ML20234D338
| ML20234D338 | |
| Person / Time | |
|---|---|
| Site: | Palisades, Turkey Point, 05000000 |
| Issue date: | 09/20/1966 |
| From: | Dinunno J Advisory Committee on Reactor Safeguards |
| To: | Advisory Committee on Reactor Safeguards |
| Shared Package | |
| ML20234C970 | List:
|
| References | |
| FOIA-87-40 NUDOCS 8707070178 | |
| Download: ML20234D338 (31) | |
Text
_-
r*
k J. J. DiNunno-S;pt mb2r 20, 1966 r;
lOTES TAKEN AT THE ACRS-AEC-INDUSTRY MEETING STRUCTURAL BEHAVIOR OF PRESTRESSED CONCRETE FOR CONTAINMENT STRUCTURES AND REACTOR VESSELS t
September 19 and 20, 1966 1717 H Street, N.W., Room 1030 '
Washington, D.C.
Current State of Development and Applications - M. Bender Passed out several charts.
Summary of Design Data for European Prestressed Concrete Dual Function Pressure Vessel for Cas-Cooled Reactors.
Used in pipe - 1942.
In 1949 ductile behavior of prestressed pipe.
]
1 Marcoule, French, made one of the first applications for containment in 1958.
J Are still in operation and have demonstrated feasibility.
HWCTR. 24f/sq.in.
l Portion of the cylindrical shell.
I Iow pressure containment for reactors.
French. Fully prestressed.
Brenn111s, France. Now in United States -- Brookwood and: Turkey Point, Pa111sedes, and General Atomies for Colorado-project.
j Showed slide of EDF-4 containment.
Stressed the following points:
1 1.
Must maintain decay heat cooling..Cannot tolerate system l
breaks. Vessel failures not tolerable.
j l
2.
Temperature control. Must be maintained. Thermal stress causing l
break.
l 3.
Deterioration of the steel itself.
4.
Earthquakes and shock loads on the anchors - tr.ust be designed against.
Not concerned for small leaks - from a public safety concept from being able to keep cool.
l l
As a leaktight containment, 1-157. compared to conventional. These are the kinds of vessels we are talking about.
/
8707070[78870527 Of PDR FOIA THOMASB7-40 PDR
_~
s.
1,
Showed slide of a pressurized-water type concrete pressure vessel.
Pointed out the liner.
If it is okay, cracks in concrete not too severe.
Liner important for leakage.
Concrete for the pressure retention or prevention of large burst.
In a pressurized-water reactor, containment. required only for the major accident.
Have the problem of surveillance to have assurance that integrity is not deteriorating.
Reliability very important.
European Practices Main thing of importance -- how they look at fuel.
In United States we drive fuel hard.
If we get loss-of-coolant accident, get fission products out.
French and British believe fission products will not get out of the fuel, so they do not drive fuel as hard.- Temperatures not so high.
British.and French are concerned about cracking.
Limit thermal; stressing.
Not certain how they know.
Design liners to be under compression Like to say is everywhere.
If they have tension loads, they are quite low.
Big openings are usually reinforced with steel.
British do not design for seismic.
French - 0.1(Og for one reactor.
Separate base from foundation.
Neopreue pad.
Do have some problems.
High loads on foundations in some casas.
Europeans have been doing considerable model testing.
Both French l
and British rely heavily on strain gauges for nonitoring and are concerned with corrosion of cables.
French grouting.
British still uncertain.
Exemplifies the uncertainty.
Must be decided for United States also.
Both British and French are interested in the elastic behavior.
Do not know how to calculate other effects.
Are using computer codes.
British and French have not established limits on crack size.
Anchorages are important for prestressing.
Several kinds presented j
in slide:
)
Freyssinet sacket (Roebling)
EDF-3 and EDF-4 (Inactive Anchoring) (SEEE Anchor)
y l-
(q q r.
.i European applications.. lUnder prestressing.
Stress relieved under L
normal operation.. Design for uniaxial. loading, but have triaxial' loading.' Conservative approach.
Limit temperatures.to 1500F in concrete.
French Criteria:-
-1.2 x Nominal Design + 1.5. Nominal 5 train.
Take 2.5 design pressure without depressurization.
Can acandJ cracking, but cracks close.
Brenn111st 3 times - no cracking.
4 times.
no breaking.
- but 'were designing to only 8 psig.
Some of the concerns in the United States are:
1..
Can gross rupture'in a large penetration occur?
(Strain on penetrations mus t not violat'e the liner.)
2.
What happens to structure under static and dynamic loads?
Wha t about the' anchors?
i 3.
Corrosion mechanisms:in the prestressed steel - what sur-veillance can be done?
4 How does the liner react with concrete?
5.
What are failure criteria?
' Factors introducing strain:
l 1.
Loading.
2.
Temperature gradients.-
3.
Shrinkage and creep.
The effects are not necessarily additive.
'Showed slide of progressive crack growth.
Are quality control methods for conventional prestressed good enough?.
l 1
m 4
W, L.
s l
I
. REACTOR VESSELS - R. O. MARSH Have 'been working on the prestressed vessel as a primary and not
- a. secondary vessel.
r Showed slide of the Marcoule reactor vessel.
46' diameter x 51' length.
Two in operation.
Slide of W/LFA.
Thermal insulation inside the liner. ~ Liner is
~l water-cooled.
Cooling tubes 96' internal diameter.
HTGR - 30' ID x 60' 3)
Similar requirements for the GA design as European concepts, d
150 F average maximum temperature in - the concrete.
'To be discussed later.
1 a
(
g,Model Testing l
1300 1000 Visible Exterior Cracking 800 g
ist Regime I
Strain i
Slide - Basic PVRC Concept.
= 5% strain change from cycling - as constructed to pressurized.
Ultimat
\\
l Effects of P
A = Inelastic Life -
)L reduction n
of pcint on the clastic Strain curve 9
I
7 :.l -_.
'[ _ - _
_. u.,._
p Loads assuming uniaxial' loading may be one-balf'those under biaxial.
Have not-yet verified.
Creep data.
University of California.
Sone data coll'ected over a.
period of 30 years.. Have been steady-state.
Question was raised.
~ of effects of cycling.
Seemed to be. consensus that cycling effects could be considerably. different.
Most significant structural property for PVRC:
1.
Multiplying of load bearing tendons.
2.
Concrete is unloaded by service pressure.
