Fatigue crack growth rates of irradiated pressure vessel steels in simulated nuclear coolant environment

Journal of Nuclear Materials 96 (1981) 261-268 0 North-Hood Pub~~~g Company

FATIGUECRACKGROWTHRATESOFIRRADIATEDPRESSUREVESSELSTEELSINSIMULATED NUCLEARCOOLANTENVIRONMENT W.H. CULLEN, H.E. WATSON, R.E. TAYLOR and F.J. LOSS Naval Research Laboratory,

Washington, DC 2037.5, USA

Received 30 June 1980

Fatigue crack growth rates, in a simulated pressurized water reactor primary loop environment, are presented for A508-2 and A533B steels in both the irradiated and unirradiated conditions. The initial results, for specific experimental conditions, constant load amplitude waveforms, and load ratio = 0.2, indicate that irradiation does not enhance fatigue crack growth rates more than the enhancement due to the environment alone. This article outlines the experimental techniques involved in this testing, and describes the results and thejl possible significance with respect to the existing safety analysis codes.

recently, irradiated, over a wide range of load, loading waveform, and water chemistry conditions. This work is expected to provide an enhanced data base for a future revision of the ASME code. Fatigue crack growth rates of irradiated and companion unirradiated A508-2 and A533B steels, in a high-temperature, pressurized, reactor-grade water environment are presented in this article. Tests were conducted for two frequencies (1 Hz, 17 mHz sinusoidal waveforms) and load and environment conditions which attempted to model start-up/shutdown, and hydro-leak-test conditions in a pressurized water reactor. While these types of cyclic loads occur relatively infrequently, compared with reactor or turbine trips, or power loadings and unloadings, they represent some the largest pressure or thermallyinduced loads, and are expected to be a major contribution to the set of forces which might extend a crack or flaw in the wall of the pressure vessel.

1. In~oduction Laboratories in several countries are conducting research to determine fatigue crack growth rates in nuclear pressure vessel steels [I] . Such research is not only important in providing continued assurance that the pressure vessel steels in use are reasonably resistant to fatiguecracking, but also will be used to provide reasonable and durable codes for design and safety analyses. Appendix A of Section XI of the ASME Boiler and Pressure Vessel Code [2] Ruies for In-service Inspection of Nuclear Power Plant Components presents suggested upper limits for fatigue crack growth, in air and water environments. These limits, or default lines, were derived on the basis of some early research (-1970) which evaluated fatigue crack growth rates in unirradiated A533B and A508-2 pressure vessel steels in air environments over a wide range of temperatures [3,4] and in high-temperature, pressurized, reactor-grade water environments [5,6]. Since that time, additional research has defined some conditions, involving aqueous environments and high mean stresses, for which crack growth rates can exceed the default lines of the Section XI code. As a consequence, in recent years there has been a concentrated effort in several laboratories to measure fatigue crack growth rates for a wide variety of pressure vessel steels, both unirradiated, and most

2. Mater@ Compact fracture or wedge opening loaded specimens 1T or 2TCT, or 2T-WOL as defined in ASTM E647-78 [7], were machined from reactor-typical materials. Chemical analysis and mechanical property studies ai room temperature on these materials in the 261

262

W.H. Cutlen et al. j Fatigue crack growth rates

pre-irradiated condition, produced the results shown in table 1. The ASO&2; codes Q71 and V82, came from two nozzle dropouts; the A.533-B, code L83 is from plate 03 of the Heavy Section Steel Technology (HSST) Program, code HT material is from Electric Power Research Institute supplies. The major face of the specimens was oriented parallel to the plate or forging surface, with the crack propagation direction collinear with the final rolling direction (plate) or normal to the applied hoop stress (for~ng). The code LS3, HT and Q71 specimens were all of the IT-CT design; the code Vg2 specimens were all 50.4 mm thick; ‘2T-CT specimens were used for the irradiated samples, 2T-WOL specimens for the unirradiated samples. Some air environment FCGR data on heats L83 and Q71 has been published previously [9,10]. The 1T-CT specimens were irradiated in the Union

