ML20086B568

Submits Addl Comments Re Valve Nozzle Failure & Recommendations for Followup Actions.Apparent Departure from Recommended Valve Installation Practices Warrant Thorough Review of Main Steam Safety Valve Nozzle Problem

ML20086B568
Person / Time
Site:	Robinson
Issue date:	05/26/1970
From:	Lofy R PARAMETER, INC.
To:	Seidle W US ATOMIC ENERGY COMMISSION (AEC)
Shared Package
ML19289B536	List: ML20086B507 ML20086B516 ML20086B524 ML20086B528 ML20086B568 ML20086B573 ML20086B591
References
NUDOCS 8311180037
Download: ML20086B568 (8)

Text

.. 1 iI.-

O ran cesaAL USE ONLY

( Pagn of 6 fardmeter, bn c.

Consulting Engineers

• Design - Analysis - Development 13545 WATERTOWN PLANK ROAD, ELM GROVE, W15 CON $1N 53122 784-7380 May 26, 1970 Mr. William C. Seidic, Senior Reactor Inspector United States Atomic Energy Commission Division of Compliance - Region II ,

230 Peachtree Street - NW Suite 818 Atlanta, Georgia 30303 Copics: Mr. L. Kornblith Mr. H. R. Denton CO-Hg.

Reference:

AEC Contract AT(11-1)-1658, Task A Consulting Services PAR : 69-70A Assignment DC-67

Dear Mr. Seidle:

Supplementing our report DC-67, this l etter contains some additional comments on the Robinson 2 valve nozzle failure and a number of recommendations for Compliance follow-up .

actions. -

1. In our report we pointed out that failure to consider the valve " popping".-thrust load was an apparent short-coming in meeting the general design requirements of the Power Piping code, B-31.1.0 as far as the sizing of the nozzle on the main steam lines was concerned.

Beyond this,the installation appears to violate the T.

recommended practice of the valve manufacturer in s _ ,_ ,,,.

number of ways:

a) Typically, the valve manufacturer (Crosby W1've and Gage Company) specifies the minimum cross

/ ti " f "*ld r*9"ir'd * " ""* hi" 7/9/'/d **tablishing the inside and outside diameters

"^1"* bY of ppW es the valve extension at the point of attachment.

j As these valves, in_a fossil plant, would normal-ly be mounted on a steam drum or heavy section

,/ nozzle extension, the weld joint would essentially N

0311180037 700922 PDR ADOCK 05000261

[QR -Q$G)/\[  % h fp endix B t,0 cA tic /

.o c

Pags 2 of 6 fa ra m eler, bn c.

- C O N S U LTIN G ENGINtits ELM OROVE, Wi$ con $lN Mr. William C. Seidle May 26, 1970 develop the full strength of the valve body if wolded up to full thickness.

Reference:

( WE are enclosing copies of pagos 6 &

7 of Crosby Catalog 402-SC which show typical mounting arrangements and dim-ensions for their HCA Series valves.

At the time of this writing, we have not yet received a bulletin on the HCU Series used at Robinson.)

From the figure and notes on Page 7 of the abouc reference, you will note that the "B" minus "A" diameters represent the minimum attachment cross section reconmonded by the manufacturer. If the HCU-65W valvo connection is similarly specified, we might conclude that the attachment weld ( to a stub nozzle or steam manifold) should occupy the full thicks.ess of the valve wall ( the 6.065" Dia.

to the 9" Dia. at the taper shown as approximately 0

45 on Attachment #2 to DC-65). Please note the following sentonces on page 7 of the reference:

" "B" dimensions are recommended minimum to withstand stresses induced by pressure and by bending moments due to reaction when valves arc discharging. Schedule pine must not be used for safety valve nozzles on boilers be-cause of insufficient wall thickness.8#

(underlining added)

We will check the recommended attachment area when further design information on the valves is re-ceived and follow up with a definite interpretation.

b) Note from the figures on pages 6&7 of the above reference, that recommended practice is to mount the discharge stack structurally isolated from the valve. As we understand the design j the stacks at Robinson are welded to the flange of the discharge elbow cantilever mounted to the valve body and no external support is provided. This feature, while not recommended, would in itself not be to serious if stress levels were kept low.

~ .

o o Page 3 cf 6 farbm eler, bn c.

CO N S U LTIN G ENGINEERS ELM OROVE, WISCONSIN l Mr. William C. Seidle l May 26, 1970 1

I

c) The bulletin for the HCA S eries limits the moment l arm "L" from the valvo centerline to the discharge conterline to 24". The limitation for the HCU Ser -

ies will be checked and reported along with the recommended weld area discussed above. We under-stand from a telephone discussion with Mr. G. W.

