ML19309C563

App 5C to TMI-1 PSAR, Design Criteria for Reactor Bldg. Includes Revisions 1-11

ML19309C563
Person / Time
Site:	Three Mile Island
Issue date:	05/01/1967
From:	JERSEY CENTRAL POWER & LIGHT CO., METROPOLITAN EDISON CO.
To:
References
NUDOCS 8004080792
Download: ML19309C563 (5)

Text

I O

kl APPENDIX 5C DESIGN CRITERIA FCE REACTCR SUILDING 1 GENERAL The prestressed concrete Reacter Building vill be designed to have a lov stra elastic response to all conceivable leads, thereby ensuring that the integrit of the vapor barrier is never breeched. The intent of the design is also to provide a = ode of failure should the vessel ever be tested to destruction of ductile rather than a brittle -a-ner. The design vill be further based upon various ec=binations of factored leads that are based upon factors by which loads are increased to approach the ILnit of an elastic respense. These facters are developed in a sLnilar =anner to the Ultimate Strength Design pre visiens of ACI 318-63 where facters are applied for tnese facters outlined in the "Cc==entary on Suilding Ccde Require =ents for Reinforced Concrete,"

ACI 318-63.

In the case of this design wherein a =cre exact analysis is perfor=ed than conte = plated by ACI 318, the load facters pri=arily provide for a safety

=argin on the applied loads. The Reactor Building vill be analyzed to ensure proper perfor=ance of all ec=ponents includi 5 the liner, cencrete shell, and reinforce =ent under the following leading conditions:

a. During construction but prior to prestressing s_,/ b. During prestressing

c. t nor=al operating conditions

d. At test conditions

e. At factcred leads 2 ME"' ECD CF ANALYSIS GENERAL The shell of the Reacter Suilding vill be analyzed to deter =ine all stresses,

=c=ents , shears, and deflections dua to the static and dynamic loads listed in Section 1.2, Appendix 53, of the original PSAR.

STATIC SCLUTION The static lead stresses and deflections that are in a thin, elastic shell of revolution are calculated by an exact nu=erical solutica of the general bending theory of shells. This analysis e= ploys the differential equations derived by E. Reissner and published in the "American Journal of Mathenatics, Vol. 63, 19hl, pp. 177-lSL. These equations are generally accepted as the standard ones for the analysis of thin shells of revolution. The equations

/~'N given by E. Reissner are based en the linear theory of elasticity, and they

\2 take into account the tending as well as the =e=brane action of the shell.

s 80040807 N0003 001 -

3C-1 (Revised IC-2-67)

O he =ethod of solution is the =ultiseg=ent method of direct integration, which 2 s capable of calculating the exact solution of an arbitrary thin, elastic hell of revolution when subjected to any given edge, surface, and temperature cads. This =ethod of analysis was published in the " Journal of Applied echanics," Vol. 31, 196k, pp. L67 kT6 and has found vide application by =any ngineers concerned with the analysis of thin shells of revolution.

he actual calculation of the stresses produced in the shell and foundation as carried cut by = cans of a ec=puter progra= vritten by Professor A. Kalnins f Lehigh University, Sethlehem, Pennsylvania. This ec=puter progra= =akes se of the exact equations given by E. Reissner, and solves the= by =eans of he =ultiseg=ent =ethod =entioned above. The progra= can solve up to four ayers in a shell and thece layers can have different elastic and ther=al reperties and can vary in thickness in the seridional direction. Applied loads an vary in seridional and circu=ferential directions.

INAMIC SOLUTICU he stresses and displacenents of the response of a shell of revolution to the xcitation of an earthquake can be calculated by superi= posing the nor=al cdes of free-vibration of the shell. The = odes of vibration are calculated y =eans of the general. bending theory of shells derived by E. Reissner. The ranslatory inertia terms in the ner=al, =eridional, and circu=ferential irection of the shell are taken into account. The = ass distribution is the etual = ass distribution of the shell and no approx 1=ations are =ade. E. h eissner's shell theory is such that it predicts exactly the ec=plete spectru=

f natural frequencies of the shell without any approxi=ations.

ne differential equations given by E. Reissner are solved by =eans of the altiseg=ent direct integration =ethod of solving eigenvalue problems, which as published by A. Kalnins in the " Journal of the Accustical Society of

erica," Vol. 36, 196k, pp. 1355-1365 According to this method, the igenvalue problem of a shell of revolution is reduced to the solution of a requency equation which vanishes at a natural frequency. The frequency quation consists of exact solutions of E. Reissner's equations, and no pproxi=ations are =ade.

.e calculation of the natural frequencies and the corresponding =cde shapes f each =cde of free-vibration is perfor=ed by =eans of a ec=puter progra=

ritten by A. Kalnins. The cc=puter progra= has been used for the calculttien f the dynamic characteristics of =any types of shells of revolution and its nsults have been verified with experi=ents on =any occasions (a listing of revious applications is attached). The progra= calculates the natural requencies of any rotational 1y sy==etric thin shell within a given frequency

terval and gives all the stresses, stress-corresponding to a natural requency, consultants and displacements at any prescribed point en the aridian of the shell.

te no=nal =cdes of free-vibration need only be added in order to construct l

e responco of the shell to an earthquake. The relationship between free-l !3 ration and a given excitation is given by the folleving equation:

' ^ '

0003 002 5C-2 (Revised 10-2-c7)

19 d, b) Y(x,t) = Yi(x) Ci Sv1 Wi Ni i=1 where Y(x, ) = funda= ental variables of the response i