3.
Tendon loads barely change under pressure loading.
Have been looking at their tendon designs from standpoint of corro-sion, creep, relaxation - inspectability desired to determine or test for deteriorating effects.
Are doing model tests - (2).
Includes cycling loads.
Have taken-
{
pressures. up to observe cracking in model fl.
In second model are t
.also adding thermal loads.
These are also doing cycling.
'Showed. phot'os (slides) of model tests.
Ultimate strain.
Went to 1930 psig, crack at 950 psig.- Got some strand breakage.
Did not unload tendons.
Have not looked at these as truly modes.
Are really checking calculational techniques.
Scale 1:4.
PERFORMANCE AND SERVICE EXPERIENCES - B. C. GERWICK Experience with prestressed concrete - buildings, bridges, etc."
50th Anniversary celebrated in Paris.
1920 - Few applications 1930 - U.S. in tanks 1950's - Large scale uses everywhere, including United States.
Wide variety of uses now:
1.
Bridges - precact, pre-tensioned.
Europe - longer spaa, cast in-place bridges.
- Rhine, Bendorf, 700' England.
Preshrunk. 1000' Arch bridge in Australia.
e..
6 --
j t
2.
Buildings.
United States and-Russia.
Precast and pre-1 tensioned.
Many -'are now precast and -post-tensioned.
Sidney Opera House.
-3.
Piling for bridges and foundations.
200' piles:in Lake i
. Maracat bo.
Dams in Arkansas River.
I
'4.
Machinery foundations.
England - rocket test stands.
United States - forging and hammers.
5.
Dams.
Anchoring into rock.
t 6.
Retaining bulkheads.
Tendons anchored in rock.
Ungrouted in important sectiors.
7.
Water tanks,. petroleum - (using liners).
8.
Railroad ties (many cycles).
9.
Runways for airports.
Particularly extensions over water-ways.
10.
Pipe-lines.
Also high pressure. pen stocks.
Pres tressin'g by post-tensioning.
l 11.
Chimneys, towers, transmission line poles.
Experience'shows a number of problem areas:
1.
Cyclic loading with respect to fatigue.
Many tests by American Association of railroads.
Pile driving stresses.
As long as the yield point of the tendons was not exceeded, cracks closed.
6 5 x 10 cycles of loading reported on a machinery founda--
tion.
w 9
--______.____.________--_._---__--___--____-______a
<t
- r e.
'4 p',
h Have had. some ' experience under earthquske conditions..Tashkent,7 Niagata, Anchorage.
.4 g reported. - Had. 30 prestressed ' structures.. As failures of grouted prestressed members.-
Sore failures were:
Needed'better' shear connections.
Four seasons building. Unbonded, wedge-type anchors, building did collapse, j
30% of tendons did fail.-
r l'
' Substantially, all experience has been with bonded-grouted tendons.
International Federation of Prestressing. Believe bonding and grouting should be required in seismic zones.
There are many applications of unbonded cables.
ACI is now studying.
Corrosion -. generally, tendons have had no trouble. Have been enough cases. of -
corrosion to be of some concern:.
' Chemical. Electrolysis. Generally porous concrete.
Presence of impurities. Chloride or fluoride. Also. presence of metals.
Sulphides.
Can be troublesome with just moist air.
1 Present trends. California Division of Highways.
Tendons installed unrusted.
1 Af ter installation. ' Keep out moisture. Af ter stressing, apply lubricant or grouting within 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />. Protect the anchor. Concrete alone not enough.
The concrete is important.. Volume change. Limiting the water.
Precasting.
q Creep.
Important. Choice of aggregates, i
i Steels - minor amount of stress relaxation with good.
)
Construction practices important:
1.
Ducts - keep straight.
2.
Joints - seal to keep grout infiltration.
d 3.
Stressing - friction values not as high as previously anticipated.
One end practices.
4.
Grouting or grease - An art.
Is important. Recently good specifications developed.
European.
La Guardia Airport.
Practices were good.
England - major tendons are Xrayed.
i l
'e q
1 Anchorage ends. Delieved grouting is extremely important for assuring better dynamic behavior - even where remainder of tendon is greased.
Pertinent. Experience:
1.
' Corrosion. Europeans believe grouting okay. Trying use of Epoxy coatings.
Lack thereof.
2..
Seismic behavior studies (international).
Rotation.
3.
Expansion cements.
Shrinkage compensation.
I 4.
High and low' temperature behavior.
5.
Biaxial and triaxial test conditions.
In general, behavior under multiaxial conservative.
6.
High strength concretes.
7.
Highly stressed concrete.
50-70% of concrete strength.
8.
. Dynamic wave phenomenon.
9.
England - Method of stabilizing concrete by repeated loadings.
10._ Prestressed and precast elements are placed into cast-in-place concrete..
5th World Congress on Prestressed Concrete - Applications illustrate how widely.
S. Bush - Question: Nitrates and Nitrides - What about these?
Answer:
Are bad outside concrete but not had when protected by concrete. Most cases of corrosion can be traced as to how treated during construction.
Rust - Some believe that once started will likely to con-tinue after installation.
Palladino: General Comments - Not a public meeting, speak freely.
ACRS has a difficult situation. Want to explore limitations.
Need the help.
Questions:
1.
Regarding experience, have there been f ailures other than earthquakes?
Corrosion. A number of pipe failures. Calcium chloride in the gunnite.
Some tanks. Were wrapped and gunnite.
______.__-___.._________-____.__-_______.______.____.-___,__.__--__M
~
i 1
l6 '
-. ~
2.
Pr'actices for inspecting tendons after installation.
Probably a neglected area.
3.-
Two tank f ailures reported.
Corrosion due to exposure of wires to chemicals in the contents.
Sulphides.
(Two tanks out cf 3,000).
4.
Series of f ailures in Germany.
Planks. Very small diameter. wires.
Piano wires.
Poor concrete coverage.
A number of schools, barns, i
etc. had catastrophic failures. An " alumina" cement and the thin coat. Alumina cement caused considerable more trouble than~ Portland cements. Chlorides have been the major difficulty.
(Comrent by E. Hognestad) 5.
Gerwick: Keeping steel clean and controlling cements are important.
" Proper mixes" and " aggregate.s". High quality mixes.
Architect engineers need to identify:
a.
Does do more than strength.
b.
High quality concrete under anchorages.
c.
