IAEA-126

BASIC STRUCTURAL DESIGN PHILOSOPHY,CRITERIA AND SAFETY

OF CONCRETE REACTOR PRESSURE VESSELS

REPORT OF A PANELSPONSORED BY THE

INTERNATIONAL ATOMIC ENERGY AGENCYAND HELD IN VIENNA,9-13 FEBRUARY 1970

A TECHNICAL REPORT PUBLISHED BY THEINTERNATIONAL ATOMIC ENERGY AGENCY, VIENNA, 1970

The IAEA does not maintain stocks of reports in this series. However,microfiche copies of these reports can be obtained from

INIS Microfiche ClearinghouseInternational Atomic Energy AgencyKdrntner Ring 11P.O. Box 590A- 1011 Vienna, Austria

on prepayment of US $0.65 or against one IAEA microfiche service coupon.

PLEASE BE AWARE THATALL OF THE MISSING PAGES IN THIS DOCUMENT

WERE ORIGINALLY BLANK

FOFW0HD

A panel on"Baeic Structural Design Philosophy Criteria andSafety of Concrets Reactor Pressure Vessels" was held by theInternational Atomic Energy Agency on 9 to 13 February 19?0at Agency Headquarters. A total of 34 specialists representing14 countries and two international organizations participatedin the discussions.

Since the first two prestrsssed concrete reactor pressurevessels at Mar-coule, Prance, were built in 195^» the technologyof concrete pressure vessels for nuclear power reactor applicationhas-developed rapidly and there are now 15 vessels of this type inoperation or under construction» In addition, there are eightconcrete reactor pressure vessels known to be in the planningstage in the United Kingdom, France, U.S.A. and Federal Republic

è '

of Germany. It is also known that several other countrieshave already started very extensive studies and research forusing prestreesed concrete reactor pressure vessels.

Although the problem of vessel availability is not yetcritical for the PWH and BWH systems, it is clear that thepotential for continued growth in unit rating will ultimatelybe limited "by shop-fabricated vessels and that, at that time,either field fabrication of steel vessels or prestreesed concretevessels will bs required. Of these two alternatives, "based onexperience in the United Kingdom end France, and taken into considera-tion, integrated nuclear power station's concepts, prestreesedconcrete vessels now appear to offer excellent possibilities.

In view of the growing significance of this type of vesselin relation to the future of nuclear pov/er plants, the Agency-has organized a Panel on this subject with the"purpose ofreviewing the latest pertinent deveJ opme'nts, ' to facilitate theexchange of informations and, mainly, to discuss the variousaspacts of the design philosophy criteria,'economics and safetyof prestressed concrete pressure vessels as '#ell as to formulategeneral guidelines concerning the -objectives.

It is hoped that this collection of nine papers, togetherwith the conclusions and recommendations which were worked outby the Panel Meeting, will be of interest to reactor designersand to the authorities concerned with the safe working ofnuclear pressure vessels. The texts of these papers have beensupplied by the Panel members and no editing has "been done by theAgency.

The Agency is grateful to the authors of papers, to all theparticipants of the Panel for their contributions to the discussionsand, it would particularly like to ejtpress its thanks to -theChairman of the Panel, Mr. I. Bavidson, of the United KingdomAtomic Energy Authority, for his guidance of the discussionsin the most productive manner.

CONTENTS

. > General;. Analysis for Servies Load Conditions» Limit Design. Ul t i mat e Be s i gn, Subjects of Particular Interest for Further Research or

Dcvelopmewt Study

". . Structural Design Philosophy and Criteria for Concrete HeaetorVessels ~ U.S. Practice 17W. RcckenhauserContribution from the United Kingdom 53I. Davidson

. > Work on Reactor Pressure Vessels of Prestressed Concrete inYugoslavia (>1B. PetrovicProblems and Perspectives of Prestressed Concrete Pressui'eVessels - Franch Experiance up to 1970 o9D. Costes

;3, Teat Stand for the Prestressed Concrete Vessels Containingthe Keliusi Loop 85L.K. Komoli

•J. Ultimate Design, Experience fronn Small Dimension Models Testing 99P. Scotto

V. Report on the Starting of a Coordinated Work Programme forProstresssecl Concrete Reactor Preesxire Vessels in the FederalRepublic of Germany 115T , Jae ger

•'•', Design Philosophy and Criteria of Safety of Prestressed ConcretePressure Vessels ~ Practice in Czechoslovakia 179M. IDavidSwedish Development Work on Prestreased Concrete Pressure Vesselsfor Water Reactors 213S.

J-''22 _»i* Fro^r'i:ftne 243'L3I;2S~JZ> List of P":.ot:i cip'inhs £47

Sumaaary Report ^and Recoamendatiojggof. the Panel Meeting on/"Basic.Structural fiesiign PhirlQsophy_Criteria and Safety of Concrete Heaotor Pressure _Vgssels"

I. Genejral

1.1. Following the discussions of tra Panel which met in Vienna from9 to 13 February 1970 the present, report was prepared' with a viewto setting out the general requirements and principles whichappear to be applicable in this field. Attention is' drawn to theadvantage of farther work in certain areas, and recommendationsare made regarding further international collaboration. Becausethis is a rapidly developing field of study,, the present reportmust "be regarded as provisional.

1.2. Reactor vessels perfora the function of containing the nuolearreactor, the primary coolant, which is normally under pressure,and various othsr cbsponants and equipment essential to theoperation of the reactor. The vessel must perform this duty for

*its design life under all noriaal and foreseeable abnormal conditionsof operation, with a degree of reliability such as to precludeany -unacceptable risk to the public.

1.3. The vessel is loaded by the primary coolant pressure and by theeffect of temperature-induced strains in the various structuralcomponents. The basic principle of a prestressed concrets vesselis that, for a range of predetermined loads, including normaloperation, the concrete is maintained by the tendons in a state ofnet compression across any section of the vessel, The admissiblestate of stress and strain in the concrete may be influenced bypassive reinforcement.

1.4. The prestressed concrete structure may be furnished with featuressuch as penetrations and associated closures, an impermeable liner,insulation, a cooling system, passive reinforcement and means forpressure relief.

1.5» This application of prestressed concrete is novel in many respects}consequently, existing standards are of limited application,Progress is, however, being made in the preparation of nationaland international standards for PCRPVa.

1.6. Early development in this field was associated with gas-cooledreactors (OCRs). The potential, of the concspt f c .- other types,such as liquid cooled and/or moderated reactors, is now appreciated.It is not expected that the basic principles of PCRPV design willbe influenced by the type of reactor contained. Variations inengineering applications must be expected.

1,7» Sfce vessel must be capable of performing its function under certainpredetermined conditions of interaction between the contained reactorand the vessel. It follows that, not only must the vessel providethe necessary standard or' containment under predetermined reactorconditions, but there. must also be an acceptably low probabilitythat vessel behaviour will of itself induce an unsafe state in thereactor.

1.8. One incentive for adopting PCRPVs is the economic advantage to beexpected thereby. This advantage arises, for example, from:(a) the ability to contain large reactors or reactor systems

with the acceptance of high pressures?(b) simplification of a plant layout}(c) the fact that a highly developed steel fabrication industry

is not necessary,

1.9» Attractive features from the safet,y point of view include:(a) physical isolation of the stseJ prestressing tendons and

reinforcement from sources of heat and radiation and fromthe primary coolant;

(b) the high degree of redundancy in the preetressing system;

(c) the possibility in many cases that tendon loads mayba measured and reset;

(d) the possibility of removing tendons during operationfor inspection or replacement;

(e) inherent ability to withstand seismic shock.

1.10. 1-1 some cases the conventional practice of grouting tendons isadopted in order to provide sotae protection against corrosionand an alternative anchorage,, These .advantages must be setagainst the inability to test, inspect and replace individualtendons.

1.11.. Where a penetration as required bo provide access for services,equipment etc , it is customary to provide a purpose-built removableclosure. The design objective in such cases is that the standardof integrity of the closure and its attachment should be at leastas good as that of the main structure? the presence of a penetrationshould not prejudice the necessary integrity of the main structure,

1.12. Zt is necessary to provide some raeans of limiting ths effect on thestructure of an excessive rise in internal operating pressure. Theaiagnitude and rate of a postulated pressure rise can be determinedonly by reference to the characteristics of the particular reactorcontained. It is common practice to provide automatic ventingdevices, such as safety valves, for this purpose. An alternativeis to design the vessel so that it is self-venting by partialstructura] faiïure. It is generally recognized that this alternativecannot be relied upon at present*

1.13. It is customary and advantageous to make provision for verificationof .the state of the vessel. This may be done fey installed instrumenta-tion and/or by periodic in-service inspection» Such measures, whichare taken in a manner appropriate to the particular situation, serveto verify the vessel integrity ana to confirm the design criteria.

1.14. Design philosophy is based on the recognition of two or morephases in -vessel response to increasing pressure. Wie designobjective is to ensure that a particular response to the imposedload can b© achieved in each pbase and that this behaviour isconsistent with the appropriate operational and predeterminedfault conditions.

1.15. Over a range of pressure and temperature, including- normal operatingconditions, the vessel will respond to short-term variations inpressure in an elastic nurimîer, This facilitates machine analysisof stress and strain .in the structure, Longe r-ter-m stress andstrain are affected by shrinkage and creep of the concrete, re-laxation of tendons and possibly fatigue, ïn -fois range of responsethe affects of short- and long-ierifi behaviour can be combined todemonstrate that stresses and strains are limited to acceptablevalues.

1.16. Beyond the elastic range the response becomes increasingly inelasticand non-linear. Tho vessel would not be expected to enter this phaseexcept under the most severe overpressure fault conditions. In thisphase tho structure is stable but may experience permanent damage.It is in this phase of vessel response that some limit states occur,

1.17. The ultimate load condition, in which the vessel is incapable ofsustaining- any further Increase of internal pressure, is a furtherlimit state, Evaluation of tM ultimate load provides a measure ofthe factor of safety above design conditions.

1.18» The methods of design verification which are referred to in paragraphs1.14 to 1.17 are described more fully below* it should be noted thatit is the usual practice to use two different methods of designverification simultaneously.

II. Analyses foi*, Service Load_ Conditio.ns_2.1.Loading conditions

Stresses, strains and deflections in the vessel structureîîhould be analysed for all relevant combinations of mechanicaland thermal loads which can arise under normal service conditionsthroughout its life.

The tsndon forces adopted in each analysis should includeallowances for the most severe effects of friction and lossof prestress. Account should, be taken of all significantloadings applied to the structure, including stresses arisingas a result of construction procedures and normal operatingtransients. Any significant effects of penetrations and -lineron the vessel structure should be taker, into considération,

2.2.J Analysis

The analysis method selected for each loading condition shouldtake appropriate account of time and temperature dependentcharacteristics of the concrete and bsve clue regard for thecomplexity of the design and loading conditions and the accuracyrequired.

Each analysis should provide adequate details of the stressesinduced in the concrete, in the passive reinforcement and strainsin the liner to enable the acceptability of the design to beassessed.

For the purpose of the analyses covering the prestressingforces and dead loads at completion of construction, arid undertest pressure, the concrete may be assumed to be a linearelastic material»

For all other service load conditions, including start-up andshut-dowi both initially and at the end of the vessel life,the stress-strain characteristics used for the concrete shouldtake account of the age, temperature and time under load.

2_,j_._ Minimum design près tressNotwithstanding the acceptability of stresses in the concreteunder normal service load conditions a net cornpressive forceshould be maintained across any section of the vessel under apressure which exceeds the maximum normal service load pressureby a suitable margin.

2.4» Tendon anchorage zone designConcrete supporting tendon anchorages should be reinforced inaccordance with applicable codes v^here existing, 'The safety ofthe vessel structure is specially dependent upon the integrityof the prestressing tendon system. Suitable tests should becarried out on representative tendons and anchorages in combinationunder support conditions representative of those obtained inthe pressure vessel.

2.3. CrackingIt is considered that lire: ted cracking may be accepted provideddue regard is paid to any significant redistribution of the stresseswhich may arise and the integrity and leak tightness of theliner are not impaired.

2.6. Concrete gtrèse concentra.tiensWhere local concentrations of stress occur, due to the presenceof embedments or other discontinuities in the vessel geometry,these should be assessed individually. Where such stresses are

very local and can be shown to "be self-limiting, they may bedisregarded» Where they are more extensive, due regard shouldbe paid to the effects of increased creep rates or tensiJecracking on the distribution of stresses in the vessel concreteand the influence which they taay have on the strains in thevessel liner and the stress distribution arising in the vesselunder conditions of shut-down.

Comment on gialti-a.xj.a.j jîompresslye stresses in concrete

2.7. A certain amount of evidence is available to demonstrate thatfor short-term loading, under multi-axial corepressive stresses,concrete can safely withstand higher compressave stresses thanare generally accepted under uniazial loading. It is generallyaccepted, however, that if one of the stresses approaches zero(or becomes marginal I?/ tensile) the effect can be nullified orreversed. There is a lack of knowledge regarding long-termloading unuer multi-axial compressive stresses.

2.8. In view of the above, it is recommended that for tne time being,if any advantage is to be taken of multi-axial compressive stressconditions?- The loading condition must be short-term only.

The minimum stress value irust ne suown to oe sulstantiallycompressive beyond all reasonable doubt.Careful attention must be paid to the effects of any increasedstrain on the stress distribution un<5er subsequent reduced orreversed load conditions.

Ill,r Limit Design3.1. Five limit states are recognized for the present:

- Limit of instantaneous linear elastic response. Definesthe upper end of the range through which the overallresponse of the vessel to short-term loads remains essentiallylinear and reversible. Minor, localized clacking of theconcrete may occur.

- Limit of instantaneous, reversible overall structuralresponse. Similar to item 1. Defines the upper end ofthe range over which the vessel response remains reversiblealthough no longer linear.

- Limit of permissible deformation (short-term and/or long-term),Represents the largest oeformation under wbjch the containedreactor system wil] still function properly. The limitusually applies to penetrations und other such parts of thevease"1 where relatively close tolerances must "be preserved,

- Limit of" liner defect stability. Defines the upper end ofthe deflection regime in which liner integrity can bereasonably assured. For liner deformations beyond thisrange, such defects as may be present in the liner, maypropagate through the wall. Siraila^jy, highly stressedareas may lead to local liner faj'iure. This could resultin substantial leakage into the vessel concrete, andconsequently, to crack pressurisation,

- Ultimate strengtn limit. , As defjne in more detail inPart IV "Ultimate Design", tftis represents the ultimateload carrying capability of the vssaeT struct ire. It isevaluated wjtnout regard to the credibility or manner inwhich this load might come about.

3*2. The limit of linear elastic response and the limit ofpermissible permanent deformations are considered significantin terras of evaluating normal and upset loading conditions.An upset condition is defined as a transient operatingstate resulting from a single operator or equipment malfunction.Similarly, the limit of reversibility is related to the conceptof (potentially repetitive) emergency loading conditions»For reactor systems where major leakage from the vessel con-stitutes the design basis accident, that is, where linerfailure represents the faulted condition, the lirait of linerdefect stability 'becomes the significant consideration. Thus,the reactor systems design and vessel design 'interact directlyin the assessment of limit states three and four.

3.3. The designer, in evaluating the various limit states of the•vessel should provide suitable margins of safety with respectto the loading conditions and their probability of occurrence,against which the limits are assessed.

3.4. At present, quasi-elastic and visso-elasti.c analysis methods,as well as structural model tests are used to estimate theselimits. Because of the inherent uncertain ties, considerabledesign conservatism results, much of which may be unnecessary.

3.5» Finite element techniques are under development to predict vesselbehaviour in the aneiastic range. The methods consider crackingof the structure, but so far only on an axisymuzetrical basis,Chess finite element techniques should be expanded to threedimensions. In addition, meaningful failure criteria forconcrete under multiaxial loading, and defect stabilitycriteria for liner ftaiberiais, should be formulated. Furthermore,additional materials characterisations work, should be performedfor concrete in the elevated-temperature visco-elastic regime.

IV. Ultimate Design

,3 . Outline

The ultimate design is a. simple method of ensuring that thereis an adequate margin of safety between the design pressureand the pressure at which the vessel will fail. In calculatingthe failure pressure of the Tr3S^el it is assumed that the linerwill remain intact up to the failure pressure and that the over-load pressure is applied tu a new and cold vessel. The safetyfactor chosen should allow for the degree of predictabilityand the gradualness of the failure mode and should he adequateto cover any possible time-dependent deterioration of thevessel materials.

4+J?*, _ Re 1 e van oe

•fbte ultimate load calculations are not in^nded to representa realistic condition. The calculations do not take accountof possible gross liner rupture at a pressure somewhat lowerthan the vessel failure pressure wftich could, result in prematurevessel failure due to pressure acting on the cracked concrete.Keither does the approach consider a possible weakening of thevessel due to time dependent material degradation. The approachdoea not calculate the precise margin available against arealistic condition. It merely gives an indication of the marginavailable against a hypothetical over-preesuriaation under theoondit ': ons assumed aho ve .

4.3» Method of Gal cu] ation

The method of calculation depends on a knowledge of the mechanismsof failure of the vessel. This information is normally obtainedfrom appropriate model tests to destruction.

It is essential that the mechanism of failure is fully understoodand is shown to be progressive.

10

IK the case of flft>ru.ra'i fa i lure , it- is posai r/e to calculate,

with smf f~ cie>i t accuracy, the r>ressure at v,*hlcV tho vesseJwill finally f a l ? ,

If a shear modo of fa i lure ayr l iea • tr» a s 'aY; , the xredic4 '!.bâ !t i tyoT this may be sufTioier.tly u r -ne r tu i r j to require a higher loadf .actor.

The moâel shou'H ix su.*Ti c i s / i ^ i y rspre-'sntati. '^ and its dimensionsso chosen thai th-3 resu'lvT ar-i not diKtcrtec by . inabi l i ty torepr'-jnent adc-.qii.ately such ,<?yeclf*3 feat-are^ as» for inDtaiioe,tendon anchor-aces.

It Is v'iob posa: M ft to scale accurst-''} y th*j l ine r because its'oenaviour (s^ioh as pul, ' out stran^th of ct^dp, etc,} is Jioti inearî y r f t la ted to dimensions,

4.1;, Pressure i« Cracks

The u l t imate approach useô to da te has no i ta'^en I n t o account

gross liner r ' jpvare at pressures lover thu?i tr.e u ^ t i m a ^ e pressure,Cross l iner ructure v/nuld not DO axpectftd before the concretestructure stii '-s to become a mechanism. Th i s gross crackingof the concrete starts close to the f ina l u l t i m a t e ' collapsepressure and bancp the f ina l i 'a ' lurs r^°e:;sure w i l l be only

s ^ i n a J i y effected: but //hei.hei- the vessel •ictual.lj ^ a - l s or'.'y re lessee the pressure w h i J s ' c remaJ n i.ng l;\r^ely intact

wil l de send greatly upon tne extent to winch the cracks becomepressurined and t'^e pressu.-s level at the ti;ao or gross linerruo ture,

This sub;}©-}"*; Is covf-rod by Limi t State (Part T i l )

1.1

It is important to ensure that tb.e design of penetrations is such.t.;;at the penetration and its interaction with the concrete doesnot reduce the load factor of the structure*

12

V'._ . Sublets of ..p_a_yt.i.oulja ijn_tiejr t._ ar]r../urthe]r. researoh ordevelopment study

The Panel has listed the subjects briefly described below, in noparticular order of importance, since it is felt that betterknowledge might result in greater economy with equal safetyor greater safety without increased cost.

Jj ,_ Pe jLvat:lou ojf_ currently._ opera ting PCRPYs

Vessels currently operating offer the best opportunity of checkingcertain assumptions and calculation methods. It would, therefore,be desirable if appropriate data could be obtained and made generallyavailable. Opportunities might arise during- commissioning tests,repeat pressure Lests and inservice inspections the latter should

• provide opportunities for examining tendon relaxation, tendoncondition, linsr and insulation conditions, concrete strains, etc,The instruments and facilities required for such purposes shouldbe considered so that provision may be made in the design specifi-cations; the inservice collection' of the data would be theresponsibility of the operating personnel.The development of instruments having suitable long-term stability -in so far as they are not already available - is recommended ae asubject of further study,

5•2, Concept of hot 1iner

In this concept the thermal insulation is placed outside theliner. The liner works at the temperature of the adjacent partof the primary fluid.

13

The associated problems are:'a) Liner material?b) anchorage of 1 iner, buckling stability and fatigue in

general areas and at. special points and penetrations:c) insulation material?d) behaviour of the whole structure if the structural

concrete is at a high température.

3.3. External prestressing g^steme

With the trend towards increased pressures in reactor pressurevessels and tne adoption of PCRPVs for reactor systems otherthan GCHs, it is increasingly difficult to «achieve the high densitiesof hoop près tress required to satisfy the designer, A number ofexternal hoop prestresn systems have been developed recently or arestill in the development stage, ""he Panel feels that more informa-tion about thesa systems ought to be made availaole to vesseldesigners, including information about the following:a) Ultimate load behaviour5b) measurement of prestress levsl. and its adequate maintenance;c) corrosion protection5;d) effect of the system on station building design and schedule;e) economics,

cj.4» Load- oar ryingoapa o i ty analy si s

Methods of analysis of the load-carrying capacity of PCKPVs forintegrated reactor systeia designs developing towards geometricallymore complicated shapes may not give sufficiently good estimates.

14

It is recommended that further development of methods beattempted with the aim of achieving more reliable toolsfor the ultimate carrying capacity analysis of FGRPVs.

5.» 5.» Effect ofa Jjg aj ^

The tendency of PCRPVs to become thicker- walled in relationto inner diameters has intensified the need for furtherinformation about possible pressures gradients in concretedue to porosity or cracks in the liner and concrets.

3.6 • Sliort~term and Ipng- term sjr n j oj s ^

The short-term strength of concrete under triaxial loading conditionsat normal and elevated temperatures is not sufficiently well knonm.It is recommended that emphasis "be placed on the continuing of three-axial short-term strength testa. IJva-luation of the results of suchtests ehould lead to the formulation of statistically reliableparametric failure hypotheses for a range of typical concretes ofinterest for PGRPV construction.

The ratio of long-term to short-term strength of concrete is knownonly for the uniaxial états of stress (it is about 0.75-0.80).Almost nothing is known about this ratio for nsultiaxial states ofstress.It is recommended that investigations of the strength of concreteunder constant long-term multiaxial loading be performed. Theobjective should b© to correlate findings on the "strength loss*1under long-term loading referred to the short-terra strength, withconsideration of multiaxial states of stress.

15

It is recommandée that in all caees the strains of the specimensbe measured. It is also recommended shat the effect of size ofspecimen "be eiaminad.It Is further considered that 1 jug-terre teats under slow cyclingof loading and temperature are of great importance.

5.7.» Cyclic Ipading of cg org ta and

The future development of load-following high-temperature reactorswith direct-cycle gas turln ies may require that the coolant pressurein the PCBPVs - and porhapa the coolant temperature ~ be cycled. Therange of pressures may be L-stwesn 0.25 and 1.0, and fche total numberof cycles may 'be 25 000 in the life of the PCRPV.It is not yet clear what the above requirements would imply in termsof range of stress or strain cycles in the steel linsr, tha concreteor the attachments. It is, however, envisaged that these «ay besufficiently important to require further knowledge of the abilityof the concrete and steel liner to survive the cyclic loading inthe environment which will "be required,

Long-term strains in concrete.

Further studies of concrete material properties would bs useful,particularly in the following areas?a; Relaxation modulus and Pois£>n'e ratio, in th>. temperature

range and for concrete of interest to PCHPVs designers.Special emphasis should "be plsopd on the. current and predictedfuture trend towards higher temperatures and the so-called hotliner concepts}

16

"b) interdependence of relaxation modulus and Poieson's ratiowith Hiul'ti axial stress (strain) fields;

c) moisture migration in mass concrete as affected by (high)temperatures, the presence of impermeable barriers andlocal injection, of water. Of particular interest is theinfluence of moisture on the oroperties of the concrete.

Radiation from the reactor core may cause deterioration in thesurrounding concrete, through the direct infjuence of irradiationon concrete properties and through a secondary effect, theproduction of beat with, ensuing temperature gradients. Someresults of irradiation tests on concrete specimens have beenpublished} they seem to indicate that significant damage mayoccur if the neutron or gamma dose is sufficiently large,Internal shielding may be sufficient to keep the irradiationdose within safe limits during the whole reactor life. Furtherwork on the effect of irradiation on concrete ia, therefore, required,

3 « 1 0 • JTendpn oorros ion

The structural integrity of a concrete pressure vessel depends, toa very great extent, on the integrity of the prestressing tendons.It îs, therefore, of vital importance to ensure that the tendonsdo not corrode. In the case of u.igrouted tendons, several commercialproducts are available to prevent corrosion. Many firms. r throughoutthe world - specialise in preventing the corrosion of s ea] .in alltypes of environment. Research in this subject is generally conductedby specialist commercial companies whose proditcts should be thoroughlytested before acceptance. Attention should also be paid to thereliability of grouting,

17

Tn concluding their Report, the Panel would like to make somerecommendations which they believe would be valuable in consoli-dating the work done by the Par;el and continuing it in thefu.ture.

1) It is recommended that the work initiated by the Pane]in this rapidly expanding field coiud "be continued withadvantage to the Member States if the IAEA would establish,under its au sin ces, a Working Croup ox* its equivalent onprestressed concrete reactor pressure vessels. Such aGroup, at subsequent meetings .and through other regularcontacts, coulr maice further progress regarding specificproblems, such as are listed in Part %

2) It is also recommended that the IAEA could usefully facilitatethe interchange of relevant information. For instance, theycould circulate a bibliographical listing of reports andpublications which were Rent to them. Also, it would beparticularly valuable to 'nave such a listing circulatednot only of research reports, but also of major experimentswhich are planned or a)ready in progress. It is envisagedthat the members of the Working Group oouM assist in thesetasks,

3} Finally, it is recommended that in certain fields, particularlythose mentionad in Part 5» it way be useful if the IAEA couldfacilitate the publication of review papers, and monographeon selected topics suggested from time to tirée by the WorkingGroup»

18

STRUCTURAL BESIGÏÏ PHILOSOPHY ANS CRITEHIAPOH COECRETE BEACTQK VESSELS - U.S. PRACTICE

W. Rockenhauser

Prestreased concrete reactor pressure vessels technology isvery young and codified. Standards have not yet been established.However, efforts are being made in the U.S., as well as in Europe,to write uniform "tentative criteria*' or "recommended practices"which could eventually become tue codes for design ani coastructionof these vessels. Much of the material presented in the paper isdrawn from the content of tentative criteria and from the designstandards for the first U.S. PCPV, the Ft, St. Vrain reactor, Itis followed by a comparison of European and U.S. standards and abrief discussion of the relative merits of the various approaches,

A PCRV comprises five major constituents: the basic concretestructure, the post-tens!oning system, the nonprestressed reinforcement,the liner and the thermal control system (insulation and ooo]ingsystem). The paper concentrates on the first three components.

19

1. INTRODUCTION AND DEFINITIONS

Concrète E.eactoj; Vessels as defined here, and as distinguishedfrom the secondary containaientG art structures which contain areactor or reactor (primary) system directly without the use ofanother intervening pressure barrJer. As such they are continuallysubjected to the pressure of the primary system whenever the reactoris in operation.

These vessels are commonly known as "piestressed concretepressure vessels for rp-îctoiV or more briefiy, as prestressed con-crete reactor vesseJs (PCEV's); hereafter this abréviation will beused. A PCRV comprises five major constituents: (1) the basicconcrete structure; (2) the post-tonsioning system; (3) the non-prestressed'reinforcement ; (f\ ) the Ijner; (5) tne thermal controlsystem (insulation and cooling system). This paper will be confinedto the first three», components.

PCRV technology is very young; and codified, across-the-boardstandards have not. yst been established. However, efforts are under-way in the U. S. 5 as well as in Europe, to write uniform "tentativecriteria" or "recommended practices'* which could eventually becomethe codes for design and construction of these vessels. Much of thematerial presented here is drawn from the content of these tentativecriteria and from thr> design standards for the first U. S. PCRV, theFt. St. Vrain reactor. It is followed by a comparison of Europeanwith U. S. standards and a brief discussion of the relative meritsof the various approaches.

2. PCRV STRUCTURAI, RESPONSE

The response of a correctly designed PCRV to an internal pressureload is illustrated in Figur? 1. The figure shiows three distinctregimes of structural behavior, which has in fact been observed onactual vessels. (Depending on the point of ineasurensent, the distinc-tion between the second and the third regime is not always as clear asshovvi in the figure).

The regimes can be explained qualitatively as follows:Regime 1. Starting from a compressive strain (Imposed by the

tendon system), the vessel overall response, i.e., strain, is linearlyelastic slightly beyond the point identified as RP on the ordinate.In this regime, minor cracking of the concrete may occur as a resultof thermal loads or local discontinuity stresses, but basically thevessel behaves as a monolithic structure.

20

Pressure

ULTIMATE STRUCTURAL STRENGTH

DeformationFigure 1 Idealized General Load-deformation relationship of a PCRV

Regime 2. With further increase in pressure, the general tensilestress in the concrete reaches the point where major cracking commences.The concrete continues to crack throughout Regime. 2 with the structuralsteel still responding elastically. The overall response of the vesselis, therefore, still elastic (although no longer linearly so) and thevessel deformations are more or less reversible.

Regime 3. At even higher pressure, concrete will be extensivelycracked and the structural steel is loaded into its plastic range.Vessel deformation increases very rapidly wiTb increased pressures untilthe ultimate, strength of the structure is reached.

The foregoing simplified description assumes a short term responsein which time dependent deformations are not a factor. During the designlife of a vessel, considerable irreversible deformation (primarilyresulting from creep and shrinkage of the concrete) occurs. This resultsin a concomitant reduction in prestress which inust be factored into thedesign. Creep and shrinkage deformation interacts with the short termresponse of the vessel only to the extent that (because of the reductionin prestress) it changes the point at which the concrete enters thetensile regime under increasing load. Il: thus results in the somewhatlower end-of-life response curve also shown in Figure 1.

Thermal loads (thermal gradients through the. walls and heads) donot result in significant deformations in a structure as highly redundantas a PCRV. They have little direct effect on the response of a vesselto internal pressure although the presence of elastic thermal stressesmay result in a slight lowering of the pressure at which the vessel ceasesto be linearly elastic. Temperature does have an indirect effect, how-ever, in that it influences creep and shrinkage rates and thus increasesthe long term changes.

3. DESIGN PHILOSOPHY; DESIGN/ANALYSIS APPROACH

Design Requirements, Definitions of FajLlure

As in the case of other structures, a PCRV, to perform its functionadequately, must carry the imposed loads "safely", and its deformationsmust remain within allowable limits ur.der these Loads. In addition, itmust remain impervious to the. fluid it contains.

These requirements, imply three definite types of failure: (1) struc-ture failure; (2) excessive deformation ; (3) failure by leakage. Theintent of the design process, and the intent of the criteria on the basisof which one judges the design, is then to ensure that none of these typesof failures can occur during the service history of the structure.

22

Bas ic PCR VjDe s i grt Philo s o phy

To accomplish these aims, PCRV'ss (at .least to date) have beendesigned to be fully prestressed against the highest pressure which •the vessel is expected to encounter in seivice. The question ofwhich normal or abnormal pres&ure conditions need to be consideredin this context, is, of. course , related to design of the reactorsystem used inside the vessel, the pressure or absence of pressurerelief devices, the overall plant design, and other factors. Itcan, therefore, be discussed meaningfully only as part of the overallreactor system design philosophy.

In addition, the design usually intends to furnish some factorof over-strength beyond the highest anticipated load levels, regard-less of any credibility considerations. This is accomplished byproviding more structural steel than would be needed for service loadsalone.

Within the range of anticipated loads, vessel deformations areusually small enough as not to affect significantly the functioningof the PCRV. On the other hand, long-term viscoelastic response mustbe evaluated to ensure that creep and shrinkage deformation does notrender the vessel unusable. Aside from the associated reduction inprestress forces, the most critical factors in this respect tend to bethe position, alignment, and dimensional accuracy of penetrations usedfor refueling machinery or removable primary-system components.

Thermal loads are,, only partially balanced by prestressing (overand above the prestressing required for pressure) ; the remainder istaken into account fay appropriate use of bonded mild-steel reinforce-•rnent. The ratio of over-prestressing to reinforcing varies considerablybetween designs. However, the liner (and, therefore, the concreteimmediately adjacent) is kept in compression under all foreseeablenormal and abnormal operating conditions throughout the design life ofthe vessel, this compression, may actually become rather pronounced,particularly toward the end of the design life, and appropriate pre-cautions are therefore taken to avoid the possibility of liner buckling.Design Methods

In current U. S. practice, the design philosophy of Section 3 isimplemented by the simultaneous use of several design/analysis procedureswhich can be categorized under three main headings.

Elastic Ana^lysis^Cworkinjg _jsj gss_j esi gp }

The elastic response of the vessel to mechanical and thermal loadsis evaluated by appropriate methods and the stresses thus calculated arejudged against working stress criteria to ensure that the vessel responsewill be essentially elastic, under all credible combinations of operating»accident, and environmental loads.

23

Liaiit .Analysis

>,oad factors are applied to predicted JcaKimunt loads; thenallovible limit criteria are s.ttisfied for the«e factored luads,usinp apsropriat<? limit design tecnniques. This 1-, done to ensurethat rh& structural bo»avirr of th*> vessel beyond th« range ^f

3-3*ds is, in fact, predictable «m«.l to pr'o«?idv« suitable over-aargjns.

Creep Analysis (Viacoelaitlc Analygif )• <n»«nn»^i*i mn«ii*. *inm •• •* «HMI*»MI 44»*. . •.••••i * *> , »»i ••»<•. i^mnimi n •JiiK^i •««•

Upjvcr aid lower bound evaluations art» »ado o" th* p«?rnaneutdéfera» irions which ca« be ocpectea during tî»t> servi e îife of vh<«veasc. , on the basis of creep analyse? which arc usually carried outin conjtujction with toe elastic oiaiysis

4, LOABMC UON^niOHS AV) «via»ÎK Bfk'D,*'& CONCiBEAiffl iN TtT

Loads

1*4»^ loadings oa a if7-^' '•#& b<2 "> tutorize* in (-vvera1 vay&; for," *: (1) Construe» ton Loada vs. «.ijeratiag. 2 »«<»»; (?) IOQ<!S

by tue ayct'j'n >)"$ «»i>r^roame"tjî l.)acs; O) afcnatt'csl vs.loads. Xhe cate4or*zôtioQi$ obvio»^lv» o\*-t 1 Ap» but /o*

tHelrss K^C to persil th« applicatJon of <H f forent allowables»to tt'jW.irv construe t ii»n ^ado t >an would be- usr-d f'^r r-Kr^ice icitis,and & that fhc-xiaa) lodd^ ts&v bA tr«aff«' dirt.rentJ.y if» t'i«? analysisfrom •'ethanic^l loads. /">so, jf^ie 'o^s m.iy »<«ed 10 "hoaly . the iim^w *.«*!«, n s .