Yi (x) = funda= ental variables of the ith =cde

.w Wi = natural freq,uency of the i'" =cde

~u Ni = constant for the i*" =cde Sv1 = =aximu= velocity fran the response spectrum for a single degree of freedc= system for a given value of Wi for the ith =cce j For analysis purposes the Reacter Building shell is divided into structural 1 parts, and each part is divided into a specified nusber of seg=ents. l The Static Analysis rtd Dyna =ic Analysis have been used by the folleving )

cc=panies for the an Aysis of thin shells:

1. Martin Cc=pany - Orlande, Florida

2. Pratt and Whitney - Aircraft, East Eartford, Conn.

3. Central 7'- +-< city Generating 3 card - Lenden, Ingland The Static Analysis has been evaluated by E. Kraus, in Welding Research Council Sulletin, No. 108, September 1965 l

2 The Dynamic Analysis has described and its results ec= pared to experiment ty:

J. J. Williams, " Natural Drcught Cccling Towers - Ferry bridge and after,"

in the Institution of Civil Engineers publication, 12 June 1967.

The nonaxisy==etric leads imposed upon the Reactor Building base slab vill not have a contributing influence upon the design of the shell; therefore, the fcundation slab vill be des'.gned for nonaxisy==etric leads by censidering a circular slab on an elastic fcundation.

t 3 LCADING STAGES l

3.1 CURING CONSS.UCTION SUT PRIOR TO PRISTRESSING The Reactor Building vill be designed as a conventional reinforced concrete structure subjected to dead, live, vind, and construction leads with allevable stresses in accordance with the linits established by ACI 318-63.

3.2 DURING PRISTRISSING d'

The Reacter Building vill be designed for prestress leads and vill be checked Oc insure that the concrete stress vill not exceed 26f'c at initial transfer.

0003 003 3C-2a (Revised 10-2-67)

Stre::ssa dua to shrinkage, creep, and elastic shortening of concrete vill be taken into account, and flexural creep tending to relieve bending stresses O will also be considered. All remaining stresses vili be in accordance with ACI 318-63, Chapter 26.

3.3 AT NORMAL OPERATING CONDITIONS The loads due to nomal operating conditions are:

a. External pressure of 2.5 psig

b. Operating temperature transients

c. Dead load

d. Live load

e. Prestress load

f. Seismic load

g. Snow and vind load The stresses in the concrete and reinforcing steel resulting from these loads will be in accordance with ACI 318-63, Chapter 26. The stresses and strains vill be suen that the integrity of the liner vill be maintained.

3.h TEST LOADS The Reactor Building vill be designed to function under the following loads at test conditions :

a. Internal pressure of 1.15 times accident pressure

b. Dead load

c. Live load

d. Prestress load i

e. Temperature transients at test conditions I

The allowable stresses vill be in accordance with ACI 318-63 Chapter 12 and 26 The vessel vill be adequately instrumented to verify during the pressure test that the structural response of the principal strength elements is consistent with the design.

35 AT FAC"0 RED LOADS The building vill be checked for the factortid loads and load combination given in 'ppendix 5A, and canpared with the yield strength of the structure.

The load capacity of the structure is defined for our design, as the upper limit of elastic behavior of the effective load carrying structural materials.

3 5C-3 -

t; .;

I I - - -

I For steels , (both prestressed and non-prestressed) this limit is considered to be the guaranteed minimus yield strength. For concrete, the yield strength is limited by the ulti= ate values of shear (as a =easure of diagonal tension) and bond per ACI 318-63, and the 28 day ultimate compressive strength for flexure ( f' c) . A further definition of " load capacity" is that deforsation of the structure which vill not cause ecmpressive strain in the steel liner plate to exceed 0.005 in/in, nor average tensile strains to exceed that corresponding to the minimum yield stress.

The load capacity of all load carrying structural elements vill be reduced by a capacity reduction factor (Q) as stated in the basic structural design criteria. This factor vill provide for "the possibility that s=all adverse variations in material strengths , work =anship , dimensions , control, and degree of supervision while individually within required tolerances and the limits of good practice, occasionally may combine to result in under-capacity" (refer ACI 318-63, p.66 - footnote) .

T's allovable tensile capacity of concrete for membrane stresses (i.e. ,

excluding all flexural and thermal stresses) due to the factored loads vill be 3 V fc . The allovable tensile capacity of concrete for =aximum fiber stresses to the factored loads ' including the thermal load plus other secondary effects will be 6 N/IE". Where tensile fiber stresses exceed the allovable, mild steel reinforce =ent vill be added on the basis of cracked-section design. The addition of mild steel reinforce-ment and the increase in steel stress will be determined in a manner stnilar to that contained in ACI 505-5h " Specification for the Design and Construction of Reinforced Concrete Chimneys." The minimum steel on the exposed face of the concrete vill be 0.15 percent of the cross- lh ,

sectional area of the concrete as indicated in Paragraph 15 of Appendix 53 to the PSAR.

The cracking limit of the concrete in principal tension vill be governed by the allowable values of the shear as a measure of diagonal (principal) tension. The allovable shear values vill be as follows:

a. When membrane tension exists or when membrane ec=pression is less than 100 psi, the section vill be designed to the ultimate shear provisions of Chapter 17 of the ACI Code 318-63. Where sheer re-inforcement is recuired sufficient prestressed force vill be pro-vided so that the net membrane force remains in compression or zero tension so as to result in a condition analagous to that covered in Chapter 17

b. When aembrane co=pression of greater than 100 psi exists, the prin-

! cipal membrane tension vill be limited by the ultimate shear provision of Chapter 26 of the ACI Code 318-63.

1 O

t' . , . .

0003 305 5C k (Revised 1-8-68)