Do not generste high heat of dehydration.
d.
Keep water down - avoid excess shrinkage.
e.
Control temperature of mixes - many need to put ice.
f.
Ia a big pressure vessel may have to measure temperature.
3' no t thick - 16' would be.
In dams may put cooling coils. Use thermo-couples or resistance therm.
6.
In congested areas may be voids - What happens?
Answer: Will show under the prestressing.
Also " nondestructive" testing techniques.
Uniformity is important. Getting every batch the same.
Compaction may be a problem. Must be workable, yet not too watery.
Hognestad: Documentation of good construction practices can be done.
One of the structural engineers should be on the job.
7.
Okrent: How do you find that an error or a poor construction practice has not been made?
Answer:
The stress on the plate and concrete under will be maximum at the prestressing. Before concrete shrinkage and steel relaxation.
California Highway Department mentioned detailed specifications.
8.
Fistedis: Does not agree with grouting tendons. Thinks that inspectability is important and is possible for nuclear applications.
t Seismic problems are not the same for nonnuclear buildings and nuclear facilities.
9
-~ :s o
i po :
r l-
- 10 P
- 9.u Building in' Texas.. ' UnbondedL tendons. Very brittle' materials. - Possi-bility for catastrophic.
Bonded -' Structure and prestressing acts as:
a ductile material. Unbonded - acts as a brittle.
r
-Rebuttal: Situation on the concrete.was bending. Most cases.. c Tension.-
Grouting ' leads-to local stress. Unbonded disturbs loading.
- 10. Kesler:. Can the committee better define the kinds' ef-situations of concern?
Palladino - - Answer:. ' Accident. Must raintain leakage. Few" cycles -
impesed for containment.
Primary' ve'sse1~ - now operating cycles could be several thousand..Perhaps:
backed up by the containment.-
. Vessel might act-for both purposec '- operating cycles 'plus the once in -
the lifetime accident.
- 11. ' Benders Possibilities. of openings blowing out...Is it-conceivable to getla split through.the anchor?
(Answers noc clear.)
- 12. Bush - Question. Reactivity rsep. Very lar5e-increment of energy.
Gases. tend to compress.
100-1,000 MW-sec range. -Several hundred pounds
~
of TNT. What would happen.to concrete?
Fistedis - Answer A case has been investigated. Assumed up to 500 pounds of' TNT.
Insulation between concrete.
Shock wave to liner.
. Attenuated in the insulation. Most effect lost in liquid movement. Gas hubble.
Shock wave in microseconds, liquid moves in millisecond, gas bubble in' tenth-seconds.
Strands do not see the strain until concrete cracks. Mass of concrete cannot be accelerated fast enough to develop anything but low strain rates (in the order of' ten).
~
- 13. Palladino - Question:
Can Marsh summarize criteria for failure mode design?
Marsh - Answer:
(1)
Classify into two the kinds of things one does
^ "
but rot realistic f ailure conditions.
(2) bbreal - opening.
s.
If failures occur, they must be on a ductile basis.
1 b.
Failures of component parts should not lead to further failure.
i
-.e~
p l
7..
n 1
.Q 'f
.l.
m hV :
7
11 -
l
. 3.
.mm.
-o,<>r.a e
r*+
=,
i ;
c.., Failures from. conditions which might be experienced: -
'(1) Corrosion (2)
Inelastic deformations.
d.
Margins of safety.
L Failure Pressure l:
. Normal Operating Pressure Numbers'in order of 2 have been discussed for some more recent.
designs.
APPLICABILITY OF EXISTING CODES - R. C. REESE
.n ACI-318.63. ACI writes saterial. Becomes's code when accepted by regulatory groups. There is an ' international group - small, trying to set forth a " code of practice".. CEB.
The ACI code covers normal and. prestressed concrete. There is a Prestressed.
Concrete Institute.
CEB - European Comittee' for Concrete. Writes best. techniques and philosophies
,of design. ~ ACI' code - no way of 'getting " code case" considerations.
Reliability of these documents - great care given to get the best: possible jobsJ on these codes.
ACI has no great restrictions on membership. 49 members. Chosen for. geographica'l distribution and knowledge.
If'no contributions are asked to step aside.- ACI has many 'subconsnittees.
Standard is a consensus - has much safeguards and criticism.
Applicability.
To nuclear reactors is lowest common denominator of the group.
Needs almost unanimous. opinion.. One strong-minded viewpoint 'not justification j
for not proceeding. Approximately 90% of committee in f avor the standards are i
No way for the committee to interpret cases.
enacted.
Code is largely empirical.
Produces highly reliable results. Not easy to relate various relationship.
1 l
Now trying to review the existing code to see what can be updated.
Hope to I
incorporate some additional. results of experiments.
In itself is not a completely sufficient code for the design of pressure vessels.
Needs supplementation.
)
Perhaps by 1970 some changes.
6
I P g',
l*
_ _U.12 ' ~
~
- j ACI Committee 349.. Working with ASME. to do some further revision work. - Not sure' that a " code" is what ' is needed or. desirable at this. point.
ACI may come out with sowe : series of _ reports dealing with "reconnended practice".' Safety provisions.. Empirical methods. _ Gave some very. amusing illustrations of, unusual' design requirements.
Point 'made was a. need for assessing the unusual.
ACI-349 wants to work with ' the nuclear people. to better define the conditions and consider; probabilities.
How likely are the circumstances for which we are designing.
Bush - Question ' Lag of codes behind good practice., Does ACI' run into this?
Reese - Answer:
- To some extent but, generally speaking,i lag has been 'cnt the conservative side;- that is, trends.have. been to decrease margins.. On the other. hand, if saf ety had-been. compromised, code changes could be made faster.
Each coantry has its own standards.. Have a-different evaluation of life and' property.
' Polladino - Questions - Does compliance with the ACI code assure quality, emphasis made by previous speakers?
Reese
. Answer:
Not necessarily so.. The code is minimum, but-more is needed.
Questions -Where could one go to get the "more" that is needed?
Answer:^
Papers and reports.
1 Shortcomings:
'I. ' Area of high temperatures.
2.
Openings.
3.
Code does not deal specifically with corrosion. Grouting not specifically required.
Nothing on lubricants.
a ACI-349 has much draf t material for consideration. Will be next year before papers l start coming through the ACI Journal.
ACI Subcommittee on "Untonded Post-7ensioning" should have something in the form of a draft out in October.
