ML20084E331

Ro:On 741227-28,cracks Discovered in Feedwater Spargers. Caused by Flow Induced Vibration.Predominant Failure Mode Appears to Be Fatigue Cracking.Corrective Actions Include Liquid Penetrant Exam & Sparger Replacement

ML20084E331
Person / Time
Site:	Dresden
Issue date:	01/27/1975
From:	Stephenson B COMMONWEALTH EDISON CO.
To:	James Keppler NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION III)
References
1071, 43-75, NUDOCS 8304150020
Download: ML20084E331 (13)

Text

{{#Wiki_filter:y Yl'g C:m] National Plaza. Chicago. Ilhnojs0 .wnith Edis n / mj One Fv Address Reply to: Post Office Box.767 Chicago, linnoic S0690 BBS Ltr #43-75 Dresden Nuclear Power Station. - R. R. #1 7 - Morris, Illinois 60450 \\ January 27, 1975 .N s;3(}t'.l~,..y/.,

• * )

,\\ .- \\ l-! f. {i

• [ f'j.l}q [;

Mr. James G. Kcppler, Regional Director Directorate of Regulatory Operations-Region III sce U. S. Nuclear Regulatory Comission 4 b Q'[ 799 Roosevelt Road c,, jy / Glen Ellyn, Illinois 60137

SUBJECT:

REFORT OF UJ' JSUAL EVD:T PER SECTION 6.6.C OF THE TE9TICAL SPECIFICATIONS FEEDWATER SPARGER FAILURES AEC DCCKEP Ni2Ghlt 50-237

References:

1) Regulatory Guide 1.16 Rev. 1 Appendix A

2) Notification of Region III of URC Regulatory Operations Telephone: Fr. P. Johnson, 0900 hours c,n December 28, 1975 steport Date: January 2/, ly/>

Occurrence Date: December 27 and 28, 1974 Facility: Dresden Nuclear Power Station, Morris, Illinois IDD:TIFICATION OF OCCUR 3D!CE The cracks discovered in the feedwater spargers constitute an unusual event because they present a substantial variance from perfor=ance specifications contained in the safety analysis report. CONDITIONS PRIOR TO OCCURRDICE The cracks were discovered during a scheduled feedwater sparger inspection -while the unit was in a Refueling outage. DESCRIPPION OF OCCURRENCE A feedwater sparger inspection was performed in response to a General Electric reco=endation (FDI #B7/57145). The inspection consisted of using an underwater TV cer.cra to examine areas including:

1) welds of six inch schedule 40 header pipes to the cleven inch junction box pipe, 2) wcld of janction box to thermal sleeve, 3) contact of bearing bars-to vessel wall, 4) pin engagement with clevis 8304150020 750127 PDR ADOCK 05000237 S

PDR [

1. N 0 1975

i January 27, 1975 l 'M6 Jamns G. Ksppl p ' 1' \\ ends of each sparger and 5) cladded feedwater nozzle blend radii area on the l reactor vescel. The inspection was performed from 4 P.M. December 27 to 4 A.M. December 28, 1974 and was recorded on video tape. The inspection revealed two cracks. The first crack was located on the upper l part of ths right side header pipe to junction box weld area of the south-west quadrant sparger. The crack (see Figures I and II ) appeared I to be relatively straight, extending about 90 degrees around the pipe circum-I ference. The second crack was located on the upper part of the left side header i pipe to junction box weld area of the northeast quadrant header. This crack (see Figures I and II ) appeared to be a little more jagged than,the first crack, near the weld area and extending about 200 degrees around the pipe circumference. 1 It was also discovered that the bearing bars (preload spacer), on the right side of the southeast quadrant sparger and on the right side of the northwest quadrant sparger were not in contact with the vessel wall. No other discrep-l ancies were observed. The existance of the cracks was confirmed by additional personnel. The cracks l were viewed by Comonwealth Edison metallurgists and a copy of the video tape i of the inspection was given to General Electric for their review. I ] DESIGUATION OF APPARENT CAUSE OF OCCURRENCE 1 The TV inspection and the General Electric cold flow tests indicated that the l primary cause of the crackine was that the denien of the feedwatar appegava rendered them susceptable to a condition of flow induced vibration. Installation / l Construction and varying service conditions may also have been contributing causes of the cracking. j ANALYSIS OF OCCURRENCE Eased on the underwater TV inspection, the predominate failure mode appears l to be fatigue cracking. The video tapes cf the sparger inspection were also viewed by General Electric personnel and concensus of opinion was that the Dresden 2 sparger cracks looked very similar to the sparger cracks at another EWR plant (Ref. #2). The cracked spargers st this other EWR plant were metal-t lurgically examined and determined to be transgranular (fatigue cracks). As viewed through the underwater TV camera, the bearing bar on each of two spargers appeared to be out of contact with the vessel wall by a slight amount. Although there are several possible explanations for this, the actual cause is not'known at the present time. This condition has been observed at other EWR plants and its presence does not alter the conclusion that the sparger cracks are caused by fatigue due to flow induced vibrations. Numerous full-scale cold-flow tests (Ref. #1) have been conducted at the j General Electric test facility in San Jose on feedwater spargers of several different configurations to determine the cause of vibration. Results of these testa (see FigureIV) have shown that unstable flow-induced vibration occurs as a function of the following variables:

l 27, 1975 U O, January 'Mrs James G. Kapplem - -The pressure differential between the sparger inlet and discharge. -The average radial cap, which permits leakage flow between the inside of the thermal sleeve and the outside of the feedwater nozzle. -The amount of damping present in the system, particularly at the thermal sleeve-to-nozzle interface. It is known from the cold-flow tests that leakage flow plays an important roll in sparger vibration because the tests show that the sparger will vibrate when only leakage flow is present. However, it is not known whether the radial gap is important pricarily as it affects the amount of leakage flow, or as it affects damping, since gap changes inevitably change both these variables in tests. The importance of damping was demonstrated in a test of the spargers (without proload) in which a lifting force was applied at the tee box while the' point of instability was determined. The tests at the GE sparger test facility show that there is a relationship between thermal sleeve / nozzle leakage and sparger vibration. For the cold flow sparger tests, the vessel nozzle inside diameter at the thermal sleeve fit was increased in increments for the performance of leakage tests and correspond-ing full flow tests. These tests indicate that for a given sparger flow, the leakage is directly proportional to the average radial gap between the nozzle and thermal sleeve. (See figure I). Although the Dresden 2 thermal sleeve was designed to allow only a 0.003-inch racial gap, Ine error cana of the 8thbinty boundaries '(Pigure-IV), the operating pressure drop across the sparger of about 16 nei __, and the methods of construc-tion and installation which could produce a much larger radial gap than specified makes it entirely possible that the Dresden 2 spargers experienced flow induced vibration. The actual radial gaps found on another B'dR unit with cracked spargers (Ref. #2) were significantly greater than specified. The unusual service itnd environmental conditions of changing feedwater flow rate, the temperature difference between the feedwater and the water in the vessel, and the movement of water within the vessel also contribute to the problem by producing thermal cycling, thermal stress, and effects on the damping. This event did not cause any personnel injuries, personnel exposures, or release of radioactive materials. It is concluded that this event did not endanger the health and safety of the public. In October and November of 1972, General Electric and Cc=monwealth Edison jointly performed an evaluation and wrote a report concerning the Dresden and Quad Citics feedwater spargers. The report included evaluation of the consequences of a feedwater sparger failure. This evaluation, given in its entirety in Appendix A, specifically considers enthalpy variation at the core inlet, the blockage of a fuel element, and effect of broked pieces on the core spray header. The evaluation concluded that the consequence of sparger cracking when position is maintained is not of safety significance. Even gross failure of the sparger would not result in an immediate safety concern. An independent safety evaluation of a sparser failure from incipient through complete failure was given in reference #2. In this reference, the safety ~. -~-

1 Mr.' James G. Kappl January 27, 1975 consequences of the following events were considered: 1. fuel bundle flow blockage by small pieces; 2. damage of ECCS core spray line due to a broken sparger falling against it; 3. riamage to the jet pump from a falling sparger; 4 disengagement of the thermal sleeve from the feedwater nozzle; 5 operation of the HPCI sybsystem with a failed feedwater sparger. There are no major changes in the new Design 4 sparger (to be discussed in'the next section) from a safety standpoint. The conclusion is still that the safety consequences of a feedwater sparger failure are acceptable. The cause of feedwater sparger cracking has been essentially determined by the GE cold-flow tests and the special instrumentation on the Design 3 spargers of another E*a plant (Ref. #2). The design features of the Design 4 spargers and their successful operation, as indicated by the special instrumentation, show that a solution has been found for the identified problem. Based on tho above analyses and the successful inspection of the Drcsden 3 spargers during the previous refueling outage, it is concluded that the renewed operation of Dresden 2 and the continued operation of Dresden 3 are justified with the corrective actions to be taken as specified in the next coction. CORRECTIVE ACTION The program for corrective action is to 1) liquid penetrant examine the accessible l portion of the nozzle blend radius of each feedwater nozzle, 2) further inspect the old Dresden 2 spargers and determine the actual radial gap when they are removed,.3) replace all four Dresden 2 spargers this outage with new spargers of the new design (very cimilar to the Design 4 spargers in Ref 2), 4) inspect the feedwater spargers during its next scheduled refueling outage and 5) inspect (including a liquid penetrant examination of the nozzle blend radii) the Dresden 3 feedwater spargers during its next scheduled refueling outage. The salient design features of the Design 4 feedwater sparger are (see figure V):

1) Interference fit between the sparger thermal sleeve and vessel safe end.

