
| FOIA-87-20 NUDOCS 8706170360 | |
| Download: ML20215B540 (3) | |
Text
I
.f j-1 p& /
4 f
4: ~ '.
5
?...
bec: JRShea i
JAN' 2 81987
.RDHauber HJFAulkner
. Woodruff, IE i
RSSenseney 1
Tc1bfectir.11 d sto" KDBurke
, File-Japan
]
l Ur. Sunan Kanda
)
PifFCtor
')
Nuclear Power Sefetv Administration Division Agency of Natural Resources and Eneroy, NITI Inkya, Jepan Fax:. P1/3/501-Ofu Vsrify:
E01-1511
.)
r
Dear t'r. Kreda:
)
~
Thank 'yr;u for v(ur latter Of DeCFmber 27,19f6 on the Surry Unit 2 acciden+.
I apprcciato receiving the ereinsure vou provided en tha Sacordarv Pipiric Systre Aoino Effect irsnoctions n# Kensei Nuclear Power Plants and beve sont it te the technical staf f for their use.
The PRC Arcident investioatiar Taar Depcrt ce Surry is scheduled to be conple+ee ir early February and I will send you 'a copy as soon as it is availatle, very of the cuestiem you raisad in your letter will be ane,wered in detpil ir the repert. Howaver, attached ere some praliminary answers to vrur cuestiers.
l
.,J1'3".8W' moneR D.58'bac Ronald D. Hauber Assistant Director for Internatienal Cooperation A#fice of Interna +,ioral Prorren's i
Attachrart:
Prelicirery answers to oustio n trensritted in Kanch li/07/&F letter to Hauber i&C T..infaesir.ilrc 201/4M 761' Verifv:
49?-7371 8706170360 870611 PDR FOIA ZWELLINGB7-20 PDR LU i
..#.....S............
.g^ W..."I..E...........
o,,,c, >
m r.8.u s.b......RM,d r,u,y,,,,
,,9,ul,hm,,,,,,q.,,,,, u,h,.........
, r,,,,,,
-">.v.t.v.n......
.. ~.....v.d 82........
.ugs.........
..........f *.'~'
.f 2. &.
ucson -.osomnewouc OFFICIAL RECORD COPY t u.s o.o tassoo.m
'g.
1/28/87 Preliminary Response to Questions Raised in 12/27/86 Kanda Letter to Hauber.
(Complete information will be provided in the NRC/AIT report.)
1.
VEPC0 mircoscopic observation does not indicate cavitation.' 'AIT j
report will-provide complete detail of the pipe's condition.
2.
Ruptured pipe was A-234 grade WPB carbon stael, 18-inch pipe with 1/2-inch wall.
3.
Temperature 380 degrees Fahrenheit, pressure 367 PSI, flow at elbow 17-feet per second, chemistry showed low oxygen.
4 AIT report will answer cuestion in detail.
l 5.
AIT report will answer question in detail.
6.
Upstream: 24-inch ccr.censate header, spool "T" off header at 90 degrees, d
.18-inch diameter pipe (90 degree elbow to spool is elbow that ruptured).
All piping horizontal. Complete piping layout will be contained-in AIT report.
7.
No knowledge of any such inspection.
I 8.
Not aware of any thinning prior to rupture. Unit 1 inspection has now revealed thinning in condensate elbows, feedwater pump suction "T" and elbows, feedwater pump discharge line to first point header, elbows "T" and. reducer, feedwater dise;se ge header "T" and elbow, recirculation line to condenser elbow and valve.
.9.
VEPC0 reports that a change in water chemistry had taken place.
- 10. Overstress in elbow caused by wall thinning followed by unstable crack tearing.
- 11. No vibration, however main steam trip valve had closed 41 seconds prior to pipe failure.
- 12. Yes.
- 13. Modest pressure transient.
~
- 14. VEPC0 reports materials used in construction, configuration of piping, fluid velocity, water chemistry, and temperature above 200 degrees Fahrenheit contributed to wall thinning.
- 15. No.
I
- 16. No.
i gune,...s.,EB.
F 02 '87 11: 05 N..RC.M.[5,5.A.G.E., CENTER BETHE5DR..MD P.001 5
=
- -- > u a asa m e r,.,ni t w o:
N ACM 0370 FAC64 MILE TMAN8MITTAL. REQUEST oAva jg merunN o ioiN.6,0 seNoen l4 'd f)f.
h vas ll l No a
403 5 488 70
//'\\
s PAcalMILE Me No.
vsRIPicATION PHONE NO.
NO. OF WAnasfinetusfing Dem.
'L C'-Y' st.2-wr/
ava-ermo
""~ '~7"*""
w MEM488 PROM MAM pace #Msta PMoma NO.
VERIPICATION PMONE No.
NYW fed 09 0 JRhd W MonkJ Sul DING OPPICE PMONE NO, MAIL STOP PASCSDENCE A;//,hs o24Jd 02 w
- ~'
' ~ ~ ~ ~
=
ASCElvs0 VI' TAAMSMITTSO i
.p....
1 n
- p...,.......
RECf0NAL ADwtN6TRATOR DEPUTY AM.itE.7xATort For4 8?.0c l e/n
[
FEB.02 '87 11:05 NRC MESSAGE CENTER BETHE5DR MD P.002 i
k Um4TeOsTAfs8
[
NUCLEAR REGULATORY COMs4lSSION s
t wasmusrom.o.c. semen k...*
.m aoas i
MEMORANDUM FOR: Themis P. Spets. Director Division of Safety Review and Oversight F1t0M:
Robert J. Bosnak. Chief Engineering Issues Branch Division of Safety Review and Oversight SU8 JECT:
REPORT ON THE OUTCOE OF THE NRR TECHNICAL METING HELD ON 1/16/87 TO CON 51 DER THE SENERIC IMPLICATIONS OF THE SURRY FEEDWATER LIM FAILURE On January 15, 1987, the NRC staff invited experts from several engineering disciplines (piping design, metallurgy nondestructive examination, water chemistry, corrosion, and fluid mechanics) to participate in a technical panel discussion on the pammeters believed to have had an important role in the December g.1986 pipe break at the Surry-2 plant and the means to predict and mitigets the effects of erosion-corrosion in piping systems. The cause of the Surry-2 faibre which occurred in the feeduster piping near the suction side of o% of the feeownter pumps has been identified as pipe thinning from erosion-corrosion, however, at-least one panel member believes that cavitation erosion cannot be completely excluded. The actual failure of the thinned pipe well resulted frem a system pressure transient, not a classical water hanner event.
The complex interactions of the individual variables which affect the erosion-corrosion phenomenon ad influence the rate at whiah it proceeds, have not been thoroughly established by available research activity in this country or abroad.
The panel discussion elaborated on the role of those parameters which could have potentially contributed to erosion-corrosion in the feedwater piping system. A copy of the meeting agenda is attached.
The technical emperts on the panel has several observations and recornendations to make and they are suuustrized in this memorandum. A more detailed set of meeting minutes is under preparation and will be forwarded to all attendees when completed.
Observations
- 1) The phenseena associated with piping wear due to erosion, erosion-corrosion, or cavitation, have not been recognized as significant problems by piping designera and, consequently, they de little to accommodate piping wear due to theA thenomena in design. Some architect-engineers provide guidance in system e= sign specifications to minisiae erosion-corrosion by limiting maximum bulk flow velocity in water filled systems. such as feedwater and t vice water, but most architect-engineers focus primarily on meeting code a6...imum pipe wall requirements, with some variable added well thickness, and on meeting stipulated pressure drop rMuirements for equipment
,, m! v if M'
,q um.m f
. ll j (
i
~
/y
y pw s.g y-p m.
p a
4
(;
FEB.02 '87 11:06 HRC RESSRGE CENTER BETHE5DA RD P.003 Themis P. 5peis operability. The Heat Euchange Institute has provided guidelines to avoid erosion and erosion-corrosion in condensers and feedwater heaters.
- 2) The NRC initiative on omitting the dynamic effects of postulated ipe ruptures,whenevertechnicallyjustified,shouldnotbeaffected$ythe Surry failure. Safore leak before break analyses can be applied to any system..it must be demonstrated that phenomena which may challenge the pipe integrity, such as, water hassner, corrosion, cmep, fatigue, erosion, erosion-corrosion or other environmental conditions are either unlikely to occur. or have been evaluated and do not challenge the pipe integrity over I
the life of the plant. This is discussed in the broad scope rule change to GDC 4 and in the proposed new SRP section.
- 3) The amount of erosion-corrosion damage and the rate at which it proceeds is a compler phenomenon depending on a number of variables. In carbon steel systems the process is dependent on the dissolution of the metallic evide (magnetite) protective surface layer. This rate of wear decreases with an increase in oxygen concentration and with an increase in pH above g.2.
Flow velocity and regions of turbulence caused by pipe fittings, particularly.
30' splitting tees, also abet the wear process. Temperature plays a role in the process with a peak in the amount of dans6a occurring in the 250-340'F l
range. Erosion-ccerosion wear can be significantly reduced by material selection (increased Cr), and even the No may possibly reduce the rate at whicgmsence of small enounts of Cr, Co.
damage occurs.
- 4) In the case of Surry and other U.S. plants, the three most important variables influencing the corrosion-arrosion process are material local fluid velocity / turbulence and water chemistry /pH. The panel could not identify which of these three variables Jes controlling but believed that the Surry specific piping configuration with its high bulk flow velocity i
and turbulence at the splitting tee and close-coupled elbow confWuration had a significant role. The role of the past operating water chen,1stry conditions is uncertain. It is possible that most of the damage could have occurred sarly in life, with an appreciably lower damage rate following
'nstallation of the condensata polishing system.
- 5) Although erosion-corrosion is not a new or unknown phenomenon. it has i
received relatively 11tth study in the U.S. because incidence of recorded fa11eres has been low and because the relationship of the variables influencing.the process is comples. Erosion-corrosion failures in two phase systems and erosion in single phase systems containing suspended solids were noted to be such more prevalent end consequently, have received more attention in operation and maintenance than the single phase erosion-corrosion phenomenon. Some utilities have programs to control and mitigate l
damage in two phase systems such as steam estraction lines and turbine wet steam piping and in single ase systems containing suspended solids, such as silt in service water p1 ng. It is possible that single phase system erosion-corrosion failures ich did not result in personnel injury may have been experienced in non-nuclear applications without records being made of their failure and repair. Erosion-corrosion or erosion in service water systems would not likely result in a catastrophic failure due to the low operating temperature and pressure of the systems.
l
\\
n FEB.02 '87 11:07 HRC MESSAGE CENTER BETHE50R MD P.004 4
Themis p. Speis secos.endations
- 1) The'pene1 believes that adjusteents to pH or ovygen content fran levels now in i
use to protect steam generators, should not be sede without a thorough study of the possible global effects of such changes on the entire system.
Often changes are made to fir a specific problem in a nuclear plant without considering the total ramifications of the fix. The addition of hydrogen to BWR systems was cited as an erwaple of another proposed fir which requires careful study. Long term water chemistry needs for nuclear
. plants, particularly, a better understanding of the effects of pH and oxidizing or reducing environments on seterials of system components, was cited as a possible topic for additional research. Another important topic mentioned for additional research was the effect of local fluid dynamics in typical pipe configurations.
- 2) The panel believes that the factors influencin the rate at which the erosion-corrosion phenomenon proceeds in sing 1 phase systems cannot be ranked because the available current quantitat we data are insufficient.
However, the panel believes that the information which is available relative i
to the identity of.the involved parameters and how they influence erosion-corrosion in single phase systems should be pmsented to the utilities so that they)may formulate programs to ultrasonically measure esisting (baseline pipe well thickness at susceptible locations in specific systems.
Screening criteria, such as that developed by Virginia Power. should be used
)
on the most susceptible systems to identify potential locations of maximum The need for periodic monitoring could be determined from the results wear.
of such baseline measurements.
- 3) Data resulting from baseline wall thickness measurements taken as a result of recommendation (2) should be cortslated with flow velocity, turbulence, water chemistry. temperature, actual material chemical competition and installed original wall thickness to the estent that such czta are available. The data should be centrally collected and processed to increase overall knowledge and understanding of erosion-corrosion. Quantitative erosion corrosion data for single phase systans appear to be more generally known and available in Europe. and apparently receive some consideration there in the design recess. The Netherlands. France. the Federal Republic of Germary and the et Union were mentioned as having joint government-utility sponsored on erosion-corrosion.
- 4) The American Society of Mechanical Engineers (ASME) should consider the need for providing appropriate guidance to system designers on the subject of erosion and erosion-coprosion in its conventional pressure piping and nuclear piping codes and standards. Additionally the Subcomunittee on Nuclear Inservice Inspection (8/C XI) nf the ASME Soiler and pressure Vessel Committee and the ASME groups active in plant aging and life entension matters should be sede aware of the need to consider tiquiring pipe wall l
thickness measurement in their respective programs. Dr. 5. H. Bush.
Chairman of $/C XI. agreed to carry this recomunndation to $/C XI at their meeting during the week of 1/1g/87. One panel member expressed particular l
concern with possible erosion-corrosion in BWR sein steam piping where two phase conditions are prevalent, and suggested that $/C XI review both single l
phase and two phase systems. The Chairman of the ASME B31.1. Power Piping i
8 l
l i
I
l
+-
w u..