Cable T lis^s fhe ty*>e<s »i l"$i»s, wj'«icui tog rd lu cri «which ar«" co««i^f»r»«i in tit. vôwst î ^»i^^. &ia'-î »xit«iu'*o.v remarks,

lax to soiae of these loadinp ^i-'^i.t-i^ C>îlo\ '{

Dead '.oad

aft» je«t4i«a ly \j"'.ta, inJ aro <• «n idcr*,* in fh#but fcv comoari&ou with r î» . oressui-» !>.,«*<:! tho> urc notSignî.u.snt, Hovev«-£» ue^d 'o<»ds ' • c- -i iaport.nt c<«»ri»t exit ion fotthe c?~ti<*l!y coj'v'pted v«*»s2* .it v«r*ous .t Ot * »

corre»poto«linr*' evi>-j-*.if «^ «ire r^dc

Loadgj P^pe ^ a d & and

Live loads, Pipe Lruul* aad Ra^cttcnv .11-2 significant factors»in th»-1 detail design 8tf»?t t^.es t-lativ^4y local ef»fo t t » i^ea ro a«consi îcred. mhia involves» priaoi»r.»ly, thi- dcsigt. ar«} analysid of pene-trations and penetration

TABLE I

LOADING O

a) Desd Loads

Weight of th% PGRV i t s e l f : we'lgbl of pur t s of v^ r sc ' . ^. variouscwjstjruct-iori sta[/fr !3 3no equipment subs:, ructurc-c; r-.;,ap->r^ri iy (e.g..,

).:i£ ccn-';i act.ion) or perwsïi^ni ly i,cn.,nu.d o^ or v*ti.ia the PCRV,

Caused by nquipineni moun ted 011 wr w i th in r h f t >'CRV - refuel ing machinery,ro ra f i rg TMchii.-af}' , pn r, ••,:"- I v

11

Caused by T"<enBài K^pavçlon , Cold Spr rig, P^pc ?di lure&,

sbj',.; vf»i"i Lu/, îrcffi va:u<.i%tt (ic;. Oc.rt«ii-. seacror sysGPia'j) to**

aused by t J2iparatur^ f - i y d ' ^ n l throughout 1 1' • PP

î, ^.on

lo irl.3

* Tne terra Loaaxn.fr Co^dltioi' AJ. <3et i.o.fxd )»fcre i"«.J ."iis ^T1. «i^ f«t-i; vh;i-.hin strefia' ^jid 'or ^tr;jir»<-.; and, t o - r a f o r t j i* net i - > " L t o i <.o r.^chfi'.u <J

bee text

Pressure LoadsCavity pressure may vary all the way from vacuum, to values above

the peak working pressure unc'er certain postulated abnormal conditions.European practice has been to design for a pressure level higher thanthe anticipated p*»ak working pr^ss ire by come arbitrary percentage.U. Sv design practice uses a reference pressure "TU''1 as the significantload. This is defined as the. highest pressure which uust he consideredin the design o£ trie pTrtJcu'ar reactor plant tx be housed inside thevesse). Up to this prefi^urt» rbe averr^p stress in the concrete acrossany section of the vess«.i nust be kept comnresBive.£;j irj&nineota.l__Condit. ions

In the U. S. A. :>as beeow» established practice in nuclear powerplant design to consi-iei fw- severity levels t-iher. >~<s storing environ-mental disturbances* into <•? d.vsiga. M^t*1 noterbl} , ftr.s applies toseismic effects. A pla^L (<>fd c h ' . t e f o r f c , th»? PC'1..'.'), wl?J be evaluatedfor an "operations.!"'' eaxthouake %-hich r>e"omes i> r *- -"'f the loading condi-tions considered in the working «tress c'esJvn; and .=-, "destga"* earthquakewhich is factored ••'nto the Huit design* T'>e d«»Jg*i eai thq»jiake or safeshutdown (Ksfu/'bancG is neutrally tak>'-n î9 nuit .'pie o£ c be def t i fu dis-turbancft "7hcre the n>uJ t ip] ic°t ion f-u*tT d«pcndp '»u xhe spec i f ic plantsite.

Locio[__Corcb 5 nat ion a

lu both the worKing srrecs d«si.go sad tîu. Jiuût. de*ipv>, certaincombinatirms of the .loading conditjn:;s of Tabit i. rMur,t be rcnbidered.In the working slxvss dcoif^» this iicludes ' l.i < oribf nations wbich canreasonably occur darL.g tne iiCe of the plant. Tht coc.birac ions are.lasted iu Tf»bJ.f ÏÎ. S imi lar ly» for lirait dostg.j purposes, specificload factor équations. wb3"ch wi iL be discussed in Section 1} h#-ve beententatively agreed upon. These s re intandoJ fc predict, the elasticlimit of the structure.

5. CRITERIA FOR ELASTIC BJ sWN

,a general., the cri teria for th« "Working *!trcs8 Lv*-3ign" areformulated fonventlonally in terms of «illuw bi«i stressée, expressed as.€r?ctions of standard ïiaterial ptrcper t i f tp . ïh-».se aïe c/Jincter or rubestrength for concrete azvd y ie ld or tensile atr^vi^lh lor P teoJ . Inlarge neasute, w*»tl established al.3owav1e "aiues «rf» u<?eJ, primarilytfaoae of. ACT. 318-63. (Cor respond ing to B.S. CFM15 in Britain). Som«deviatlonrs Iron these acceptée* sfandards t.^.v.t ncj^ssary however; andnew crirorJa had to b^ formoïaced . t/ »sïî:.:ît spr?s-?:ueru of f.^nperpture,•shrinkage aad creep effects.

The Individual val^fcj? >»sed for aJlov.nble stres:. «Vf diuri^seâ iusome detail beto.w. Tb*> values given, represent, curie it U. S. (>ta.t 11'°.

stip-ilatel hy D. S, regulating

26

TABLE IT

Load Combinations

• Period of Design Life

I,.- ^Construction

~*\ ^ iPressure Pr as tress Dead & PipeLive

| 1 Loads"~ """ '^' "•- \- '- ' - - -

2. Son-pressurized Vessel£_ter Completion ofPresrressiRgOpération

•

Reactions'

1

Wind- r • "

Tempera- Creep Shrinkageture______ : . • _ .-L. ._ _ .

/ / i /'j-.- _. _._... ..... . !..__.........

VF Effective »

* $3, Nou-pressurlaed Vessel j

ertcr Initî&I Heating V?. Effective /i

ColdSpring |

GoldSpring

*

. i /</

'

/

JEEarth-quake./

/

i

^i ! i ! i

1 ! i'A. Fressur:J2ed Vessel aftatr;Initiai HiaLin- j ÏJ.5 SPai Effective / i /

[5. Pressurized ''Vesafcl i - - . 1 .

i if \ // !

i j• at Time of Start of NW? Effective / ' / !

Operation

6. rjessurized Vessel atEnd of. Design Life

7. Son-pressurized Vesselat £IK? of Design Life

RP

</ \ •/

.„!: _ .__!_. ..„. L_ ._ . . .JEffective i /

, / ;. ;J _ . J . _ ' .

Effective ^''

/'

/1 ., J... L..

'

/

'

j

/'

ConcreteGeneral Consjlderationsof ConcreteAllowable concrete stresses quoted in AC1 318-63 generally are

expressed as .fractions of the 28-dsy cylinder strength f'c. Thisis du.;e for several reasons. One 1s just, time-honored practice,Another is the realization that structures are often fully loadedonly when they are considerably older than one month; the increasein strength over the 28-day value riien provides ? margin coveringuncertainties in analysis, material properties, loading conditionsand 30 on.

In the case of PCRVs, the luain service load (internal pressure)usually is applied only after the vessel is at least a year old; evenpost-tensioning may not occur for many months after fL;//;'! completionof the vessel. The- loads are known accurately, Moreover,, methodso£ analysis are much more refined than those used in everyday concretetechnology. Therefore, stresses are caJcalated with high accuracy;and prestress losses, particularly those occurring while the concreteis young, are calculated with great care. Consequently, tiui use ofthe 28-day strength as n clesjgu basis is o£ questionably validity andunnecessarily conversative. This is recognized, and ttept a^es UP to90 days have been specified (Section 8). The most ration»! basis wouJrtseem to be to use the strength of concrete at the time of post-tonstoo-ing.

Typical (28 day) concrète strengths for vessels completed or underconstruction f a.l 1 into the range of 4500 psi to 6000 pai.Types of Allowab 1 e Streajs es Jîpecil'ied

The stress analysis methods employed in the deuign of PCRV's(fini te-eietuent or f'inite-di f tecence calculauions) ^ Lve detailedresults on magnitude and orInatation of stresses at msny locationsthro'!.:.;hout the structure. ThereEora „ prlncipai-sf-esK criteria can bespecified directly and. no shear stras- allowables need 1o i>e given.

For the first PCRV's connp.rvrtti.ve .lilowafoîe ftveas values werespecified for "general stresses" (soiuewhat comparable to membranestresses in sheila) i.e., «tresses affecting an. area or voluisd largeenough to have overall structural significance. Different higher valueswere permitted for more- "local stresses" vrhich could be coiisidereidanologous to secondary or discontinuity stresses in stop! pressure vtigyodesign. The obvious difficulty in thiu Ls Kow to define quantitativelythe distinction between "general." and "locfjl" for a -struct.-or*? as coif-i>.lexas a PCRV. The trend now seems to be tc get away from "^e^craL strt-ss"criteria and to have only one limit., higher than the customary 0.45 f'c.which must be satisfied everywhere except at poinLs which are obviousstress concentrations.

Multiaxidlity EffectsMost of the concrete in a PCRV is subjected Lo multiaxial Load-

ing. It is well known that concrete strength increases with increasingdegree of confinement, and an effort is made to take advantage of thischaracteristic,

In a sense, even the early approach?*» which permitted higherlocal stresses can be viewed in this J ight , A more- systematic attemptis beinç m?de with the Pjirt St. Vrain vessel, wherr the. principal allowable v..jtnp res give stress </a!ue is adjusted by a multiaxial strengthfactor "C" in a manner presented in some detail Jn Appendix 1.

For a aniaxial stress lie Id C-l; lor tru«, î.ydroslat Je compressionC - 2.71. It is worth acting that C is nut a f aï Dure criterion, butsimply a design tool intended i:c gi.v<3 rerdleî Ic or const.1 votive results.

_ _ _ _Val_ues

PrlMcipajl Allowable Coiftpressive SLjcef»*> 0,45 C t*c

This al lowable app^ips, to 'ill opcrstiti£ &tr£'»ses including thosevhiiJh occur when the PCRV 1" s ' jbjsotttJ to pr« i>t re j j& ^ri'l thermal gradientduring reactor shutdown p^i^ i l . <n.i:d 10 nuera^i i>i"e«-ur'j^ up to Refereacc-Pressui e during plc-rt o^tLut ion . T'-c Liu-it uf U.^5 C f IF net to btexceeded when r.he ten-'^^ris a? f s*.ress"d to O.ô of thei*. guoianteed ulti-mate ceïislle strength. The dl j--»v j b J o p c i n c t p s i l'umpressive stress isderived fron the valu? gLve^ in ACI-31B, P^ra^japh 7605 , which is vjiidfor prestressed concrete çlroc-t 'uret j where In? Ptt i ioJfi oi andïysis sreprescribed by the Code.. The fac.or (", l<-i iuclaJs'J lor the previouslys ta ted raascus and i& to be c>L';f mined p^r Ap j i rndLx A. Whore high localstresses aie Incliceted adiacer,' r^ s r ruc tu r r l discun*- Jnui LifcS , and thusc ïear ïy result ivosi stress c OPO' LU. rat ion e f f e c t s , the values prediccsdhy eldstii. ana lysis are c i a t - regarded sxncc «..giiituu ing c^pcrjence hasshown tuat such local coiidltioiïo oc< or in Ail «•"••nctures wifhouL causingdetrimental e f f e c t s .

';ncrefe has a i-ensilo otreiigLh cf 1/12 Lo 1/8 nf the 28 day com-press, on strength For a 6000 p&J "Likrete, tlje I ensile sî.rt?ngf}> shouldtherefore bo 500 to 750 psi. ïyplrailv, the andésite concrete iut theFort St. Vrain PCRV has exhibited a Lenstie «freugth of around 600 osi,i.c», about__i/10 of the conp)es&iv*« streiigchs. éor f', - fOOO pal» thevalue S'T' . givos an a] lovable tensiJ*.- stress uf 233

29

Reliance is «oL pioced on tot. tensile res.! SB tance of r.he concretewhen the computed tenslif stresses exceed 3/t'c, In such cases, it isassumed that concrete cracks will develop and, therefore, reinforcementis provided to carry the entire tensile load.

Bearing Stress Under Tendon Anchorake«" ——— • ——— • —— •.——--- — — —— -— — —— —- ——— — -- —— — ' —— - O.fj f .A,/ 0but not greater than f'

*

j 'ils allowable stress Is taken dirent.! y from Uie ACT -.3 IS, Paragraph2605., where A'b and A,., ;-jre as defined in ACT-318. Paragraph 2 GOO, forlow A'k/Atj ti.is formula results in conservative permissible values. Forlarge Afj /AL values the ion-uila roarers cxcossivtpe.rmip.'ïibio concretestresses. Therefore } the iv.axi.aum limit "f. ff'c is also specified whichagain is a very conservât i 12 VHÎUC. Guyon reports in hi H hook "î'estress eCoiicî'et_«. ., Page 155. that t-,:r an inleriia.i t'rictjon angj.t- ol" 3'5a , a formuladeveJoped by 1-1. Cd.quof g'j.yc-c an uJ.tj'aate bearing nre.'î.surt- pqudl to about11 times tlu. uniaxiai coh>pr'.iss5ve H trees' tlu The c.rit crjc--1'» of 0.6 f „i,. _____ . ' '-vA!, /A., i.s wide.lv -.îaetl «j»d t'Uert- .-ire no kno-^a vea^oivw fo.' doubting itsc: ons erv& ti sm •.

Stetij

In general t.ÎHi convanLiOi i -^T prov-îBÎoi îS of AOT~^li< c;jn reodily beadopted. The- applicable paragr^pus ace a? follows:

/vppj .icdbj.i! i'ara.

Tensile 8 tresses

This implies

1003

For deformed bars No. 1! and wni.ilJcT" witha yield strength of 60,000 prf.i or mAllowable Strc-ôs . , . „ . . . .For "M otl'er reiuforc,au7ents :Allow^.blt; Sv.rt-.ss . . . . . . . . . . .

_Compre_s3ive_ Sj.rest.eR

"aragrauh .1102 i.s ime-decl ro -..icjiccompression of steal iu bendju^; .Tuni-bers. Although no! Jirocfi> Dralogous,che paragraph way be used i:i-re =IK veil.Thus, where steei is Latendo.d to fnucf-jo'.!as corapres s i ve r t; i nf or c emen l, :Allowable coaprcsuj on . . . . . . . . . .

2«i ,000 psi

?u>000 ps.I

1102

24,000 psi

s

Bond Stresses 1301Tendon_StressesTendon Wires - Allowable Mean Strèsses^The provisions of ACI-31S, Paragraph 2606» are met:

Temporary jacking forceAllowable stresses ...... r ... 0.8 £'

Immediately after transferor anchoringAllowable stress . . . . . . . . . . . 0.7 fg

Effective Prèstress 0.6 f1gAllowable stress . . . . . tb« smaller of

0.6 fsyIn addition, the maximum temporary jacking force is iiraited to 1.be

yield stress of (f }, and the maximum stress after anchoring is limitedto 0.9 f _t> These two limitations are intended to covet prestressingsteels which are not manufactured to the acquirements of AS'CM A-421, ystwould be acceptable for use in a PCRV, (ASTM A-A21 specifies a («Lnrwirnyield strength of 0.8 f5 }STendon Anchors

Ko; general criteria are available. It is specified that the actualend anchor assemblies must withstand, without failure, at least 1,2 xGUTS under static as well as cyclic loads; this strength to be demonstratedby appropriate tests.

6. CRITERIA FOR THERMAL LOADING AND THELASÏIC EFFECTS

Thermal loads (temperature gradients), creep aixd shrinkage areclosely interrelated. All three may cause stresses; in addition, creepmay relieve stresses caused by the other effects. To be most meaningful,a design calculation should, therefore, consider these, factor together,as well as their interaction, which is difficult at besi:.

Furthermore, relatively little quantitative information exists aboutthese phenomena, particularly at temperatures above the 150° - 200° Frange.

In view of this, several techniques are generally used simultaneouslywhen considering these factors in the design of a PGRV: i. Thermal loadsthemselves are Lirai ted to values known to be safe; ?.. Upper-bound esti-mates are made of the permanent deformations to be expected as a resultof creap and shrinkage;, 3. Stresses resulting from temperature gradients,creep and shrinkage are assessed on an elastic, nodifiad elastic, orviscoelastic basis and the calculated values are limitée* to values con-sidered safe and reasonable.

31

Criteria for Thermal LoadsFour quantities are usually specified by designers. Numerical

values vary somewhat from case to case, but fall within the rangesgiven below. The four parameters are:

o Effective concrete temperatureat (or close to) concrete-linerinterface 100 - 170 F

o Local maximum concretetemperature 130 - 200 F

o Maximum temperature differencebetween inside and outside ofconcrete 40 - 70 F

o Maximum rate of temperature riseduring startup Order of 5-20 F/week

Deformation CrjiteriaKo generally valid criteria can be given. Under t;he worst possi-*

ble combination of tolerance stack-up, anticipated load history, andcreep rates, deformations at. ead-of-life :nust b<= srn.aU enough for thereactor system to remain functional. In audition, the loss of initialprestrees must be no greater than the. value factored into the design.

thermal and_ "r>'gp 3i.regs J ^ L°_PJLEarly designs evaluated thermal {stresses on a purely elastic basis

without considerixîg creep effects. It was found that:., for the designto remain within practical and economic limits, adjustments had to be.made. Some designers calculated thermal stresses using a fictitiouselastic modulus whose numerical value was established empirically.Others relaxed the allowable concrete tensile stress value for combinedmechanical and thermal loads.

In U. S. practice, a strong effort is made to calculate actualstresses in the- vessel throughout its life and. load history as realistic-ally as possible, using the best creep and shrinkage data available torthe specific concrete to be used. These stresses are then limited tothe allowable values previously established as a basis for cilsstic design..LIMIT DESIGN CRITERIA

In the design of early PGRV's, elastic design-analysis was supple-mented only by an attempt to establish a gross structural overload factorbased on what was taken to be the ultimate load carrying capability ofthe structure.

32

ID U. S. practice, this has been replaced by a dual approach,both parts o£ which have been called limit design techniques. Actually,the first is a true limit design: the second is a variant of the pre-viously mentioned ultimate load analysis carried out for a different,and more meaningful purpose.

In view of this., it might have been more accurate to talk in thispresentation about four methods used simultaneously in the design ofa PCRV: elastic, viscoelastic, lirait, and ultiinate-joad design-analysis.

Limit Design (Prediction of E3-astic Limit)

As stated previous3.y, load factors are applied to predicted maximumloads; and stipulated limit criteria are then satisfied for certain com-binations of these factored loads (analogous to the manner in whichallowable stresses are satisfied ir\ the elastic design for all reasonablecombinations of actual loads),

The intent is to set the limit conditions and the correspondingcriteria such that compliance with them assures predictable (responseof the 'vessel. Local large deformations are uu.i-ike.ly and prematurefailure in the liner, piping., or internal equipment due to excessivevessel deformations should not occur.

The specific limit condition equations now in use are:a. BL H- RP •••>- £' -T XL

b. DL + 1.5 EP -f 1.5 TL

where

DL = dead loadKP =• reference pressureE' = load resulting from "design"

earthquakeTL - thermal load

Elastic response under Limit Conditions a is intended to ensure asatisfactory seismic design of the PCi'V. The reason for selection oflimi't Condition b is that some reserve load capacity should exist beyondthe maximum anticipated loads.

The allowable criteria corresponding tc the above limit conditionsare as f o3lows :

ConcretePrincipal Corspreasi.y^Stress 0.6 C f c

33

The use of an increased allowable concrete compression under LimitConditions a and b is justified by the philosophy implicit in the adop-tion of the limit design concept where it is intended to ensure elasticresponse at Reference Pressure and design temperature by demonstratingthat elastic response is to be expected even under higher loading condi-tions. Thus, the allowable level of concrete stress can be the elasticlimit in compression for concrete. îfewnisn^ ' states that concrete maybe P ^ected to respond elastically in uni axial compression up to about50 to 60% of die unLaxiaJ ultimate strength: hence, the use of the upperlimit 0.6 f'c sterna justified. ACI-318-63 aJso makes provision for auincrease in allowable stress of 33-1/2^ for improbable lord combinations.If this increase is applied to the concrete allowable used in elasticdesign (0.45 C f ), the quoted Ugure of 0.6 C F' is obtained.Bearing Stress Under TCP don Anchors 0.7 f 7 A.' /A,————— ——— —————————— —————— t. b b

The allowable stress permitted for bearing stresses io Increasedby only 17% over the elastic design allowable. This is a very conserva-tive approach considering tht maximum fotresset> possible ft these rein-forced and confined zones as. pointed out in the discussion on the allow-able bearing stress under tendon anchorages fcr operating lo«<ds.

SteelLiner: PrincinaJ tensile stress 0.8 fsyReinforcing bars: Principal tensile stre&s 0.9 fsyPrès tress inp; wire: 0,95 fb sy

The selection of 0.9 i< _ is thought to provide an adequate marginof safety against inelastic action In the iinnr and tjonu^d reinfoi'cement.The action of the bonded reinforcement is well understood and is moreakin to that o£ flexural r&in Lorcement than cjtber categories. The capacityreduction factor selected of 0.9 coulronno ! n the 4C1--3J3 recommendationfor such s tee.!.

A reduction factor of 0.95 is considered justifiable icr prestressingwire because of the stringent qua! Lty control required a_n its manvtfacture,and the in situ load testing vhich occurs during pr*"3tre£sing.

Ultimate Load EvaJv^L-iC;?,As regards j n tun t , this may be.st be corap<jrccl w t f c h the ul t imate load

capacity equation given IP ACi~318-63, S- jcLion 1506. }Iov;evcr 5 only asingle, large load i a< - fo r oi 2.1 is applied Co the load RP, Trie justifi-cation for tne d l f f p r e ^ ' G iie^ in the f^ct char the dead weight of thePCRV results in negligible f-tress compared vita the ir.teinal pressure,thus allowing the effect of dead load to be neglected. The higher loadfactor applied to the Load RP actually results in .1 nore conservativedesign than ACJ-318 would stipulate.

The ultimate strength criteria of ACI-318 are directed primarilyto framed structures, where yield of reinforcement implies the formationof plastic hinges and collapse mechanisms. Yield in the wall reinforce-ment of the PCRV does not mean that a collapse mechanism has formed and,provided the deflection characteristics are investigated under increasingload, there is no reason to state that ultimate capacity has been reachedsince the tensile failure strength of reinforcement is much greater thanthe yield strength. Conventional analyses do not poc^pps this ability,making the attaining of yield stress a convenient point to halt an analysisalthough a reserve of strength exists beyond this point.

The allowable stresses for this conditions are:Concrete Principal Compressive Stress 0.85 fTendons 0.95 GUTSBonded reinforcement GUTS

except in heads whore steelstresses will be limited to 0.9 fsy

ID areas exhibiting primarily membrane action, the concrete willof course, be fully cracked. Where structural action is more complex,e.g.» in the heads, compressive stresses are limited to 0.85 f , avalue which is accepted bv A,CI-318-63 (Section 1900) as being applicableto members subjected to combined axial compression and bending. Whereit car» he clearly demons treated that a condition of triaxial compressivestress exists and the material is confined, an increase of this compressivestress allowable is justified similar to that provided under Limit Condi-tion 1, i.e. , 0.85 C f .c

b . Tendons and Bonded ReinforcementAs the ultimate limit condition is a definition of the minimum

cavity pressure required to fail the vessel, the use of minimum guaranteedultimate tensile strength seems warranted for bonded reinforcement. Àvalue of 0.95 GUTS is used for tendons. The use of this value is justi-fied by the results of European tests on full scale tendons reported inreference 27. Their authors report ultimate load tests on B.B.R.Vw 'unitscontaining 121 wires of 7 mm diameter. The results of tests on straighttendons indicated the development of 99.7% GUTS and the curved tendonsgave strengths of 96.5% GUTS.

The allowable stress of 0,9 f for bonded reinforcement in the heads,combined with the concrete allowab?! of 0.85 f, ensures that the thickplate nature of the structural response of the fèeads is not lost by theformation of large cracks. It also provides a margin of safety in thehead analysis which is necessary since the action of the heads is amenableto precise analysis than other portions of the structure.

35

8. COMPARISON OF VARIOUS DESIGN BASES

If one compares such design criteria information as has becomeavailable from different design organizations, considerable-variationwill be found. Design evaluation and progressive changes have takenplace within each organization as well.

Therefore, a single overall comparative evaluation cannot meaning-fully be made.

Table III is a compilation of parameters for which a comparison mayhave value. Brief explanatory notes follow:Pressures

In European practice,, the design pressure DP .(at which static balancebetween tendon force and internal pressure force occurs) is selected at10% above normal working pressure (NWP); and in the most recent Frenchvessels (Bugey and Fessenheim) as low as 6% above the NWP. This comparesto a corresponding value of 23% in the U. S. Moreover, in the U. S.practice this pressure (called Reference Pressure, RP), is higher thanthe highest pressure which can actually occur in the system. The Europeandesigns permit the maximum pressure to exceed the design pressure byanother 10%, giving a true maximum pressure of 1.21 x NWP and correspond-ing with it the possibility of net tension in the vessel walls, which maylead to partial cracking of the concrete. ,Load Factors

A basic characteristic of the European ultimate load evaluations, aswell as the limit conditions factored into the more recent French designs,is the a priori assumption of a certain mode of structural behavior orfailure respectively. This is exemplified by the analytical, approachesgiven in references 28 and 29. The analysis offers no method to ascertainwhether the postulated structural behavior would actually occur. Carryingout similar calculations for the Fort St. Vrain vessel would give ultimateload factors between 3.0 and 3.3.

In general, one can conclude that this type of. load 'factor evaluationis of limited use and can, in fact, be somewhat misleading since the resultdepends strongly on the assumptions and calculation techniques used.

The U. S. approach has been to base limit calculations as well asultimate load factor calculations on methods which mathematically pi-edict(rather then assume) the structural response of the vessel.-1 A correspond-ing numerical load factor of "« 2.5 x NWP has been selected.Concrete Temperatures

All temperature values shown in the table are sufficiently low asto eliminate the likelihood of significant material deterioration. U.K.practice parallels the British approach; both are somewhat more conserva-tive then the French designers*

36

TABLE III

\

P C . R . V . DESIGN COMPARISON- "

1 PQE55UQËNQBMAL WORklNGPGE55U2E (WWP)

maximum crediblysystem -pressures

, PCES5UÎSE (D R)BEFEGEMCSPJ2E.35UEE (CJP)INITIAL PEOOFdPIR)TE5T PCE5SUEE

2LLOAD FACTORUMiTCOMDîTICW la.tlAAJTCOMDtTSOK! îbUMSTCOWDITiOW L

FT 51 V2AINPSIG. W.W.R688

O345„_

845 1.23972 1.41

DL*IÎ3N\VP*E'+TLDL+L84H\YP*i5TL^(5EE TEXT) .PAGE vfl~25

3%KZNG°U*f!*"E 600OpREAMR£*at age ^days)

j EFFECTIVE4-.-COWCEETE TÊMR(F"}

5 STRESS „.„ .,.ceiTeaîA (PS.I.)

TtJMPOSAKY PCIWC1PAL"COWPBESSION

. .___ __ ._. _ PBWCIPALTEMSfOW

OP£J2AIING PS^tCIPALCGMPBES5IOWCQWTOEiSiON

(LOCAL)PeiMCIPAL

~ TEWSIOW

- - TCMDWWEABEAE1MG ;

SJEA8 AWCUOSS

6 LÏNF-P.. TUtC^WESS

130°

„_» ^-

) *^~""

u6C*^T:<feO.6fc AVERAGE

-V4-*

OLDBU2YP 5 Î G NWP350

1.2'1385 I.IO_ ^

WYLFAP51G384

423_ _ .

443 I.Z65 |437

NWP

1.21I.IO

_ _

1.165j

___ . . . ...,

1 i 55~T 3.5-0

4700 (25)

131'

EDF-4-&P S I G385

426

469

NWPOUNGÊWES5ÔP5IG455

)

I.IO |478{ —

'l.ZI ip^O

WWP

1.21I.IO_ ,

1.265

JDL+UONfWP+TLl{DL+t3ZWWPH5TL|

! !25 1.92 j 1000 2.75

4-700 (56).. _ .

95s 6'FKMUMK

2

pSifc s X40OJ

0.4'ifc5 2000

5600 {90}-

^ 167°._.

05^cs 29 8O

II9Ô 2.75,

470O (28)

13!"*-

497 |035C-, (9ÔO

jassfi» 4OOO jojfc 5 s&o

NO CC!T£2iA

i/z"

LOCAL-,1 — 5OO

500

5500NO CQITEGIA

V4*

!9S

0.53C*'C

AT LIN£~K_ iOQAT OUTSIDE- 50O

FC.EWCH FRESTEES&EdCOWCEETE CODE 1

MO CCiTEPiA | NO C!2ITCC?IA

s* Vz"

&UGEY - 1PSiG NWP637

G76 --I.06._

NONE 1 - -

i6SO 265

580 O (28)

ISO"

NO CQITEKSA

FE5SENMEIMPSIG NWP435

464 I.O6.

NONE

a.«ID6NWP+TLDW3ZN\VW.25TLiiOO Z53

5400 (Z8)

.e?fc s 36OOZÔ4

Q4fé s Zi6OQ8fé s 432O

170

NO CC.ITEE1A"

1* 1

= Factor which takes ïnto account the principal conpressavcstresses at the location under consideration

£' = Cylinder strength at (days) specif.ted

Al lowab le. Stress es

In ail instances, it is recognized thai the established allovjabj.esof conventional s truc tarai technology cannot be used Verbatim in ameaningful manner. Higher values are permitted in compression Forlocalized stresses, or where concrete is confined. British designers ,are permitted to exceed the compressive allowables of Code U.S. C? 115in locations where two or three dimensional c oppressive stresses ar? pre-sent, provided the excess is justified "expérimentai. ly or in the light ofexisting knowledge". Also, calculated tonsile stressas up to 500 pai arepermitted provided the net force across tJie ve.ase.1 wall section regainsccrapressive.

In general, allowable tensile stresser, in unreinf creed areas rangefrom 'x . /T7" to 7.3/ET"".

Allowable comptess^va bearing stresse,--, at anchor a vary between0.6 f and 0.93 Ir ' 3; in conventional European practice/., In somePC8VT3crelatively large high-strength concrete anchor bearing blocksare used. While these would seem to reduce significantly the bearingloads ou the PCRV, the blocks th ems elves require considerable develop-ment, and no standards are available. ACT.-3iti (which is «adhered to byU. S, practice) appeals to be the only code prescribing allowable stressethat: account for tendon spacing and hole size.

Two other factors, uot listed on Table III, are worth

Relatively similar criteria appear to have bt?en used, by all designers,A summary comparison is given in Tabla TV. An obvions difference of opinionexists on the question of tendon grouting. The tendons in the Frenchvessels have been grouted. In British and U. S» practice, corrosion isachieved by other means, and tendons are J.&ft unjrroute:' so as to make tendonreplacement possible should it ever be required.

The U. S. criteria imply a conservative approach as regarda the usaof reinforcing steel, In. Great Britain, fch»? C.E.G.B. specifications con-tain no sf-ipulations regarding i:he use of bonded reinforcement. Actualpractice has varied considerably between different designs. The Oldburyvessels have very little bonded reinforcf.meiit and a h.iy.h degree of pre-s tress; the Wylfa vessels, oa the or.her hi-ind, coo tain a considerable ajcountof reinforcement:. Other cylindrical designs include a moderate proportionof bonded steel. The early French vessels alao had little rein forcement. .Later French designs have involved munh higher amounts of bonded niild steel,mainly for purposes of crack control,

38

TABLE IV

Crinaria for Prestre^sing Systems

LIMIT OR ITEM U, S.Practice

Européen

Average axial tensile stress duringjackingAverage, axial tensile stress afteranchoring% Elongation at failure in totaltendon length

30' of straight tendon10" wire length

Corrosion protection

TendonsType

EfficiencyStraight

Uirveci

Tubes

Anchor hardwareAnchor ultimate,percents oftendon ultimate

0.8 £'

0,7 £'

3.54.0

oil or greasesystem

unbonded

100%

95%Solid

0.85

0.7 £'

1.5 tc 3.0

cernant-sandgrout,and greasesystems

unbonded andbonded

100%

Solid andflexible

39

ConclusionAs staged in the introduction, professional groups, both here

and abroad, are working toward the- formulation of such criteria.Drafts are being circulated in France, the U. K. and the U. S. withinnuclear and PCRV oriented groups. It is anticipated that in the nextfew years, we will ^ee the emergence of reasonably unified PCRVcriteria standards.

40

APPENDIX A

Many attempts have been made to derive a single failurecriterion for concrete subjected to the many kinds of combined stresses, sofar without success. The leading candidate theories, Mohr's theory and theoctaii^dral shear stress theory, have been shown to be inappropriate byBresler and Pister (Réf. 1 and 2) and Bellamy (Réf. 3), respectively. Inthe latter work Bellamy makes the point that criteria based on the principalstresses themselves, rather than derivatives of them, are likely to be themost successful. The Mohr theory can be the basis of a satisfactory criterion(Réf. 4) for triaxial compression when the two lesser principal stresses areequal, but fails if one of them approaches zero (Refs. 2 and 5)(or becomesnegative). The point has been made by several investigators that the failurecriterion should be representable by a surface of revolution, or the like,in principal stress space whose axis is inclined equally with respect tothe three principal stress axes. No simply expressed surface of revolutionappears to be capable of representing results of triaxial compression testsin which the least principal stress varies from high values right down tozero, apparently because of a change in mode of failure in the low minimumprincipal stress r.egion which may be due to the large disparity between thetensile and compressive strengths of concrete.

