ML19323F977
| ML19323F977 | |
| Person / Time | |
|---|---|
| Site: | Arkansas Nuclear  |
| Issue date: | 05/15/1980 |
| From: | Clark R Office of Nuclear Reactor Regulation |
| To: | Cavanaugh W ARKANSAS POWER & LIGHT CO. |
| References | |
| NUDOCS 8005300048 | |
| Download: ML19323F977 (3) | |
Text
[fA C> -
yneg
[1
,,o, UNITED STATES
+
3d i
NUCLEAR REGULATORY COMMISSION
.C WASHINGTON, D. C. 20555 v
May 15, 1980 Mh Docke!*tto". 50-368 g_
3n+hs' Mr. William Cavanaugh, III Vice President, Generation and Construction Arkansas Power & Light Cocpany P. O. Box 551 Little Rock, Arkansas 72203
Dear Mr. Cavanaugh:
On December 28, 1979 the NRC Office of Inspection and Enforcement (IE) issued Information Notice No. 79-37 that discussed the discovery of cracks in the keyway and bore sections of discs in Westinghouse low-pressure turbines. A copy of this Information Notice with an errata sheet is enclosed.
Subsequently, all licensee /
users of low-pressure turbines manufactured by General Electric were invited to meet with the NRC staff and representatives of the vendor on January 9,1980 to discuss the probability of disc cracking in these turbines. A summary of this meeting and the General Electric Conpany's presentation are also enclosed with this letter.
At the time of the January 9 meeting General Electric did not have any recent results of ultrasonic inspections of its low-pressure turbines. Since that date full UT inspections have been performed on six rotors at five nuclear power plants.
Some indications in the keyway region have been reported in discs at three of these plants. General Electric personnel believe that these indications were caused by water erosion rather than by stress corrosion.
The staff desires to learn more about the underiying reasons for the indications found and the probable rate of growth of these indications and their effects on turbine disc integrity.
For this purpose we request that you provide the information sought in Enclosure 3 to this letter and address its safety significance. Under the provisions of 10 CFR 50.54(f) your response is requested within 30 days of the receipt of this letter.
A copy of this letter is being telecopied to you, along with Enclosure 3.
It is my understanding that additional UT inspections are to be performed by General Electric in the near future. We encourage this action as being the only certain means of determining the integrity of turbine discs. We also recomend that if you have not already done so, you develop a schedule for performing a full UT inspection of at least one of your low-pressure turbines during the next major outage of your plant.
THIS DOCUMENT CONTAINS P00R QUAL.lTY PAGES 800s300 o_4 7
This request for generic information was approved by GA0 under clearance number B-180225 (S79014); this clearance expires June 30, 1980.
Sincerely.
J((
Robert A. Clark, Chief Operating Reactors Branch #3 Division of Licensing
Enclosures:
1.
Information Bulletin 79-37 2.
Meeting Summary 3.
Information Requests cc:
See next page 4
i
=
=
Arkansas Pow:r & Light Company ccw/ enclosure (s):
D Ma[1ag. L ce s on 1
Arkansas Power & Light Company Office of Radiatior. Programs i
P. O. Box 551 (AW-459)
Little Rock, Arkansas 72203 U. S. Environmental Protection Agency Crystal Mall #2 Mr. James P. O'Hanlon Arlington, Virginia 20460 General Manager Arkansas Nuclear One U. S. Environmental Protection Agency P. O. Box 608 Region VI Office Russellville, Arkansas 72801 ATTN:
EIS COORDINATOR l
1201 Elm Street First International Building S
u ar u atory Commission Dallas, Texas 75270 P. O. Box 2090 Russellville, Arkarisas 72801
~
Mr. Robert B. Borsum Director, Bureau of Environmental Babcock & Wilcox Health Services Nuclear Power Generation Division 4815 West Markham Street Suite 420, 7735 Old Georgetown Road Little Rock, Arkansas 72201 Bethesda, Maryland 20014 Troy B. Conner, Jr., Esq.
Conner, Moore & Corber 1747 Pennsylvania Avenue, N.W.
Washington, D.C.
20006 Arkansas Polytechnic College Russellville, Arkansas 72801 Honorable Ermil Grant Acting County Judge of Pope County Pope County Courthouse Russellville, Arkansas 72801 Mr. Paul F. Levy, Director Arkansas Department of Energy 3000 Kavanaugh Little Rock, Arkansas 72205
.we wp em==
.w
9 i
UNITED STATES ISINS NO.:
6870 NUCLEAR REGULATORY COMMISSION Accession No.:
i OFFICE OF INSPECTION AND ENFORCEMENT 7910250525 l
WASHINGTON, D.C.
20555 Dece,ber 28, 1979 IE Information Notice No. 79-37 l
CRACKING IN LOW PRESSURE TUR8INE DISCS Oescription of Circumstances:
An anonymous letter was received by the Director of the Office of Inspection and 17, 1979 which alleged possible violation of Part 10 CFA Enforcement, on November 50.55e and/or 10 CFR 21 Regulations concerning reportanility of recently discovered stress corrosion cracking in Westinghouse 1800 rpm low pressure turbine discs.
Westinghouse had made a presentation on the turbine disc cracking to electric utility executives on October 30, 1979.
I Telephone discussions between the NRC staff and Westinghouse's Turbine Division on November 20, 1979 established that cracking, attributed to stress corrosion phenomena, had been found in the keyway areas of several LP turbine discs at operating plants and that inservice inspection techniques (i.e., in situ ultra-sonic examinationJ for crack detection have been developed and are being imple-mented in the field.
The Office of Inspection and Enforceaent was also notified on November 20, 1979 that during the current overhaul of Commonwealth Edison's Zion Unit 1 LP turbine, ultrasonic examination revealed embedded cracks located on the inlet side on the disc bore area where no cracks had been previously observed. Ultrasonic measurements indicate this disc bore cracking is of greater depth than the keyway cracks found to date.
According to Westinghouse, these bore cracks have been metallurgically examihed and preliminary findings show them not to be typical of classical stress corrosion cracking observed in the keyways.
The probable cracking mechanism and impact on disc integrity is being further evaluated by Westinghouse.
A meeting was held on De'cember 17, 1979 between the NRC staff, Westinghouse l
and utility representatives to discuss the disc cracking problem, repair alter-natives, turbine missile evaluation, inspection techniques and plant inspection priorities.
In response to the staffs' request, Westinghouse provided the staff an updated report on December 21, 1979 regarding the current field inspection program that included a list of nuclear power plants already inspected, recom-mended inspection schedules and pertinent information related to LP turbines
~
Inspections to date have identified turbine where cracks have been cbserved.
disc cracks at Surry Unit' 2, Point Beach Unit 2, Palisades, Indian Point Unit 3 and Zion Unit 1.
All units except Point Beach Unit 2 will make repairs before Point 8each returned to power on December 23, 1979 the plants return to power.
with a small crack in the No. 2 disc of LP Turbine No. 2.
An analysis by Westinghouse indicated that the observed crack will not attain critical dimensions The NRC staff is evaluati,g during 28 additional months of turbine operation.
the turbine inspection results and analysis by Westinghouse.
l
w ev i
l IE Information Notice No. 79-37 December 28, 1979 Page 2 of 2 i
Westinghouse also notified the staff that extrapolation of information obtained from Indian Point Unit 3 inspection and analysis indicates that disc cracking j
could be significant at Indian Point Unit 2 and the turbines should be inspected sooner than the spring outage of 1980.
The NRC staff is currently reviewing Consolidated Edison's plans for prompt evaluation of this potential problem at this unit.
i lists the PWR plants having Westinghouse 1500/1800 rpm turbines.
The AA category represents those turbines which appear to have Lhe earliest need for inspection. With the exception of Yankee Rowe, Westinghouse has i
recommended to utilities that inspection of these machines be completed by the Spring 1980 outage period. The Rowe unit is uninspectable by the present ultrasonic techniques due to its design. Westinghouse has recommended the remaining machines of the Category A plants be inspected as their service periods approach five years or in the event significant corrosion problems become evident during this time. The NRC staff is currently reviewing the need for inspection of those PWR plants having other interfacing turbine designs shown in Enclosure 2.
Changes to the forementioned inspection schedules proposed by Westinghouse may be necessary as new technical information becomes available.
4 From the information available to the NRC staff at this time it appears that cracking may be more generically widespread in turbine discs (e.g., keyways and bore areas) than previously observed.
It is important to note that'the UT inspections performed by Westinghouse thus far were essentially limited to the keyways (disc outlet) of selected discs whereas the Zion Unit 1 inspection results indicate that examination of the disc bore section must be taken into account. Also, Westinghouse is currently re-evaluating their previously estimated l
turbine missile energies based on recent missile test results from model symmetric and non-symmetric missile impact tests.
Their preliminary findings, although subject to change, now indicate possible higher missile exit energies in some cases than previously expected.
This Information Notice is provided as an early notification of a possibly significant matter, the allegations and the generic safety implications of which j
are currently undergoing review by the NRC staff.
It is expected that recipients j
will review the information applicable to their facilities.
If NRC evaluations l
so indicate, further licensee actions may be requested or required.
Embedded cracking in keyways'and disc bore areas have been observed only in Westinghouse i
LP turbines thus far.
However, the NRC staff believes that turbines of other manufacturers should be included in consideration of this problem..
No written response to this Information Notice is required.
If you have any j
questions regarding this matter, please contact the Director of the appropriate NRC Regional Office.
Enclosures:
As stated 5
= - -
.,.-c---
O O
CATEGORY AA UTILITY STATION UNIT Florida, P&L Turkey Point 3
Consolidated ED.
Indian Point 2
PASNY Indian Point
'3-Arkansas P&L Russellville 1
VEPCO Surry 1-Carolina P&L Robinson 2
So. Calif. Ed.
San Onofre 1
Yankee A.P.
Rowe 1:
1 Wisc. Mich. Pwr.
Point Beach l'
14 Consumers Pwr.
' Palisades
=
Commonwealth Ed.
Zion 1 1
Commonwealth Ed.
Zion.2 2'
Florida P&L Turkey Point 4
Nebraska PPD Cooper
- 1.
Wisc. Mich. Pwr.
Point Beach 2
Maine Yankee Bailey Point 1
Rochester G&E Ginna 1-Northern States Prairie Island 1
Wisc. P.S.
Kewaunee c
1.
s.
l
)
l s
1 l
l
. (Continued)
Page 2 of 3 CATEGORY A i
UTILITY STATItj UNIT Alabama Power Farley 1
Alabama Power Farley 2
Baltimore G&L Calvert Cliffs 2
Carolina P&L Harris 1
Carolina P&L Harris 2
Carolina P&L Harris 3
Carolina P&L Harris 4
Cincinnati G&E Zimmer 1
Comonwealth Ed.
Byron 1
Commonwealth Ed.
Byron 2
Comanwealth Ed.
Braidwood 1
Comonwealth Ed.
Bra.idwood 2
Connecticut Yankee Haddam Neck 1
Duke Power McGuire 1
Ouke Power McGuire 2
Duquesne Lt.
Shippingport 1
Ouquesne Lt.
Beaver Valley 1
'Ouquesne Lt.
Beaver Valley 2
Flordia Power Corp.
Crystal River 3
Flordia Power & Lt.
St. Lucie 1
Flordia Power & Lt.
St. Lucie 2
Houston L&P Se Texas 1
Houston L&P So. Texas 2
Louisiana'P&L Waterford 3
Hetropolitan Ed.
Three Mile Island 2
Northern States Pwr.
Prairie Island 2
Pub. Service E&G Salem 1
Puk Service E&G Sales 2
Pacific G&E Diablo Canyon 1
Pacific G&E Diablo Canyon 2-4
. (Continued)
Page 3 of 3 CATEGORY A UTILITY STATION UNIT P.S. Indiana Marble Hill 1
P.S. Indiana Marble Hill 2
Puget. Sound P&L Skagit 1
SMUO.