_c l
FEB.02 '87 11:08 HRC MESSAGE CENTER BETHESDA MD P.005 Themis P. Speis '
Section Comittee, indicated that he would present the information developed at the meeting'to the 831.1 Cassittee since they now have under development a proposed non-mandatory appendix on operation and maintenance. And they may wish to include guidance on well thinning. The Chairman of the ASME Boiler and Pressure Vessel Committee, the Chairman of the ASME Power Piping Section Committee (331.1), the Director of ASME's Nuclear codes and Standards and a representative of the National Board of Boiler and Pressure Vessel Inspectors were invited guests and were present for the meeting.
EPRI was represented at the 1/15/87 NRR technical meeting. EPRI has since advised us that they have assembled a technical task force and are in the process of issuing an informative technical white paper which will contain general recessendations on the subject of erosion-corrosion. INP0 was also represented at the meeting. INP0 issued a Significant Event Report (SER) on the Surry incident on January 7,1987 and they have advised us that their 1
evaluation is continuing.
Virginia Power will make a workshop presentation for utilities on February 10 I
simultaneously at sir locations (NRC city locations) throughout the US to share infomation within the utility industry of its program relative to the feed-water line failure and subsequent corrective actions.
On January 16, 1987 Jack Rosenthal of IE was contacted. It was agreed that a follow-up IE Information Notice (!N) to IN 86-106 should be prepared presenting j
the inforststion developed at the 1/16/87 NRR technical meeting. The Inforumtion Notice should focus on criteria to select potentially susceptible locations to erosion-corrosion in s measurement of pipe wall thickness,pecific single phase systems for the as discussed in recessendation #2 and #3.
In addition, the IN should again mention the failures which have occurred in two phate wet steam piping so that utilities understand that these lines continue to require their attention.
R. Johnson of E!B and D. Tereo of DPL A (temporarily assistine DSR0) will aid in the development of the new 1N. Both individuals assisted In setting up the 1/15/87 inseting and are now working on completing more detailed meeting minutes, preparing for an ACRS full connittee 1
meeting on February 5, 1987 and a Commission briefing on February 17, 1967.
l-Robdrt J. Bosnak, Chief Engineering Issues Branch Division of Safety Review and Oversight Enclosure Moeting Agenda j
l 1
1 i
t
[gb* *ECQ),
UNITED STATES l
E' NUCLEAR REGULATORY COMMISSION
^
g i
5 j
W ASHINGTON, D. C. 20655 i
s FEB 0 4 987 1
MEMORANDUM FOR: Themis P. Speis, Director Division of Safety Review and Oversight i
FROM:
Robert J. Bosnak, Chief Engineering Issues Branch l
Division of Safety Review and Oversight
SUBJECT:
MEETING MINUTES OF 1/15/87 TECHNICAL PANEL DISCUSSION ON SURRY-2 PIPE FAILURE IMPLICATIONS
Reference:
Memorandum from R. J. Bosnak to T. P. Speis dated January 30, 1987 In the above referenced memorandum, the observations and recomendations of the technical experts who participated in the January 15, 1987 panel discussion on the generic implications of the Surry-2 feedwater pipe failure were summarized. The purpose of this memorandum is to provide a more detailed Sununary of the technical discussions between and the presentations by the panel experts.
l Enclosed is a summary of the meeting minutes, the list of meeting attendees, and copies of the slides presented by the panel experts. The attached meeting minutes will be distributed to all attendees who requested a copy.
/L Rob t J. Bosnak, Chief Engineering Issues Branch Division of Safety Review and Oversight
Enclosures:
As stated cc:
V. Stello H. Denton R. Vollmer J. Taylor E. Beckjord R. Fraley C. Heltemes P. Shewnon G. Arlotto Internal Distribution List External Distribution List i
1 l
F t
L F,,
rors r,-ov t
!i d*/04
R.
MEETf MG SIMMARY ON THE SURRY ? P!PE FAILURE IMpl1CATf 0NS 1
~
0n.1anuary. 15, 1087, the NPC staff invited industry experts from various engineering. disciplines -- piping design, metallurgy, non-destructive testing, water chemistry, and fluid mechanics - to participate in a technical panel discussion on the parameters believed to have had an important role in the December 9.- 1986 pipe break at the Surry-2 plant and the means to predict _ and i
miticate the effects of erosion-corrosion in piping systems. The cause of the L
Surry ? pipe break -- which occurred in the feedwater pipino near the suction side of Feedwater Pump A-- has been identified as pipe thinnino from erosion-corrosion. The actual failure of the thinned pipe wall resulted from a system pressure surce although not a classical water hamer event. The complex i
interactions of the individual variables which affect the erosion-corrosion phenomenon and influence the rate at which it proceeds have not been thoroughly established by available research activity in this country or abroad. The penel discussion elaborated on the role of those parameters which could have potentially contributed to erosion-corrosion in the feedwater piping system. A list of meetino attendees is provided in Attachment 1 to this sumary.
Maior Factors Contributing to pipino Erosion Corrosion 1
1.
i"pino Design (Presented by D._ Landers, E. Rranch, F. Rodabauch) 1 The configuration of the piping at Surry-2 where the break occurred is believed to have played a ma.ior role in establishing conditions which promoted erosion-corrosion. The pipe failure occurred on the intrados of
/
a 90 degree flono radius 1 18-in diameter elbow which branched off a M-inch diameter flow-splitting tee. The break initiated in the pipe material (A2?a Grade WPR) -- a low alloy carbon steel -- and not in the I
weld. The average bulk flow velocity of the water was calculated to be 17 f t/sec in the 18-inch branch and 12 ft/see in the 74-inch header.The elbow-tee configuration has been identified as an undesirable desian arranoemert which is believed to have caused direct flow impingement on the inside elbow intrados and to have established secondary flow paths in the elbow. causing even higher turbulent flow velocities. Industry practice typically limits bulk flow velocities in single phase (wateri systems to less than 10 ft/sec. However, flow velocities as high as 50 ft/see were reported in some piping systems in foreign facilities with no apparent detrimental effects. Thus, bulk flow velocity alone in sinole phase fwaterl systems does not appear to be the dominant factor but can play an important role in erosion-corrosion in the presence of other interactive factors.
The use of a 45-degree lateral fitting would have reduced the direct flow impingement effects and minimized the flow turbulence in the elbow.
The pipino design stresses in the elbow due to pressure, dead weight, and thermal expansion were low and it is not believed that they contributed tn the pipe wall degradation.
'I
-m-m.___._
___--~__________.___--_.___.___-m._m___--_-
m.-__m
l
(
s i
i i
1 TheLimplications of the Surry-2 pipe failure were discussed with respect to the proposed broad' scope rule on the. " leak-before-break" (LRP) concept.
The surrv-2 mode of failure was initiated by high membrane stress, resulting from wall thinnino and was similar to that found in piping pressure burst tests where the break generally occurs almost instantaneously. In the LRB concept, the break is assumed to initiate in.
I a crack in the pipe wall and with a high degree of probability, the failure of.the piping pressure boundary will be signaled by a detectable
~
' leak providing ample time to shutdown and repair the leak. However, the "no-crack" type of break can occur in piping systems for reasons such as 11 overpressure 2) reduced wall thickness, and 31 overtemperature as reported in NUREG/CP-4305, " Comments.on the Leak-Before-Rreak Concept'for-Muclear Power Plant Pipino Systems," by E.C. Rodabauch ' August 14R51
't was noted that the feedwater pipinq system where the Surry-? failure occurred, would not have been a candidate for the LRR concept because, as discussed in the broad scope rule and in the proposed new SRP section many factors which would have been used in the LRR analysis are not predictable-
' fe.g. fatigue, stress corrosion, water-hammer, material traceability).
The phenomena associated with piping wear due to erosion, erosion-corrosion or cavitation have not been recognized as.significant problern by piping designers. Consequently, they do little to accommodate piping wear due to these phenomena in design. Some architect-engineers provide guidance in system design specifications to minimize erosion-corrosion such as limiting maximum bulk flow velocity in water filled systems (e.g. feedwater and service water 1
.However, most architect-engineers focus primarily on meeting code minimum pipe wall.
requirements with some variable additional wall thickness added, and on meeting stipulated pressure drop requirements for equipment operability.
P.
Fluid Dynamics (presented by P. Griffith, R. Keck - see Attachment ?)
Oxide dissolution is believed to play a key role in the mechanism of '
erosion-corrosion wear and is hichly interactive with flow velocity. An increase in flow velocity generally. tends to increase erosion-corrosion rate in carbon steel piping although the effect is more pronounced in two-phase *Jow conditions. Experimental tests have further estabitsbed that locas flow velocities in an elbow can he 2-3 times higher than bulk flow velocities.
Pased on the system operating temperature and pressure at the time of the pipe failure (380'F/367 psi), the conditions for cavitation to exist in the elbow was not likely. However, cavitation-erosion cannot be completely excluded as a contributory factor under different operatino conditions which may have existed during reduced flow or single-pump opera tion.
The temperature of the feedwater at Surry-2 (380*F) may have enhanced the rate of erosion-corrosion in the carbon steel pipe elbow.
1 l
_ _ _. 1 The temperature effects of erosion-corrosion on carbon steel are greatest
.in the 750"-340"F range. Below 750*F and above 340'F, erosion-corrosion wear rates decrease rapidly (see Attachment 4).
3.
Piping faterial (presented by S. Bush - see Attachment 3)
(
Carbon steel can be vulnerable to erosion-corrosion when certain un-favorable conditions are present. However, by increasino the alloy content (e.g. chromium, molybdenum, copper), its resistance to erosion-corrosion can he improved significantly. Field experience has shown that the use of Pi Cr-1 Mo steel improves piping resistance to erosion-corrosion by a factor of four. Chemical analyses of the failed pipe elbow from Surrv-? have disclosed unusually low amounts of alloying elemcnts-j particularly chromium (less than 0.0? percent 1 Austenitic stainless steel has been proven to be highly resistant to erosion-corrosion.
4 Water Chemistry (presented by O..lonas - see Attachment 41 l
l i
Water chemistry is believed to have been another important factor in.
causing the pipe wall thinning at Surry 7 The erosion-corrosinn wear rate of. carbon steel is create'.t when the pH levels are between 7 and 9 or below pH 5.
Little attack is found for pH values of 6 and 10. Erosion-
)
corrosion rates drop sharply at pH levels above 9.2.
The Surry-2 pH i
levels were reported to have been maintained between pH 8.8 and 9.2, l
however, local values could vary sionificantly.
It has also been noted that during the initial years of operation, poor control of water chemistry and condenser in-leakage may have contributed to the degraded feedwater piping in Surry-2 especial.ly at locations of high flow velocity. Subsequent to 1Q81, condensate polishinq units have been used at Surry-2 to remove impurities from the condensate.
A preliminary findino by the Surry-2 licensee that extremely low oxygen content in its secondary side water contributed to the recent pipe
)
break has been questioned by some panel experts. Althouch it is known that oxygen content above 100 parts per billion (ppb) is beneficial for neutral water, since this improves repassivation of carbon and low alloy steels, many fossil plants and forcion nuclear plants have operated at extremely low 0Fygen Content With no evidence of piping erosion-Corrosion in the feedwater piping. At Surry-7, the oxygen content has been maintained at 1 ppb to minimite steam generator tube degradation.
Predictive and Preventive Measures to Control Frosion Corrosion in Pipino Systems (presented by 5. Bush, P. Griffith,.R. Keck.- see Attachments 5,6, and 7)
- Althouch' erosion-corrosion is' not a new or unknown phenomenon, it has received relatively little study in the U.S. because incidents of recorded failure has been low and because the relationship between the variables influencing the process is complex. Erosion-corrosion failures in two phase systems and erosion in single phase svstems containing suspended solids were noted to be:
much more prevalent. Consecuently. they have received more attention in operation cnd maintenance than erosion-corrosion in single phase systems.' Some utilities have procrams to control and mittoate damage in two-phase systems
.such as steam extraction lines and turbine wet steam piping and in single phase systems containing suspended solids such as service water piping. It is possible.that sinole phase system erosinn-corrosion failures which did not result in personnel injury may have been experienced in non-nuclear applications without records being made of their failure and repair.
Erosion-corrosion or erosion in service water systems would not likely result ~
-in a catastrophic failure due to the low operating temperature and pressure of the svstem.
The use of zero degree ultrasonic beam technioue is accurate for measuring pipe wall thickness. CLrrently, the ASME Code and ANST P31.1 Code do not reautre any inservice inspections specifically for measuring pipe wall thickness. Althouch.
not recuired to do so, some utilities had elected to perform wall thickness
)
measuremente routinely in piping systens where erosion-corrosion due to wet steam or raw water impurities had been found to cause severe pipe wall.
degradation. From a volumetric context, the ASME. Code Section XI requires examinations in Code Class l'and ? piping systems, but examination is limited to the weld area and heat-affected zone using 45 and 60 degree she.ar wave techniques. There are no inservice volumetric requirements.for ASME Code Class 3 pipino nor any inservice inspection requirements for ANST R31.1 piping i
i
- systems, j
The ASME Code Section XI is currently developing rules for plant life extension, inservice inspection requirements beyond current weld examinations to account for the effects of erosion-corrosion on piping systems is a topic which should be addressed by the consnittee.