I 1
l DESIGN BASES - C. P. SIESS 4
Practice of using conventional rules for nuclear - questions raised because of:
)
)
1.
' Difference of consequences.
2.
Long-term requirements.
Second question not necessarily a difference -- even bridges and other structures also must be built for 30-40 years. Differences in consequences of f ailures --
not certain he understands this aspect fully.
)
~~
6 4
1 Basis for " currently used design code rules".
Two major sets:
1963 ACI Building Code and 1961 Edition for Highway Bridges (also one for railroad bridges).
Both i
of these codes embody safety provisions. Are in many ways similar; in other aspects are quite different. Basis for their development. Both of these criteria i
consider what can happen with an individual member--not e structure.
Also deal with what happens to beams, torsion; e.g., not treated.
In buildipgs and bridges, loads are determined or assumed. Moments or shears are determined by analysis. Resistance to moments or shears is determined as a function of geometry.
Saf ety f actors are introduced for uncertainties:
Dead loads x 1.5 Live loads x f Must allow for increase in loads and the load effects. Mocents and shears computed by specified formulae.
Strength reduction f actors; e.g.
Moment factor x.9 Shear f actor x
.85 Based on material and geometry differences. For bridges the strength reduction factors for moment and shear not used.
Not possible to compare directly these criteria. One must consider how the live and dead load f actors are used and the uncertainties.
i Even for conventional structures, the " uncertainty" f actors reflect the use; e.g.,
j schools might warrant more conservative f actors.
There has been consideration j
l to increase " safety" factors where people were living:
I 1
design load = a function of the consequences of f ailure.
I mean load Safety factors. Mean to compensate for uncertainties, undersize, etc. Have not been rigorously established.
Represent the consensus of competent engineers.
l Collective judgment based upon experience. Much of the experience related to I
conventional concrete struct sres rather than prestressed.
Prestressed research on beams and this done on scale nodels. For outnumbering columns or slabs.
Load f actors largely the same for conventional building (ordinary reinforced concrete) and prestressed concrete.
No unequivocal position, but no confidence that use of ACI codes will lead to quality and safety factors required for nuclear installations.
Experience and basis for codes have been largely on beams.
For conventional reinforcement.
Formulae are written in terms of For prestressed in terms of ultimate strength.
The overall safety factors derived come out to about 1.83.
Load f actors and design f actors cannot be treated separately. By "under reinforcing" steel governs the behavior.
Steel characteristics dominate rather than brittle behavior of concrete alone.
l A
7 p
c l'
(
lA.
i n
- 14 l
ANALYTICAL DESIGN TECHNIQUES AND MODEL TESTI!C - E. IK)GNESTAD l
Two major points - viewpoint of a structural engineer:
1.
Concrete probably the most " butchered" material but serves us well g:neral. The point is - need sound engineering and sound construction.*
2.
Details are most often controlling. Must guard against no modes of f ailure.
We must see that our designs include a capability to handle overloads and buckling.
Enough ductility to readjust loading. Must have enough tolerance-to allow for some cracking.
In butidings we do a stress analysis. Hold tensions to within certain predefined limits. In a pressure vessel, the ultimate strength of the main members can be computed.
A stress analysis to study the conditions under load.
Complicated since three dimensional. Using elastic theory.
Soue 12,0007 equations.
Major difficulty af ter some analyses made -- what are the strength f actors?
Not at all certain.
Forced to go to models.
Slides:
Emphasized the importance and usefulness of models.
Mentioned Japanese nodeling of a dam for seismic studies. Also mentioned a certain Italian capability to model and test.
Urged more than just theory.
Theory okay but check with modeling.
PCA Laboratories not impressed with possibilities for use of prestressed concrete for pressure vessels. Much larger needs for highways.
Recognize importance of needed power.
Question: (Koerner) Would modeling be a good idea for pressure vessels for reinforced structures?
Answer:
Yes, for the primary pressure vessels -- not conventional or secondary vessels. At least not as a general rule.
In a highly unusual case even a builder would probably recommend.
If it is too f ar f rom experience presently existing.
Question: (Koerner) Since containment is tested before use on 4, full segle, are models necessary?
8-Antwer:
Probably not test full scale to destruction - would provide assurance.
Question: Could you scale a 1/4" liner?
Answer:
Could pose real difficulty, but cautioned against too small a model in such cases.
s.
s
, Question: Has modoling been done in England for large towers that have failed?
Answer:
England has made up' one model but has not yet tested. Are now applying wind loads to try to_ explain.
FATIGUE AND CRACK GROWTH - G. WIlrIER Safety factors. History of the17 development. Originally, overload and under strength factors were used(d) but in a somewhat different way.
Came from statistical studies.
Under strength
.65 for concrete
.85 for steel Dead Load
- 1.5 To reflect more correctly what.might be expected realistically. Under strength factors were adjusted and.so were the load f actors. Point is that nuclear des.igners should consider their own knowledge about loadings and develop pertinent factors.
Fatigue and crack growth. Most of the "up" data on other prestressed structurcs.
Also on beams.
So far as cracking, the concern by the ordinary designer is considerably dif f erent f rom what the nuclear application involves..In prestressed -
cracking represents a change in flexural behavior of a beam.
Nuclear interest in gas-tightness but depending primarily upon the liner; therefore, cracks that do not affect the liner are not of major concern.
Biological shielding-against neutrons - Some decrease if a long through crack.
Presumes this is not a large factor of importance.
Cornell has been doing work on cracks. Microcracking - material problem due to aggregates. Macrocrackt.ng - due to stressing.
8 57 At 307. microcracking beginning of inelastic.
( 70%
At 707. much stronger inelastic behavior.
30%
At 85% cracks become continuous.
Between 70-857. concrete develops its
" sustained strength".
c Strain j
1 For repeated loading (cycles 10'), the 70-85% also repeated load limits.
)
i
--mm___.,.____.
l
- 16 Macrocracking - flexural and tension cracks.- For prestressed, uniaxial. Cracking occurs at 400-600 psi stress.
So small that not visible but does show as a reduced flexural rigidity of the beam. Cracks become visible at 900-1000 psi.
In a vessel, the cracking may not be significant.
The steel takes up the load.
W = Crack Width =
f/CA ( f -3)
= 1000f/sq.in. (Reinforced Concrete) f, = stress in steel C = cover Work has been done on putting small wires right into the mix. 3% by volume.
j Increased the tensile strength by twice the morter had without. Also changed the behavior to somewhat greater ductility.