Rancho Seco 1
TVA Sequoyah 1
TVA Sequoyah 2
TVA Watts Bar 1
TVA Watts Bar 2
VEPCO North Anna 1
VEPCO North Anna 2
VEICO North Anna 3
VEPCO North Anna 4
VEPCO North Anna 2
WPPSS Hanford 2
WPPSS WNPS 1
WPPSS WNPS 3
WPPSS WNPS 4
WPPSS WNPS 5
0
UTILITY STATION UNIT Ouke Power Co.
Oconee 1
Ouke Power Co.
Oconee 2
Ouke Power Co.
Oconee 3
OPPD Ft. Calhoun 1
Baltimore Electric
& Gas Calvert Cliffs Metropolitan Edison Three Mile Island 1
Indiana & Michigan Electric D.C. Cook 1
Indiana & Michigan Electric D.C. Cook 2
Northeast Utilities Millstone 2
Portland General Electric Trojan 1
Toledo Edison Davis Besse 1
Arkansas Power &
Light Arkansas Nuclear One 2
- WPPSS, Hanford 1
4
- e
l i
f Errata Sheet For IE Information Notice No. 79-37 Page 1, paragraph 2, line 9: Change " inlet" to " outlet" Page 1, paragraph 3, lines 9 and 10: Change " Point Beach Unit 2" to
" Point Beach Unit 1" l
1 Page 1, paragraph 3' line 10: After Point Beach Unit 1 add an asterisk
("*") footnote and place a note at the bottom of the page as follows:
"* Wisconsin Electric Power Company orally notified the NRC project manager on Novenber 5 that turbine disc cracking had been observed at Point Beach Unit 1.", line 4: Change "Russellville" to "ANO", line 8: Change to read:
" Yankee Atomic Electric, Yankee Rowe", lines 9 and 15: Change "Wisc Mich Pwr" to "Wisc Elec Pwr", line 16: Change to read " Maine Yankee Atomic Pwr, Maine Yankee", page 3, line 13: Delete as redundant reference to North Anna 2 O
~
dd$e@a A
.o (W
[t'
f.
b M $^d21 % 1 S.
[
b, UNITED STATES NUCLEAR REGULATORY COMMISSION
! l.
, i 'I s.
wAsmucTom. o. c. 2 osse 3
s_,
rE3Ru m z 1 an MEMORANDUM FOR:
A. Schwencer, Chief Operating Reactors Branch #1, 00R j
FROM:
W. J. Ross, Project Manager Operating Reactors Branch #1, 00R
SUBJECT:
MEETING WITH GENERAL ELECTRIC RELATED TO TURBINE DISC CRACKS r
At the staff's request, representatives of General Electric met with the staff and licensee / user of G.E. turbines on January 9,1980 to discuss the design of and operational experience of low pressure turbines. A list of attendees at the non-proprietary and proprietary sessions is attached.
1 In the non-proprietary session General Electric's personnel discussed the following topics:
(a) turbine wheel (disc) integrity; (b) minimizing wheel (discs) bursts.
Copies of the slides used in this presentation &
attached in Enclosure 3.
Turbine Wheel Integrity There has been no indication of cracks in the bore regions of G.E. Icw pressure turbine wheels. This experience includes the operation of 35 turbine:: at nuclear generating plans (22 BWRs-of which only three have actually b,een inspected) and 4000 wheels at 235 fossil units (percent of wheels inspected was not established at the meeting). A G.E. turbine has the following design characteristics:
(1) 14 wheels or discs per rotor with 38" or 43" active length buckets or blades.
(2) Rectangular axial keyway and circumferential locking ring minimize rotation of wheels on the turbine shaft (i.e.
minimize stress).
(3) Feedwater systems on nuclear plants are provided with full-flow dominera11zer, (experience with fossil plants has been good even with poor water chemistry).
General Electric referenced a 1973 memorandum that postulated the proba-bilities of turbine missiles as folicws:
N f
_\\ f
,h, L \\!R gs
P (<127% normal turbine speed) 2.6x10~7
=
1.5x10~7 P (runaway)
=
~~~
4.1x10~0 P (lifetime total)
~
=
1.4x10~
P (annual average)
=
General Electric continues to reconenend that its users perform UT inspections of wheel bores at 6-year intervals. A satisfactory UT test has ben developed for this purpose. To date the three UT tests that have been performed were on nuclear turbines that had averaged about 3 years of operating experience.
Minimizing Wheel Bursts A brief review of actions taken by G.E. to eitminate the formation of cracks in turbine wheels was presented and included the following:
- 1) Forging process designed to eliminate internal cracks.
- 2) New wheels are inspected by visual UT and magnetic particle techniques.
- 3) New wheels are tested at 120% operating speed.
- 4) Tolerance of defects caused by stress and corrosion maximized through choice of material.
- 5) Provision for UT toting of wheels after installation and use.
General Electric's personnel provided the following responses to questions fron the audience.
1.
Retention of a 6-year interval for inspection of turbines was justified by operating experienes and crack growth rate studies.
2.
Three G.E. turbines at nuclear plants (1 BWR and 2 PWR) have been inspected by UT and no indications observed.
These turbines had seen about 3 years of service.
3.
All wheels of turbines at nuclear sites are inspectable in sites l
(without removal from the turbine). Approximately 5 days are required to inspect one 14-wheel rotor.
Four additional days would be needed to complete the inspection of a second rotor at j
the same site.
4.
The length of a crack is postulated by G.E. to be 4 to 5 times the depth detemined by UT.
5.
Overspeed devices on G.E. turbines are testable under load and retain their protective capability during the test.
l 1
6.
The G.E. representatives were not aware of any customer experience where overspeed emergency systems have ever been used.
7.
Minimum defect size observable by UT as practiced by G.E.
is as small as 30 mils in new wheels.
8.
G.E. has little data related to chemical analysis of deposits in cracks because few have been observed. Chloride has been observed in pin cracks.
9.
Cogarisons of stress corrosion cracking between turbine discs (3.5% Ni C1 Mo V) and in 304 ss pipe are not appropriate because they have different fluids (water and steam) and different materials.
10.
G.E.'s wheel keyways are not shielded from steam flow chemicals (Westinghouse is considering such protection)
- 11. Thermal and vibrational stresses on turbine wheels are considered to be very small in comparision to design capability.
12.
G.E. does not presently have specific teams of inspectors mobilized for inspecting turbines.
William J. Ross, Project Manager Operating Reactors Branch #1 Division Of Ooerating Reactors Attachments:
Attendees Slides used in G.E. presentation 2
t i
l e
ATTENDEES 1
GENERAL ELECTRIC fiRC R. S. Couchman W. JUoss D. P. Timo R. E. Johnson J. J. Hinchey M. L. Boyle W. J. Kaehler A. Taboada H. T. Watanase J. J. Zudans R. D. Brugge W. P. Gannill K. G. Hoge T. Ippolito W. J. Collins E. G. Arndt I. A. Peltier K. M. Cage i
F. Clemenson C. D. Seller M. Wohl R. W. Klecker K. R. Wichsan W. S. Hazelton V. Noonan i
e
,,vg.9 7 g e e v.n UTILITY NAME Becntel Power Corporation
/
Paul Nastick C. W. Hultman
- 'lPPS Portland General Electric J. E. McEwen Stone and Webster John Coombe American Electric Power Richard G. Kriftner Paul H. Barton Duke Power Company Northern States Power Conpany Craig F. Nierode Otakar Jonas Westinghouse Consuners Power Company James B. Lewis Robert L. Smith Yankee Atomic Electric Conpany Jersey Central Power and Light Company Jim Knuber Tennessee Valley Authority Larry A. Johnson Norman H. Gaffin Philadelphia Electric Company Ma rtin J. McCormi ch, J r.
Philadelphia Electric Company L. Erik Titland Baltimore Gas and Electric Company Baltimore Gas and Electric Company j
Ronald O'Hara Baltimore Gas and Electric Company 1
R. Niall M. Hunt Baltimore Gas and Electric Company i
R. G. Clisham Omaha Public Power District S. F. Binderup C. C. Seitz Metropolitan Edison Company P. K. Colvert Commonwealth Edision Conpany R. J. Tammi nga Commonwealth Edision Company W. G. Cl a rk. J r.
Westinghouse R. E. Warner Westinghouse V. S. Anderson Westinghouse B. B. Seth Westinghouse J. F. Etzweiler American Electric Power
k e
S NUCLEAR STEAM TURBINE WHEEL RELIABILITY T
INTRODUCTION This paper deals with shrunk-on bucket wheels used in steam turbines manufactured by the Large Steam Turbine-Generator Department of General Electric for use with nuclear reactor cycles.
In particular, General Electric's efforts aimed at avoiding a wheel burst as the result of stress corrosion cracking are described.
These efforts include steam purity recommendations, the use of optimum design, material ' selection and acceptance practices, and the introduction of a computerized in-service wheel bore ultrasonic test.
In February,1978, TIL-857 was issued outlining recommended in-service inspection practices for nuclear steam turbine rotors manufactured by General Electric.
Recommendations include a complete ultrasonic examination of the shrunk-on wheel bores at about 6-year intervals.
OVERVIEW There are two possible mechanisns for initiating and/or growing cracks in nuclear wheels in service:
1.
Stress Cycling 2.
Stress Corrosion Cracking The likelihood of initiating and/or growing a crack due to stress cycling associated with starts, stops, ur load changes is small.
Variations in stress amplitude resulting from operating transients are too low to produce significant crack growth, in the unlikely event that a defect exists in the whee's when it is placed in-service. The manner in which wheels are forged essentially precludes the possibility of producing an internal crack-like defect in the plane normal to the maximum stress (the axial-radial plane). Furthermore, in addition to a complete visual and magnetic particle inspection of all bore and external surfaces,
all nadern nuclear wheel forgings are subjected to a stringent 100% volumetric ul rason '- inspection at the time of manufacture. All nuclear wheels are spin tested during manufacture at 20% overspeed(although on some.early units a few wheel.s withou buckets attached), which further minimizes the probability of having ~an undetected crack or crack-like flaw with a dritical size which wo'Uld lead to spontaneous propagation at normal rotating speeds.
Thus, the major source of concern with respect to the in-service initiation and growth of cracks is that associated with stress corrosion. Although we have not observed stress corrosion.(or other) cracking ir, the bores of any of our fossil or nuclear wheels made of modern materials, the possibility of initiation and propagation of a stre:s corrosion crack at the bore of a shrunk-on wheel cannot be entirely discounted. We and other turbine manufacturers have experienced stress corrosion cracks in the wheel dovetail region of integral rotors made of essentially the same material as that used in nuclear LP wheels. Recently, stress corrosion cracks were detected on the periphery of two modern GE shrunk-on wheels, operating in a fossil plant, which had been exposed to heavy caustic deposits.
In addition, machines of both domestic and foreign design (not GE) have suffered cracks in '%e bore ragion of shrunk-on wheels, leadino to wheel bursts in several Cases.
A considerable number of steps have been taken to reduce the probability of a wheel burst due to stress corrosion cracking. Laboratory tests and service exparience have indicated that consistent high levels of steam purity provides the best protection against stress corrosion cracking. Steam purity recommendations have been published in GEX-72281, attached. Modern wheels manufactured for nuclear
and fossil turbines are maJe from the highest quality vacuum poured NiCrMoV forgings which am heat treated to obtain an optimum combination of strength, ductility and toughness. Each forging must pass a stringent material acceptance procedum before being considered for use as a wheel. The procedure includes the non-destructive inspections previously discussed, along with laboratory tests to verify that material properties fall within specifications. Wheel geometries have been chosen to maintain the lowest level of operating stress.
Although a considerable effort has been made to maintain superior fracture toughness properties, and to minimize the likelihood for stress corrosion cracking, laboratory and field experiences indicate that the possibility for a wheel burst due to stress corrosion cracking caanot be discounted. For modern GE design nuclear shrunk-on wheels, the materia! crack size in the rim region which would lead to a wheel burst is very large, approaching the axial thickness of the wheels.