The effectiveness of ultrasonic measurements in piping could be enhanced by predictive methods which can locate potential areas of maximum wear. The Department of Mechanical Engineering of MIT (Cambridge, MA) has been developing a computer program to predict location and extent of wear in two-phase piping systems.
The methods are also applicable to single phase piping systems.
Using the MIT computer program, predictions were obtained for the Surry-?
feedwater piping system conditions. The results showed that the degree of pipe wall thinning was highly sensitive to pH levels. Changes in pH levels from R.4 4
to 8.4 could result in a variation of the remainino pipe wall thickness from 0.4 inches to 0.02 inches. (The initial elbow pipe wall thickness at Surry-?
was approximately 0.5 inches 1 I
I i
1
-5 In other studies, as reported in EPRI report NP-3944, " Erosion / Corrosion in Nuclear Plant Steam Piping: Causes and Inspection, Program Guidelines," April 1985, tbc influence of the flow path configuration for various fittinos has established a rance ci Mrical value) from 0.04 (least hannful flow occurring in straight pipes). to 1.0 (most harmfn* flow occuring in flow-splitting tees) to be used for calculatino erosion-co~ssion wear in two-phase (wet steam 1 piping systems. The Surry-? elbow-te6 configuration albeit not in a two-phase system would have been predicted as an arranoement potentially susceptible to severe erosion-corrosion wear based on the high empirical value of the elbow-tee configuration.
Conclusions and Recomendations by the Panel
- Althouah the effects of erosion-corrosion in wet steam piping systems (e.n.
extraction steam) and in raw water piping systems (e.g. service water) have been a recognized problem in several operating nuclear plants - and several utilities have established programs to monitor pipe wall degradation in these
-4 systems - the Surry-2 pipe failure is unusual because it occurred in a single phase (water) systen where erosion-corrosion had not been previously believed to be a concern.
The panel of industry experts concluded that no single parameter was responsible for causing the Surry-2 feedwater pipe failure. Rather the failure resulted Gom a synergistic effect of several adverse parameters occurring together. The panel concluded that the factors influencing the rate 1
at which the erosion-corrosion phenomenon proceeds in sinole phase systems, in j
general, cannot he ranked because the currently available quantitative data are j
insufficient. However, in the case of Surry-2, the panel believes that of these p4rameters,- the most contributory were 1) the elbow-tee piping confiouration
~
which caused locally high turbulent flow velocities, 21 the rance of pH levels j
maintained 3) the peor control of water chemistry especially during the initial years of operation and, 4) the carbon steel piping used also, to a lesser degree, 5) the system temperature, 61 the low level of alloyina elements' in the pipe elbow 7) the hiah bulk flow velocity, and 81 the lack of oxygen content in the water. All these parameters are believed to have contributed to establishina unfavorable conditions for erosion-corrosion to occur in the feedwater pipe elbow at Surry-2.
j Overall system changes such as an increase in system pH content or the use of a piping material with increased amounts of chromium for the feedwater system may have prevented the failure at Surry-2 However, a chance in one system parameter may have an unforeseen adverse effect on a different aspect of the system. The panel believes that ad.iustments to pH or oxygen content from that now in use for steam generator performance should not be made without a thorough study of the possible global effects of such changes on the entire system. Changes are sometimes made to fix a specific problem in a nuclear j
plant without considering the total remification of the fix. The addition of hydrogen to RWR systems was cited as an example of another proposed fix which requires careful study. Long term water chemistry needs for nuclear plants,
)
particularly, the effects of pH and the impact of oxidizing or reducino environments on materials of system components was cited as a possible topic for additional research. Thus, extreme care and forethought is warranted prior to initiating changes which 3
i
- f; -
i have an overall system impact. In this light, the advantages of a program which integrates predictive analysis, periodic monitoring of pipe wall degradation, and replacement (when necessary1 becomes a reasonable and viable alternative to drastic system changes.
l Since the Surry-P pipe failure, severa.
scensees at operatino ruclear plants l
have conducted pipe wall measurements in their fcedwater piping systems. It is expected that the results of these tests will provide an initial " quick-look" assessment to determine whether erosion-corrosion is a problem endemic to the feedwater system and warrants a reconsideration of certain system parameters.
i The panel believes that the infonnation which now is or will soon be available relative to the identity of the involved parameters and how they influence erosion-corrosion in single phase systems thould be presented to other utilities se that they may formulate programs to ultrasonically measure existing pipe wall thickness at susceptible locations in specific svstems.
Screening criteria such as those developed by Virginia Power should be used on the most susceptible systems to identify potential locations of maximum wear.
The need for periodic monitoring could be determined from the results of such baseline measurements, in the long term, a reconenendation by the panel identified a need to 1) develop a uniform set of screening criteria to establish a comprehensive data base of erosion-corrosion parameters in pipino systems and 2) develop generic guidelines to ensure the use of consistent techniques to predict and control erosion-corrosion in piping systems. The generic quidelines would include establishing critical erosion-corrosion parameters, predicting and measuring pipe wall degradation at potential areas of high wear rate, and evaluating and documenting the results of erosion-corrosion wear.
I I
I
)
l
l
'N I
Attendence List,for.lanuary 15, 1987 Meeting Name' DrqaniEation NRC/RegTo H J j
s
.>1^
W.1. Ross-l E.B. Rranch F
,U Sargent & i.mey A.R. Herdt N2C/ Region il B.P. Crowley NRC/ Region U D.F.. Landers Teledyne Engineering Services Rooer Woodruff NRC/IE E.R. Johnson Westinghouse F.1F tric Corp.
i Robert Schueler National Board IColumbus, OH)
E.J. Brown NRC/AE00
'Novak Zuber NPC/RES D.L. Basdekas-NRC/RFS L';
'?
Georae J. Brazill INPO i
Ca rl 'W." Hi rs t Westinghouse Electric Clinton Wolfe
?'
Westinghouse Flectric Larry Miller Virginia power L. Connor..
DSA 4
G.L.Parra)1 Virginia Power M. Tahmtodi virointa State Corp. Com.
W.S. Ha ulton NRC/NRR W. Ruberry Richmond Times-Dispatch s
P. Griffith MIT (Cambridge, MA)
R. Keck MIT P. Miller Virginia Power Frank Walters Pahcock & Wilcox S Paul Cortland
\\
NPC '
t W. Shack i
Argonne National Laboratory T.F. Kassner Argonne Mathnal Laboratory L.S. Gifford Cereral Electric Rindi Chexal EMd W.P. Mikesell Robert L. Cloud Associates / Chairmen of ASME BAPV Comittee Otakar Jonas Jonas, Inc.
T.S. Lubnon MPR AssocLetes M.D. Beaumont Westinchouse (Bettesdai M.A. Schopoman Fledda Power A Lkht Co.
Milton Vacins
%C/RES
,1 Heikki o ponen EC/fMR e
Yoichi Toco PE/AE00 Tetsu Imai locan Electric Power Info. Center H.M. Fontecilla Virginia Power R.W. Eaker Duke Power David W. Smith Duke Power.
E. Marshall Weaver M e Power Pobin ilones EPail Eve Fotopoulos SERCH Lice 7 sing, Rechtel Carl Cra.ikowski Brookhaven Pational Lab.
John P. Weeks Brookhaven National Lab.
3 Gregory Brown Stone A Webster Engineerina Corp.
Dan Donoahue Stone & 4ebster Engineering Coro.
l
(Attachment I continued)
Name Organfration Tdward F. Gerwin Chairman B31.1/ Pullman Power Products Michael Specter Washinoton Post June Lino ASME Brian Jordan McGraw-Hill J.E. Richardson NRC/RES Dan Guzy NRC/RES
'J. O' Brian NRC/RES H. Naninnen MEA Rudolph Behrens Pumptom. Inc.
Pete Eselgroth NRC/ Region i M. Hartzman NRC/NRR
.T. Sullivan NRC/NRR Cliff Henson Virginia Power Robert Baer NRC/IE Paul Cortland NRC/IE l
Tom Kennedy DOE J.P. Zimmer DOE A. Taboada NRC/RES
. lack Strosnider NRC/ Region !
D. Terao NRC/NRR R. Bosnak NRC/NRR E.C. Rodabauch ECR Associates S. Bush Review & Synthesis R. Johnson NRC/NRR 4
i I
.I
1 i
Af%.c 6 nt A OSre9es)
{
(
PREDICTION AND MITIGATION OF
^
EROSIVE-CORROSIVE WEAR IN STEAM EXTRACTION PIPING SYSTEMS PRESENTATION TO THE NUCLEAR REGULATORY COMMISSION JANUARY 15, 1986 BY RICHAnn 6. KECK, RESEARCH ASSISTANT PETER GRIFFITH, PROFESSOR DEPARTMENT OF MECHANICAL ENGINEERING MASSACHUSETTS INSTITUTE OF TECHNOLOGY CAMBRIDGE, MA, 02139 (617) 253-2248 i
RESEARCH GOAL:
FREDICT LOCATION AND EXTENT OF WEAR IN EXTRACTION PIPING; RECOMMEND SOLUTIONS TO HIGH-WEAR PROBLEMS.
OUTLINE OF PRESENTATI0M:
1 INTRODUCTION AND PROBLEM DEFINITION 11 DESCRIPTION OF THE Two WEAR MECHANISMS 111 WEAR LOCATION GUIC* LINES IV.
EXAMPLES OF WEAR PR. DICTION AND MITIGATION
{
V.
CONCLUSIONS AND RECOMMENDATIONS 4
I
i 1
1 PROBLEM DEFINITION
{
s EXTRACTION PIPING SYSTEM CHARACTERISTICS-t i
k P = 5-100 PSIA T = 150-450 'F X = 85-95%
DtA = 12-36" WALL = 3/8-5/8" MASS FLOW = 105 LB/NR V, = 75-175 FPS PH = 7-9
[0 ) < 20 PPB
{
2 l
FUNDAMENTAL IDEAS:
k
- PIPE CORRODES, FORMS PROTECT!VE LAYER OF IRON OXIDE (FE 0 ).
34
- CORROS!ON RATE INVERSELY RELATED TO OXIDE TNICKNESS.
I
- LIQUID IN LINE REMOVES PROTECTIVE OXIDE LAYER.
- MORE STEEL CORRODES, FORMS OXIDE, IS REMOVED, FORMS MORE, ETC.
- COMBINED EFFECT OF CHEMICAL AND MECHANICAL WEAR WORSE THA EITHER MECHANISM ACTING ALONE.
i WEAR CHARACTERISTICS:
- MAXtMUM WEAR RATE = 1 MM/YR
=>
10 YEARS a 1/4"
- WORST IN BENDS, TEES AND FITTINGS 5"
- LOCALIZED, NON-UNIFORM, RANDOM
- UNUSUAL TEMPERATURE DEPENDENCE:
f
- WEAR IS STEADY-STATE.
I~
~
\\
ko
.No b [*F]
i i
[
t 1
I i
)
l 1
EAR MECHANISMS
]
1 OXIDE DISSOLUTION:
i
- MAGNETITE HAS FINITE SOLUBILITY IN HIGH TEMPERATURE WATER.
- WATER HAS LOW CONCENTRATION OF DISSOLVED IRON.
- Flow 0F WATER =>. CONVECTIVE MASS TRANSPORT OF DISSOLVED IRON.
- VALID FOR SINGLE AND TWO-PHASE FLOWS.
. i
_y
_7 WAI[R QF
/F
.. e,...... :...f. f2*....i... :...t* f 1* f
- 'i.,;;. '. i.:g..?.SS.'.;;. '.%%...'::((.1.::::..'::'.?.i ?:iM
\\\\\\\\\\\\\\\\\\\\\\\\\\
METAL SANCHEZ (1984) DISSOLUTION WEAR MODEL:
di'=
Ce 0 1+(L+
)
K D
d
- TEMPERATURE EFFECTS IN TERMS:
(1) K (2) CE (3) 0
- MATCHES CHARACTERISTIC TEMPERATURE CURVE (USING POROSITY REDUCTIO
/
I i
.-......am.
_d
)(EAR MECHANISMS
- 2. DROPLET IMPACT WEAR:
- PURELY MECHANICAL WEAR MECHANISM.
- LIQUID DRUPS BECOME ENTRAINED IN THE CORE VAPOR FLOW.
IN A BEND, DROPS DON'T TURN CORNER, IMPACT ON OUTER BEND
- REPEATED IMPACTS CAUSES FATIGUE DAMAGE AND WEAR.
- VALID ONLY FOR TWO-PHASE FLOWS.
ox!x e
i :':..'..-l 9:...1 :.<
e-+ e-+ ).:.:n a
jg
'l. ;;'l:l u outo o s s
'}'t,$ ' A SANCHEZ (1984) DROPLET IMPACT WEAR MODEL:
4 th= Cgrh(1 - x) F F V h
(PE )2 A c
c 1
- DOES NOT MATCH CHARACTERISTIC TEMPERATURE CURVE.