Not clear whether results obtained on mortar would hold for concrete. Believes will get some beneficial effect.
Suggest some development along these lines might well pay off.
Under sustained Icads, cracks do grow but not much. May be in the order of 30-40% f rom that et which cracking initicily occurred. Crack widths are measured at the surf aces.
No experimental evidence to mock the situation of vertical prestreesing but hoop stresses.
(Modules of elasticity are the sane.)
Answer:
The prestressing vertically should retard by Poisson's effect the cracking from the circumferential loading.
Not too much. A factor of maybe-
.2 might be experienced.
Cracks you see on the outside. Are opened to the steel but reducing in width from the surface.
CEB - last year care to the conclusion not enough on cracking of prestressing to comput with any definitive guidance.
3 4
Low cycle fatigue (10 -10 cyclesg. For nuclear vgssels we are talking about low cycle. Few data.
Sgme at 10, not much at 10 and lower.
Fatigue strenSth of concrete after 2 x 10 cycles. Approximately 70%. Better somewhat in case of stress gradient (approximately 15%).
Prestressed concrete. Cracking tensile strength goes down considerably under cyclic loading.
Can develop as low as no computed tensile strength.
Anchorages of unbonded strands may be subject to f atigue loadings. German tests. Wedge anchors.
Reduced strengths in the order of 1/2.
These were run 5
3 in the crder 10 cycles. Would guess at 10 reduction would probably have been considerably less.
Illustrates effect and importance of anchorages. " Hoop anchorages" provide little reduction on f atigue strength.
e
s
~_
,I Cracking in anchorages.
Stress fields.
Due toL having so many. anchors in proximity makes the stress fields very complex. An area which should be con-sidered in any specific case.
Some tensioni ests. Plain tension members.
t Prof. Winters suggested that his simplified formula could be used for micro-crack predictions on a Brookwood-type design. Useful only for'effect on the liner, except for information of design of course.
Corrosion is more offset by other factors than cracking.
The f actor C was mentioned.
Healing. Cornell's experience limited, but Winter expressed skepticism.
Required high humidity.' Before healing.
Cerrosion likely to take place.
CONCRETE - B. MATHER-Concrete denotes a vague set of properties that can vary as widely as rocks.
Characteristics umst note explicitly be defined than just " concrete".
A mixture of aggregates, cements, water, ad-mixtures.
The only one uniquely defined is vatar. The other ingredients can greatly influence the characteristics of the combination.
Portland cement, type 2, very good, but not necessary.
A homogenous product.
It is not impossible to get lots in which successive lot s vary 2-1.
Ratio of water to cement that ' determines what the strength will be.
Foint - All concrete should be " tailor-made".
If specifications are specific,
.it should be possible to achieve.
Range in properties. Characteristics are not constant with time. For a nuclear application. Forms are already quice full. Makes the problem of assuring lack of voids, etc. s real one, Concrete continues to gain strength with age.
Does eventually reach an optimum.
If concrete stays damp (1007. humidity), it should increase in strength.
If concrete stays dry, probably stays at the 28-day strength.
If an " unfavorable" environment, could deteriorate. At a relative humidity of 50%, it has been calculated that it takes 100 years to dry out to a depth of one foot.
Dataat40F,70,effectofgemperature,butthis Calculation and tests at 72 F.
Expectfinear point not well established.
F, and 100 F.
Isothermal environ-ment.
No data at temperatures of 150 F under a temperature gradient.
b l
jo
~
- 18'-
i
(
I' The phenomena that cause concrete deterioration - Chemical reactions f rom the environment. Freezing and thawing under a water environment. Drying and shrinkage occurs, but should be controllable.
Thermal gradient effecg. Heat should not produce deterioration per se. Ac least at levels of 150 F.
Referred to work of Ross in England on thermal effects. Heat of hydration.
Large masses.
Insulate first winter. Keep peaks down as a way of accelerating average mass temperature toward steady-state.
PRESTRESSING STEELS - H. K. PRESTON Post-tensioning tendons. Variety of forms.
Seven vire strends.
Strands Approximately.5 to
.6" Bars - Made of high strength steels. 80' lengths and given permanent elongation, then stress relieved.
Strands - 7 wire.
250,000 psi strands.
New on market, 270,000 psi.
Slides shown. ' Strand f abrication process.
Each wire
.16".
Rods are butt welded.
Elongation - 3.5%.
Tensiles run on average approximately 10% above specified.
Resistance to stress corrosion.
ACI codes prohibit quenched and tenpered steels.
These are more subject to stress corrosion.
A mininum relaxation under load is important.
Stress' loss versus tim is a straight line on a log-log curve.
Af ter 707. of ultimate for 18 hours2.083333e-4 days <br />0.005 hours <br />2.97619e-5 weeks <br />6.849e-6 months <br /> 3 f "E I0 and other load reduced 20,000 psi at 0 ars below initial Elevated temperature influences strength loss only when held for extended periods of time.
4
W ;.
. c..
4
" Stabilized Strands'" - Developed in England Somerset Wire Company Stress in Temperature-Ultimate Stress Relieyed Stabilized 68 F 62.5
.10 1.8~
70 F 70.
11.5 2.5 120 F 70.
17.0 5.6-140 F 70.
'70.
Clean wire strands. European experience with rust. Were not stress-relievedL wires. May have been the cause of getting further nisting af ter grouting..
PRESTRESSI!C SYSTEMS AND ANCHORAGES - M. A. S0 ZEN
^
Discussion to center upon:
a..
Anchorages.
b.
Structural Response of Bonded versus Unbonded.
c.
Inspectability.
Slides:
Several basic types of anchorages.
s.
Wedge-type with threaded inserts into slots.
b.
Bearing each strand attached.
c.
A "tankshead" type arrangement.
Bonded versus Unbonded. Grouting may not necessarily mean bonding although grouting can be a form of bonding.
Bonded p
-Unbonded e
e
- i '-,
L In general, under pressure unbonded will develop cracks earlier than a bonded.
Slide: Transver se - Ioad s.
'in Bonded Members Concrete C" :3
. Bonded a
Steel J h-1 Unbonded A
L R---
92 The same movement in concrete taken by a much longer section~of steel and therefore'c'is less.
l Unbonded member is weaker and more flexible. Bonded will absorb more energy.