Thus, such cracks, should they exist, would have a high probability of detection by surface inspections. This is not the case in the bore region where, because of higher stress levels, a crack could gmw to a dangerous size prior to breaking through to an accessible surface. For this reason, G.E. has dueloped an ultra-sonic test which searches for cracks at the wheel bore and keyway surfaces, along with material in the vicinity of the bore.
TIL-857, which was issued in February,1978, outlines our recomendations for periodic inspection of General Electric steam turtine rotors operating in nuclear power plants. As described in TIL-857, we recomend that a complete magnetic particle inspection of the shrunk-on wheels should be performed during any outage when the turbine section is open. In addition, we reconsend a more extensive test at about 6 year intervals, which should include an ultrasonic inspection of the shrunk-on wheels using the wheel bore ultrasonic test mentioned above.
I A great amount of development work on stress corrosion cracking has been conducted, and our understanding of this phenomenon is much improved, although still inadequate to predict the precise time required for crack initiation and the rate of crack growth in a cormsive environment. It is therefore impcssible to specify absolute " safe" inspection intervals to preclude the possibility of initiating and growing a' crack to critical size between inspections. Recognizing this, we nevertheless believe that periodic inspections, as described in TIL-857, will greatly reduce the probability of a wheel burst.
The remainder of this report describes in greater detail the service experience of G.E. wheels in nuclear and fossil plants, along with the steps which have been taken to understand and reduce the chances for a wheel burst. Also included is a description of the wheel bom ultrasonic test, and the experimental program which helped to develop and verify the test capabilities.
SERVICE EXPERIENCE At the present time, 42 (35 domestic) large steam turbine-generator uni _ts manufacture by G.E. are operating in nuclear power plants. The 1579 shrunk-on wheels in these units have accumulated over 9,000 wheel years of service without having experienced a failam. To date, 46 of these wheels have received an in-service wheel bom ultrasonic inspection with the result that no crack'like indications have been found.
4,
Of the large steam turbine generators operating in fossil plants, 235 G.E. units have shrunk-on wheels. This ac. 'unts for over 4000 wheels which have accumulated over 130,000 wheel years of se.vice.
To date, only one wheel failure has been experienced, and this ogcurred on a multiple stage Curtis design wheel which was made from an 1850 F austentized CrMcV material. The wheel had been in-service for roughly 30 years and was scheduled for replacement.
Metallurgical analysis of the fracture showed that cracks, probably caused by creep-rupture, initiated in pin bushing holes which extend radially from the bore surface. Unlike nuclear shrunk-on wheels which are used strictly in low temgerature applications, this wheel operated at temperatures on the order of 900 F.
The wheel fragments did not penetrate the turbine casing.
Fourteen shrunk-on wheels from two GE units, not manufactured from a modern NiCrPbV material, experienced stress corrosion cracking early in the 1950's.
The cracks, which initiated in pin bushing holes, were detected within.several years after the units went into service. (Nuclear wheels do not have pin-bushing holes.) An investigation revealed that a heavy build-up of deposits had formed on the turbine as a result of a large percentage of make-up water. To remove these deposits, the custoper washed the turbine with a mild caustic solution. Caustic leaked into the pin holes, and concentrated during continued operation-resulting in the formation of stress corrosion cracks. The wheels were replaced and the units placed back into service.
Since this time, a total of 260 fossil wheels from 19 G.E. large steam turbines have received a full inspection.
166 of which were disassembled frcm the shafts while 94 have been inspected with the wheel bore ultrasonic test.
Wheel disassembly was largely performed in conjunction with TIL-647, which called for unstacking particular built-up rotors to improve shaft and wheel geometry.
While unstacked, the wheels were given a complete magnetic particle inspection of bore and peripheral surfaces. No cracks were found in these wheels. The wheel bore ultrasonic inspections performed to date on fossi1 wheels, resulted either from customer requests or from G.E. recommendations as the result of known steam chemistry upsets. No crack-like indications have been detected in bore or keyway regions.
Recently, a magnetic particle examination revealed stress corrosion cracks on the periphery of two wheels operating in a fossil unit. This unit was found to have heavy deposits throughout, the chemical analysis of which revealed that they were largely formed by caustic deposition. Stress corrosion cracks were found on the wheels and at other locations in the high pressure and low pressure sections. Cracks on the wheel peripherles were found to be shallow.
The ultrasonic inspection of the wheel bores and keyways revealed no indications, implying that if cracks existed, they were shallow.
Experience of other manufacturers with wheel stress corrosion cracking in nuclear units has received considerable attention in recent years. Pe rhaps the best documented is the British experience at the Hinkley Point A nuclear power station. In Septancer,1969, turbine generator No. 5 suffered a catastrophic wheel burst, which studies found to be the result of stress corrosion cracks reaching a critical size in the wheel keywayd.
A detailed follow-up study by the CEGB revealed stress corrosion cracks in a i
considerable nunber of wheels having semi-circular keyways. The 3 CrMoV material used in many of these wheels was found to be temper embrittitd and highly susceptible to stress corrosion cracking.
Bursts of large steam turbine wheels manufactured by other suppliers and operating in the United States are known to have occurred in fossil plants.
Recent investigations have also revealed stress corrosion cracking in some nuclear steam turbine wheels.
LABORATORY PROGRAM An extensive laboratory program has been underway to better quantify the resistance of wheel materials to stress corrosion cracking. The program has concentrated on relating mechanical, material and electrochemical parameters to the processes of crack initiation and growth. Most of the work to date has focused on caustic cracking, although other environments have been studied, including high puri ty water. Caustic appears to represent the greatest threat to G.E.
wheels operating in nuclear and fossil power plants.
A considerable number of testing procedures have been utilized to investigate the resistance of NiCrMoV steels to stress corrosion cracking.
Figure 1 shows results from dead weight load tests performed in caustic solutions under various conditions. Figure 2 illustrates the test apparatus with the environmental container renoved so the smooth tensile type specimens can be observed. Scanning electron micrographs of a wheel material tested to failure in caustic are shown in Figure 3.
Quantative measureients of the stress corrosion crack propagation rate in wheel alloys have been made using fracture mechanics type specimens (Figure 4).
Generally, stress corrosion crack growth rates are depicted by plots of the type shown in Figure 5.
The crack propagation rate is plotted versus the crack tip stress intensity factor, which is a function of the applied stress, the crack size and the geometry. A family of such curves are necessary to describe the variation of crack propagation rate with corrosion potential, temperature and caustic concentration.
Of particular interest are the threshold stress intensity (K,3), the plateau or second stage, and the third stage of crack growth. The threshdTd stress intensity is the limit below which stress corrosion cracks will not propagate.
7c plateau or second stage is the region of relatively stable crack growth, over a range of stress intensity.
Crack growth in region 3 is sharply accelerated as the crack tip stress intensity increases. Testing in the laboratory under controlled conditions has indicated that the threshold stress intensity in caustic is low, j
and in fact may be virtually non-existent. Figure 6 shows the typical range of stage 2 crack propagation rates versus temperature measured on wheel materials in 40% NaOH and maintained near the optimum corrosion potential to accelerate crack growth.
Corrosion potential has been found to be one of the most important parameters influencing the resistance of NiCrMoV wheel materials to caustic stress corrosion cracking.
In the laboratory, corrosion potential can be controlled, thus providing a convenient means to run accelerated stress corrosion tests.
Varying the potential from optimum, however, signficiantly reduces stress corrosion susceptibility. When uncontrolled, the fre'e corrosion potential in caustic generally lies outside the range of maximum severity, except under transi ent conditions.
Additional electrochemical studies have shown however that trace quantities of certain compounds, such as PbO, Cu0 and NANO,, can influence the corrosion potential. It is possible that a critical quantity of trace compounds,
if present in caustic deposits, will shift the potential towards an undesirable
. region. Thus, this is a possible mechanism for explainin,g why fos,sil_ steam turbines, with caustic deposits,have not in all cases experienced stress corrosion cracking.
Tests have been performed to evaluate the potential for stress corrosion crack propagation in high purity water. Such an environment is reasonably typical of condensate in the wet stages of steam turbines operating in plants with BWR reactors. The tests have shown that although stress corrosinc cracks can propagate in this environment, maximum rates of propagation have 91..erally been lower, by a factor of 100 to 1000, then the maximum rates in 40% caustic.
Similar results have been published by the British, who tested NiCrMcV materials both in water and"high quality" power plant steam.
GENERAL CONCLUSIONS Based on laboratory test results and service experience, some general conclusions have been reached with regard to the potential for stress corrosion cracking in shrunk-on wheels.
1.
Concentrated deposits, such as caustic, represent a significant threat to steam turbine components because of the possibility of developing stress corrosion cracks. The likelihood of cracks caused by caustic deposition is dependent on the corrosion potential, the temperature, the causti.
concentration, the operating stress level, and the materials, and physical characteristics. Corrosion potential can be influenced by the presence of trace impurity compounds.
2.
Stress corrosion resistance generally decreases as the material tensile strength increases.
3.
There is considerable heat to heat scatter in resistance to stress corrosion cracking within a given material specification.
4.
Modern large steam turbines require materials which cannot be made completely inmune to stress corrosion cracking. Proper control of water chemistry remains the best way to guard against stress corrosion cracking.
5.
Some stress corrosion cracking has been observed in NiCrMcV laboratory specimens exposed to pure water and wet steam enviromnents. Maximum crack growth rates measured in these environments,however,are significantly less than those measured in more corrosive environments such as caustic.
To date, G.E. service experience on steam turbine components manufactured from materials typical of modern wheel alloys has never linked stress corrosion cracking to pure steam or pure steam condensate.
In all cases, stress corrosion cracks in these materials have been associated with concentrated caustic deposits. We believe,however,that at higher tensile strengths and/or at stress levels beyond those currently found in G.E. wheels,
a significantly greater potential for stress corrosion cracking in relatively non-aggressive environments does exist.
6.
Locally aggressive environments may develop in surface pits or in regions of the turbine where steam flow is restricted. Wheel keyways are regions where th 's can potentially occur.
l 7.
In a contaminated steam environment cracks can grow by a stress corrosion mechanism, to critical size.
l
STEAM PURITY The need for good control of water chemistry in nuclear, as well as fossil l
fired steam turbine power plants, is generally recognized and the concentration of impurities in such systems is generally held to very low levels. Due to effective concentrating mechanisms operative in steam turbines low levels of impurities can be transformed into concentrated solutions, however.
There are three major concentrating mechanisms operative in a turbine, and they are briefly described i
below. Some of these mechanisms may be operative to a greater extent in PWR plants than in BWR plants.
1.
Deposition from Superheated Steam As steam expands through the turbine, the solubility of impurities in the steam decreases. Deposits form in the turbine when the concentration of impurities exceeds their solubility limit. The impurity concentration in a deposit can be far greater than its concentration in the steam.
Concentrated caustic and chloride solutions are typical of deposits which can form by deposition from superheated steam. Deposits of this type formupstream of the Wilson line.
2.
Acid Concentration at the Wilson Line Turbine steam may also be contaminated with organic or inorganic aci ds. Unlike the situation for caustic and chloride described above, acids are very soluble in superheated steam and do not exceed their solubility. At the Wilsen line, however, these acids become enriched in the first drops of water that are formed. This results in a situation in which low levels of acid in the steam can result in a concentrated acid solution at the early moisture region.
3.
Evaporation and Orying Another concentrating mechanism which occurs in steam turbines is the formation of concentrating solutions by the evaporation of water from dilute solutions. Far example, during a cold start-up ste'am will condense on metallic surfaces which are at a lower temperature than the steam temperature. As the metallic part heats up, the moisture evaporates and practically all of the impurities are left. This concentrating mechanism is not a significant problem for parts with exposed surfaces because the impurities wil1 be redissolved by the large volume of steam flow during operation. In stagnant areas, however, like the spaces between buckets and wheels br in the crevices between wheels a'nd a shaft-)
the solutions can concentrate and remain for long periods during operation.