1
- VERY STRONG DEPENDENCE ON THE DROPLET VELOCITY => 4TH POWER.
- SENSITIVE TO THE OXIDE MATERIAL PROPERTIES => 2ND POWER.
REAR LOCATION GUIDELINES:
A DISSOLUTION KE&&:
- 1. INSIDE OF BENDS:
(A)-SECONDARY FLOWS: PAIR OF COUNTER-ROTATING HELICAL VORTICES.
I TRANSPORT LIQUID FROM OUTSIDE TO INSIDE BEND RADIUS.
PRESENT IN BOTH SINGLE-AND TWO-PHASE FLOWS.
MAY CAUSE " TIGER S~nRIPING" IN EXTRACTION PIPING.
(B) SEPARATION AND REATTACHMENT: LEADS TO ENHANCED TURBULENCE AND INCREASED MASS TRANSFER COEFFICIENT.
4 WORST POINT: BETWEEN 90-110' AROUND THE BEND.
- 2. WELD-WAKE REGIONS:
UNIFORM WEAR REGION DOWNSTREAM OF A CIRCUMFERENTIAL WELD.
CAUSED BY LIQUID SEPARATION AND ENHANCED TURBULENCE.
TWO-PHASE EXPERIMENTS AT MIT SHOW 3X GREATER MASS TRANSFER.
IDEAL LOCATION FOR WEAR MONITORING WITH UT.
- 3. WATERLINE REGIONS:
i CONTINUOUS WEAR REGION, GENERALLY FOUND ON BOTTOM OF P!PE.
i 1
NON-ENTRAINED LIQUID IN LOCALLY STRATIFIED (TWO-PHASE) FLOW, CAUSES OX1DE DISSOLUTION IN STRAIGHT SECTIONS AFTER BENDS.
i DROPLET IMPACT ME&g:
j
- 1. OUTSIDE OF BENDS:
DUE TO LARGER MOMENTUM, DROPS IMPACT ON OUTER RADIUS OF BENDS.
WEAR REGIONS ARE ELLIPTICAL, WITH "DI AMETER" UNDER TWO INCHES.
EXACT LOCATION VARIES WITH DROP VELOCITY, DROP SIZE, BEND RADIUS.
SMALL-SCALE LAB EXT >ERIMENTS SHOW WORST POINT 70-90' AROUND BEND.
RANDOM LOCATION MAKES UT VERY DIFFICULT.
l.
I l
l-l l
1 1
l i
l CROSS-0VER CURVE FOR TYPICAL SYSTEM AT 300 F
-1.5.
Dissolution With PH = 9.0 1
-2.0 -
Droplet Impact log (Wear)
Wear Dissolution With 2.5 pH = 9.5
-3.0 =
1 1
l g /
g
[g 100 125 150 175 200 VELOCITY (fps)
l l
CROSS-0VER CUAVE FOR TYPICAL SYSTEM AT 353 F l
l
-1.5 i
-2,0 Dissolution With pH = 9.0 log (Wear)
Droplet Impact
-2.5 Wea Dissolution With pH= 9.
-3.0 i
m i
,/
100 125 150 175 200 VELOCITY (fps)
l COMPUTER PROGRAM
- 1. INPUT VARIABLES:
1 i
(1)
TF:
TEMPERATURE (*F)
(2)
MD0i:
MASS FLOW RATE (LBM/HR) i (3)
X:
QUALITY (UNITLESS)
(4)
DIA:
PIPE DIAMETER (FT) i (5)
PH:
SYSTEM PH AT 25 'C (PH UNITS)
(6)
CR, MO, 00:
WT% OF CHROME, MOLYBDENUM, AND COPPER IN P!PE.'
I 1
- 2. OUTPUT QUANTITIES:
(1)
V:
SUPERFICIAL 6AS VELOCITY (FPS) g (2)
F,:
EN',RAINMENT FRACTION (UNITLESS)
(3)
DDELDT:
DROPLET IMPACT WEAR PREDICTION (INCHES /YR)
/(4)
DMDT:
DISSOLUTION WEAR PREDICTION (BASE CASE) (INCHES /Y (4A) DMDT1:
DISSOLur!ON NEAR CIRCUMFERENTIAL WELD (48) DMDT2:
DISSOLUTION NEAR WELD DROP (4C) DMDT3:
DISSOLUTION IN WATERLINE
INPUT hl COMPUTE Ftuto PROPERTIES A
\\/
ExeTS V
A Da0PLET IMPACT NEAR DISSOLUTION WEAR 1
V M LEMENT YES lilGH WEAR RATE?
HITIGATION STRATEG1ES V
OUTPUT i
d Temperature
- 350*F.
Pressure
.450 psia l
Mass Flow Rate 13,000 gpm = 5,787,000 lbm/hr-System pH 8.9 9 25'C Pipe Diameter 18" nominal Pipe Wall 1/2" nominal Pipe Steel A106B and A234 (1030 low-carbon steel)
Total Run Time 76085 hrs.
Table 1: System Conditions for Surry Unit 2 Feedwater Piping 2
of "b#h O.4- -
WRtL D/cW4M 3
Cacus)
O. 2.
- 0. I o.o 8.4 gf gb
$7 gg B.9 pNf2f'C )
~#
~
~ 129_.?M MMf/WM. ll2.' WAU f.Lbtd7~ RdV-7/ML* 4' ?@W*UC1-9
l f
MITIGATION STRATEGIES i
(1)
ALLOYING THE PIPE STEEL:
- THEORIES UNCERTAIN OF EXACT PROTECTIVE MECHANISM OPERATING.
- TO IMPLEMENT, REPLACE EXISTING PIPE WITH LOW I ALLOYED STEELS.
STUDIES SHOW WEAR REDUCED BY 100X WITH 21 CR, 1% MO t 1% Cu.
(2)
OXY 6EN INJECTION:
- LEADS TO FORMATION OF STRONGER, LESS SOLUBLE OXIDE (HEMATITE).
- TO IMPLEMENT, REQUIRES CONTINUOUS MONITORING IN EACH LINE.
- CAN REDUCE WEAR BY FACTOR OF 3-10X.
l (3)- MOISTURE REMOVAL:
- WATER CAUSES THE WEAR => REMOVING IT'ALL WILL HALT THE WEAR.
- TO IMPLEMENT, USE MOISTURE SEPARATORS AT EARLY EXTRACTION LINE.
- VERY EFFECTIVE FOR DROP MODEL; NOT WELL CAPTURED IN DISSOLUTION (4)
INCREASE SYSTEM PH:
- SOLUBILITY OF MAGNETITE HIGHLY PH DEPENDENT.
)
- TO IMPLEMENT, REQUIRES GOOD KNOWLEDGE OF PH IN EXTRACTION LINES.
- ONLY EFFECTIVE FOR DISSOLUTION WEAR MECHANISM.
(5)
DECREASE SYSTEM VELOCITY:
- DECREASE MOMENTUM OF IMPACTING DROPS.
- TO IMPLEMENT, LARGER PIPE DIAMETER OR PLANT RE-DESIGN.
- VERY EFFECTIVE ONLY FOR PROPLET IMPACT WEAR MECHANISM.
i i
EXAMPLES'0F MlTIGATION STRATEGIES FROM WEAR COMPUTER PR CASE
' DISSOLUTION WEAR DROPLET IMPACT WEAR WEAR RATE
% FROM BASE WEAR RATE
% FROM BASE BASE CASE 0.013 IN/YR.
N/A 0.048 IN/YR N/A
' ALLOYING 0.000126 IN/YR
-99.0 0.00046 IN/YR
-99.0 i.
0.00189 IN/YR
-85.4 0.017 IN/YR
-64.6 i
RE.
0.013 IN/YR 0.0 0.014 IN/YR
-70.8 RAl'SE SYSTEM PH 0.0013 IN/YR
-90.0 0.048 IN/YR 0.0 CR E
0.0092 IN/YR
-13.0 0.0018 IN/YR
-96.3 g
y KEY:
j (1) ALLOYING:
2% CR, 11 Mo, 1% CU (2) 0xYGEN INJECTION:
INCREASE FROM 20 PPB TO 200 PPB i
(3) MOISTURE REMOVAL:
INCREASE QUALITY FROM 93 TO 98%
l (4) RAISE SYSTEM PH:
INCREASE BY 2 UNITS (5) DECREASE VELOCITY:
DECREASED BY 33%
l l
{
CONCLUSI M M RECOMMENDATIONS:
WEAR MONITORIN6:
- USE FIXED REFERENCE POINTS IN EACH LINE TO MONITOR WEAR RATE.
- MEASURE INITIAL WALL THICKNESSES WHEN INSTALLING NEW PIPE.
- DISSOLUTION WEAR: WORST AT 90-110' AROUND INSIDE RADIUS OF BEND.
- TAKE ADVAN" AGE OF LOCALIZED DISSOLUTION WEAR LOCATIONS.
- DROPLET IMPACT WEAR: WORST AT 70-90' AROUND OUTSIDE RADIUS OF BEND WEAR PREDICTIONS:
- MODELS CORRECTLY IDENTIFY CONDITIONS WITH WORST WEAR.
- ERRORS IN PREDICTIONS:
OFF BY FACTOR OF 2-3
- USE MODEL RESULTS AS GUIDE TO WEAR LOCATION AND MITIGATION STRATEGI
- NEED TO MEASURE ACTUAL PH AND [0 3 IN EACH LINE.
2 WEAR MITIGATION:
- MilST DETERMINE DOMINANT WEAR MECHANISM BEFORE IMPLEMENTING STRATEG
{
- ALLOYING IS GENERALLY THE BEST LONG-TERM SOLUTION TO THE PROBLEM.
- OTHER STRATEGIES REQUIRE MONITORING, PLANT DESIGN CHANGES, ETC.
FUTURE WORK:
- OXYGEN CONTENT: I4EED EXPERIMENTS ON WEAR AS FUNCTION OF TEMPERATURE
- NEED TO SEPARATE LITERATURE RESULTS FOR DISSOLUTION VS. DROPLETS.
- WEAR MODELS:
DETERMINE AND CORRECT SOURCE OF ERRORS.
- DROPLET IMPACT WEAR:
MORE EXPERIMENTS NEEDED FOR BETTER UNDERSTAND!NI
- WORN PIPE EXAMINATION: NEED TO CATALOGUE OBSERVED WEAR PATTEPNS.
- WEAR LOCATION: LARGE SCALE EXPERIMENTS NEEDED FOR BETTER PREDICTIO
~
tj(yd;v "o
8 P
7 L
E E
1 E
R C
H R
A A
-A O
,R
)
T F
N HT
,E N
O P D
S E
T I
D E
N T
,N 3
N E
TEA E
A 8
A V
2 G
O E
I I N
R V
V TE A
A P
O A
C M
M R
C
,N I
P EE 6
E M
NRL 0
B F
I OUU 1
Y O
ITB C
A F
SAR I
A MS O
ORU T
S
(
R RET S
T S
O R
E P I
C A
R T
O M,
M E
E C
T FER I
F H
P A
C OTE T
F C
P F
A Y
P E
U O
F E
,A O
S C
D X
TEL L
E S4 AR E
A S
Rxxxx E
RUE R
I L
R E
U T
S1233>
IL GSI E
N A
GO NET S
t T
E E
B IRE R
i A
S D
M B
C N
G T
D Y
Y NF E
,A C
L L
L EO UYM A
I O
L I
LT T
F M
M A
o RO FIEN T
M EM NCHE E
E
, N O1 Pl IOTS S
N M
M-I X-L E
E E2 R o N I
U ER SE
,R H
L I
I/CM2 C
RVNP T
M R14l-D4 O
E E
O E/R L/
TWGS S
R PR1RC E1 COYN A
H XC C6 I
ALXO B
C E1291 F2 FFOI
?
5 2713 6
3' 0,
3,7 9 5, 0,92,8, 5
2 7961 0007.O06 1
40 1
Q 21 1
- 4. 2 2 9
7874 5920
- 0. 9 2 8,.
8 9,
4,7 8 2
5 4 8 0' 0 0,8 84 0007 1
1 5
1 1
12 9
6929 1
4 49,02 0,
- 0037808, 1
1 1
5 3
4113 1
450 1
1 2
8 2
4 87
- r.
9,45 8,
90, 0. 71,
- 0087004, 0
2 1
0 1 3 8
4 1
4 480 4
2 4
12 2
ga 9
7770 3
387 5
8,7,02 8,
7, 8,3,0 0, 5
1, 0,43 420
- 204, 9
3 1
9287 4
0 -
0047 8
.~
178 1
5 62 6
3 3
84 9
8 3
2 3
3 12 1,-
843 5
3539 87 4,
00001,1 1,2, 5
5, 1 3 3,0, 9
11, 8
0 4
0 0001 007880,0 8
3 1
3 420 8
1 N
12 1
3 4314 7
49 4
7, 10012,1,1,1, 8
8 8, 3, 2. 0 8.
2.-7, 5
[
7 3
1 3
1 0001 0066103 1
3 460 3
2 2
W.N-257 6
2 9325 1
n.