For purposes of PCRV design, this difficulty can be over-come by use of a criterion which makes no pretense of indicating actualfailure throughout the range of its usefulness, but which does representfailure conditions reasonably well in regions where unnecessary conservatismwould be uneconomic, and which is conservative, or even extremely conserva-tive, in regions where economics are not affected. This approach basicallyresults in close prediction of failure in the region where the least principalstress is small and increasing conservatism as the value of the leastprincipal stress is increased. This characteristic is appropriate too inorder to make allowance for the somewhat uncertain effects of pore waterpressure on failure» the effect being small at the lower levels of confiningpressure, but becoming larger at larger confining pressures, as shown byAkroyd (Ref. 6).

Close prediction of failure in the region where the leastprincipal stress is small, demands careful consideration of the work thathas been done on failure of concrete in biaxial compression. Hilsdorf(Réf. 7) has reviewed and discussed quite comprehensively the existingknowledge on this subject and contributed one or two experimental results.Vile (Réf. 8) has also published work on this subject. Although Hilsdorfclaims that lower values of failing stress for specimens under biaxial com-pression will be found when more representative means of loading specimensare employed, it does not seem likely that this will be found to be asubstantial reduction in view of the awareness of this problem and the careta&en in this respect by Vile and Wastlund (Ref 9)

41

The best available claua on failure under biaxial compression;by Vile (Re£. 8}, Wastlund (Réf. 9) and Weigler and Becker (Réf. 10), areplotted in FÏR. V.I,A.I as full-ling curves. The curve labeled "Vile" is theone drawn by him, though the scatter of his results would permit one to drawa curve almost equally well about halfway between his curve and the curve ofWeigler and Becker. The ooints representing equibiaxial compression on thecurves of Vile and of Weigier and Becker are quite well confirmed by Glomb(Réf. 11).

Chirm and Zimmerman (Réf. 12) have summarized prior inves-tigations of failure of concrete under triaxial compression and presentedtheir own work on this subject. Theirs appears to be the only investigationof failure of concrete for both the cases of axial pressure predominatingand confining pressure predominating. They also present data on the uniaxialcompressive strength of concrete remaining after it has been subjected to'high triaxial stresses. The outcome of all these investigations, includingChinn and Zimmerman's, is that for axial pressure predominating the maximumprincipal stress at failure is conservatively given by f£ plus four timesthe confining pressure. Ackroyd (Réf. 6) has found that this is only trueup to about 3 f£ for water saturated concrete « For the case of confiningpressure predominating Chinn and Zimmerman found that the -maximum principalstress at failure is approximately fco plus three times the axial pressurewhere f is the equibiaxial failing stress as determined in their tests.

The criterion chosen for use in PCRV analysis, valid forboth biaxial or triaxial compression is the paraboloidal expression(a,2 + c 2 + a 2) - (a o + o o + a a ) - .25 f (a + a + a ) - .75 f'2 - 0i 2 3 } 2 i 3 2 3 C I 2 3 ^where validity is confined to a < 3 f.* \ c

o. - Maximum principal stresso_ = Intermediate principal stresso «= Minimum pri^o.i-oal strp-?c'3Compression is negative, tension positive

For the case of biaxial compression, a£ <* 0, the abovecriterion yields (o 2 + o 2) - c^ o3 - .25 ££ <Oj + o3) = ,75 f,!2 which is _plotted as the interrupted curve in Fig. .A.I. It will be seen that thiscurve lies slightly on the conservative side of the envelope formed by thecurves of Wastlund and Weigler and Becker.

42

For triaxial compression with axial pressure predominating,the criterion is plotted in Fig. .A.2 where it is compared with c^ = f£ H- 4 03,i.e., the two lesser principal stresses are equal. For triaxial compressionwith confining pressure predominating, i.e., the two greater principal stressesare equal. The criterion is plotted in Fig. ,.A.3 where it is compared witho as i.i5 f + 3 o3 where 1.15 f'c is tl-e equibiaxial stress predicted by thecriterion.

The assumed failure criterion represents the biaxial com-pressive strength of concrete fairly well and represents the triaxialstrength conservatively. The degree of conservatism increases as the con-fining stress, and thus the assumed failing stress, increases. The completecriterion is shown graphically in Fig. ,A.4.

ACI-318 sets an allowable stress value (0.45 f^) forextreme fibers without consideration of multiaxial stresses or the higherstresses that occur at discontinuities since such stresses are not predictedby the analysis methods used. The finite element methods used for the PCRVdesign predict in some detail the multiaxial stresses that occur throughoutthe vessel. These predicted stresses are then compared with acceptable ACI-318allowable values which are modified to take into account the multiaxial state•of stress by the utilization of the factor "C". The factor "C" is determinedas follows:

The failure criterion equation is solved for o , the maximumprincipal stress at assumed failure.

1. If all three principal stresses (c , o , op are in compression:

calculate -jr anc* TTc ca

determine C « -rf from Fig. V.I.A.4c

[a | - 0.45 C f£

2. If a « tension, assume a « 00 ^

calculate -77 and -rr * 0c c

determine C - -rr f^om Fig. V.I.A.4c

|a,| £ 0.45 C f'

43

3. If o2 and 0- tension, o, •» a, =• 0C = 1

Jo ! 0.45 C f^

It is common design pr .ctice for ordinary biaxially stressedstructures to use the full allowable concrete compressive stress (0.45 f^)when the stress in a transverse direction is tension. Under such conditionsconcrete cracks will develop which will reduce the tensile stress fieldsin the concrete. The compressive stresses are carried by "compression colur.ns,"which cannot buckle because of the confinement provided by adjacent concrete.Lateral restrain may also be provided where bonded reinforcement is used.

In triaxially stressed structures where tension occurs, asimilar condition exists. Uniaxial tension with biaxial compression produces"compression planes" which are loaded biaxially in compression. Biaxialtension with uniaxial compression produces "compression columns" which sus-tain full uniaxial compressive loads providing confinement is provided.

It is recognized that this criterion depends on the resultsof test specimens being directly related to the structure where confinementis provided, and that biaxial determinations used as evidence in Fig. .A.Iare not further influenced by platen effects yet uncvaluated by the referencedresearchers. Although these assumptions are perhaps chailengeabie, they areused here in an effort to be as realistic as current failure technology permits.

44

ILr

WE1GLER & 3ECKÊR

0 0.2 0.4 0.6 0.8 1.0 1.2 1 . i* 1.6

f 'c

Fig, STRENGTHS OF CONCRETE STîRJKCTpP TO RTAXIAT.STRESS IN TERMS OF UNÏAXIAL COMPRESSIVE STRENGTH f '

,- - o

f 'c

Fig. .A .2—TK1AXIAL COMPRESSION WITH AXIAL PRESSURBPKWITNATTNG

46

r-1- O

3 -

2 -

0,2 0.6 0.8 1.0

ffc

Fig. .A.3—TRIAXIAL COMPRESSION WITH CONFINING PRESSUREPREDOMINATING

47

•Til- <"

0.2 -

0 0.2 0.4 0.6 0.8 1.0 1.2 l.t 1.6 1.8 2.0 2.2 2.fc 2.6 2,

"sifc

PRINCIPAL STRESS CURVES FOR DETERXIMMG FACTOR C

48

REFERENCES

1. Bresler, B, and K, S. Pister, Faj lut e of Plain, Concrete,, under Combined.St_resses_, ASCE Trans. 122, pp. 1049-1068, 1957»

2. Bresler, B. and K. S, Pister, Str: ft&th oj; Concret ejjnder ..Stresses^ AGI Journal 30 5 No. 3S p-t>, 321-345, 1958.

3. Bellamy, C. J., Str e ng t h o f Cone re te un de r Comb i ne d _S t r e sjs , AGI Journal58, No. 4, pp, 367-380, 1961.

4. Balmer, G, G. , Shearing . ..Strength pj L Conerevi:e .under _ljljgh Triaxlal Stres^s-. Compuca t ion pj Kohr s Erive. ope ..aj__a_ Ji' IYJl» ^'" ^* Dept. of Interior,Bureau of Reclamation, Laboratory Report Ko. SP-23, pp. 1-13, 1949.

5. McHenry, D. and J. Karnis g t rerig th^ of n Corije re t e^ under Comb Ine.d Te ns_i l_e.rar.d^ompr_esslve.iiSt.res:s. Prôc. AGI -54, ppf 829-839, 1958.

6. Akroyd, T. S, W» , Concre te jande.r Tr iaxial^ , S t :res _s^ Magazine of ConcreteResearch 13, No. 39, pp. 111-118, 1901.

7. Hilsdorf, H., De ternunat io^-oj t^&^J^y^al^§trejn.gtln_ p f Cone re te ,Deutscher Ausschuss fur Scahlbeton, Heft 173, Berlin, 1965.

8» VilGj G, W. D. , Strength _o_f... Concrete under^ Short- Term Static Biaxial^S.treAg.1 Paper P-2, International Conf. on the Structure of ConcreteImperial College, 1965.

9. Wastlund, G. , New Evidence Regarding the Basic Strength Properties ofConcrete,, Betong 1937, Keft 3, pp, 189-205»

10. Weigler, K. and Gf Becker, Exper .ing_nt .s or\ .the k_7racture and De j pjrmaticmBehavior of Concrete at Biaxial Loading» Deutscher Ausschuss furStahlbeton, Heft 157, Berlin, 1963.

11. G 1 crab, Die Ausnutzbarkelt Zwaiacb- ; tger^^Drtick^s^igkelt Des Bétons i,nFlachentragwerken, Congress F.I.P.S.L. Pa, 1, Berlin, 1958.

12. Chinn} J. and R. M. Ziranerrnan, B ehavi or _iof_ PI a .in ._CQRC.re it e^ynde r Var lous,.Hi.gjh_T.r_i.ax.lal....Cp_mpjr_aj_sl._qni o_andjjnj ConjlAt o ns Air Force WeaponsLaboratory, Tech. Report WL TR 64-163, August 1965.

13. University of Illinois, Bulletin Sérias No. 405, Vol. 50t No. 29, 1952.

49

14. Heft 5/54, Vèro'f fentlichungen des Deutshen Stahlbau Verbandee, 1954,V

15. Heft 138, Deutsher Ausschuss flir Stahibeton, Uber die Grundlagen desVerbundes ZIrisehin Stahl und Bebern, Dr,-Juj. Callus Rehm,

16. England, G. L. , Lpns.._TcrTO 'Ifoe.mal.j gejAeA-lX'Striic.t.uresit Conf. on Prestr'essed Concrete. Pressura Vessels , London,1967, Group p, papsr 34 ,

17. 0' Connor, H. £. and J, L. M. Morrison, The. Effect .of. Meaai.iiStre5S_onLjbguInternational Conferenceon Fatigua of Metals, AShfi, 1956,

18.' Design Data ~ Nelson Concrete Anchor Studs, Manual No. 21, Nelson StudWelding Division of Gregory Industries, Inc.j Lorain, Ohio,

19. Da vis, H» S.> fJ c.ts_p ._Htgh __Tretiip.e.jr.aturg_.j|xpp.&ure. .Qn.._pQ.ncy.&te, AmericanKuclear Society Meeting, Gatlinburg, Tennessee, June 1965.20. Cottrell, A. H., Theg.ry_ of .Jrj. tl Fraetj.ira _in^Stsel and Its, App.l icatlon

to^ Rad 1 a'c i on. %b r 1 t_t 1 ement . " onferenc'e. on Brittle Fracture, p. 1,Culcheth Laboratories, England, November 1, 1957.

21. Pellinij W. S. and ?» P, Puzak, Practical Consideyatigns^inLaboratoryVessels., U. S. Naval Research Report 6030, November 1963.

22 » The_ TStru.c t ura 1 Us e of^Frest r e s sed... Concr e fcef in Buj. Id injgs.» The Council forCodes of Practice^ British Standards Institution, CP 115» 1959.

23 « Sjate. _cf. .;grh_e ^ Nuclear.À Critical Review of the Literature, ORNL-TM.-312.

24. General Atomic Staff, P.re.s_tr.es_s_ed__C_o.ncrGeneral Atomic report GA-7097, October 25, 1966.

23» General Atomic Staff, | 'res très s&ô i _ Co rj te React.or Ves _s e 1, Mode 1 2 1General Atomic report GA~7150, November 4, 1966.

26, Newman, K. , The Structure .A?J.-.. n 1 ee.?M J'XCB I --vP--CP.P.nS. eJ .eJ»Proceedings of luternational Symposium oa the Theory of Arch Dams,Southampton, April 1964.

27 . Ros, M, R. and ?, E< Speck, L£r _Tend_on.g__for jr e s ur e JVe.s.sjg I s_in.SiSiÊâL- SSSïLâSâÊiSSS.» Conference -on Prestreseed Concrete PressureVessels, Hfjrch 1967 1 Group E^ Paper 25.

50

28. Harris, A. J. and J. D. Hay, Rupture Design . of. _ jth,e._0j4bury. VesselsConference on Prestressed Concrete Pressure Vessels, Institution ofCivil Engineers, London, March, 1967, Group F, Paper 29.

29. Finigan, A. , .Ult_ima te__ Analy s 1 sx._p £f .the,, Dufig,e fi g ^ Vags. ! . Conferenceon Prestressed Concrete Pressure Vessels» Institution of Civil Engineers,London, March, 1967, Group F, Paper 31.

30. Anthony, R. D. , Development: o Sj:_a.î:.iat.oir JRe._( u._iTreiine_n_tsj_for. Redactor Vessels,Conference on Prestressed Concrete Pressure Vessels, Institution of CivilEngineers, London, March, 1967, Group Bt Paper 9

51

CONTKIBUTIQN K30M THE UH1TED KINGDOM

I. David?on

The following organizations are concerned in the UK:~ twodesign and construction companies, two customer organizations, theInspectorate of Nuclear Installations» and the Atomic Energy Authority*

The basic philosophy in the UK is to provide adequate pressurerelief valves on each PGRV, to design for the working loads, and fora nominal ultimate load factor, to test models, to carry out a pressuretest and to monitor each vessel throughout its life.

A committee of the British Standards Institution has almostcompleted the first Code of Practice, in line with the philosophysummarised above. It is hoped to extond the Code to cover LimitState Design.

It is suggested that more knowledge about the long termproperties of concrete under working conditions would "be advantageous,and could be facilitated by international collaboration.

53

Introduction1. In the UK there are two reactor design and construction companies, BritishNuclear Design and Construction, and The Wuclear Power Group, within which thedesign and construction of PCRVs is carried out by Messrs. Taylor Woodrow andMessrs. KcAlpine respectively. These two companies have carried out extensivedevelopment programmes both theoretical and experimental, including the testingto destruction of numerous model pressure vessels, and have completed and arecurrently constructing several PCRVs. The two customer organisations, the CentralElectricity Generating Board and the South of Scotland Electricity Board,purchasethe PCRVs from the two construction groups, and the CSGB carries out theoreticalinvestigations into stress analysis and experimental work on the properties ofconcrete. The Inspectorate of Nuclear Installations, Ministry of Technology, isresponsible for issuing licences and, therefore, must be satisfied on all safetyaspects relating to PCRVs. The Atomic Energy Authority is also interested inmatters affecting the safety of nuclear power stations and their current PCRV exper:mental programme mainly examines structures and materials under overload conditions.2» In the UK to date *f PCRVs have been constructed; two of these at Oldbury havebeen in operation for some tv/o years and the other two at Wylfa are at present ueinjcommissioned. A further 8 vessels are at various stages of construction and willall be completed and commissioned within the next three or four years»Basic Philosophy and Current DesignCriteria3, The basic philosophy to ensure the safety of PCRVs is described below:

(a) Reactor safeguards including safety relief valves are provided such

that under any conceivable fault conditions the pressure inside thevessel cannot rise in excess of the safety valve setting of 1.10 xdesign pressure, the design pressure being 1.10 x working pressure»

54

(b) Comprehensive elastic and visco-elastic stress analyses are carriedout for all possible combinations of loadings to ensure that the stressesthroughout the life of the vessel remain within conservative acceptablelimits»

(c) An ultiraate load analysis is carried out to show that with increasingpressure at ambient temperature, and assuming that the liner remainsintact, the vessel is capable of sustaining at least 2-J x design pressurewithout failure»

(d) Fully instrumented model tests are carried out, at pressures up to atleast the ultimate pressure, to show that the design behaves as anticipatedby the calculations.

(e) A pressure test at 1.15 x design pressure is carried out on completionof construction to verify that the vessel as built behaves as predictedby the calculations and the model tests.

(f) The vessel is monitored throughout its life, paying particular attentionto the loads in the tendons and the freedom of these tendons from anyform of corrosion.

4. Although, at the time of the design of the first PCRVs in the UK (at Oldbury),-carried out in 1959 to 1961, computer programmes for stress analyses were notavailable, the first vessels being designed to an ultimate load criterion of not lessthan 3» subsequent checks with the computer programmes now available, both short terraelastic and long terra visco~elastic, have shown that the stresses in these firstvessels are within acceptable limits. Detailed stress analyses have been carriedout at the design stage on all subsequent vessels»5» The purpose of design criteria is to provide sufficient margins to cover threesorts of uncertainties?- loadings, calculations, and strengths. It is consideredthat the provision of relief valves and other safeguards in the UK type of gas-cooled, reactors removes any possibility of significant overpressure, and temperatureexcursions are necessarily slow» Therefore, the design criteria are required todeal only with errors in calculation and unforeseen loss of strength in the structure*The elastic plus visco-elastic design method, with appropriate code stresses, is

55

the classical and most direct way of providing such margins. The ultimate loadfactor is only a different way of providing the same margins because it is neverintended that the vessel will actually be subjected to the ultimate pressure»6. All the PCRVs built to date have been tested in the form of scale models,instrumented to measure strains and deflections and pressurised hydraulically inthe elastic regime and up to at least the intended ultimate load factor. In somecases thermal loads were also applied. It has been found that the elastic behaviouris in close agreement with the analysis for the cold model. Of courset temperaturegradients and creep have effects on the stresses which, whilst they can be estimatedby calculation, can not be completely measured in models.?. During the last 3 years a small Committee set up by the British StandardsInstitution has been drafting a British Standard for PCRVs based on current practicein the UK, and it is hoped that the final draft will be completed in a few monthsfrom now. The draft Standard covers in detail the design, construction, inspection,testing and surveillance during operation of the concrete pressure vessel completeincluding the concrete» prestressing tendons, bonded re-inforcement» liner, penetra-tions, closures, relief devices and permanent instrumentation. The insulation andcooling system are also included» but in brief outline only. The design sectionsett; out the requirements for the loadings to be considered and the stress analysesto be carried out, with guidance on the methods to be adopted for the short termelastic, the long term viaco-elastie and the ultimate load cases. Permissiblestresses for both the concrete and the steel are specified. Whilst much of theDesign Section of the Standard is of a mandatory nature there are, inevitably atthis present formative stage of PCRV design, some Sections which are of an advisorynature with further guidance being provided in the form of appendices. One of therequirements is that the design should be undertaken by a qualified engineerthoroughly experienced in the field of PCRVs.8. In preparing the Standard it has been appreciated that in the concrete structuresfield internationally there is a move towards adopting limit state design. CurrentBritish pressure vessel practice considers in effect two limit states, that of

56

permissible stresses under normal loading condition's and that of collapse of thestructure» As an alternative to the present design methods a more comprehensivelimit state approach is given in very brief outline only in the draft Standard,It is recognised that much more work is required to examine the application ofa full limit state approach to the design of PCRVs. The Drafting Committee intendto study this aspect further after publication of the present draft, with a viewto incorporating, in the form of an amendment, a fuller limit state approach shouldit be found to be practical.9. In the Materials Section reference has been made, where appropriate, to existingBri'L-i-ih 3U--ndar> s but the requirements relating -pacifically to PCRVs, particularly

on material selection, testing and quality control have been set out in some detail.Safety Considerations10. It is considered that the basic philosophy and criteria currently being usedin the UK, and which will shortly be published in the form of a British Standard,

provide adequate safeguards for PCRVs used in conjunction viith gas-cooled reactors

with designs incorporating safety relief valves and other devices to limit the

pressure within the vessel- The safety of the vessels depends on:-(a) comprehensive stress analysis confirmed by model, testing limiting the

stresb',o to acceptaole levels throughout the life of the vessel,

(i,) an adequate nsargin between normal loadings and those required to cause

failure, confirmed by stress analysis and model testing.

(c) careful selection of materials and control of quality throughoutconstruction.

(d) confirmation that the vessel as built behaves as predicted.

(e) knowledge of the behaviour of concrete throughout the life of thevessel.

(f) knowledge of the behaviour of the prestressing tendons throughout the

life of the vessel.

Lon g-1em 3e haviour11. As has already been said, the theoretical analysis of the long term behaviourof a PCJRV under mechanical and thermal loads is now feasible, provided the behaviour

57

of the concrete is known. la this respect the material cannot be divorced fromthe structure because moisture movement, which is important as regards creep andshrinkaget is largely conditioned by the sise of the structure and by its teraperaturdistribution» Moreover, many of the experiments which are necessary cannot beperformed on a small scales and long times are required» Under these circumstancesthe value of large-scale long-term experiments is obvious, as from them some under-standing of the phenomena may be obtainsdj and ths ong-term behaviour of any otherconcrete in any other structure may b© deduced after relatively short-term tests»12* The failure of concrets under multi-axial stresses is a subject which hasengagod very much attention for many yeara^ but there still seems to be room fordiscussion concerning the safe long-tens multi-axial working stresses which mayb » appropriate in a thick-walled structure such as a PCRV» A further distinctionmay be appropriate as between mechanical loads and strain-induced loads.13» Concrete studies of the types just mentioned ara being carried out in severalcountries, and it could be of mutual benefit if some steps could be taken to avoidduplication, and to share the resulting information. It is therefore suggestedthat ths IAEA Secretariat might consider setting up an information office, whichwould list the nature of the experiments being carried out in the different ccuntrieAs a further step a snail panel of experts might be able to help in co-ordinatingsuch experimental work, and in considering the results obtained»Economicsl*t. The development of PGHVs has already shown economic benefit in relation togas-cooled nuclear power stations. The concept is relatively simple, materialsare cheap and the techniques of construction are well established. It is unlikelythat future developments will show large economic benefit, but appreciable savingscould perhaps be made if it proved possible to operate the inside of a vessel ata higher température than is usual at present»Conclusions15. The basic philosophy and current practice in the field of PCRVs in the UK

G'

have been briefly outlined and will be published shortly in the form of a BritishStandard» It has been suggested that in the area of concrete as a material»

58

international collaboration would be rewarding and the IAEA, mighc like to consiûersetting up an information centre and perhaps a small permanent Panel with theintention of co-ordinating such international collaboration» The ISO are alreadylooking into the question of standardising, design and construction procedures forPCSVs internationally and there would seem to be no requirement for further work onthis aspect.

59

W'W O'fi REACTOR V£8S!CLS

C08 CBT?TF; 17; ruo

5. PetrcviS

The des'g.a for -••;• two- v^al • oon-.T^ ta r«aci"crr p

.'•K)3sj. hi^s been a>37elnp-jd it V'^ros!•-•. i-i» 7'h~ #:* r/e^t ' ta 1 •/.'5s?^5 is

p."63tri?ft?e<I "'.'it;h st-:.j'. cabj.:1.? .:« ;,hs JB^a] wr-iy, '";:ere are no s^b^os

in the inrerns . veHse ; soO. t!-• ï vesse": :c ..;:-fK -,r-osseo b:/ a hyira.:] icsysteia» Jnser ta r i oecwc-.-i '.r-e outer r,r;.o -jhs inner vo- ' - t fe ' s . 0:-i; O'ii L? . t jone

sao'fj th-it, fo:.- S'jch a -/c^-ftol a;?icepT;ior., eo îi.o-pJ c:; or.ild ha m<-i"j<~ for

about 23 percent in cables ar..? about 40 percent in concrete, usii";*: at •

ibe same time the same prestresain^r as for single walled vessels. Atpresent,, the raodsl 01' such a vessel type is und<?r construction In th*>Insti tute of trig 3R Serbia for Testing ?lateria]s3 in Boograd, Theresults of mathematical analyses of the action oi% seidmic forces arediscussed »

61

Ho Kuclear Power Plant has aei yet fo&en coastrueted inYugoslavia, i^s construction,, ho^sver, is being consideredfor tiie next years* ïhersfore it cannot be th£ creation ofan owa expedience in thie field, szcept for sanie investiga-tions made in the past, y a are»

Kelyicg oa the SwediseJa conception of the BKW k-saetcr,studios were started hers on the hoiae possibilities for theconstructiou or a près tressed concrete pressure vessel for a200 to 500 MW Power Plant.

In the first phase of this research, fairly detaileddesigns for the eaid pressure vessel were developed, leadingto a better understanding of the problems to be faced within this connection*

One of the most serious problems, set before any con-structor of prèstressed concrete pressure vesselst concernsthe temperature in tiie vessel and the stresses caused "by itsgradient,

It is a well known fact that» by enlarging fch© wallthickness, thermal stresses are not efficiently reduced,which means that the thermal stresses have to be covered byprestressing correspondingly the wholo concrete cross section-Besides the waste of steel, this increases the difficultiesin the location of steel caules and leads to extremely highpressure stresses on the internal margin© of the concretecross section at the moment when the re-actor is put out, Nor-mally, a solution for this problem is sought in the construc-tion of a sufficiently efficient insulation v/hich is to pro-tect the concrete against the high temperature reigning in-side the vessel and to reduce the teaperaturs difference be-txveen the in- and outside nargines to an acceptable gradient.

62

But, on principle, the temperature gradient, could be ni&ci«3 ac-ceptable aiao 7/1 tjQ s les -3 3ffieleut grid cheaper Inside insula-tion» provide à we acL'p": a cheap outside insulation,, ^as-ing thei.eiiiv>era«u,re at •;£•" out-side surt'scs "Do tfee d^sir^d b.oagi'.ta !<b.i«iv ou Id naturally Incr~s»e -~ae temperature of tU>? c^cr^te as :,v?hcle co a FSOL-O nlgh^r level» J'llie dsftfi/^Jjiatlon of ths £ afiipsr-£t-,u'o, afc vliio.û a ooa^r^le preesu^e vessel could be technically

secce^nful.iy ano. scon-omit-, ally designed, ia> no dounty a com-pies ts-:ii£? and it t - a n be solved only after lorv-, and ex-vencicests,

soc howevar^ 01 the future oMLwer to this ques-fciorij w^ think that;, ^ith so'-;y lûaulatiar- z.r..z .v.oagiîM; of theou^'.-'a^d M£^;se.l Cur/aoa-, alor.;; vrl, î:,:i t;ic inside ineulaticnj usedsr f:j,i% fcuonoraiss could b« obtaiiied which wou3 n , ju0tify thap^icstion of ui^ch a solution, and w era Id yield at any rate anc.>2 convenient c t a i f v 01 stresses.

Oorjsidejred. wao also the question of two-wall concrètevoxels The- conclusion wa^ reached that, by inserting a Ly»dralic eystes bctweazi the extor.as.i and internal vessel , onecould, exorcise an active influence upon the whole state of

in tlis vesssi»

•£o tida purpoc.0 a design for a two-wall pleasure vessel'K.-j^e.l has been deve loped» !J!he external vessel is prestresaedx'rltli steel câblas in the usual way. In the intovnt'i vesseltJ: o-\- t..::-; .no cables» and ti*is vessel is prestressed by a hy~dr-.-vul'.c R/Ptosï, .'inôer-ted between the outer and the inner ves-~fi$;ls, Ihe './: 11 'crloknesG of bobh vossele»f tiie quantity of cables:.,^ \\li-: t 'Ttsrr.ftl T^t'sel, r^ria the î^yaraulic pressure in. the sys-tc.:. -;-.- '.«^-'\j i;-i^ vessels have be-vn determined so. as to satisfythe icllowiiv; coad.itiore;

( • - < ) it full lot-diA'a LM: the roactor feLs ^tress&B at thecr ls jca l pointy in both vessels oo^ix: ôcvvr to î:ero (or to oeJ.r i;:i',un àJtorrïJ r>ed prc\"3uxe; »o tbat th.e whole .^yr.teœ is fully{XTjprosccrl.

ib> y/hen Vrv- r^<?ctor is ^,at c>,).t bat trie vessel ie stillhot, the critical stresses in both y easels do not evened. I;);®

63

desired pressure* This Is ootained oy a reduction ofthe pressure in the isiddi® hydraulic system to a determinedquantity (e.g. to 1/2 of the pressure necessary for prestress-ing the ¥ ess el &t full load) .

Simultaneously the inserted partition can serve also asa température regulator in the vessel both, when setting thevessel into operation ana in the COUPS© of its running,

Calculations show that, in this way» économies oould belaade for about 25 psrc©at in câblas and about 40 percent inconcrete» using at the same fciias the same prestressing as forsingle walled vessels. The Ijydraulie system in the partitionwill costfc of course, a csrt&in amount which reduces the effectof the said economies.

To such a vessel conception, besides the rejsarks concern-the introduction of the hydraulic system which becomes an

element in the functioning of the vessel, remarks areusually made which are common to all two-wall constructions,insofar such, constructions intend to jaake economies in cables.it la the question of the safety factor at failure,

It is commonly known that the ultimata bearing capacityof concrete veassls practically depends exclusively on the quan-tity of steel cables in the vessel * In some countries a safetyfactor &+ :?ci.!v"-~ jr L«3 is requested for pressure v@s&ela. Thisis the usual measure applied to cone t ructions of bridges andbuildings in près tressed concrete and it is always met with, ifthe norms are rsspected concerning the allowed ptresses in con-crete and steel*,

if we consider a pressure vsasel constructionj abiding by all the other conditions, makes economies for

a determined quantity of steelj the request that the safety fac-tor at failure should attain 2,5 will possibly not be met with.In our conditions, one half of the cable economies is due tcthe geometry (smaller cable circle) and this part has no influ-ence upon the reduction of the safety co«fficiea.t at failure.In other words, the sal' -sty coefficient for cables at failurewould be 2,25,

64

It la a question whether aucn a demand Is jusfcifled-O'ji1 c^:Uaio.i:-.9 the- e&fd tv of p^essiire vessels sLould BC~ be

ated by the loss v^i'ler- tears the cs.jl ^s but l*y the 3. o,-,dwiv.U.'a p.L*'.';l'ace?3 fiii;:;:.,;.' • * j- ••: ^n.ftl with ^issuic-fss becomes ^suff

. he

Tv-.<3 said, model r»f s double vessel IIP s bear! developed orsthe cojaeeptloij of preser^.a^; l::;^- cfe/st^ t-gain&-c oecu,x-rence offissures» cler«-lsj it should be kept in ^acd that it is Indis-beno^ble io préserva a^ainr:t riasursy o.al,y tne inaei •ysasel,whi*5 tf in me c-'ita.v vres^^l^ xn case of an a c-rid*.ni, fissures eanbe aiimitteu. By I^^raasing tlie prs.issyri8 1^ the hydraulic systemof t'-:e partition, it i? possi'b'ft ir> (••.âi^tai.a the inr>er vesselin a create without ^isgui"^ - rVv^ j at & c^rr.?iu«rabi 2 i^jroase cfthe pressure wif-".^* the */ass©l: 7rli*i-ein. i-at ortte.^ vessel willsuffer acre or i^cs fisùu.:.-iiiâs 02 ^\lll remain nntouctied dspend-iiig on the qu&jatitj for rhicl: tiie pressure- ', r «ne ve^wl ho.8been raiaed. In short, it IB possible to pi 00 £!•,••£ tiie Innervessel to l>0 intao1;s wlti;o'v". jrisaur-ss,, pryctic.*j.lj qi-, ;,te ru*arthe pint of failure of tlj.e outer Td-sB-sli on condition that thehydraulic sys^-e^ of th,B pa.riitr.on. lr not too tru'-rt,

This properly vi sucb & tvv'o wa1 lad y £•;£-•-• ;. , r?ajne3.y, thatits safety factor aça.lsst o-^c-ur-r.'-ncti oi fî/j^'.-7-es c&r» oe raisedconsiderably above- >,>.-.> usual V£,lue:.;., w&ea it eeHe.'.itluIly d3.i-fsi-ent from tiie one-wall veswsls ted ^aices il*o me .loi "-eptingworth ¥?hii<s»

At prtssijt, the Insfciu.-te ci t";e b .- Serb La Tor TeMaterials ia .3elgra&9 is o«-.\;. pied wit A t&s coris true 1:1 c.«suaii a model in. order "so verify the advanced statemeuts,

Parallslly --,=?ith this t^sts ars bei&x /nawir^e £or cable-;; &x increassô. températures « -i'sst-fôd wer-s v7i0 9 raa of Austrian origins» Juecr^uival properties of i'he Jre9

5 inclue1 ln>r rslosatiojj, vere invoirtigated e.^ t'srapsratu.v-es of 50-100-150 and 20. CC rlu-ough a period of 30 daye.

Based upon these tegts we coaclude thai., at J.eaat whatthe wire ia conosr-nedj. tb.e tc^ipc- rature of the concrete, in

wkLeh. t&e- wire;?-. c.ra placed, <>oald be some^h^t raised, -Pheincrease of tï-^ relaxation at t-life température <>x 50 °C is vsrysmall as compare. with tb/'j rslâX-Tfcicn at fee roe.v temperature ,f£ir''.s increase is tolerAoX^ ev&a nt a temper^'aa^ of 10'} '"'0,. ,vbhigher tempers, tu^es «-;!£• relaya ::i;>ti cons.:. os:.-snl.y IriC.^essti'r sotliaf: It Is a q^T^i'lou whc-th-;r e fur-the-?- raAsiss, o£ tb.s «fjapttr-aturev is technically and ftooso/iiically justified.

