ML20008E421

Evaluation of Yankee Vessel Cladding Penetrations

ML20008E421
Person / Time
Site:	Yankee Rowe
Issue date:	10/15/1965
From:	WESTINGHOUSE ELECTRIC COMPANY, DIV OF CBS CORP.
To:
References
WCAP-2855, NUDOCS 8101070056
Download: ML20008E421 (81)

Text

. - -. -.

i WCAP-2855 I

i e

l

\\

$1 e

k l

hj i

I

,g-ff bS/

i 44

- hf EVALUATION OF YANKEE VESSEE -CLADDING '

.Q l e cc;g i

PENETEATIONS 3

l s'

Ill%

l

~ 'Lfp

/,

P y9

=

l

-c;7' j OCT 191965 > $5 A

stsancM i

Q.

y Att, 5EcTIM j

tcati CLEM 8

l a

cn i

October 15, 1965 W ETINGHOUSE ELECTRIC CORPORATION Atomic Pcwer Division P. O. Box 355 Pittsburgh 30, Pennsylvania

I I

i I

i Table of Contents 1

1 j

Section Title t

l I

Summary of Analysis and Conclusions I

II Mechanical Considerations III Corrosion of Vescel Base Metal l

IV Hydrogen Distribution and Concentration V

Hydrogen Embrittlement Appendix A Vessel Material and :>rocess Information I

i l

l I

L

. s

... -. - __ -.~.-_

1.

SUMARY OF ANALYSIS AND CONCLUSIONS i

Introduction During the 1965 refueling of the Yankee reactor, all fuel was unloaded, and the core barrel and core support structureir"e removed to permit a general inspec-tion.

It was found that one material irradiation capsule, which had been known l

to be loose two years previously, had broken up and released Charpv irpect ;p et-mens and other debris into the lower head area.

Apparently in this process it j

had worn several small areas in the lower head cladding in the NW quadrant.

i The lower head material is A302 B steel, clad by the B & W intermittent spot welding process with.109" thick stainless steel plates, ASTM 240. This clad-i ding process results in a welded bond over approximately 70% of the area, leav-ing the rest of the cladding - steel interface unbond id.

The unbonded areas 1

are interconnected, so in the event of a cladding defect, 30% of th" interface I

of one 4' x 8' sheet could be exposed to reactor coolant.

Although there were numerous scratches and gouges in the cladding, the carbon steel was apparently exposed only in two small adjacent areas. The total area of carbon steel exposed was only about 2 equare inches. As the capsules that came loose and presumably wore these penetration in the cladding were originally situated in the shell area behind the thermal shield, the possibility of their having caused similar penetrations of the cladding in this area has been consid-ered.

Inspection of this area is most difficult because the vessel is offset, and clearance is less than 2". Considering flow direction and forces involved it is very unlikely that any wear deep enough to penetrate the cladding could have occurred in this area.

Casts of the vorn areas in the lower head were made using a curable cilicone rub-ber compound. A photograph of a plaster replica is shown in Figure 1, where the L

exposed carbon steel areas are shown in white.

t l

Repair of these penetrated areas vould be very difficult, as the lower head is f

under about 55 feet of water during shutdown conditions.

An analysis of the l

possible problems that might arise as a result of continued operations with these i

penetrations was made, and is the subject of this report.

I-l

' re ?

4 e

r l'

rh

'?

E '.

ip

. c,.

u 's t,4 4

id!

$$Nf.j*g,^w I

\\

l~<fjy 2

>~%

+fhfei y htp 55/6 f'h 'if "2yih lr N

3 av i

.l$

,v

'l 4 $bEh

@f' hf

• k d-dy&,.QVt&%

x

1. ,5Qpti.

4c 3lf;{ysy _-

b':;;,,

?..

r j' ty"diff{&ll$f 4%

4 A*q S

'l,, jW..

,P kh j

- 4tp o3. j,hh g

"$yQ qm, h;

p:w k.M;

~,fn;J [(:q a g

~ jef t

n 3M

. U33'n kh$

M,.mygg g ;;

.,1 a;

y r

kA

hld, l} ',- Q g'-12l..

a

-:h sm 3

t A

Figure 1 Photograph of plaster cast of silicone replica of cladding penetrations (raised detail is due to folds in cilicone replica). Expeced carbon steel areas are painted white.

-Actual Size 9

.. _ _. _ _ _ _ _.. _.. _ _ _ _. _ _ _ _ _ _ _ ~. _.

i This Sect ion sicmarizem t N analysis made of all poscible problems that can be l

L forec,ee n at thi: time.

hd ai l s o f t he analy+ 1 cal methods and logic used to ar-l l

rive n'. t hee conclualonn are given in the later Gections.

It is concluded that I

t.here are no t,ar-t y problems involved in continued plant operations with the clad-ding pe v + rat l o*is.

Mechantral and St ruc+. ural Erfeet.s The n!rirm orig!nal thickness of t he A302 B steel as measured after forming of the hemisphere wun h.1.10" This compares with a minimum allowable thickness, as i

l required by Sect. ion VILL of the A::ME Code, of 3.T74" and, as required by Section III, of comething less than 3.0". The :ratximum depth of the defect has been meas-1 i

ured from the impressions taken as not nore :.han.115".

If it is assu-ed that

.010" of general veur has taken place near t ne penetrations, t hen the reduction in the thicknens of the A302 B steel is approximately.016".

This leaves a pre-l sent thickneas of h.09h" and at least.

3" of material availftble before the mini-mum al'lo<able dimenolon is reached.

On the basis of analyses summarized in later L

sections,fhia is considered to be a more than adequate corrosion allowance for the lifet ime of t.t.e plant.

A fat igue analysis considering all transient effects which might be expected dur-ing the life of the plant has been performed to detemine what stress concentra-tion fact or at the clad penet ration 3 vould be acceptable. This acceptable value uns detemined to be a factor of 9 Detailed analysis of the actual contours of j

i the penetrst ion reveals no very small radii and no area representing a potentially j

large at rear ccccentration. The stress concentration factor resulting from the penetrations gives a factor of 3.1 for the minimum radius of 1/8".

Corrosion and Coolant Chemist:ry Effect.s Penetration Rate l

The total penetration by corrosion of the exposed A302 B steel is estimated to be less than.1" for the rest of the plant lifetime. This is made up of 25 years of hot operation, with a penetration of 1 mil per year, and 30 cold shutdowns of 2 months, each, with a penetration of 1.5 mils per shutdown. As 4

i I

J l

1-2 e

• w-ee---

. ~. -.

,,r.

m-

vas chown above, there is presently.3" over minimum thickness; so a total corrosion penetration of.1" is acceptable, The penetration rate during normal operation used in the previous enalysis is based on a corrosion rate of approximately 150 mg/dm -mo.

This rate in turn is based on experimental information available to Westinghouse which was developed at conditions of temperature, flow and boric acid concentra-tion quite similar to those at Yanke. No allowance is included for gal-vanic corrosion since with the high resistivity and lov oxygen conditions of normal operation this is not expected to be an appreciable effect.

The penetration rate during refueling is based on a corrosion rate of ap-proximately 1500 mg/dm -mo.

This corrosion rate is a conservative estimate including an allowance for the effect of galvanic corrosion which may take place in the aerated water present in the vessel during this period.

It seems clear that any galvanic corrosion vill take place primarily at the aera of the clad penetrations and vill very rapidly diminish away from this area because of polarization of the stainless steel cathode due to the build-up of hydrogen in the region between the stainless steel and the carbon steel.

B.

Contribution to Crud Level and Coolant Activity If the entire unbonded area of A302 B steel within a 4 by 8 foot section, the maxir of exposure from a cladding hole, in considered to be cor-roding at <. conservative rate of 150 mg/dm -mo, there would be a release of 13 5 grams of iron per month or 18.7 grams of crud. This amount would in-crease the crud load by only 1% and cause an increase in activity of the coolant of even less than this amount during operation.

During cold shutdown, if a corrosion rate of 150 mg/dm -mo is assumed under the cladding 38 grams of crud vill be produced during a two month shutdown from the 4 by 8 foot plate. This will contribute a negligible amount com-pared to the amount produced during normal operation for a two month period, so woul not be a problem from crud load or activity considerations.

C.

Crevice Corrosion During normal operation there vill be no oxygen in the coolant and crevice I-3

l I

i i

corrosinn dots no* nor. ally cccur in t h. absence of oxygen. During cold i

fn. t d o.m r c *tli + 1 c.* u., d a*.a available show ' hat crevice corrosien rates vill i

f Ws tin t9 general corrosion rat e.

Alt hough t here is n. ore oxy-b" e s i

l cer. in t he general coo'. ant, t he availnbility of oxygen leird t he tightly fitting ntainl" ; clad vill be very limit ed.

l f

D.

f;t rees Corrocicn l

';o tecPa.nie exist c in t he area of the penetrat ion for the ccncent ration i

t j

of alkall vithout which st resc corrosion vil] not take place.

Furthermore,

[

)

c:cnium hydroxide, t he cnly alkaline additive normally used in the main 4

j coolant, is not subject to concentrat icn because of ito volat ilit y.

i l

1

]

Hydrogen 1:nbritt lement of A 302 B Steel j

i i

j Under specific canditions well document ed in the technical literature, the me-chanical proper *1es of st eel can be seriously impaired. This is co =only refer-red t o as " hydrogen embrittlement", a t er used generally to cover several speci-fic types of hydrogen prceluced phenomena. These can be categorized as follows:

(

1.

Temporary embri.ttlement caused by hydrogen in solution in the steel.

Re-l r

. oval of the hydrogen restores the steel to its no" cal state.

2.

Pe=an-nt hydregen dam 6e caused by formation of voids and blisters fro l

high internal pressures of colecular hydrogen, and, at high temperatures l

an1 pressur-s, decar'.rarizrstion of the at eel by formation of methane from t re carbon-hydrogen reaction. Voids ar.1 blisters can also be produced by the formation of methane, as it diffuses very slowly through steel.

Possible sources of hydrogen in the area of the penetrations are:

l l

j 1.

Norcal overpressure of hydrogen used to inhibit radiolytic decompcsition.

l j

2.

Radiolytically prcducel hydrogen.

l l

3 Hydrogen formed by the corrosion reaction.

l An exhe.-;tive analysis of *he quantities of hydrc ?en available from t hese cources

[

nnd tb effect< which tPcse quant it ies of hydrogen might have cn the structural I

propert les of tt: vessel has b - n performed. This analysis shows that the hydro-l l

1 l

IL i

i

_ _ _ _ _ _ _ _ _ _ _ _ _ ~., _ _

_ _ _ - _ _ ~ _.. - -. - - _ _ _

l i

I i

1 gen made available by corrocion between the stainless cladding and the A302 B ves-cel vall reprecents the major cource.

I i

l At operating temperatures and coolant conditions, the corrosion rate and the dif-l i

fucion rate vf hydrogen through the veccel vall are cuch that the maximum hydro-gen concentration in the steel of the lover head is approximately 0.' prm. This amount I

is at least an order of magnitude below the concentration that could cause reduc-

[

i tion in ductility even at lov temperatures. The partial prescure of hydrogen equi-valent to this conce. tration is only 21 poi, not enough to cauce any concern a-l I

l Furthermore, experience /1 bout decarburization reactions.

has shown that no i

permanent damage to steels of the A302 B type has been caused by even high pres-I sures of hydrogen at temperatures belov 600F, so it is clear that no permanent damage vill occur. During cool down, thiu hydrogen concentration vill not change appreciably, thus no problem can exist during reactor shutdown.

f l

i During cold shutdown, the water in the veccel is aerated. The corrosion rate at

[

the penetration vill therefore be higher (1500 mg/dm -mo).

This howcVer will be

{

2 1

galvanic corrosion, therefore the corrosion H vill be released at the etainless 2

steel curface and not enter the carbon steel.

In the gap however galvanic cor-rosion will soon be eliminated due to buildup of H and consumption of 0 '

I" 2

2 this case the corrocion H vill be generated at the carbon steel surface.

De-2 tailed analysis of the fraction of this H which will enter the carbon steel baced 2

dissolved in the steel l

on experimental data shows that the maximum level of H2 vill be no more than 1.2 ppm, which vill cause no problems with respect to safe i

I operations. The possible increased cusceptibility of irradiated steel to hydro-gen embrittlement has been considered. The only effect that appears pertinent is a possible increase in the tensile ductile-brittle transition temperature of no more than 40F above that caused by radiation alone.

In summary, an intencive rigorous analysis has been made of all possible dele-terious effects of the worn areas in the lover head of the reactor vessel. The conclusions are that the presence of the defective areas will in no way cause

(

operational or reactor safety problems.

l I-5

l l

i I

I i

l l

Heterence.,

l l

\\

I 1.

Nelcen, G.

A., "Bevisicno to Pub 11ched Chart; Metalc for High Pressure Hydro-j genation Planta," Journal of Engineering for Industry, Trans. ASME, Series B, 6

j Vol. 81, 1959 t

t 1

1 I

l I

i, I-l l

\\

l l

I i

i i

[

I e

p i

l l

I-6

i l

l

(

l II.

!ECHANICAL CONSIDERATIONS i.

l l

i Mechanical Considerations l

i

[

i An analysis has been conducted on the effect of the wear areas and penetration I

I into the base metal on the mechanical safety of the vessel during its subsequent 4

life. The minimum thickness of the bottom head tase metal as specified by the j

design drawings for the vessel is 3-7/8 in. not including a resistance velded i

I stainless sheet cladding of.109 in, thickness. The minimum allowable Section VIII thickness requirement for the head is 3 774 in.

l A replica of the worn areas has been made, that shove a maximum penetration from the surface of the cladding of.115".

i Inspection reports of the as-formed lower head show a minimum total thickness of l

h-7/32" or 4. 219". This gives an original A302 B thickness of L.110 after sub-tracting for the.109 thick stainless steel. A penetraticn of 0.016" into the l

carbon steel including.010" of general wear would reduce the A302 B thickness to 4.09h" giving at least 0.3" of wall thickness for corrosion before reducing the vall below the 3 77h" minimum vall thickness according to Section VIII of the ASSE Code, and less than 3" required under Section III.

i L

A similar analysis has been made for the vessel area behind the thermal shield.

l The plate thicknesses as received from the mill for the two shell courses in this area were 8-1/8 although the drawin g called for 7.875" minimum thickness. The re-quired per Section VIII is T.8015".

Note attached mill test reports.

I The effect of the stress raiser caused by the wear mark on the head was also eval-i uated. The stress raiser effect of this wear area on the head is dependent upon the contour and geometric shape of the actual depresssion in the base metal.

A fatigue analysis has been conducted assuming various stress concentration factors, to determine the maximum stress intensity which can be tolerated at this defect.

Based on the evaluation of the material in the bottom head under the defect by EAPD chemistry and metallurgical personnel the material at the defect is consid-ered to remain ductile throughout the life of the plant in the fatigue analysis.

The following pressure and temperature transients occurring simultaneously were j

utilized in this analysis.

i II-l

Transient Duration No. of Cycles oTemperature APressure 1.

Heatup100F/hr (4.3 hra) 200

-T F 2500 2.

Cool down 100 F/hr (4.3 hrs) 200

+7 F

-2500 3

Plantloading48F/hr (0 5 hrs) 15,000

-3 6 F 4.