Slides: Comparability - bonded versus unbonded.
Photo - showing bonded beams develop staller cracks, but more of them'.
Greater ductility.
l Slide:
Ratio of measured'to computed to ultimate.
Point - know how to calculate the. flexural strength of bonded and unbonded beams.
Slide:
Moment Versus Average Curvature M
M
- 3 xtL As long as one is low down in the stresses (before cracking), one cannot observe much dif ference between bonded or unbonded.
Slide:
Shrinkage plus creep. Main point - Concrete will creep to approximately i
2-4 times initial strain as a function' of time. Believe this to be l
predictable.
i l
l l
l l
l
E r
.4 a
21.
s.-.
Slides Rate of Creep l
'20%'7 days.
807.;1 year ! 50%
I
' Slides Behavior of Beans Transverse and Time-Dependent Loadings - Mid Span Deflection versus Time.
Points Can predict time. dependent creep of
. prestressed beams.
Slide: Mid Span Deflectior, versus Time.
Point s Can predict reasonably.
I If-inspection 'is desired, not justifiable for time-dependent creep and ahrinkage.
Believe.can be predicted well enough in advance.
Does not have too much f aith. in load testing as a measure of structural integrity.
But since must test for leakage, can instrument for structural understanding.
But if check of failure modes is desired, go to model testing.
Question:
(Fistedis) Argonne questioned the unbonded versus bonded implications presented by the speaker Newmark.
Argument really out so far on the elongation range beyond rcsim of interest for nuclear applications.
Preston:
The bonded cable will eventually develop its specific elongation before breaking as well as the unbonded.
Hognestad For practical purposes - the choice is not nearly so great.
Newmark :
One has_ to design whether one is designing for a blast shield or a pressure vessel function.
i Okrent:
Does not see the explosion bearing capacity as a predominant interest i
for the next few years.
Primary function.
1 Anderson: Bonded - smaller cracks - better assurance for liner. Anchorages -
better secured and less stressing.
Koerner:
Should decide upon design requirements.
Each has advantages. Where crack limits are tinportant, then decision must aim toward such an l
objective.
Marsh:
Where there is a large percentage of steel, the bonded and unbonded curves show little diff erence.
Sozen agreed.
Bush:
Arguments about gross deformation are academic for applications under consideration. One must move f airly f ar down the curve.
L.
I
s E
- e:,
Bender:
Raised question as to whether bonding 'does provide a backup for j
the anchors.
Solicited comments.
Answer:
Preston - Tests they ran sho[ed grouting performed' the same holding function of the anchor, j
Gerwick - Have had some experience in-pile driving where -
anchors removed - very little.
E Anders'n - 8' of grouting in 6" hole developed 800 kips.
e o
STEEL CORROSION PROTECTION - I. CORNET Expressed agreement' that grouting may not be' bonded.
l-
.Unbonded tendons have been protected since 1952 with gresses.
Enough known.-
g.
- Properly applied, they will protect.
Allows for inspection. Can replace l
grease.
Compositions - several mentioned.
Cathodic protection - 1.05 volts relative mentioned as one value.
Is commonly employed.
Can be controlled and instrumented.
To protect steel, concrete or mortar should havet 1.
Proper. composition:
Type 2 Portland Cement.
Low water mixtures.
Low permeability.
Restricted chloride and sulphide contents.
2.
Proper thickness.
Suggests at least 2-1/2" thickness over steel in salt water environment.
3.
Proper placement.
4.
Environment not aggressive or environment controlled.
Cathodic effects, etc.
Corrosion effects can cause cracking.
i e
1 l
0 Slideo:
California - large water tanks. Date back to 30's.
Show cracks on the surface. Mcde in slabs, placed like barrel staves and bounded by hoop-like cables. Mentioned his article:
" Corrosion of Prestressed Concrete Tanks".
New York Sludge Tank. Collapse. Suggested a hydrogen-sulphide embrittlement of wires.
Cement SlurgyTank Failure. Wire corroded. Ductile failure.
(By
+
bond - stress transferred in 18-20").
Penetration - Showing leakages. Poor practice in water tank design.
(Variety of other slides shown illustrating the aggressive nature of com,e of the. fluids held by the tanks.)
Report on work at University of California. Corrosion of steel in concrete.
Showed electrolysis induced corrosion causing cracks in concrete specimens.
Repeated salt water immersions.
Questions (Bush) 1.
Has bacteria played a corrosion role?
Answer:
Pollution plus sulphate reducing bacteria can create hydrogen sulphide and cause adverse effects. Quite a lot ic known about this kind of problem.
2.
Stress corrosion - Have clearly validated cases been observed?
Answer: Grouting lef t out in one industrial (coal mining) application.
Anchorage trouble.
Sulphides the main off ender.,
3.
Stray currents.
Under conditions what are the rates involved? Can these occur over months?
Antwer:
Effects can be quite raoid in some cases. Mention was made of the high voltage DC developments. These could be a future source of difficulty.
Situation is being watched closely.
Question: (Pallodino) Experience with Erease protected cables.
Anewer!
Principles are well known.
Been used since 1952. Many applications.
Fingerprints on machine parts. One can get seepage through cracks.
Bflieves good steel can be protected by gretses for many years.
Question: (Okrent) - Cracks in concrete - What effects on corrosion?
Answer:
Deitruyn h4s published much material on this subject.
Opinion expressed that cracks enhance chance of corrosion.
Coastline, such as California, a hostile environment.
20 miles inland.
The sea ef f ects should not be present.
j Cornet expressed belief that crack effect may be overemphasized.
California has " salt fogs".
1 hh_
4 o;-
m, 4.
-. ~...
Question: (Bender)
-Brought up Marcoule cable experience - What led to these difficulties?
Answer:
'Did not know the details.
Koerner: Keep in mind that cracks should not be present in a pre-stressed containment vessel. May open up only under the accident environment.
Questions (Levine) - Brookwood - Prestressed tendons in the ducts - then inside ducts filled with grease - therefore should cracks. in concrete affect tendons?
Answer:
No.
Tendons are in ducts. Well inside. Reinforced steel on outside for crack control.
Hognestad: Much work done overseas of corrosion through cracks.
Eight years investigation by Corps of Engineers - Results: If crack <.01 near the steel, nay get corrosion - otherwise not.