Rainmuth, CO:Hg., that the offset dimension for the subject valve installation was quoted in the meeting at We stinghous , Pittsburgh, as 30".

It is normally good engineering practice, when specify-ing a manufacturer's product which is required to pro-vide warranted performance (such as the safety valve),

to obtain their approval of the application. In the i course of your further investigation, it might be im-portant to determine if Ebasco obtained Crosby's ap- h((3 proval of this and other applications of safety valves used in the Robinson-2 and other plants. WE feel that the conclusions of our report DC-67 and the apparent departure from recommended valve installation practices described above warrantsa thorough review of the eval- k untion and disposition of the main steam safety valve l nozzle problem and other similar valves in nuclear safety related systems by the AEC Compliance.

2. It is assumed that a formal report of the failure will be submitted by the licensee together with his pro-posed repair or modification program. It is recommend I ed that his evaluation of the cause of failure be re-viewed in terms of the stress related probability of failure described in our report DC-67, and in light of recommended design practice as well as for the metal-lurgical considerations.

From the presentation by Westinghouse personnel in the May 15 meeting, I got the impression that a thorough and comprehensive metallurgical report would be submit-ed. It also appeared that the conclusions of the metal-lurgical report would probably be essentially negative l as far as the cause of failure is concerned. '

,e ,, .i.

.- j

'l Pags 4 cf 6 fa ra m eler, bn c.

CON SULTINo ENGINEERS e un onovt, WISCONSIN Mr. William C. Seidle May 26, 1970 To the writer it was not obvious at the meeting, that Ebasco had plans to report formally on their stress analysis as part of the entire failure analysis program.

We feel that such an analysis would probably turn out to be the salient part of the failure investigation; and if the licensee does not intend to present a com-bined stress-metallurgical evaluation, he should be encouraged to do so. The analysis should not rule out the valve discharge or " popping" mode of loading.

Convincing evidence must be developed to show that the valve did not pop, if this conviction is maintained.

The analysis in our report was prepared using assumptions and speculations as to the thrtat loads and specific dimensions of the discharge pipe. In a review of the applicant's analysis we would expect to find specific loads, dimensions, pressures, and temperatures and the references therefor.

3. Two possible contributors to the stress field in the vicinity of the failure would be the residual stress in the wcid (not stress releived) and the restraint of the backing ring (fused to the root of the weld).

Their effect should be treated at least qualitatively in a comprehensive analysis of the failure. The type and size of the backing ring should be described and i

the welding procedure reviewed in this regard. The location of the split (if any) and tacks for the back-ing ring would also be of interest.

i f WE have concluded on the basis of our trial analysis, and

' the non-authenticated information obtained during the May 15th meeting describing the present installation and the l circumstances of the failure at Robinson-2, that there NO exists a high probability of further failures with this design. Of course the surface condition and geometry of A the nozzles ( in the transition area), (presence of under-cuts, stress risers, etc. ), the ccndition of individual welds (and backing rings), the materisis properties of the

g g Pags 5 of 6 fa rbnn eler, bn c.

CONSULTING ENGINEER $

sta GaovE, wasCONSIN Mr. William C. Seidle May 25, 1970 installed components, ambient conditions, handling loads experienced,and many other indeterminants are all factors which could determine whether a valve fails.. The number of times a valve pops will certasiy increase its likelihood of failure. For these cyclic loads at high stresses, low cycle fatigre might become an initiating cause of failure. In any case, the connection appears to be i stressed much too high under the discharge condition to '

have any confidence in its realiability.. It should be pointed out that our trial analysis showed the high-est stressed point to be adjacent to the valve trans-ition section section just above the weld. The section just below the weld where failure was experienced was stressed just slightly lower.

No doubt a "fix" will be proposed which will either rein-force the valve attachment and/or provide additional a external support to the discharge pipe. The new design should be supported by a thorough analytical treatment and be reviewed by AEC Compliance. If the valve is welded to the doubler plate on the steam line, care should be i taken in the detail design to assure that a full pene-tration connection of sufficient area is made to the l steam pipe itself. The reinforcement of the opening i should be checked and structural discontinuity stresses l between the valve and the steam line calculated. In

! other words, it is not obvious that merely eliminat-ing the nozzle and/or providing a heavier attachment weld will assure an adequate design. A fundamental des-ign analysis for pressure, thermal, seismic," popping",

wind and weight loads must be made as a minimum check-off to validate the new design.