14332 7,
85448,4 9,8, 2,7,3 e
6 4
11536 5
004710,5 3
1 t
1 21 3
4 1
1 440 t
^
e 22 1
l AP W-46096 351 1 5 885 n
8 2 9 4,8 9 4,
0.9.6,9 5, 1,7 6 2, N^
16210,1 0059038 e
5 5
41100 4
t 422 5
t 1
~
e 24 1
d 1
n 86494 2
11 18 0,87 C
e 5
1 9 5,7,2 8
- 5.0.0,5 6 0 0 8 7 1 0,5 S-he 4
5 45421 8
06336 2
1 84 3
3 592 7
480 0
d b
84 1
12-8
.n-r 1
o u.
v 943 1
.a 1
th 2,
1.0 1,1 1, 0,*
5 0087503 c
3 0
00001 1
i 2
n
+
480 0
9 12 8
7 l
J e
b 11 76 78 3
0000 5,597 1, 3
'h 0
000 0079327 m
2 2
c
+
480 4
12 i
. 3 gp 3737 86 n
5 3,6.1
- 0. 3. 2 n
0 8532 -
002740 2 i
1
&6, 2
651 480 4
e 1 1 1 2 3
ta rs
,/-
. 9m0
.h g
c s
/s eC h A/c/
r i
k a
/m m
b*
gS 0 r
r u
t
.dh I
b
,p p p e
t 9
l
.n6 a
n i
e p8I H 0A-r id t
v v
p e
e b,
t t
M i
i e
e e
e s
k k
m h
g g
i i
h c
d 5
d 5
)
a E
s n
n C
i b
b
- ~ it Z
l i
i r.
w 0 N w
0 i
ug N
r t
r 0*
e N
h 4
o h
4 c
1 1
t 2 e
W u i (l p
c f
49 49 s
M c
f t
M k
t t.
t e egh eo s
f S
h e
o 2.3 o o u e
o
- 2. 3 oo u r
s f
a zt t r t27 MMC et ehiehg uh r
1 u
G s 2.7 oMMC G
c t
z s
e k73or k p a WGt c
Sn ec k73 r rI rI e3tMCCN r3tMCCN l
s t e di ums-r e
e r
n e0 e i H ri s b9 e S 5305 r
a V
tS f
e W3R1111 DTEpOlE WSR11115305 si 1
T e
L_
.; } * ~ ~
i
,t,.
g 4 tr..
I 4
1
\\
5 M
s 1 }.
u f
.O g.
k3; a
-aik g
,n asco PIi s
1 r
n; CCC9
,e
,' j So g
x s.
m V.
\\
/
- 'fvi'=
e%
w9
\\
4
/
s 1
.o
\\
g*/ "/ /
- n. 9 E S. R
\\
e E
~
=
(/
fiM C jQ*i i
t
..w-(-x%
k i
c
(
s w n.
A x
is N
s S
is4 1
N D**'* T','S 9$
3 3"
i i
.) E gg
$ *4.e, a "S
Wg 4
a.
o s
e.
4-===
m {#O cBE
.o
,f
/
t
1 i
i l
j i
n'
(
)
att
-l M
5 a
.! t ES\\'- J
.W\\
4 1 of aJ A
i a
.m
.m 3 R.i e;
8(
\\
b-1 3k\\-
i*l
"\\h T n.
a E
j
'N \\\\
! ;i s \\\\
l 5
\\)Rf l
R 3
.\\.
1 s
N s
- 8... m..
. \\
s s\\se \\
a 3
N Es 4
3 j
3 Ji JLe s
s n=
3 h
em
\\vI l6 4
-Jt'v g
,pa1 4
%.%['
b i
i 1
1 Wear Aete i
) ast n.2 t'L.
r..w, ca h
- - rn e s
8 - me n i
pH = g
- Ha 30. _._ 0,<s,g, s
a <*pum y
'N##3I 4"
(G
/
5 e
p* f
}
O I
l 2
-O-*~~~~~~~~~
I u cau. w l
-g,,8 % Nb5
"~~~~~~
1 a;6 i
O=. ggg 0.5 0.2,
O
- i n
Rg. 7
-C5
&. * *tial Wear Due to E
/
- +o u q s.,2 0.c
,"/ % /$
)9 // c,9, G,, s,p e
/
c
~
J
1
,/ (. % /vo 4.
Specific Material Wear Rate a 1000 f
{
500 si-N l
cal h--
p-40bar 15Mo3 g_ yg g
y-39m/s 0 -Content 55pg/kg 50 w
.L ma.~N Acc#J.ii 10 to Resck
- -75*C
e 5
v-1.5m/s 02-Content =20 g/kg g
[ ~ 2 'l b%
\\e.
t
)
si35.s' - I{
05
- ,l 13CrMo44-4 01 i
O-c1even a vai..
'6 7
5 9
to 11 Fig.8 H
- P Yaim Eliect of pH Value on Material Wear Due to Erosion-Corrosion I
c,,
2 i
Speede WiedalWew Rate i
1000 a
500t}
M
,00 E
24 bY P
~
4
%eP Acconbag y-Sa35 8 gg m-
- =75*C k
._ v - 1,5 m/ s i
(
pH-7 i
I L
1 p-40bar
- 8 l
1 1
0-120*C -
\\
'-13CrWo40 v = 35m/s
-- h 'I -.....___._d=.2'.hr C*)
pH i
0.5 i5u.3 i
- g-T O - C
.<s.4 vano.
i 0
100 260 360 400 pghg 500
- 0'**
- Rg.9 Effect of Oxygen Content on Matetial Wear Due to Erosion-Conosion 9
f 4
a..
L I
o l
4 specific Meiwialwon state
-Soo l
- 9 em3h ano M Untrested Specumens(Ground) l E!ZD Sp==neas with Prosectin t,yw sadwed in wenn at 200 b r.sorc M Specunens with Protectin Layw l
b?"'d h 53am at h,45W 300 i
l 200 - ---
.=
\\
100-0-
- ~
3"3I#
" "* 3 s scum 2 5 u oMe 44 Fig.10 Resistance of Magnetite Protective Layers to Erosion-Corrosion I
p -40 bar.4= 180% v-20 m/s. Neuttal Mode (pH = 7.O Coateat < 5p g/kg) f
)to hw i
e i
'i
St37.2 + 500pn; i Metco 33 Layer 5t 37.2 + 500pm l MLetw
, ummanus X100 Niil189 X20Cel3 0-X8C NiMo12 10C MoS10 f.
GS 16CelAo910 i
s 13C:Mo44 i"
,g. % ' o, c.,,,.
30CrMoNIV411 g
i
< s,p, j
O s.s
<seth 26Nic MoV115 0
F mees 15NicuMoNb5 l
15Mo3 umnummumusimmusummuumusummuuuus i
RSt37.2
)
umummuuuuu muumusumuunum Material V10 50 100 150pg/cm1h 200
* Specific Metodel Wear Rate plg,4 Material Wear Due to Erosion-Corrosion of Various Materials p-40 bar,d-180 C,v=20 m/s l
e.
Awk~aW
,7 l
wo Control erosion / corrosion 1
of steels in wet steam The combined phenomenon reduces the integrity of piping and major cycle components, and is a prime contributor to sludge and crud buildup in both drum boilers and nuclear steam systems By Otakar Jonas, P.E., Jonas Inc Erosion / corrosion of carbon steel thickness in several pressurized water and corrosion, with corrosive action initi-exposed to wet steam.in both reactors (PWRs)-to perforation, in at sted by crosion of the protective metal-nuclear and fossil-fueled power-least one case-and has been responsible oxide layer from a metal surface. With-plants has long been a matter of concern. for partial or complete disappearance of out this protective layer-mostly mag-It is most pronounced in carbon steel moisture-separator chevrons after sever. netite (Fe,0.)-carbon steel is vulner-components with high velocity, turbulent al years of operation. It is identified with able to general corrosion / dissolution, flow of low pH moisture containing a the accumulation of large quantitics of even in weakly acidic water. The acids high concentration of CO, (or other oxides in boiling water reactors (BWRs) most likely to be present are carbonic, seid-forming anions)." The phenome-and PWR steam generators, causing short chain organic-acetic, formic, non is also associated with high-velocity additional corrosion and contamination ete-and possibly hydrochloric, hydro-water under similar hydraulic and chem-problems. Erosion / corrosion is also fre-fluoric, and sulfuric.
ical conditions.
quently observed in feedwater heaters, Carbonic acid forms from reaction of Known for decades to exist in con. vertical deserstors, BWR steam separa-water with con. The latter enters the dens 6te return lines, erosion / corrosion is tors, and wet termns of steam turbines-steam cycle with inleaking air, or results now recognized as the cause of a new set The term cru.un/ corrosion is applied from decomposition of carbonate and of problems. It has reduced pipe-wall to the interaction of mechanical wear organic impurities. Organic (and certain inorganic) acids can be generated by decomposition of organic impurines, cbclants and polymers used in boiler-t /2Mo siw i
6ous materiale from water treatment, and i n exchange res-8m muumuus erosion /corroeien ans.
in 356F water mov-
$5-
.~::e of a moisture film I
Caroon sid' I
ing.: es.g esf ec, depends on its velocity, turbulence, and
)
Certon siw i
500 M, h h
+ 600,m i
for three typical 3ono Metco 33 i
pH/ oxygen combl*
CarM steel Carbon steel I
+ 600,m ruchel I
1000-l f*8 0, conteni yfpuo (416 t 18Cr sanwuess MY soo,oike
~
0808 HI I
m 96 e 6 ag/k0
~
O g
m?
e 6 80/k0 13Cr stairness 100 l
8/88' l
2 1/ dC#-
1 t
aaru o
.n.
ia is i o,eo. T i 30-I tuo siews sA4i4 o,eos e
- 3. Temperatears of. $
- Cr.r/s Mo
$s$,$?g
~
fect on ero-
,o_
LAtt3 Graar 712) l l
IC# te2Mo stee*
I sion/ corrosion is W
Greatest in 286 I
366F ran0s. Cond6 3
t f/4Cr Mo Ct Mo M V steel I
tions: 580 pelo.115 (Af f3 Grece 722)
?t/6ec,
- pHm7, m c, wo v sisei I
0e80 ao'ho. *x-i i
l posure time =200 I
m e
B
- I E'
O.3 4 4(u Mo-ND steel I oa a
ao a ',o
,;o
- c 2.
'22
.2 r spec.c m.ie,.s.ee,,.ie.,0/c,i,.-n, P i2,,,y,,,,,u,, g2,
q Power. March,seS
.l chemical composition. For each n aterial fron oxides removed by erosion are applied to units having condensate poi.
and temperature, there seems ta tw a transported and deposited in various ishers incorporating cation resins oper-critical velocity below which ero-parts of the cycle. In PWRs. moisture-ated in the ammonia form.
sion/ corrosion is negligible. At higher esperator drains are pumped forward to Occurrence of crosion/ corrosion can velocities, damage occurs in regions of the steam generators? As a result, much be detected by monitoring pipe wall turbulence-such as tube ends tn heat of the iron oxides prcduced and released thickness-either manually or automati-exchangers, now obstacles pretent in reach the steam generators, eere they cally,with the unit on-line-and by peri-pipes, and points where the flow t hanges usually deposit on the lowet.sbesheets odic inspection of moisture separator direction (such as at moisture-separator and form crud and sludge.
reheaters and other components. These chevrons). Under the indicated condi-
&luble impurities, such at: chlorides, techniques can be supplemented by tions, the rate of wear depends on the caustic, and phosphate, can con:entrate esamination of chemical data and review material's composition-with th's pres-in the sludge pile, often resulting in of steam generator studr,e accumulation ence of chromium providing a beneficial stress corrosion, thinning, and pitting of history. A cycle chemical transport effect-as well as temperature. flow steam-generator tubes.
study with a material balance can reh-velocity, and chemical compositio n of the Transport of these oxides into PWR ably identify locations where the rate of wster in contact with the surface steam generators and boiling water reac-erosion / corrosion is excessive.
iavolved; oxygen concentration and pH tors can be reduced by magnetic filtra-Ed6ted by sheldon D strause also have pronounced effects.
tion of moisture separator drains or feed.
n,,,,,,,,,
Material wear rates are shown in Fig water, or by rerouting the drains to the 1 H e Huammm ene p aimub. inm a for nine stects and Iwo plated carbon condenser. Oxides prodaced by ero-een=e een a men pH venue in ves experisace esconowy steels. Water moving at 65.6 ft/sec, with sion/ corrosion of piping can also lead to E *,', T,,,,,',"cE' 366F temperature and three typical solid particle erosion (cutting) of bah wesw enso mry m. L_--- -
1 swano nce pH/ oxygen combinations, was used.8 turbine control-stage blades and valv-aw y,sca,seur. aremon, pg The highest erosion / corrosion rates were ing.