Some of the incidents that can result in contamination of nuclear turbines are:
1.
Use of cleaning fluids with unacceptable levels of caustic, chloride and sulfur for removing protective materials used during shipment.
2.
Cleaning solutions used to remove deposits from the turbine or associated cosponents.
i 3.
Contaminated exhaust hood sprays used to control temperature in the low-pmssure hoods during low load operation.
. 4 Improper operation and/or regeneration of feedwater demineralizers.
1 5.
Contamination from condenser cooling water source because of condenser tube failures.
Stress corrosion cracking has been found in fossil steam turbine components from each of these causes but not generally have shrunk on wheels been effected.
GEK-72281 outlines General Electric steam purity recompendations as applicable to the turbine-generator unit. Other requirements may be applicable to maintenance of auxiliary components such as reactor canponents, steam piping, etc.
WHEEL DESIGN AND MATERIAL SELECTION Modern wheels manufactured by G.E. are made from vacuum poured NiCrMoy forgings which are heat treated to obtain the desired strength, ductility and toughness.
The chemical composition and strength level of wheel forgings are chosen to best meet the service requirements of the wheel. To optimize stress corrosion resistance, tensile strengths are maintained at the lowest levels sufficient to adequately withstand operating stresses. Optimized material chemistry and processing results in wheel materials having superior toughness properties and thus superior resistance to brittle fracture.
In the bore region of shrunk-on wheels, induced stresses principally result from interference between the wheel and shaft, and the centrifugal forces of the buckets and wheel.
Figure 7 shows the variation of tangential bore stress with rotational speed. The total stress, which is the sum of the shrink and centrifugal stresses is reasonably constant up to the speed at which shrink is lost.
Centrifugal stresses hay ( been minimized in General Electric wheels by optimizing the wheel shapes and mounting only one row of buckets on each wheel. Reducing centrifugal stresses lowers the magnitude of shaft-to-wheel interference necessary to maintain shrink at normal. operating speed. Thermal stresses induced by steady state and transient steam conditions are generally small in comparison with shrink and centrifugal stresses.
In built-up rotors manufactured by G.E. and other manufacturers, shrunk-on wheels are keyed to the turbine shaft. This assures that even if shrink happens to be lost during transient operation, the wheels will not rotate relative to the shaft. Since keyways act as stress concentrators, local stresses in a wheel keyway are higher than nominal bore surface stresses. The magnitude of stress concentration is dependent on the size and shape of the keyway, which differs from manufacturtr to manufacturer.
ULTliASONIC TEST DESCRIPTION The wheel bore ultrasonic test searches for radial-axial cracks in the vicinity of the wheel.boreand keyway of shrunk-on wheels (Figure 8).
Due to the complex wheel geometry, every suitable wheel surface is used to ensure maximum inspection of the wheel bore and keyway. Tests are made from the wheel webs, hubs,
and hub faces when accessible as shown in Figurg 3,10, and 11 Twin pitch-catch probes are used from the wheel webs and faces aim a single probe technique is used from the wheel hubs. Various beam angles are used to project the ultrasonic beam I
. to the bore from the test surface in a way designed to detect radial axial defects.
By varying the transducer positions in the manner of Figure 12 the axial length of the bore is tested. To obtain maximum coverage, each wheel may receive as many as forty individual scans with twenty in each opposing circumferential direction.
On each scan, proper operation is monitored and sensitivity checks are made by measumment of the keyway signal aglitude.
Figure 13 shows a block diagram of the standard test setup. The rotor is rotated slowly within the casing using either the turning gear or an auxiliary turning device. Automated manipulator ams, which are mounted on the horizontal joint, position the transducers at precise locations and angles which have been pre-selected to provide the best assessment of each portion of the bore.
If they are available, the test can also be performed with the rotor setup on power rolls.
Ultrasonic data is processed by the computer and graphically displayed for operator review. Final results are stored on a magnetic disk for permanent retention.
While developing this procedure, it was found that score marks which are sometimes produced on wheel boresduring the rotor assembly may produce disproportionately large ultrasonic indication amplitudes. A means of discriminating these signals from crack signals was obviously required.
It was found that different test frequencies produced a more proportionate response to flaw size.
Consequently, a dual frequency technique was adopted as shown in Figure 14.
OETECTION OF STRESS CORROSION CRACKS Stress corrosion cracks have historically proved difficult to detect with ultrasonic means. The tightly closed, intergranular cracks are filled with corrodent and corrosion products and usually exhibit poor reflectivity for ultra-sound. Although machined discontinuities were used in the early development of the test method, it was recognized that there was a need to establish the detectability on actual stress corrosion cracks of the size being sought.
An extensive program of produci@ stress corrosion cracks in full size wheels (Figure 15)and large ring specinens (Figure 16) was undertaken to verify the test method. Stresses near yield were generated by shrinking the wheels on to a stub shaft which was heated to further create an additional source of stregs.
The rings were stressed with a hydraulic jack. A 40% NaOH solution at 100 C was circulated through the wheel and ring keyways to produce stress corrosion cracks. The potential of the solution was controlled to accelerate corrosive attack.
To date, a total of three wheels and two large rings have been artifically cracked. The first wheel was exposed to 40% caustic for approximately one year and later broken open (Figure 17 and 18). Ultrasonic indication amplitudes equal to that of the keyway itself were observed from these cracks as shown in Figure 19 and 20. The deepest crack in this wheel was 1 1/4 inch (32 m). Numerous smaller cracks with approximately 1/4" (6.4 mm) depth were also present near the axial end of the keyway and wem detected from the wheel hub.
Another wheel was exposed with the objective of producing smaller cracks nearer to the detection limit of the ultrasonic test. After eight months of
. I exposum, ultrasonic tests of the wheel revealed significant cracking (Figum 21). A section containing the keyway was removed and magnetic particle tested (Figure 22) revealing an extensive network of cracks.
Smaller cracks were successfully induced in two ring shaped specimens maximum 0.18 inches (4.6 m) ques detected these cracks which ranged in depth to a Again, the ultrascmic techni with an average depth of 0.040 inches (1 m).
Figure 23 is a metallographic section showing some of the cracks.
DISCUSSION AND
SUMMARY
Shnank-on wheels manufa::tured by General Electric have established an excellent record of trouble free service. Efforts to minimize operating stresses and optimized material properties undoubtedly contribute to this record. However, laboratory tests and field experience have demonstrated that wheel materials, operating at stress levels found in shrunk-on wheels, could develop stress corrosion cracks in-service. Cracks which might initiate in the bore region could propagate and, if left undetected, could result in a wheel burst.
It therefore is prudent to inspect shrunk-on wheels periodically.
General Electric has developed an ultrasonic test, which can be performed with the wheels in place, and can inspect the critical bore and keyway regions.
In February,1978. TIL-857 was issued recommending inspection practices for 1500 and 1800 RPM nuclear turbine rotors.
It was, and still is recomended, that a wheel sonic inspection of nuclear shrunk-on wheels should be perfomed at about six year intervals. This, in conjunction with 1:aintaining steam purity levels, as recomended in GEK-72281, will significantly reduca the probability of a wheel burst.
I i
1 i
I I
i I
Reference 1.
D. Kalderon, Steam Turbine Failure at Hinkley Point "A", Proc. Instn.
Mech. Engrs., Vol. 186 31/72, 1972.
2.
J.L. Gray, Investigation Into the Consequences of the Failure of a Turbine-Generator at Hinkley Point "A" Power Station, Proc. Instn. Mech.
Engrs., Vol. 186 32/72, 1972.
3.
T.G. McCord, B.W. Bussert, R.M. Curran, G.C.Gould, Stress Corrosion Cracking of Steam Turbine Materials, General Electric Report GER-2883, 1975.
4.
F. Amnirato, G.C. Wheeler, Ultrasonic Inspection of In-Service Shrunk-On i
Turbine Wheels, General Electric Report 79MPL405,1979.
S.
B.W. Bussert, R.M. Curran, and G.C. Gould, The Effect of Water Chemistry on the Reliability of Modern large Steam Turtines, General Electric Report GER-3086,1978 6.
T. Cowgill and K.E. Robbins, Understanding the Observed Effects of.
Erosion and Corrosion in Steam Turbines, Power September,1976.
7.
R.M. Curran, B.R. Seguin, J.F. Quinlan, S. Toney, Stress Corrosion Cracking of Steam Turbine Materials, Southeastern Electric Exchange, Florida, April 21-22, 1 % 9.
(Proceedings were not published and reprints are no longer available).
1 i
l l
t l
SCC DATA LOAD VERSUS TIME TO FAILURE 40%, 212*F-960 mV (Vs. LAZARAN ELECTRODE) 28%.150*F-960 mV (Vs. LAZARAN ELECTRODE) 28%.150*F UNCONTROLLED POTE.NTIAL 1 W !EK 1 MONTH 1 YEAR 2 YEARS I
I I
\\
k o
@o oo o
o 110
\\
\\
100 4
- 48..
O oo o
o oo Cooo o =
\\
\\
\\
b w -
\\
\\
\\
o\\o 8.
es. 4 e 90..
80 -
=
=
\\
\\
i m
5
\\
\\
5o
\\
\\
\\
\\
w
\\
w 1 I lit i
i i I Ii11 1
I I I Ilit i
100 1000 10.000 TIME TO FAILURE IN HOURS FIGURE 1: TIME TO FAILURE VERSUS APPUED STRESS FOR WHEEL MATERIALS IN CAUSTIC. THE SOUD UNE IS THE LEAST SQUARES FIT TO THE 212*F,40% CAUSTIC DATA. THE DOTTED UNES ARE DRAWN PARALLEL TO THE SOUD UNE AND ROUGHLY THROUGH THE MEAN FOR THE O.THER TW CONDmONS TO ILLUSTRATE THE EFFECT OF CHANGING CONDmONS ON TIME TO FAILURE.
^
i l
t
)lp
-"**Sw:
i.~*
1 I
j.s
'I { I.[
~ -
~
p 8ON A
l l
j
,g
.l k
mgL wn lY%w.k.
l
'M.
., ~.
A.
e
~ '
63 ~.. f J:-
..;.-lf-}
.[
lo i
j M.
I FIGURE 2.
APPARATUS USED FOR UNIAXIAL DEAD WEIGHT LOAD TESTING
STRESS CORROSION FAILURE
,, '%f
, h g. a'"
A As i
s-l h
T*r r
4 u.'
(a) 200X
.r i
4 f
g.
l
,q *'
s (b) 500X h v' f
. p ;. _'
Yd Figure 3: Scanning electron micrographs of NiCrMoV sample-fiicrograph (a) shows intergranular SCC near edge.
Micrograph (b) shows transgranular tear at center of micrograph (a).
A
.- - - - -, - - ~ - -, - - - - - - -
j l., -
CRACK GROWTH SPECIMENS l
- 5.,,
df b.
+
i 3
-2
~
3 4*
.s
'l.
Y 1
s-4 "b
,g.,',h
'I11IIi!!!Iiti.l&8l{fI!!4&866 6e g j' i
-L p%.
(g,
e' A*
x.g.
.r..
..g,.
- a..,
s
.7 up.
- g-
'g.
1 i
s
~.a. n,
- .w,.
e.
v.,
s.
'In V..*%
- D I
- ~.
.tf.g.fe-
~fN.
M
.l i
y
~~
~.
k 1 *
- f. ' '
- ygif1;
. ~, f g, < ',' "
ef i
,a m,
l l
l FIGURE 4: IT AND 4T MODIFIED WOL STRESS CORROSION CRACKING SPECIMEN 1
1
.1
i STRESS CORROSION CRACK GROWTH CURVE
=
3 z
z 9
9 o
o E
E E
Ii e
6!