Plantunloading72F/hr (0 33 hra) 15,000

-5.4 F 5

+10% step change in power instantaneous 5,000

+12 F

-130 6.

-10% step change in power instantaneous 5,000

-12 F

+110 T.

Scram accident instantaneous 50

-60 F

-90 8.

Normal scram instantaneous 1,050

-32 F

+550 9

Steady state power instantaneous infinite

+6 F

+100 fluctuation The above transients chosen for this analysis are conservative estimates covering all modes of normal and emergency operation believed to be significant. The num-ber of times each transient is assumed to occur for the fatigue analysis is based upon conservative estimates of the number of occurrences which can be anticipated for each event. The actual number of experienced transients is expected to be significantly less than the above number of transients as borne out by past experi-ence at Yankee Rove. The "d temperature" column in the above tabulation is the actual difference between the mean temperature and the temperature at the inside surface of the bottom head for each transient condition.

The stress at the bottom head defect area was then calculated for each transient condition. Note that the stress is calculated as a function of stress concentra-tion factor (K), code allowed vall thickness to compensate for any loss of vall thickness due to corrosion, pressure fluctuation and temperature change.

l MI MI Transient Pressure Stress Thermal Stress Cycles 1.

Head up 18,900 K

-1860 K 200 2.

Cool down

-18,900 K

+1860 K 200 3

Plant loading

+1080 K 15,000

-1620 K 15,000 4.

Plant unloading 5

+10%powerstep

-985 K

+3600 K 5,000 6.

-10% power

+838 K

-3600 K 5,000 T.

Scram accident

+1J50 K

-18,000 K 50 8.

Normal scram

-680 K

-9600 K 1,050 9

Steady State T T55 K

+ 1800 K infinite II-2

l I

I l

1 Various values for the stress concentration factors were then applied on a trial 1

and error basis to establish a peak combined stress for each transient condition.

(

f l

For each value of assumed stress concentration factor, the minimum and maximum I

values of fluctuating stress were determined by combining the transient stress conditions in a manner to establish the maximum stress difference for the maxi-mum number of predicted cycles.

In this manner it was determined that a theore-r 1

tical ctress concentration factor of 9 may be tolerated and still maintain the i

1 fatigue usage factor nt the defect below the allowable value of 1.

The contour of the defective areas has been determined by making a replica.

It I

is clear from the cor. figuration of the replica that there are no cracks or sharp stress raisers at the surface of the defects.

Based on the dimensions of these t

defects these can be considered as holes thru the cladding thickness and shallow grooves in the base metal. Since the hole geometry and proximity effect of the i

defects are restricted to the cladding and we are no longer concerned with the integrity of the cladding, the resulting stress concentration factors in the clad-ding which are less than 5 could be eliminated from consideration.

Even if the factors were considered as applicable they would still be well below the value of 9 shown to be acceptable.

The minimum groove radius found in the defect was 1/8 inch which would result in a maximum theoretical stress concentration factor of 3.1 in the base metal which is well below the allovable value of 9 It is therefore concluded that there is no concern over the mechanical aspects of cladding penetrations in the Yankee Beactor Vessel.

L l

l l

I I

)

II -3 i

~

l 1

i i

c I

m

-ac-

-ar

-.. -.. r - u c- ' I.a L*n c f

.<E i ML

- r-.

L VTT xD 1 L s. ur 4f J

=.

6 i1;.

I k

i i

i l

A.

Introducticn and Summary 1

corrcaicn queeticn p: red Ic the existing penetrations in the vessel The i

clad has two facets, namely 1

I 1.

the evaluatien of correei:n of the veccel metal in direct centact i

j with the ecolant at the clad penetraticns; I

l 2.

the evaluaticn cf corrcsien cf the vessel metal under the clad to recult of the penetratione I

which s+ agnant ecolan+ has access as i.

and the disecntinucus nature of the clad to vessel welds.

(

i "hece questions must be evaluated at both operating conditions and shutdown cenditiens. An additicnal factor of great significance, which must be ecnnidered, is the fraction of n genetnted by corrosien which will be i

retained in or abscrbed by the bare metal under these conditions.

I l

i The caolant at cperating conditienn will be a water solution of the following composition at about 500 F:

i Barcn O to 1500 ppm as H.EO, 3 3 NH, O to 10 ppm 3

30 cc (STF)/kg H,0 (~2 7 ppm)

H

~

2 c

O t 0.1 ppm g

l C1,

< 0.1 ppm c

At shutdown conditiens, the bulk coolant will be at about 100*F or less f

with the following compcsition:

i Baron 2200 - 2h00 ppm as H,E0_,

{

3 3 H

trace due to radiclysis 2

0 as established by radiolysis of the water and ecntact i

with ateccphere, ~ L - 8 ppm

[

i i

H,0 as established Ly radiolysis, ~ 2 - 4 ppm 9

u u l

[

i i

III-l i

l l

P

.i

i l

l 1

I i

NH 0 - 10 ppe depending on shutdown procedure, but of no 3

i great ecnsequence i

e e

C1

< Q'l EE=

2 l

The foregoing specificaticas apply generally to the bulk coolant to i

which the vessel metal at the clad penetrations will be exposed.

During

(

chutdcun, the coolant will be essentially stagnant. Durir.g operatior., the coolant will be flowing, but the exact flow conditions at the clad I

penetration at the lcwer vessel hemisphere cannot be precisely defined; ficv eddies and turbulence will no dvubt exist, but an equivalent flow velocity of five feet per second seems a reasonably conservative estimate.

t Special consideratiens must be given to the conditions within the clad-vessel gap.

Clearly here the ecolant is stagnant. Accordingly, corrosion data pertinent to stagnant rystems apply; furthermore, changes in bulk coolant compositicn are not immediately reflected in tk gap (and conversely) being controlled by diffusion processes perhaps. The latter factor is of special importance with respect to the oxygen and hydrogen percxide content of the gap coolant at chutdown conditions. While some oxygen and/or hydrogen peroxide might be generated initially by radiolysis during shutdown, as these compounds are ecucumed by the corrosion process, an excess of H wi1l 2

necessarily accumulate in the gap.

The continued accumulation of H2 "ilI then either cuppress further radiolysis, or perhaps accumulate to such an extent that sufficient gas pressure will be generated to force most of the soluticn from the gap (cnly a few at=cspheres pressure would be required at shutdown).

In either case (0 free, or gas, atmosphere) corrosion would be 2

insignificant as will be shown.

The fraction of corronicn-genertted hydrcgen which will enter the carbon steel will be largely gcrer. _ ; by two factors, namely:

(1) the point of generation of the hydrogen, and (2) the relative rates of effusion of the

(

hydrogen from the surface and diffusion into the steel.

l l

III-2 l

i

l l

t

]

1 i

1 i

(

"he Point of Hydrcgen Generaticn - This will depend cn l.

i 1,

i whethe: the corrcEicn is largely galvanic in nature, or due i

to local electrochemical action. That is, in galvanic i

.I corrcsicn, the corrcding metal is 'inadic and the tcre ncble metal, in this cace the stainless steel clad, serves as the f

cathode.

In the cystem under considerati n, the ancdic reaction

)

would te I

i I

++

i l

+ e electr:nc

[

re re j

I

( subse quent reactions leading to the formation of Fe;0, need I

not be ccnsidered in this discussicn), while the cathodic reacticn i

i could be either or both cf the follcwing, depending cn the 02 ccncentraticn:

?

l 2H + 2 electront -) EH (surface )-) H2(

lant) (effusien)

[

i l

2H (dissolved in SS) (diffusien) or H O + 1/. 0 + 2 elec t r:ns --> 2 CH- (coolant) 2 2

I In the forter case, it is clear that the corrcsion generated r

hydrogen is found en the stainless steel surface and is cnly l

available for inccrporaticn in the carbon steel to the extent l

that it dissclves in the ecolant; this source will be shown to be af minor significance.

In the latter case, the corrosien process I

l consumes 0 rather than generating H hence is of no ccncern frc=

2

' tis pcint cf view.

j i

Clearly, at shutdown the 0 censuming corrosien prccess will 2

centrol corrosien at the vessel steel directly exposed to the I

i l

0 rich ecolant at the penetration.

In the gaps, hcwever, as I

2 I

noted above, O availability wculd be 1cw and the H generating l

2 2

i i

precess would tend to occur. To the extent that cerrosicn l

continued to be galvanic, hcwever, H in rporaticn in the carben 2

steel would be negligible. Actually, however, it is found that the i

stainless steel surface quickly polarizes (i.e., the formation of

(

F l

l III-3 e

9 4

1 4

i l

H at the stainless steel surface essentially stcps) and galvanic 2

j corrocicn ceases. Any further corrocicn occurs by local electro-l chemical action en the carbon steel itself.

In essence, the cathodic (H generating) renetion cecurs at local active sites en the carbcn 2

j steel, hence the H atocs are found at the carbcn steel surface and j

are potentially available for diffusicn into the steel.

This is the 1

l ccnditicn which w;uld prevail in the gap during shutdown and both in i

the gap and at the penetration at cperating conditicns.

1

)

2.

The Eelative Eates of Effusien and Oiffucian of Ccrrosien - Generated l

l H at Carbon Steel Surfaces

- Hydregen dissolved in steel is in the l

2 1

atetic rnther than colecular fonn; molecular hydrogen rapidly effuses l

frc= the steel surface to the envircncent. Accordingly, the relative i'

rates of effusien of corrocicn generated hydrogen (which is atomic) i j

frc= the surface and diffusicn into the metal will be determined in i

part by the rate of combination of atcmic hydrogen to form molecular hydrogen on the surface.

In general, =clecule formation is an extremely l

l rapid process; metal datage due to corrosion at moderate te=peratures l

(blistering, etc. ) has cnly been cbserved in systems wherein the

[

corrodent contained a component which inhibits colecular for=ation.

(

6 Sulfides are the best known inhibitors.

In the absence of such an inhibitor (such as in the Yankee coolant) cost of the corrosicn generated

[

hydregen is emitted to the envienzent as gas.

Ample documentaticn exists and is presented to justify the conservative r

assumption that cn the average only 2QS of the corrosien generated

[

t H will enter the carben steel. This factor, alcng with corrosien data, 2

t is used to evaluate the H ntent of the vessel in Section IV, the 2

effect of which is discussed in Secticn V.

With the foregoing considerations, existing informaticn on the corrosien i

f of carbon steel is reviewed in the following sections resulting in conservative l

estimates of the expected corrosien of the Yankee vessel under the various

{

ccnditions as summarized in Table 1.

{

l l

l III L l

l i

t i

Y t

I

.~...

l 1

i i

i 1

2 I

- t s.m..;--.

1 i

s_

l 1

i i

,.1. c. h.

M.

t

i. 4.,;..i-m.+ 4,

m.. d..a.

t ** D ' ' t.

m--

's%a y nAL....

J t

4 4 9 4... r. e

/*y

~

.,u

... t.. a 3, I a." _. ~

cJ

,,r..,,,.,

,,;... ~.j s

i.,

1 c,

.. c -. -

.., 3

. c, s

._e.

4.

,3.

)

l s

a....w...

,c m.

v,

l l

)

.s........ c.4 _ e,e.r.o. v 3. g.a.

1.

7.,

o.. c_. c -

y.,.....,m.

.m...

g.

c,,. 4 n... -.

..s m

t t

2._

e. 4. c

w.. - e..,..,-., 3,.~ m

c.. n a..

.2.

.n.

t..y...

u...

,r e, a,o4. s- -

y.

r..,t n..on-r.. 9. - -

r. n.. -a t,1,..... 4.., n.m. _;. m, e...m....:,

,,.m_...

si.

s i

I

. a t. _ c i... m.1 ~_4 a '.. - + ".. s.' ~. a.. i c u. 4

3. r.a_. a 2' l

, o_

,.e c,..., 4.,-..

c4 ~.

. e a.

.m m

I, 3,_3. r., e e.s " a += '..

l n a 4 4,,..,1. 2.,.

4

. 4..

.c 4 r.c.,.

.u.-n

,, 3

,.._,e_.-1.

e m.a m

., e g.....e e

.t., a n t a.-

i v

u..

4. s.. s a, y y. v, y 3, + 4 r..,,

m..

m.

.t.

4.

r_ c, ot

,s, e n a......

a.. a,..,

..c.

c +us. 0.. e s s e, + t=. M. a m.i "..e- ('.' ra e-4 i.

. > +- s.. m.

'e m. o.

..s.

+ >,.

wr+.

e.,v.rs + 4...

+; -

n m m. t>

n

/,

~. i u. 4.-. s a.

,. 4

.c o'__,

/ 0. s~.,;-

.,. 4, c...

az,e rm t-c

+,

c t.

.,,,.2_,

j m

+ -

.. 1 m..

m m,..e,% '.._, [..,

.q, c e,, h, m e. y., a i.r.e.

.i 2 )

v,_m, 4...

,c

p. 4 4,a c.

a c m.

_.f,,.

m n..m l

c.,).c..,-.

. 4.%.....

)' b..i^ s.**

cc..e ~* a'i ~.

s^.

.' x' u i. d l

a.

4.

c,.4a,

.,.g.,

o c.p c.

c w

n

&c.. a.,)

n.

cv-o <.w..,n r 1.,,.

+m>. o. x,,

.e

..e o

._., - u. a..a e c_c

+...,4..

..m e.c c.e..-

.s m

.m u.1 v,.-. cego..

j 3'

... i. 4 m.. s.i

.s.c

, a p. c.,~-.. c,.. - ma y

~

mc., ug m...e

.m a _m s

3. _.,. ~3..,,. e 6F.

k'

.. c.. ~. c

.m. n

,, c 4 ec.

s I

-. 1., a,_ 4.,a..

-..a,,..,,.4.

a.,=.,.,,..,c,. v

n,.

n ~

(.r._, -.. m..,-.. a. v. a, c.m,

+o b'.;

I m:

s t

I a - s,,

s r

1

.;u. v y

.s. i. c r. )

3.0v,

m. u, g.,.4 t...o.

l pr s

0xyge n deaersted I Onditi nr'

. ~/3'~~

'.'. e. r.r ' e i n.r.s +m.c, ~3 a

m.m t.re 1.,, 4 +.".

t

,c.,'.

__ n' c-s-0

.cm.c

.y A r L

p l

e

*.

ce I.xp=ed ! rpy : tqcsed i Descaled Exrcsed,rpy Exposedj rp:.

~.

c heurn

..3.0

- 2 ~-,

.-c v.c-

.u G8k hnurs i

- 25L

- 230**

1.5 l

r'n+

c' "u. r. u* 4. c.,

~.#

t o-o c i ~..

.n,.n..

4...>d

- "4 ~

1

- n.-

m.

r b

. a +. e

'a' < *s.

t a 4 am

..s 3

..m Ve l.w i t'.

.-f Ecric Acid Scluticn, frec 'a' AFT-CF-D2.

< s. a _l a. e + a m.a. u 40. a 1' l.4a.

en.m~.a 4

.* It e sn. *' <' c e.. + F,a*.

+m>.a, c-4-

4

~-~a

. +,

s^..#

~+

o c

mu

.n. m 5

Co I

g,y" "ypgu, Ira h I1 g .*

g*

• b.. { /, 9.,* v$ ^m.c 4. g,.'.

ac. t n r-4 4 + g.,*= 4.. m a

.e.r.

c3 gy*y*s^w } g {g,b.+s

.-4,.

v s s u s. u change.

r. r r

'-)

s

. ~... _ _

l l

l 1

i l

[

.i 2.

