ML19344F188

Responds to Request for Comments Re Sequoyah Containment. Forwards Info Re Stringer & Skin Interaction.Concludes NRC Calculations Overestimated Beneficial Effect of Stringers

ML19344F188
Person / Time
Site:	Sequoyah
Issue date:	08/18/1980
From:	Hubbard H R&D ASSOCIATES
To:	Gilinsky V NRC COMMISSION (OCM)
References
NUDOCS 8009120578
Download: ML19344F188 (8)

Text

R O D ASSOCIATES Post Office Box 9695 U. .-'"1

. 0N

. Dl* Marina del Rey, M.T20UDENCE

. 3 California 90291 18 August 1980 Nuclear Regulatory Commission M/S 1149 1717 H Street, N.W.

Washington, D. C. 20555 Attention: Commissioner Victor Gilinsky

Dear Commissioner Gilinsky:

In response to your request for comment on the further calculations of NRC staff concerning the Sequoyah contain-ment, we have prepared the enclosed brief note. Our con-clusion is that the staff calculation has overestimated the beneficial effect of the stringers because the induced stresses in the nearby containment skin have been ignored.

If a better estimate of containment strength is desired, our previous recommendations of a carefully designed set of experiments and calculations still seems the best course.

If there are any questions about the enclosed work, we will be glad to try to answer them.

Very truly yours,

,- l -

• /.jgttt.it(I.V 4 .

Harmon W. Hubbard HWH/dl Enclosure l

1 l

l l

l 8009120 5M f

4640 ADMIRALTY WAY

• MARINA DEL REY, CALIFORNIA 90291

• TELEPHONE. (213) 822-1715

t NOTE ON STRINGER AND SKIN INTERACTION (SEQUOYA3 CONTAINMENT VESSEL)

Introduction:

The NRC memorandum, 07 August 1980, Goutam Bagchi to F. P. Schauer on the " Analysis of Sequoyah Contain-ment Capacity examines the effect of stringers on the contain-ment strength assuming that the stringer can act as a beam with fully constrined ends and a fully plastic moment capacity of the beam cr,oss section.

The previous RDA analysis bounded the calculations between the cases of:

1. Fully rigid rings and stringers, leading to the var-ious panel solutions with critical stress reached at pressures below those of the membrane analysis and in particular with local stress concentrations at the middle of the long edges of the panel.

2. Shell/ stiffener analysis following aircraft fuselage analysis methods and leading to the 27 psi center panel membrane stresses.

It was our opinion that the stringers could not greatly contribute to the membrane strength since as beams they would likely suffer elastic instability (a horizontal beam with encastre ends, 1/2" x 9 1/2" deep and 10 ft long will twist under its own weight). However in view of the Bagchi analysis we have taken a quick look at what might be expected in the interaction of the skin and stiffeners.

Critique: When bending occurs in the stringer the skin must also be considered to be a part of the beam and the beam section can be considered to be a T-section with the skin as the end flange (Figure 1). How much of the skin will be effective is debatable I (it clearly will not be tne whole 48") but for some distance either side of the stringer the stress in the skin will be the

TENSILE y STRINGER

/o r P

NEUTRAL s

AXIS s / KIN l l y

\o 8

f F l i

\

i f t. _t

' ' COMPRESSIV5' EFFECTIVE "Ta FLANGE Figure 1. Fully Plastic Moment Applied 9

I T

4 same as that in the stringer. When the fully plastic =o=ent condition is reached the neutral axis will have descended so=e way toward the skin and for sore distance either side of the stringer the skin will have a compressive stress cc *E"*1 *C that in the stringer; i.e., (ce ) sk., n = (c c ) s. 4 ge. , _,.I (yie.3, co=pressive), and this stress will go across the full thickness.

Now consider the panel assuming Bagchi's condition is reached (Figure 2) .

Stringer - -

N "x "x '#c

~ t ti Pressure = P A 3

/

Ring /

Figure 2. Ele =ent Definition l

l At the skin centre A (Figure 3) :

i ist princ. S 1 =e g 2nd prine. S2"Cx 3rd prine. S3 = P/2 i hence the Von Mises stress at A is given by: -

2 e =c y = c.2 A c X

- c.c X WA  :

l l

l c- .

with a = ,f therefore (cg) =

f. c and (cx) =h fe g O S x= i og A -c 6"82 X

~

S = P/2 % 0 (P < < Sy, S2 3

Figure 3. Mid Panel Element "A" In the skin next to the stringer (Figure 4) , the stress element will be:

1st princ. Sy=og 2nd princ. S2 " "x + C c 3rd princ. S3 = P/2 R: 0 (P << Sy, S2

A X C 0

9 B o g =S y X C S3 = P/2 % 0 (P << S ,S) 2 Figure 4. Edge Element "B" But ( c6 ) B ", 0)A #y and (ox )B " I#x}A " Y and (oc) cI Y B stringer Hence, Sy = C y

Sg =

h - 1, a, S =0 3

and the Von Mises stress at B:

a = S y +S ~ S 2 I2 1

t 2

+ -1

• U y \! - 1 -

s Vi + t i - V! + 1 - t i + Vi

" + +1~

"y\

,,vor

~

= 1.414 o Y

If Bagchi's condition is reached then the skin adjacent to the stringer will have a Von Mises equivalent stress 41.4% in excess of o , which is impossible unless strain hardening occurs.

y Thus, if the stringer support is as great as Bagchi postu-lates then the skin will fail adjacent to the stringer at a lower imposed pressure (Biezeno's Effect) or the. stringer can-not support the skin to the extent that Bagchi postulates and the skin will fail at the centre of the panel at a pressure not much higher than that for our value of pure membrane failure.

' L2 conclusion,. we believe that the value of 27 psi is probably as close as can be obtained by simple analysis and this agrees with Sandia's figure. If more detailed definition is required then our prior recommendations for a detailed finite Biezeno and Grammel's " Engineering Dynamics" Part III, Section 14 states "It is known that stiffening ribs may actually weaken a construction they are meant to reinforce" and an example is detailed to "show how dangerous the design of stiffening ribs may be." In the example quoted a compressive load is induced in a cover plate (membrane) adjacent to the stiffening ribs which combined with the membrane loading leads to an actual stress increase due to the addition of the ribs. They summarize by saying "Although the calculation is only a rough approxima-tion the result proves that the stiffening ribs have in fact weakened the design quite considerably." .

L .

, element code analysis will hold. Further such analyses should be backed up with model tests. The variability in the answers derived by the different simple analyses depends entirely on the assumptions which are made regarding panel edge conditions.

Hence, the model test should probably include several adjacent panels to the actual test panel. If the stringers do take any appreciable bending loads then we must clearly be concerned about local skin bending stresses and a practical test is almost mandatory to i,lluminate this point.

I i

l l

l I

(

' ~

-- . . , _ _ _ _ _ _ . _ _ . _ . . _ . . .