On the o'cher hand, ii' by «)aa ajct^xral ?.nsu:ia ïloiï ovessel the température ^ra-lierit 01., its our.raaes woula b£:

for about 2G - 30 °C., the rherrA.1 •.-.i-i""^^ and the qtity of steel for their covering would be cop.^iasrsbly reoucedSince, with all that, ws r"maii- 7d.ti-.ir1. toe- ,ilta\, te or" temperatur« of 50-70 uCj w« «'^alù .o-:-t ttxpsct ^.n assent Is.-, change ofthe "tiahav.loup o:1 r-\& steel fo.- i-ho ct^ls;:.-

»

.ly with the wire trnj t fSj c<uacrete investigationsat inc^^aaed temperatuz5^?.. were organised too» jA aeries of ^e^tsof ., selective ciiaraster i7.l,th aiicro-concrets has been organizedwiiarain csiaents w^ra seleo&ed whicjJ-i :vould. come into considéra-tion for further invastig-ationc-î, ïn Aks «oursea£l preparationsof tests referring to the behaviour of concrete at inc. reapedtempératures which shall star*, in 197^ and which v?ill give, aswe hopes» a», answer to the quesvioz-1 at what te/apery,vure pressurevessels of prestressed concrete can bo t «clinically and éally designed y.n-1 realiaed*

^ a gre?/v part- of th* Yugoslav territoj;/ IB poteij.~exposed to th<s :-;c!;io:? of seismic forces of destructive

power and since fche location of future nuclsar power pit /ce ispossible for the greater part just in ev.ch regions^ the analysisof the action of seicrolo forces upon auch i^^i- allât ions ha# beenapproached too. So f£.r» sucn Iviveaf-iga biurjo were ma.de & A O J - U ™cively by means of K\atlie;..ai;lco'.l analyses, but the constructionaud test of a certain iiLiaiber of s^iifipie dynainic Models ars pro-vided for.

Basjeci upon tae satheKatiool analyses ^ade so fai\. w»should sayf that r.aseive eoncrete presrjuro veôsela, il they arewell fourded, ore not directly enoangered. by tïie action of seis

66

•aie- f" orcey . BUÏ tkis ne -xi net rrfer to tae elemtmtB of tiie

v3B0ei, ^'or thé ^ors^ïi* a -mà?. £:*:•• attention was give-u to tàe:;li;ô'.> o?' ,!:;: i/ebîiviour cf ae iaflaiiysac'-.L'- ol -îir.eiits , as weïl aaof rae tars for n>ieckl;=p: •',?:••; n>-" l . :"#•-; oui; ^ha r- ertctor.,

/ '">ftLy;;aà wereni^ vn^-Mat:: cal modela of tudsst) oscillations?wliil e moisi Iijvestigatlo.-i GJ. the u;-,hav:loi3r of tne core elements

'tlvvua&s la pr-vided i'or.

DfOJ

strana

hydraultk tacks

( 4=1

68

Transj s ted fro?s Frc h ~*;

Fît SNOW S.".-2HJ «SCg UP TO

Af~ier an his tor.: ::•« j s\:r-ve,y .;f frsiich b-j.OK.i.r rotund ic ?d : -V : r , ,J;he paper c'wscr; r-c^ Mis coo^, ; L-J-.Î ''.ed oro«i",r,: c'j.cTiad out ox% ]ongff-""r:s c-Vfes?;int'\ cie /.-- , cpœ.5r t o [' i;e?. ocnc:-er,cs, ^eat trajisfcr a tulles

ar^a d^velop:s^al- of ii-v-c'..-,,1;: OA lechnolo^iss.» ça! ou la iior. programme.1.',OK^-_Hureïne>it/s on ac tua l vssj j&I^, ar;d sciiaa deaie-': pr -^bJeras^

Tua paper :lj.aoys:;t?s h n C - . - s .in a as Pety prot iams in nucle«rapi.'.i ic-3 ï.' yn. The gênera; s a f e t y yi, -\ n i •* vd is very r i igh, bur t.he

a'. 'ilitj to warn hf-^'ori ''ai. lu;1?* o.ue ^,o vc-.iixening: of some cc'-joon;r:it

i s cr: t « ' way a es t-V . '.. ched ,

ïhe fr&nc;-; o'rfici^l re-ru.'! ati or: ( ira^' t ""or go.:; reactor PC?*/1 s

i:3 Kursin?..v?.Z9ci. wL~L..*i ,--. pf i . ' îonal cp . in :on abou t ^.he fu tu re ev,»at ion

of î/'ia'ï "-sx t. The oc-onoraics l'or 7 'L 'p - / ' , ^ are aor- f ieRted.

I. BACKGROUND1.1.0. It was decided in 1955 "to use prestressed concrete as the material

for the pressure vessels of the Marcoule G.2 and G.3 reactors (forproducing plutonium and electricity) for the following reasons

1.1.1. The vessel, with a diameter of 14 m, had to withstand a pressureof 15 bar, but the only" vessels of that diameter which had beenbuilt up to that time were those at Calder Hall, for 7 bar pressure.Any solution involved extensive extrapolation from the existingtechnology.

1.1.2. The impressive results which prestressed concrete had yielded inother fields and the special experience of the French engineeringand design firms created a bias in favour of this technique, andits advantages were confirmed during the preliminary studies.

1.1.3» A particular advantage was that the vessel could be built in avery short time. There was a lapse of only 12 months between thedecision to adopt the technique and the commencement of pouring.The completion of the first vessel took another 18 months (early1958).

1.2.0. However, in view of the rather high cost of concrete vessels, theElectricité de Prance (E.D.P.) decided to choose steel for itsEDF.l and EDF.2 reactors at Chinon. This solution proved to beexpensive and time-consuming because of alterations in the techniquesof execution, so that when the question came up again for theEDF.3 reactor, the choice fell on prestressed concrete. Thefollowing factors were taken into account:

1.2.1. Prestressing with conventional cables and the vertical arrangementof fuel channels were intended, to make the execution simpler thanat Marcoule, which had external hoops and horizontal fuel channels?the use of internal heat insulation was designed to simplify gascirculation;

70

1.2.2. Above all, it was found that prestressed concrete lent itselfmuch more easily to extrapolation to larger dimensions thansteel, a factor which is indispensable in developing graphite-gas reactors.

1.3*0. In practice, the only difficulties in constructing the vesselfor the EDF.3 reactor related to heat insulation. Onoe a varietyof solutions had been worked out to the latter problems, thetechnique had achieved maturity. It has been applied abroad,and in three new vessels in France.The graph!te-COp series has now been abandoned; there are noconstruction plans yet for the graphite-helium series (HTGR); theheavy-water and natural-water series do not, at present, useprestressed concrete, and fast gas-cooled reactors are not beingstudied in France. There are no other construction plans underway.

I.4.O. This could mean a serious lag for the French industry.However, the apprehension does not seem to be warranted, and forthe following reasons:

1*4*1. Engineering design offices and companies have been studying otherapplications of prestressed concrete, particularly for reactorcontainment.

1.4.2. Study programmes continue to be carried out by E.D.F., CEA,contractors and design offices. These programmes are now beingco-ordinated and considerable funds have been rm,de available.

1.4.3. The co-ordinated programme of research includes basic studies aswell as the analysis of technological developments which can beadapted to new series.

1.5.0. The present problems are to co-ordinate the French programme ofbasic studies with the corresponding programmes of other countriesand to bring the various standards and regulations into harmony.In addition, French industry can make a contribution to techno-logical developments which are useful to other countries.Below we give further details on the French programme beforepassing on to the technical problems proper.

71

II. ORGANIZATION OP FRENCH RESEARCH

2.1.0. Up to 1969, research was carried out in Prance within the fol'lowingframework:

2.1.1. E.D.P. entrusted various civil engineering firms with the designand construction of its four pressure vessels in Prance as well aswith the design of the Pessenheim vessels, which are not to bebuilt'. In its Grenoble and Renardières laboratoriesf it hascarried out theoretical studies on the types of concrete underconsideration (from the point of view of rupture, creept moisturemigration and instrumentation). It took part in the analysis oftests on models and in the development of measurement systems forthe vessels, for the surveillance of which it is responsible.The E.D.F. Centre at Chatou has developed calculation programmes,and E.D.P. has promoted the testing of some original prestressingtechniques (the two-layer method and PATIN prestressing byexpansion). It is responsible for operating and exercisingsurveillance over the pressure vessels of its reactors.

2.1.2. The French Atomic Energy Commission (CEA) operates and exercisessurveillance over the Marcoule reactor pressure vessels /2_/.It has carried out safety studies on the processes of rupturein the case of tendon weakening and asked the laboratory of theEcole Polytechnique to study triaxial deformation .in a certainmicroconcrete. The Saclay Concrete Laboratory has developedspecial concretes for radiological protection, heat insulationand high-resistance. It has al^o carried out structural studieswith the help of the laboratory of the CEBTP.

2.1.3* The following French firms are dealing with pressure vessels:- Pour construction firms (Grands Travaux de Marseille,

Campenon-Bernard, CITRA, Société Générale d'Entreprises);Engineering firms (SEEE, STUP, Coyne and Bellier, Bertin,etc.) ;Industrial firms (CAPL, Neyrpic, etc.)»

"During recent years, these organizations have carried out theoreticaland technological research, either under contract to E.D.F., CEAor EURATOM or financed by their own funds.

72

This research included:

- Studies on the pressure vessels under construction, withreference to the suitability of the materials and methods,and tests on large models;

- Studies of methods (two-layer pressure vesse]t prestressingoy expansion, "boiling hee -y-water reactors (in relation toCEA invitation to contractors to provide designs), removablecovers, pressure vessels with hot liners, hoops with steelbands, insulation, etc.).

2.1.4. A committee comprising representatives of the above-mentionedorganizations was set up by the Service des Mines to prepareofficial regulations relating to pressure vessels. Theseregulations have now been submitted for Government approval.

2.1.5» Another committee has prepared, at the request of the InternationalOrganization for Standardardization (ISO) and in collaborationwith British and German committees, a report suggesting whatstandard-zation work the ISO could do in this field.

2.2.0. Since 19&9 there has been official co-ordination between CEA,E.D.F. and SECN (Société d'Etudes des Caissons Nucléaires(Company for the Study of Nuclear Pressure Vessels)), organizedby the four specialized firms. The studies are financed, on anequal basis, by E.D.P. and CEA and, to a large extent by SECN.These studies are carried o-i/fc by the three contracting partiesor are entrusted to specialized laboratories.

The specific safety studies (processes of structural rupture)continue to be conducted by CEA. Other studies, theoretical ortechnological, may also continue to be carried out independentlyby con ractors. However, it is the co-ordinated efforts ofthese bodies that constitute the bulk of the French programme.

III. FRENCH RESEARCH PROGRAMME3.1.0. The overall co-ordinated programme includes:

A. General studies and researchI. Concrete

(]) Study of the properties of concretes:

73

- Instantaneous deformation and rupture undertriaxial loading (under way);

- Creep under triaxial loading (in preparation);Programmed monoaxial creep (under way);

- Moisture migration (under way);Relaxation of thermal stresses.

(2) Study of high-resistance concrete (under way).(3) Study of insulation concrete (under way).

II* Theoretical study of heat insulation:- Convection in porous media with a permeable hot

face (under way);- Natural convection in a cell assembly;

Convection and phase change in porous media.III. Hot liners for water-cooled or high-temperature reactors

Economic study of the method. Preparation of assembliesfor the fatigue-testing of liners and anchorages and corrosion-testing of anchorages (under way).IV. Industrial insulation (CAPL, fibrous insulation)

V. Calculation programmes:Three-dimensional programme (under way);

Two-dimensional programme with plasticity (£3t_/(temporarily suspended).

VI. Interpretation of reactor measurements.

Study of pressure vessels- General design according to the reactor type

Components, in particular penetrations and covers,cooling systems (under way)

- Designs and models

74

3«3«0. The co-ordinated French programme does not cover all the problemsthat may arise. The selection which has teen made on the basisof the economic and safety aspects of the various studies, canstill be extended.The French organizations are interested in any internationalexchange of ideas relating tc all the problems in this field.

3.4«0. We shall beg-in our analysis of the technical problems with asummary of the main properties of prestressed concrete constructionas far as nuclear pressure vessels are concerned.

IV. REINFORCED AND PRESTRESSED CONCRETES FOR REACTORS4*1.0. A distinction can be made between the following types of

construction:4.1.1. Standard reinforced concrete.4.1.2. Partially prestressed concrete containing hard-steel tendons,

which neutralize a part of the tensile forces acting on thestructure.

4.1.3* Prestressei concrete proper, in which the tendons balance allthe permanent and transient tensile forces, leaving only com-pressive stresses or, in some cases, low tensile stresses whichcannot lead to cracking.We should, however, point out that higher tensile stresses cannotgenerally be avoided in localized zones like those of tendonanchorages where the tensile stresses are transverse to strongcompression. In this case, tho concrete is reinforced withpassive reinforcement steel in the direction of these tensileforces.

4.2.0. Reactor containment buildings, which arc usually under low loadsand are subjected only rarely to high overloading, may be builtof reinforced concrete or partially prestressed concrete, wherethe prestressing tendons are adapted to permanent loading.Fully prestressed concrete, however, has the advantage of beingleaktight and thus obviates the need for a metallic leakprooflining.

75

4.3«0. A reactor pressure vessel, on the other hand., is normally subjectto the nominal working pressure; there is not likely to be anyheavy overload. The aim is to avoid cracking during normaloperation, first, because the high-tensile reinforcement whichis required by the economics of the project would then be moredifficult to prctsct againct corrosion and, second, because thesteel liner is DCB+ applied to an uncracked wall. These con-siderations ordinarily lead to the selection of prestressedconcrete proper.

4.4«0. By comparison with the prestressed concrete used extensively incivil engineering construction, this type of prestressed concreterequires special study for the following reasons:

4«4«1« In view of the seriousness of any possible rupture, safetyconsiderations become a matter of national importance.

4»4»2. Because of the cost of the plant connected to the pressure ven*"".the latter must have very high operational reliability; this canbe ensured only be relatively elaborate and costly arrangements.

4.4«3« Because •< t operates under pressure,, the technological arrangementsand the stresses exerted on the materials are different from thoseof conventional installations in which beams, columns and thinslabs predominate. More particularly, the concrete is used ingreat thicknesses and the stresses and strains must be known inthe three principal directions.

4.4«4« Because of the rapid development of the technique, the time neededfor finding a bac-: c solution is relatively short.

V. SAFETY OF PRESTRESSET) CONCRETE PRB-'SSïïRÏÏ VESSELS5.1.0. The advantages, in terms of safety, of concrete pressure vessels

over their only rivals, steel vessels, would appear to be thefollowing:

5.1.1. Resistance to pressure is ensured, first, by the concrete, whichdistributes forces without serious risk to itself (except in thecase of certain désigne in which heavy shearing stresses canoccur) and, second, by a system of metallic tie bars which arenot interconnected and are therefore not, very responsive to highstress concentrations; moreover, since there are generally largenumbers of them, they are not sensitive to the rupture of anyone of them.

76

5»1.2. These metallic components are put in place without welding whichenables one to be certain of the strength characteristics of thematerial.

5.1.3« The pressure vessel can be viewed from the outside and incipientcracks can thus be detected.

5.1.4. Since the pressure vessel is a massive and cool structure, alarge number of monitoring sensors may be included in it.

5.2.0. The purpose of the safety criteria will be to ensure that thestructure can give proper warning1 of deterioration before thereis any risk of rupture and that its resistance is maintained,mainly because the tendons and leak-tight lining remain intact.

5.3.0. Some concern may be felt concerning the possibilities of long-term failure due to the inaccuracy of the assumptions madeconcerning the behaviour of materials. Aparb from the possibilityof generalized corrosion of the tendons, we do not think this isa serious riskr since:

5.3.1. Deterioration of the concrete would have to be considerable tohave any mechanical effect, except perhaps in the case of largeslabs not reinforced by tendons, an arrangement which so far hasbeen avoided in French reactors. Such slabs should in our opinionbe reinforced with passive steel until there is some definitewarning sign before rupture takes place.

• *M%

5.3.2. Reduction of prestressing by relaxation of steel stress resultsonly in a risk of strain, for which warning is given. <!jn thewhole, this relaxation can be foreseen and be measured on tendonsfitted with instrumentation»

5.3.3. Irregular distrioutions of tempe rat ure can give rise to undetectedhot points only of a very localized type. At such a point thermalfatigue of the steel liner is possible, followed by cracks withslight leakage of hot fluid. This fluid will pass between theliner and the concrete and will disturb the thermocouples^in thearea before any ap reciable pressure occurs within the concrete.Because of the precautions that are taken, this pressure wouldnot caus' rupture. We would advocate, however, that a drainagenetwork be installed behind the steel liner to detect this sortof deterioration earlier.

77

5«3«4« Because of the shrinkage and creep of the concrete, the relaxationof the steel and the settling of the bearings, the pressurevessel can develop in a way that is not fully foreseen. Inparticular, tensions can appear in the concrete. Ib is veryunlikely, however, that macrocracking could appear in areasthrough which prestressing te.idons pass (areas which are heavilyreinforced), thus giving rise to corrosion of the tendons. Norcan such macrocracking take place behind the steel liner whichis applied to the concrete and causes it to creep transversely;that is it cannot take place unless the overall deformation ofthe pressure vessel is extensive, in which case it can be detectedbeforehand. The essential safety components, the reinforcementsand the steel liner should not be weakened. u,'e believe in anycase that such deformations can be at least approximately foreseenand that reinforcement steel should be used so as to preclude anyraacrocracking.

5.4*0. We give below a summary of the proposed French regulations forthe safety of pressure vessels.

VI. SUMMARY OF FRENCH REGULATIONS

The arrangement by articles is the same as in the regulations, butin brackets we give additional information taken from the documentationaccompanying the relevant decree:

Article 1 - Date of entry into force (proposed for 1970)The decree applies -co pressure vessels containing a non-corrosive

gas (and not for example to pressure vessels for water reactors)^Article 2 - Definition of the manufacturer who is solely responsible.(See Article 26 for the supervising engineer.);

Article 3 - Exemption from identity marking and hydraulic testsjArticle 4 - Definition of the documentation to be submitted beforeconstruction;Article 5 - The pressure vessel is built of prestressed concrete,having a steel liner and a means of thermal protection. Irradiationlimits.Article 6 - Materials and devices used

(j) Prestressing system. Prestressing method to be chosen by the

78

manufacturer on his own liability, subject to justificationof the action taken. Elongation of the wires without necking;at least 4.5$. (List of the tendon properties, the values forwhich must be guaranteed),

(2) Types of concrete (qualitative recommendations);Article 7 - Engineering' design(1) Under working conditions, the three-dimensional concrete stresses

must fall within the "safety volume" obtained by applying afactor of /3 to the "range of resistance to rupture" in thespace °it °2, a'}. The tangential stresses on the inner faceare compressave.Under thermal load without pressure, the reduction is 0.5.In special areas, a figure of 0.6 is allowed, provided that somereinforcement is used. (List of stresses, earthquakes excluded.Definition of the design pressure P. Strength range obtainedfrom Caquot's intrinsic curve, modified if necessary by theresults of triaxial tests. Elastic design (obligatory), plasticdesign (desirable)),

(2) F'or transitory thermal stresses a reduction factor of 0.5 isapplied to the concrete strength range,

(3) Safety factor of 2.5 against rupture due to excess pressure,(4) In case of rupture of leak-tightness and occurrence of pressure

gradients between external ; nd internal atmosphere, no ruptureof steel reinforcements.(Checking under increased stresses: recommended only, increaseby a factor of 1.2 for pressure and 1.5 for temperature variations.The "extended" safety range is 0.7 R for compression- and 0.95 Rfor tension.);

Articles 8 to 11_ Model testsModel of at least 1/6 scale, unless the pressure vessel comes

very close to a preceding- one in size.

Hydraulic test for rupture;Articles12 to 16 Construction

Concreting programme, sampling. Prestressing programme.

79

Reduction of strength range to 0.4 during construction. Protectionagainst corrosion of steel reinforcements. Initial prestress:limited to EQ ^ (0.1$ yield stress) for steel and 0.8 R for concreteor 0.85 R provided there is no adverse effect on the corrodibility.

Steel liner. Bonding condition. Mild or semi-hard steel,elongation 25% on L -- 5.65/ T* Thickness 25 mm for the nominalthickness.

System of thermal protection: avoid any degradation andcorrosion of materialjArticles 1? and 18 Final test

Gas pressure of 1.1 P, overall deformation limited on the basisof calculation and model results;Articles 19 and 20 Monitoring and safety devices

P, T,£ , displacements, tension of reinforcements. Valves.Filtration.Articles 21 to 28 Use, maintenance and operational monitoring

Maintaining in operational condition, checks on any changesoccurring.

Operating instructions, operating log.

The supervising engineer is the District (geographical) ChiefMining Engineer. He ^an draw on the services of various experts(in particular for cc-ordinatin^ his control measures with thegeneral regulations for nuclear facilities).

Administrative provisions.

VII. THE DESIRABLE DSVSLOPMCl.'T OF REGULATION: A PERSONAL OPINIOM

7«1.0. It has "been pointed out that the regulations are relativelyflexible; many provisions, some of them important, may be agreedbetween the manufacturer and the supervising authority. Thisis a fairly general characteristic of French official regulations,which specify essentially the results to be obtained and not themet^s to be used. In the case of those pressure vessels for

80

which a control test at 1.5 P is no longer applied, it has "beennecessary to go into a certain amount of detail, but the regula-tions still do not go into as much detail as certain foreignspecifications.If very much more precise specifications are not to be issued,we may at least expect the compilation of a detailed memorandumcovering, in particular, the general field of French technicalregulations and specifications as they apply to each part of thestructure. This memorandum could be issued when the regulationsare extended to cover water reactors.

7.240. The limit state calculation should be defined; it should describeat various loading levels the permissible behaviour of the variou,materials (assessment to destruction).

7«3«0. The concept of the safety factor against ruptureIt is possible that reliability studies will make it possible torelate this factor to the probability of catastrophic rupture.However, such rupture without warning is only conceivable in caseof generalized corrosion of the tendons (excluding the case ofexplosion of the nuclear core).. The risk of rupture depends onlyin part on the number and tension of the tendons. Rather thanspecifying a safety factor for excess pressure a -direct definitionshould be given of the conditions necessary to prevent corrosionboth in the case of grouted and ungrouted tendons.

7.4.O. Level of the excess pressure testThe level of 1.1 P is very low, since the aim is not to affectadversely the quality of the structure. This risk should bedefined exactly and balanced against the probative value of thetest.

7.5»^- Periodic testsThe principle underlying periodic tests is that of concentratingthe existing risks of rupture in situations of lower potentialdanger. It would be desirable to examine how far this principlecan be applied to our pressure vessels and to lay down a retestprogramme in the light of the findings.

81

7.6.0. ModelsLarge-scale models of graphite-gas reactors have produced veryinteresting results but these relate to the technological ratherthan the strictly safety aspect. Methods of demonstrating theaccuracy of predictions based on "assessment to destruction"will probably have to be laid down. If the production of a modelis to remain obligatory in future, it seems preferable to laydown the absolute dimensions of the model rather than its scalein illation to the pressure vessel.

7.7»0. Stress limitationsAn improved knowledge of the conditions of deterioration andcreep in concrete will enable us to analyse better the role ofpassive reinforcement steel and to specify permissible stresses,taking into account the possible passive reinforcement.

Analysis of the corrodibility of steels as a function of theoverall environmental and load conditions will probably enablethe stress limitations to be more precisely defined.

VIII. IGViilLOKO-j'f PKOoP^CTS

The extent of the French effort in this field shows that in spiteof the present slack period in regard to construction projects, the variousF:rench organisations concerned are convinced of the value of prestresscdconcrete pressure vessels, especially for light-water reactors and high-ternperature gas reactors.

Basic research projects, e.g. on the behaviour of concrete undertriaxial stress, are thought to be oi' general interest even outside thenuclear field and it has appeared reasonable to allocate considerablefunds to them i-jithout evaluating the expected advantages in quantitativeterras. In the case of technical development, on the other hand, largecredits are only allocated af'ùer an economic study on the desirabilityof each solution has been made.

An example of this is the comparative study presented to ^(JKATCli [_on water reactors, vihich concluded in favour of concrete for reactorsabove 600 _?i(e). This s udy provided one of the reasons for our generalemphasis on the techniques involved, each of which nevertheless requiresseparate economic evaluation.

82

REFERENCES

DAMBRINE et alo "Les caissons en "béton précontraint dans leprogramme français de réacteurs de puissance", 19&3» (Prestressedconcrète pressure vessels ir/. the French power reactor programme).Bulletin d'informations scientifiques et techniques, No. 75) (CEA).

/~2_/ BUPAYf F., "Les résultats pratiques obtenus sur nos caissons en"béton précontraint au "bout de 11 années d'exploitation". (Practicalresults obtained with our prestressed concrete vessels during 11 years»operation)* EURATOM meeting, Brussels, 18-20 November 1969»

/"\_7 On request from DEP/GTSP, CEA, Saclay.Commission des Communautés Européennes* Journée d'information du3 octobre 1969» Communications SSEE»

83

TSS'jf ST/.îIÏ.' :*OF. TRE::SSTSSSSBJÎ Of>;,{J5'ri'i-£ JEiïSL CCî?7ArmTC TBS ïil-LIUSî J.OOF

Î,.E, Kor.cli

The plur'-.j.a lavg» T :•-!;> ITIiV'-MoOd" "i;-... . i:e r o '.1 o^v \r;g a

feaii-res; rot ' j : n e r '>"•'•' C, , $*.sv:>.t--i - ."r. >;..-• vi« in ibc. concrèteyesso" -J?ai'i i UP tv ".'i''^'",} a/'d - /,'?."•.•.' s ' r.g1- -cc-o L 'ir.g aye Lsm Jr. vhs w*illto raciintaiii t sLipsraTVJ-e ~:t i ' ï j ; - - - v ^ . , i lovs ' ! ! . Vr--'1 i^lr-vs f^. i: .S: D work

underway: '^6'" "ï-e .'Tjar.-'iT,;,!..'?* ;_,£•:•;" ' ;,'tinp c:t'",Gi*t;"t!-., " ïr'soturs.". oonoreteand. ors stress !,<;£: ,?to ; ; - j s, ft 1^-rais-"' Uî-^p&^a^i'-.r&s, h">«;b. Ternie rature

instruîssntafc'j ou, ï:he vesoiï.'. ;.£• TO be t-rsci'oci 9. i i.a? a i t - , ->f tbeHe&otor Cevrssr l7":-. : .?er . ; \ -Trf , A;i;;1.ris., ta*" .& l 'V?O f vj- r-.-? tested v/i fchhot vater 19? I •''•-.'ate- re :>•-••. c^r ca«ai t.i o,i«) nrjci ^ t i - r 4 Hélium-tests

19*7'-„ Design and ere,C':. -.v;) i Kr;-ik.';oro.?,u .'"orHchur-gs- u-id Pa^g'ê

rsbH & Go. in coilau:; r-'-i'..; ' n ?/! U. Cs ter reich i se.'ie ^^j' .iorîg-esel" soha/t ftir

1. Prefatory NotePor more than a decade the enterprises associated inthe REAKTORBAU FORSCHUNGS- UND BAÏÏGESELLSCHAFT m.b.H.& Co. (RPB), viz. the construction firms

Ed, Ast & Co* IngénieursIng, Mayreder, Kraus & Co.Allgemeine Baugesellschaft -A. Porr AktiengesellschaftH. Relia & Co. Baugesellschaft"Universale" Hoch- und TiefbauAktiengesellschaft,

have concerned themselves with those questions andproblems which confront the construction industry inconnection with any work in the field of nuclear en-gineering. More particularly these problems involvethe construction of research- and teaching reactors,the planning of nuclear power stations as well as thetechnology of shielding concretes and special-purposeconcretes for the thermal shielding of reactor pres-sure vessels of pre-stressed concrete. In pursuanceof these endeavours, the construction firms -mentionedabove decided in 1969 to build a test stand for ahelium high-temperature loop as a joint venture with"Osterreichische Studiengesellschaft ftir Atomenergie"(the Austrian Society for the Study of Nuclear Energy)and a number of well-known Austrian industrial enter-prises. This loop will be installed in a large pres-sure vessel of pre-stressed concrete which is to beerected on the terrain of the Seibersdorf ReactorCentre, Within the scope of this joint project, RPBare to be responsible for the planning, design, con-struction and the operational testing of the pre-stressed concrete vessel.

2. The Basic Concept of the.._Pre-s_tressed. Concrete VesselThe guiding principle in evolving the overall conceptof this vessel was the desire to make a contribution

86

towards improving the safety of reactor pressurevessels made of pre-stressed concrete. Accordinglythe design concepts of the pre-stressed concretevessels "built so far were subjected to a detailedand exhaustive analysis, which showed that the dif-ficulties, and also risks, are to be found, aboveall, in two spheres, viz.a) in the necessity to protect the pre-stressed con-

crete and the steel liner from the high temperatureof the reactor coolant by means of some thermalinsulation on the inside of the liner, and

b) in the deformations which pre~stressed concreteis liable to undergo in the course of time as theresult of shrinkage, creep, moisture movement (de-siccation) and the consequent changes of strain.

In their essential aspects the effects produced bythese two facts are well known and will here be onlybriefly summarized.

The effect of the thermal insulation arranged withinthe pressure cell is that the steel liner, which isalso called the "cold liner" , remains at a low tem-perature level. On the other hand, however, the lineris penetrated by the reactor coolant so that, at de-fective spots of the thermal insulation, the temper-ature of the liner may rise to that of the reactorcoolant, - an operating contingency for which it isimpossible to make allowance in designing the linerand which, owing to overstraining, would of necessityresult in damage to the liner. What makes this situ-ation even worse is the fact that there is no way ofdiscovering incipient defects, as the liner is ef-fectively concealed by the thermal insulation and isthus not accessible for purposes of superficial .in-spection.

87

The long-time variability of pre-stressed concreteexerts a noticeable effect on the state of stress,the changes of shape and, thus, on the safety ofthe vessel. Since any changes of the properties ofthe concrete depend to a very substantial extent onthe degree of desiccation of the concrete of thevessel, viiich, however, can only be determined inadvance with great difficulty, whereas its variabi-lity in the course of time can hardly be predictedat all at the present time, it is obviously extremelydifficult to appraise this effect by means of com-putation.

In order to overcome these difficulties, two possiblesolutions have been investigated:

1 , The thermal insulation is arranged externally tothe steel liner with the result that the latterassumes the temperature of the reactor coolant.This is the well-known case of the hot liner»

2. Even before the reactor is put into operation, thepre-stressed concrete of the vessel is extensivelystabilized, so that the changes occurring in thecourse of service remain very b'

The investigations conducted so far have shown thatthe hot-liner concept can, indeed, be implemented.The main difficulty in this connection io that, inthe course of each temperature cycle of the reactor,the liner, which is unable to expand in relation tothe pre-stressed concrete that remains cold, is sub-jected to very substantial constrained stresses andtends to local buckling. With a larger number of tem-perature cycles there thus arises the danger of fatigueor repeated-stress failure, The main problem in con-nection with a hot liner is, therefore, to find outthe permissible number of temperature cycles in thecore during the scheduled life Lime of the reactor.

88

The permissible number of load cycles depends, in thiscase, on the difference in temperature between theliner and the pre-stressed concrete as well as on thevariation in their thermal expansion» The smaller thedifference in temperature is, and the less the linertries to expand, in relation to the pre-stressed con-crete , the smaller are the constrained stresses andthe danger of buckling, and the larger will, therefore,be the permissible number of load (temperature) cycles.

There are thus two ways of attaining a sufficientlyhigh number of load cycles, viz, by

a) keeping the difference in temperature between theliner and the concrete at a low value. This can onlybe done by raising the temperature of the concrete,since the temperature of the hot liner is given bythe operation of the reactor;

b) by developing concrete with a large, and steel forthe liner with a small, coefficient of thermal ex-pansion.

The more important influence in this respect is hereexerted by an increase in the temperature of the con-crete. For this reason a concrete temperature of morethan 100°C was chosen, i.e. a temperature much higherthan any used hitherto in structures of this kind.The temperature of the concrete is to be controlledby means of an imbedded pipe system and should permititself to be adapted within certain limits to the tem-perature cycles of the liner.

In addition, however, the increase of the wall temper-ature also offers the possibility of accelerating thedrying out of the concrete. Creep, too, is speeded upby the high temperature, so that a large part of thetension losses caused by creep can be compensated forby re-stressing even before the vessel is put intoservice»

89

3. Design of .the Mode?^ Vessel

Service pressure: Pg = 100 atmTest pressure; Pq = 1 1 5 atmService température: Tmax = 300°G (at the liner)

3.. Dimensions^The vessel» which is of cylindrical shape, isvertically positioned. It is equipped with abottom plate made of concrete as well as witha steel cover,Internal diameter: d^ = 1.50 mExternal diameter: da = 3.60 mHeight, inside: h^ = 8.00 mHeight, outside: ha =11,40 m

_3.3. P.e eJi.ra.T.LonsjThe model vessel features an axial penetration inthe bottom plate intended to accommodate the blowerwith a diameter of 0.65 m, two radial penetrationswith a diameter of 0,30 m and an upper opening overthe entire cross-section of the vessel with a dia-meter of 1,50 m. All possible kinds of penetrationsare thus represented in the experiment. As regardstheir design as well as in terms of figures, thesepenetrations can be allowed for in laying out thevessel.

The vessel consists - in radial arrangement - ofthe following layers, seen in the order of theirsuccession from the pressure cell towards the out-side :Steel liner: 5 mm in thicknessInsulating concrete: 20 cm in thicknessPre-stressed concrete: 85 cm in thickness.

Axially the vessel is laid out analogously.90

S,v sterns^So as to permit the temperature of the concrete tobe regulated, the vessel is equipped with, two cool-ing systems, viz,a primary cooling systems, which is arranged atthe boundary surface between the insulating- aridthe pre-stressed concrete,, anda secondary cooling system, which is laid outwithin the pre-etressed concrete.

It is worth mentioning in this connection that thesetwo systems can be operated both for purposes ofcooling and heating»

When the vessel is put into operation, the lineris warmed up to a temperature of 300°C« At thesame time the pre-stressed concrete is» with theaid of the two cooling/heating systems, warmed upat a constant temperature gradient of 20°C to suchan extent that its temperature amounts to 120°C atthe contact surface with the insulating concrete,and to 100°C at the underside*

In regular service and throughout all temperaturecycles during which the temperature of the linerdoes not drop below 120°C, the temperature of thepre-stressed concrete remains at a constant level.It is only if the temperature of the liner fallsbelow 120°C, that the temperature of the pre-stressed concrete is also lowered by means of thetwo cooling systems to such an extent that no ten-sile stresses or strains arise in the liner.

2. truct uralm Analy si £ :The structural analysis of pre-stressed-concretepressure vessels is a many-sided task that is de-termined by numerous parameters. It is character-ized in a substantial manner by the fact that the

91

basic conditions customary in all other cases withregard to the relations of the dimensions to oneanother as well as to the uniformity of the materi-als of construction, and the temperature level dohere no longer apply» The methods of calculation,adopted for this purpose are, therefore, based onother premises such as, to gi^e an example, on themethod of finite elements or on the dynamic relax-ation process. Detailed reports on this subject aretime and again su&ruixted at meetings of experts inthis field. In the following the methods adopted byRPB in this respect are to be explained in "brief.