Limited Access Corrosion Data a.

Galvanic and General Corrosion i

i The corrosion rates for carbon steel in semi-stagnant or stagnant l

fluid are presented in the following table. These velocity conditions would simulate the exposure of the carbcn steel under the cladding surrounding the hole in the Yankee reactor prescure vessel. The corrocion ratas again represent both galvanic and general corrosion 7

due to the method of testing.

Temperature 650*F /2a 650*F /2a 600*F/1 600*F/_2 l

1 2

Flow Velocity semi-static semi-static semi-static semi-static Test Duration 28 days 28 days 30 days 120 days t

Boric Acid 0

0.1 molar 0.07 molar 0.1 molar l

0xygen

< 0.1 ppm

< 0.1 ppm

< 0.1 ppe

< 0.1 pps

(

Hydrogen none added 100 ml/kg 100 ml/kg 100 ml/kg l

Exposed Weight Change

-21mg/dm

- 5 mg/dm

- 12 mg/dm

-100mg/dm max.

2 F

Descaled Weight Change 139mg/dm 98mg/dm Average Corrosien Fate lh9 mg/dm -mo*

105 mg/dm -mc*

12 mg/dm -mo**

25 =g/dm -mo**

Fanetration Fate 0 9 mpy

• 0.6 mpy

• 0.08 mpy **

O.7 mpy **

f i

• Descaled Corrosion Fate Corrosion Pate Determinui From Exposed Weight Data and is Low Due to the l

Weight of Oxide on the Specimen I

i 3

Interpretation of Data a.

Geneml and Galvanic Corrosion A test time of about 10' hours, an average boric acid concentration of about 0.05 M, a temperature of about 500*F and a flcv velocity between O and 10 feet per second would be required to simulate the exposure con-ditions of the carbon steel base metal at the hole in the stainless steel cladding of the Yankee pressure vessel. Since mmt of these conditions l

l are less severe than those reported for free access corrosion in the i

2 i

first table, the corrosion rate would be much less than the 497 mg/dm,,,

for the 36 fps flow velocity and probably near the extrapolated 28 mg/dm -

l l

Ill-6 l

i

j 1

i I

l l

l I

l mo for the 2 fps flew conditions. A reasonable maxi =um value at 10 fps for the corrosicn of the pressure vessel base metal at the hole l

2

~

in the cladding would be 150 mg/dm -mo for a penetration of 0 9 mils / year. The corrosion of the base metal carbon steel

~

under the cladding and surrounding the hole in the cladding could be ecnsidered cecurring in a stagnant area and therefore the cemi-static tests reported in the table above should be I

applicable. The corrosien rate again is dependent on exposure j

i time and temperature but not nearly as dependent en fluid boric l

1 acid concentratien as at the hole since the ionic movement would l

l be diffusion controlled in the crevice. The carbon steel cor-I rosien rates reported in the semi-static tests ranged from j

149 mg/dm -mo to abcut 30 mg/d=2 2

-mo.

One can assume that the f

l maximum galvanic and general corrosion rate in the carbon steel i

area with limited access to fluid would be 150 mg/dm -mo for 0 9 l

mils / year penetration or the same as for the free access corrosion.

L b.

Crevice Corrosien i

The effect of crevices cn corrosion is reported not to be a r

problem in systems with low oxygen contents on the order of 0.lk f

ppm.[3a Since the Yankee reactor operates with oxygen levels below f

this value, the effect cf crevices should be regligible during hot

(

l cpe rati on.

l c.

Stress Corrosion Cracking i

No problems exist for the stress corrosion of carbon steel since no high alkali concentrations will be present or any nitrates which can cause stress cracking of carbon steel.

l d.

Production of Corrosion Products i

i Assuming the maximum corrosien rate for the carbon steel of i

150 mg/dm -mo and the metal release rate equal to the corrosion rate and 1.h ratio of corrosion product / metal then the corrosion product release rate would be 210 mg/dm -mo.

In the Yankee i

l III-7 i

6

-. - _ _.=..

l i

I i

L t

.l i

l pressure vessel bott cm head the total area of carbon eteel that i

i would be exposed by a hele through the cladding if no resistance I

welds were precent wculd be.42 f t or 297 dm Since only 30%

l l

of this area ic unbended; then the total corrosion product re-leased per month wculd be 18.7 gms/ month.

t I

t 1

r I

t I

i-I l

l 1

t I

j I

I i

t i

l t

b l

l l

t t

I s

li I

e f

III A

(

t i

1 i

i r

i l

t C.

Reactor Chutdown Conditions 1

1 1.

Free Access Corrosion Data e

[

i a.

General Corrosien f

Carbon si, eel corrosion results obtained at 200*F in non-boiling f

~

beaker tests containing 0.16 molar (1730 ppm B) boric acid solution vere reported by Howells and Vaughan The exposed weight change van 1565 mg/d=2 for a total test time of 450 hours0.00521 days <br />0.125 hours <br />7.440476e-4 weeks <br />1.71225e-4 months <br /> resulting in I

2535 mg/dm -mo.

Data obtained at WAPD indicated that the room temperature exposed weight change of carbon steel in 0.28 molar (3000 ppm B) i 2

l boricacidsolution(airsaturated)was315mg/dm for 216 hours0.0025 days <br />0.06 hours <br />3.571429e-4 weeks <br />8.2188e-5 months <br /> 2

testing. This would be 1050 =g/dm -mo or less than half the rate reported by Howells and Vaughan at the higher temperature.

b.

Galvanic Corrosion Data i

Couples of lead and carbon steel exposed to solutions of lithium and potassium tetraborate and lithium metaborate at 180 F contain-ing equilibrium oxygen from air indicated some galvanic corrosion occurred on the carbon steel The amounts of this corrosion are shown below for 1000 hours0.0116 days <br />0.278 hours <br />0.00165 weeks <br />3.805e-4 months <br /> testing.

Solution Uncoupled Coupled t o Lead 2

0 5 W/0 boron 10mg/dm 280mg/dm as potassium tetraborate 2

O.5 W/0 boron 75mg/dm 590mg/dm as lithium tetraborate 2

0.5 W/0 boron 30mg/dm 3ho mg/dm as lithium metaborate i

l Galvanic corrosion tests of 43h0 steel electrically coupled to b n pure water, with oxygen present, stainless steel i

indicated that galvanic corrosion was the predominant form of attack. This did not appear to be true in high pH solution (pH-10 with lithium hydroxide). The results of these tests are presented below.

l T -9 4

vw--e-

l i

I

(

Galvanic Corrosion of 43h0 Steel Coupled to Stainless Steel l

Test Conditions Test No.

Galvanic Attack Total Attack

[

Temp. - IkO*F l

LiCH - None 2

2 02-0.5 ppm 1

545mg/dm 627mg/dm Duration - 90 days r

Temp. - 180 F l

L10H - None 02-2 ppm 2

687mg/dm 645mg/dm Duration - 14 days Temp. - lho F LiOH - 6.9 ppm F

2 2

02-0.5 ppm 3

12 mg/dm 20mg/dm Duration - 82 days The room temperatare corrosion rate observed for carbon steel specimens directly coupled to stainless steel in an air saturated 0.28 molar boric acid solution (3000 ppm B) for 168 hourr at WAPD l

2 vas 29 mg/dm -da (0.43 mila penetration per month). Uncoupled carbon steel specimens exposed to identical conditions corroded at e

35 mg/dm -da (0.52 mils penetration per month).

t c.

Crevice Corrosion Crevices appeared to have little effect on the total corrosion of carbon steel in air saturated-room temperature solutions of 0.28 molar (3000 ppm B) boric acid solutions as indicated by a WAPD test where the general corrosion rate and crevice corrosion rate sere nearly identical at about 870 mg/dm -mo after 216 hours0.0025 days <br />0.06 hours <br />3.571429e-4 weeks <br />8.2188e-5 months <br />. This is further confirmed by tests in oxygenated water at 180*F for 1000 2

hours where the corrosion rate in the crevice was about 550 mg/dm and the general corrosion rate was about 800 mg/dm -

2.

Limited Access Corrosion Data a.

General Corrosion Corrosion experiments performed at WAPD on carbon steel exposed to a 0.28 molar boric acid solution initially containing air but l

I 1

• -20 l

I t

vacuu degassed for 30 minutes at about LOO *F irdicated a cetal attack rate of 133 rg/dm for a 13 day test. At this rate of attack the corrosien rate would be 300 eg/dm -=o or 0.15 mils penetration per conth.

l b.

Galvanic Corrosion i

Carbon steel coupons were coupled to stainless steel and

(

1==ersed in a 0.26 colar boric acid solution with the air removed

(

I by vacuum degassing. These couples were tested in this selution l

for 13 days with a resulting corrosion of 1h7 mg/d: or nearly the f

t same as the general corrosion for the same period, l

i 3.

Interpretation of Data f

L a.

General Corrosion I,

The general carben steel corrosien data indicate the corrorien l

I rate expected in the free access area (hole in clcd) would be l

2 2

between 2535 mg/dm -mo, determined at 200 F, and 105C =g/dm,=o, detemined at about 75'F.

A reasonable maximum value for this I

corrosion at Yankee conditiens of ~ 100 F would be 1500 mg/dm -mo resulting in a penetration of 0 75 mils per tenth.

l Although the corrosion rate for carben steel in the low cxygen 2

solution of boric acid was observed to be 133 rg/dm ir. 13 days, i

the rate of attack decreased drastically after the first day as I

determined by visual observation of the hydrogen bubC.es released frc= the steel-water reaction. This large initial attack in a short test would greatly increase the average cerrosicn rate. A reasonable corrosien rate for a 2 month period would be about 2

150 mg/dm -mo if the assu=ption is cade that 65% of the total corrosion occurred during the first day due to oxygen depolarization from some oxygen remaining in the solution or on the natal surface.

Thusthe133mg/d=2 corroded in 1.~t days would include a corrosion attackof87mg/dm for 1 day and 46 mg/dm for the renaining 12 I

2 days resulting in a corrosion rate of 3.8 eg/dm -da.

This would i

t i

l O

-11 i

I I

i

result in a penetration of 0.15 mils in 2 =cnths or an average penetration rate of 0.8 mils per month.

b.

Calvanic Corrosion The low te=perature galvanic corrosion rate of carbon steel coupled to stainless steel in pure water in the presence of oxygen appears to be nearly equal ta the total attack rate. The same situation appears to exist in boric acid solution where the general corrosion rate near the galvanic cell is decreased to a low v.lue.

This is possibly due to the entire carbon steel specimen becoming quite anodic thereby eliminating the normal local anode-cathode areas on the specimen itself which contribute to the general corrosion attack. The taximum total corrosion (galvanic plus general) vculd be nearly equal te the separate general or galvanic corrosien mtes which would be 1500 mg/dm -to or 0 75 mils per month pene-tration. The galvanic corrosicn of carbon steel coupled to stain-less steel in boric acid solution in the absence of cxygen (limited acce, s) chould be quite low due to polarization of the cell from lack of oxygen. Thus the galvanic corrosion rate of carbon steel in the crevice should be low and the total corrosion (galvanic plus general) would be about equal to the general corrosion rate of 150 mg/dm -to resulting in a penetration of 0.08 inches per month, c.

Crevice Corrosion The total corrosion rates reported in the previous section include crevice cerrosion since no acceleration of corrosive attack was observed due to crevices, d.

Stress Corrosion Cracking Stress corrosion cracking of carbon steel at low temperatures is not considered a problem.

e.

Production of Corrosion Products The total corrosion product as metal fr= a L x 8 foot section (total area exposed from hole in cladding) for the assu=ed corrosion rate of 150 mg/dm -mo in the crevice would be 45 grams /conthor c

. $2

f c. r iOf mf the

• otal area (n n-bend"d area) 13 5 grama/= nth.

The rntril2t inn f re. t he carbcn et eel corracicn at the hole would be negligible.

D.

N-

r-of Fvdr gen Felease1 by Corresion s *

.c a ' ?rpti'n

-f hydrcgen intr a met al due to corrocton of the

• • e t il i n :1 m ttt er of c ncider.ible inductrial significance. Where corrosien rates are high, and conditions are favorable for the re-tenti n - f t ho relewed hydrcgen in t.he metal, deleterious effects en the metal are cbrerveci even at law temperatures.

Evanc/7, pre-cents a comprehensive review cf these problems. Generally speaking tk.ere are

• hree c igni ficant factars which lead to high hydrogen con-tente in correding met als.

These are:

a.

Acceleration of enrrccien leading to increased source of hyd r <: gen.

su f> ice, which retards the for-b.

Pnicrning of 'N n

4 mat ien of molecular b a: rgen and desorpticn of the hydrogen generated, c.

Physical c;ntairmen' cf dissolved hydrcgen by impervious curfaces cuch as the tin of tinplate.

In the precent instance none of the three factors which are usu-ally

• spensible for industrial di

'culties are operable.

Pather, we are here cencerned with the fraction of corrosien hydrogen which will enter the pressure vecrel steel during well defined and un-agreceive conditicns. Beycnd the obv.cus fact - (based en visible relesse cf hydrogen) that certainly less than 100 percent of the cerrosien hydrcgen doen enter the steel, no quantitative recults s p ;.,

1 be avatlable for the specific conditions of interest.

Measurements reported in the literature on the transport through eteel of corrocicn or c c r -.v*ically released hydrogen by Morris, H e it h,"/c and DeBeer and Fast'/10 are typical. All of these studies involve trancpcrr through thin membranes of steel. Although not usually stated by the authors, it is to be expected that the hydro-111-13

l gen accumulation in the samples employed, would be quite low because of the high permeability of the thin samples. Thus, the possible effect of the hydrogen accumulated in the steel, on the ability of the hydrogen to enter the steel is not reflected in these studies.

In contrast to the above, Hudson!

found, that on totally immersed

~

specimens at 100 F the average fraction of the corrosion hydrogen retained in the specimen decreased markedly, as corrosion progressed.

His results were correlated by the expression

^

Log 8

~

10 10 where Y is the average percentage of corrosion hydrogen in the speci-men, and X is the total corrosion, milligrams per decimeter square, l

The relationships obtained were essentially indpendent of acid (H SO hcl,H #0 ) used.

rr si n rate, as varied by using various j

2 4 3 4 acid concentrations of 0.05 normal to 10 normal,and total corrosion of20.to2000mg/dm. A single regression line, with little error, l

can be used to represent the complete set of data for the non-oxidi-I zing acids. The results obtained with HNO3 ( **

"8 siderable lover than for the non-oxidizing acids, as would be ex-pected, since the hydrogen formed per rdt of corrosion vould be lover. Of note also is the fact that the fraction of hydrogen re-tained, all other factors the same, decreased up to 194 F, the high-est temperature at which measurements were made.