We hope to follow up on this memo with some published-i dimensional information and thrust data on the valves l

shortly.

I

-o o

Pag 2 6 of 6 a r a rn e le r, bn c.

CONSULTING ENGINttes ELM OROVE, WISCONSIN Mr. Willian C. Seidle May 26, 1970 If you have any questions or comments on our report or the contents of this letter, or require additional investigation in any specific area, please do not hesitate to contact the writer or Mr. Walter Foley at our office at any time.

Very truly yours, P 4ETER, Inc.

\, .

a.-

C -

Richard A. Lofy, P.E.

Consulting Engineer l

DC-67afile

7

. .- q A /

Foreword rating for Safety valves is only 90% of the actual Safety Valves are intended to open amt close with- flow.' The total pressure drop shouhl not exceed 2%

in a narrow prewure range; therefore imtallatiom re- of the valve act pressume to avoid rapid cycling of

• quire careful and anurate planning both as to inlet the valve or chatter. The wall thickness of the inlet and diuharge piping. The higher the operating pres- piping mmt be heavy enough to sesist bending mo-sure and the greater she vahe capacity, the more ments due to reaction when the valve discharges. See critical becomes the need for proper design of, the Reaction Charts, pages 13 - 16.

installation. .

Discharge Piping inlet Piping The discharge piping from Safety valves shot!!d Safety Valves should always be mounted in a be equal in size to, or larger than, the nominal valve vertical position diratly on noules having a well- outlet and shouhl be as simple and direct as possible.

rounded approach that provides smooth, unobstructed llroken connections with gencrous clearances (see il-flow from the vessel or line to the valve. A Safety lustration) are prefeired. Suitable Drip Pan Elbows ...

Valve should NEVER be installed on a fitting having are shown on page 28, Catalog 402. The discharge pipe an inside diameter smaller than the inlet connecion above the Drip Pan should be farmly anchored to the to_ the valve, or on excessively long fittings.

building structure and should be sired adequately to The pressure drop ouurring in the inlet piping avoid " blow back" of steam from around the Drip '-

between the valve and pressure soune should be com-puted at actual flow of the valve. It is well to re- Pan into the boiler room when valve is discharging.

Design comiderations are included in an article member that the ASME Iloiter Code (Section 1) "Sidng Ven't ~i liping for Safety Vahes" by Mr. Max

9' FIXED SUPPORT "'. P"

' '" """$' W "E A g ANCHORED TO DUILDING suomng m ober $1. Qre pap, sng extends

. STRUCTURE to a considerable length, it may be necewa_ry to in.

i ,

crease the pipe size at the upper part of the stack.

t

' If Maxim Silemers are used to reduce the noise of

['

r. f

• the discharging Safety Valves (see page 29, Catalog

(

. ( , %ys ;,L [ ) ._ " t." 24" MAX. .

t

} 402), there will be no irurcased back prenure imposed

' due to the Silencer.

s[' } ,

Where flexible hose connections are used in place

( DISCHARGE i

!_. [ f

~

of or in conjunction with Drip Pans, they must be

( -

l

, of sufficient length and flexii>ility so as to absorb ex-c' M]Q DR:P PAN $ kr . parbion and impose no sir.iim on the valve outlet hw n -

T- p( when the prewure venel is at normal operating

.,. 1 pressure.

,g g4- ) .

Main

' ] .; 'l [ f 4 1 1 Where solid piping must be used, expansion bends should be employed to accomplish the same

,,O' CROS8Y Am-( ,

(

t 3

i f ptirpose,

{ Vl= Reaction Force grcm3 9- g in Pounds NOTE: AHow sufficient Drains are located in the valve bodies and should

] i space to prevent Q "Jg,j',

be piped to a drainage system to remove condensate

• S-Shutest posait>le "

from the valve bodies. Separate drains are required pen on the dis-

1. ** ' ' ^ Recommended to drain the Drip Pan and on certain types of installa-fr$"not to e"x. *i"i m"

• b[Su$ INdY. tions additional drainage from the discharge piping '

y

• 'di -

cced 2% cf valve tE"*" '""

'[',

tic,ns of expan-must be provided.

r Hydrostatic Testing of Boilers ,

L _ -@g j Ilydrostatic Testing of Iloilers should be com-s.

Y T t

• ?

1%OUNDED SMOOTH [ pleted before flanged inlet Safety Valves are installed, the boiler openings being blanked off.