SEISMIC CONSIDERATIONS - N. M. NEWMARK Hazards of an earthquake. Always increases with our experience. One of major problems. What is the intensity for which we must design? How do we get this from history? Not sure cf what the maximum might be.
El Centro 15' displacement.
14"/sec. velocity. 1/3 g. for acceleration.
What is the maxitnum that one can predict. Uncertain. Perhaps, several feet ground displacement. Accelerations up to.5 to.7 g.
Velocity 2-3 ft/sec.
Faults.
Parmanent displacement. Conceivable several fee *;. Ground motions more slowly than the oscillation of motions.
Response of a structure. Depends upon its natural frequency, dampening effects and coupling to the ground. Mathematical predictions may not adequately account for the coupling loss.
For reactors we may be 6-10 times more conservative in the response. Also several' times more conservative for the assumed earthquake
' design basis.
" Response spectrum". _ A plot of one or more of pseudoacceleration versus fre-quency(7).
Can obtain maximum force or tnaximum energy for which we must design.
If natural frequency is > 5 cps, then the acceleration is crucial parameter for design. For 1/2 to 2 cycles, the structure absorber energy. Maximum ground velocity becomes determined.
Below.2 cycles. Maximum transient dis-placement is the crucial design consideration.
4.
4 7
l: :> ;
- b1
<r
! l,
- Theref are three-types of procedures for design
1.
Most' complicated
. Force time histories. Do several computer. analyses.
The dampening characteristics of the structure have great 'effect on the results.-
- 2.. Model analysis for each of the modes of vibration.
Explore the time history for each mode and look for maximum. May expand by also'using.
. probability considerations. Mostly linear analysis but can take some-nonlinear effects into account.
- 3. ' Empirical procedures - Methods now used for large buildings generally absorption. of energy;in the inelastic range..
Probles Areas Because of. uncertainty about the actual bonding achieved, would still want good anchor ' designs..There will be stress concentrations at the anchor; but unless the load _is raised considerably, there is no significant change in the stress in the tendons.
Cracking. Only prevent if loading does not produce tensions in the same section where tension exists because of prestressing. Under dynamic loading, prestressing with some eccentricity can produce undesirable effects.
Seismic loads can stress in the reverse direction. Cracking ind ultimate load depends upon whether designing for a loss of function or just safe shutdown.
Pene tra tion s.
Present points of stress concentrations.
Not enough research to allow us to proceed other than very conservatively.
Some of the problems of reverse loadings might be handled by " hinging". Creates need for hinges to handle moments and shears. Hinges may not be okay for seismic j
conditions.
I Joints.
Sidewall to base and cylinder to dome.
Potential problem areas.
We need positive anchors.
)
We need corrosion protection.
If we choose, can provide for inspectability.
We must provide resistance to loading in different directions.
Need further studies on penetrat ions, joints, and hinges.
Seismic environment analysis, model tests, and perhaps full scale tests.
1
H e
26 -
Question: (Okrent) - In active earthquake zone, any distinction of bonded versus unbonded?.
Answer:
Inspectability great advantage.
Is not rapporting the " consensus" opinion (not clear what this meant).
Grouting may not always provide corrosion protection.
Could get good bonding using unsmooth conduit or ducts.
(Consensus seemed to be one could not build without ducts.)
Question: (Palladino) - Confidence in computed results of. seismic effects?
Answer:
(Newmark) - Confidence in methods for assumptions used.
Not necessarily confident in earthquake " zones" that have been identified.
Answer (Steinbrugger) - Uncertainties require conservatism in assumptions for seismic design. Matters of judgment, indeed, involved.
Can easily be off by an order of magnitude.
Could affect the decisions of bonded versus unbonded aspects.
Question: (Bender) - What is effect of loss of vertical prestress due to seismic effects?
l Answer:-
(Newmark) - So long as liner remains intact should be no significant I
problem.
PROIOTYPE TESTS OF VESSEL MODELS - W. ROCKENHAUSER Two demonstration models. Full size components used - full sized aggregates and full sized steels. Models were developed. Check that the approach (design and analysis) is reasonable.
Are not scale models.
Slides:
Photos of the models at General Atomics.
Stress analysis done for the design was rather extensive.
Structure subdivided l
into many sections. Discussed some of calculations made.
STRUCTURAL PROOFTESTS AND IN-SERVICE SURVEILLANCE - A. ANDERSON Showed slides depicting a wide variety of prestressed concrete applications.
Surveillance.
Should start with construction.
Proof tests of vessels. Might q
be done in the post-tensioning of the cables. Jacks pull cables temporarily up to 80% of tensile strength (catalog ultimate), then relax down to somewhat lower values.
Anderson suggested taking up to 200 K psi as a test load and then seat at 135 K psi.
s
~
[
s O
v,,
Cracking load criterion has been discarded in many kinds of applications. May want to revive for nuclear applications.
By taking tendons up initially, this could be equivalent to f actors of 1.5 to 1.6 of the load pressure.
Leave 57.
tendons unbonded.
Place upon load cells. Quarterly tests for first year.
Decrease interval with time.
Annually after several years.
PENETRATIONS AND LINERS - I. ARTHUR Slide:
Liner and Penetration Arrangement.
Liner. Function.
Primary barrier to the reactor coolant in a pressure vessel case.
Isolates coolant from the concrete. Transmit internal pressure load to the concrete.
Minimal requi rement for displaying structural strength.
Liner also acts to keep with temperature control of the inside surface of the concrete.
//
[
Slide: Typical Liner Cross Section.
Thermal
/O Insulation Stub Anchor
//
Cooling Tubes
~%
C h
Liner. One objective is leaktightness.
The concrete itself in the pressure vessel is also a barrier to leakage of fission products.
In the GA tests --
Several holes bored in liner. Helium used to pressurize at 700 psi. Measured leakage at 70 pounds. Results showed concrete does provide some significant holdup even with a leaking liner.
Liner. Objective is to make it conform to the concrete.
Liner ductility is therefore important.
Liner is thus strain-controlled and not stress-controlled per more conventional pressure vessel desibn criteria. Ductility requirements.
We must assure brittle behavior not possible. We have to consider effects of irradiation.
NDT types of considerations.
Probably considerations of type given to conventional nuclear prest,ure vessels will suffice for liners.
Tensile forces and shear forces between liner anchor stubs and concrete. Quite a bit of experiments and data on the behavior of the interface between liner and concrete. Techniques reasonably well understood.