In seeking an explanation for the remarkable consistency of such a vide we of data, attention was directed toward the effect of nydrogen accumulation on the instantaneous fraction of corrosion hy-drogen entering the metal. It is to be expected that as the hydro-gen content of the metal increases, the surface concentration of hy-drogen atoms vill increase proportionately. A discharged hydrogen atom vill therefore have a higher probability of forming a hydro-

. gen molecule, on the surface of a metal with a high hydrogen con-centraticn, than on the surface of a metal with a low concentra-tion. A transformation of Hudson's relationship,. has been obtained III-14

which expresses the fraction of the instantaneously released hydro-gen, which enters tne steel, as a function of p, the concentration of the hydrogen at the surface of the steel, ppm. Diffusion cal-culations chov that the hydrogen concentration at the center of IIudson's specinens differed from the surface concentrations by less than a factor of two in one hour, and by less than two percent in one day. For practical purposes therefore, the average concentra-tions can then be taken as the concentrations at the corroding sur-face, with little error.

The transformacion enployed is as follows:

Iludsons equation rewritten as a function of F, the fraction hy-drogen absorbed, and P, the hydrogen equivalent to the total cor-rosion, is log F = -0. 21 - 0. 62 P or FP.62 0.616

=

Now this can be rewritten as,

-0 32 (FP)P

= 0.616 or FP = 0.616 P.32 0

Differentiating with respect to time, ve obtain d(FP)

(0 38)(0.616)P

.2 g

=

dt dt But the time derP.ative of the hydrogen content is the in-stantaneous rate of accunulation of hydrogen, equal to the instan-taneous fraction of hydrogen absorbed, f, times the corrosion rate, dP dt Therefore

~*

G.235 P f

=

and

~*

f = 0.235 P To express this relationship in terms of the instantaneous hydrogen III-15

content p, we use the relationship, that F is equal to p/P, and is given by F = 0.616 P Thus y =

0.616 P ed y

07 0.616 Substituting we obtain

-0.62 0.235 0 3d f

=

O. 16 which reduces to

-1.63 0.1062 p f

=

Typical results are P(Ppm) f 05 0 33 1

0.106 o

2 0.034 For the range of hydrogen fractions calculated in Section IV, it will be seen that an average fraction f, of 0.2 is conservative.

III-16

E.

Evaluation for Yankee Eeactor Pressure Vessel a.

Total Corresten Penetration The maximum total corrosion penetration rate at reactor operating conditions would be the same both at the hole and the crevice sur-rounding the hole. This rate would be 0 9 mils per year.

For a lifetime expectancy of 25 years this vould be 22 5 mils.

The maximum total corrosion penetration rate during shutdown conditions would be 15 mils at the hole for a 2 =cnth shutdown and 0.15 mils in the crevice surrounding the hole. Assuming 30 shutdowns durin6 the pressure vessel lifetime, the total corrosien penetration at the hole would be 45 mils. The maximum penetration for the pressure vessel lifetime would then be 45 + 22 5 or 67 5 mils.

This 67 5 mil decrease in the thickness of the carben steel base metal in the pressure vessel vall is much less than the 0 3 inches which was determined to be the maximum decrease allovable for corrosion before reducing the vall below the 3 774 inch minimum vall thickness required according to Secticn VIII of the ASME code.

b.

Contribution of Corrosien Product to Crud Levels and Coolant Activity The maximum corrosien product released per month at reactor operating conditions and at chutdevn conditions vould be about lk grams /

month. The crud released per month from the stain 13ss stcel expose.

to Yankee primary coolant conditions at a release rate of 2 mg/d=2 -mo as metal vould result in 1400 grams of metal per month.

The contribution of the 14 gram / month (1%) to either the crud level or coolant radicactivity is negligible.

III-17 O

~_.-_- - _.-_

References 1.

WAPD C(PC)-18, " Corrosion of Structural Materials in Chemical Control i

Solutions", P. E. Brown (August, 1954), pg. 23 and 26.

2.

TID-1027 Technical Progress Report, " Pressurized Water Reactor Program for the Period August 26, 1954 to October 7, 1954," pg. 65 2a. TID-10028 Technical Progress Report, " Pressurized Water Reactor Program for the Period October 7,1954 to November 18, 1954".

4 3.

WAPD-CP-532-R-1, " Loop Operation with H 803 - High Velocity - Chemical 3

Engineering Section" (June 30,1954).

I 3a. " Corrosion and Wear Handbook for Water Cooled Reactors", Edited by I

D. J. De Paul (1957), pg.147 4.

Research Report 7301, " Corrosion of Reactor Materials in Boric Acid Solutions", E. Howells and L. H. Vaughan, (1960).

5 CERD-SIC-106, " Primary Shield Materials Study", G. P. Lewis (1956).

6.

WCAP-1844, "The Galvanic Behavior of Materials in Reactor Coolants",

D. G. San =arone (1961).

7 "The Corrosion and Oxidation of Metals: Scientific Principles and Practical Applications," Ulich R. Evans, Edward Arnold, Ltd., London, 1960.

8.

"The Diffusion of Hydrogen Through Mild Steel Sheet During Acid Cor-rosion," T. N. Morris, Journal of the Society of Chemical Industry, Transactions and Communications, 1,17(1935).

9 "An Experimental Investigation of the Diffusion of Electrolytic Hydrogen Through Metals," H. R. Heath, British Journal of Applied

Physics, 3,13(1952).

10.

"The Diffusion Von Wasserstoff Durch Eisen Bei Zimmertemperatur,"

J. H. DeBoer and J. D. Fast, Rec. Trav. Chim. 28, 984 (1939).

11.

" Hydrogen Absorption by and Dissolution Rate of Low-Carbon Steel in Sulfuric, Hydrochloric Phosp'; d.c and Nitric Acids Corrosion, R. M.

'Iudson, 20, #7, 245t (1964).

III-18

IV, liYDROGEN DISTRIBUTION AIT CC:!CI' CJC. (

In the next section, the potential effects of hydrogen embrittle-ment are to be presented and vill be shown to be dependent on the con-centration of hydrogen in the steel. This section vill present calcu-lations to show the amounts of hydrogen which may be formed or to which the Yankee vessel base metal may be exposed and the resultant concen-trations or pressures.

The first discussion vill pertain to the possible cources of hydro-gen to which the bace metal could be exposed. From these, it is pos-sible to compute the hydrogen concentrations and/or pressures using videly accepted technology concerning diffusion and effusion of hydro-gen through iron and steels. These values vill be comparad with the allovable limitations in Section V.

For the sake of completeness, some additional calculations have been c:ade, relative to the total hydrogen produced from corrosion com-pared to that contained in the coolant, and to the amount of boron that could be present in the void between clad and base metal. These con-siderations, as will be shown, are of minor consequence to future plant operation.

A.

Parameters Used in Calculations For the analysis which follows, certain parameters have been selected as applicable to the Yankee vessel. Where there is a choice, the most conservative parameter is selected, such as to give a " worst case" result. These are as follovc:

Veusel Clad 304S/S,0.109 inch Vessel Base Metal A302B Carbon Steel On Bottom 4 inch thickness On Side 8 inch thicknces Radius of Bott:,c 56.5 inches Surface Area of Bottom Hemisphere li 2

Total Hemisphere 2.0 x 10 sq. in., 1294 da 3

a 4 ft. x 8 ft. Sheet 4.6 x 10 sq. in.,

29T dm 3

e 30%of4x8 Sheet 1 38 x 10 sq. in.,

89 dm

-IV-1

Void Volume Between Clad and Base Metal l

(Assume 0.010 inch gap over 30% of area)

Total Hemisphere 61 cu. in., 1.0 liters k ft. x 8 ft. Sheet 14 cu. in., 0.23 liters Clad and Base Metal Temperature Hot Operation 500*F Cold Shutdown 100*F Bace Metal Corrosion Fates 2

Hot Operation 150 mg/dm,,,(=,x,)

Cold Shutdown Beneath Clad 150mg/dm-mo(max.)

Exposed at Clad Defect 1500 mg/dc -mo (max. )

B.

Sources of Hydrogen Hydrogen may be available to the carbor steel base metal from three

sources, i.e.,

corrosion, dissolved in coolant, and radiolysis. As will be shown, the only significant source vill be that from corrosion.

4 1.

Hydrogen From Corrosion Iron is known to react with water according to the following equation:

3Fe + kH O

) Fe O3 g + kH 2

2 The principal product, even in the presence of some oxygen is magnetic Fe 0. The amount of hydrogen produced, therefore, 34 may be computed from stoichiometry and a known rate of corrusion, i.e.,

-3 (Y,'mgFe/dm-mo)(10 gm/ms)(4/3 mol H /m 1 Fe)(2.24 x 10 cc(STP)/mol) 2 N=

2 2 2

(55.85gmFe/mol)(10cm/dm)(720hr/mo) p = 7.hk x 10' Y p = ccH2 (STP)/cm -hr Y = corrosien rate, mg/dm -co IV-2

2 vhere at 150 mg/dm -mo, the hydrogen production is 1.12 x 10-3 cc(STP)/cm -hr., or for 89 d:

exposed surface area, the total hydrogen produced would be 10 cc(STP)/hr. The corrosien rate of 150 mg/dm -=o represents that for the base metal beneath the clad at either hot or cold conditiens.

For the base metal 2

exposed at the clad defect, 150 mg/dm -mo is still representative for the maximum corrosion rate at elevated temperature. During cold shutdown ecclitions, however, corrosion of the exposed metal may be as high as 500 mg/dm -=o, as shown earlier.

In this case, however, hydrogen is released at the stainless steel surface

--.3a corrmir-

' r ely CC nnic It is reascnably assumed, therefore, that calculations of hydrogen concentrations in the base metal carben steel need not include galvanic hydrogen but be limited to 150 mg/dm -mo corrosion rates at all temperatures.

2.

Hydrogen Dissolved in Coolag j

Following standard recommended practice, Yankee adds hydrogen to the primary coolant, to the extent of about 30 ce(STP)/kg, for corrosien and radiolysis control.

This represents a source of hydrcgen to the carbcn steel base metal which will add to that produced by corrosien. The partial pressure of hydregen within the coolant at 500'F is derived frcm Henry's Lav where, P

H H

E 2

2 Thus, for an cfg 6200 psi /(mol/ liter),

a P"2 (6200) 30 ce(STP)/kg x ( (kg/l)

=

2.2h

s. 10 ce(STP)/ mole 8 3 psi

(= 0 57 atmospheres)

P

=

H2 conservatively assuming the water density is unity. The consequence of this pressure of hydregen in contact with the carbon steel base me'al vill be a n to be inc3gnifi m nt; vMn ecmpared t o that avail-aM e frce ~: racien.

IV >

i

l 3

Hydrogen From Endiolysis The possible increase in hydrogen at the surface of a base metal, due to radiolytic decczposition of water, has been considered by many, all of whom conclude this to be an insignificant source.

The mechanis=s of water decomposition by reactor radiation are J

very complex and, therefore, are not ful2y explained such that an analytical solution is possible. However, certain evidence would indicate that radiolysis need not be considered as materially affecting hydrogen concentrations at the steel surface.

It is a well documented rule that the presence of excess hydrogen vill suppress radiolysia and completely inhibit the net production of hydrogen and oxygen. This is the situation which exists for the clad and base metal interface. Hydrogen present at the base metal surface will undoubtedly prevent a further, net, buildup of hydrogen from radiolysis.

One may argue, however, that the primary product of radiolysis is the H atom, which, supposedly, vill be adsorbed to 6

a large extent by the steel. In conflict with this hypothesis, is the observation that tubing water exposed to high reactor fluxes, such as zirconium and aluminum, sh7e no increase in hydrcgen content, compared with similar sections held cut-of-pile. The con-clusion is, therefore, that there is no enhancement of hyd-ogen up-take due to reactor radiatien. Moreover, the metal areas in question at Yankee, i.e., adjacent to the thermal snield and the vessel bottom, receiveminuteradiationdoses,estimatedtobe0.03 watts /cc(max.)

and ~ 0, respectively, during full power operatien.

The most susceptible area for enhanced attack due (indirectly) to the presence of radiation would possibly be the base metal exposed directly to the reactor coolant during cold shutdown conditions. As a resul.t of residual radiation of core structure parts and fuel, while the reactor coolant is aerated, hydrogen peroxide will be formed which may accelerate corrosion of the exposed base metal. This attack, however, would be restricted to that area in direct contact with the TV-4

i coolant.

Areas of metal beneath the clad, even though vetted with coolant, could not be similarly attacked since diffusion of oxygen or peroxide would be very slow, and the direction of water flow within the gap, if at all, would be toward the defect from the cmall generation of hydrogen from corrosion, which would also serve to eliminate radiolysis in the gap.

C.

Hydrogen Accumulation in the Carbon Steel Base Metal The concentration, or partial pressure, of hydrogen in the base metal may be computed using either of two methods.

One method utilizes data rela-tive to the diffusion of hydrogen through steel, whereas the other is based on effusion rates for hydrogen through steel.

Both methods depend on some knowledge of the quantity of hydrogen availnble to the metal surface, and other constants related to temperature and metal composition. An analysis of the Yankee situation vill be given below using both methods to demonstrate the mnr N = hydrogen concentration that can result in the vessel wall.

For these computations, it is necessary to define the basis for se-lecting the quantity of hydrogen available to the steel surface. The only significant source of hydrogen vill be that produced by base metal corro-sion. All of the hydrogen produced will not diffuse into the base metal, but rather, the majority will be evolved to the gap between base metal and clad, as was shown in Section III.

Both steady state and transient analyses will be presented. At 500 F only steady state (vorst case) values vill be calculated using the conservative assumption that all corrosion hydrogen diffuses into the base metal. In this case it is shown that the maximum H n ntration in the 2

steel vill be 0.32 ppm at the inner vessel surface, decreasing linearly to zero at the outer surface.

At 100 F, both steady state and transient calculations will be shown. Using the experimental results presented in Section III on the frac-tion of corrosion hydrogen which enters the steel, it will be shown that at steady state conditions, the r h H e neentration $n the steel vill be 2

2 ppm, under which conditions only 3.h% of the corrosion H enters the steel.

2 IV-5

Transient analysis using the conservative assumption that 20'% of the cor-rosion H nters the steel show that the maximum H n ntration in the 2

2 steel vill be only 1.2 ppm after a four month shutdown.

A final consideration that should be Ir.ade is that although only 30 percent of the area beneath cladding is available for corrosion, the hydro-gen which penetrates to the base metal vill diffuse through 100 percent of the area. This assumption is discussed in the Appendix to this section, and is shown to introduce only about a 15% error.

Thus, with the considerations given above, the hydrogen flux at 2

thecarbonsteelsurfaceisasfollows(foracorrosionrateof150mg/dm-mo):

% = 1.12 x 10-3 ce(STP)/cm -hr x 0 30 x f

= 3.36 x 10- fec(STP)/cm-hr

= %.f where f is the fraction of corrosion hydrogen entering the metal.

% = total H gnerated/cm outer vessel surface 2

1.

Concentration From Diffusicn. Rates; Steady State The surface concentration of hydrogen, i.e.,

the inner vall surface at which corrosion occurs, is given as a function of the hydrogen diffusion coefficient, as follows:

Co (1) a o

D which assumes the outer surface hydrogen concentration is zero.

In equation (1) the quantities are 3

surface concentration of hydrogen, ccH(STP)/cmFe Co

=

2 f%

hydrogenfluxthroughsteel,ccH(STP)/cm-hr

=

9 2

X thickness of steel, em

=

D diffusion coefficient of hydrogrn in steel, em /hr

=

IV-6 4

The thickness of carbon steel base metal is nominally 8.0 inches on the side, or 20 3 cm.