THRESHOLD STRESS INTENSITY KTH l
STRESS INTENSmf (K) i FIGURE 5: STRESS CORROSION CRACK GROWTH CURVE
I CRACK GROWTH RATE IN CAUSTIC 10 1
=.
a*
4 h
O ENVIRONMENT:
40% NaOH w"
V =.900 mV (Ve h
K = 25-46 KSI 1
g g
g
.01
=
150 175 200 225 250 TEMPERATURE (4)
FIGURE 6: MEASURED STRESS CORROS80N CRACK GROWTH RATE (REGION N) V FOR NiCrMoV WHEEL MATEMALS. CORROSION POTWmAL MAINTAINED NEAR O
O I
l WHEEL song gynggg l.
TOTAL I
l
/
=
2 l
l
%g i
i I
(RPM /1000P F4URE 7: SHRUNK ON WHEEL gm a gg l
---m__A_
_o-_
-r
_e.
ma-m.
m._._-.
---_.m-
-a
--a_.__.-
_w w
9 f
yff W
%R$p fft
' ~^
g s '*
l f
5
~~
l
- j'
/
a_id s
' ~ '
{
t~'
\\
\\ j
~
W i
s
'g
~
r l
s
.d NbY
(, ',.
Bt?.,$
~
FIGURE 8
w
Y&
GMAR W4VE TEG7 0FBUCKET WMEl$
MANAL/ RAD /ALRAM
,~
~N[('s DUALTRANSDUCERS
's)p FROM THE WEB
\\
I 6
/
C m
a 3
i I
h y
E % N
~
W o
X 3
g Ri cs %
W
$@3
~
V
$gg gu
~
eQ,A,, m3,l
, z,
p-
/
ri e
FIGURE 10
SMM W4VE TECT OFBUCKET WMELG menayman nm N'x, 's DUAL TRANSDUCERS fCylf.\\
FROM HUB FACE
)
n
/
[
i:l
a wa a
a m
-+m m-w a.
a.-
a-.s--
j i
i i
i i
l W
I I
O
- c...........
i Ou i
y O
E
- .. *
- w...,
i
- ==....,**==.
v-3 1
-- r.>
g Km V=
C *---------*
O we E
{ g........
E FIGURE 12
WHEEL TEST BLOCK D180R8M J
TRANSDUCER O
e PULSER /
H CRT l
C-RECEIVER V.
[,
- \\ T1--
n POSITION SIGNAL COORDINATES PROCESSING I T COMPUTER DISK STORAGE y
a DISPLAY CRT 3 r 5
1 C
TERMINAL 3
n
LP WHEEL BORE TEST 100 FREQUENCY 2 m
/
e0
/
60 AMPLITUDE
(%)
40 20 3
0 E
0.100 0.200 0.E0 E
RADIAL EXTENT OF BORE REFLECTOR (INCHES) g
j
~
i CRACKING A FULL SIZE WHEEL 4
N
' f'
..L..
e.
-m-+=
a,
. 8 **
7.y f... '
t
/.-
/
e
/ ^\\
.,;;~.
\\
4' f
~
't g'
T
~"
% Ah
=
s i
4 FIGURE 15
l
[
CRACKING A 1.ARGE RING SPECIMEN B" W
< {
W]
.1 a-f i
J k
1n l
< 7 'E g
4
- h.. B_
/ 5
- l 5G i
m,.,,
!(
t
.....h h,
l i=
O O
... _...g _.
'I[iE+
"I 6.
.m.
t i
c.
e 4,
FIGURE 16
4 KEYWAY STRESS congag, CRACK d~
,,4 -
r-a
!.t
. 9p,.;,g,:ljt..
p.
yit is.,.
f#
y,
^
g.
% g= p
(
g")j:
k jt n
!A p
h.
a L.
,l k
I l
x I
=
i
~..'.
i. I' 7~.g, Q\\, -
k
-; ' t.,
7,
- ' ?ib:
R i
- +.. l)
_,i
?
h y
1 i
N
- ~
f
,y
,g
'U' 1l0, l. j.g j
2 1
j.
~
i
l l
I STRESS CORROSION CRACK IN KEYWAY l
)
.s T nve. as
. s, I
%k:A-,'.m.,
p.w-w
-.Q.,,.
g
~
c.
3 :
ee s....
y
~
/ g., 'Q. T h:, 'l
' N D.
- 4DW..
rk ;::,?
+..%
)
j
.M.w.
n(.
y.
. s.'... 4 i
f
,W
.t:>
1
- hn.
-[f 1
t:.
- r s
%.+
A:. i f.
(.Q w, hsle 4e,*.;[A,,v.:;,
l r$ *.UTl.3 d
w r
M
' W. g ~ f%h.. g
+ pe,;
g,
%.. h j:
s '.TA.*,
A 4-
++
- p
+
c
_*a
= p.
s
,r, I
7 ~.,..
' 1 7
a[ s ~,.- 5 "h ~..- h A'j
~
4
)
g,,.
44'-
.?.
g.- A-h:
.tt'
- o. !
' d Me.- 3:: :..' 3!.i,"n- -T.x
.1 c.2 2
s f..
{ ~
.,,ff h*e h'
h'QfgP x 4.![.}N
.g
' Y g -~ D '. W. w,,.,
.~
,.f -
. L,,..'...,.
. g _..
c..,
i :r. ',",,.
.,t,.*.:.sp..%..'7
.,3
- g, 1..
4 r
ff' N
. \\_
j
.p yb
% l...
) ::ti.,, (hl ',
N TF-T.h.
fg_.M i
u.
Y
~.
.n. - [49;..'
.f
'1
....?.- y y.
<. f.j ;~y}::,...pl s...
f
.f.
o y,
l s-s we 1
- }
!> 5 2,*t i
- * *
- i ' g * -
- a. ~ k,
,u
. -. w, g;,k'
=.~
a.pggy
.x,
.. ?.
AMk..{4.a.
^*
2
.g
.)*
FIGURE 18
a m,.
-a m,
+
N
-'QT"'
,7,6 TA",,,
- 57. " +.. W"yTr
'IW A FUE 4-
'7"i -,p i
_,c.--
. i. y,g:
.+,,g ;~
- g-Mgy vt.;
- .,.,,eg,t h : r.J.*
sms.k, h* *hi,,
hop w y,,..,..,,;
y 4
p,., s.4
}@.'h;'ll.+i a;
't
~
=
j' Q'"';.
?2 q
y*
g.i ix g, q
- j+ = y %g )gp
~M j
~
~
c n
~
ch;6.
[..!}.'. *
~$
'g
/
W ' d, h
h_%.
@1
% I..
ea j
e y
Jpy g-4 s
1 r
$wm;agQ -pp2 1
s'$ggdMpff_ g p4
~
si m
.,.y. s a
._*?
ff-($
r BER
?.%?!i M!M EEE 2_
d? '9M
& Mag 3EME
+=
p d A y~,
geg3EEEN Etiiila s s a g g g/e M
"i 1.M.2 M$
.3MMEEEE.,a3)g13$
y gd.-
siam
' w.
n..
au av na r, Aq l
$ gap)y1pr=@iN $hih. W l'
A w nw Wah FIGURE '19 l
1
O 9
T e
i d*
8 3 y, W A
- " M' y g*.8; g u:. p r m pr yf R.
Y 9%
/
Jlg
_ g P,<
p-
.e J,
@!a
' s ng
. 2)
,i s,
s g_,
'f f
r u
5 i
e re.
,g. : r.q.
O 4
f g.
e a
.o i~
_.5
-J
- 5
_4 ry r 4*_
- 1 y y$. a.
v.
f j
y.
?
(.
iffQB QQR
.) '
?
p R
7dag 53 p
e ggy gu e lEksk ida
.sa a4.,:'
6 MMP $4_E 555 4lll 9,
7 FIGURE 20
b I
e I
k WHEEL SPECIMEN h
\\
y q. y-4 y g g g g g A.
i y.
- u, y*,
, p+,... m,g.eg;c - My MPs,%
c.
.ju,,.y..
..p
- q
.- qy. 2,.
?..ya.
y y...
10/16/78 CONDITION
- -g W xy %.
' $19?M ?
..'?.,J va?,... [.'..,,. g.a' j.. a.Mk.3 -.. -
\\
m.
2 x.
i
?N..
k n h y ;.J Y
- y
.e y y'...
. s.
e-.
- ..3p. p r.+ e g..
.o
. gc,.
. =vwy... $,-..
a.
a
, - 7:9, p.-
wy,9
,4... g -.,
ce -
~
- v. :... g 3, 9.,;.
....-m.
g.
- 3,-
,...,e d.. + v y.... ;. :
.~
., s e..g.
~
N" r
+
4, 4.
+
'4-r
- h.59 e
+,_
S s.
,e-eq..
6*, 7; w. g]f a 4. - e..
. g; -
a Mjhe
', r - * %. " 1 1j
.5
'D
? ;f.4.s.h. L,. u'"Y T1f$k{f $?$f5f..W P,* $ m.l L;
.y
,y y
BASE CONDITION
^ -
1 -.i
'~
4 4h ~ f {.W@$ ? % WW, g -M, v, gA
' # h4
, g-3 A
.-1
. :c E,.c 3 r.W.M@, J-
-t 3y 2/2/78 r
e K4n
. ~.~ q,.n.J.; A.1 E g -<v o.,<c '
- 5
,,.~ ;-
g:. -;..
- -, - e, c
-+
m,
",.. '.~~l,N f.Th' (^r*j.7,4[-;n,;.j,.,( h
.N,h..l. f.. ' N.7 - ' ;
2
' 8 l
~
M. jf.mys.99 '..
wy. 7t 7
,QA ~ '. ' i. - 1' >.'.h ' ! : '[
d
.. N.,...,.
,p,..<._...'s-
..'g
~e.
..,; ;,. '. ' - A 4 4...
,p r;
,s.,.
..w
% g..... 3 p.
-s
.a}'.. &. :7.4.;' : - m j % p,...
e in *
(
>; W'a_,.., _., ~. > s M."Q Q
'*ig. f 7
- 1 i
1 i
4
-.m.
I i
j l
l STRESS CORROSION CRACKS IN KEYWAY
,,_ T (
'sY
~'
y
'.s m -
gS. '. ry S' z,* - -
57.-
q.~*k_.
,.- :. R_.g; s.
o P' d '.&,%
%;;.m f, h:..a i
- D ', ' a : % '
li-l*
.M.- (%
'f. %
L 4
g d. 4 v -e. dW,s, t.. *-,
,;1 d
- 7. a t,p,y,,.-c
)-
5m
. -e.
N.,s.
%...o
- W n.p 3(.,. +
W..
u.h,Jy;o +,.
~
.x
.,. e. ~.
' 'v..%a,..* 3Nii w:n y;
~,s
. ~
' ': ^
.y n.
2*:
' 34
. ;f 4. <:,mp,:M" v.:.?
1.i
^' @
.,[
~.
k
- .4
- y.. c' y,#4.d;.
$s
? + -p,
', 3,y /-r -
sg f
.4-a h6
. ?.?'b l
fQ?~,[:." q{.-l.'*yL, a-f.> a.
,.e-
.h f
~ *
,4 '.
ff.,' r * -
mu:
. ' &,vl _..h,.?'t.%.
&s., F c *:.
1 n
.e s
.?e. q.;;G,f) 1
- t:
g'N
'.. Ke n
- A r;
r.
3
' ' wg,,//.
j.s
.:. p v.( ;)
-.g d.Sh
.f65M y 7*.9 f$F Si,'s,.-
- .
- ': - l
.,.3
<n74 k-
~
n m]'
Qq:f; m! :.$
- ,y i
-a' dj N
. s,
- w.-
'F'.,*(.c 4.x.g..
'I :
w.j,2-
' '22'
,a
.;.w.