Wes4remseressmanna, vos.
observed on,0.50 molybdenum and car-Suffiolent det and knowledge are
,,8) "*4,8, g Ln.nm,,,.,n,o,,es bon steels in neutral water voth low now available to mitigate these prob-oxygen content. For many environment lems. The following remedies may be g g/Aose gg=e,i*
combinations, the highest rates occurred applied:
m,,,,,,,,,,,,,,,,,,,,,,,,,,,,
at temperatures between 266F and 366F.
e increasing the moisture pH.
assron isso as seen in Fig. 2. Moreover, the effect e Reducing flow velocity below the 4 d,***",",*",,8,',*",**',*",M'*"*""*'"**'"-
increases exponentially with flow veloct-critical level.
manwassurme, nar/moane ermposann tese ty (Fig 3). Under constant flow velocity e Changing the material of construc-g8Q and chemistry, the rate is alniost con-t,on from carbon steel to a steel with i
stant with tirne. (Note: All figores are higher chromium content.
,eso n,nnn n, aneroy aacmi, waner casemary sh from Reference 2.)
a 1.owering the moisture content of
@ m m. aseme of enmientry en aroman.
Independent of oxygen content, ero-wet steam, estremen as amm m mener and wet seenm. isso sion/ corrosion rates drop sharply at pH All of these approaches have been f """* anorer ammeiy n=sene mee nee levels above 9.2 (Fig 4). To provide an tested in operating units. In ferrous 7 m evenese me o remer. cerraman a pe.er.
operating margin, therefore, a pH of 9.6 PWR systems using AVT chemistry, g g y a8# P"aum or higher is recomrnended for all volatile increasing pH to about 9.6 and control-a e 6 a,enmaso. J av rweses Jr. eid 7 o aireuo.
treatment (AVT) systems. An Os con-ling air inleakage to reduce CO repre.
pr inen w menar means m v wasm centration above 100 ppb is apphrently sents the easiest remedy. It offers the r d*" '""c"ons'e"e"Ier Ys beneficial for neutral water tince this added benefits of improved cortcoion a o Jonen meresse et ampurny sensentranane in improves the repassivation of carbon and protection of feedwater heaters, steam 7,,,g' M" Mien, low alloy steels (Fig 5).
generators, and turbines. It can also be p r9 6000 1000 1000 t/2 1000 f
~
(A 1,Mo 1 Ofece i1, 600 Y
- ,*ll ll p.6a0 pe carbon tar 6 26 Amo
,6' 100- =
t,00
]
h/ sec 1
1x-c';pp
/
>,0. "s ee.
,c,n-t t
<ws 9,
- ' rsa, ca.
a so-
- /e i
t g
1/?MO Ig3 9 eor Tff) {10-j
. 10 p
i 10-Carbon steer b~
g I/scrMo (A213 Greer 722) e d
1 - j,f,[
~ }
\\
l T *6rF l 'k FCr 1//W J1 e am;'
i i - w<
32 8 65 6 90 4 1312 ft/sec 01 j
C..am.e_ _.
a5 l0 a
a a
ec }
,a. <a,,, c, r,,
0.1-Flow velocHy. m/sec or ft/sec
~
L Ple1 sing water increases mrater6el-loss 6
8 9
10 11 0
t00 200 300 400 600
~~
rete exponentWty with flow woocity.
pH Oxygen content, so /k0 Conditions: 580 psig. 356F. p@ 7. 07 6
- 4. Decrosair g pH reduwe material wear,
- 5. Oxygen sentent above 100 ppb gives ag/kg. exposure time = 200 fv particutarly below OH = 9.2 maximum steel protectkm in neutral water power. heeren toes 103
b7
~\\
t NET F
P I
O 0
X 6
O E
T T EBU D
B A
N U
S A
R Q
A O
E E
5 D
O D
S4 3
A T
O L
C O
H S
E E
R T
S B
N N
T I
A O
I N W L
N L
D O
C O
L E
C D
I I2 E
L R
T W
R V
E O
A D
G W
F N
M N
E D
I A
A D
N M
S M
E A
O T
A B1 O
R N
X R
I F
E E
E S
E X
E S
Z T
E C
R A
E5 R
I G
L R
I R
E C
O S2 U
T D
F A
E Q
EO R
OV E
O S
F TA R
U R
F T
O W
L E
N D
C O
Z S
E I
ER I
V M
M I
TA R
C A
E IE T
I X
R MH E
N E
I IS O
U L
U S
G Q
E L
A N
E SE O
A I
R MR V
T P
P AG L
I O
XE O
U P
N ED N
1l l
L
uh er 1 &
An..
LIST OF THESIS ON CORR 0SION-EROSION j
Department of Mechanical Engineering Massachusetts Institute of Technology t
Cambridge,MA, 02139 Professor Feter Griffith (617) 253-2248 Berkow, J.M., "The Use of Novel Bend Geometries to Reduce or Eliminate Erosive-Corrosive Wear in Steam Extraction Lines " M.S. Thesis, June 1984 deFreitas, G., " Dissolution Rates in the Wake of a Welding Backup Ring," M.E.
Thesis, February,1986.
Dernbach, E.A., "An Experimental Study of Waterline Dissolution Rates," B.S.
Thesis, June 1986.
Gawlik, K.G., "The Determination of Wear Patterns in Pipes Conveying Wet Steam," B.S. Thesis, June,1985.
Keck, R.G., " Prediction and Mitigation of Erosive-Corrosive Wear in Steam Extraction Piping Systems," Ph.D. Thesis, February,1987 Kinnare, T.C.,
" Dissolution Rates in the Wake of a Weld Drop," B.S. Thesis, June 1986 Pack, B.G., " Prediction of Maximum Wear Location in Steam Extraction Lines of Power Plants," M.S. Thesis, February,1984 Sanc.hez-Caldera, L.E., "The Mechanism of Corrosion-Erosion in Steam Extraction Lines of Power Stations," Ph.D. Thesis, June, 1984 Vu, H.V., " Erosive-Corrosive Wear in Steam Extraction Lines of Power Plants,"
M.S. Thesis, June 1982.
l l
l l
l 1
(
I
[
t
.q l
p e / r. j. i
'7 '
(~". r -
)
MASSACHUSETTS INSTITUTE OF TECHNOLOGY Department of Mechanical Engineering Cambridge, Massachusetts 02139 I
l THE MECHANISM OF CORROSION-EROSION IN STEAM EXTRACTION LINES OF Ple>3R STATIONS SUBMITTED T0 THE TRANSACTIONS OF ASME i
LUIS EFRAIN SANCHEZ-CALDERA PETER GRIFFITH ERNEST RABINOWICZ l
l i
August 1986 l
Presently at Mixalloy Corp., 14 Brenc Dr., Hudson, MA 01749.
- Professors at the Mechanical Engineering Dept. of the Massachusetts Institute of Technology.
l t
.~
3
a 2
ABSTRACT Cerrosion-erosion occurs in steam extraction piping made of low. carbon steel that conveys wet steam. The rate of metal removal peaks at 150 C. and it is most sever 6 on the inside and outside of bends and in the vicinity.of fittings.
A theory is established by which three processes are shown to give rise to the observed peak in the metal removal rate: the oxidation
- reaction, the mass transfer coefficient which governs at 150 C, and the diffusion resistance of the oxide layer which governs at higher-temperatures.
The results of the derived model agree well with the experimental data in predicting wear rates and in establishing the temperature and the location of maximum material removal.
3
- 1. INTRODUCTION The specific problem that generated the work here reported has been described by Vu (1982) [1].
Vu's thesis showed wall thickness data for several bends taken from power stations.
These
- pipes, which experienced wear rates (material removal rates) of the order of one mm. per year (0.04 in/y), operated at
~ around 90 % quality and at 0.4 MPa (60 psi).
The most severe wear was sometimes on the inside of the bend and sometimes on the outside.
l Vu's' experimental work showed
- that, as a result of the l
secondary flow in an elbow (see Fig. 1), a water film flows in the inside of a bend.
The mass transfer to this film was found to be responsible for the removal of ferrous ions from the inside of a
bend.
Tnis phenomenon in many cases leads to more material loss than that caused on the outside of the bend by water droplet impact.
In order to understand the above process it is necessary to investigate the chemical aspects linkod to the phenomenon of corrosion-erosion in single-phase water.
Since the regions of high wear were those where the liquid film was found, this investigation has dealt only with single phase corrosion-erosion.
Other.cudies in the area of corrosion-erosion in two-phase flow have also been undertaken by the M.I.T. group [2,3].
a J
u tn yr e
a m
sd e
e k
n n t
a g
a or e ce n
r r
ip f
t et t
o ss
~
r a
e m
n a
p i
I Ie rw l
w e ow n
t t
n e
al o o
l o
ol i
l gh Wff p
i t
o c
eg a
r i
r D
Rh t
I xe w
="
m ae t
s a
b D
_~
fo wob le n
e wf a
o t
o n
~
a l
i f
r d
s f
n o
n w
o rb o r
i a
Ha d
b e
g e
nel r
w od e
m n
r o
ci a
l s n e
i e n f
i s
gh S. i o W
eg t
i e
r Rh e
l 1
po e
r r
D u
g iF
9 5
To understand the corrosion-erosion process, the specific goals of the research were to characterize the wear, under a common range of operating conditions, of a typical material from which steam extraction lines are made (SA 155 Gr.C) and to propose a method for estinating the wear rate as a function of
- velocity, temperature, and pH.
To achieve.these goals, a high temperature and high pressure loop was designed and constructed.
Its operation has aid the characterization of a specific material for different temperatures, 300 to 450 K (100 to 320 F) and different velocities up to 10 m/s (30 ft/s).
2.
Corrosion-Erosion Model Derivation This section gives a
background on corrosion-erosion and describes how the.model used to predict wear rates was derived.
2.1, Corrosion-Erosion Background Water conditions in a power station are such that the dissolved oxygen concentration is very low, of the order of 100 ppb.
Under these conditions, the final product during che oxidation of steel is a porous oxide layer of magnetite (Fe 0 )
J 34
)
(4),
The pores in the film were associated with the dissolution of iron at the metal oxide interface and its transport into
{
solution in the ferrous form.
I l
i j
6 Experiments with low carbon steel in water at a temperature above 150 C have shown that the first step in the oxidation of steel is the formation of the hydroxide of iron Fe(OH)
The 2
following reactions, which take place at the. metal-oxide interface, account for the generation ~of this hydroxide [5,6,7]:
2+
Component reactions.
Fe + Fe
+ 2e,
(1)
+
H 0 + 2II + 20H,
(2) 2
+
2H + 2 e- + H, (3) i 2+
Fe(OH),
(4)
Fe
+ 20H
+
2 Overall reaction:
Fe +
2H 0 + Fe(OH)
+H.
(5) 2 2
2 The hydroxide produced during the first step of the oxidation of iron can form magnetite through the Schikorr reaction [8],
3Fe(OH) + Fe 0
+H
+ 2H 0.
(6) 2 34 2
2 The rate of the above reaction increases greatly at temperatures close to 175 C (350 F) (7].
Important research was carried out by Sweeton and Baes (1970)
(9].
They showed that the oxide magnetite can lead to
++
+
the formation of the following type of compounds: Fe
, Fe(OH)
f i
1 7
-14 g
Fe(OH)
Fe(OH) which are soluble in water (10 Ig/IgH 0).
2 3
2 Their experimental data on the solubility of magnetite as a function of temperature and pH allowed them to compute thermodynsmic data (AS,4H) for-calculating the. equilibrium constant for the following reaction:
i
+
(2-b)+
Fe O + 3(2-b)H e H (g) + Fe(OH)
+ (4-3b)H 0 (7),
a 34 2
b 2
)
where b can have the values of 0, 1, 2, and 3.
The work of Sweeton and Baes is very useful, since it provides-a method for determining the concentration of the hydroxide species in equilibrium, and it suggests that these hydroxides of iron are the compounds by which iron is lost in a dynamic system (flowing water).
The net iron removal rate will depend on the concentration of soluble iron species at the oxide-water interface and on the physical-chemical characteristics of the water.
This is the phenomenon of corrosion-erosion of l
steel.-
2.2.
Recent Work on Cerrosion-Erosion In order to study the phenomenon of corrosion-erosion researchers have mede use of laboratcry rigs or have inserted their specimens inside the pipes of power plants.
A complete susmary of these investigations can be found in Ref. 10.
l 8
Some of the investigators have tried to obtain a model for estimating corrosion-erosion damage.
The model of Keller.(1974)
-[11]
is empirical ar.d the main disadvantage is that it does not include the influence of the chemistry of the water (pH, oxygen concentration).
The models of Bignold, et al. (1981) [12] and Berge (1982)
(13]
fail to provide values of the essential constants needed to estimate the wear from their relations.
One reason for.not having obtained a workable expression is the lack of a physically. realistic model that expresses clearly the influence of both the hydrodynamics and the kinetics of the chemical reactions insolved.
- This, in turn, is linked to the fact that experimental work has made use of very complicated geometries (bends, impinging plates, orifices).
In these cases the hydrodynamics by themselves are difficult to establish; and, t
it makes it even more difficult to determine the way in which the hydrodynamics influence the reaction kinetics.
The influence on the wear rate of
- velocity, oxygen,
- material, pH and temperature of the water are described in detail elsewhere
[14];
nevertheless, it is important to note that experimental data on corrosion-erosion show that wear rates follow a rectilinear let with time.