The diffusion coefficient for hydrogen through pure iron is reported by Johnson to be 1.4 x 10-3,-3200/r Above 200 C, D

=

1.2 x 10- e' Below 200 C, D

=

6.8 x 10-E em/seeor0.2h5cm/hr Thus, at 500 F, D

=

-T 2

-N 2

3.8 x 10 em/seeor13.68x10 cm /hr and at 100 F, D

=

Therefore, the surface concentration of hydrogen crising from cor-rosion, ut.ilizing equation (1) and the constants given above, is 830fp At 500 F, 8 in.,

Co

=

3 At f# = 3 36 x 10-ccH(STP)/cm-hr, Co 0.028 ccH (STP)cm y,

=

2 2

0 32 ppm *

=

At 100 F, consideration must be given to the fact that not all of the corrosion H enters the steel.

In Section III it was shown 2

that

-1.63 0.1062C (2) f

=

where C is the H n ntration in pps in the surface of the cor-2 roding metal. Therefore, from Equations (1) and (2)

Co

=

o "D

11.4*

D from which 3

0.174cc(STP)/cm or 1 99 ppa at 100 F Co

=

2.

Concentration From Effusion Eates Data presented by Webb shows that hydrogen effusion through iron and steels may be described by 1

g, mPa

_q/g e

X 3

11.4 x Co, ccH (STP)/cm Fe.

ppm H

=

2 IV-T

where

% = hydrogen flux, ce(STP)/cm -hr permeationconstant,ce(STP)-=m/hr-cm-atm m =

Iqdrugen pressure, atmospheres P

=

metal thickness, em X

=

activationenergy, cal /mol Q

=

I For iron, Webb gives Q 144; thus, with x = 203 mm 9,100 and m

=

=

for the 8 inch thick carbon steel vessel vall, b

1 32 x 10~ P at 500 F

=

2.Th x 10-T/

o at 100 7

=

850, 16,100 and m For 304 stainless steel, webb gives Q

=

=

frcm which is calculated (for X 2 77 mm)

=

T. R x 10 M at M

=

33 9

1 57 x 10 P at 100 F

=

SS From permeation considerations, therefore, and considering that hydrogen diffeses through the carbon steel only, for f# = 3 36 x 10~

U 2 (STP)/cm -hr, at 500 F, the hydrogen partial pressure required ccH to sustain this flux would be 6.4 atmospheres, or 95 psi.

If, at 500 F, the effusion of hydrogen through stainless steel is also taken into account, the pressure of hydrogen then becomes about 2 atmospheres, or 30 psi.

~9 3 36 x 10 ccH(STP)/cm-hr, (taking f = 0.1 At 100 F, and with f/

=

2 merely for purposes of illustration), the hydrogen pressure required 4

to sustain this flux through 8 inch thick steel would be 1 50 x 10 atmospheres. If effusion through both the carbon steel and stain-less steel clad is considered for 100 F, the calculated pressure to

~5 sustain a total flux of 3 36 x 10 ccH(STP)/cm-hrwouldbenearly 2

the same as for carbon steel alone.

In order to compare the results of effusion calculations with those by diffusion, use is made of the solubility relation for hydrogen IV-8

in iron. Hydrogen solubility in iron is described by Sievert's equation, i.e.,

Co kP e

=

There is no data available for carbon steels, but it can be assumed that hydrogen solubility in iron and carbon steels have equivalent

~

values. Johnson gives values for the constants in Sievert't, equation as, 59.4 k

=

6360 E

=

and hydrogen concentration in iron, ppm Co

=

P hydrogen partial pressure, atmospheres

=

3h At 500 F, Co 0.132P and at 100 F, Co = 1.99 x 10 p

=

Thus, for values of hydrogen pressure calculated above, (considering effusion only through the base metal) 0.132x(6.4)I 500 F, Co

=

0 33 ppm

=

(1.99 x 10-3)(1,5gx10)k 4

100 F, Co

=

0.2h ppa

=

The values for hydrogen concentration at 500 F from diffusion and effusion calculations show excellent agreement. It is clear how-ever, that at 100 F, the values given by Webb and Johnson are in disagreement. Since there is no known reason to doubt any of the published values, with which the above calculations were made, it remains a question as to which would be more correct. This anomaly only points out the lack of suitable data concerning hydro-gen diffusion and effusion at lower temperatures. The calculations suffice, however, to show that only low concentrations can be cx-pected for steady-state conditions. However, during the most a6-gressive period - cold shutdown - steady state calculations shov IV-9

t that a concentration of 2 ppm hydrogen could be attained. Since this value may be borderline, with respect to hydrogen embrittle-ment, it is noteworthy to consider the actual concentration which would be attained during a shutdown by transient analysis.

3 Transient Analysis Calculations show that hydrogen damage vill certainly not occur at operating conditions, i.e. 500 F, assuming that all of the hy-drogen generated due to corrosion (which is shown to N the major source of hydrogen diffuses through the vessel. At shutdown con-ditions on the other hand, somewhat higher concentrations are ex-pected. The calculations given belov vill show, however, that long times are required to attain steady-state, or conversely, that during a cold shutdown period of even four months duration, the hydrogen concentration in the base metal vill be very small.

For a semi-infinite solid (X > o) with a constant plane source,

%(t)ccH/m-se,atX 0, the concentration at the surface

=

2 (whic;1 is the maximum) is given by

- 1 d(t')dt'

~ (g D)1 Co J

(t-t')s (3) 2 g

For a constant source %, this reduces to 1

2%(t/77'D)f (4)

Co

=

which can be conservatively used to estimate the time dependence of the concentration at the inner vessel vall for a vessel of finite thickness.

It was shown in Section III and above that % is actimlly a fune-

' tion of Co, and that p = %.f where f rapidly decreases from D.69 at initial shutdown conditions (Co 0 32 ppm from hot

=

operation) to 0.106 at C --1 ppm and to 0.03h at c 2 ppm.

=

3 g

Assuming 69% of the corrosion hydrogen enters the steel, use of Equation 4 shows that the surface concentration after only one week of operation would be IV-10

\\}

/

4 /

168 x 11.4 x 0.69 2 x 3 36 x 10 trx 13.66 x 10_4 C

=

1.04 ppm

=

at which time only 10% of the corrosion hydrogen could enter the steel. Using 20% of the total corrosion H as the average for a 2

four month shutdown is certainly conservative, therefore. Using 20% of the H flux in Equation (4) shows that Co 1.2 ppm after

=

2 rr four month shutdown compared to the steady state (infinite time) value of 2 ppm.

It vill be shown in a later section that a hydrogen concentration of 1.2 ppm is acceptable. However, it is prudent to note that the average concentration is much lover. This average is given as t/x C

=

Av6 3

2 where x is the et2 of metal per em andpandtarehydrogenflux inec/cm-hrandtimeinhours,respectively.

(Note the initial hydrogen concentration is assumed zero rather than the small quantity of hydrogen which may be present in the steel following high temperature operation.) Thus, C 0.11 ppm after four

=

months.

This shows the hydrogen computed from diffusion considert eions to be largely located at the inner surface.

IV-ll

D.

Other Consideraticns 1.

Pressure Increase Between Clad and Base Metal (Worst Case)

A limiting calculation can be made to show that in a worst case condition, the hydrogen from corrosion could accumulate and cause a large pressure between clad and base metal if it is assumed that no diffusion through the metal were to occur and if the clad defect were completely sealed.

For a cor-rosion rate of 150 mg/dm -mo, the hydrcgen production from 89 dm exposed area under the clad would be 10 ce(STP)/ hour.

The p escure generation rate is found from Henry's Law to be (Acc/hr)(q()

A psi _

4 h*

~~

(2.24 x 10 cc/mol)(Volume H O in gap)

^

2 At 100"F, q is approximately 1 9 x 10" psi /(mol/ liter) and H

the void volume in the gap is taken to be 0.23 liters. Thus, the pressure increase would be 37 psi / hour.

2.

Boron Solution in Void Between Clad and Base Metal The void between clad and base metal can be considered as a trap for boric acid solutions. Assume this void is filled during reacter shutdown with ~ 3000 ppm boron in solution.

Assume further, the boron solution is expelled during reactor operation due to some thermal or hydraulic change in the system.

The resultant, instantaneous, change in the core boron concen-tration will be:

(3000 ppm B)(0.23 kg void volume)

AB n

core (2400 kg core volume) 0.29 ppm

=

The reactivity change, at 0.000% ok/k ppm, would be, therefore, 0.0017 percent, in the worst case.

LV-12

Summary of Results The foregoing assessment of hydrogen generation, hydrogen solubility and hydrogen pressures given for corroding stainless steel may be summarized where assumptions made are still applicable. The assumptions and criteria are:

a.

Hydrogen is available to the carbon steel vessel base metal from corrosion, from the coolant, and possibly from radiolysis.

b.

Corrosion of hydrogen by water proceeds to form magnetite, Fe U '34 according to 3Fe + kH O -+ F 3 4 + k"2 U

2 c.

The corrosion rate, for illustrative purposes only, was taken to be 150 mg/dm -mo at 500 F and 150 mg/dm -mo at 100 F.

There is some evidence which indicates these rates to be conservatively high, even in pure boric acid media.

d.

All corrosion hydrogen is assumed to enter the steel at 500 F, i.e.,

the corrosion product layer is completely impervious. This, undoubtedly, is a very conservative assumption. At 100 F, 20 percent of the total corrosion hydrogen is taken to enter the steel, based on conservative analysis of available data.

e.

For calculation of hydrogen concentrations in the vessel, hydrogen generated by corrosion is assumed to diffuse through the vessel vall, resulting in a linear decrease in concentration to an outside surface concentration of zero.

The results of the calculations are as follows:

a.

Sources of Hydrogen (1)

Hydrogen From Corrosion

-3 150mg/dm-mo-1.1 x 10 ccH(STP)/cm-hr 2

-(2)

Hydrogen Dissolved in Coolant L At' 500 F, 30 ccH /kg coolant 8 psi 2

At100F,30ccH/kgcoolant

~15 psi 2

-(3)

Hydrogen From Radiolysis of Water None IV-13

b.

Hydrogen Accumulation in the Carbon Steel Base Metal From Corrosion (1) From Diffusion Considerations: Steady State a

-b 2

At500*F,30%of150mg/dm-mo % = 3 36 x lo cc/cm -hr 8 inch: 0 32 ppm At 100*F, 1% of 150 mg/dm -mo

/ = 1.17 x 10 cc/cm-hr 8 inch:

2.0 ppm (2) From Effusion Considerations 2

At 500*F, 30% of 150 mg/dm -mo (effusion through carbon steel only) 8 inch: 97 psi (0 33 ppm) 2 At 500*F, 30% of 150 mg/dm -mo (effusion through SS clad and base metal) 8 inch (base metal) + 0.1 inch (clad): ~ 30 psi (0.19 ppm) 2 At 100*F, 3% of 150 mg/dm -mo (effusion through carbon steel or carbon steel and stainless steel) 8 inch: 1 5 x 10 psi (0.24 ppm)

(3) From Transient Analysis,

-At 100*F:

1.2 ppm maximum surface concentration after four months c.

Other Considerations (1) Pressure Increase Between Clad and Base Metal (Worst Case) 100*F,150mg/dm-mo,89dm H pr duction: 10cc(STP)/ hour 2

Pressure increase (between clad and base metal - assuming no escape):

37 psi / hour (2) Boron Solution in Void Between Clad and Base Metal Maximum A Boron in Core Solution: 0.29 ppm Maximum'Ak/kfromABoron: 0.002 percent IV-14

APPENDIX In the diffusion calculations, it is assumed that the flux is uniform in a direction normal to the vessel wall. Thus the actual diffusion flux is as-illustrated schematically:

E d' Weld Zones g

Vessel

//

\\

7

\\\\ //

\\

p Wall ll l l l ll l Qll l l l ll lll ll l llll lIl1lllllllllllllilllllllll Approximately ll l 11 I I I I Il l I I I I I I ll 1 I I I I ll Wif m Flux An estimate of the effect the higher flux in the non-uniform diffusion region would have on the calculated concentrations at the inner surface of the vessel can be made as follows:

Approximate the actual condition as follows:

C'

)

)

(

C p

)

o Inner --)

03" i b "*#8#"*

Surface

(-- f

)

(

(1 - f)

)

(

Vessel Wall Thickness X That is, the g flux in a fraction f of the vessel vall is 3.33 times the average flux through the outer surface. Thus, D(C' -C) y i

fX D(C -o) y o

(1 - f) X IV-15

from which x v (1 + 2.33 f)

C'

=

o D

or o (1 + 2.33 f)

C' C

=

where C8 is the R concentration at the surface with non-uniform diffusion 2

and C is that assuming uniform diffusion.

g Since the welded zones and open spaces in the clad are of ecmparable size and about 1/2" or less in diameter, it is reasonable to assume that approxi-mate uniformity of flux would be reached about 1/4" from the surface. knee in the worst case (X = 4 in), f = (1/4)/4 = 1/16, and O' = 1.15 C. That g

is the actual surface concentration would be 15% higher than the values used, which is no doubt well within the accuracy of the diffusion coefficient values.

IV-16

References 1.

E. W. Johnson and M. L. Hill, "The Diffusivity of Hydrogen in Alpha Iron", AIME Trans., 218, 1104-1112 (1960).

2.

R. W. Webb, " Hydrogen Permeation Through Metals", NAA-SR-98hh, (1964).

3 M. L. Hill and E. W. Johnson, "The Solubility of Hydrogen in Alpha Iren", AIME Trans., 221, 622-628 (1961).

4.

Carslaw and Jaeger, " Conduction of Heat in Solids", Oxford,1950,

p. 222.

6 O'.

IV-lT

V.

HYDROGEN F24BRI'ITLIMITT Introduction One of the questions often raised with respect to the safe operation of unclad carbon steel reactor components is that of possible hydrogen embrittlement of the steel.

Although the maximum hydrogen concentrations calculated in this study are very lov -

3 ppm at operating conditions and 1.2 ppm after a four months period at shutdown conditions - the seriousness of the problem showed hydrogen embrit-tlement exist warrants a discussion of the topic in some depth.

There are many ways in which hydrogen can harm the properties of steel. These 2

may be conver11ently separated into permanent and temporary effects. Permanent damage remains even after the hydrogen is removed from the steel, while temporary embrittlement completely disappears upon removal of the > hydrogen by diffusion.

Permanent Damage There are several types of pernanent damage well documented in the technical lit-erature. Most of them fall into one of the following categories.

A.

Blistering Under conditions where steel is highly charged with hydrogen either elec-trolytically or by corrosion in the, presence of " poisons" such as S, Se, P, et c. large quantities of hydrogen can collect as molecular hydrogen in defects in the steel, building up a: pressure which vill form blisters on the surface. The steel is: permanently damaged mechanically by this' action.

This phenomenon only occurs under special conditions where very large quan-tities of hydrogen are forced 'into-the steel. These conditions are too far removed from the present case'to be of concern here.

B.