.;4 y-f-
h
'(,,
l$,,"
+
I, j' N-p m
A,Qaf a p-_-
. k.,,3.l,
yns 4-up%,. c..
q
%,,a,p. y
~
- t.?. e
... j s
{.',.,.,j
- t. -
- y lt.
' ~{
'). y pg. ;
a T,4 T-s s 4y r
?..
e
- i.., "
- ~.,..
g.,
- l' j
'ed
,,4
.,. "g.m.
. ~, s,Q -
4:
.3 -.
.,.,3, *,
C. T
- f. 4..
J; w, -
l$r-?J.
%~
.....~s
. w...-.,
.a
.s t l
l FIGURE 22
i l
METALLOGRAPHIC SECTION OF CRACKS l
l
'n
,j
,f 1,
8 s
I i
l i
W1/16 w l
Figure 23. Metallographic Section 8X As-Polished
.i l
l
CEK-46527A, PERIODIC OPERATIONAL TEST SUmtARY
SUMMARY
OF TESTS TO BE PERFORMED DAILY
SUMMARY
OF ACTION FOLLOWING
SUMMARY
OF TEST UNSUCCESSFUL TEST Fully close the main stop valves and combined Shut down immediately by unloading and then valves by sequence teettag at the EHC Test tripping from the EHC panel. DO NOT OPEN Panel.
the generator breaker until ZERO or slightly NEGATIVE load has been reached. The cause of For details see " Flow Control" in Volume III.
the problem should be corrected before restart-ing. In all cases of malfunction, the operator must make his decisions with a thorough knowl-edge of the system and act in the best interests 1
of safe operation and minimizing potential dam-age to the turbine (e. g. Water Induction may be a problem LI an open steam path exists to the turbine).
Test for movement of the extraction mk Isolate the extraction line immediately. For de-valves provided with positive assist Jev.ees, tails, see " Extraction Check Valves", and in-resugate also per " Extraction Check Valves" in For details, see " Extraction Chee. Valves" Volume 1.
In Volume 1.
Check the EHC fluid pump motor current.
Follow prowedure in section IV-D of " Hydraulic Power Unit for Electro-Hydraulic Control Sys-For details, see " Hydraulic Power. Unit for tems" la Volume 1. Take pump out of ' service Electro-Hydraulic Control Systems" in and investigate if required.
Volume 1.
Check the mechanical filter condition indicators Change filter elements if any indicators or on the EHC hydraulic pump suction strait ers gauges show a change is required per section (and auxiliary pump strainer when applicable)
IV-C and IV-G of " Hydraulic Power Unit for and the pressure drop acrose the Fullers-Electro-Hydraulic Control Systetrs", in Voi-Earth filters in the EHC hydraulic system to ume 1.
ensure that all filters are clean and function-ing normally.
For details, see "Hydram!!c Power Unit for Electro-Hydrasile Control Systems" in Vo!-
ume1.
l A
i 5
PERIODIC OPERATIONAL TEST SUMh!ARY.GEK-46527A
SUMMARY
OF TESTS TO DE PERFORMED WEEKLY l
SUMMARY
OF ACTION TO DE TAKEN
SUMMARY
OF TEST ON AN UNSUCCESSFUL TEST Fully test ALL Main Turbine steam valves and Shut down immediately by unloading and tripping OBSERVE the travel of the valve stems and from the EHC panel. DO NOT OPEN the gen-linkages locally. It is recogr.ized on nuclear erator breaker until ZERO or slightly NEGATIVE fueled plants that it may not be practical to ap-Ioad has been reached. The cause of the problem preach the valve during valve testing on a should be corrected before restarting. In all weekly basis due to the high radiation level, cases of malfunction. the operator must make Nevertheless. each valve should be observed his decisions with a thorough knowled.te of the from a safe distance once a week. during valve system and act in the best interests of safe op-testing, to check for changes in noise, vibra-eration and minimizing potential damege to the tion and other behador.
turbine. (s.g.
water induction may be a prob-lem if an open steam path exists to the turbine).
For details, see " Flow Control" in Volume III.
Perform the Mechanical Overspeed trip test at Unioad the machine from the EHC panel. Open the EHC Panel to teet for operation of the Over-the generator breaker when ZERO or NEGATIVE speed trip device and Mechanical Trip Valve.
load has been reached "'an perform the checks outlined in " Trip and Monitoring" in Volume III For detalls, see " Trip and Monitoring" in before shutting down to correct the problem.
Volume III.
Perform the. Mechanical Trip Piston Test at Unload the machine immediately (within one week the EHC panel to test for electrical activation if test on electrical trip is successful) from the
, of the trip mechanism.
EHC panel. Open the generator breaker when ZERO or slightly NEGATIVE load has been For detalls, see " Trip and Monitoring'! in reached then perform the checks outlined in Volume III.
" Trip and Monitormg" in Volume III before shut-ting down to cerect the problem.
Perform the Electrical Trip Test at the EHC Unload the machine immediately (within one week panel to test for operation of the Electrical
!! test on mechanical trip piston is successful)
Trip Valve.
from the EHC panel. Open the generator breaker when ZERO or slightly NEGATIVE load has been For details, see " Trip and Monitoring" in reached then perform the checks outilned in Volume III.
" Trip and Monitoring" in Volume !!I before shut-ting down to correct the problem.
Perform the " BACKUP OVERSPEED TRIP Go through the trouble shooting scheme in " Trip TEST" at the EHC panel to teet the 2 out of 3 and Monitoring" in Volume III. Shut down should logic circuits.
only be accomplished after unloading at the EHC panel. The generator breaker should not be For details, see " Trip and Monitoring" in opened with any load on the machine.
Volume 111.
Continued on page 4 i
3 l
L
l GEK-46327A. PERIODIC OPERATIONAL TEST SUMMAftY
SUMMARY
OF TESTS TO BE PERFORMED WEEKLY (CONTINUED)
I
SUMMARY
OF TEST I
SUMMARY
OF ACTION TO BE TAKEN t
ON AN UNSUCCESSFUL TEST Perform the Power load unbalance test at the Reduce load to under 40% maximum unit load or EHC panel to check for correct operation, go into the standby mode before replacing the power-load unbalance board with the factory spare For details. see " Rate Sensitive Power load per " Rate Sensitive power load unbalance analog unbalance analog and logic circuits" la Vol-and logic circuits" in Volume III. When in ume M.
Standby, load should be limited to 50% of maxi-mum unit load on units with trip anticipators and l
80% of ms.ximum load on units without trip antici-pators. Individual units may have a higher per-missible load. Consult with your General Electric Representallye for this load point.
Test the Thrust Bearing Wear Detector for Investigate immediately and reset or repair satisfactory trip points and operation, within one week. While the device is out of ser-vice, avoid maximum load and switching in or For details. see " Thrust Bearing Wear Detec-out of feedwater heaters.
tor Testing" and " Thrust Bearing Wear Detec-For details. see " Thrust Bearing Wear Detector" l
La Volume 1.
tor" in Volume 1.
Test automatic starting of ALL motor driven Investigate and correct immediately malfunctions pumps by actuation of their pressure switches, of all DC motor driven pumps. Investigate and and exercise each standby pump.
correct within one week malfunctions of all AC For details, see " Automatic Pump Starting m t r driven Pumps. For details. see " Auto-Weekly" in Volume 1.
[
Pump StartW NW.
m at
' Test for alarm annunciation on the oil tank Investigate immediately and repair within one level gauge, week. Check oillevel once per shift. Replenish For details, see " Oil Level Gauge Testing" n rmal lents as necessary. Forms see in Volume 1.
.. Oil hul Gauge TestW.
i Check the air gap on the silver brushes in the Replace the silver brushes and/or operate per front standard for wear and wear rate.
" Removable shaft grounding device" in Volume 1.
For detalls see " Removable shaft grounding device" in Volume 1.
Perform the EVA test if early valving is Replace with the factory spare per "Early valve
- provided, actuation analog and logic circuits" in Volume III.
For details, see "garly Valve Actuation Analog and logic circuits" in Volume III.
Check thst the air dryer on the hydraalle power Reactivate or change the desiccant immediately, unit has active desiccant.
For details. see " Hydraulic Power Unit for Electrohydraulic Control Systems" in Volume 1.
l 4
i
I IN-SERVICE INSPECTION OF 1500 & 1800 RPM NUCLEAR TURBINE ROTORS
~
TIL-857 dated 2/17/78 i
PURPOSE l'
The purpose of this technical information letter is to give recem-mandations for inspecting all 1500 and 1800 RPM nuclear turbine ro-i tors and, in particular, to announce the availability of a newly de-veloped test for sonically inspecting shrunk-on turbine wheels.
The mechanisms for initiating and/or growing cracks in nuclear shrunk-on wheels are also described, with particular reference to stress cor-rosion and recommendations for steam purity.
INTRODUCTION Nearly all of the turbine-generators produced to date by the Large Steam Turbine-Generator Department for use with nuclear reactor cy-cles are tandem-compound units with a rotation speed of 1500 or 1800 RPM.
These units are constructed with integral rotors (rotor mach-ined from a single forging), and/or built-up rotors. (shaft with shrunk-l on wheels and couplings).
In most nuclear turbines the HP rotor is of 1
integral construction and the low pressure rotors are of built-up de-sign, although there were a few early exceptions to this general con-figuration.
The recommended inspections to be conducted on nuclear turbine rotors, the available inspection techniques, and the recom-mended intervals for such inspections are described below.
1 The Large Steam Turbine-Generator Department made plans several years ago to develop a means of inspecting the critical regions of shrunk-on nuclear turbine wheels.without removing them from the turbine shaf t.
This development is now complete.
We now have available the capabili-ty of ultrasonically insoecting for cracks in the vicinity of the key-way and the bore of nuclear wheels with the wheels in place.
These critical regions have previously been impossible to inspect without removing wheels, using available nondestructive tests.
INTEGRAL ROTOR INSPECTION We recommend that nuclear integral rotors be given a thorough wxternal inspection at each outage when the rotor is exposed.
This inspection should include a complete magnetic particle test of all external sur-faces, including rotor, buckets, packings, journals, and couplings.
Normal visual inspections should also be conducted at this time.
We recommend a more complete inspection of the rotor at apprpximately 10-year intervals.
This should include magnetic particle and ultra-sonic inspections from the rotor periphery and from the bore.
A son-ic test of the wheel dovetails on each stage should also be perform-ed at this time.
BUILT-UP ROTOR INSPECTION Ouring any outage.when a turbine section is open, the built-up rotor should also be given a thorough inspection.
This inspection should include a complete magnetic particle test of all external surfaces, in-L
In-Service Inspection of 1500 & 1800 RPM Nuclear Turbine Rotors (TTL 857) - Continued BUILT-UP ROTOR INSPECTION - Continued cluding shaft, wheels, buckets, packings, journals, couplings, and gears.
Last stage erosion shields should be given a red dye inspec-tion, and all finger dovetail pins should be sonically inspected.
We recommend a more extensive test at about 6-year intervals, includ-ing the ultrasonic inspection of tangential entry dovetails, and an ultrasonic inspection of the shrunk-on wheels.
The inaccessible wheel bore and keyway surfaces, and the material in the vicinity of the bore, should be inspected using the recently developed ultrasonic test described in the following paragraphs.
The ultrasonic test of the bore and keyway regions must be conducted with the rotor being turned slowly at a constant speed.
Specially designed ultrasonic transducers positioned on the hub and the web of the wheel are used to transmit ultrasound toward the bore.
If a crack is present, a portion of the ultrasonic energy is reflected, ei-ther back to a dual transmitting / receiving crystal assembly, or to a receiver located at the appropriate location of the wheel.
An analy-sis of the reflected signal is made to determine whether a crack is present.
The material within 2 or 3 inches of the bore, including the keyway, is inspected by continuously varying the location of the transmitting and receiving crystals.
Rotor turning can be accomplished in some cases with the turning gear.
In other cases it may be necessary to make special provisions or modi-fications to achieve the required speed.