Here too, as is generally customary, simplifiedmethods are employed in performing the preliminarycalculations. However, the concept of the pressurevessel described bere differs from similar investi-gations in that the load-bearing part is subjectedto much higher temperatures and in that the thermalinsulation, which is arranged behind the liner, it-self forms part of the load-bearing structure. Fromthis layout alone it already becomes evident thatthe properties of the materials are, within thethickness of the wall, subject to diverse and im-portant changes in the course of time» i?)ven for thepurpose of preliminary calculations, these changescan by no means be disregarded any longer, in par-ticular also since the stressing of the innermostelement, vis. the hot liner. Is decisively deter-mined by the deformabil^ty of the pre-stressed con-crete part, ïhis effect is so far™reaching that thestresses likely to occur in the liner must be takeninto account even at the construction stage proper,which is decisively influenced by this necessity.

The properties of the materials in accordance bothwith the test results on hand and a study of theliterature have thus been depicted depending on thetime and temperature as a baois for the computation,This can be accomplished by means of simple functions

92

which can be readily programmed with a satisfac-tory degree of approximation. As a result one thusknows at any time and at any desired spot in ac-cordance with the temperature distribution thecharacteristic values for the deformation, viz,the modulus of elasticity, creep characteristics.,Poisson'a ratio etc», so that the development ofthe play of forces and of the changes of shape canbe taken in at a glance» A more comprehensive com-putation programme, which is set up with the aidof finite element3, will then furnish the requiredinsights and opportunities for comparison with thetest results,

The distribution cf the temperature can be deter-mined, and thus extensively influenced, by theservice temperature» ambient temperature, insula-tions, the coefficients of thermal conductivityand, in particular, by the cooling systems.

For the purpose of preliminary computations, whichenabled the first fundamental decisions to be madewith regard to the dimensions, the cylindrical shellwas split up into finite, cylinder-shaped elements,for which the equilibrium and continuity or certain,deformations were specified. To these elements cor-respond the diversified or variable material pro-perties and temperature conditions. This calculationprocess then provided not only the basis for dimen-sioning the pre-stressing forces in axial and radialdirection, but also the basic values for those con-siderations which concern the marginal disturbancesof the stress behaviour at the thick-walled cylinderwhen the latter is connected to the bottom plate oralso at the top cover. Last not least this also en-abled the fatigue analysis of the liner, which couldthus be carried out in accordance with the pertinentASME-speeifications. The loading conditions to whichconsideration was given in this connection comprisenot only the operating processes within the helium

93

loop, but have been chosen in such a way that theyalso satisfy the conditions occurring in the reac-tor of a nuclear power station* It should be pointedout in this connection that the findings obtainedapply not only to gas-cooled but also to water-cooled reactors»

_3.8Provisions have been made to arrange 50 measuringpoints in the insulating concrete and 200 straingauging points in the pre-streased concrete» Overand above this the deformations of the liner andof the steel used for tensioning are measured» Atthe same time the temperature is also measured atall expansion points»

4. Prelijninary InvestigationsThe novel kind of problems which this vessel designconcept entails must be seen in what - for highlystressed structural elements - is the unusually hightemperature range. This necessitates a series of pre-liminary investigations and development tasks, especi-ally for the liner, in the field of technology and asregards the insulating- and pre-stressed concrete. Partof this preliminary work was begun as early as two yearago and is currently neariag the stage of completionrequired for its actual implementation.

.1 • .1 Iiijle£ » _Sjt e e^l^Co^ ve^r^and^P^n^et r ajb i£ns_

The technological investigation of the liner andits anchoring in the insulating concrete, of thesteel cover at the top of the vessel, the steelcomponents of the penetrations as well as of thedesign and manufacture of the elements of thisvessel, are carried out by VÔEST ("Vereinigte Oster»reichische Eisen- und Stahlwerke Aktiengesellschaft ;

94

4. 4

The planning, design and manufacture of the cooling/heating systems is carried out by Simmering-Graz-Pauker Aktiengesellschaft.

Here it was necessary to develop a pressure-resistantand pressure-transmitting concrete which is reason-ably resistant at cyclically changing temperaturesof up to more than 400°C.

Little is known about the technological propertiesand behaviour of pre-stressed concrete in the en-visaged temperature range, i.e. up to 120°C. Forthis reason a series of tests is currently beingconducted with a view to ascertaining the materialproperties and characteristics in dependence on thetemperature. Over and above this investigations arebeing made as to the possibility of producing a kindof concrete which, by desiccation and re-stressing»is far-reachingly stabilized,

Measuring the deformation in the insulating- andpre-stressed concrete as well as the measurementof humidity in the latter is exceedingly difficultwithin the envisaged temperature range. It is, infact, a task which demands the development of spe-cial methods of measurement suitable for this pur-pose as well as the careful selection and tryingout of the commercially available instruments whichare to be used» Investigations in this respect arebeing made in cooperation with the Technical Uni-versity of Vienna,

95

5,___ RecapitulationIn the light of the general, and essential, develop-ment in the field of nuclear energy, a decision has"been taken by RFB to cooperate with organizationsinterested in the progress of nuclear reactor en-gineering (SG-AE) and firms belonging to other branches(VO'EST and SGP) in building a large-scale test standat the Seibersdorf Reactor Centre and participate,within the framework of the aimed-at larger experi-mental installation of a helium loop,, by assuming theresponsibility for the design and construction of thepre-stressed concrete vessel required for that purpose,

The preliminary investigations and the developmentwork involved could be completed by the end of 1969to such an extent as to permit the construction, in1970, of the experimental vessel of pre-stressed con-crete in accordance with a modern concept.

One may feel confident that the implementation of thisproject will represent a useful contribution towardsimproved reactor safety in connection with the use ofpressure vessels made of pre-stressed concrete.

Vienna, 29 December 1969

96

1- . '4&. _ > âe.^$ti I II

sv ' , - ; . - - y*' -S. '. .'<-\\*.; /I Ûi5s!ïte|?psq •:• ¥ml!i!ti ;•»v -,:?'"•- • ';,;—-*So° '•• •'-;>

I1fv:.r-.'-v-;: ^- ! J --vrt';.-••

Btrom: Small Diraeras 1 on

F, Scotto

The paper is related to the experience of the Author in thefield of prestressed concrete pressura vessel modeJs, built and tested"or BKSL State Electricity Oesier-ating Board b,y the "Istittrto">p era mem ale Modelli a Strutture - ÏSMES" of Bergamo, Ita^y, within'he framework of the EURATOM. THÏR Contracta and ENEL research work.

The paper outlines the "behaviour of the structure as thepress-are increases, starting from working condition up to structuralcol]apse k

The "nearly elastic", tbs "definitely anelastic" and thecollapse limits of the structure are discussed with regard to theconcretet the liner and the cable steel behaviour.

99

1. Linerijijandr_it8_ influertce on the struoture safety limingThe liner undoubtedly représenta a vital safety element for

the vessel. Indeed, it serves the fundamental purpose of forming animpervious barrier to the cooling gas. The reactor operability dependson its integrity»

The vessel design calculations are basad on the assumptionthat the gas pressure is always exerted on the internal surface ofthe liner. As far as we know, the possibility of a leakage of high-temperature gas on to the concrete walls following liner rupture isnever taken into account in the calculation assumptions,

A confirmation of this statement, lies in the conventionalmethod to define the safety margin of the structure to rupture assumedin the design calculations. The primary cause for collapse is assumedto be the failure of the prestressing cables.

For the sake of simplicity, cylindrical vessels - which arethe most commonly used - are assumed to come apart into cylindricalwalls and slabs in the ultimate conditions. The gas pressure isentrusted separately either to the circumferential or to the verticalcable system only, and the pressure which determines cable failure isdefined for the two cases. The requirement ie that the 'ratio betweenthe pressure determining cable failure and the operating pressure isnever to be lower than about 2.5.

The failure pressure is assumed to act on the internal surfaceof the structure and thus liner integrity is taken for granted untilthe vessel collapses»

Actually» if we evaluate cable elongation to rupture, v/e findthat, in the grouted cable vessels built so far, the elongation is onthe order of one meter. This is also true for vessels with cables ofhigh unit capacity (whose adherence to the grouting mortar cannot betaken into account), that is, for most of the vessels recently designed.

100

It is easy to infer, and the experience with model testsconfirms it, that the liner currently designed will not retain itsintegrity and gastightness up to the cable failure pressure. There-fore, the safety limits as above defined are merely conventional.The vessel will collapse not at the ultimate collapse pressure of thestructure coincident with the failure of the prestressing cables, butat the liner failure pressure which is lower.

It is essential to consider the behaviour of the liner inthe ultimate conditions. This investigation, however, would entailthe knowledge of the topology and magnitude of the cracking on theinternal surface of the concrete in contact with the liner. In other7/ords, it would be necessary to have a clear idea of the cracking modeof the structure for increasing pressures up to the collapse pressure.This behaviour cannot be defined on a theoretical "basis, and thereforeit is necessary to resort to experimentation.

We believe we have obtained valid information on this pointfrom the experiments performed on small models of prestressed concretecylindrical vessels (scale 1:20) for high-temperature gas reactors.It may be interesting to note that these models were provided with asupplementary annealed-copper liner which allowed them to be tested upto pressures corresponding to the cable failure pressure vathout thepressurizing- fluid (water) penetrating the concrete»

2. Ultimate tests on sraell soal_e modelsLet us define as ultimate tests all those that are intended

to clarify the behaviour of the structure as the pressure increases,starting from working conditions up to collapse. The main assumptionsrequired for the performance of these tests concern:

a) the time, referred to the pressure vessel life, atwhich the incident occurs?

b) the rate of increase of the test pressure.

101

With regard to point (a), the stressing condition, thestrain distribution and the related deformation of the structurechange with time because of concrjte shrinkage and creep and steelrelaxation. We must therefore know the history of loading andheating of the actual structure, starting from the concrete pouring(taken as zero time), namely, prestressing, test pressurization,depressurization during the preoperational tests, heating andpressurization at start-up, depressuriaations, if any, duringoperation, etc.

As the pressure increased, the structure passed from a"nearly elastic" to a "definitely anelastic" behaviour. Theseparation line between these two behaviours, is a matter ofdefinition. The upper limit of the "nearly elastic" behaviourcan be defined as follows!

a) For the concrete?(i) The highest pressure P at which the de-

""" Cformation measured on the model remainsproportional to the pressure in everypoint of the structure.

(ii) The pressure Pf at which the first crackscan be detected visually on the externalsurface and by means of crack detectorson the internal surface

(iii) The pressure P at which the structurestill presents a satisfactory reversibilityof its deformation. This pressure can b'eeven higher than the pressure Pf definedabove.

b) For the cables $The pressure Ps at which the yield point of thesteel cable wires is first reached.

102

The models also reveal that the range of"definitely anelastie" "behaviour can "be furthersubdivided into two regions»a) Regi on of small deformati on where extensive

areas of the concrete work nsainly in the anelas-tic range and also present some cracking, butwhere the steel wire have not yet reached theiryield point.

b) Region of great deformation in which the struc-ture becomes disconnected and some of the steelwires in one prestressing system or other haveexceeded the yield point. The upper and lowerlimits of this region are respectively Ps and PR(collapse pressure).Under these circumstances, the selection of the reference

time of the incident has a determining effect on the elastic limitsof the structure. Conversely, it is reasonable to assume that itwill not affect the collapse limit (PR), as the latter depends onlyon the ultimate strength of the cable systems.

With regard to the test pressure increas rate, weshould bear in mind that in the stress-strain diagram for concrete,the point corresponding to the ultimate strength under prolongedloading lies "below the point corresponding to the quick applicationof the load. This behaviour is known for monoaxial compressions,but I believe it is not known for triaxial conditions involvingtension.

For the THTR model, built for THTR German Association,we decided to stimulate an overpressure incident occurring at aboutone year from startup. This seemed to be a reasonable compromisebetween the advisability of considering both the average and the

103

most unfavourable conditions of the structure, particularly asconcerns thermal stressing. To do this, the temperature differencesacross the walls was assunssd. to be one half of the reference value.The pressure was to "be increased at a rate of nearly 142 Ib/sq. in.(10 a tic) every three minutes.

The foregoing may give an idea of the difficultiesencountered even for the mere definition of ths structure safetymargins. For instance, with regard to the safety margin K0 toticollapse of the model, defined aa the ratio between the modelcollapse pressure PR and the working pressure Pw, the modelsdemonstrated that the simplifying assumptions described in PartOne are inadequate.

The TETE model was designed for a working pressurePW a 569 lb/sq.in- (40 atm) and collapsed at 2730 lb/sq.in. (192 atm)because of failure of the vertical cable system. The pressurescorresponding to the simplifying assumptions were 2546 lb/sq.in.(179 atm) for the vertical cables and 2185 lb/sq.in. (153.6 atm)for the hooping cables» According to these assumptions, the structureshould have collapsed earlier, and the hooping cables should havefailed before the vertical cables»

The explanation of this apparent inconsistency isthat as the pressure rises the structure warps and releases itshyporstatic restraints so that the failure of one cable systemrather than the other depends exclusively on the mode of rupturernd disconnection5 in other words, the cable system first reachingits ultimate elongation because of the deformation will be the firsttrs fail. In addition,, déformation is opposed to a greater or lesserdegree by the steel reinforcements, cable ducts and especially bytha concrete in the unfailed areas.

An important result of the THTR model rupture testwas that although the slabs got, disconnected as shown in Fig. 1,they did not suffer any cracking. The slabs were rebuilt and

104

tested separately in a simply supported condition. Lateral pre-compression along the perimeter of the slabs was 1507 Ib/sq.in.(106 atm); rupture occurred at 3910 lb/sq.in. (2?5 atra).

Pig. 2 shows a section of the slab along a diametralplane after the rupture test. The grad.ual transformation of theslab into an inverse dome through successive cracking is evident.Prom these results it would appear that the rupture is imputablemore to a dome-collapse effect than to a punching effect. Itmay therefore be advisable to carefully reconsider the importanceattached to the punching effect in the calculations for the ulti-mate strength of the structure and for the sizing of the slabs.

ConclusionsIt should be emphasized that this paper refers to

experience acquired from two small-scale (l:20) cylindrical vesselmodels? and therefore it would be premature to draw conclusions.However, it may be stated that as concerns the elastic field, thecomparison between the experimental results and the computer calcu-lations shows a fairly satisfactory agreement. This consistencyconfirms that the model accurately schematized the actual structureand that the calculation methods were quite applicable. Therefore,there seem to be no reasons for doubting that the model will bejust as representative of the structure also in the anelastic field,

Of course» a final judgement on this point will bepossible when a larger number of test results are available on thebehaviour of numerous models of the same and/or different scale.Hence, it becomes a matter of mere economics.

At any rate, it is ray opinion that these small modelsare to be considered a real calculation tool, as well as a valid

105

means of experimental checking. Their cost is moderate - on theorder of Lstg 30}000 - and the aine months required for their con-struction and testing represent an additional advantage, as modeltesting can "be carried out during the design stage of the actualvessel.

106

*• ' '

108

ANNEX I

'.«"he main activities of Italian Organizations in the field of prestreáscd'

concrete pressure vessels can be deduced from the following "Contracts"

implemented on bobalf of International Organizations and "Papers" presented

to specialized Conferences and Symposiums :

1. CCWTRACTS,

1. I Client :

O. li. C. D. - High Temperature Reactor Project Dragon

„ 1.1.1 Subject of the Contract :

1250 WW (th) H. T. G. R (High Temperature Gas Reactor)preliminary design

1. 1. 2 Performed by :

Joint Group: AGIP Nucleare (1) Italy - P. C. P. V. DesignINDA TOM (Prance) - Heat Exchangers Design

1.1. 3 Fina] Report, December 1 963

1. 2 Client :

EURATOM

1. 2.1 Subject of the Contract ;

H. T. G. R. System Assessment Study

1. 2. 2 Performed by :

Joint Group :ENEL (Italy) - INDA TOM (Prance)

1. 2. 3 Final Report, December 1965

(1) - The AGIP Nucluare experts were grouped in the Italian State ElectricityBoard - ENEL on January 1st, 1905,

»na

T. H. T. R. Association - Bro-vn fio very /KruppKeactorbou GMWi (West Gci i.j.-.ny)

3. 1 Subject o; the .Contr»_c i. :

330 MW(cl) Pebble bed Thorium High TemperatureGas ReactorUp to built prestresKOd concrete pressure vessel

3- 2 Perforated by :

Joint Group : ENEL (Italy) 40% - INDATOM (Prance) 40%Fried Ki-upj) Univorsalbau (W. Germany) 20%

3- 3 Final Report, December 1967

i Client :

EURATOM

*• J Subject of the Contract :

Programma di studio sulla tecnología del cnlcestruzzosouoposto a sollecita'/.ioni teriíiiche e a radiazioni

^ ^ Perforniod by :

SNAM - S. p. A. Attivitá AG1P Nucleare

Í. 3 Rapporto n. 7 - Con/'erenza Bruxelles, 18-20 Novero-bre 19G9'

110

2. PAPERS

2. 1 The Institution of Civil .Engineers - Conference onPrcstressed Concrete Pressure: Vos-colsLondon, 13-17 _Mar«:h 196'»

Paper 27 - Dr. F. Scolto -• 35NEL, Italy :"An Improved System of /looping Obles1' (ENJRL pateni)

Papor 63 - Dr. E. Torielli, Dr. P. Rocca - ENEL (Italy) :"Thermal Problems in Pressure Vessel Design"

2. 2 EURATOM - Convegno di informazione sui problem! dicontenitori pressurizzatj in calcestruzzo precompressoe del loro isolamento termscoBruxelles, Concert Noble,?-8 Novembre 1967

Ing. E. Torielli - Ing. F. Scotto - ENEL (Italy):"Prestressed Concrete Pressure Vessel for NucleaiReactors - Experiments on Small Models"

2. 3 The European Nuclear Energy Agency - Symposium enthe Technology of Integrated Primary Circuits for PowerReactorsParis, 20-22 May 1968

Paper EN-1/21 - Bologna - Morelli - Pagani - Scotto :"Enel's Experience in the Design of Prestressed ConcretePressure Vessel Components".

2. 4 The British Nuclear Energy Society - Conference on ModelTechniques for Prostressed Concrete Pressure VesselsLondon, 10-11 July 1969

Paper 4 - P. Scotto - ENEL (Italy) :"Techniques for Rupture Testing of Prestressed ConcretePressure Vessel Models"

111

Problems relevant to the thermal insulation of Gas Heactors were

studied by the Commission of European Community in J&pra, Italy.

The experience of the Panel Member covers all the points listed

in the Agenda of the Panel. Said experience arises from the work devel-

oped as Civil Engineer responsible of the Civil Items of the Contracts

1.1, 1. 2, 1. 3 above-mentioned.

The more interesting contributions of Dr. Ing. P. Scotto to the

studies on the prestressed concrete pressure vessels are the following :

a) - As for the Contract 1.1, the design adoption of 660 t cables

and the promotion of the activities for their achievement at the B. B. R.

Society of Zurich. Owing to the intelligent and far-sighted collaboration

of the Managers of said Society, these cables are now operating with a

great influence on the simplicity and economy of the prestressed concrete

pressure vessel constructions,

b) - As for the Contract 1. 2, the study of facilities in order to

realize openings in the upper cap slabs to vertically withdraw the heat

exchangers. The operation consists in drilling the concrete in pre-arranged

areas and in sealing it by means of special pre-casted plugs properly

grouted and fixed.

112

c) - Aa lor the Coruroct 1. 3 :

1 - The development of a per.icuiar pr.v.*>rn oí hoopi.-_ cablcfe, vr:c!

not only numruzeri ihe ova'i/i'.g <.-f.octs of the cucboring ÍH*»d

reactions against Ihe outside «rurCaoc o" cyliuor cal v/elis, but

also utilizes I!, use reactions lo provide» additiorv.l prestrors'ng

forces. This concept has been applied by the T. 1U T. R. Asso-

ciation for its up to date design of prcstressec1 concreie pressure

vessel.

2 - The proposal of tests on small scale awl low cost models fiu- ihe

analysis of the behavior of the p. c. p. v. structure with special

reference to the anelastic fields and i o the collapse of the struc-

tural components (concrete and cableo).

These studies, besides demonsirating the interest and the availability

of this investigation method, have put in evic^-nce the acíu: i morgiu.» of -the

structure and the unexpected valuc.s of material sirength.

The researches under development aim to more exaciJy know Ilia

actuai safety limits of the vessels and the development based on hypothesis

more effective.

As regards the Contract 1. 4 performed by S\AIV1 S. p. A. AGIP Su-

eleare, this is framed into a program of studies and researches relevant

to the technology of the concrete under combined stresses subjected Ic

temperature and radiations. The concrete specimens were irradiated in

113

the swimming pool of the CAMEN - Galileo Galilei Reactor (San PÍero

a Grado ~ Pú-a, Italy).

The achievement of this research was ihe selection oí a special

concrete called B. H. T. (Betón Ifoute Tempéraiure) that seems segable

to be utilized as structural materiaJ up to temperatures of the order of

500°C. The B. H. T. concrete, together with another t3'pe of concrete

called Standard, that is suitable to withstand temperatures of the order

jof 300°C, were subjected at a thermal integrated neutronic flux of abaut

1020 n. cm"2.

The maximum temperature reached by the irradiated concreto

samples was 280°C.

At present, the field explored by the above-mentioned studies

is lacking of information and therefore these studies give the possibility

to take into consideration more advanced solutions of prestressed con-

crete pressure vessels.

114

REPORT ON THE STARTING OF A COORDINATED WORK PROGRAM FOR PRESTRBSSEDCONCRETE REACTOR PRESSURE VESSELS IK THE FEDERAL REPUBLIC 07 GERMANY

Thomas A» Jaeger

An optimum design of a. PCPV for gas-cooled or water-cooledpower reactor systems can be achieved only if the reactor vesselis not considered as just a construction project, but if it isinterpreted as a functional unit incorporated in the specificpower reactor system with interrelated functional, safety andeconomic connections with the system. The paper describee theconcept of a basic programme for research development of pre-stressed concrete reactor pressure vessels in the Federal Republicof Germany, as well as the basic structural design philosophy,criteria and safety of PCPV. Several special engineering develop-ments for PCPV (e.g. multilayer, prefabricated vessel structures,and prestressing systems) are described. Detailed descriptionsof "Joint Development Programmes of Prestressed Concrete PressureVessels for BWR" (AEG-Teiefuhken and Fried.Krupp GnbH), "PrestressingSystem Developments using Different Methods", "Computational Methods"and "Instrumentation Assessment Programme" are given in theappendixes.

115

Part 1. The Concept of a Baa i o Program for Researchand Developmentfor Frestressed Concrete fteaotor Pressure Vessels in JheP.E.G., - a Discussion with Respect to Basic StructuralDesign Philosoghy Criteria and Safety of PCRPVs

Since several years, there is a marked trend to large unit sizesof nuclear power plants. The driving motivation for ever larger unitsizes comes from the fact that with increased capacities capital costsand operating costs per MW of electric power installed go down morerapidly with nuclear plants than with conventional plants.

With gas-cooled and water-cooled reactors, the natural economictrend to very large unit sizes of power reactors is opposed by techno-logical difficulties inherent in monobloo and laminated steel pressurevessel construction. The advent and rapid development of prestressedconcrete reactor pressure vessels offers several great advantagesí

1. The possibility of attaining vessel dimensions in combinationwith internal pressures which are not feasible in a practicalway when using steel;

2. considerable freedom in the realization of any vessel shapeknown to be advantageous in a structural and functional sense;and

3* increased safety a) by the peculiar structural characteristicsof prestressed concrete designs and b) by the possibility ofarranging boilers and ducts within the pressure vessel itself(integrated design concept), with consequent advantages inflexibility of siting.

The prestressed concrete pressure vessel concept, therefore, offersa significant step forward in performance, safety and economy of gas-cooledand water-cooled power reactors, - for the gas-cooled power reactordevelopment line, there is no alternative to PCRPVs.

Recognizing the great importance which the development of thetechnology of prestressed concrete pressure vessels bears for the improve-ment of economics and safety of gas-cooled and water-cooled power

116

reactors, - raoreovet-, to make very large-capacity plants feasible at all -the Burdesminis-terium für Blidune uno 'flies en schaft (Federal Ministry forEducation and Science ) of the Federal Repibiic o-r Germany is sponsoringa we'.l-coordinate'i, rather comprehensive research aad development programfor prestressed concr-d te pressure vessels for gas-cooled and water-cooledpower reactors which is embedded in the foreseeable German reac'tor strategy.The program has bee»n set up by the Bandeeanstalt für Uatevialprüfung (l).Several industrié*! companies, university institutes and government institutesare participa tirj£ I* this H & 5 program which has been started at the endof last yea*-.

The principle oo v/bicn the program concept in br.sed, *3 a classi-fication of the research -. n^o tnrec- categories:

I. FundanenfcH! rII. Reaotov-t^pe '•water-cooled resp, gp.s~coo'iei) oriented

research,TIT. Inri ividua"! reactor-project orientated research,

*Alth as much shifting ol' weights tow-arris fundamental research as possible.

I. Trie focal po^n:-s o;' the fmdansnt.?J research are?(l) Ge'.e"*a3 sjste~s axiaDysis, sa-'ety and economy parametric

studies to asoess She full potencial of the prestressedconcrete pre'ssure vessel concept for tae modern develop-ment lines of gas-cooled and water-cooled power reactors,and studies to establish design criteria,

(?) Materials studies on concrete behaviour under triaxialstates of stress, about creep of concrete and moisturemovement at elevated tsmpera tures . etc..

(3) Development of a comprehensive and flexible computerprogram utilizing the "finite e'e^e.-it matnod'* for theanalysis of stress, strain, temperature and moisturedistributions In thres-di ruens tonal composite structuresof arbitrary shape, consiJering nonlinear, inelastictemperature- an-.i time -de penden b materials behaviour.

117

(4) Selection and installation of measurement-instrumentation,and suitable procedures of commissioning.

II* The reactor-type oriented research focusses on the development)design and analysis of liners, insulation, and cooling systemsfor gas-cooled reap, water-cooled reactors.

III. The research connected with individual reactor projects needsonly be concerned with particularly complicated structuraldetails, due to the above-mentioned successive shifting of alltopics of more general character to the more general and funda-mental categories II and I,

It is intended to explore the possibilities of achieving an inter-national cooperation in the categories I and II.

The concept will be discussed in the following with respect to someaspects of basic structural design philosophy, criteria and safety ofPGRPVs.

It is clear that an optimum design of a prestressed concrete reactorpressure vessel for gas-cooled or water-cooled power reactor systems canbe achieved only if the reactor vessel is not considered as just a con-struction project during design and thus is treated in a conventional manner,but when it is interpreted as a functional unit incorporated in the specificpower reactor system with interrelated functional (including technicalinstallation principles), safety and economic connections with the system.

thus, the creative cooperative thinking of the civil engineer forconstruction is needed in the development of the overall design of a powerreactor system making use of a prestressed concrete reactor pressure vessel.This means, that the civil engineer should be involved in an advisorycapacity at an early stage of development of the draft of a system design.Only close collaboration of the participating specialists of a reactordevelopment group can produce true technological and economical success.

118

For a truly creative coi labora* fon o" tbe civil engineer incharge of construct lor. in the development; cf the over-ill design ofa po'.ver reactor system vdtb optimum utilization of all potentialstructural possibilities offered hy a .-restrecsed concrete vesseldesign from the standpoint of a functionality, installation engineering,safety engineering and economy, a sufficient insight of the civilengineering inte the p~obJe¡,í3 of reactrr and shielding pvsioa, in-strumenta* to-:, heat- eivsirieeriAg aria rafety engineering of powerreactor piante, naturally is a necessary prerequisite.

<*n>y such insight can «p.vs r":-;e to ideas in civil engineeringwhich, e /en though ou ¿vvardl,v siuple, can lead to th-3 d«'scovery o¿' nevirpüssi'bÁli'jief? of r.ys>-;,S':i dasLgr, and of íncrüasii-ig t.ce capability ofpov/^r reactor aiit.-», ana thus, result iu stvaetiufts considerable over-all 1 cost reductions in AJÍ inciviJaal caso a>*<i even in ne*.v objectivesin strategic reactor planning.

The good safety diario IsritsfScfc cf reactor pressure vessels ofprestressed concrete construction permit one SD expect simplificationsof the primary system and reductions of piant wccts of gas~coo3edand water-cooled power reactor systems. Safety aspects, largedimensioning possibilities as «el! as a Cfcrtain ^'lerjb' lity ofconstruction schedules, wnlcjh exists an a function of the structural «design, raay a] so introduce PCRPV design Into fields of reactor powerunits with econoaica?. superiority where the possible limits of asingle-shell steel container have not y<t been reached.

In order to achieve the required flexibility c.f PCRPV design,which io considered indispensable, spccJaJ e^Ipb^^sis must be placedin the near future on basic researoi\ in the fields of materials, - I (2) -,and computing methods, - 1 (3) -. ^ie very close attention ofresearch to date to actual projects on an international scaJe limitsthe applicability ar>d genera1 useful noss of findings obtained at highcosts to a regrettable extent.

119

The same basic approach, in my opinion, will also give maximumsupport to the establishment of detailed design criteria by facilita-ting systematic parameter studies which — if sufficiently detailedand based on good knowledge of the complex materials behaviour —would make the overall terrain clearly visible.

An important part of the R 5s D program with regard to safety isproject I (4), aiming at a full utilization of all PCRPVs pressurevessels .to be constructed in the next few years in order to obtainscientific data for use in project design and calculations of futurePCRPVs (projects after expiration of the present program): in otherwords, existing vessels should be used as fully as possible as 1:1models for scientific purposes. Therefore, it has been recommendedin the program that pertinent requirements be set for the participatingconstruction companies and utilities which naturally demands fullfinancial support of the part of the instrumentation installation andcorresponding recording and scientific analysis going beyond necessarycommissioning tests and operating controls. Further' large-scale modeltests would become superfluous with such a program aimed to achieve anadjustment of computational methods to a high degree of realism and torapidly accumulate a body of data which would allow the development ofas well safe as economic design criteria.

Testing of a construction project and control of the operationalbehaviour and the safety of a PCBPV, and beyond that furnishing scientificdata, requires:

(1) Use of the most suitable measuring devices for the specialconditions;

(2) use of recording methods which are adapted to the specialconditions (number of test points in the order of 10 , testdata in the order of 10 ).

120

following measurements should be carried out:a) Displacement measurements on the inside and outer vessel

surfaces;b) strain measurements within the concrete structure, on the

rods of unprestressed reinforcements, on tension cables, onthe liner and it& anchors, on conduits and on the installedcooling pipe system;

c) crack distribution and width measurements in the concretejd) stress measurements in the concrete;e) force measurements at anchors of tension cables without

bonding«f) temperature measurementstg) moisture measurements.The measuring devices of the system are installed und<?r unfavourable

conditions and with the progress of construction are often considerablyexposed to the risk of mechanical damage. Nevertheless, a reliableindication over long time periods is necessary while some types ofinstruments will be operating under conditions which are unfavourablefor them (broad range of temperature fluctuations, variable moisturecontent).

Por this reason a systematic and comprehensive program for instru-mentation assessment has been set up which will be performed at variouslaboratories, but with the main body of the work to be done at theDivision 6.1 of the BAM which is also in charge of the close coordinationof the overall program. The instrumentation assessment program is outlinedin some detail in Appendix D.III,

121

Par i 2. The Work of a CoffliKifctoe for jS^jyio^r.ai.jiiipePigij_.TOrl.te'>..a fogPrestiré sg eel C c a r e f c e Re a o tor Preasure

In 196ft in the F.R.G. a committee for eFtatJissiJn# design recommen-dations for prestressed concrete reactor pressure vessels for nuclearpower reactors has boeri formeJ «-ithin the fram* of the " I?achnormen-ausschuss Kerntecnnik (PNKe) im Deutschen Kormer-va-cr :u¿s (DKA)".committee is in cormnunisafcion wxfch the ISO Tea'\«ioal Coinm1tte€ISO/TO 85 "Kuclear Snerf^", Si:bcrtnmittee 3 "Rpaoot S^^ety",Group "Prestressed Conc-eb? Reactor Pi*essure V

Our committee up to now has held several c«ci long, however, adraft design which cotTu be presente'! nere has not ye - been achieved.

The proposed division for the jQcotanen'ia.tlcn.s on design criteriafor prestressed concrete reactoi pressure vessels :<? as folJofirs:

1. General2. Terminology and Definitions3. Materials

3.1. Concrete3.2. Reinforoing ana Prestreasin,? Sfc*»el3.3. Liner Materials3.4. Thermal Cnsuiatioa Mateiiglc3.5. Cooling System Structur' Material s

&. . Des j gn4.1. Elastic Stress Analysis4.2. Inelastic Stress Analysis4.3. Limit Analysis4.4. System Relief Devices

5. Manufacture, Workmanship ard Construction6. lasting and Instruaie-ílation7. Inspection ^InstrtJiDer.t.ation and Procedure"^

The work on ÍÁS draft oesign of recommerdatjons is presently proceedingin a number of study groups.

122

Part 3. Some Special Sngineering Developments for Pre stressed Con óreteReactor Pressure Vessels

3.»1.« General

Ihe diverse structural possibilities and problems of PCRPVs have notyet been investigated systematically with full inclusion of all factorsof conceivable system designs of

1) Gas-cooled power reactor development, and2) Water-cooled power reactor development,

with simultaneous consideration of variants in construction to obtaina reduction of the construction and installation periods.

Within a fi & 5 program studies are being started in which con-struction possibilities and problems of PCHPVs will be investigatedfrom an overall economy aspect with systematic inclusion of all factorsof conceivable system designs and simultaneous consideration of variantsin construction possibilities in a collaboration of experts in reactorsystem design, prestressed concrete construction as well as experiencedbuilders.3.2. Multilayer and Prefabricated Vessejl Structures

The thick-walled PCRPVs to date have been built of massive monolithicon-site placed concrete. Designs (l) for multilayer containers witha pressure- transferring liquid intermediate layer, and (2) for containersof precast concrete design have been realized in model form arid tested.3,2.1. Multi-layer PCRPV

Ifce company Pried. Krupp G.ro.b.H. Universalbau, Essen, has developeda new concept in PCHPV design: a multi-layer (double-wall) vessel withintermediate water gap under automatic pressure equalization with thereactor cooling medium. Each layer in the composite vessel has assigned

123

its special function. The arrangement, composition and function of theindividual layers proceeding from 3r.side to outside is as follows:

The first inner layer primarily serves as a radiation shield. Itis a wall made of high temperature resistant concrete of comparativelyhigh density and incorporating a relatively high content of hydrogenin the aggregates whinh must be oheraically firmly bound up to ratherhigh temperatures. This concrete wail is backed by a layer of thermalinsulating materi&o.. In order to ensure a gas- and water tight seal,the two initial layers are interposed oelween two thin steel liner plates.The next layer is a water-filled gap. The pressure of the water in thisgap is either totally or partially self-equilibrating with the reactorcoolant pressure. The heal passing inte the water gap through the internalwall and thermal insulation layer is removed by a secondary system ofcooling pipes located ir the water gap. The outer wall is the load bearingpree tressed concrete pressure vessel -which 3s outpide o-** thermal and radi-ation influences.