2) Forged-welded tee between the thermal sleeve and the sparger headers.

3) Different size and location of exit holes in the sparger.

4) Scheduled 80 rather than schedule 40 304 stainless steel spargers.

The first design feature was based on several tests showing there was no vibra- ' tion under any flow conditions when the thermal sleeve was tightly fitted to the safe end (small radial gap).

' Mr. Ja:nen G.,Keppl O 3a uari 27, 1975 t l The second design feature reduces peak stress levels in the toe by a factor of i 4 due to smaller stress concentrations. This is due to the use of full penetration welds, more unifom sections, and large radii at the junction of the ] header pipes and the themal sleeve. l The third design feature was incorporated to lower the pressure drop at rated i flow from 16 psi for the first designs to 11 psi for the new design which increases the stability margin (See Figure IV). l The fourth design feature will increase the strength of the spargers by increasing the thickness of the pipe walls. l FAILURE DATA i j Previous inspections of the Dresden 2 and 3 spargers did not reveal any sparger da:nage. Sparger datsge has been noted in at least one other B'G unit (Ref 2). j The cracks we.re found in the area where six inch schedule 40 header pipes are { welded to the junction box pipe (see figures I, II, and III). The sparscrs i are made of 304 stainless steel (schedule 40). 4 l The documents references in the letter are: 1. General Electric "Feedwater Sparger Cold Flow Vibration Tests" ) (NEDO-20554, June 197).). ) e. In11 stone Interim steport on Feedwater sparger rallure including addendum 1, 2, 3 and 4; and special report, Chlorido Intrusion i ] Incident. 4 s k 1, l { L 4 i l I i

F i., t re I Old Droeden II Pwec.ra ter Sn :r;.7 rs s D "D f)4er (40h SS) [f ( j ./ s s.'. .%/ ^. / a / NQ ,/ .. f Cracked 'delds --- /,, / ~ ~ /,. / r /. N z j: -t, s. K._

y

,4 c=. 0 'Clevic Pin v. st ~. Vecac1.a11 -y tr,h P'.... Y- !/ /f f 'sl / j ',, N. lend Jadius o Pearing Ear (Spacer) / l ( N. ;' 'l s l. - Veccel *.'oz7.le i,.; ,-s r s L) l Q Jur.ction Box (Tee Box) [,.. e s I $/ N ,)! l 's i: -s t L. N 'lormal L.lceve ..dadial Gip / . Note: . parger arr.s are r.ot ot' couel length e o

_y. .r De.II l ict_u_ren of Crack (Is..' O.~er) flote: These t.wo pictures of 'the saIne crack are mirr'or images of the actual sparger r.nd tt:crofore, do not indicate tr.e correct orientaticr..

• Y

'l _ @Q QS M,M N D**. - - += ~ ~~-,-*e-m- ... g 9 = T i I N ~) I ~ b"* hf.* h '-4.,_. 2 m __m 9

• q q'*y w

99 ppg p_-- _ dlWheWM%thWMM" ~~ - - ~ ~ - - - I i I .) t i t p-d I .., su - 9 l i. . _aaA_.a 11 d-J i a - - mu

d _Fi.m re III Pic *uren c f CracP. ('. .._ - n_.r ar)_ t :(&.,.i c 'ure3 i i of t.hc :x*c cr sch *trr Error i*':". *ca of the ?!nte: Thhne ....,.r... .w 1.,,,.,,......... e e n.. ro.. orientatior.. 9 V **w"W~ ~..-_.-a^.-.--. ^ _ a T*: ~ ~ *-^ a..., -.-- T m3 W -.'M-=}~+^ -} -j.i, _. - L:(-~ ~ M. L ', ~ ~ - Y =.. ~.. =.T -7x.' -- S w .3;. F. --P+2. y. 1 gi % T. 1 m l i.! a, I l l t i 1 4 i s. m es 'e eg., .4 1 M h*h9*w e e e, 44-...m Wt w.m Aw h g '~ = =&=: .- -- - -.. - J =:.-----.--------- ^ - - = -:- r pm- = - F I ' E s b ~ 1~ e t h-h..* E f b 1 1 i

r. a"