The exact speed requirement and related recommendations for the specific machine to be tested will be furnished prior to the outage.
The turbine owner may wish to pur-chase a set of powered rolls.
These could afford the added advantage of permitting the rotor to be tested away, from the turbine, so that other maintenance can be accomplished concurrently.
The I&SE Service Engineer can provide the functional description of such rolls.
The expected elapsed time for the wheel bore ultrasonic test is about five days for the first rotor and four days for each additional rotor inspection performed sequentially at the same site.
This time does not include that required to prepare the wheels for the testing -
cleaning the wheels, removing grease, rust, loose scale, etc., to per-mit close coupling between the ultrasonic transducers and the surface.
i The IsSE Service Engineer can discuss the required cleaning, and how it may be best accomplished.
The internal portion of the wheel away from the bore, which is not in-spected during this test, is of much less concern.
This is because a flaw of unacceptable size and location in the wheel, as manufactured, 1s unlikely.
Th, manner in which the wheels are forged essentially precludes the possibility of producing an internal crack-like defect in the plane normal to the maximum stress (the axial-radial plane).
Furthermore, all modern nuclear wheel forgings are sqbjected to a stringent 1004 volumetric ultrasonic inspection at the time of manu-
In-Service Inspection of i 1500 & 1800 RPM Nuclear Turbine Rotors (TIL 857) - Continued
)
i BUILT-UP ROTOR INSPECTION - Continued
)
i i
facture.
All nuclear wheels are spin tested during manufacture at 20%
overspeed, which further minimizes the probability of having an unde-tected crack or crack-like flaw with a critical size which would lead to spontaneous propagation at normal rotation speeds.
Thus, the likeli-hood of a modern nuclear wheel entering service with an unacceptable defect is low.
It therefore becomes more important to concentrate on the potential for the initiation and growth of cracks in service.
There are two possible mechanisms for initiating and/or growing cracks in nuclear wheels in service:
1.
Stress Cycling.
2.
The 11kelihc 4 of initiating and/or growing a crack due to the stress cycling assoclited with starts, stops, or load changes is small.
The variation in stress amplitude resulting from operating transients is too low to produce significant crack growth, in the unlikely event that a defect exists in the wheel when it is placed in service.
Thus, the major source of concern with respect to the in-service initiation and growth of cracks is that associated with stress corrosion.
Although we have not observed stress corrosion (or other) cracking in the bores of any of our fossil or nuclear wheels made of modern material, the pos-sibility of initiation and propagation of a stress corrosica crack at the bore of a shrunk-on wheel cannot be entirely discounted.
We and other turbine manufacturers have experienced stress corrosion cracks in the wheel dovetail region of integral rotors made of essentially the same material as that used in nuclear LP wheels.
For modern GE design nuclear shrunk-on wheels, the material crack size in the rim region which would lead to wheel bursting is very large, approaching the axial thickness of the wheels.
Thus, such cracks, shoulc. they exist, would have a high probability of detection by surface inspections.
This is not the case in the bore region where, because of higher stress levels, a crack could grow to a dangerous size prior to breaking through to an accessible surface.
The ultrasonic test permits the inspection of this region of the wheel.
A great amount of development work on stress corrosion cracking has been conducted, and our understanding of this phenomenon is much im-proved, although still inadequate to predict the precise time required for crack initiation and the rate of crack growth in a corrosive en-vironment.
Stress corrosion crack initiation and growth is a complex process, in-fluenced by many factors such as material pro)erties, stress levels, en-vironment,etc. and there is still considerabla uneartainty about th2ir interaction.
Data generated to date show a great deal of scatter on
-~
.s In-Service Inspection of 1500 & 1800 RPM Nuclear Turbine Rotors (TIL 857) - Continued BUILT-UP ROTOR INSPECTION - Continued both crack initiation times and growth rates, so that it is impossible to specify absolutely " safe" inspection intervals to preclude the pos-sibility of initiating and growing a crack to critical size between in-spections.
Recognizing that periodic inspections at reasonable intervals cannot provide absolate protection against a wheel burst, we nevertheless be-lieve that,such inspections will greatly reduce the probability of such After having considered this and other factors, we con-occurrences.
clude that inspection should be conducted at about 6-year intervals, as described above.
l The inspections can be coordinated with reactor refueling schedules and/
or sectionalized maintenance plans.
TURBINE STEAM PURITY j
We believe the control of steam purity is the most positive way of pro-tecting against stress corrosion cracking.
Numerous studies have been j
made over the years to determine realistically achievable staan chemi-J stry, and attempts have been made to relate impurity levels to the stress corrosien susceptibility of turbine materials.
While much work remains to be done in this area, the attached instruction, GEK-63430, describes our judgment on the approach which we currently feel is workable and prudent.
Wo are not at present recommending a general inspection program for fos-cil turbine shrunk-on wheels.
Ws may recommend occasionally that certain wheels be inspected, depending on specific circumstances.
Requests for inspecting fossil wheels will be honored to the extent of our inspection capacity, but priority will be gir.-an to nuclear wheels.
The information furnished in this technical information letter is offered In view of this by General Electric as a service to your organization.
and since operation of your plant involves many factors not within our and since operation is within your control and responsibility, knowledge, it should be understood that General Electric accepts no liability in negligence or otherwise as a result of your application of this inform-ation.
t
l i
l l
INSTRUCTIONS csx.:2281 I%Y' l
I (New information, Ausst 1979)
I
\\
i STEAM PURITY - STRESS CORROSION CRACKING n
ek ).l}O
- \\
t
\\
9v
'- Q
-)
7 O,u L
\\
o0 B
These instructiorw do not purport to cone all details or veristoons in equroment nor to provide for every possoble contingency to be met in connection wrth installation, operation or meintenance. Should further j
informetion be denied or should particuler problems ense which are not covered sufficient!y for the purchaser's purponen, the metter should be referred to the General Electric Company.
7
l l
GEK.72281. STEAM PURITY - STRESS CORROSION CRACKING CONTENTS PAGE I.
G EN E R AL D ESCRIPTION.................................................... 3 -
II. OPER AT*0N AL RECOMM END ATIONS............................................ 3 A. Once.Through Steam Supply Systems............................................ 3
.3 B. Drum Type Steam Supply Systems...................
.4 C. S te am ?un ty M o nitonng..................................................
111. M AINTEN ANCE RECOMM ENDATIONS........................................
...5
.5 A. Tur b ine Deposits............................................
i I
I I
l t
I I
i l
2
STEAM PURITY - STRESS CORROSION CRACKING. GEK.72281 1.
GENERAL DESCRIPTION remain dissolved in the steam and pass into the turbine. For these systems. the Utilities have always controlled boder water water punty input to the boiler is a good chemistry to prevent corrosion and depoesta measure of the output steam punty. We m the boder, which can result in tube fadures, conducted a water chemistry survey from and to prevent deposits in the turbine,which 1975 to 1977 in which questionnaires and decrease unit output and lower efDciency.
plant visits were used to assess current m.
Sporadic instances of strees corrosion cracking dustry practices related to feedwater treat (SCC) in turbines indicate tha*,,in addition to ment, botler water chemistry and steam steps to prevent boiler corrosion and turbine punty measurements. The survey results deposita, the water chemistry must be con.
from 50 once.through steam generators trolled to prevent the intruduction of corro.
indicated that about 80% of the units con.
save contaminants into the turbine which can tinuously monitoring sodium and cation cause SCC.
conductivity of the final feedwater achieve typical values of 3 ppb or less sodium and The moet serious corrosive contaminants are 0.2 pmho/cm or less cation conductivity.
caustic, chlorides, and sulate (which decom.
In those units reported to have been oper.
poses into hydrogen, sulade). Due to power-ated withm these limits, no major stress ful concentrating mechamsms operative in corrosion cracking incidents have occurred turbines and the aggressive nature of corrosive and only a minor amount of pitting cor-contaminants in high concentrations, it is rosion has been observed.
necessary to restrict these contaminants to very low levels in the steam. The substitution In recognition that turbine operators need of hydrazine for sodium sulfite as an oxygen some margin on the feedwater chenustry scavenger has essentially eliminated problems limits to accoun' for start.ups, shutdowns, due to sulfite. The elimination of chlorides and system upsets, we have adopted the and caustic is not as easy. Chlorides are almost followmg practical steam purity recom.
always present in the condenser cooling water meridations which should provide adequate and condenser leaks permit chlonde to enter protection from senous SCC incidents:
the condensate stream. Caustic may be pre-sent intentionally from chemical a<*Jitions to R k mommended Mcit: ne stam punty the botler or unintentionally fivm improper be mamtomed at the lowest practicallevel operation and/or regeneration of condensate of containants not to excud 10 pph N<a polishers or make up domineralisers.
and catson conductivity of 0.2 umho/cm unns n mcilopm ns;dunng cabnomcil The steam punty required to prevent corro-pm n, for shon penods not exueding sive deposits in utility turbines is not pre-100 hours per incident and accumulating sent'y known. However, correlations between 500 houn or im in 4: 12 monM opmung Said service experience and utility water time,6.0 pp6 Na and 0.5 umho/cm should chemistry practices has enabled the General not be exmded; and dunns megency Electric Company to formulate steam purity
- n IHon8 / r pend 4 24 houn orlese guidelines that, if followed, are likely to avoid with occumulati n n t exceeding 100 major SCC incidents. These guidelv are oun in a 12 month opmuns Hme,10 0 described in detail below.
ppb and cation conductivnty of 1.0 umhol em should not be exceeded.
II. OPERATIONAL, RECOMMENDATIONS B. Drum Type Steam Supply Systems A. Onos Through Steerp Supply Systerns A major difference between drum type Minimialns the level of feedwater con-units over once.through doetgns is the drum taminants la estremely important for once-boiler's ability to separata dissolved solide through type boilers since essentially all from the steem due to the otrong affinity the impurttles dissolved in the feedwater of the solids for the liquid phes..
3
GEK.72281. STEAM PURITY - STRESS CORROSION CRACKING There is a much lower incidence of SCC C. Steem Purity Monitoring on units operated with drum type boilers.
Since these boilers generally are not well Most of the senous mstances of tur-mstrumented, it is not possible to relate bine corrosion damage for both once-this better performance to steam punty.
through and drum type botiers are aso-ciated with accidents or upset conditions.
For example, ie once through systems.
We would expect that drum botters oper.
improper regeneration of deep bed po-ated on the zero solids or all volatile treet.
lishers operated on the ammonia cycle can ment system would readily meet our introduce caustic into the feedwater. In recommended limits on sodium and cation drum type systems, high drum levels, conductivity listed above fotonce through foaming, or defective steam separator steam supply systems and these limits baffles can significantly increase the should be adhered to.
amo mt of carryover. Operation of any boiler turbme system with severe conden-Some drum type boilers may be operated ser leaks can eventually introduce chlondes with a precision type control in which into the turbine. Avoiding such instances caustic is added to the boiler water to requires constant attention to condenser achieve the desired pH. Any carryover leakage, demineralizer effluen. punty, and would result m the mtroduction of steam punty.
caustic, a very CorrCaiVe Contaminant, mto the turbine. In order to minimize The measurement of steam punty in units corrosion damage to the turbine, drum operated with drum type boilers is not as boilers operated with precision control straightforward as for once through types should deliver steam to the turbine that because of separation in the drum and the meets the sodium and cation conductivity need for steam sampling.
Limits recommended above for once-thmagh systems.
Steam sampling techniques and instrumen.
tation for drum boilers should be used to For drum type units operated with the provide assurance agamst the type of coordinated phosphate boiler water treet-chemical upsets desenbed above.
ment, it is not evident what levels of sodium and cation conductivity are achiev-Saturated steam sampling at thedrum may able in the steam. The lower incidence of be more readily accomplished than super-sonous corrosion damage for such units heated steam sampling. Although the ab-suggests that steem purity levels are com-solute values of steem punty measure-parable to thoes found for once-through ments can be inaccurate because of boilers or that deposits containing corro-nonrepresentative steam sampling, such sive contaminants are buffered by phoe-steem purity momtonng is extremely phates. Sodium phosphates are not be-useful in detecting trends or step changes lieved to be corrosive to turbine matenals-in the chemical carryover.