Decarburization At elevated temperatures and' high partial pressures of hydrogen, a reaction between the hydrogen and carbon in the steel to form methane (CH )'does oc-g V -1

~ -..-

L cur. Damage to the steel is caused both by reduction in the properties due to decarburization and formation of methane at high pressures in the steel, r

This phenomenon has been studied by a large number of investigators, and the requisite pressures and temperatures have been well established. In a summary of practical experience,

~ the lowest operating temperature of pro-cess equipment that failed by this type of hydrogen damage was 555F, at a hydrogen partial pressure of 500 pai. The operating conditions for the Yan-kee vessel are 5 0F and a partial pressure of hydrogen (corresponding to 30ce(STP)/

H O ) of 8.5 psi.

Furthermore, it has been well 2

established that eteels alloyed with carbide formers such as Cr, Mo, V, etc. are more resistant to this form of hydrogen attack, because their car-bides are more stable.

In fact, the A302 B type of steel, containing.5%

~

Mo, of which the Yankee vessel is made is shown by operating experience to be immune to failure by this mechanism belov 600F, no matter how high the hydrogen partial pressure.

It is clear that there is no reason to expect permanent hydrogen damage to I

ot :ur in the Yankee vessel as a result of the operating conditions and the L

. amounts of hydrogen that vill be picked up by the steel during operation.

Temporary Embrittlement The presence of hydrogen dissolved in steel in appreciable amounts has been shown to cause reduction in ductility and resistance to fracture without causing permanent damage. That is, when the hydrogen is removed, the steel reverts to its original mechanical properties. This effect is generally noticed only under static or slow strain rate loading.

i-At strain rates'such as are associated with impact tests or fast brittle fracture, only very slight effects are noticed, if any. Therefore, the ductile-brittle transition temperature as-detentined by impact tests is not affected significantly by hydrogen in the steel.

t

. There are two rather different manifestations of temporary hydrogen' embrittlement:

1.

Delayed Failure 2.

Reduction in Post-Yield Ductility-i V-2 z

These are sufficiently different that they should be discussed separately.

A.

Delayed Failure One of the most vell known types of hydrogen embrittlement is the phenomenon usually known as delayed failure - or sometimes called static fatigue./3,4,5,6 This commonly occurs in high strength steel components charged with hydrogen -

for example during electroplating.

Notched parts are particularly suscep-tible. The characteristic of this problem is that' the part vill fail in a brittle manner when subjected to a continuous static load. The higher the strength of the steel, the more susceptible it vill be to this type of hy-drogen embrittlement.

Lov strength steels, below about 130,000 yield strength,/18 are generally considered not to be affected, as stress of yield strength magnitude or higher must be applied before failure vill oc-cur. This effect is dependent on the hydrogen concentration, and the mini-mum critical stress belov which failure vill not occur is higher, the lower the strength of the steel and the higher the temperature. Steigerwald,etal./19 shows that in high strength 4340 steel a critical minimum concentration of 5 ppm hydrogen is necessary to cause delayed failure at -321F.

It is probable that this minimum concentration is somewhat higher at higher temperatures. Time to failure under given stress and hydrogen content is also reduced by higher temperatures, indicating that diffusion of dissolved hydrogen to the notch or crack tip is a controlling factor.

Consistent with the fact that hydrogen embrittlement is a slow strain rate phenomenon, the failure does not occur instantaneously. There is a delay or incubation time before cracks appear, then a period of slow brittle crack growth until the remaining section cannot support the load, at which time it breaks without influence of the hydrogen.

As this type of hydrogen embrittlement is only found in high strength steels, it is not considered to be a possible failure mechanism in the lov yield strength A302 B steel in the Yankee reactor vessel, even 11 the hydrogen con-tent were to be a factor of ten higher than the calculated maximums. Fur-ther, the vessel vill only be. stressed to significant levels (less than two-thirds of. yield at operating conditions) at temperatures where the delayed failure mechanism requires very high stresses to be effective, even with very high strength steels.

V-3

B.

Reduction in Post-Yield Ductility Hydrogen in steel also changes its " necking" characteristics during a ten-sile test, reducing the ductility as measured by total elongation and re-duction in area.! '

The yield strength is not significantly affected, although the breaking st. ess is - as would be expected by the change in plastic properties. The term " embrittlement" is semantically too severe to be applied to this phenomenon, because the ductility is not reduced to zero, but typically may be reduced to half of its original value. This is another temporary effect, as after the hydrogen diffuses out of the steel the original plastic properties are regained in full.

Again, this type of hydrogen induced effect on mechanical properties of steel is much more pronounced on high strength, lov tougLness steels, and low strength steels of the A302 B type with high toughness are affected very little by hydrogen on this manner, and comparatively large amounts of hy-drogen are required to show the effect.

As the stresses in the Yankee pressure vessel are always well below the yield strength, and the maximum hydrogen concentrations are well below levels re-quired to cause significant effects even in high strength steel, this type of hydrogen embrittlement would not be a pertinent problem in the Yankee vessel.

Effect of Radiation The radiation level at the lover head is extremely low, and the total integrated fast (21 mev) for the plant lifetime is in the order of 1 x 10 nyt.

This is well below the dose necessary to cause any noticeable effects on material proper-

~

ties, so can be ignored in this analysis.

Although the possibility of cladding penetrations in the shell area behind the thermal shield are remote, the A302 B material there vill be affected by radia-tion, so a discussion of possible hydrogen effects on irradiated steel is of interest.

The present maximum dose received by the vessel shell is 1.6 x 10 nyt (> l nev),

which' v111 cause an increase in the impact transition temperature of approximately 90F, and only a very slight increase in yield strength or reduction V-4

ir ductility. At the end of plant life (30 years) it is estimated that a total 9

dose of 1.2 x 10 will be reached. A realistic increase in transition tempera-ture of 235F for this dose at the Yankee operating temperature of 500F is indi-cated by data in the literature /9-and unpublished Westinghouse work.

It is presently estimated that thic dose will increase the yield strength from a nomi-nal value of 50,000 PSI to 75-90,000 PSI, with corresponding decreases in ductility.!

Radiation effects pertinent to the hydrogen embrittlement question are of two types.

A.

Possible effect on corrosion rate and hydrogen diffusion coefficient, affecting the maximum hydrogen concentrations.

B.

Interacting or additive effects of radiation damage and hydrogen embrittlement.

)

4 v

A.

Corrosion Rate and Hydrogen Diffusion Coefficient l

The possible effects of radiation and radiation damage on the corrosion of l

i t

4 metals has been studied by several investigators. Effects on Zircaloy have

[

/10,11 l

been noted but the in-pile increases in corrosion rate have been i

shown to be more likely related to the presence of radiolytic oxygen in the test loops than an actual change in the corrosion properties./12 Sim-l l

l ilarly, the rate of hydrogen pickup is not affected by radiation.

/11 LaQue and Cordovi

' reviewed the corrosion behavior of austenitic and nar-l tensit.ic stainless steel, Inconel X, Zirculoy 2 and carbon steel in presence and absence of nuclear radiation. The test conditions were 116 C (600 r),

21 flowing water, pH 8. 5-9. 5 ammoniated water, and integrated flux of 1.1 x 10 l

They shew no significant changes in average corrosion rates for SA212 carbon steel for the in-pile or out-of-pile testo under these ecnditions.

i Three types of steel including A212, A302, and Tl were corrosion tested both in-pile in MTR and out-of-ui te in 575 F to 600 F water and 2200 psi for 79 days by Linnenborn et al./lb-- They observed that there did not ajpear to be i

any increase in corrosion due Lo radiation.

Therefore it is concluded that there will be no acceleration of corrosion or i

l l

increase in estimated hydrogen pickup by the steel as a result of radiation j

e f fect s.

It has been postulated that radiation damage is similar to cold working with respect tc changes in physical and mechanical properties of materials.

Although gross effects sometimes appear similar, the type of lattice defects produced - point defects with an associated stress field sice of the order l

100 Angstroms --

different than the dislocations and dislocation tangles produced by cold working.

It is not correct therefore tc infer that ir-l radiated steel will act in the same manner as cold worked steel.

It is worthwhile to consider one possible effect, however.

It has been shown/15s,that cold working decreases the diffusivity and increases the sol-s s s ubility of 'hydrog'en-ir steel.

It is postulated that this occurs because lattice defects caused by cold working act as trapping sites for hydrogen.

l l

i l

V-6 i

l 1

I s

)

If radiation damage acts in this manner, the calculated maximum concentra-tions of hydrogen in the Yankee vessel will be increased.

For the steady state operating case, a decrease in the diffusion coefficient by an order of magnitude would increase the maximum concentration in a linear o

manner, bringing the maximum concentration from 0 3 ppm to 3 0 ppm.

For the transient case, occurring during shutdown, the increase vill be inversely pro-r portioned to the square root of the diffusion coefficient, increasing the con-l centration after 4 months from 1.2 ppm to 3.8 ppm.

l t

Even if this aaditional hydrogen is effective in causing hydrogen embrittle-ment, these levels are still below those that could conceivably cause trouble in any way.

It is not probable that the extra hydrogen, trapped at lattice defects, will be available for diffusing under a stress gradient to cause temporary hydrogen embrittlement effects.

B.

Interacting or Additive Effects Effect of Radiation on Hydrogen Embrittlement of Steel i

In the evaluation of potential hydrogen e=brittlement of steel under the combined effects of corrosion and radiatien Harries corrosion tested steel tensile specimens at 250*C (482 F) both in-pile and out-of-pile for 19 107 days. The neutron dose was approximately 1 x 10 neutron dose (fission) n/cm. There were no permanent hydrogen induced defects under the combined-effects experiment. Harries showed that the only property changes were those produced by fast neutron atom displacements as shown by the comparable test results of the ' In-Pile" isolated-from-water and in-contact-with-water speci-mens.

Significantly, permanent hydrogen induced defects vere not formed. With time and temperatures as low as room temperature, the hydrogen in the steel will diffuse out through the vessal plate. As Elsea and Fletcherb ndi-i cate at very high strain rates, the embrittling effects of hydrogen are mini-mized. They state that in general, notched-bar impact values of steel are not reduced by hydrogen and that the impact transition temperature is virtually unaffected.

In the case of consideration of fracture toughness of material regarding brit-tle fracture, the high strain rates associated with fast brittle fracture are not consistent with the time required for diffusion of hydrogen to the crack tip or nuclei.

The combined effects of radiation and corrosion on the fracture toughness of reactor vessel steels including SA302 Grade B and SA212 vere evaluated in a comprehensive study by W BAPL Five heats of SA212 and six heats of SA302 Grade B vith varying thermal histories were tested. The temperature ranged from 450 to 520 F.

Two test facilities were used. The dosage range was 0

from 2 x 10 to 2 x 10 nyt> 1 mev.

All test specimens were exposed to high-pH, lov oxygen water for periods up to 9 months. No pe manent damage effects were observed due to the corrosion effects. The shifts in transi-tion temperature reflected the normally-expected transition shift effects.

The effects of irradiation on the solubility, diffusion and permeability of steels vere reviewed by Harries and Broomfield They concluded that the effects would be small. Harries and Broomfield stated that the inter-nal cavities as small as the irradiation induced clusters of vacancies as sites for hydrogen precipitation vould be unlikely.

The most pertinent work on poscible interactions between radiation damage

/18 and hydrogen embrittlement reported to date is that by Broomfic,1d Notched and smooth bar tensile specimens of 1%Cr, 5%Mo steel vere irradiated then charged with hydrogen and tested. Comparisons of mechanical properties were made between irradiated and non-irradiated, charged and uncharged with hydrogen. The conclusions reached were that no interaction effects were found. An increase in the tensile ductile-brittle transition temperature was noted in both irradiated and non-irradiated caterial charged with hy-drogen, based on increases in % brittle fracture. On the basis of these data, we have estimated a maximum increase in the irradiated transition temperature of 40F due to the 2 ppm of hydrogen (1.4 cc/100gr) used in the experiment. Broomfield also ran notched bar delayed failure tests on the 3 rradiated steel, with no failures, and concludes that the irradiation did not make the steel susceptible to delayed failure when charge with 2 pps of hydrogen and loaded to stress levels above the.5% proof stress of the ma-terial.

V-8 l

f 1

l

l It is concluded therefore, that on the basis of information available and a technical evaluation of possible interactions, that there vill be no signi-ficant differences in response to hydrogen embrittlement conditions due to irradiation of the Yankee pressure vessel steel.

i I

l l

I l

V_9

1 I

Re ferences 1.

Nelson, G.

A., " Revisions to Published Chart; Metals for High Pressure Hydro-l genation Plants," Journal of Engineering for Industry, Trans. ASME, Series B, Vol. 81, 1959 1

i 2.

Partridge, E. P., " Hydrogen Damage in Power Boilers," Journal of Engineering i'or Power, Trans. of ASME, Series A, Vol. 86, No. 3, July 1964 I

l 3

Steinman, J.

B., H. C. VanNess and G. S. Ansell of Rennselaer Polytechnic Institute, " Studies of Embrittlement of 4140 Steel by High Pressure Hydrogen,"

PVRC Interim Report, July 1964 4

Harries, D.

R., " Hydrogen Embrittlement of a 1 vt % TO Cr 1.0 5 vt T Mo Pressure Vessel Steel in High Pressure Water," Journal of Nuclear Materials, May-June 1964, Volume 12, 5

Elsea, A. R. and E. E. Fletcher, "The Problem of Hydrogen in Steel," DMIC Memorandum No. 180, October 1963 6.

Elsea, A. R. and E. E. Fletc her, " Hydrogen-Induced, Delayed, Brittle Failures of High Strength Steels,' DMIC Report 196, January 20, 1964.

T.

Fast, J. D. and D. J. van Ooijen, " Hydrogen in Iron and Steel," Phillips Technical Review, Part I, Vol. ah, No. u, 1962/1963; Part II, Vol. 24, No.

8,1962/1963 8.

Harries, D. R. and G. H. Broomfield, " Hydrogen Embrittlement of Steel Pres-sure Vessels in Pressurized Water Reactor Systems," ALJ.E-R 4194, September 1962.

l 9

Steele, L. R. and J. R. Hawthorne, "New Information on Neutron Embrittlement 1

)

and Embrittlement Relief of Reactor Pressure Vessel Steels," NRL Report 6160, October 1964.

10.

Dalgaard, S.

B., " Factors Affecting Oxidation and Hydriding of Zirconium j

Alloys," ANS Transactions, November 1963, Vol. 6.

I l

Appendix A i

l Process Sheet for Yankee Vessel Bottom Head i

1 The bottom head of the Yankee reactor vessel was made from a hot formed SA302 Grade B modified (low alloy steel; attached Mill test) plate which was clad with SA240 Grade S (0.22 to 0.45% carbon, Type 304 stainless steel sheet). The base l

plate was hd" thick; the stainles7 steel plate was 0.109" thick. The cladding l

operation was performed in accordance with B & W specification W 801 - Resis-tance Welded Clad Plate Material Manufacturing and Testing. The essential operations are listed below:

l 1.

Processing of Stainless Steel sets i

a.

Ultrasonic inspection of plate by W-882 I

b.

Shear to size c.

Etch one side of stainless steel sheet f

d.

Polish other side of stainless steel sheet 2.

Processing of Carbon Steel Base Plate a.

Ultrasonic inspection of plate per W-882 l

b.

Shot blant and polish with abrasive belt l

l 3

Resistance Weld Clad per W-801 l

a.

Thickness of clad plate was 4-5/16 at thinnest section 4.

Pre-heat and manual are veld seams between stainless steel plates per W-800 l

leaving one inch gap for a subsequent leak test.