{ f n t, I l 'P-%~ N i e + s N 1 N. ~ 4 h

• 3 a4 l

e.ee. = m... %..a

(, Figure IV. Conditions for Sparger Instability Based on Cold Flow Testing _I 20 Unst abl e i 18 ~ 16 Stability boundary for Design 4 sparcer, Indeter.i m te with normal installation ste.bilite C 14 Ely 12 .c C C 10 s i e wc E 8 c) E Stability boundary 6 - based on reduced dec.'r.ing 4 (friction) test series 4 2 Stable 0 e i i i i i i i i 0 2 4 6 8 10 12 14 16 18 20 Operatinn AP, psi

s %.*g O g g 's ? .N, 's s s \\ s \\ -\\ s, s \\ N, \\ \\ i \\. p e. g O ~; P i \\ \\\\ ,p \\ ,4 O \\ r3 C)

• )

U C \\, l'; / s \\ e \\ e, \\ L C J .\\ C \\ H e / ]!e 'N -t g F:\\ ^; M-x a 4, i* -9 me t

e,. % L A

c-

• e.

,o _s; t {

,,iy

,- ~~.... 3~:.._. 1 Q- [ {*-- - .... s / ;jj ][ } 5 \\ t . x.- P +-s,., f =- a 1 O 5< l I o 1 l O Q i o....- e u e \\ o p .,3 i. b I 4 [3 h / j'~~ i oo b r-4 g r-4 i y= l / eC oo xu o~ j n i j.. es f ~ 0 i /e i /.' ( ,.c a ') . z ".. - -./ w

e O 3anuarF 27, 1975 fr.JamesG..KepplO 11-l i APPEIDIX A Consecuences of Feedwater Scarger Failure The safety aspects of Feedwater Sparger failure are addressed as follows: i A. Enthalpy variation at the core inlet. B. Possibility of small pieces of the sparger blocking a fuel element i orifice. I l C. Effect of large pieces of the borken sparger on the core' spray header. l A. Non-Unifomity in Core Inlet Enthalpy l If one or more Feedwater Spargers should partially or completely break, there will be a non-unifom temperature distribution of the coolant in the downcomer annulus. Because of incomplete mixing in the lower plenum, non-uniformity in the core inlet enthalpy would exist. As stated in references (1) and (2) core power assymetries would result. The references also conclude that these assymetries do not present a safety problem. The rationale for this conclusion is that the calculation of MCHFR in the highest power region will be in the conservative direction, i.e., calculated KCHFR will always be less than the actual MCHFR. This is further explained in the following paragraph. If a particular region in the core is supplied with coolant at an enthalpy lower than the average inlet cnthalpy, the power in that region will be greater than i .the average power. Because incore instrumentation indicates actual oower and MCHFR is calculated using the average inlet enthalpy, the MCHFR in the region of higher power will always be calculated at a value less than actual. This calculation always results in a conservative ECHFR and assures adequate margin for all transients and postulated accidents. B. Effect of Small Pieces in the Reactor If the sparger junction box should separate from one of the sparger ams, a small piece could concievably break loose. A concern with a s=all piece is that associated with potential fuel bundle flow blockage. Since the piece must be sucked in through the annular passage around the jet pump no::le, the maximumsizeofthepiecewouldbelessthan2["by2j",byd" thick. The chances of the piece flowing into the jet pump are small..However, for purposes of determining potential consequences, this is assumed to occur. The safety analysis for this piece is considered to be identical to the safety analysis and test perfomed for the Quad Citiesjet pump washer and was used as a guide in performing the sparger piece evaluation discussed below. A detailed study of flow blockage in a E'nd has been made in a GE Topical Report (1) on file in the Public Document Room. As stated in that report, based on analyses of high power density fuel operating at 18.5 kw/ft: 1 I a) It would take mare than a 90% area blockage to cause a MCHFR less than 1.0: therefore, no fuel rod damage occurs. I l m y ,e iy.-e. m y ,+.--,-m = w mw m.m---+ - - + r 1

, Mr* James G. Kepple'T ' January 27, 1975 A,/ s b) If the blockage were more than 905, clad melt and fuel crumbling weald occur. This would lead to high radiation sensed by the main steam line radiation monitors which would scram and isolate the reactor. Offsite doses rc=ain less than 10CFR20 limits. I Based on the infor=ation concerning the maximum size of the sparger. piece, the following conclusions are drawn:

1) Because the fuel bundle orifice diameters are 1.425 and 2.262" and the cimum surface of the sparger piece is 2 5 inches square, it is possible for significant blockage to occur if the piece were carried to the orifice properly oriented.