The concept has been described in some detail e.g. in ref. (2).Result? from relevant model tests using a 1 :•> mode1 corresponding toa PCBPV of dimensions as used for the Bue^y 1 nuclear power plant havebeen describee, in ref (3).

.2,2, PrefabricatedSiemens AG, Erla&gen, has been developing a PCRPV for water-cooled

power reactors made of prefabricated parts.This structural concept and the model for testing have been described

in some detail e.g. an ref. (4), Some structural problems have been studiedin further detail on a ring model representing an area cut out horizontallyof the cylindrical part of the vessel, see ref. (5). - See also Appendix A. II.

«3. Prestressing SystemsThe structural and construction problems of PCnPVs concern primarily

the arrangement of the large number of largr tensloning members and thedevelopment of a structurally and economically satisfactory prestressing

124

system. Two principal aspects «.«ply to the design of the prestressingsystem;

1) The prestressing members should be arranged in a way that thevolume for the placement and consolidation of concrete and forthe installation of unprestressed reinforcement will no.t beexcessively compartmentalized, i.e. large tension members ingrouped arrangement are favorable in this respect.

2) The arrangement oí the tension members must be based on theavailable surfaces for anchor arrangement with considerationof the access requirements for the prestressing presses.

The latter problem may he eliminated for the hoop prestressingsystem by using anchor-free prestressings Either individual tensionmembers placed in rings around the cylindrical surface and locked intoclosed hoops "by tension clamps or wire winding prestressing. Developmentwork for such a system is described in Appendix B.

125

Part 4. Development of Kumeriojdjgejb^^^ Analysis forPrestressed Cpnctete Reactor Pressure Vessel£

f/i" problem of struotnral analysis of t PCRPF can be de '3red as anonlin-*.-, time-dependent and te.n^3ratujre~o.ependent three-dimensionalbonwdi*ry value problem of a compre? composite structure. The concrete

is to be considered as a nonhomogeneous noniinea^, viscoelastic"«triable in its properties under the inflnenoep of loading con-

aitio u , ten>perature and moisture fields as veil as fciTie end becominginoren t^£ly anisotropio by craok fornmion. Thus, a reliable calculationof a ox.,vorting structure sm&fc be based on realistic assumptions concerningthe ircaustic material bebaxiour as well as bonding effects and the historyof the mechanical loads and of the coupled temperature and moisture field.

1i«c, requirements for the computing methods and principles for theselection of suitable methods have been briefly discussed in ref. (6 ) jthe reader is referred to Appendix C.

Referencos

(1) Bundesanstalt fur Mate rial prüfung, Pachgruppe ?.2i Grundsatz-Pr^gramm der deutschen Porsohung und Bntwicklung fur Spannbeton-Pe-^ktordruckbehalter. Berlin, Version of 1?. April 1?69; (EnglishTranslation: Proposed basic German Research and DevelopmentProgram for Reactor Pressutu Vessels of Pres tressed Concrete,Preliminary Vercion of 1? March 1969; cienfcjf jc TranslationService, STS Order No. 7¿,21 í v r Oak Rioge National Laboratory,13 «iune 1969.

(2) B. -sraer, P.: Multi-layer (Double-wall) Prestressed Concrete Pressure" bel. Nuclear Engineering and Design 5 (1967), p. 183.

H.P.: Results of Model T-asts on Multilayer pressureVieseis. 2nd Conference on Fres tressed Concrete Reactor Pressure'ocsels and Their Thermal Insulation, Commission of the Europeanw-otmrunities, Brussels, Tf-20 November 196$.

.0^.«R, H,? QRJHL, H.t Spannbeton-ReaVtordrucVbehalter fur 100 atüj nendruok. Technische Uberwachung 7 0966"), No, 1, p. 10.

126

(5) BISSLER, B.: Reactor Pressure Vessels Made from PrefabricatedParts — Part Model for Testing Design Improvements. 2nd Con-ference on Frestressed Concrete Heactor Pressure Vessels ana •Their Thermal Insulation, Commission of the European Communities,Brussels, 18~20 November 1969.

(6) BRAKD3S, K.s Numerical Methods for Calculating Prestressed ConcreteReactor Pressure Vessels, -ibid.

127

A* -_StruoturaJ Systemo and Model Structures

I, Joint Development Program "Pres¡tressed Concrete Pressure Vesselsfor Boiling Water Beactors" of Ag!CrTeiei>unken and Fried. Krupp GmbH*'

1,_ Introduction

The Allgemeine Blefctricitats-Gesellschaft (ATDCt-Telefunken) and theFried.Krupp GmbH (?K) have jointly established a cooperative developmentprogram aiming to achieve the technical and economical feasibility ofthe use of prestressed. coders fee pressure vessels for "boiling water powerreactora.

The work to be done in this development program has been grouped intothree successive sections:

a) Systems analysis,b) Conceptual design study.c) Reference design.

The tasks are:

a) Engineering (structural analysis and deejgn).b) Tests.o) Safety considerations.

The integral program is being jointly conducted by AEG-Telefunken «andPried. Krupp GmbK (PK) with the following main departments of PK partici-pating:

PK GmbH Industriebau und Maschnnenfabriken Essen,HauptabteiJung Xerntechnik,

PK GmbH Zentralinstitut fur Porschung und Sntwicklung,PK GmbH UniversalbauPK GmbH Maschinen~ und Stahlbau Rheinhausen.

The responsibility of steering the vrhole program has been assigned to thePried. iCrupp GmbH Industriebau und Jfaschinenfabriken Kssen, HauptabteilungKerntechnik.

TJCommunicated by G. Sohafstall and G. Lukacs, Pried. Krupp GmbH Industriebauund Maschinen-fabriken Essen, Hauptabteilung Kerntechnik, Essen.

128

«?, Systems AnalysisWithin the frame of the systems analysis tbe boundary conditions —

particularly the nuclear, thermodynamical and safeguard engineering designdata —« are evaluated with respect to the prestressed concrete pressurevessel itself as well as to speeda^ associated internals (hot liner andthermal insulation). The fuel element loading/unloading procedure andthe arrangement of the reactor components within the prestressed concretepressure vessel are analyzed.

On the basis of the determination of the boundary conditions thetesting conditions and the arrangement of the projected tests are settled.

Finally, the possibilities o? utilization of available reactor compo-nents (as they are used with steel pressure vessels BftHs) are evaluated,the eventually required structural modifications are studied, and thevolume of the necessary mechanical engineering design and development workis estimated.

1« Conceptual Design StudyOn the basis of the results of the systems analysis a conceptual design

study is performed, the purpose of which is to find a suitable combinationof the layout of the reactor primary system and of the design of the pre-'stressed concrete pressure vepcel.

This part of the program is essentially characterized by a close inter-*action between testing work and structural design work.

Starting from the evaluation of the layout conditions and of the testresults the layout and mechanical-structural design of the whole primarysystem is worked out. In this phase as well isolated components as thetotal layout are considered. If necessary, mechanical components will benewly designed or modified. With some of the components, e.g. the fuelelement loading/unloading machine, this may require the performance of tests,

129

In order to find a technical and economical satisfactory solution,beyond the study of the prestressed concrete pressure vessel and itsinternals, an overall study of the primary system, respectively of

***tbe whole equipment arranged within the containment will be performedincluding an analysis of construction procedures.

Coupled with these purely engineering studies, economical analysesare to be performed in order to arrive at an optimum concept,

4. TestsIn the frame of the total development program the following

tests are scheduled:a) Tests on the "hot liner".b) Tests on the thermal insulation»c) Concreting tests for the bottom closure of the vessel.d) Photoelastic analyses.e) Testing of the wire winding procedure.f) Operational testing of special components of the fuel element

loading/unloading machine under simulated reactor conditions,

5. Reference DesignAfter finishing the conceptual design study inclusive the tests,

the work on the reference design is .-..tarted, considering a boiling waterreactor nuclear power plant of about 1100 We generating capacity.

The reference design will yield;a) an integration of the prestressed concrete pressure vessel

into the primary system of a boiling water reactor with dueregard to economical and safeguarding viewpoints,

b) the proving of the technical feasibility of this new system,c) a cost evaluation for the assessment of the economics.BWR power plants of differing generating capacities shall be studied

in comparison with the 1100 MWe reference design, in order to determinethe influence of the parameter groups for further extrapolations.

The results of the overall development project will be describedin a final report.

130

II. Peyelgpment rk for r&s^^seá jiwonc^e^e Pressure yessels forWater-Cooled Power Kcaotors at Siemens AO.

* • Some Conclusions from the JPejr e j,opmea PCBPV Structures

As a resume of the development work done at Siemens AO in recentyears it can be stated that the construction method for prestressedconcrete reactor pressure vessels using prefabricated concrete blockscan be regarded as completely suitable for vessels to be designed forinternal pressures up to 100 atmospheres under the requirement of onlya few penetrations through top and bottom closures» In such oases themethod of anchorfree circumferential presfcressiag - in our opinion -offers considerable advantages which are mainly due to the lack ofcomplicated end anchorages. The problem of radial penetrations throughthe cylinder-walls can be solved without too much difficulties.

771 th increasing internal pressure, however, in our opinion there isa limit reached rather soon beyond which it is no longer possible toapply the radial jacking forces to ths concrete. ?/ith the operationalpressures of pressurized water reactors of about 16*0 atmospheres thislimit is definitely surpassed, so that such a vessel can only be con-structed in monolithic concrete.

The development of suitable vsrire winding met\ods for prestressedconcrete pressure vessels wruch is being pushed forward at severalplaces will probably lead to a decrease of the incentive for vesseldesigns using prefabricated blocks, which mainly lies in the lack ofanchoring ribs, and with relatively minor modifications we can changeover to using the wire winding method.

+) Communicated by R. Rissler, Zentralabteilung Technik, Reaktor-teohnik, Siemens AC, Erlangón

131

The adaptation of the* pres ¡.r«;*se<l concre-j* pressure vessel conceptdeveloped at Siemens AO io the reqm reraents of pressurised writer reactorsmainly leads to -he fo .lowing ¡nod j float ions:

(1) Turning a'vay rr-oj:-: the deaign w-sing prefabricated Vooks toa monolithic concrete design utilizing a nuicable anchor-freeprestrassing amhod.

(2) Turning away from the pl'ug closures The necessary adaptationof the vessel structure to the well proves concept of theSiemens presr.ufixed water po-.v<si» reactors \n the case of lightwater realtors requires complete reaorraMlity of the topclosure of the vessel and penetration o*" 60 to 80 control rodsthrough the top closure.- 'A'J th heavy water reactors some300 to 40C loading pe/ietrat.ona taunt be provided in the topclosure in considerably ooser arrangement than it is usual inthe case of gas-cooled i;ower reaotorr,, ~ Thus, both reactortypes require a severely perforated vessel closure which, underthe imposed conditions can be fabricated only of steel.

2« Present Trend of Beyelopipent ^Worjc

Due to tne reasons outlined in paragraph 1 we are presently startingdevelopment work for the following conceptt

The concrete pressure vesseJ is provided with a steel closure simi-lar to the type of closure used for steel pressure vessels. However, taisclosure is not bolted on a flange but it in supported under 45 to aboveby a series of hall-type support «aemoers zesting1 on cast steel hearings(similar to bridge bearings) which trans-cit the forces to a preatressedconcrete ring on top of xhe cjrlínórical part of the vessel.

The essential feature of too support5r,# structure is that the ring isseparated fron tbe cylinder wall by a sliding ,'joint hut is prestressed againstit by vertical prestressing cables. In the ring the vartlcnl and Hori-zontal prestressing is arranged in a way that in horizontal directioncentrical prestressing is applied to tne ring but that the resultant forcefrom both prestressing directions meets with i be- v«ctor of the supportingforce in a point so that under no loading condition oonstr-'iints are actingon the cylinder ¿rail which ie. already highly stressed by internal pressureand temperature gradients. The water-light lining ^ n Cnis area infacilitated b¿; a sys bena of bellows.

132

III. PCRPV Model for the 300~M?/e Thorium High Temperature Reactor .Prpjeo*

At present, a 1 : 5 scale model of the prestressed concrete reactorpressure vessel for the }00-líWé TrtTR prototype nuclear power plant isunder construction at Pried. Krupp GmbH Universalbau. The studies tobe performed with this model structure will be in concern of thefollowing:

Determination of the limit of elastic behaviour of the vesselstructure.

Strains, deformations and stress concentrations in the surroundingsof large vessel penetrations (for blowers and for removal of heatexchangers).

Stress field variations under long-time continuous and alternatingloadings.

Cracking and fracture phenomena with increased loading up to theultimate load.

Temperature distributions in zones of complicated geometry.Mastering of complicated concreting procedures, mainly in the domain

of the control rod penetrations in the top slab.Effectiveness of prestressing with cables of high prestressing

force capacity with large bending angles.Besides of operational loading conditions setup and emergency

loading conditions will be simulated.!Ehe results of this model testing are not only of specific interest

for the TSTR prototype vessel but beyond that they will be of quitegeneral interest for the design and analysis of further vessel structures.

+} Communicated by P. Bremer, Pried. Krupp GmbH Universalbau, Essen

133

APPENDIX B. -Prestressing Systgic Development

I. The Wire Winding System "PRISMA" for BuRP'/s1. .Wire Winding Prestressing

A multilayer circumferential wire winding prestreseing system hasbeen developed by the company PhiJipp Bclzraann AG, The system usesan equidistant arrangement of separated prestressing wire lajers(Pig. B~l), The spaces between the layers facilitate inspection andmaintenance, provided that the spaces are made wide --noxigh. Platbars are used as spacers. The use of spacers impljes that the wirewinding prestressing runa polygonalIy with resultant concentrated forcesbeing transferred to the concrete vessel at the bents (Fig. E-2) . Bymeans of several wire turns side by s:;3e, which are named "rows" aprismatic shell surface is formed (Pig. B-3) which has given the name"PRISMA" to the system.

2. Sejctions of 9/ire WindingThe speed of the vrire winding machine is res trie ten n;-/ several causes.

But since the winding of the prestressing reinforcement in general lieson the critical path the time available for bhe vrindlsg operation islimited. Therefore, for the winding of the prxsma reinforcement, a wirewinding machine is used which faciHtatas simultaneous vinding of 4 wirea.The 4-wire tendon is named "tensaon .ng tape".

A possible way of wire winding prestressing is to cover completelythe surface of the PCPV with terisioning tapes and to adapt the requiredprestress by different numbers of layers. However, it is advantageousto subdivide the prestressing in bo several -sections of wire winding.Then rails for the wire winding machine can be mounted between these sections,

Each section of wire winding is acting like a large hoop tendon with-out anchorages, which transfers the deflection forces at each bent viapressure distribution plates onto the concrete. Tae winding operation can .be interrupted so that a partial prestressing is possible.

+) Communicated by L. Mühe, Philipp HojZKsnn AC, Prankfurt/Main

134

3L. J'In-between-Tan chorares"Several "in-between-anchorages" are provided for each prestreesing

ring where a complete bond is set up between all wires of one windingsection. At those points the free force, of an eventually broken wireis transfered to the other wires. A wire rupture thus affects only asector of the prisma pre^+ressin» "between adjacent anchorage ledges.

Favourable "in-l tween-anohorages" are facilitated by means of theprestressing system "KA" wh.1oh has internafcionally proven its reliabilityover many years (l). Between the clamping plates, tensioning tapes arearranged (Fig. B-4) so that a packet is built up layer by layer. Ttymenas of pressure plates, clamping forces are transferred to this packet»In this way a bond, due to friction ané shear, is set up between the wiresand the clamping and pressure ledges. The tension force of the wires isthereby transferred to the adjacent plates. It is characteristic of theclamping anchorages that the required clamping forces are not dependenton the number of layers, but they are depending on the number of wires ineach layer and its friction and bond properties. The clamping forces areproduced by means of high-strength prestressed clamping bolts which havea large elastical deformation under the effect of the full clamping force.Thereby it is secured that the clamping pressure required to anchor theultimate tensional wire load remains effective also after a small settingof the clamping packet»

The described "in-between-anchorages" which are designated as anchorageledges (Fig. B-5) consist of a pressure distribution plate, base ledge,clamping ledges, pressure ledge a;.d clamping bolts. The example of Fig. B-6shows the use of ribbed oval wire with 20 layers of teasioning bands.(Fig.B-7)

The diverting ledges are built up (Fig. B-8) in accordance with theanchorage ledges. Instead of clamping bolts there are a reduced number ofspacer bolts. These ensure the position of base and in-between-ledges andthe space between the tensioning tapes. The spacer bolts are arranged ina staggered manner so that in general only one pair of bolts is adjacentto the tensioning tape over several diverting ledges.

135

4¿ ____ Spiral Winding; Step Winding

In Pig. B-9 a \7-sided srism? reinforcement is «hownj there thefour anchorage ledges are marked "1" ?nd the 13 diverting ledges are marked "2M.The developed elevation of the "barrel shows a spiral run of one layer of windingThat is why this arrangement is called "spiral winding". The winding operationstarts at the bottom. Tho "beginning ar.¿ the end of the tensioning tape isfixed "by separate clamping anchorages (described later).

The disadvantages of spiral winding stem from the expensive fixingof the difficult leading of the wire winding machine and the necessity towind the next \zyer in a descending spiral Therefore the prisma systemuses the so called "step winding", which is e^ow» in Pig. B-10. By thismethod 5 anchorage ledges and 1? diverting ledges are arranged directly onthe ad.laoent noints of the 17-sided polygon.

Between these two anchorage ledges the tensioninp- tañe stems from onerow to the other. The wire winding machino runs exactly horizontally(starting from the "bottom). In the field bet*reen the t'-'o adjacent clampingledges the tensioning tape step? to the next row. After having reached theur>per row the tensioning tape runs in the opposite direction and the secondlayer is anplied. As shown in Pig. B-ll th<% "bent forces are neutralised"between two layers at tho step.

Coupling of Wire: Anchorages for Start ing and Piniohing

Tn order to produce a '.-rindins «oction of nrisma reinforcement,extremely long wires are needed. Therefore it is necessary to couple shorterlengths of wire which is rolled up in transport coils. Bibbed oval shapedwire is coupled by using a pressing tube joint (Pig. B-12) which can easilypass through the wire winding machine.

Anchorages at the tensioning tape are needed at the start (Fig.B-13)and the end of each winding section. The anchorages* at the start becomesuperfluous once the clamping bolts at the anchorage ledge have beentightened, but they are left there thus providing redundancy.

136

The anchorage at the start is a standard clamping anchorage of theprestressing system KA with two clamping holts. It is mounted in thesection of the step change in front of the anchorage ledge. It isarranged in a niche ir. a way that the first tensiocing tape can passover. The horizontal forces will he transferred by flat bars whichare embedded in concrete.

The anchorage of the end of the tensioning tape can "be mountedwithout any additional operation on a clamping ledge (see Fig.B-14).The same procedure is used when the wire winding operation is inter-ruptea, as it is necessary, for instance, for successive partialprestressing.

6, Construction Procedure

With the proceeding of the concreting work pressure distributionplates and rails for the wire winding machines are built i». Afterpouring the concrete, round Jink chains - which are used as tensioningmembers - are placed into the rails and the wire winding machine ismounted. At the same tisce the clamping holts are fastened and the spacerbolts are welded.

The undercarriage of the wire winding machine hangs in the round linkchain and is supported onto the rails. On the machine there are grippingdevices for the wires, the brakes, the engine as well as the controllerand measuring mechanism in order to fix the tensil*» force.

In accordance with the statical requirements the wire winding oper.it i oncan he interrupted within one wire winding section at any anchorage» letgeand can he continued later. After one winding section has been coveredwith the required number <*? wires the pressure ledges will be fixed andpressed together by clamping bolts.

References? 1.) L. Mtthe: Kleramverankerung vereinfachtSpannverfahren

Bau und Bauindustrie 10/3962, S.394 - 400L. MQhe: Spannglieder mit grosser BinaelepannkraftBau und Bauindustrie 6/1965, S.282 - 299

137

2.) H. Wasoheddti 3t8.hle fur Bautelle aua SpannbetonBetonstein-Zeitung 9/3968.

138

Report on the Starting of a Coordinated Work Programfor Prestressed Concrete Reactor Pressure Vessels in

the Pederá] Republic of Germany

APPENDIX C - Computational MethodsS»«.«»ssss:*«ssie«:s*s«re3ss 33t!ssis:srss!nsisss.«e¡ssc«s«ss:te

^- dimensional Matria: PislacementSystem ASKA ( TSD. Stuttgart) *)

The Finite Element Matrix Displacement Method, which was developed in1954/55 [5-J TO,a originally intended for the complex static and dynamicproblems of the aircraft and spacecraft industry. However, due to its general! t,the method has found application to a vast array of structural engineeringproblems. The rapid development of large high speed computers in conjunctionwith graphical input/output devices has raaáe the Finite Element Method evenmore popular (see e.g. L^J* /"-L/. Tt must be recognized, neverthe'ess,that the use cf this general method demands a highly sophisticated software.At the ISD, the general matrix displacement method has been implemented inthe problem-oriented language ASKA (see e.g. /"X./, /. 5_7 of which severalversions have been developed. Ihe most general version £&J is writtenin FORTRAN 4j it includes flange, beam, membrane, ring, plate and shellelements in addition to a number of elements for tfce analysis of generalthree-dimensional problems,

For the structural analysis of prestressed concrete reactor pressurevessels (PCRV), work has been begur. at TSD on a special version of AS£A.In addition to the linear thermo~elastic problems encountered in the PCHVstructures, this version will efficiently handle vis co-el as tic problemswhich arise due to elasto-plastic behaviour of the steel parts of the structure,or due to creep. Also, the anisotropic material properties of reinforced andpossibly cracked concrete are taken into account. This ASKA version, inconjunction with powerful third-generation computers, will allow for theefficient analysis of all critical structural aspects of PCRV's.

Communicated by J.H. Argyris, K,?. Buck and £. Grieger,Institut fur StatiJt un<) Dynsmik der Luft- und Raumfahrt-konstruktionen (iSP), Urnversitab Stuttgart

139

Por a deeper,understanding of the theory the reader 1s referred tothe references compiled in refs. /"4-T» Z~5_7« ** should be stressed thatASKA is problem oriented and thereby relieves the engineer of tedious andunnecessary details, allowing him to depict his structure with an elegantdescriptive language and guide the calculations by simple statements. Thestructure to be analysed could come from different fields of engineering, suchas civil, marine, aeronautical, nuclear *»tc. In all oases, however, theASKA solution procedure employs fche samo --veil defined steps. First thephysical model of the structure - also often callee the idealisation - hasto be selected. This idealisation process is decisive for a successfulanalysis, and one tends without previous experience to be too basty at thisstage. It should not be forgotten that ASKA .yieMs the solution to the'idealised model and not the actual structure.

Having defined the idealised structure, the necessary input data mustbe prepared:

1. The tapoiogical description, which can be comfortably generatedthrough the problem-oriented descriptive language contained in ASKA5

2. the element properties, i.e. cross-sectional data, material properties}3. the co-ordinates of tne chosen nodal points, and4. the external loading systems.

The latter include temperature loading if so required. The input datacan optionally be scrutinised by the system, and the user will be informedof any discrepancies encountered. Throughout the checking procedures, ASKAdoes not stop at the first error found, t-tt continues to try co clarify as muchas possible per run. Should, however, a fata] error occur that would makethe remaining calculations meaningless, ASKA automatically terminates the run.

The ASKA system has been designed to keep a running check of all calculations,digesting intermediate results and proceeding only when these are found to bevalid for the physical or numerical sta.te they represent. Such checks may,of course, be time-consuming but experience sbovsrs t*\at they are absolutelynecessary in a proper software system. Moreover, the man-hours and computer-hours needed to try to locate an error - that as soon as it is found usually isonly too obvious - in a system without running checks do very often exceed theamount of time spent at checking the calculations and data in a properly designedsystem anyhow; not to speak of the psychological consequences of frustratinghours spent at searching for these unavoidable human errors.

140

fo achieve maximum flexibility for the user when applying ASKA, eachseparate step (i.e. an ASKA Statement) means the end. of a certain well-definedtask and the computations can therefore be terminated after having executedany ASKA Statement, followed by a 3ater restart at the user's convenience.

In the design of ASKA, a modular concept was strived at throughout.Therefore, ASKA can easily be updated in terms of either structural ornumerical innovations to reflect tie latest state-of-the-art. At present,an interaction between ASKA and digigraphic devices is being programmed tooptimise and ease the communication between man and computer. Furthermore,this is also a vital step in the direction of an effective Computer-Aided-Design (CAB)-system.

Already at the initial stages of our problem-oriented computer languagethree-dimensional elements, such as TET4 and ÍW3T10, (Pig, l) - were included.The TET4 element - a tetrahedron with four nodal points with three degreesoffreedom each - is based on a constant strain and stress distribution. Due tothis simple assumption, this type of element as a rule reproduces only theglobal behaviour of the structure. For a more accurate stress distributionthe TET 10 element based on a complete set of second order polynomials(linear strain and stress distribution) is more suitable. A consistentdevelopment in the direction of higher-order elements leads to the TBT20element based on a complete third order poiynoroiai for the displacements orparabolic distribution for the strains. The nodal points of this elementare not only placed at- the edges but also at the centroide of the four faces.Each nodal point again has three degrees of freedom so that the total numberof freedoms is 60.

Considerable advance was subsequently achieved by the introduction ofelements with curved boundaries like the second and third-order elements TEC 10and TBC 20 respectively (Pig. 1). Por these elements the interpolationdescribing the displacements is also used to map the general curved elementonto a regular straight-edged "parent element". A smaller number of theseelements are needed to idealise regions with steep gradients, which arefrequently associated with geometrically complex shapes. The topoligioaldescription of complex idealised structures is often facilitated by buildingso-called macro elements such as general cubes consisting ol' six tetrahedronelements. Incidentally, the tetrahedra are the only three-dimensional

141

displacement elements allowing for a concise and really "olean" theoreticalderivation; it is moreover possible to base these elements on completeinterpolation polynomials of arbitrary order for both geometry and displacements»

A different family of general three-dimensional cube elements (which alsohas its two-dimensional counterpart) is based on the Lagrange interpolation;these elements are denoted by LUMINA (Fig C-2) in ASKA. Tae faces of eachcube may be curved. Together with the general curved triangular prism elementPENTA (Fig.C-3), these elements may be used for flexible and efficient ideali-sation of structures such as occur in PCRV's. The number of necessary elementsin the structure clearly decreases if the order of the strain distributionadopted for each element increases.

Por axisymmetric structures It as convenient to devise special ringelements. One of these families of elements is the TRIAX groups theseelements consist of rings with triangular cross section (see Pig. C-4, 5)the edges of which can be curved or straight. It contains elements basedon a first (THIAX 3), second (TRIAX 6) and third (TRTAX 10) order polynomialinterpolation for the-displacements (and geometry) across the radial cross-section. For the idealisation of membrane shells, e.g. the liner in nuclearreactor vessels, the FLAX family of elements has been developed (Pig. C-6),which also may be used for conveniently describing distributed pressureloading acting on TRIAX elements. The TRIAX and FLAX elements may begeneralised to account for non-axisymmetric loading actin on axisymmetricstructures: the general loading is represented by a Fourier series and thestructure analysed independently for each harmonic of the series. However,this device is limited to truly rxtsymmetr.' c structures whose materialproperties do not vary in circumferential direction. In the presence ofsevere temperature gradients this may not be the case, and for such axisymmetriostructures, the SECT family of elements was developed (see Pig. C-?).

The above list forms only an extract of some new element developmentsfor the matrix displacement method which are of particular importance toPCRV structure.

We include two examples to illustrate the application of above elements.The first example illustrates the analysis of an axisymmetric prestressedconcrete pressure vessel. The configuration is an oversimplified academicone, but the problem does include the essential features of practical

142

prestressed concrete reactor vessels. The details of the problems are givenin Pig. C-8 together wit* the chosen idealisation. For the concrete, TRIAX 10elements were employed, while the steel liner was idealised witn FLAX 4elements. The natter elements were also used (wi fch zero stiffness) to applythe prestresfling pressure due to the meridional tenlons, whereas the forcedue to the circumferential ana the vei Asa": prestreH»in¿ cables are directlyentered in the load matrix. Further reinforcements could easily be includedbut are omitted for the sake of sirtiplici ty. Vhe ca-culated deformationsara shown in Pig. C-9 whlcu also contains ooni.our lines 01' the circumferentialnormal stress in the concrete. F5g.C~10 contains a comparison with resultsobtained by an analy«i 3 base'! on the si<np3er TRIAX 3 elements and the samo nodalpoints as in Fig. C-8 (and thus the same number of unicnowns); the superiorityof the more sophisticated e^enenbc; is thereby well i!1ustratsd.

To illustrate the application of the LUMINA element to a three-dimensionalpressure vessel analysis, we include as a second example ^he thick-walled pipejunction iepicted in Fig. C-ll. This structure is loaded by internal pressure,to which end loads are added to simulate the prsssure acting r»n the end caps.The structure is idealisfid by HSXS 27 elements, i.e. 1/TMTNA elements with a3 x 3 x 3 grid of noda" r>oints in each element. The calculated displacementsare shown in ?ig. C~12 for the two symmetry p3ar.es, ar¿3 the von Mises equivalentstress is plotted in Fir. C~i} for the longitudinal p'iane of symmetry.

As indicated above, the appliontion of the matrix displacement method ienot limited to linear problems of stress and deformation analysis, but provesa powerful tool for the solution of non-linear problems a,:so. In PCHV analysis,non linearity is mainly due to tae non-linear material properties, whereasgeometric non-linearity (large deflections) seems to l-e of secondary importanceonjy. Step-wise procedures are used to solve these non-linear problems bythe matrix displacement method.

Several approaches are known to determine ^h& elasto-plastic behaviourof the steel mem&ers -'n ^CRVs, i.e. the liner and the prestreosing members (seej/o.J7, /"XT'5- The simplest approach is the tan^ntial stiffness method.This method is ba^ea on the c> las to-plastic material stiffness, and elasto-plastic element stiffnesses are cal cuín te-i anew in each step, depending on thestress in each step anci the load history. In PCRV applications, where the

143

analysis can usually be oased on small strains and iisp]aoements, theinitial load technique in general proves advantageous. ?he initial loadsare defined for each element in such a manner tnat they suppress elastioallythe initial strains or stresses. If one chooses to -work, with initialstrains, the incremental e lasto-plastic strains are derived from the incrementalstresses. On the other hai-1, working f~om initiaJ ¿tresses, The incrementalstresses are calculated from tae inoremen al e3asto-plastic strains» Thisstep, which la itself is noa-linear, is ic. both oases best performed byiterative methods. It bns been bitown that tne initial strain method «aydiverge, especially for Mw/Ty redundant sy^teits, ?ae range of applicabilitymay, however, be extendel considerably if a sophisticated iterative deter-mination of the incremental «J%raj.ns io used, For ideally plastic materials,the Iftiiial strain appruae'x canso' be used, Tr.«? initial stress method hasbeen proved in £"(>J *^ converge in a1.] caf.rfs (even for ideally plistic material),as long as the structvre or part tnereoC is not statically determinate.

The approaches outlined above for ela^tc-plastic analyses may in principiebe employed to take account of creep strains also, as occur in concrete. AtISO, work is currently in progress in an attempt to simplify the analysis ofcreep. In any case, «riven the (rather complex) material properties, i.e. thetime dependence of creep or shrinkage strains dependent on stress state,temperature and other state variables, it is possible to conduct an analysisof the time-dependent stress and deformation behaviour of two- or three-dimensional FCRV structures based on the matrix displacement method.

In addition to the plastic and creep -strains, the deformation characteristicsof POBV structures are decisively determined by the development of cracks inthe concrete. This material non-linearity (the material mainly sustains onlycompressive principal stresses) is to be simulated in the finite element analysisby introduction of anisotropic materJai properties; it should be noted thatthe finite element method requires no limitation on the anisotropy (no fixeddirections of orthotropy). The analysis will apain pi*oceed step-wise.

It is seen then that the matrix displacement approach of the finiteelement method is ideally suited to handle the complex problems encountered inPCRV structural analysis. It seems fair to state that it ie, at least for thetime being, also the only method capable of so general a tasK.

144

References

ARGYRIS, J.1" t Energy Theorems and Structural Analysis.Aircraft Engineering, 3954 ~ 1955: a- so Buiterworths, London. I960

RASHID, Y.R. r On Computational Methods in Solid 'Jlechanics andStress Analysis. Proceeding Conference on the Effective Use ofComputers in the Nuclear InMisfcry, Knoxville, Tenn., 21 - 2} April,U.S. A. E.G., Biv.Techn.Inr., CONF-690401, p.2UDANS, Z. : Survey of Advanced Structural Design Analysis Techniques.Nuclear Engineering and Design 10 (1969), p. 400

ARGYRIS, J.H. : ASKA - Automatic System for Kinematic Analysis.Nuclear Engineering and Design 10 (1969), p,441

ARGYRIS, J.E., GRIBPBR, 1,: Three-dimensional Elastic and Bias to-Plastic Analysis of Reactor Pressxire Vessels. Paper presented at2nd Conference on Prestressed Concreto Reactor Pressure Vessels andTheir Thermal Insulation, Commission of the Suropean Ooraarunities,Brussels, 18 - 20 November, 1969,

ARGYRIS, H.H., SCHARPF, D.W. : Methods of glasto plástic Analysis,Proceedings of the ISD/ISSC Symposium on Finite Element Techniques,Stuttgart, June 1969

SCHARPP, B.V,;. : 'Die Frage der Konvergenz bei d«r Berechnung elasto-plastisch deformierbarer Tragwerke and Kontinua, Dr. Ing. thesis,University of Stuttgart, 1969

/"8J7 FUCHS, G. von, SCHPJSM, E. i ASKA - A computer System for StructuralEngineers;Proceedings of the ISD/ISSC Symposium on Finite Element Techniques.Stuttgart, June 1969,

145

Listi of Figures

C-l The TBT and TEC families of Three-dimensional Elements

C-2 LUMINA - A General Hexahedron Element

C-3 PBNTA - A Pentahedron Element

C-4 The TRIAX Fondly of Ring Elements for Axially Symmetric andHarmonically Distributed loads

C-5 The TRIAX Family of Ring Elements with Curved Edges for Axisymmetricand Harmonically Distributed Loads

C-6 The FLAX Family of Membrane Shell Elements for the Analysis ofAxisymmetric Bodies with Symmetric and Harmonically Distributed Loading

C-7 The SECT Family of Sector Elements for the Analysis of Problems withAxisymmetric Geometry and Unsymmetrio Loading

•

C-8 Prestressed Concrete Pressure Vessel ~ Idealisation

C-9 Prestreseed Concrete Pressure Vessel - Displacements and CircumferentialNormal Stress

C-10 Prestressed Concrete Pressure Vessel - Comparison of CircumferentialStress and Vertical NormalStress for Two Idealisations

C-ll Thick-walled Pipe Junction - Idealisation with Curved HBXB2? Cube Elements

C-12 Thick-walled Pipe Junction - Deformation in Two Symmetry Places

C-13 Thick-walled Pipe Junction - Equivalent Stress in Symmetry Plane.