Penetrant test was performed on seam welds and cladding area immediately adjacent to the sean veld.

5 Etress relieve at 1100 F to 1150 F for two hours, furnace cool to 600 F.

6.

Cut lower head circle for hot forming.

7 Pre-heat to 300 F, attach two lugs by velding.

4 Appendix A

l l

8.

Heat to 1800 -1850 F for two hours, raise temperature to 2150 F and press

[

l head to draving dimension.

9 Attach two thermocouples 180 apart.

c 10.

Austenitize at 1750 to 1800 F for 4} hours; spray quench to 1000 F (check-ing with tempil stick). All quenching was performed within 5 minutes after l

removal from furnace. Head was not cooled below 150 F.

11.

Stress relieve 1100 to 1150 F for 4} hours, and furnace cool to 600 F then air cool.

12.

Thickness measurements of lower head ranged froat a minimum of 4-T/32 to 4-11/16".

13 Additional stress relief operations were performed on the head during sub-sequent welding to vessel shell and find stress relief.

I 14.

Pressure leak test of cladding; seal veld at leak test ares, and penetrant inspect.

15 Magnaflux inspection of outside surface of head, l

t I

I i

1 endix A

11.

Burns, W. A. and H. P. Maffei, " Neutron Irradiation and Cold Work Effects cn Zircaloy 2 Corrosion and Hydrogen Pickup," HW-SA-3168 (1963).

12.

Pressurized Cater Reactor (PWR) Project Technical Progress Report April.A-l July 23, 1964, WAPD-MRP-lo9 t

13 LnQue, F. L. and M. A. Cordovi, " Corrosion of Pressure Current Materials in Boiling and Pressurized Water Reactors," Steels for Reactor Pressure Circuits -

Special Report #69 I and SI,1961.

14 Linnenborn, V. J., G. E. MacVeigh, H. S. Dreyer and G. M. Dinnick, " Operation of NRL-1 Carbon Steel Loop at the Materials Test Reactor," NRL Report 4798.

15 Hill, M. L. and E. W. Johnson, " Hydrogen in Cold Worked Iron-Carbon Alloy and the Mechanism of Hydregen Embrittlement," Trans. AIME, Vol. 215, August 1959 16.

g Harries, D.

R., " Hydro.7,en Embrittlement a 1 wt % Cr/O.5 vt % Mo Pressure Vessel Steel in High Pressure Water," Journal of Nuclear Materials, 12, 1, May-June 1964.

I i

l 17 Carpenter, G. F., N. R. Knopf and E. S. Byron, " Anomalous Embrittling Effects Observed During Irradiation Studies on Pressure Vessel Steels," Nuclear Science and Engineering, Vol. 19, 1964, p. 18.

18.

Broomfield, G.

H., " Hydrogen Effects in Irradiated Vessel Steel," Journal of Nuclear Materials, Vol. 16, No. 3, July 1965 19.

Steigerwald, E.

A.,

F. W. Schaller and A. R. Tvolano, "The Role of Stress in Hydrogen Induced Delayed Failure," Transactions AIME, Vol. 218, October 1960.

20.

Wessel, E.

T., " Variation in the Embrittlement of Irradiated Pressure Vessel l

Steels," W Scientific Paper 65-134 Fract-P1, April 1965 i

4 V-11

MNM0800d ave -

j m3 ~

q v < p e, 4 1-'.

-y p-

.w 7.

s..

7w>m - - y,

r 4

-*.+

[

L

.+

a

,%,,x

,3-

.<i

.-y

= a.

.i \\

,4-

.~

44 p

4.

.+.

n 4

y.y t

-J

<gf,

-.M nt y-

..g 4

a

. m

• W

,, a.

4 a

s,

?Q

,.g ?>

7,:

E.

. o-a

w -=

.g

.- 1

e i

g

- M g% W.:, -s.',,

.N.

e w

,w

• g f&g. 7 ; + -:

, e g.-

41 s

F' F 'k ',-.&

'+ %%':^

'< &. ~

.Q., J,..

S i a;.%,...~,y;* *, -

' ('d7

v

,.,. p4. T o

, j 8

, A.-

, A.

v.

1%%

3n'.>

sy.

g%.~

s

<%.. gg

.m.-

8

s. N.

- y g; 3

7 F..

e g

d' T s, g%

.4_

4 M.

.a 3.

?,

g'*

3

,.v,,i,%s l 2Lv 4.f. y.

,e

s.,.

u a

, ' 9*1 g% ~,'n, I't.h ' i. r,P.+

.+

p_

j g

'?'

-w'%%I ',

N

=

w

-.'s e % '- 4., 'w

• a ur*N

.,p.

j.

-y

s,.

'{y y~7 e q -g.,g.

- r,y,y y-3Q. ' <

.r,

,, m-w y

gp --

y.

)x...,s g-

.. s

~.

-*,.~.m t.

4..-)8,

~

..'l

+

t

,3

2. 3

1. s i1

.s,,,

e..

+

$t.-

-+

2 4

s,,M-w i

%(s. y 1,-.

.,c,, x /,,

s'

~<..~v.

. :A ; y ',

,f 'p.]N

.r,

g

~

w w

s y

. s. -

2

.t -.. c ':

m 15 1.

,s u.

p.,

.,se 73-3

'3,p.

,e=

m

'. EE t..-s",

-, _ m

,M,. m.. s.,b.4.

,' M g.,4 Q.6,. W W t

J,, f *~, ;

L

• a.<

.M i-l',,

~3, g

-*i...

(' &'$

,'i **

h s, _g

.,g * *"

"{

.,g * &.

..O...,

3 V

'=

y Q-f.

, - M '.,')

^;l T

e t

v u-D s

}t,

,y 3

• %.,.m,A1.

YW

,r

.m'(9 h

j

[ 9%'/

^*., h5W 2'P 7

II f D &,i.*r" v' W,&,.^j'l' W ^_-d-__,,j

>.* *[";Py fq N

.s 7'

4, _ W,

1. ]

r>'-

s.

v

. e s

?

A

.w.

,c-s e

m_ N w,

.' # )

i

}.."d'*

j.f...

ff' '

ll1 y,

- w.m,., :

. -,g -

1 m-w

..,.., n

. n i ' p. ", Q, *1 y J O ?. / T D h b.% % <

5 _

1' m,

l.., 'N.

$l&%s.&<L, l: l.hh j.'{,;Q. w i

&r. '!:

.~

2 <

,a cg. m ~s,p2 l, : &. __g

.cz w_ _

. y.s,u) ; g

.*. sc.

R, w ;;P T//K',@e@..,!.y. ;j@7..*W-M W +F

_-Q4M M w ~ *w w

a, o

c.

T,.

... >r

,u

r..

s_.~~

.1. (.y

[y ; 3

.m

-, s., ^*

A'6 y.4

-,,.'4.~..2-..j%

,.4

.,,, ;A.

4

.Y v ~y ~

e h 4

4 t

• wl,_,

4.,%

D,3.". j 'd,

g***

%9.

h

.A

.,.,'.e,,m=*

g c &% s a,

e y-:

q*

4-7 m 4

t, 4

t.

.,.g 3*

-7(

i s

$,h.;

wsa

.2 e

y

.m 3,

s.

- q--

1-

.x.

1

,p g7-vd

,3.,,.

.M c y,pw t.4. w i. * -

.J.,**L.s

.sd 5,

A p.

A

.s sv-9.P-

-'p-,

4N

,r

.n.'.,,...%,.-'8e g

9

.,m gS

,j

+

g 9

_.e

'14.,.,4

- ' s '"e

.m'"'.

• gg g~..g.

~

,w s4,-

e..*', A v*

"t*

r s'.k,.'.

go y

6

~ <. -

...y b

' s,.,.

,,'.J ss

..' M s.,.c. e.. M.

.-m'

.r.

M'-

__v.'.

g+>

' m

-V. 3. t,yi )q..*n_y,;. s~. m.., u,, A}..

'c w :.

w w

.J c,

.c em 4

.w

m

.-f. ;.

z _3( *. *,)

7,> s.

. 4 '. c'-

A '* ' ?

,,v, " 'a"s"l-d % 7

-4.dW 8,

, f

'8

.,M

, Jy p

wh 9

e[ (

[InI ;

,. y f k

., A,.y'.

f... t.

.n, s -.,.*y sv.

e.

r.-

I'. 5"+) w,-

• er.s.- es=. e. 3,,,

s.4...,_,+

..y.,..,' w',[ #

.- 3 vw w 1

e

=.1

g. M g. w-Ted r

g,

's.-

.ds

i. y.

4 e c,. -

o y -

g,

> 'g f. ?...w w[.. A' E. $ _-

7

. (

'h,a a t{ ;[g f..

?

.J

• G t

,r R.

74

.5 g.

s

+

s y,.%.,,, e.

r,..*ev

• ,s.,~'.N..4

, %y* ius -

xe

- - - - = -

r,

. cy.

4 'j

,.M.,* ',. '

~~*

,.g*.

$< r*

4r.-

r..

• c

- ~

4

. ~

m.nn,, o' '.'r,.=

. +.

,J.y

.s

r. %u. '..f r',

c

...c r9,

-s?

G

.'i.'.e,.'

s

.r g,

,.j s',*'

3%c

' ';.,.#t

,1

e.

,t e

r'

• f Ja a

'A-y

~

n ~

[

,f, If, _

p y.x'-

8 g#

p r a. v ;~y~ ; - m.,r.

3y,n n

w.

\\.

. M i.y.% & x

s t

.* u.

i~. C.

.v 4 N' $ bb; e.

(

'l h.

.M 4

N b M

[~

z$h

,5 $

M~

^

. ?-

W W - M Q;70 P C W ~ % m o G.%W7p y ss - ~ WMg - @%

g

w % nk:

39 3 q, Ana-h

..- a yy

,y c w n 3.M s,q x s. ;:.w,, t.,.;,.. 3

,m...,3.y.a.,, 4, c p.3

- ; y

. gg y

y

+

..,v

. <..4,.

19,7. w w.,y,,, ~. v. py,

n,,..-

as5 e..

ye.

1L.,.. -

n 4

,s; q ty,&.. -T"t4k.%,. a) h.,".N. &sfy &.u, &.

A

< o.m f.,'

,)- t ~-.y,;}f;m,b)}

c.n }..,M. ',*j h*

e

. p7.-:m,

..t-

.. ~ m.

o x

...w:.$:

.O $.*a*.s D TJ W Wa m.% y w.yc w, -

- y r_,n,3,+.p., m. u

-.. s a

+/N

.s

>. - n+

.e

..>,i.

s V y..L.

c.a s.

4 u-.

A. t.

4._ - 1v c.- 4

, o,

i4 N-c. -

4 e

i, I

41M

,d ii m.[*f,' *h,g

'.N Q<.N lq I^^,. 4. '.[k. b...ip*..i.

7^ (.. w.,.a...,, h.,.. w. s L >ha"Y[

. r.~ f -

n 3

.~r. v a-,,s

.s

,uL

),N. " % wy w mm.~~~K 'qu u,q$.:

m v.4 v ~ ;** *n ~ y y'*?n.,.' -, k yD y y' " t-6

- W

'u-43:

• A -

Reg ~3~4y w V*f S.

W

v. o fv y" '< -

s-s

m. a m

' n,'. s. p '.p u b i "

• j nIh;.

h

.e. -

• T s 1 ** - -

j y

/..

s

.t

~>-

'q na

.y y

,,3 - j,4 L

s

.t-4,

,g.n'

'. E

--h j

^.

(s,

j u,,,'

~

m.c

.o g, s

[.

i

.s 9

t

.,?, N.,.:,f.,

<8

.-. t

- t, s.. w r

7v;w.

w

.,,. s..s.s

,f,

~,.

.,D.-

s_

~..

g y*

p'e' q7 l'

.4,e

' gl J,,;

. s3 p

pf -b,,.-.,s.

g.i*d.

fig 4 % ' $85,,*'g<

g-

\\$ 4:

,. [s

',;A-

..ty. 4-L i nw '

.e - c 4,hd..,das inse@ A - d.8'H. ' h g-)

7 hgf

{

1

• yb 7

TNN0ll00d

-__Y

-h~..~

ki.

~

' Y.

n w,%L

?..

?'

..^i.

~

• '*Z *v &, Y,UA k ~b.l*Y.N~4,N.

?.f.- ( *-

f p

m.a.' f & N & *yL f.e,'hl& $

C y!

b

..e. 6M hJ v,.'

,+N ~., '. ;

A a.;; x. (*n.

x...,..a.g i.w, s',f % c.,,4.. e, J,.w.,. 9 m w n m ie p; n

s E < f d y e'. g v.4.#, M., ~"*fj+.

+

p -'*

v.,-

t+a t

-N

^

,7 f 3g v.

'r*,

4 iY,

o

,y-

+

+. ;:#

58 Ow

' n L.

3 6

,se :-M *(

-&; t

< *:c.

^7:, t_ g

g. 2 ",/ - -~. ' m. p. : %.

'5

, 1.u.e ;3 ) j KV,, a

-~,y g, f; ; ' <-.

v

. 5 %,.

3

.7

,a o g <qwf:1 +v +@,? g :wy & v;.1.w gp' y, %y m.,,n - -y)v:s e. ' 9 p l

..E ?;

W.h; Q e

s_

w... o twmmw ve s. g.y v

y; kv-,.m, m e.:

.m w p. g r. y.e,,m.,y% -

~ ~;, a u

R,. -, g y - V*+',*. ]w. - > e.

m en c

e

• ~ c

'm.m

• m s.)**,

6p,

' L

% 'h -

J-

• y s y n g; W.y.Q'.e^, ; p:.

.; a. >.;-.3< y).'q.g..V'y_.,,-,.,,-9s'

, } R p;.fa. A,.. g,,,.%.;iv m ; '4,.,. s v y.. -o; _

ye

.d*y.W

~-

% x, n & ;

,_e -L

':~

8%

r g

g y

., - ng s wg

.n~7.-

..,. a y_

.. Wsy;n S, (:

,"as.Sj d.

.;u y%w.,m? +q A i Qs.'4W~ m.,r..w. n P *N:ob v t.&..uy,. A y m.A. s ?

i9 N

,6.

J6 c

s. m,.y1

,r n : - s. vz.:-.w.

v,k 4..a :,p.e y * * y

. e! n

.o.., t 0v

.%^; >

e-g.- s.%, j"em..e opm-'-

.s, p

s

-5,-

j r

g s} $ s y n.~6 y *

.,,,g q g g

f' N3

..I 4 ;I."m

.J N

g 3.,. 4 g

e js

+e a

4 f5kN~

r e

Q,y.W Q a\\.E ' %,e.g" f" R@d 'm s M-

• a p 3.1. e. 2","D' & gd.)p.i,,*l?', -%v t'+; &gp& p%,3, T 'ig #

i

,AY*+

. g'{.

g

,.~

L,;;... ss N.

.:- w ; M"m J_;,Q fs % _

• V.

- y ' : 'g >: ~4 5 X s w m a ' ;sbn f ', x' i'

p' 'q} $. w w im

y _t*.. h

.,r ff O.

'f. c

.S 4.t' h.

g 3:-;*f}

. Y Q&y e

W 4

.p +e w, n - 7...a e,,, sem-

.e,.7,

w. n. : y.,,

.m nv y.x

-.2,LO~

v.

y., : 2 ~

w.n.m.