Need to consider possibilities for f ailure by f atigue. Nuttber of cycles in a reactor vessel could be several thousand.
Fatigue needs to be considered.
ASME,Section III, the same kind of criteria on f atiguing seetrs to be applicable.
4
4 I,_.
Construction. Rather thin.
If used too thin, one gets into fit of problees and possibility for introducing difficulties at penetrations.
1/2" or above seem to be okay for these 1,iner applications. GA using mild steel - 212B or similar, plates in thicknesses of 3/4".
Cooling system. Criteria should be centered around reliability. Perhaps, redundancy in systems.
Stress in liner. Compressive.
Desirable but not a requirement.
Protection against crack propagation.
Penetrations. Have to avoid overgeneralization.
Showed a number of slides of penetrations.
Presumably ones planned by GA.
QUALITY CONTROL - CONCRETE - C. ITSLER Inspection. Checks on the job should not direct changes.
ACI Code - Recommends engineer or representative of AI to inspect the concrete. Good inspection.
Avoldo costly errors and affords opportunity for improving process.
Indicated processes commonly inspected.
1 1
Responsibility for inspection should reside in the " engineer". He may select inspection agency.
Inspection should not be obtained on a competitive bidding j
basis but on qualifications.
1 Clear-cut and specific specifications are a "must" for a smooth job. Two ways for specifying concrete:
(1) Specify mix and (2) specify strength. Generally speaking, specifications refer to the latter. Materials are generally covered by material specifications such as ASIM.
Still allows for variability.
Engineer cannot rely on strength requirements. Too late for job control. Quality control charts are an aid for determining good job.
Engineer has representatives at the mixing station and at the placement station.
A laboratory setup checks on quality.
Inspector.
Integrity end knowledge.
Even technically trained men need on-the-job experience. Requires man with thorough knowledge of the processes. Must have authority to enforce prerequisites and stop any work not per requirements.
Control of concrete for nuclear vessels important.
Inspection requirements may not be more severe than for more conventionci uses.
Some government groups have their own inspectors. Other agencies train their men. Many inspectors not given on-the-job instructions.
l l
l L_ _
~ ;^
y
.w
. s REINFORCING RODS AND PRESTRESSING WIRES - P. RICE Not much to discuss on prestressing. Will concentrate mainly on reinforcing rods.
Variation in reinforcing rod. quality. Ten years.. 1500 tests.
Some 30 low tests.
1949-1957 m = 1000 tests e - 3000 psi = standard deviation V = 6.8 = coefficient of deviation i = 123% of specified stress = average observed stress 1958 m = 1220 1/240 below specified ' yield a - 3000 1/1000 below 96% of yield V = 6.2 9 - 123 Before rods get out, a certain rejection has already taken place. Underweight in rods controlled by the rolls. Cross section varies with roll vear but not in cross section along a rod.
a3.5% by weight is a common specification control.
Rate of loading can affect somewhat the observed tensile strengths.
After rolling, bars are sent to a. storage. Along with the bars, fabricator is supposed to send mill reports.
ASTM Specifications: A432, 431, 408, 185.
Grade marks now on bars:
Up to 60 K - generally none 60 K - bars 75 K - bars Rolled on when the deformations are rolled on.
q At present time, ASTM trying to set standards for 40, 60, and 75 K grades.
L As to specifications, ASTM coverage seems to be best available and adequate.
Inspection.
There are requirements for mill tests.
Some agencies accept these as adequate. Others conduct own inspection and test all materials. Ran survey of many highway inspection groups.
Some cause considerable disruption of con-struction.
i I
v-
.4 pe.
-.30 -
- f.,
- Suggest for inspection.. Do much atL the-fsbricator. Dolsome.at receiving but not-those better done.at. fabricator. 'Should check for gross errors, general condi--
tions, grade marks, etc. ' If using unmarked bars, take field samples and test.
?-
- 0n welding'-
- Bars are not welding material..Get mill. specifications and adjust -
. welding specifications; but, in general, do not tack welding in. the field.
Irradiation on bars - affects ' ductility.
Rapid' loading increases apparent yield..
3 Fatigue - Do not bend bars in fields of high stress.
Prestressing wires - probably the best tested.
0-GENERAL DISCUSSION Surveillance Techniques Andersons ' ' "Carlson Meters".
Load cells. Can be used for stisins and cables.
Electrical SR-4 exhibit difficulties. Carlson mete 3s embedded in concrete. Temperature compensated. 'Much use in dam 3.
Several thousand. Can be insta11ed'with reasonable cars. Needs care to well embed.
' " Lime paste" around test tendons for observations.
Anderson:
Put plugs on the concrete or ' liner. Use a gauge to measure expansion.
Fistedist Something to measure overall change in dimensions.
Strain pauges under the. anchor.
Stress Levels on Prestressing Cables
' Anchoring
- 80%
Back off
- 70%-
Creep 'and Shrinkage - 60%
. Under Load -
- 66%
Cracks appear at 1-1/2 or 2 times design load. This means first cracking at roughly 70% by losa.
For reinforced concrete pressure vessel.
Should run periodic tests at less than pressures at which cracks might appear.
Perhaps,125% design pressure for a periodic test pressure sounds reasonable but not necessarily for reinforced vessels.
Prooftests should demonstrate structural adequacy.
l L ~
l
O'
-p g,
l,
leaks in coolant lines to the liner..Should anticipate some leaks will be experienced.
Questions raised about loss of cooling of the : liner. GA indicated this could be a problem,. but the time constant of the vessel is in the order of months.
' Concrete shrinkage puts the. liner in compressive stress.
Catastrophic failures of concrete versus steel.
Anderson: Concre'.3 structures are composite and capable of having con-siderdae reserve capability.
Fistedis: Embrittlement should be a considerably less problem.
Creep.- What in/in can one expect after, say, ten years? Can be in the order of 2 x 10-4 in/in. - About 1/7 amount that steel would be taking.
Stress loss in a tendon might average 20 000 psi as a result of creep. We do 3
not know too much at temperatures of 150 F, but experts thought it could be estimated to 20-507..
Hognestedt. Pointed out that in these large structures his work would suggest that creep and shrinkage will be considerably less than experience to date would i ndi cate.
Right now creep data are not 'on triax!al situations. Data now Isrgely on uniaxial.
Believed to be on the conservative side.
I
..