146

APPSNDIX D« Research Programmes at the Bunáesanscaltfur Materialprtifung (BAM).

*

The Bun des ana tali, fur Mafcerialprlifung (BAM) is participating inthree areas of the R & 3 programme for PCEFVs? some relevant commentsgiven in the following sections.

Research on, "Sttrength and Charaoteriatioa of Fracture of Conore teunder Tfaree~Ptmensiosa1 Stress and Temperatures be fcween 20 C and ' 'C C

In prestreseed concrete reactor pressure vessels concrete has toendure long periods of Three-dimensional stress at elevated temperatures.Jut the present technical knowledge of the properties of concrete under;-ult:iaxial stress and especially at higher temperatures cannot "be regardec•s sxifficient.

The strength properties of concrete are influenced by a multitud» orparameters. For example the choice of aggregates is of great importance is*x"ar as the thermal properties and volume changes are concerned, that goes .orthe permeability and nuclear radiation attenuation properties, too. Becauseof the manifold parameters -which can influence the testing results, it isnecessary to make the tests with one certain kind of concrete, - a mixturowhich can be expected to be fairly representative for the kind of concretev.Mch probably will be used for the construction of the reactor pressurevessels. Further it is considered as absolutely necessary to align our te^tsvfith those of other research instituís so that the te«it resulta are compárenlaAlso it will be aspired to >rgaaize the research programs in such a way thatour and other institutes' test results complement each other. In this w.¿we can make the most out of the existing funds. The main partner in thicresearch is the concrete research laboratory of Pried.Krupp GmbH Universa1-^^u.

The first task in the research program is tne design and constructjnof a testing apparatus, which allows the application of stresses in 3 ortho/ron,.directions under normal and elevated temperature conditions up to 150 C. inaddition to that we plan to perform tests on concrete cubes of different

) Communicated by P. ScMnanelwit?., Pachgruppe 2.1, Bundesanstaltfür Materialprüfung (BAM), Berlin

147

Also it is intended to facilitate changing the kind and applicationof stresses e.g. to apply first stresses in ?. directions and keep them fixedwhile raising the stress in the 3rd direction continuously until failure ofthe conórete specimens.

Presently we are engaged in developing such a testing apparatus for500 Mp loading capacity in each of the three directions. There are specialproblems in the field of the oil hydraulic system and the measuring devicesas well as in the construction of the apparatus itself. Right now we arestudying different methods for applying tension to the specimens withoutcreating lateral stresses at the boundary. The diagram of Pig. B-l mayconvey an idea of the time schedule of the planned research projects.

II. Research on "Bffeot of Higher Temperatures^ on the Moisture Content*T " ~of Concrete" '

Concrete when harding shows a change in volume which is mainlyconnected with the changes of the inleraal moisture content. The state oftension caused by the moisture emission leads to deformations which arecommonly called the shrinkage of concrete. Present knowledge of the movementof moisture - espeically about the connections between distribution of moistureand the structure of concrete (type of pores, pore size distribution, etc.) -is not sufficient to estimate realistically the behaviour of concrete whenused as building material. This is especially true for the distribution andmigration of moisture in concrete elements of large volume which are subjectto thermal gradients (e.g. reactor pressure vessels). Research projects sofar were dealing with the phenomena and results only applicable to that certaintype of concrete tested with no remarks on the structure of concrete nor thedistribution of the moisture in the test material and/or the testing conditionsand environment. Therefore, the results of these tests cannot be appliedfreely.

Our- present research will include the foil owing tests: Tests on thedistribution and migration of moisture 3n concrete especially when subject tohigher temperatures. Besides tests on the subject of how the moisture contentdepends on the temperature gradient it is important to find out the influenceof the type of pores and the pore s.lse distribution as; well as the kind ofabsorption on chemical bond of water.

148

Pos. 1..__-> 6

An essential proble» is th« suitable, selection of the way o^measuring the moisture content of concrete!, especially the proper choiceof the testing probes to be vsed. Tb^s proVten will be studied as indicatedin the time schedule under "Basils'1 and "Preparations".

Poa, 7 - 14

?ihen we have selected all the Measuring devices required and aftersufficient testing of these devices we will start with the actuaJ research.The positions called "Pre~Testing" will raajnly be concerned with thelimitations of probT-ems regarding the material and its texture, which meansthe setting of boundary conditions and examinations of these conditions andthereby establishing certain rules of procedure for the v.hole project.

Poa, 15_- 16

The part called "Supplemental. Testing" serves the purpose ofevaluating ar.d reviewing the results.

149

III. - Instrumentation Assessment

. *)I,/ Purpose and Ala ofthe Instruroentataon Assessment Programme

Por the experimental checking of the calculabional methods and ofthe underlying assumptions on which the structural analysis of a prestressedconcrete reactor pressure veosel is based (inclusive of the assumptions forloading conditions), for the control of the vessel structural behaviour underoperational conditions, and in order to obtain experimental data required forthe future improvement of the design of prestressed concrete reactor pressurevessels, all relevant physical data have to be measured and processed in(l) materials investigations, (2) investigations on the structural behaviourof components and complicated details of the vessel structure, (3) mode]testing investigations, and (4) testing of tb<? operational behaviour and safetyof prestressed concrete reactor pressure vessels.

The most suitable instruments as well as measuring and data processingprocedures haije to be selected, with the requirements for measurements onactual vessel structures being the primary considerations. In the first line,this necessitates studies on tfce different measuring gauges which eventuallyhave to be further developed or some times even newly developed to meet therequirements. Such modified or nearly developed gauges then have to undergoextensive laboratory tests and are to be installed for trial in the firstsuitable prestressed concrete reactor pressure vessels as these are beingconstructed.

*) Communicated by N. Caaika, Fachgruppe 6.1, Bundesanstalt furMaterialprüfung (BAM), Berlin

150

The studies should lead to the formulation of detailed recommendationsfor the installation of measuring devices and the procedures of performing themeasurements in order to facilitate that all measurements with future pre-streased concrete reactor pressure «-essels can be performed in a uniform v-»yas they should. However, -che utilization of essential developments inmeasuring techniques should,of course, not be blocked by this ain».

2. Speo5.fi cat ions and Requirements^ for the J3auges

In different parts of prestressed concrete reactor pressure vesselcthe essential physical data are to be measured under partially very diff icul tconditions.

The parts OP which measurements are to be performed are: Steel li.-:tsrinclusive the anchora¿íep of the linor in the concrete, the cooling systemstructure (for reactor vessels with "odd liner"), penetrations tftrough tb>3vessel; thermal insulating system (for reactors with "hot liner"); taeconcrete body: prestressing wires and/or prestresssing cables and anchoragesor clamps; bonded steel reinforcement; the concrete surfacej the iowediatnsurroundings of th«* vessel.

The data to be measured are: Temperatures; nuclear radiation-moisture; prestressing forces? mechanical stresses; concrete and steel strainsinclusive of tbe bending strains of the linerj various, macroscopic deforma lions,e.g. buckling deformations of the liner, distance changes and rotations ofoperationally critica1. t>arts such as non tro"1 rod penetrations? crack widths;overall deformations and distortions* «to.

The measurement conditions firstly are determined by the roeasure^nhintervals which for the commissioning and pre-operafcional testing will be ofthe order of hours or days and for the continuous control of the operatior-\behaviour of the reactor vessel, fot- example, may extend over 20 years.Additional conditions are the disturbing parameters and their range of vaí-íoíf.ionv.ujch are corresponding with a part of the above mentioned parameters to boir-eesured. The nuclear radiation which is at the inner part of the vesselstructure sometimes is net negligible in its influences and the high tempe/aturesin insulating concrete in vessels with hot liner are particularly to be emphasized.

151

Further to be mentioned as conditions for the measurements are thegeometrical characteristics of the vessel, such as the thickness of theinsulating layer, the distances of the prestressing cables etc., as well asthe ffbterials characteristics such as Yourg's and Poisson's moduli, concreteaggregate sizes, etc.

By definition of the data to be w asured, their rar'jes of variabilityand the maximum permissible errors, and by the mentioned measuring conditionsthe fundamenta! requirements for the gauges and thereby the criieria for theirselection are essentially posea. Beyond, that shere are some further importantcriteria for the selection of gauges? ( l) Moi?turé tightness and ruggedness ofthe £Av,£es as required by the installation conditions? (2) in conjunction withthe raterials properties in tne immediate surronrdings cf the installed gaugesthe cilbUu^bances imposed by the instaJled gauges should be as sraall as possible,whic.n Indues e.g. that the gauges should not act as crack initiators in theconcrete.

As a final very iirportant criterion for the proper choice of gaugesthe costs of the measuring inntal latlcns as we"! I as the costs for preparationand performance of the measurements and their evaluations are to be mentioned.The eornomic criterion way require that during short-time testing and commissioningof the vessel a large number of relatively cheap gauges are being used, but thatthe control of the longtime operational behaviour is done by use of relativelyfew, relatively expensive gauges at we'1 selected critical points.

3.) Work Programme and Coordination

The tasks of the instrumentation assessment orogruT.nie have been dividedinto two parts: Part ] incorporates the problems the solution of which isunanimously regarded as neces?3sry by all engaged companies and institutions;Part ?. incorporates tasks the solution of" which is considered as desirable.The ."•'"at part of the work has been started, the second part presently has beenset baik.

The work to be done extends from a general literature search and thepr<?-5J.*-e formulation of tasks up to the trial installation and assessment of theter*aviour of gauges ana the other irfrtrutnentation devices in the first suitable.«•dual vessel which becomes available. The various tasks can le seer, from theacccTpanying time schedule (Fig. X>~3).

152

The research and development work has been distributed between theBAM, Berlin, and the Siemens AG. Srlargen, in cooperation with the ReactorbauForschungs- und Baugesellschaft - Kerntecbnik im Bauwesen -, Vienna. Theprogramme has been intendedJy kept open for participation of further industrialcompanies and institutions. The partitioning of the work has been done in away that closed problem complexes — which may be characterized by the physicalprinciple of the roe?surement, or by particular measuring conditions, or bythe vessel components on which measurements are to be performed — are to bestudied at one place.

*>Ttíe coordination of work in the instrumentation assessment programmein the..first line.is adapted to -the requirements of all participants? in theGerman programme,of research and development, for prestressed concrete reactor' <pressure vessels in 'a flexible way^ Results are to be made available to a!3

* ". • l -participante .as soon as they are; obtained-. -.'-JSeetings shall serve for mutualinformation"ahd,7lnspiration. The coordination is intended to be extended to• • * , ' • - , <• > -,* -, .•• f f. .-^the exchange of ,experí en ce with other, also foreign ins titutiom?.' which -'are

' , /' . Vt * '•* *' -* %

engaged in -j>ej*forming. measurements /on psreartressed concrete vees.el sttjifituresor in concrete studies» . . ' ' ' , * • ' * '

153

l i l i l í .

XlIXJLXW///////////////////S,

. B - 1. í'ultilayer wire winding prestreBsJns? systems \vithoutand with spacers.

154

j?ig. B - 2. Circular and polygonal wire winding prestressing system

155

PRISMA

B - 3. Schematic of the orestressing system"PRISMA"

156

Fig. B - 4-. Principle of the clamping anchor'.ge,

157

Spannstahl

Betonauftenkante

—~**Breite U—

B - 5. Schematic of the anchorage ledge1 t>recsure riistribution plate,2 base led-:e»5 clamping ledges,6 pressure ledge,7 clamping bolt.

158

3?if?. 5 - 6 . View of anchorage ledge

159

'7. V'ifi.v of anchorage leci^o ( construction eráronle20 loyeirn of tensj or.inq: "bands)

160

taM

i Betonaufienkantu

Pig. B - 8, Schematic of diverting ledge.1 pressure distribution plate2 base ledge3 intermediate ledges4 spacer bolt

161

Grundril) . Abwicklung

2 2

Fig. 3 - 9 . Schematic of spiral v/ire winding,

1 anchorage ledge2 diverting ledge3 clamping "bolt

162

Abwicktung

2 2 2

Soannhánder /Art fang

Fig. B - 10. Schematic of step wire xvinding,

163

1+2. Lage

_l

B - 11. Schematic of the step change,

164

B - 12. pressing tube joint for prestressing steel of type' SIGMA OVAL 40 St 145/160.

165

rig. B - 13. Anchorage at the start,

266

f\•4-'

Big. B - 14. Anchorage at the end (safeguarding anchorage)

167

B - 15» T o f if i of th'^an AGH

en applied toconcreto rrensur'"- ver.rol o p type

168

Jaeger Fig, D-.l

t-«CDCO

1969

ulsnniM

1970

of tests • Irn i

vrsUataary teats |i

| *wl*> of tMtlaK MebiBs I

I ———————

•

1971

f construction of teatiac MOUa» !

• i

j

1 Ht renort I

Itrial of testing•achín»

I

1972

remrlng *ad enríate <X concrete epeclama

1973 ;

JN \ .,. —.,....->--

tMtst eoapresaiv* strength In 2 «nd ? <UMIUKtOBS|

-testal compreasive atrensth to Zend 3 IdiBsnslona. temerature UD to 1K>°C 1' i ' -1-- ^^1 tests: coablnai oo»pr«Mlv» sad t«nsil« . |

| strsn^th. acwMl tenueratur» |

;

•2nd iwpwtj , ; •

.- I

ttatsi oastolnsd eostpressiT» and tmsiielstr*W[ttu tonenrtur» uo'to 190°C 1

- '

i 11 1 UuS. report

-

jaeger Fig. D-2

Position

BASICS

PBÍPUUTIOM3

PBS TCSTING

TBSTISO

w

SUPPLEMENTALTBSTIHO

1

?

3

456

78

910li12

1314

1516

Study of literature on the subjects concerned}Analysis of the existing material}Design and construction or acquisition of machinesand measuring devices*First report

Construction of test specimens*Testing of measuring devices under differentconditions *Second report

Construction of certain test specimens tExperimental testing} linear and areal surface/intern.measuring while applying boundary conditions;Third report

Completion of measuring devices and machines*Construction of certain test specimens}Experimental testing} linear and areal measuring to explainthe internal characteristics of moisture and change of thesecharacteristics under different conditions}Analysis of resultsFinal report

fheoretioal investigation for application of results. tolarge* volume concrete elements ttesting of a large-volume test specimen under normalconditions and environment}

69 197<

CXXX

ETTT

xxx

«T«l

xxxx

X3

5

EXX

CXX

nrxran

H*1

•

ntTIT

XX

1

X

j** }6

xxxxXXHI

Jaeger Fig. D-3

1.1 . General literature studies concerning related measurements

2,. Precise formulation of the various work tasks

3. Clarification of general problems concerning the applicationof electrical and fluíate measuring methods4. preliminary choice of the electric and flnidlo measuring prin-ciples, inclusive of the required circuitry

5. Preliminary studies of the measuring principles according to* paragraph 4, with special consideration to the gauge behaviourunder nuclear radiation and temperature influences during longtime-periods, eventually using commercially available gauges6. Selection of the transforming principles

?_. Preliminary selection of the coupling and transforming device»

8. Preliminary studies of the coupling and transforming devicesaccording to paragraph 7, under special consideration ofminimising the reactions on the object of measurement

10. Preliminary selection, evaluation, selection and laboratorytesting of moisture protection possibilities for gauges

11. Selection, evaluation, improvement or new development, andlaboratory testing of measuring gauges for the variousrequirements13. Selection of indication, recording, and evaluation techniques

13. proof testing of the measuring gauges and recording devicesin model test» and with the first suitable full-sice* prestressedconcrete reactor pressure vessels.

10. 11970

.1» 1-4» 1,7. 1*10. 1i t

i ' "

A» 1

*

ISA* 1

m.7* *00. 1•*• u

i*. i.

,

mr. i.10. 1

•

19a. i.

•

&». 1.1

V

Pig. C-2

TET10 TET 20

TEC 10

Fig*C-l

TEC 20

»1) Knotenpunkte

(m*J)Knottnpunktt

<Ut)Knotcnpunkte

GmamtzaM dtr Knol«npunfctt

Pig. C-3172

TRIAX 3 TRIAX6 TRIAX 10

Pig. C-4

TRIAX 6 TRIAX 10

Pig. C-5

173

FLAX 2L

FLAX 3

Pig. C-6

FLAX

SECT 9 SECT 16 SECT 30

. C-7

174

ore

Pig. C-10

Loading:Internal Pressure p= WOO Ibf/in2

and Equivalent End Loads Ogeto End Caos

C-ll

175

-J05 J__ j_ __ __ Syirnigoj»»*^ _ __

•ft—1500 mi» —»j

1% rxiAXiO-Elm*ni*17 Pi AX « -Eltmml* fur

vfc» • Ols

v__ s Oí

Mtn«wi>«» Vor jp»«n»t»t| S7 > • Itf upCtumtt Ridailiatt <Jt- Voroamtkabtl m 39J.io'v(i

Normalspannung inUmfangsrtchtung « [kp/mm*]

Fig. C-8 Fig. C-9

Seal»

c-12

ScaleStress r- >

0 WOOD 20WO tbf/in

Pig. 013

177

DESIGN PHILOSOPHY AND CRITERIA OP SAFETYFOR

PBBSTRB3SBD CÓNCHETE HÍESSURE VESSELSPRACTICE IN CZECHOSLOVAKIA

1C. David

saiu conclusions and resul t-s of abcut thr&e year's worfc.in Czechoslovakia. 5n development of PCPV's are described withpartícula" reference to ^he design philosophy and criteria ofsafety, Uevelopraent was made for gas-cooled reactors of «>OOííW{e)vrith beat excnan/yers placed inside tke PCPV's. The moat difficultproblem of the structural design was to find an appropriate methodfor solving requirenjents of the fine mesh of penetrations, together

n

with extremely high interna} working pressure (80 kg/cm ). Varioussmalj scale models of elastic materials in the scale of about 1 s 40•were investigated and u.ethods of measurement of strains and photo~elasticity were employed. The programme which is proposed for theshort taodel is described an I the estimation of cost of PCPV's iediscussed in general terms.

179

1.0.0. INTRODUCTION

The development of Pr-. treased Concrete PraawiraVessels (PCPV) in Czechoslovakia has heon in progresasince the year 19 7 3» a part of the coroplate task :* Nuclear Power Staticm A« % PCPVs are required for •two heavy water , gaa-coolod í"íactor3 the output ofeach tteing 500 M*(e)* The "integrated solution*1 wasproposed <md the reector «a well as beat exchangersshould be placed inside the vassal»

The adoption of heavy wster, gaa cooled roactorhas formed very unfavorabJc conditions for tha structu-ral design of the vassel* The 001*0 i& vary compact and.comparatively very small internal diameter of the vesselio required* On tbe other hand the internal pressure isextremely high* The very fino mesh of penetration forthe refueling must be designed in the upper slab» Therequirement of the fine «ash of penetrations togetherwith extremely high internal working pressure has createdthe most difficult problem of the structural design»

The development of PGPV started a'n Chechoslovakiain the time, when PCPV has already been successfullyused in France and in England, but the very specific

180

fissign perimeters» cited above, makes the Czechoslovakvessel little different from foreign designos* Thelatest had to be taken into account and spocial develop-ment program wna designed»

The main, conclusion and result» of about threeyears'work on davslopaoat of POFV will be described inthe report with particular reference to the design philo-sophy and criteria of safety*

2»0«0» STRUCTURAL ggSIQ»,

Por better understanding of the designed criteriaof safety and methods of investigation the structuraldesign will be breifly discussed at first (Ref«fig« 1;2; 3i 4| ).

With the reapeot to the required design parametersthe vessel of cylindrical shape was adopted* The heatexchangers are: situated under the core* This, layoutseems to be better than the solution with heat exchangareplaced around the cor«, which would lead to the substan-tial increas of internal diameter.

The structural design is strongly influenced by thef.way of refuelling* From different alternations of refuelling

183

*fche solution vnth fuel-handling machine placed outsidethe vessel was chosen.!?or o/sch channel in the core thespecial penetration through the upper slab had to bedesigned. Such a pari'orated lid together with highinternal pressur* has not yet been used neither inCzechoslovakia nor abroid and therafor the. investigationof this part of the vessel was designed very carfully inoder to satisfy tho prescribed safety and integrity of thevassal.

After assessment of many «dternativsa of structuralshapes and progressing systems the vessel prestressed byBBHV tendons was soleoted. Hoop tendons are 'anchored inthree vertical buttresses *

The general data of proposed PCFV are ;2Working pressure ...«.,*..,...... 80 kp/cm

Internal diameter *».**«.,....«<» 11,0 mThickness of the «rail ..«.,,«.». 9,90 mInternal height «»..,.**......*• 31,0 mThicknesa of the slab .......... 6,00 aIt should be pointed out that the vessel is very

" thick wallad" when compared witn other PCPVs.

182

This fact L-mi b* fv'ih'yf,*rs fv/.r» < h i » tabl* 1,

the retio W1.h1c1íív:*í« r»f -H-*a y/alV «?r\o ihickaen» of th«to th« intarna) fJ ia iQf» te < »- s i a J^-ví^fíCite»! for diffarent

D « , » . « » » & . , » . .... intí' r.vni ''. i «P>ÍJ t.«?x»

h » , . . . « . , « . . . „ , • > . . , » « *:hicjtn-»gs of the a.1

of IKo

POP? <1/D h/B

Old bury 0,19 0,23

Bugey X 0,32 0,44

¿tanganes» B

A3 (C'?ochoslov«kia) C^n.^ o,73

183

The theoretic! «nd particularly experimental

stress analysis become more difficult for such tickwalledstructure?* Th?. properties of mod flj.s' mat trials as the"Poiseon ratio, conditions of iyotrophy «itc* may influencethe results of sBodnXinvestiginion» of thic»kwallea vesselsmore substantially th«n of Vi»s801s w i t h sodium thickness

of the» vral'):,

3.0»0* DBSTGN FiilLGSOrilY AHD CHIggRIA. OF C

The design of PCPV in CzechosloveVJa aust satisfyall Czechoslovak t^cUnical codes and atand«ir<a« particularlythe " Directives for atructurss of preatr^asíí^ ooncrate H«All these standards were not f^signed tvith th« respect toPOP? and it was necgssary to form the special technicalcriteria for the designes of FC'PV»

The -following additional criteria of safety andrequirements has been adopted for the design of PCPV i

Working pressure »,,»,.,.. pwDesign pressure »„.»«.»•• pd' R 1,1 p wTesting pressv-r* ' pt » 1,15 pdultimate pressure pu » 2,5 * pdUltimate load factor 2,5

184

Th8 behaviour of tha vessel is required as follows:

1) Elastic behaviour of tho vessel must be ensured atnormal operational conditions and during the presauretest* All strasses of steel and concreta must bewithin allowable limita*

2) Increased working conditions ar* assumed temporarily:The pressure is increased for 20 %The temperature crossfall is inoreneed for 50 %

A

At these conditions it is allowed tp increaseallowable stresses and small but only recoverableoraoks may occur*

*' The minimal factor of safety of the vessel must beat least as prescribed and the mode of failure mustbe progressive*

4) PCPV should fail in a simple way which could bepredicated*

It mast be proved in the design, that all thesecriteria are satisfied» This can be dono by means of

»

theoretical methods*

185

Ttie most import antfeasuiapt i onsí&hould be supported byexperimental investigation of materials and completescale models»

THEORBTICAL STKfiSS i

ELASTIC.

The structures of PCPV are usually very thiokwalledand this must -be taken into account in stress-analysis*the computation is usually carried out by use of differentapproximative methods as Method of Finite Differencial,Finite Element Methods, «to» Th«» use of high-speed digitalcomputers anables very fine aoshing of the structure andthe accuracy of the results is satisfactorily,

Several computer programs for elastic stress-analysiswere compiled in ESTEBGOPROJSKT, Cssanho Slovakia.

The computer programa called Betaa - 1 and Betas 2are baaed on the Method of Dynamic Belaxation» The computerprograms Betas were designed for elastic stress analysisof cylindrical ?CPV. The effect of displacement on thestress of prestressed tendons may or may not be taken intoaccount in the computation^

186

% . .. . . » < . .. ,Further computar programs , compiled in

.».'"> . '•' • " • ' •' * '•' ;,.' ' '• «. .- ' : V>, C:./' >\ .-./•- ••:••. •Energopro5ekt ar« based on the Method oí FinitaV • • « . ' • « < • • - • * • ' :' : ! ... • , . . ; « . . • [ . . ' . . •

Elementa* The ooajputer programs using Pinit« Element• ' ' ' i • ' • • , • . • • • ' • " • " • ' ' • ' - .Method were ílaboratíá in ord«.r to check the results

obtained by the Method of Dynamic fíe.laxotioa and ea( • • . • • > . • . • • > . * • • ^a first stsp to thrse-dimsnsinBl sna ultiDafc* 3oad' ''-.'f • • ' • ' - . • - • . .. " -J.:i-analysis of the va^nel*

' ' :>.

The results obtained by the Method .of JDyno:aic! • '-. : ' ' . ' " « " \.

Relascati.ftn. were checked against th*?* obtained bythe Finite, Element M«^,hod for the oase of an cylinUri

' a ' •• ' • • .-...• • ', - . . ' • . •

¿el vessel* Both methods were in a very good agreement* , • • - • < • • . ' . -:¡ .'•• , , •' «,particularly in the cyiindrioal port of the vessel*

•' "•: Í -. ' •'»,•*Th« meshes for both methods are indicated on the fig«5and the comparison of some resulta- obtained by theMethod of Dynamic Relaxation and finita Element Method

f '• • ' :»••are presented on the fig» 6 and.. 7»

•

• The theoretical analysis o'f PCP7 must prove tha.tthe integrity of the vessel «nd tha criteria of safetyaré satisfied for all combinations of loading and during

187

*the whole life of the vessel* The stress distributionis strongly influenced by concrete creep and shrinkage,by steel relaxation by the loading history «tc. Thisproblem is encountered in all prestressed concrete struo»tures but in the case of PCFV it is more complicated dueto the influence of elevated temperatures*

Por an approximate calculation of crefp effects weusually use the computer program Betas 2. Tn the computa-tion the so called effective Young»' «Modulus is assumed*The effect of displacements on the state of stress ofprestredsed tendons is included in these calculationsand the lossea, respectivly the gains of tendon forcescan be obtained in this manner» The accuracy of resultsis comparatively good when :

fc - creep strains¿V - elastic strains

By the above mentioned method for creep-analysis onlyapproximate final values of stresses and displacementscan be obtained*

188

The spoeiol theory anr] corresponding computerprogram was derived "in cooperation with- SwedishState Pwjr Board for the calculation of the full

stress-history of tih» vessel» This work was dejsorlbadin Nuclear Engineering ana Design ( 9/1969 » 439-448)and only mnin assumption of tha method will be d'soussed

here»The full solution is obtained in a step-by-step

process* The time is divided into short intervals duringwfeich the stresses, are suppojy-ad to remain constant»According to this assumption the creep strain incrementsare calculated.after each time interval and then theincompatibility is corrected by an elastic solution.The choice of the creep function *s well as of the lengthof all time intervals is free* The Method of Dynamic Re-laxation is used for the elastic solution at any stage»The corresponding computer program is called Betas-3.

For the specific strain the following formula is usedin the computer program :

f/*(*-

189

where apocifio «tr«nrco'!u3us of alasticity as &function of tinm end temperatureconstant depending on time andtemperature

*ige 8*. loading ' '

.tiros at which ptrains ere consideredr.

T ~

The total creep and elastic strains are expressed inintegral form ;

Í* L

f C - r A'.*» ' ' <l

where i

? -

/-// ' '

190

The first integral in aq» (3) denotes the elasticstrain, while the second integral expresses the creepstrain»

In the computer program the sura of incrementa duringshort time intervals is used instead of integrals»

Several computations has been already successfullycarried oat by the program Betas 3* However, tho utilizationof this computer program is dependent on the basic data ofconcrete properties» These data are limited to some simpleloading histories» Much more complicated bwhaviour of theactual structure is met if the temperature of the concreteis pre-assumed to vary with tame» For such oases suitableexpressions for creep are not ye't available, but we believethat additional extension of the computer program Betas 3will be possible after further progress in this field»

4.3.CV gBMPEHATURB STBBSSES,

The temperature stresses- are influenced by the heatinghistory of ths vessel» The increas of temperature crcssfallresults in an increase of the temperature stresses» TheseStresses are at the same time reduced by the creep ofconcrete» The temparatur* stress analysis should be carried

191

out by the step-by-step procesa as described in thepast aoction* Such 9 calculation would be rather doubt-full as th0 knowledge of concrete properties on theconditions of transient temperature is limited at presenttime»

From above mentioned reasons the approximate solutionfor temperature stress analysis ij used in Czechoslovakia*The calculations were carried out for steady temperatureconditions with effective modulus of elasticity. The valueof effective modulus of elasticity was determinated accordingto the time of heating and the velocity of creep.lt isassumed in recent calculations :

elastic

It should be pointed out that th$ stresses calculatedas indicated above are different .from the actual temperaturestresses* Despite this fact the calculated stresses areused in combinations with other loading cases in. order to'check the prescribed criteria of safety,

192

4; «i.0,THE HBAL,STHfeM'lM OK Cy

It is well known that, the veal si rang th oí"

concrete depend» in triaxial state of stress on all

components of stress.The theoretical s tüdifs* ¡3f re&i st"efifitf of

concrete with the respect to PCr'V bw.s b«8en carried outin Caechoalex'safcia. the «!•••* :H.Í a te\á r?«aory is bat»od on

"Surface of reel strength of concrete" proposed byH.fíeimanri, The developed method tíu«bles to judge th«

stresses in the concrete and to prodlet the mode offailure of concrete, the diruction of first * cracks

etc»

On the basis o.f tfte SD*:<*tioued theory the computer-program was complied, Tbis computer program can be

joined to&ether with the progc¿ia for el&stio stressanalysis and by this menne*- a local margin of safety

cea be obtained at any point of the veasel.Further atore it is proposed to elaborate the coa».»

pater program for ultimate losó analysis »s a combina uof step-by-step elastic analysis with incrementalof iateraal pressure and the determination of cracksth« concrete on the bases of the surface of realstrength.

193

4.5.0 ULTIMATE LOAS 4HAL£?JS

It is necessary to prove in *U»e design not only

aatisfactarily behaviour of th% vessel during its lifebut to determine slae tho ujtiokato load factore of the

vessel * Further more H is required: the failure shouldbe very slow and vhe corning as cracks eee, shouldprecede the total fa 11 aro» In these considerations

«re suppose that the failure of the vessel is such astate of the vessel, whe» the further operation is not

possible. In spite of tnis t.hu failure is ©3 ready metwhen the leakage ox- exceetiive deformations ha& occured,

dispHt; that the structure hat not yet reached itsBtruetnral feilur"*<

There are sc-ver».! cauüCí>, v?í ich could i'esult in

1 ) Tine deterioration of the concrete2) The deterioration of the ^rostrcaJ) Error? in the aaMmation of Icug- terw behaviour

of the atructar»*4) The effect of temperature5) Excessive increas of intcrnel

In fact the excessive increpa or Internal pressurei u pr&ci-ioally not possible becouae ths eraergcnoy tech-noioglcAU equipment is all way a uaed for the reductionof the internal pressure, in the emergfciiey eaaa» Butwe e^P'june, that 5r the jaoa^ of failure is» progressive

194

for the case of excessive inereas of internal pressure,than the modes of failure for any other causes, citedabove, are also progressiv» The adoption of this assump-tion reduces the possible causea of failureto only oneof them-exceesive increase of internal presaure-and onlythis case is investigated.

Further more we assume that the mode of failureis progressive when the ultimate load factor is higheunough, however this assumption has not be neceasarillyvalid for any structure OP any part of the structure.

To predict the mode of failure for the flat, per-forated lids may be very difficult» Several modes offailure may be predicted and each of them could resultin a different ultimate load factor. The model investi-gations are neccessary in these cases»

According to the criteria of safety cited in thesection 3.0»0. it is required that the vessel shouldfail in a simple way which could be predicted. In prac-tice we have to design the vessel in that manner thatthe failure should occure in a cylindrical part of thevessel and not in the upper or bottom slab • It meansin other words, that the safety of the lids should behigher than of the cylindrical part of the vessel» Thisrequirement can he ensured by earful scale model investi-gation of perforated slab and the complete vessel* thescale model investigations are described in the section,

195

5.0.0 of this report*Only very simple ultímale load aualysas of the

vessel has been carried out in Czechoslovakia up todate. According to above mentioned philosophy, thefailure is assumed in the cylindrical part of thevessel •

The failure of longitudinal or hoop tendons canoccure* It has been assumed in recent calculations,

* », i

that just before the failure all internal forces are

taken up by p re a trussed tendons <.$nd mild reinforcementonly, because the concrete is already cracked»

More sophisticated ultimate load analysis is pro-posed for the future be step-by-atep process with pro-gressive increas of internal pressure*

5 • Q i.« 0 SCALfj! MQDBfr INVESTIGATIONS

.1«0 SMALL SGÁLE

Different small scale models from elastic materialsin the scale about 1:40 were investigate* in Caechoslo-vakia at the beginning v,f the development of PCPV. TheMethod of measurement of strains and photoelasticityMethod were employed» The aim of these models was tofind, the best shape of the vessel , to control thestresses against the calculations and to find oat somelocal stress concentrations e round the large penetra-tions etc*

196

knowledge which can be gained from theaemodela ia limited. This is particularly valid forvary tfcickwalled atructares where the experimentalreaulta ara influenced by different valúa of Poiaaonsratio or othar properties much mora substantiallythan in other atructuraa. Also the moda of failurecannot be represented -when the nod el i a made of thematerial different from actual concrete»

The Poiason ratio of different teating materialsia usually vary .high* The control by mean» of theore-tical methods becomes difficult aa the convergence ofthe interratiofliaethorta ia very alow for high Poiaaonratio.5T2,0 CONCRETE SCALE MODELS

Prom above listed reasons and with the reapectto very thickwalled vessel the concrete scale modelaare proposed in CzechoslovakiaTHE FlflST CÓNCHETE S&ALE MODELwaa proposed for'the investigation of elaatie aa wallaa ultimate load behaviour of the perforated lid. Fortheae raaaona was special model designed, callad shortmodel of the veaael (Ref.fig. 8j9)«

Following main assumptions ware adopted in thedesign:

197

1) The slab ia in the 8cala 1:5, but the length iamuch shorter*

2) The actual concrete ia used for the erection of themodel*

3) The relationship in bevsreen the aiee of agreggateaof coacret and the space in between the penetrationsshould be on the model the earns as on the actualvessel,

4) The assumption 2 and 3 resulta in the conclusionthat the penetration must be designed in theecale lil.The prestresting of the short nodel ia provided

by longitudinal tendone and by flat jacks in radialdirection. Flat jacks are supported by ring of rein»forced concrete»

The átate of stress in the lid waa checkd for theshort model and for the full-length vessel» The resultadoes not differ very much. Some of the stresses arepresented on the figure*10 and 11*

Following testa are proposed for the short model t1) cold tents

- different loading combinationa- creep test« test with pressure increased for 15 %- loading cycles .