A constant wear rate implies that the oxide layer provides a constant resistance to
\\
corrosion by maintaining cor.stant thickness and porosity.
l
D:
9 2.3.-
Corrosion-Erosion Model Assumptions and Derivation As-described earlier, the removal of material by corrosive-erosive' wear is the result of several phenomena that include, the initial reaction of pure iron-with water and the final step of mass transport of hydroxides into the bulk flow of water.-
l The oxide layer, formed on the iron, plays an important role by restricting the flow of water to the metal and the flow of i
hydroxides byf diffusion from. the iron surface to the water i
flow; this in turn affects the hydroxide concentration at the i
iron-oxide and oxide-water interfaces.
l The various effects that take place in the neighborhood of the metal surface are shown in Fig. 2.-
I As can be-inferred from the
- figure, the process of corrosion-erosion includes complicated chemical. kinetics and fluid-rechanics.
It is difficult at'this stage of our research to. incorporate all the processes laws in the derivation of the model;
- instead, we looked for the minimum number of variables necessary to explain the experimental data.
These-data 1
indicated that a) the wear is a linear function of time, b) the is maximum at a temperature near 150 C (300 F) and c)'
wear rate the wear increases with velocity.
l
m:
10
/..
Porous oxide layer.1 Water flowing across surface 3
a m-y A
/B C
D
.W 7
V g
s i
is TiannEl
[.
\\
p
.s Metal-oxid e -
Oxide-water interface interface Figure 2.
The different phenomena occurring during corrosion-erosion.
A.
Iron hydroxides are generated:
Fe + H O - Fe(OH) 2 + H2 2
B.
Magnetite is formed according to the Schinorr reaction:
1 1
3Fe(OH)2 Fe 034+h2+
HO 2
C.
A fraction of the hydroxides formed in step B and Hydrogen
{
generated in steps A and B diffuse through the oxide-D.
Magnetite can dissolve in the pore.
I E.
Magnetite dissolves at the oxide-water interface.
F.
Water flow removes the different species oy a convection mass transfer mechanism.
.e 11
-q With the above ideaa. in mind the following. equation was derived:-
C, O 0
1" h + (1-f) (1- + 6)-
(8) j h
D d
2 A" = wear rate in sol /cm s (to be determined),
where 3
C
= equilibrium concentration of iron species in e.
3 I
sol /cm -(obtained from Ref. 9) 2
'2 0 = porosity inice of open area /cm of metal or 3
3 cm H 0/cm(obtained experimentally [15]),
.2
-h
- mass transfer coefficient.in ca/s (obtained from a d
known~ relation for the specific geometry being considered, Berger and Hau, 1977 [16]),
K=A Exp(-E'/RT) = Reaction rate constant in ca/s 1
1 (to be obtained experimentally),
f fry-tion of oxidized metal converted into magnetite
=
at<che metai-oxide. interface (constant f= 1/2),
D = diffusion coefficient of iron hydroxides 2
in water, in cm /s (obtained from Roheenow and Choi. 1961 (17]),
6
= oxide thickness in ca.( to be obtained experimentally),
C,
= Iron species oncentration in the bulk of j
3 the fluid, in mol/cm.
i 12 I
The assumptions made to obtain Eq. 8 vere the following:
l i
Assumption 1.
A steady state is considered.
The oxide layer has been developed and its thickness ($) and porosity (0) have attained a constant value as a function of time. This i
i first hypothesis assume.-
that corrosion brings about the formation of magnetite with the same porosity as the remaining of the oxide layer.
Simultaneously, the layer is assumed to remain at a constant thickness by producing magnetite on the metal oxide interface at tha same rate that oxide dissolves at the oxide-water interface.
rnis assumption t
also makes reference to the amount of hydroxides being formed in the pore. If the resistance to corrosion is constant neither the porosity nor the layer thickness can change.
The amount of 2
sagnetite formed at the metal base has to be the same as that being dissolved in the pore and on the oxide-water interface.
The assumption in the model is that the major part of the
+
dissolution takes place on the oxide
- surface, where the H concentration is the greatest; as a. result, the diffusion 1
through the pores remaics unaffected.
j It is important to note that the werd constant in the above text makes reference to a constant as a function of time.
It will be shown later thac.n order for the model to predict the experimental data observed, porosity has to vary as a function of temperature.
i 13 l
Assumption 2.
Of the amount of iron oxidized at the metal-oxide interface, a-fraction f
is converted into magnetite and this is the same amount being removed at the oxide-water interface.
It has been assumed that f has a fixed value of 0.5.
This value is based on the static experiments of the corrosion of steel;
- where, in
- general, two oxide layers are formed and the outside oxide layer has the same amount of iron as the. inner layer [13].
Assumption 3.
There is no net circulation or flow of water inside the oxide layers therefore the simulation.of the transport of species inside the layer can be considered as a concentration AI2 fusion problem.
Assumption 4.
The water is oxygen free, and the only oxide present is magnetite.
Assumption 5.
The reaction rate at the metal-oxide interface proceeds at a rate equal to:
0 = K9(Ce - Co),
(9) i where I is qual to A Exp
(-E /RT).
This implies that the i
1 reaction at the metal surface is proportional to the potential create?
between C, the equilibrium concentration of hydroxides, and the actual concentration of hydroxides at the metal-oxide interface.
This assumption is verifiable through experimentation; our data will show that the assumption is valid.
E f 14 Assumption ~6.- The hydroxides diffuse through the pores due to differences in concentration from the metal to the water.
The i
liffusion path length is considered to be equal to the oxide thickness.
Assumption 7.
At the oxide-water interface, the concentrction at the pores is C,
but everywhere else it is 1
- different'.
The mass transfer from the pores to the water is considered to be equal to h 0(C -C.), and the mass transfer from d
1 the non-porous oxide is equal to the amount of oxide formed at the setal-oxide interface.
The above assumption permits the use of the following equations t o represent the wear race (&"):
1
" = re(C -C ),
(9) 1 e
o (1-f)s" = DO(C
-C),
(10) 1 6
o 1
(1-f)d" = h O(C -C ).
(11) 1 d
1
=
l I
This system of three equations with three unknowns (i, C, and 1
1 C ) can be solved for 6",
yielding Eq. 8.
(
o 1
(
f l
)
15 3.
Eperimental Studies The experimental work,, aimed at generating wear-rate data, was used to verify the derived model and to obtain the missing values of constants.
3.1.
Description of Experimental Apparatus A
high pressure-high temperature water loop of 2.5 cm. (one inch) 316 stainless steel pipe was constructed.
Its design f
focused on operating at temperatures between 100 to 200 C (200 to 370 F),
with a pH of 5 to 10, with lov oxygen water contents
(<200 ppb) and with velocities of up to 10 m/s (30 ft/s), see Fig.
3.
An
- annulus, a
simple geometry that allowed mass transfer coefficients to be estimated, was used as a test section, see Fig.
3.
The inner diameter of the annulus was the specimen being tested.
It was machined from a plate of A155 Gr.C55 (also equivalent to SA 285 Gr.
C),
a common material used in steam extraction lines (18].
3.2.
Experimontal Procedure minimtm of 200 h, the period believed The tests lasted a
necessary for a
constant wea:-rate to be developed [7, 13);
during this time, pH, velocity and temperature were maintained constant.
4
=
n o
mg l
f cn e
o T
no l
4 i
L so x
rr
)
.m.
e B
o cD p
c
(
i
=
=
p e
2O ht t
N s
f n e
o o
\\
t n
i
/s s t e
t c s
n e m
n s
e s
~
o e
n i
o t l
l c
f n
p s e
e m e i
T.
a ot p
t C
S
(
S fo
)
a l
- i g a i t R e y
D no) i b s
p o
h r ;
r E g u
- i P
n r n
p o
o i n s o i
t oi c
r s el r o o r S
C e Js I
r 0
)
3 e
e1 9
A T
l e
(
'o r
r 30
/
o e
u e
C g
g i
r n
F uls e a
I s s h
e s c
I re x
PV E
2 no N
r-l
{
e taeH
!l, l
l i
i l
)
17 From the weight of the npecimen at the beginning of a test, the descaled weight at the end of the test, the surface area available for wear, and the period of the test, the wear rate was calculated for every flow velocity and temperature chosen.
The method by which the oxide from the specimen was removed is explained in Ref. 14.
Besides recording the weight at the beginning and at the end of a
- test, on various occasions, the weight of a specimen was recorded several times during the course of a run.
The measurements were carried out to verify that although the initial wear rate was variable, after 200 h of operation, it was a constant function of time.
Together with the above information, specimens were analyzed under the scanning electron microscope (SEM).
The thickness of the oxide was directly measured from the pictures taken at specific magnifications.
/
3.3.
Experimental Results Typical data of weight loss as e. function of time are shown in Figure 4.
The figure verifies that the wear rate becomes a constant-in time after about 150 n and that high velocf.ty tests lead to a greater wear than do thos.; et low velocity.
d 1
I f
18 I
)
Mw.
0:
I f
8 i
20 e[
'1 M
M-o
-J 10-O j
"Z" 6 j
O*
/
h
(
O O
N
-lO -
.E O
O 50 15 0 250 TIME (h) i Figure 4.
Specimen at 395 K (250 F) pH of 5.5, O 0.57 m/s j
(1.9 ft/s) and 8 3.13 m/s (10.3 ft/s).
i
I
]
19 Wear rates calculated ~in tiie manner described previously.
are shown in Fig.,
5.
It can be noted that there is a maximus' wear rate close to 150 'C (300 F);
and'that,-except at low temperatures where the process seems to be controlled by the oxidation reaction
- rate, higher ' velocities yield higher wear.
1 SEM. micrographs, not shown here, make evident oxide layer
. thickness.
of around 1 p m.
No particular trend of oxide
' thickness with respect to velocity or temperature was found.
i 4.
Discussions of Results 1
i With the data presented' abo.ve, it is possible to verify the assumptions 'made during th'e derivation of the model as,vell as obtaining values for the unknown constants.
- First, it is important to.n'otice that Eq. 8 shows that in.
terms 1/h. + 6/D are small compared to 1/K, the-limit, vlien the the, wear. rate 'becomes equal to KC 0.
This means that the wear
'e
..is
. independent.of velocity.
This result agrees well with the' data' Presented previously - where at lo'w t'emperatures the rates of high and low velocity specimens yield approximately the,same wear rates.
. Perhaps the most critical assumption in the model is that rate is equal to K0(C -C )', where I, the reaction rate the wear oe constant.
is. assumed to equal the Arrhenit.s expression A Exp(E /RT).
1 1
e 4
n >'E 5 5
0 1
)
0 0
s
/
m 0
6 5
3 0
(
\\
s 5
ne 7
m
\\
, i 1
ce p
4 s
3 y
t i
N c
o5 l
e 5 v
0
=
w 2
oi a
l 3
l p
)
F O,
)
(
r s o/
f m e
0 r
e1 5
0 u
r
)
t u3 C1 0
a t
(
a
(
3 r
r s
/
e e n T
p p e m m
~
mei t c e
e T
. p 0
s s v
8 y
/
e t 2
t i a c r o
=
Iy l
r e a v e
wh 0
g d i 6
e h l
5 2
aQ 2
c s d 1
e n D a
/
/
5 0
4 e
r 2
u 4
2 O
g i
0 0
F 0
0 O
O tog 2s
- o:cE$
i
F i
i 21 From the data ' gathered in,our experiment, it-has been possible to plot values.of I versus 1/T and verify:the above Lassumption.
l 3
n In Eq.-
8, with values of &, h, 6 and C established, a 1
d e
{
value of I can be, calculated for every temperature and velocity.
Values of h, D C..,and 0 were obtained as described previously d
e under the explanation of the' terms to Eq. 8.
The experimental
]
N ~
\\
value's of 6 and b ' were used in determining E.
1 When plotting I as a function of 1/T, Fig. 6 was obtained.
The' figure shows the satisfactory result that I tends to follow
~
an Arrhenius expression and that the activation energy constant E.
was equal to 35,380' cal /g C which agrees well with activation 1
energy values obtained by other authors for the. corrosion of
. steel [15).
The slignment of the data to a straight line is very good considering that tha snalysis covers 3 orders of magnitude for I;. nevertheless, th,ere seems to be some discrepancy at high temperatures.
q l
I The value of the parameters used for h, D, 6 and 9 was l
d t
verified and calculated by different relations.
From this parametric study it was found that the only variable that affects the outcome of Fig. 6 is the porosity 9.
i 1
1
r Temperature (*K) 22 450 430 420 400 390
- 2' I
l...
. U e
e i
g o e
9 e e
l M
C I
e l
\\
J LnK = [(
+ LnAf I
16 cal 1 A = 2.348 x 10 s
cal E.= 35,380 g
9mol e
1
-8
,3 2.18 2.26 2.34 2.42 2.50 2.58 x 10 i
1/T (1/K)
I Figure 6. Plot of natura; logarithm of the reaction rate j
constant (K) vs. the inverse of temperature.
I l
i l
l
4 23 From this result, the following conclusion was derived and l
the 'model of corrosion-erosion for. this application evolved.
The hypothesis is that at high temperatures the fast chemical reactions lead to a decrease in porosity.