. M. a ~W+ 2:w w..ccM,& LenM. ~ 2 ? ; %., n -

a

,.,.; y -.+ ~ ~;,.ui.c

, ^ V." Ws ;

9

~,

,x.

. w$r - 3'-4

'U f -, :.~..w %, w

,-[ %,- +h-Sk

~

• .i. j I.

W w

Nya

,,M P'~,

.w.'.-

jr g

, f w, b,

e 1....,

y 3

r.'j ;%* '. e,,i, k, s. %

ys

• : j1,h ' ) h.'

A.

3 *b, e

0 8

z } s.,

o W' /- ' '..

'M

'k' 1

.m

e., /.- j M.'

d

,.-h,

(( y. d,' a.,I 2 **,,..~ w

.[s. y+y ++. %-

\\., *[~i h, e 4

m....K ".,M. t 3'r e,a/

.+*h w

t 0

% 3

.g p - - e..v.

• ~

• , q {.-t f ~1.. Q,.,., -

, e.4 4

.n.

1 -

.4 4

~

s.-s -

s

( \\

s

.q r e

p.

q EN.*'s N Q '.1,@ @. I M. [~. '. y,f'[~ h. t

]

<g..',,. 3.d 7 ' '

• s',, ^- [ k,y'l s 0

8 A y%.,; q 3 a. > g y v ~ %-y,. g QgM

,[.'s 4_p t Q.. a yx w

s mm.',,v,fp +

J.

( %j,. %... ~.. L

~,.4..-

w

=,4 se

=

s %g,k" y.

y,,_.

4 ss

'4 J

~*n y ;

• r 'sy q ) - * - ',5,. ).

3 % '.

.., f. 3 y

..y,.....,,-c_,A ",, y. A f,a,*'*>_

+<3v.,v

./..

-,.a

.~9,

..~s,,..s.,-

t er g

-,.., ;ss.,..,,

xm,, e. 2..

-,j --

.r

- 4

_g p

- y Y

N., g n' :',. --

• g i 1 O_ w n. _.,., ~#

~,.

~ m.,.n.

.w

,ff./.-~' n&

%*"e

.-d sn-i n,se ef ",m _.e J 4.N, w,h. wRt,'.-*,.'+.~W'.**Y-

.' '

• P

'J

,p 14

f. A i

'. 1.

y a,'

.. e y *-%

.f.

,y g

t E. w, A:

& d V ?,$ L W l& dn$. 0 $l. ? $ $ g y e, s m g'T ? $ v% p* ;ff';,

s.

s s..

d*;. W f..:.'^.

d

,a E.

n.'WMW MCMTM6

, a 6

., n. n..e m. ;. n.. ~t.Y. o,n M J E 9. W W W %v.:?.n-,

u.p/n. 7%.:..

ww w x,

...*~>

m,.v: m

?

m c.

p~

>. c.a,,<. 4.. s.,= a r..m, % V.+r @.~.n n.,

+

s c.c A

w

=

s.%

p

@> b..m g>;-

.u.,y r,. s e

MsC h n d'n, e 4yr- % <.

y p*$.-.. g A.. 7 g4.+c g gf s M,.., 3. w a.

m.% - :-

_, m...-w4,4h r 3

74*

-:c n p s

..y

.w. q%.: m. m f;y a..%.i.,,,~a.. v, s

, a. w.3

.g

n.c.

A we..

.. _. 3. o..x.,, j.,

.. Wi } Q.?%K,,n m~~?Q,l%rs ;9%% ' i *R',' 'l s%. s+. -:;~f "i['*W!l. 4..,./ % "#Q%- W,y x j

' **J p @p$,* D p... a s. w.

,.y.-

V

. n.r;. u.

g=

yd~ ns

• n h

ys,.. + y;- -y,s. y ?s.g. <,,g,,g.<

y+

i.,. ). a

~.

e>,.

n,

.y n s :.y:aa"'v

'g* t.

• ~ ~., +- %xc h.y **m

~ - a' en m.

~.a

. y >, e~, c w ~ >,

,a g.* k au.

.w.

%,p 3

g

• e f

f 4%

4~_

n.pA.

.-e bY b

S. pN.nj a.

'^#

n %.%'/

, ;... g, i. g.. A s.A. /., N 4., c..q. M/;.vs i

'. q,w.,..% <

tx-s 3 y

,e g

q. n

.W

, J W.. m.e'.si t, -

..... p%w.

s

~ ~. w'h f o.,

., ~

- ~.-.#

y U"

," 'I l,,... ~ - *4 e, a.~.., y, m t.,

p.f % $'

.,4

. %,Or Yb

$,4 g,

i 1.I

,9

.p'

,J8*-'fty

• A$'

g3

,4.

. ~..'.#h.m.. h F+-

W g

p e

,e s

g, e

.-s

~ -

y.e. - s.:...,.

.c..;:

w 8

D.

v

.n vep-

~.

y

.~

u t

.~1

. - ~ m n.f:%n 3

4

... w. s uc..

.m. ~ e 3 s.+,e k,,y

v.. :*

, &. a.,. Q Q ~< f.,-Q ;,.&, t w..e f *,y[,q*M

?**'?gQ 'j ju,.5[- R -~_ '

id' 'e-Q.f.

's.e.y..

f a: >

=..fk L,y' y :

+.

s,

u.

a.1 7

r.

g

-.m.

.. s

^ - ' W.[4 ' y. -

e,. &.1.9,,G 1 ; D G. % '.3~?.'.'h +2.?JS.C ) > "%.e. wm.A.c /v/+/ N.r d,n *46.a M..

' w <,.

i c:

1 W

. c a-w % e a

a n

n1 p j, m-a s

v -- -

... g.,.

y

"

3 >Q J ?my +, K3W^aQ;x.V9< %, ?nw':n%m%.~%%fN #f.

?Q q:

Q W V &r w,.Lf~; O, W. v %,i

~. -

v,

(

. w.%

.<.w.

O m

~ 2 -1, Q, r

'. ?.s p gm ;W.: O x.

- ;w y _ pm v y 4:s.,;: Y Q N ?*

w ' w &.

W.iw; wa>.,-wry z: ~ p}:w.4&_; <9, a &u :. a: c p.: %. n q.

swn -

i u w.w m y_ m o,.a..

. w,a w;. M;.n,m. W a

-c 2: L>

'. c. %a s y

- o 3 W

4...

%.,?..

w.(r

. h?p5 y.

5.M n.,( ~.,- M.w' -;., G. U p4*d.@,v. Q,*s.Tw.,-@n%{M : i.'.%m. m. $u<,(73 -.,$'

W fi.

..,m

,u..wa y-e v-4 w

s..q c.4

,M. 4. a..-:... v..v 3 i

e u

.;,.n- --,-

..J

• w 4..

e A w

w :.,%q. s s.

-- R.s...c rs w %;Qw." b P ep, m +s? % s M,o.

, V we %.,p m.,..s ? 6 - a ' }:,h,;,.,

l r, Y J. ped' $$

u%

b u.

r ;; w.

,1-w. yvm p,

c-

<-4 U e u'; + h.3

.w s

. N _ %.A * ? *f

.,h *T '- - 4

.. ad d

.A ew9Npg* r N EM s d\\

7MQ V

_#.a i0 Af5,pN3y @ ? C M *A'""****j@ W'W E.bN'S* M'fM~**7/[

• PM ' ""7 4

4

.e

.I 3 k n m's.:

-wn w v. m>a

. - - n:r n~

f m m + w w w w w";r.,w. m? A;Q %m /et p<A W) d *.,"&)G.* ~.

A

..e n.

w ro.

-- ^ ~

= v',J.?=;My W : sgn 't u m..: En, Q s

mrmM 7-R Y W P W y ;,p'M

,[1 s v.4 J u L.pf >g~Rw' p 1 weg A. Ave.+y'P.g...JW gf h%gA, 4n%.,.q,y b-

,.c*

,,c=

c

.) u

- W&s~QL)pm$faw%.&w&idd LEh;ST[h.n&w$yY&w ;q &. a.. '[v-. ',

-.J sw ni Q^

g

. u _ m e,m e,>

- A wa %

u ax r

,W:'L*f V :p.at,-f M Wyfv..g. p wp nyc.wmqv-f y -

t ac<

.m

.s.

NM5 S

h Y

W 4 9, susM%cM g.

.'r-h f**,$ N,.

hh' h%

Af

('Y*

j; h.

^0 R'

t e

-. t g~ w-.'

f

,4

% 9 g e n.A4n.y'w,.%.

+-M %.9~>%.

f m k n. M, f + <.m - a nh+,v.M.%p%p&mde r e*Wm '. L g.L'N

• e.

M *

% w q

, es

- n.

+ : s~ N

• 6 p :'-g

. ny yn

.p wmy y m~ glb.E.:,r WW*cqWMX.7M4%+M'%7 - WF:/&WY.WW

Ah, l'A CMba J M

-'., - -O t V-

.r r1 b g

~..

un.gpmym" y%ph.4.;

m*pv W

ny 2 M W. w 4 %.4w w A s r.+.)4.".=aW+ :( g% MM. mMA _s pw e 6. WiMc r i

.- - m m al g A y *% 5 W + % 3~%.,r s n.

=

5

,,cs M W

y

, yph; sq& %

n k.L -, n

- +~

sw:

m@.,..h,.jh.,, _>%[59N((m.M.v$d.,hh. @'Nh~. N+

h. k.

A.- m, -

.a**t.

.j

. %. p,., <s.

e,.g~.. ~.v~'s.y

., 3 s

x

,y.e a

-mm a

.."P.

P#-

m., ','

. w. m m, w..r.~..

,..,c,

.. n... kl,...

o* y:s'Y, i.eO'.u

'N NASic-w <

a.

~)

.~~F'We. s'kP*

wwF

,5.D'> v/ '

,p38*O.**

'% T' 4 4'e P

W b

t.-'9_.r,e.*',

,g "p & -mc...

g e$

.."=.,m.L.,

' -7 [

199

• M..gf-4'~J

,.;n-W A

% =

^

t.-3

, 4 )."

^I.

R 9. >

1

. -4 'J g

b.

- ' I C.,, *')

5A.7= NI p~

M[

.ym. < ',

c.m m',.:.m.,7 g;..: m s.s,

,x,

y,

&,p

..,a.

e tJ.,

f; g

'(

[$. n,. c,.,,. -

5 I

- n'.,m

,. ~., m n

, j'f

(, {ff ~f)?*

' ' n M}' kt.t h 5'Wh&d%.'WSWW.b?;m&%-?MfW d$fM&_Jt-

$Y s l

_ jk } -

b

~

y 9

,:,.mm a w m..

~

w.-**

mq%-~msr.g'Ms* v-

'~,,,,.,r%+,- ;F., F --,_w.

-.-,. w y.,

w.

~.,, w, et s

t o

r p

e

,,.~.4

-e v.

s.,.

s

,n

,* w g

11 y. e m.'. y'm*,g.J' e*f", $ 6,.,,

u, - - N r *

[- @ N M.m V "*e"'~ ~ ) M b M

,m,

j

.. g.

y s

s

-W c -

= q Q f

.,.j f-J.;

• i h s s

e i' ' ' ' *

'I '

h^

g,

~

$. 6 'd -

p

, s

'e

~

i I

}'

f.,',.,

' g.

,.ry.j,.,-

.w.

w

• f y% PA_

, )* k

.-6

=

?f A

g

t - ' s 1

a

.'r e s.

F,

.v..

v

,.-sa

,6

..r*

' i v,_,.

-(R.g u y. g - f[,,-.

,t 4,

.m-r

} s s.

d

- p

. 7,, 1 f, '

y

..,'n r

e (,-*

.. ' *. %4]

[

.y

).

, u

- s

.o

-~ _ &,.

.s

.Lk Rt

[h

• h~.

K,~, A.

^

~-

^*

a mt. $ -

Q Jg.

6i. 3,..

,y v*

w y

e per. [ -,i++ % h ',,,,

p.,,, @,

,,. q^* : *

-[,

if p

., + c, '%', <,. 7 k*,

f.h a 7 -

• l-

6. ~ Q

,~.f

- - 9

,y y

e 2

V

'.3.

q.,a

-,a aj

,,; ~,

,M-y e4 s

g g-4 y*

N GW

" 7.,..,; j "

,y

-I w.

v. - e.

54

-s-

. -J,

,-,w.

% - gr. - a.,p =- s, e

=

s e,

s,

• ,.,.. - 3 3

,. 1, f*.;

f-~~-

'a, A 8

w t

. y ap

,,g 4-a i.gy%.

1... p 3..,,

g 4.,

',4 % - 3. :

i

]-

, [4 i,% y g

.IL.,

y

,y r(

4 gj

, <+

3

-:..e

+

",' '.G J l V; 3

1 9

e' 1

s Q'

't.1 ~ ', '4%

JJ

.q ss

.,u

'-. n.

p q.

g -yw y.-

p g 9, -.-,w 5 x.y

ry,.

I

?

a

w.s. -

?

'.. + -

.-4v. b p, 6, ' T

,1

, O, '

4

, - 9 y

a.<

.j,

.n: c,, ~c '". " :,~ g t s w...

.y C.

, r%

• - q

.J.

<*d.',

~

..a

. e, Ma 7, p h,. -.

w

> +g t '

7-

.~M

h,

'.'A

<.e m

y eq e p, j

-L

?

_84

%. s

• =

f.,.,

s

-1 s*A

,e R,

L_

s,.

7,.

,, *=-

P

+

"N-,'*

' D-.,.3 e-M MD

[ -

g 4.,

4

[

,.A3 Nef 1

w-c-

s

~.,

.= w 4

5,

@ g P-y e

s.

t

1. #5 4,,

,%3,,,,

.e.

..,...4.

4 C

<7 (

%}

3.7 *,,

s-

..A

-%. (

a

.-4.:-.

.;r.%. ~

T

.g 4_..

a.s

. - 6

_ ' Q,..."y, K.

'" sW'

,*3 $3 -

"G fA,

vFj ky s t J' .v ...p 1.- o-

3

.,L L,,..

%.,^a>p-,

t

s..,,

, = s,w .,.? .,.^. e 5 4 A ~ ~ 3. . a, w.m, y- _..- ~ ~. ?

• j 4

A ..~L ) j .on -> av, m ,s,. s.. e..#,.. 7 a, n n $P E.? " ? e-sr, r A., .w . n* * -. c 3 - s ,g a at 3 P' .-!= p '..I, q

• +

m 9. g - I e+ e.,: }. =

.j 3 j e w

., s .-,.,. ' * ~ w ~, O

r

'k ..s - e. .~t.- a- . y g M. J ~ a,,-. m, ',, h Q ) w -w

.,,.. w

-s,A a: 74 % m

q. 9 c ;,y r m wn

t. x

__ yt .m <n, 4.f .. i

y.. -.

.c' s '- 3 i, . : *a J...,@#. f j 2 9 . 4 4, 4.. 7.,,. V.- /. a s*-* g-s - y y s < t - n , p, 3

• -,, {-

1 =.,.. dg 3,1.. ' ,g , s .n.., - f . ~~ % / /

- {$

s e g. b + w s M A s+-. t g w 4 ,4., a 1 - < ,alb ,3+...u n,.. w o (,,.,- w 'M Am o a. - + +. - .c - ,}}