2) The most likely resting place of the piece is in the reactor vessel on the, f_

bottom of the outer anulus. It may have found its way into the recirculation loop. It is possible for the lost piece to enter the lower plenum through the jet pump no::le and therefore, it could be in the bottom plenum.

3) The fluid velocities in the lower plenum in the vicinity of the orifice region is a maximum of 3 fps and is not high enough to lirt the piece up toward the core inlet orifices. The vertical velocity to suspend but not lift the Quad Cities washcr (2.25 inch diameter) was 5-6 fps as determined by test. However, even this presupposes the area is inserted into the upward velocity field. This can occur only with much hir,her velocities required to lift the washer frcm the bottom of the reactor vessel. The maximum upward velocity in the reactor lower plenum is about 6 fps.

4) If the piece is introduced into the lower plenum during operation. it will have a hign (17 Ips') downward velocity which will drive it to the bottem surface of the plenum. Further, the piece cannot easily, negotiate'the 160*

turn to upwards toward the core. Due to its higher density, it will tend to move radially away from the turn and hence into the lower velocity points at the center and remain there. For the piece to be scooped up would require a much higher surface velocity than actually exists in'the vertical direction. Therefore, it is concluded that the piece will remain in the bottom plenum if it is introduced there during operation or if it is there already.. F

5) The possibility of the piece becoming lodged at the core inlet is considered to be so highly improbable that operation can continue without safety concerns.

There is no way, because of the physical barrier which encompasses the CRD, for' the piece to find its way into the control rod drive itself. C. Effect of Io.rce Pieces on Reactor The feedwater Sparger could fail in various ways. For the purpose of this 4 analysis it is assumed that both arms of the sparger header,poparates frca the ~ ~ junction box but the thermal sleeve remains attached to the' junction box. The I main concern is the possible interaction between the failed pieces and the internal core spray distribution piping. i 'I I

~ I [d.JamesG.Kepple January 27, 1975 iy: 1 If the Feedwater Sparger fails as described above the jet of the feedwater would force the junction box against the shroud head bolts. A lateral movement of the tee box of approximately four inches would result. Since the thermal 4 i sleevo to the junction is 27 inches long, this assembly will not disengage from j the presstire vessel. If the junction box (11 inches in diameter) were properly i ~ ~ oriented it could pass through the space (15 inchos) between two studs. In j this event the Junction box would travel an additional five inches and butt ' " ^ { p .up against the steam separator standpipes. The junction box still could not ' y. separate from the pressure vessel. !"r The Feedwater Sp rger arms and the internal coro spray distribution piping '; (both 6" schedule 40 pipes) are separated 14 inches center to center or approxi-mately 7 inches wall to wall. If the failed feedwater sparger arms should sag i from its restraining brackets and rub against the internal core spray distribution f~ piping, the resultant fretting action could in time wear a hole in this distri-bution piping. However, the wall thickness of the header is 0.280 inches so that acco timo would be required for this to occur. ~

-~. _

)- Failure Detection i The irrnediate offect of a crack in a Feodwater Sparger junction box is a leakage - path 'for feedwater flow. The leakage flow or its effect cannot be detected nor has safety significance unless there is a gross failure. For a gross change a in flow pattern, douncomer flow distribution anomalitics maybe detecbed by a change in the vessel water level indication and/or the core neutron flux 4 instrumentation. The pattern and magnitude of these observed changes are unknown and can vary depending on the type of failure (e.g. one or two ams broken), j enet lecation Of the failurt rclat h tv um instrumentation, anc plant operating - ~ l conditions. In chort, some change is expected but the mcgnitude of these changes cannot be quantified. FOOTNOTES 1 i (1) Consequences of a Postulated Flow Blockage Incident in a Boiling Water Reactor' NEDO-10174. j (2) C.uad Cities has a maximum linear heat generation rate limit of 17 5 kw/ft. J 7 ' _REFFitENCES (1) " Reactor Asy= metrical Neutron Flux Distribution", Dresden Nuclear Power I - Station Unit 2 Special Report No. 6, Farch, 17, 1971. h (2) Supplementary Infomation to Special Report No. 6, February 17, 1972. (h) '" Consequences of a Postulated Flow Blockage Incident in a Boiling Water l Reactor", NED0-10174, Fay 1970. , [ Sincerely c d d 'e ficnson:W j B. B.L ,1 r Superintcnt n V Dresden Nuclear Power Station g BBS:RWC:smp n V l 1 t =-}}