The complete model
- therefore takes the following forms C,0 p=
s
(
}
1 1
1 6
g + (1-f)'(g + g) d 17 where I - 2.35x10 Exp (-35,380/RT), in SI units, and 0 = 0.05% at T < 422 I (300 F) 0 = a linear function of temperature between 422 < T < 449 I (300 to 350 F) with values 0.05% > 0 > 0.01%,, respectively.
With this last equation a procedure to estimate the wear rate is now available.
Figures 7
and 8 show the wear rate calculated vs.
the
.errerimental values.
Agreement between experimental and theoretical values is good.
- 5. Summary and Conclusions s
Systematic measurements were perforned concerning material wear due to corrosion-erosion of specimens in single-phase
gk*
5 0
1 0
O.
0 6
-n 3
o
~
iso 5
rr.
7, o5 c
1 5
0 e
4 h =
t 3
H f p o
s s t n l e u m ci e c 0
r e 2
p e s h
=
3
)
F t y t
(
d i 7L.
n c a o e
l 0
r r
a e u
t v s.
)
5 0
t ad w C
0 a
1 o
(
r el 3
e t
T p a e r h m
t e
r a r T
e o wf 0
e,
h 2 8
t 1 2
n e q e E w
t e l b e d
0 n o 6
o m s
5 2
i n 2
r o 1
a i p s m o o r C e
/
0 7
4 e
4 2
2 ru 0
0
~O g
i 0
0 F
0 0
^ O NC-aOE sOe$
l ll
.3 [-
5 "ta 0
1 0
0 n
o i
0 s
\\
6 o
r 3
ro c5 5
7 e 5 h
1 t i lp f
0 o
4 s
s n 3
t e l m ui s c e e r p s
e h y 0
t t 2
i d c
)
3 n o F
al e
(
a v t
e a h r
d g 0
u eh i
)C 0
5 t
t
(
0 a
a e 1
r r h T
3 e
r
. t p
a r e o m
wf e
e T
h 2 t 1 0
n 8
e q 2
e E w
t el b ed n o o m s
0 i n r o 5
6 a i 2
2 p s m o 1
o r C e 8
0 er 4
u g
4 2
2 i
0 0
O F
0 0
0 O
T o e >.%E.]y 53
q 1
26 liquid water with a pH of 5.5 and a two different velocities.
In particular, the effecta of the fluid temperature on a low carbon steel, A155 Gr. CSS, were analyzed.
q During the work two goals were pursued.
The first' goal was' j
to study the state-of-the-art regarding-the phenomenon of
)
corrosion erosion.
Achieving the above objective' led-to the conclusion.that the_ problems occurring from corrosive-erosive wear can.be considerably diminished or even eliminated.if the systems.' a r e properly designed and constructed.
This includes i
the. use-of higher alloying metal concentration. in the steel,
.1 the use of higher s t.e a m - quality, the reduction of the flow l
. velocity, an increase in the pH talue fros.5.5 to 9, avoidance j
l of operation of carbon steels at temperatures around 150 C (300 F) and good workmanship in the connections of pipe fittings.
4 l
Since..the improvement of material composition is an expensive design solution, it may be good practice to use alloy f
st' eels 'only for the - c ri'tical temperature range. For plants already in
. operation and experiencing corrosion-erosion
- problems, further work to suggest inexpensive protective saterial combinations with low risk of galvanic corrosion is i
necessary.
The second goal was to develop a model to estimate wear
]
pten.
The model was developed on the basis of the description of the corrosion of iron given in Ref. 7; its final form is
i 27 S ven; in
.Eq.
12.
It is a simple model that uses a minimum i
number of variables and phenomenological laws to explain the-experimental findings.
With our experimental data the model haw been subjected to initial verification.
.The study. has shown that the most i
important assumptions
- made, i.e.,
a constant wear rate, and a reaction, rate constant following an Arrhenius function with 1
temperature, have been proven appropriate.
It is now required to fGrther validate the model by obtaining. wear rate data for j
different velocities, oxygen
- contents, pH's and other I
t geometries.
Future work on topics related to corrosion-erosion is
)
necessary particularly in the area of porosity as a function of i
environmental conditions, mass transfer coefficients of films and sprays over curved walls and corrosion-erosion by droplet impingement.
Some of this research is being conducted by the M.I.T.
corrosion-erosion group
[2].
In addition, further worP is needed to develop a procedure to estimate film and spray characteristics of annular flow in large diameter piping systems containing two-phase flow (3].
l'
b 28 With_ the knowledge gathered from the proposed research it would. be possible to employ with more confidence or to upgrade the model developed here.
.This would make it possible to
- identify high wear locations, and, consequently, take suitable mesexres to-prevent plant shutdowns or accidents.
This last point should be sufficient to warrant further. studies on the subject.
Acknowledgements We. thank Boston Edison for having. sponsored this research,.
l as well as CONACYT and the Instituto de Investigaciones Electricas, Mexico for their financial assistance.
l l
f l
29 REFERENCES i
[1]
Vu, Hung
" Erosive-Corrosive Wear in Steam Extraction Lines i
, of Power Plants", MIT, M.E.,
M.Sc. Thesis (June, 1982).
I i
[2]
- Berkow, H.
- Jonathan, "The Use of Novel Bend Geometr!.
to Reduce or Eliminate Erosive-Corrosive Wear in Steam Ext.
. tion Lines ', M.E., M.Sc. Thesis, MIT (June, 1984).
j t
[3]
Germano de Freitas,rDissolution Rate in the Wake-of a Welding Backup Ri:4g",3M.E., Engineering Degree Thesis (1986).
[4]
Castle, J. and T. Thomson " Stability of Ferrous Hydroxide in Aqueous Suspension at 300 C", L. Agl. Chem. V. 17 (June, 1967) pp. 177, 178.
[5]
Berge, Ph. "Mecanisme de L'Oxydation des Aciers dans L' Eau a Haute -Temperature et de la Formation des Depots'd'oxydes",
Can g ttas d ' Er'menov111e (Ermenoville, France: March, 1972).
l
[6]
- Gadiyar, H.
S.
and N.
S.
D.
Elayathu
" Corrosion and Magnetite Growth on Carbon Steels in Water at 310 C",
Corrosion-EACE, V. 36, No. 6 (June 1980) pp. 306-312.
[ _
[7]
Heitman H.
G.
and W. Kastner " Erosion-Corrosion in Water-
,[7 p Steam
- Cycles, Causto and Counter Measures",
Jfdl j
v
+
i f
traftwerp.stechnik, V. 62. Heft 3 (March 1982) pp. 211-219.
1
[8] Jehi orr.G.
Z.,
+nore. alls. Chem.
V.212, 533 (1933).
4
/
I r
?
\\
6 r
i e
~-
I 1
i-30
[9]
- Sweeton, F.
and C.
Baes "The Solubility of Magnetite and Hydrolysis of Ferrous Ion in Aqueous Solutions at Elevated Teaperatures",
J Chem.
Thermodynamics.
V. 2 (1970) pp. 479 -
2 1
500.
[10]
Berge, Ph. and F. Ihan " Corrosion-Erosion of Steels in High Temperature Water and Wet Steam" insta12A E.D.F. (Les Renardiers:
i May, 1982).
[11]
- Keller, H.
" Erosion-Corrosion in Wet Steam Turbines" VGB
]
Iraftwerkstechnik, V. 54, Heft 5 (May.1974) pp. 12 - 21.
{
-[12 ]
- Bignold, G.
J.,
E. Carbett, R. Carasey and I. S. Woolsey
" Tackling. Erosion-Corrosion in Nuclear Steam Generating Plants",
Nuclear Engineering International., V. 26, No. 314 (June 1981).
[13]
- Berge, Ph., C. Ribon and P. Saint Paul "Effect of Hydrogen j
i on the Corrosion of Steels in High Temperature Water",
Corrosion-NACE, V. 32, No. 6 (June, 1976) pp. 223-228.
f 1
[14]
Sanchez-Caldera, Luis
" Corrosion-Erosion in Steam Extraction Lines of Power Stations", MIT, M.E. Doctoral Thesis (June, 1984).
[15]
- Surman, P.
L. " Gas Phase Transport in the Oxidation of Fe and Steel", corrosion Science. V. 9, (1969) pp. 771 - 777.
l
[16]
- Berger, F.
P.
and K. Hau, Heat and _ Mass Transfer, V. 20,
)
(1977) p. 1185.
[17]
- Rohsenow, W.
and Harry Choi, 32A1, Mass, tai Momentum Transfyr (Canada: Prentice-Hall, Inc., l961)
[18] ASME Standards.
Annual Book of Standards.
Part 10-G1 (1981),Section II - Part H (1974).
i I
t l
(
l l
VIRGINIA Et,EcrRIC AND Powen COMPANY Ricnwono,VanorwrA 2026 a
i March 6, 1987 I
V.L.STawAar l
Vics Paseinswv Nuctsaa or RAM O N S Dr. J. Nelson Grace Serial No.87-108 Regional Administrator NO:vih U. S. Nuclear Regulatory Commission Docket No.
50-281
{
101 Marietta Street, N. W.
License No. DPR-37 Suite 2900 l
Atlanta, Georgia 30323 i
Dear Dr. Grace:
l
}
VIRGINIA ELECTRIC AND POWER COMPANY l
SURRY POWER STATION UNIT 2
]
FEEDWATER PIPING REPAIR l
In our January 14, 1987 letter (Seri.a1 No.87-009) which transmitted the Eurry l
Unit 2 Reactor Trip and Feedwater Pipe Failure Report, Virginia Electric and Power Company indicated that the feedwater and condensate piping that did not meet our acceptance criterion would be replaced prior to restarting Unit 2.
The acceptance criterion assures that the pipe wall thickness will remain above code wall thickness over the next eighteen months.
Recent inspections i
of the steam generator feedwater lines (WFPD-109, 113, and 117) inside i
l containment have indicated areas of wall thinning. Using the above acceptance I
criterion, line WFFD-109 (C steam generator) is acceptable. However, areas in
)
line WFFD-117 and 113 (A and B steam generators respectively) do not meet the acceptance criterion and require corrective action. Rather t'
~ placing the affected portions of the A and B steam generator feedwater
- repairs will be made in accordance with ASME,Section XI, Article IWA 4000 and performed in
- accordance with the Design Specification and Construction Code.
Background
information, repair technique, expected erosion / corrosion rates and inspection plans are provided below.
The wall thinning was observed in the feedwater line at the speci piece between the loop seal and reducer just upstream of the steam generator. The measuremerts and affected areas are included in the Attachments. These spool pieces were installed du' ring the Unit 2 Steam Generator Replacement Progrcm in the fall of 1979. The "B" feedwater line spool piece was not concentric as originally received.
During installation, the spool piece required weld buildup to return the piece to acceptable wall thickness. After veld buildup of the B feedwater line spool piece, the weld and spool piece were machined to accommodate radiography.
l For each affected area the repair will consist of weld buildup to a wall thickness of approximately.740 to 1.10 inches. In no case will the actual weld buildup exceed 0.75 inch. Base line NDE data vill be established for the veld repaired wall. TtQs repair technique will preserve this data point for further erosion / corrosion evaluation and provide actual data for. design and pipe replacement.
The weld buildup areas will be blended with the existing material to remove discontinuties. Both lines WFFD-117 and WFFD '13 will have the entire spool piece built up.
cG /).sh 2* m. I, C '7 ~ ? (
pp37y
-- y,4y%Qmw
=
y
... ~..
.. ~...
,6
- .. F Erosion / corrosion rates for the 3 spool pieces were calculated based on 5
-years of operation. The erosion / corrosion rate for the B feedwater line spool <
piece was.069 inch / year whereas A and C feedwater lines were.054 and.047 inch / year respectively. The erosion / corrosion rate for the B feedwater line spool piece was calculated using an assumed wall thickness of.835 inch based on originally installed. pipe measurements. Based' on,this erosion / corrosion rate, the decision was made to build up the entire spool piece.
l Approximately 6 months after startup, the Unit will be shutdown for feedwater piping inspection and re-evaluation of erosion / corrosion rates.
If you have any questions, please contact me.
Very truly yours.
Ot. Sh W. L. Stew' art Attachments cc:
U. S. Nuclear Regulatory Commission Attention: Document Control Desk Washington. D.C.
20555 Mr. W.'E. Holland NRC Senior Resident Inspector Surry Power Station Mr. Chandu P. Patel NRC Surry Project Manager JWR Project. Directorate No. 2 Division of PWR Licensing No. 2 l
l 2
+
9 e
o
1 I
1 e
n 1
1 1
Projected I
Minimum Erosion /
Measured Corrosion Area of Code Feedline Wall Rate' Min. Reading Min. Wall' I
"A" S/G
.550"
.054"/yr 6" x 7"
.499" WFPD 117 "B" S/G
. 490" *
.069"/hr(1) 1" x 1" (2)
.499" WFPD 113 4
y "C" S/G
.590"
.047"/yr 13" x 4"
.499" WFPD 109 (1) This rate was determined using an assumed wall thickness of 0.835 inch
' based on an originally installed pipe eccentricity.
l (2) Four additional measurements indicate very localized erosion / corrosion and are below code minimum wall thickness requirements.
Other measurements onthesjoolpiece.indicatewallthinningbutareabovethe minimum code wall thickness requirement.
6 l
>