It is possible that the steam chemistr/
limitations on units unmg coordinated To prevent the introduction of corrosive phosphates do not need to be as stringent contaminants into the turbine we recom-as those recommended for once-through mend that sodium and cation conductivity systems and industry programs need to be be monitored. Controlling the sodium to established to determine appropnate limits low levels insures that the corrosive com.
for units using this type of treatment. For pounds sodium hydroxide and sodium these roecons we are not specifying steam chloride are controlled. Limitmg the ca-purity limits for drufa units using coor-tion conductivity is intended to provide dinated phosphate but we do recommend a measure of protection against some of monitoring the steam purity, and careful the other potentially corrosive contarni-attention to feedwater control.
nants. In the event that a reliable low level 1
STEAM PURITY -STRESS CORROSION CR ACKING GEK 72281 chlonde analyzer suitable for power plant contaminants have been meroduced mto use becomes available, we would strongly the unit. These deposit analyses provide recommend its use for steem purity the information required forlogical recom-monitoring.
mendations regarding the nondestructive exammation of critical turbme compo.
nents and for formuisting corrective ac.
111. MAINTENANCE RECOMMENDATIONS tions to eliminate the source of con-taminants.
A. Turtnine Deposits We would recommend that turbine de.
During turbine inspections the unit posits be taken and analyzed dunny every should be carefully inspected for de-inspection. The results should be reviewed posits. Analyses of turbine deposits can with LSTG Engineenng through Product provide an euly warning that corrosive Service.
0 l
l l
t
(
t l
S
e 9
i GENER AL $ ELECTRIC G
D
O
~
CEMEAI. ELECTRIC - DGESTIC NUCLEAR TURS1NE-GEMRATOR UNITS WITR C.E. BOILLW WATER REACTORS OPERATING 08 UNDER CONSTRUCTION TUR31NE STAGES SERVICE CUST M E STATTON/ UNIT RATING TYPE REHEAT DATE Commeewesich Edieos Dresdes 1 210 TC2F-38 4/15/60 hreer Castral Power & Light Oyster Creek 1 641 TC6F-38 2
9/23/69 Niagere Mohawk Power Ries Mile Pt 1 620 TC6F-38 2
11/9/69 Commonwealth Edises Dresdes 2 410 TC6F-38 4/13/70 Northeast Utilities Milletone Pt 1 650 TC4F-43 11/29/70 Northera States Power Meeticello 1 543 TC4F-38 J/5/71 Comonwealth Edieos Draedes 3 810 TC6F-38 7/22/71 Commeeveelch Ediese Q 4d Cities 1 010 TC6F-38 4/12/72 Com onwealth Edisce Quad Cities 2 810 TC6F-38 5/23/72 Bostos Ediese Pilaria 1 653 TC4F-43 7/19/72 VIraset Tashee 7t.Nuc.Pomer 1 537 TC4F-34 9/20/72 Tennessee Valley Authority Browse Ferry 1 1098 TC6F-43 10/15/73 Philadelphie Electric Peachbotton 2 1098 TC6F-43 2/16/74 towe Electric Light 4 Power Arecid 1 See TC4F-38 2
5/19/74 Teasessee Valley Authority Brovee Ferry 2 1099 TC6F-43 8/28/74 Philadelphie Electric Peachbott:a 3 1098 TC6F-43 9/1/74 Coor81e Power Match 1 809 TC4F-43 2
11/11/74 PASNT Fitspatrick 1 850 TC4F-43 2
2/1/75 Caro 11ee Power & Light Brusewick 2 849 TCAF-43 2
4/29/75 Tennessee Valley Authority trouse Ferry 3 1091 TC6F-43 9/12/76 Caro 11am Power & L18ht Braswick 1 449 TC4F-43 2
12/6/76 Georgio Power Esteh 2 817 TC4F-43 2
9/22/78 Ca noewealth Edieos LaSalle 1 1147 TC4F-38 2
Shipped Pennsylveste Power 6 Light Susquehenea 2 1085 TC6F-38 Shipped Pennsylveste Power & Light Sesquehemse 1 1045 TC6F-38 Shipped Comouveelth Ediese laSalle 2 1147 TC6F 38 2
Shipped Long Island Lighties Shoreham 1 847 TC4F-43 2
Sh1PP*d Cleveleed Electric 111minaties Ferry 1 1253 TC6F-43 2
ShiPP*d 1111 note Power C11stos Power 1 985 TC47-43 1
Shipped Wiegare Mohawk Power Wise Mile Pt 2 1166 TC6F-38 g
Shipped
%1f Statee Utilities Elver Beed 1 998 TC4F-43 1
Shipped Cleveland Electric 111minaties Perry 2 1233 TC6F-43 2
To be shipped Ph11edelphie Electric Linarick 1 1092 TC6F-18 ShiPP*d Mthern ladiana Public Service Saily Nuclear 1 684 TC4F-38 2
To be shipped Puelle Service Electric & Gee Hope Creek 1 1118 TC6F-38 Shipped Public Service ca. of Oklahome Black Fou 1 1180 TC6F-43 To be shipped bit 3 cat se Utilities elver Seed T 994 TC4F-43 1
To be shipped
- ublic Service Electric & Ces liepe Creek 2 1110 TC6F-30 Shipped Public Service Co. of Oklahoma Black Fou 2 1180 TC6F-43 To be shipped Philadelphie Electric Limerick 2 1092 TC6F-38 Shipped 111& note Power C11ston Power 2 985 TC4F-43 1
to be shipped I
i
o
~
CDERAF, ELECTRIC - DOMESTIC WCCtJAR TCR81NE-GENERATOR UNITS WITH FRES$URIZED WATER REACTCBS OPERATileC OR UNDER CONSTRUCTION TUR8tNE STAGES REACTCR SERVICE CUSTDM ER STATION /tTNIT Raft 3C TTPE RIHEAT MFC DATE Wxtlaston Public Power Reaford Sta 2 422 TC4F-43 CE e/18/66 Useki:ston Fublic Power Beaford Sta 1 422 TC4F-43 GE 6/12/66 Duke Power oconee 1 A87 TC6F-38 2
SW 5/6/73 canha Public Power District Ft. Calhoes 1 481 TC4F-38 CZ 8/23/73 Duka Power ocomme 2 847 TC6F-38 2
8W 12/5/73 tattimera Ces 6 Electric Calvert 1 890 TC6F-38 2
CE 3/6/74 Met rope Liten Edison 3 Itile tele 1 837 TC6F-34 SW 6/19/74 Duke Power Ocesee 3 893 TC6F-38 2
SW 9/18/74 lediana Michigan Electric Cask 1 1089 TCLF-43 1
W 2/10/75 northeast Utilities Millstone Ft 2 481 TC4F-43 2
CE 11/9/75 Psrtland Co eral Electric Trojas i 1178 TC6F-36 2
W 12/22/73 Talads Edison Devia-Besse 1 925 TC4F-43 2
BW 8/19/77 12/26/78 Arkanese Power 4 Light Arkaasee Nuc 2 943 TC4F-43 2
CE South Caro 11em Electric 6 Can Sumeer 1 954 TC4F-43 1
W Shipped Cameusere Power Midland 2 852 TC4F-43 2
SW Shipped Duke Power Cateerbe 1 1205 TC6F-43 2
W Shipped Publis Service Co. of N.N.
Seatrook 1 1197 TC4F-43 1
W Shipped Com: users Power Midland 1 SOS TC2F-43 2
gg Shipped U2 ton Electria Celieuey 1
!!92 TC4F-38 2
W shipped
- trisona Public Service Pale Verde 1 1359 TC6F-43 2
CE Shipped Kamees Ces & Electric Wolf Creek 1 1192 TC6F-38 2
W Shipped Duks Power Catsube 2 1205 TC6F-43 2
W To be shipped Georgia Power Vestle 1 1137 TC6F-38 i
W To be shipped Public Service Co. of 5.5.
Seabrook 2 1197 TC6F-43 1
W Shipped Arizona Public Service tela verde 2 1359 TC6F-43 2
CZ To be shioned nocnettsr Gee o Electric Sterling suc 1192 TC6F-38 2
W To be shipped T*1eds Edseos Davis Resee 2 914 TC4F-43 2
8g To be shipped To be shipped Northern Statee Power ces 1 1194 TC6F-38 2
g
- Ariscos Ahlic Service e
Verde 3 1335 TC6F-43 2
CE To be shipped Duke Power Cherokee 1 1341 TC6F-43 2
Cg To be shipped Tennessee falley Authority Tellow Creek 1 1339 TC6F-43 2
CE To be shipped sortumet Utilities Millstone 3 1209 TC6F-43 1
W Shipped i
Teoressee valley authority Tellow Creek 2 1339 TC6F-43 2
CE To be shipped Unios Electric Ca11ameT 2 1192 TC6F-38 2
W To be shipped Tsleds Edisco Devia lesse 3 914 TC4F-43 2
8W to be shipped Duks Fower Cherokee 2 1341 TC6F-43 2
CE To be shipped
, Long teland Lighting Jamesport 1 1196 TC6F-43 1
W T3 be shipped Georgia Power Vogtle 2 1137 TC6F-38 1
W To be shipped Duka Power Cherokee 3 1341 TC6F-43 2
CE To be shipped Long Islead Lighting Jamesport 2 1196 TC6F-43 1
W To be shipped 1
l 1
l i
)
I 1
7
~
REQUEST FOR If1 FORMATION RELATED TO TURBINE DISCS SITE SPECIFIC GENERAL QUESTIONS - To Be Completed in 30 Days _
I.
Provide the following information for each LP turbine:
A.
Turbine type B.
Number of hours of operation for each LP turbine at time of last turbine inspection or if not inspected, postulated to turbine inspection C.
Number of turbine trips and overspeeds D.
For each disc:
1.
type of material including material specifications 2.
tensile properties data 3.
toughness properties data including Fracture Appearance Transition Temperature and Charpy upper steel energy and temperature 4.
keyway temperatures 5.
critical crack size and basis for the calculation 6.
calculated bore and keyway stress at operating design overspeed 7.
calculated K c data l
8.
minimum yield strength specified for each disc II.
Provide details of the results of any completed inservice inspection of LP turbine rotors, including areas examined, since issuance of an operating license. For each indication detected, provide details of the location of the indication, its orientation, size, and postulated cause.
Provide the nominIl water chemistry conditions for each LP turbine and III.
describe any condenser inleakages or other significant changes in water chemistry to this point in its operating life.
IV.
If your plant has not been inspected, describe your proposed schedule and approach to ensure that turbine cracking does not exist in your turbine.
V.
If your plant has been inspected and plans to return or has returned to power with cracks or other defects, provide your proposed schedule for the next turbine inspection and the basis for this inspection schedule, including costulated defect growth rate.
VI.
Indicate wl, ther an analysis and evaluation reqardinq turbine minile.
have been pe. formed for your plant and provided to the sta ff.
[r such an analysis and evaluation has been performed and reported, please provide appropriate references to the available documentation.
In the e vent that such studies have not been made, consideration should be given to scheduling such an ection, u
i
- GD;ERIC QUESTI0'is - To Be Comoleted in 30 Days Describe what quality control and inspection procedures are used for I.
the disc bore and keyway areas.
Provide details of the General Electric repair / replacement procedures II.
for faulty discs.
What immediate and long term actions are being taken by General Electric III.
What to minimize future " water cutting" problems with turbine discs?
actions are being recommended to utilities to minimize " water cutting" j
of_df s,qs ?
--- ~
Describe fabrication and heat treatment sequence for discs, including IV.
thermal exposure ~during shrinking operations.
i L.
A -