198

2) hot testa- different loading combinations- cycles of pressure for different temperaturecrossfall

- cycles of temperature- creep tests

3) ultimate load teatThe erection of the model has been started at the

end of the last year.Before the realization of the actual vessel the

investigation of the seconed concrete scale model isproposed» The seconed model is the complete/model ofthe vessel* This model should prove the elastic andultimate load behaviour of the complete vessel* .

6«0»0 TESTS 07 MATERIALSt ' . - C»

. i .The materials, .as concrete and steel are in PCPVsubjected to a very .unusual conditions, to long termincreased,temperature anc irradiation. The materialtests .has, to be .carried out in order to have eunoughinformations about the behaviour of concrete and steelat these conditions»

The knowledge of concrete properties under ele-vated temperatures ie limited and therfore the pro-tection of the concret against the high temperaturemust be provided* The thermal insulation on the inner'

199

face together with the cooling system should unsurethat the temperature of the concrete will not exceed70°C.

The teste of creep of concrete under elevatedtemperatures has been in progrese in Czechoslovakia.for about 4 months» The specimens are tested at 25°C,40°C and 70°C.

The irradiation tests of concrete and steel areprepared for the near future.

7fO»0 THgESTIMATION OP THE COST OF PCPYThe total cost of the whole Power Station must

be compared for the determination of benefit which canbe gaind by utilisation of PCPV. The PCPV enables thecompact and economic integral design. In this mannerPCPV can contribute towards more economic design ofthe whole Power Station» For thai-more some other advan-tages be gaind when PCPV is used as higher safety etc.

' According to the esftimatioaea carried out inCzechoclovakia for Nuclear Power Station with two

»

heavy-water, gas cooled reactors 500 ISW(e) each, theutilization of PCPV reduced the total cost of the PowerStation to about 82 %.

200

8*0.0 CONCLUSION

The design philosophy and the proposed criteriaof safety as well as the structural desi/m itself hasto be usually'supported by o model investigations.

The amount of models and particularly big concretemodels which cnn be tested is very limited with therespect to very high costs of these tests. From thatpoint of view we suggest this exchange of experienceas a very fruitful1.

Further permanent cooperation in this field shoultbe provided in the frames of IAEA for the future.

201

Fitf. F-F

"•"vrv

i ' //;//1 1 I I / / / /

hi-

lli ¡ "" ¡ni j iiiiiiüiiiiünii

_ ,__ _ 11 po

H-;• v - —i-tt--iJ.'/r;;! :

Ui:;Vx\

\ \ \ \ M

i ' ll i l ' i l i I l11, l. i t .

-"**"

,.'!>$Dj i i &p*^rrr-f^f1 i i 1' i ; i '' ! i i

rpttn

202

s

usucJ7T

irj—u

E

r

-1490_?B

Fie'. 2. A-A

203

204

C-C

205

£

ki

-5

Sí

206

Áp/trn*+ 120

\-8D

-20

-4ü

-60

-80

-100

too•0

Ftg. 7.

60

-to

50

\V ffff

-•*=$

._ Off.

— r£M

kp/cm

too00

•

Ftf.A Short model. SfC//'0/> •' 1 ' f

f-

tooCO

1

><->*

1 *l

\

\- ¿—

* ü K , «*«•

$/*& /f/xr

_i *nv,. j_!Tjr r!L7»rTrwi"x'

I'M -._ 2M...

Jb/> 2 - 2 '

*¿* 9* oré model.

"***'

\

•\ V

.*>

211

SO 40 ¿0 SO 100

6*

i

4r

Ir

t

t1

*

\

•

A

/y.O 20•________*________A________i

I I Stkp I cm*W ¿0 80 100 212

SWEDISH DEVELOPMENT ON PRESTRESSEDCÓNCHETE PRESSURE VESSELS POR WATER REACTORS

Shankar Menon

A project to design, build and test a model of a PCPVsuitable for water reactors is in progress. The project isbacked by the atomic energy organizations of Denmark, Finland,Norway and Sweden, as well as by utilities and the reactorindustry in Sweden. The model, which has dimensions 2 m internaldiameter and 4 m internal height, is characterised by its removabletop head and its insulation system. The model has been hydraulicallypressure tested at temperatures of 25°C and 70°C at pressures upto 95 bar. The tests included a 4<*week test at 70°C and 85 bar(design pressure). The results have been very satisfactory.Insulation tests are due to start in March 1970.

213

Swedish Development Work on Frestressed ConcretePressure Vessels for Water Reactors

1. Introduction

At present, the Swedish nuclear power programme is based entirely onwater reactors. After the first two heavy water research reactors —Agesta, operational in 1964 and Marviken which is due to go on linein 1971 — the commercial power programme consists of light waterreactors» This programme is quite considerable and includes 4 BWRs(totalling 2360 MWe) and a PWR of 830 MWe under construction or onorder»

When prestressed concrete pressure vessels (PCFVs) began to be adopteduniversally for gas-cooled reactors, a study [YJ organised jointly byAB Atomenergi, the Swedish State Power Board and ASEA (now ASEA-ATOM)showed that such vessels were feasible for water reactors. The studyalso indicated that the advantages of safety and economy that hadbrought about their use with GCRs would also be present in water reac-tor PCPVs.

A comprehensive development programme development was lauched to studythe problems that are specifically associated with the use of PCPVsfor water reactors. The main areas of this development work were theproblems of access and thermal insulation. This development programmehas resulted in:

i) a design for a removable lid that allows access for refuelling(in the case of light water reactors) or service (for light andheavy water reactors, specially integrated designs)

2.< a thermal insulation system that permits the use of insulationsdeveloped mainly for gas-cooled reactors PCPVs and

3; an improved calculation programme for thickwalled pressure vessels

214

A joint initiative has been taken by the four Scandinavian countriesto design, build and test a 2-m (inner diameter) model of a water re-actor PCPV designed on the above lines. The aims of the project areto test and demonstrate the designs of the removable lid and the in-sulation system, to verify the calculation prograrañes, and to demon-strate the safety characteristics of the PCPV. The model has been builtand has undergone pressure test at room temperature and at 70 C withgood results.

As the design philosophy is based on the ASME Criteria for ConcreteVessels which has been reported on elsewhere, this paper containsmainly an account of the special developments mentioned above, followedby a description of the model and the experimental results obtained sofar* Much of the material in this paper was presented at the First In-ternational Conference on Pressure Vessel Technology at Delft* 1969jjfj •

2. Preliminary Studies

The technology of PCPVs for gas-cooled reactors has been extensivelyreported elsewhere [3]* The following account is limited, therefore ,to the special features of PCPVs for water reactors.

2.1

St.ee 1 pressure vessels for light water reactors are provided withflanged lids» Such reactors are refuelled during shutdown conditionswith the lid removed. One basic requirement, therefore, for a PCPVfor water reactors is a detachable closure, which, when removed,allows access to the entire core of the reactor. By this means, water-reactor refuelling technology could be utilized without change. Oneimportant consideration in the design of the closure mechanisms is thatthe removal and refitting of the closure should be rapid in operationin order to minimize the downtime of the reactor. Another important consideration is that the basic redundance of the PCPV design, a signifi-cant advantage from the safety point of view, is not compromised in thedesign of the lid.

215

A design satisfying the requirements listed above is shown in Fig* 1.The load from the internal pressure acting on the bottom of the lid istransmitted to the cylindrical part of the vessel by a large number ofstruts. Each strut is fitted with a wedge to minimize tolerance problems*The concentrated forces at both ends of the struts are distributed intothe concrete by means of heavy-section cast-steel rings. The steel ringat the top of tba cylindrical part of tt 2 vessel is reinforced by webslocated directly opposite the points where the forces from the strutsact. These forces from the struts on the ring are taken up by the verticalprestressing cables that are anchored in the cast ring itself and thetangential prestressing cables at the top of the concrete cylinder.

Sealing is effected by two toroidal seals in series located near the top ofthe lid. The rings have bolted flanges and are held at temperatures under70°C, which allows the use of rubber seals under the flanges. The highlocation of the seals also means that the internal pressure in the vesselprovides the prestress required for the lid. The design pressure for boththese toroidal seals is the design pressure for the reactor. The space inbetween the seals is monitored for leakage. This space also contains amissile shield that protects the secondary toroid from damage if and whenthe primary seal should fail.

Earlier designs of removable lids have generally been characterised eitherby large deflections of the toroidal seals or (in designs where such sealsare avoided) by large outage periods for lid removal and refitting. Thepresent design has tried to avoid both these difficulties.

Studies have been made to estimate the downtime for removal and refittingof the lid for a 6-m inner diameter pressure vessel. The time for removalis 12 h, while that for refitting is 16 h. As the PCPV is insulated inter-nally, no time need be allowed for the lid details to cool to managabletemperatures before the removal operation starts. Thus, the downtime forlid removal and refitting compares very favorably with corresponding steelpressure vessel times. The weights of the main items handled are:1) lid: 325 tons2) strut: 500 kg3) wedge: 50 kg

216

In regard to the safety characteristics of the design» ample redundanceis achieved by the large number of cables and struts that take up theload from the internal pressure on the lid. On a 6-m diameter pressurevessel with 85-bar design pressure, there are 120 vertical cables (eachof 450 tons nominal size) and 40 struts. A local failure of any of thesecomponents has no significant effect on the safety of the vessel. In thecast-steel ring at the top of tae cylindrical part of the vessel, thereare transverse ribs located directly opposite the points of action ofthe strut forces, and thus the risks of crack propagation are minimal.The location of the ring behind several meters of concrete is also suchthat it is not exposed to any significant nuclear radiation. The ring isanchored to the concrete both by the vertical cables as well as by exten-sive bonded reinforcement. The type of steel used in the cast sectionsis a high-strength, austenitic-martensitic stainless steel» which has a

2 'yield strength of 55 to 60 kg/mm . The maximum stress in the castings at2the design pressure is about 10 kg/mm .

One basic principle in the dimensioning of the lid components is thatthey should not fail before the concrete in the barrel of the vessel.According to calculations, the barrel has a factor of safety of about3 while the lid and the bottom slab have factors of safety of about 4*6.

The walls of PCPV's for gas-cooled reactors are kept at acceptable tem-peratures ( < about 70°C at the inner face) by means of insulation appliedto the inner face of the steel 1ining of the concrete cavity. In orderto utilize the large amount of experience available for such insulation,-the use of the same kind of insulation has been adopted.

In a water reactor PCPV the insulation operates normally in a gas atmos-phere very similar to that in gas-cooled reactors. The gas is separatedfrom the steam and water in the reactor vessel by means of two casings,one suspended from the top and the other attached to the bottom slab.In earlier designs, the two casings overlapped at the water level in thereactor vessel to form a water seal. During normal operation, pressurebalance between the gas and the reactor atmospheres is obtained automatic-ally by changes in level in the water seal. The surface of the water seal

217

is cooled in order to minimize condensation in the insulation* The in-". . .'sulation itself can be of the stainless-steel, foil-mesh types developedfor gas-cooled reactors in England and France. Satisfactory small-scaletests on this system have been carried out on both British and Frenchinsulation types at the Studsvik Research Establishment of AB Atomenergiand have been reported elsewhere [4, 5].

Very briefly, in these tests the thermal properties of the insulationhave been studied in dry N-, in N. over a water surface at temperaturesup to saturation, and finally with the insulation water-filled* The highesttest pressure was 80 bar.Recent developments have been1) The water seal has been shifted to the bottom of the vessel*2} The insulation, instead of being applied to the lining of the pressure

vessel, is fitted on the casing that separates the gas from the steam-water in the reactor vessel (Fig* 2).

These developments, in conjunction with that of the removable lid, makeit possible to remove the casing complete with insulation from the reactorvessel for inspection and maintenance. In addition, the lining is alsoaccessible for inspection.

The adoption of a completely water-filled insulation on the casing has alsobeen studied» The principles of this design are shown in Fig. 3. Ourstudies have indicated that completely water-filled insulations are thebest solution for light water reactors. The gas-gap insulation is appli-cable to heavy water systems.

The horizontal surface of the bottom slab is insulated by material similarto the type used on the casing. Here, however, the thermal insulationproblems are of a minor order compared to the cylindrical part and the lidof the vessel. Therefore, the insulation can have thicker foils and mesh,which are better as far as corrosion is concerned, although less effectivethermally* The penetrations of the bottom slab are double-walled with con-ventional fiber-glass insulation operating at atmospheric pressure.

218

2*3

In. the absence of an established code for the design of PCPVs, the designcriteria for the model have been based mainly on the ASME Criteria forConcrete Vessels [6J. Briefly,

1) The vessel is designed for elastic response to all possible combinationsof loads during operation.

2) It oust show a progressive mode of failure under increasing overpressure,i.e., large deflections to warn against impending failure.

3) The vessel must have an acceptable safety margin against failure.

In addition, because the vessel has a removable lid, an increased marginof safety is required for the locking details associated with the lid inorder to assure that failure would occur first in the cylindrical part ofthe vessel.

Calculations have been made on a vessel with an inner diameter of 6 m usingan axisymmetric programme based on the dynamic relaxation method [?] . Amongthe cases considered were the following:1) Dead Load (DL)2) DL + Prestress (PS) + Reference Pressure (RP, 85 bar) .3) DL + PS + RP + Temperature Load with creep taken into consideration (TL)4) DL + PS + TL5) Limit Condition 1, i.e., DL + PS + l.SRP + 1.5TL6) Failure Condition, i.e., DL * PS * 2.1RP (The safety factor 2.1

represents a factor of 2.55 when reckoned against the normal workingpressure of 70 bar)

The calculation programme can take into account varying compositions ofsteel and concrete in each coordinate direction in each mesh unit. Duringthe study of the approach to failure, when. the concrete tensile stress ina mesh unit reaches a certain value (to denote cracking), the load is trans-ferred entirely to the reinforcing steel in the unit. By making the stresscalculations at 1.5, 1.8 and 2.1 times, the reference pressure, the pro-gressivity of the cracking can be studied.

The validity of the lid design for very large reactor, units has been studiedby making calculations for a vessel with an inside diameter of 10.5 m anda reference pressure of 130 bar.

219

3. Scandinavian PCPV Model Project.-

<A project to design, build,1 and test a model of a FCPV suitable'for waterreactors has been started jointly by the four Scandinavian countries. Theproject is backed by the Atomic Energy organizations of Denmark, Finland,Norway and Sweden, as well as private and stat-owned power companies andthe manufacturing industry in Sweden. The designs of the renovable lid andthe insulation system will be tested, PCPV calculation programmes will be'verified, and the vessel will finally be pressurized to failure to demon-strate the safety characteristics of the PCPV. The project will also pro-vide Scandinavian engineers the means of acquiring practical experiencein the building of PCPVs.

The general practice in England and France has been to test jthe model PCPVand the thermal insulation separately. The Scandinavian model project isone of the first instances where the civil engineering and thermal insula-tion experiments will be performed on an "integrated" model.

3.1 The jnodel

The model is based mainly on the 6~m diameter vessel for a 750 MWE BUR(Boiling Water Reactor) but can also be considered as a model of a600 KWe BHWR (Boiling Heavy Water Reactor).

A vertical section through the model vessel is shown in Fig. 4. The vesselhas an inside diameter of 2 m and an inside height of 4 m. The referencepressure is 85 bar. The cylindrical wall thickness is 1.1 m and that of thebottom slab and lid is 1.2 nu

The bottom slab is perforated to represent control rod, pump, steam out-let and:other penetrations for a BWR. The concreting of the bottom slabhas been the subject of a special study both for the full-scale prototypeand the model vessels. The concrete in the model has a 28-day cylinder2strength of 450 kg/cm»

t iThe Birkenuaier, Brandestini, Ros and Vogt (BBRV) system is used for pre-streasing the model. The vertical cables are units of 50 tons nominalforce while horizontal prestressing is applied by 170 ton units anchoredin four vertical ribs. " : *'

220

The nominal cable forces in the tangential and axial directions areabout 1.5 and 1.75 times the R.P. load» respectively. In addition there

j iis bonded reinforcement that can take up 0.6 and 0.8 times the R.P. load,respectively, in the tangential and axial directions (at yield point insteel).

The lid is held in place by 40 struts» each fitted with a vedge for ad-justments. The model has a scaled-down toroidal seal that is bolted onto the lid and the top of the barrel of the vessel.

The inside of the vessel is lined by a stainless-steel-clad mild-steellining (7+3 mm). The bottom slab lining» however, is of stainless steel.Cooling pipes are welded on to the concrete side of the lining»

Insulation of the CAFL type is applied externally on' to the removablecasing inside the model vessel. The insulation thickness is 25 mm (with-out cover plates). The stainless-steel lining of the bottom slab is alsoinsulated by means of CAFL insulation. Here, however, the thickness ofthe insulation is 100 mm, and the foils and mesh are of thicker material.

The model can be pressurized by a system that can deliver either hot wateror saturated steam. There is a circulating system that draws out the coolerwater at the bottom of the model and pumps it back to spray nozzles in thesteam space, thereby ensuring even temperature conditions in the vessel.There are also systems for level control, gas supply, etc.

Figures 5 to 9 show the model during various stages of construction.

3.2 Instrumentatio&

The model is instrumented as follows:1) 250 vibrating wire strain gauges located in the concrete along two

planes of symmetry2) 170 resistance strain gauges on the lining, penetrations, reinforcement,

lid struts, toroid ring and steel castings3) 36 vibrating wire deformation gauges fitted on a frame mounted on a

plane of symmetry4) 24 cable dynamometers5) 120 thermocouples for concrete temperatures6) 170 thermocouples on the hot and cold faces of the insulation and on

the lining and penetrations

221

The auxiliary systems are suitably instrumented. All of the model andv *some of the system instrumentation has automatic registration.

3.3

The program is divided into three phases:1-) elastic tests I and II, 2) insulatitn tests, and 3) ultimate test.

jSlasjtic Jtests I

These have just been completed and the first results are given later inthis paper* During this phase, the function of the lid was checked underoperating condidtions. Measured values of strain, deflection and tempera-ture were compared with calculated ones. The tests started with a co dwater pressure test up to 95 bar. After about 2 weeks at 85 bar (designpressure), the tests were repeated with water at 70 C up to 92 bar»followed by a 28-day test at 85 bar and 70°C.

These are due to start in the beginning of March this year. The thermalproperties of the insulation will be studied under gas and water-filledconditions up to 85 bar (saturation temperatures at hot face). Tempera-ture and strain measurements in concrete will continue. Some insulationaccident conditions will be simulated.

Elastic tests II

The insulation and internals wjll be removed. The model will be taken tothe elastic limit. Time permitting, creep tests will be performed at 128 bar(1.5 HP) with 50°C water.

JJltimate_tesjt

This is to check the safety factor and to check that large deflections areobtained before failure. The model will be pressurized hydraulically. Thevarious pressures and temperatures during the tests mentioned above areshown diagrammatically in Fig. 10. The entire test programme is planned tobe completed by July 1970.

222

jPresjtres ingj, The 240 cables were prestressed in two stages. The measuredstrains showed good agreement with calculated values. Friction coeffi-cients of 0.12 and 0.14 were obtained for the vertical and horizontalcables respectively (the cables were greased).

oj:d_p£essjure jtests: The vessel ¿as taken to 95 bax (10 bar over designpressure) in 14-bar stages. The struts of the lid showed almost perfectload-sharing» with a spread of i 5%. The compressive stress in then was2about 10 kg/mm .locking details.

2about 10 kg/mm . Examination after the tests showed no marking on the

The toroidal seal showed fully elastic behavior. The maximum stress was2 2about 32 kg/mm in a material with an elastic limit of about 60 kg/mm .

Fig. 11 shows the radial deflection of the outer face of the vessel wallat design pressure, while figs. 12 - 14 show the tangential» radial andaxial strains at the mid-section of the vessel wall at the same pressure.

Hot £fes¿ure-tesjts_ajt ¿PC ijttiier face Jtemjpjgrature: The vessel was pres-surized to design pressure with cold water (at 25 C) and then the waterwas heated to 70 C in stages. Steady state temperature conditions werereached in the vessel wall after about 7 days. Figures IS and 16 showthe variation of strain with time at a few interesting points in thevessel wall and the lid.

A report on the test results is under preparation jY] .

Two major studies are in progress. Both are organised on a Scandinavianbasis, i.e. with Denmark, Norway and Finland taking part along withSweden. One of these studies concerns a 750 MWe boiling water reactorwhile the other one is on a 800 MWe boiling heavy water reactor.

The aims of the BWR study are

a) to investigate the requirements for a plant with a concrete pressurevessel so that satisfactory safety philosophies can be formulated forvarious reactor locations (including urban siting)

223

b) to project the pressure vessel and the plant in sufficient detailso as to be able to study how the concrete vessel affects the lay-out and design of the station

c) to update the cost comparison made in 1966-67 between BWR plantswith steel and concrete pressure vessels. This cost comparison willtake into account not only the dir ct vessel costs but also costsof containment and other station details that would be affected bythe choice of type of vessel.

The study is well advanced and is expected to be completed in springthis year.

The BHWR study is a technical and economical evaluation being carriedout by the Scandinavian countries in co-operation with the NorthernBechtel Corporation, U.S.A. This study is expected to be ready by Julythis year.

224

REFERENCES

[l] C Sundquist et al., "A Study of the Feasibility of Using.PrestressedConcrete Pressure Vessels for Hater Reactors11 7,7.1966.

[2] S Menon, Prestressed Concrete Pressure Vessels for Water Reactors,International Conference on Pressure Vessel Technology, Delft, 1969»

[3} Conference on Prestressed Concrete Pressure Vessels, Institution ofCivil Engineers, London, March 13-17, 1967.

[4] B Ringstad and A Skinstad, "Performance of Insulation for PrestressedConcrete Pressure Vessels for Boiling Heavy Water Reactors", NuclearEngineering 4 Design, vol. 1, 1967, pp 177-182. ;.•C.

[5] B Ringstad, B Alexander and A Skinstad, "Metal Foil Insulation forPrestressed Concrete Pressure Vessels for Water Reactors", ENEA

N'Symposium on the Technology of Integrated Primary Circuits f o> PowerReactors, Paris, May 20-22, 1968. . ' - . > -

[6] Prestressed Concrete Vessels for Nuclear Reactors, Draft, ASME,New York, November 1, 1967.

[T] T Tarandi, Computer Programmes and 'their Application for Tempera-ture and Stress Calculations of Reactor Pressure Vessels, Interna-tional Conference on Pressure Vessel Technology, Delft, 1969*

[ft] T Tarandi, "Pressure and Temperature tests on Scandinavian PCPVModel", AB Atomenergi, 1970 (under preparation).

225

!'

F<G. 1 REMOVABLE LID

226

LINING COOLINGSYSTEM

INSULATIONGASLINING

iTER OUTLET

Fig. 2 Thermal Insulation PrincipleGas gap insulation

227

LINING COOLINGSYSTEM

WATER (colLINING ——

^" CONCRETE/ WATER <sa-/ turated)y/, _ _

COOLING WÍTERINLETWATER (colfd)BAFFLEINSULATIONCOOLING WATER OUTLET

Fig. 3 Thermal Insulation PrincipleWater-filled insulation

228

Fig 4 MODEL VESSEL229

Fig. 5. Bottom síitb penetrations before conerc ' t in&

230

iFig. t> Bottom slab after concreting

231

Fig. 7. View of construction work after weldingof lining to bottom slab.

232

8. View of model vessel

233

Fig. 9. Lid closures details

234

toCO

Pressureü££

2»-

2CO-

wo-

ISO-

140 -

120-

100-$2-AS -»•0-

w-

40-

20-

0-

J •SL

-278»-J££?-20»-we»-C9»-TO*-«0»-SO*-30» 20»

70»

I 20»

tttar Pn»ssure

-j i j •• i , • | i | . | . . |6 0 10 12 14 16 It 20

/C

}r~T

22

W!»

M M

Itar

l_tt

ri\J

' i ' r r24 2t 2t

1 1 ' 130 32

£*,«, „.

«tbor

'

-380?-21Ef-2MF-2«?~W»•460*-1201

-90°-«O»-30»

Pressurebar-220

-208

-103

-«0

-140

-120

-100-92-ts-80

-60

-40

-20

-01 —— P 1 | 1 J 1 T |

34 36 3t 40 42 Weeks

Fig. Id. PRESSURE AND TEMPERATURE DIAGRAM FOREXPERIMENTAL PROGRAM

/,

H—H-0,5 ma

Fig. 11Cold pressure tesé: Outer facedeflections, P « 84 bar overpressure

236

e • 10s

400'

300-

200"

100"

k,VNit

III

I

Calculated (prof. Zienkiewicz)Calculated (Atomenerei)

^W

44

Measured at minimum section/ -"- maximum -"-

\ 1 ? E

I ^ ^ iIIIII

I•ka __

I " ÍI ÍI Ilob Distance from

inner face, cm

Fig. 12 Cold pressure test: Tangential strains ataid-section of vessel» P * 84 bar overpressure

237

10

200 -

100 -

CalculatedMeasured at adnirewa sectionMeasured at maximum section

ISO I100 Distance from

inner face, co

Fig. 13 Cold pressure test: Axial strains atmid-section of vessel» P - 84 bar overpressure

238

Measured at minimum sectionMeasured at maxiaum section

Distance frominner face» .cm

Fig. 14 Cold pressure test: Radial strains atodd-section of vessel, P « 84 bar overpressure

239

Steady state temperaturein vessel wall

Calculated values

Inner face temp. T. m 70 COuter face temp. TQ » 20°C

30 daysTime from start of hot test

Fig. 15 Measured strains in vessel wall during hot test

-400

-300

-200

-100

Inner face T. * 70 COuter face X * 20°CoO Calculated Value

dav30

Time front start ofhot test

-100

16 Measured «trains in vessel lidduring hot test

241

Annex 1

The,I.A.E.A. Panel on theBasic Structural Design Philosophy Criteria and Safety

of Concrete Reactor Pressure VesselsVienna.. .9..-. 1.3 February 197Q

Programme

Monday» 9 February 1970Session I t 9.30-12.00 a.m.

Opening of the Panel by Prof. I.S. Zheludev,Deputy Director General, Department ofTechnical Operations, IAEAStatement by the Chairman of the Panel,Mr. I. -DavidsonAdoption of the AgendaPaper by Kr. W. Eockenhauser "StructuralDesign Philosophy and Criteria for ConcreteReactor Vessels - U.S. Practice"Paper by Lír. L.H. Koraoli "PrestressedConcrete Vessel Containing the Heli» Loop"

Session 11$ 2.00 - 5.00 p.m.Paper by Mr. I. Davidson "Contribution fromthe United Kingdom"Paper by Mr. B, Petrovié "Work on ReactorPressure Vessels of Prestressed Concrete inYugoslavia"Paper by l£r. D. Costes "Problems and Prospectsof PreFtressed Concrete Pressure Vessels —French Experience up to 1970"Paper by l£r. T.A. Jager "Report on theStarting of a Coordinated Y-'ork Programme forPrestressed Concrete Reactor Pressure Vesselsin the Federal P.epublio of Germany"

(Questions and discussions to the papers presented will take placeimmediately after each paper has been presented)

Cocktail Party at 6.00 p.rn*

243

Tuesday, 10 February 1970

Session III: 9.00 - 12.00 a.m.

Session IV : 2.00 - 5.00 p.m.

Paper "by S. Ken.on "Swedish DevelopmentWork on Pros tressed Concrete PressureVessels for Water Reactors"Paper by Ur. M. David "Design Philosophyand Criteria of Safety of PrestressedConcrete Pressure Vessels - Practice i»Czechoslovakia"Paper by Mr. F. Sootto "Ultimate Design -Experience from Saall Dimension Models TestinStatements by other participants ifthey wish to «itcV-e them (representativesof international organisations andobservers)General discussions

Discussions to the problems of basicstructural design philosophy and to thecapital cost of the PCPV

Wednesday, 11 February 1970Session V : 9.00 - 12.00 a.m.

Draft of the Report on the subjectsdiscussed in Session IV

Afternoon Free

244

Thursday,, 12 FebruarySession VI : 9.00 - 12.00 a.m.

Session VII: 2.00 - 5.00 p.m.

Discussion on the safety problems ofPCPV (possible causes of failure forlor:{j~term operation and time consequences,inservice inspections, administrativerequirements etc.) and identification ofareas requiring further study in researchor experimental development

Draft of the Report on the subjectdiscussed in Session VI

Friday, 13 February 1970Session VII: 9.00 a.m. - 1.00 p.m.

Discussion of the conclusions andrecommendations and approval of theReport of the Panel Meeting-Closing of the Panel

245

Annex 2

LIST OF PARTICIPANTS

AUSTRIA

Mr. Ludwig Komoli

Mr. Hans Griiam

Mr. Gerhard Huber

Mr. Josef Kemeth

Mr. August Nesitka

BELGIUMMr. P. Charlier

CSSR

Mr. M. David

DENMARKMr. Arne Pedersen

FflANCBMr. D. Costes

Mr. René Bordet

Reaktorbau Forschungs- und BaugesellschaftA-2444 Seibersdorf, Reaktorzentrum

Oesterreichische Studiengesellschaft furAtonienergie

Lenaugasse 10A-1080 Vienna

Oesterreichische Studierigesallschaf 1 fxirAtomenergie

Lenaugasse 10A-1080 Vienna

Reaktorbau Porschungs- und BaugesellschaftA-2444 Seibersdorf

Prokurist der Bauunterriahraung Ed.Ast & Co.GraaBurgring 16

Ingénieur à la Belgonucléaire35> rue des ColoniesBruxelles I

State Design InstituteEnergoprojectBuben ska 1Prague 7

DAEC Research EstablishmentEiso

DEP/GT3P Centre d'Etudes Nucléaires de SaclayBoite Postale Ii'o. 291, Gif-sur-YvetteDivision Génie Civil du Service des Etudes

et Projets Thermiques et Nucléairesd'Klectricité de ï^

247

FEDERAL REPUBLIC OF GERMANT

Mr. T. Jaeger

Mr. P. Bremer

Mr. G. Hohnerlein

Mr. Peter Eisaler

ITALY

Mr. P. Scotto

KQir.VAY

Mr. E. Lenschow

SP.AIjjT

Mr. ]>.S. Korena

SVfEDEN

tîr. Shank ar Lien on

Mr. Emil Bachhofner

Bundssar.stalt fur MaterialprufungUnter den Eichen 8?Berlin-Dahlem

Fried. Krupp Universalbau GtenbHFrobrihauserstrasse 95PostTach 93243 Essen 1

Pried. Krupp KerntechnikZollstras'ôô 6843 Sssen i

Siemens AGEriangen

ENÊLState Electricity Generating Board

Technical University of NorwayTrondheim

Junta de Energia NuclearCiudad UniversitariaMadrid 3

AB AtomenergiP.O. Box 43041Stockholm 43Atoinlcraf tkonsortie tBox 174611187 Stockholm

248

SWITZERLANDMr. A. Rosli

Mr. Hans Hofoiann

Chef de la Section Béton armé etconstructions en béton du Laboratoire

' 'fédéral d'essai des Matériaux et derecherches pour l'industrie, la

' construction et les arts et métiers(E!£PA)CH-8600 BiïbendorfSUISELECTHABoite PostaleCH-4000 Baie 10

Mr. I. Davidson

Mr. J.-Bowen

Mr. J.P« Brovm

Dr. J.D. Brunton

Mr. H.B. Cochrane

Mr. D. Kc. D. Eadie

Mr. D.R. Fryer

Mr. A.J. Williams

Thermal ReactorsU.K.A.E.A.Sisley, iVarrington, Lancs

U.K.A.E.A.Risley, Wa'rringtort, LancsCentral Electricity Generating BoardLand House20 Newgate StreetLondon, E.G. 1Taylor V/oolrow Construction Ltd.345 Ruislip RoadSouthall, MiddlesexCentral Electricity Generating BoardWalden House24 Cathedral PlaceLondon, E.G. 4The Nuclear Power Group Ltd.KnutsfordCheshireMinistry of TechnologyInspectorate of Huclear InstallationsThames House SouthMillbankLondon, S.W.British Nuclear Design & Construction Ltd*Cambridge RoadWhetstoneLES 3LH Leicestershire

249

U.S.A.

Mr. W. Rockenhauser

.YUGOSLAVIA.

Mr. BoSko Petrovié

O.E.C.D.

Mr. A.H. Kirikead

EURATOM

l£r. J. Reyncn

Mr. E. Benaler

Hc?ctor DevelopmentP/ffi Systems DivioionV/ostinghouse Electric Corp.Box 355Pittsburgh, Pennsylvania 15230

Institut aa ispitivanje Materijala Sr.Srbije

Bul. Vojvode MiSi6a 43Belgrad

O.E.C.D. High Temperature Reactor ProjectA.E.B.j \VinfrithNr» DorchesterDorset

CCB Ispx-aCaselle Postale 1VareseItalyDirection Générale Recherche Générale etTechnologie

BruxellesBelgium

I ASAMr. S. Havel Scientific Secretary

Division of KucJear Power & Reactors

250