,k, , . ,L, LENGm j ff " w

, CROSIW Style IICA Valves for supercritical pres- '

> Sures with welded inlets are shipped with Ilydrostatic  ;

Recommended installation Test Plug and O. Rings for facilitating hydrostatic w G testing of the boiler. This also applies to all welded

' inlet

%A.,e - bmi P31s "R". nrife ild W HUN R'cheai f 9!dhe't' a't t k

.: Nr.ho. vue 3La dar. liv /s, -) _J

s O -

)

c o

SCOPE OF DESIGN

, ._. c . ._ re,Nw '

i " ' ' ' ' :J ' *W ! _, ,

~

,c, i C l'JH' C T n

' '- ., , 3.g ; -. g.~  ! 'g . , g ) ,

i T .

,,_ { g , , g , 7;, p : {g .,,; y

.#l l I l , . , .

4'00 PSIG Amt 5000 l'SIG PHESSURE CL ASSES - WEl DED INLET ONLY 450DT6000 HCWiDF HCATISW G 03 03~ TWelded 4"-300  ;

4500& 5000 HCA 117W HCA 11CW H" 0.785 2" -Welded ' 6" 300 ,[ 4 f .[

~

4500 & 5000 HCA 1'17W HCA 118W J 1.237 2" Welded G"-300 " ,.

4500 & 5000 HCA 117W HCA 118W - g K? 1.833 2%" Welded 6"-300

1. 4, j ;

4500 & 5000 HCA 117W

• HCA 118W K' 2 2.545 2%" Welded 6"-300 ic -

4500 & 5000 HCA 117W HCA 118W M- 3.00 3"-Welded 8"-200 Y4 4500 & 5000 HCA 117W HCA 118W ~ M2 3.07G 3" -Welded 8"-300 4 '

, NOTE: P,otective Ho kis are availabie for att vaives for oEtdoor l> oiler service. See Caialog 402, sage 27 '**

For Capaceiv Tables see pages 18 thru 22 For Dunensions and Materals see belovv and p ,ge 8.

GENERAL. DIMENSIONS Unches) AND WElGHTS (Lbs.)

l  ;  ! tm .;

4 A

- F I *4 *l i '

[ .u ,g l p; i[ -

5 ,

e , [ ~D ~ fy* uv, I t: 3 r

""  % l {

- i ,- ,- -

e . .w Is n-in ,ir) m ';i, Jr.i !!

4500 PSIG Amt 5000 PblG PHESSUHE CLASSES: INLET - WELDEO ~ OUILET - 300 LB. USAS HF 6 */ ' 4500 & 50'00 2 4 4% 4% 8 6% 33% 40%

% 11 .190 ,

? 4500 & 5000 2 5 5% G, 11 8% 41 ' 48  % 313 c 3G04

]a .r E 4500 & 5000 2 , 5% 6 6 11 8% 44 51  % ,

,13 'i 410 .'

+ Mu 4500 & 5000 2% 5%- 7 7% 12 10 00 ' 70  % 14 N 660'

!~"'

s'

, , . 4500 & 5000 2% 6 7 7% 12 * '10 60 70 '"% ' '14 #

"680I 4500 & 5000 3 7 8 8  %' 2 000 I 13 ,; 10% G2 72 ;10 4500 6 5000 3 7 8 8 13 - 10% 62 72  % 10 030 i FIXED SUPPORT

, - ANCHoHED To 8UIL ING Body and inlet material: Style HCA 117W, A217 WCG, sinucmRE r'....

%, m a..

i s-i,

, Style HCA 118W, A217 WC9. Materials for temperatures over 10500 F on application.

3 "B" dimensions are recommended minimum to with-M gI(;;[ g . stand stresses induced by pressure and by bending Cs _

L~ wvax j ' moments due to reaction wh6n valves are discharging. '

, l Larger values of B may be used if desired. Schedule pipe CJ'  : *pst not, be used for safety valve nonles on boilers DISCHARGE -

hQ, PAPE N l because of ir,suf ficient wall thickness.

- m; DRIP PAN - "A" dimensions may b.e supplied %" oversize where!nlet

' C, ,O piping as such that excessive pressure drop can occur in piping so that the ring 1.D. is not less than "A" H --.

DRAIN dimension.

• ( I

( CROSBY --m Charts showing outlet reaction forces are available on

( l pages 13-10.

k The moment arm (distance between vertical Center line P-

'_-Ti of valve and of outlet pipe) should be as short as

, E possible, and not greater than 24".

Q v LJ A n -+- 77

. Mu tuw a.-_-- -