Carbide Research Reactor. Two capsules of specimens were irradiated for this and associated studies; specimens L83-1, L83-6, and Q71-1 received an average fluence of 3.4 X 10” neutrons/cm’ (1 MeV) in a 1240 h experiment. Specimen Q71-9 received 4.5 X 10” neutrons/cm’ (1 MeV) in a 1374 h experiment. Specimen V82-1 was irradiated in the research reactor at the State University of New York at Buffalo and received a dose of 2.46 X 10” neutrons/cm2 in a 744 h experiment. Irradiation temperatures, produced by gamma heating of the specimens, were constantly monitored and held at 288°C (+15’C). Thus, the total dose and irradiation conditions were intended to result in material which would simulate end-of-life condition at quarter-thickness of the wall of a preszurized water reactor. On a laboratory scale, these tests are carried out in

Table 1 Material chemistry, heat treatment and mechanical properties

Material

C

Mn

P

s

Si

Ni

Cr

MO

CU

V

A533B-1 (J-83)

0.22

1.37

0.008

0.008

0.22

0.66

0.15

0.54

0.18

0.02

0.19

0.69

0.007

0.009

0.31

0.82

0.38

0.62

0.01

0.13

0.20

0.60

0.013

0.012

0.20

0.73

0.35

0.56

-

-

0.20

0.60

0.013

0.012

0.20

0.73

0.35

0.56

-

-.---

A508-2

(Q7) A508-2 0’82) A.508-2 (RI

Primary heat treatment

A533B class 1 (HSST-03)

Normalized at 870-9SO’C (1600-1750°F) for 4 h, air cooled; austenitized at 820-900°C (ISOO-1650°F) for 4 h, quenched in agitated water; tempered at 660 * 14°C (1225 + 25°F) for 4 h, furnace cooled

A 508 class 2 (forging)

Austenitized 840°C (1550°F) for 9 h; water quenched; tempered 650°C (1210OF) for 12 h, air cooled; stress relief annealed 66O’C (1225°F) for 20 h, furnace cooled 3O”C/h max. (55”Flh max.) 0.2% offset yield strength [MPa (ksi)]

Ultimate tensile strength [MPa (ksi)]

TL

472 (68.5)

620 (90.0) a)

TL TL

538 (78.0) 5.55 (80.5)

680 (98.6) 692 (100.4)

A508-2 iv821

TL

475 (68.9)

A508-2 (R)

TL

462 (66.9)

Material

A533B-1 (L83) A508-2

(Q71)

Direction

a) Estimated from data for this plate found in ref. [ 81.

Reduction in area

Elongation

[%I

ml 19

-

-

636 (92.2)

59.5

24.9

606 (88)

73.5

25.5

W.H. Cullen et al. /Fatigue crack growth rates

autoclaves, really small pressure vessels, which are controlled so as to pressurize and heat the reactorgrade water. Each specimen is instrumented with a displacement gage for measuring crack mouth opening, The load rods pass through water-cooled sliding seals to the load cells and actuators which are external to the autoclaves. The current crack length (a) in each specimen is computed by measuring the crackmouth opening (6), as a function of increasing load (P), and relating this ratio to an experimentally predetermined relationship between S/P and crack length. At the beginning of a test, the crack length, as inferred from the compliance, is corrected by adjusting the elastic modulus value used in this calculation, to the optically measured surface trace of the fatigue precrack. At the conclusion of a test, the specimen is broken open, and the final, optically measured crack length is compared to the final value as computed from the 6/P measurement. Any discrepancies, which in this series of tests ranged from -0.2 to to.75 mm are reconciled by resealing the data with the error correction applied in proportion to the cyclic count for each crack length measurement. The tests were run under constant cyclic load controi, with initial AK values of about 18 MPG for the irradiated tests, and 22 or 27.5 MPa& (depending on the individual tests) for the unirradiated specimens. Details of this equipment and these procedures can be found in ref. [I]. In order to accurately simulate the pressurized water reactor coolant environment, the water in the test system is carefully monitored and regulated to yield the specifications shown in table 2. Water is

Table 2 Water chemistry spec~~cations Boron (as boric acid) Lithium (as lithium hydroxide) Chloride ions Fluoride ions Dissolved oxygen Dissolved hydrogen (saturation)

1000 ppm 1 ppm CO.15 ppm X0.10 ppm -1 ppb cm3 /kg 30-50 water

AU other metallic or ionic species should be at about trace levels. Some iron, both in solid and soluble form is the inevitable result of a corroding specimen.

263

deoxygenated by continuously bubbling hydrogen gas throu~ the contents of the feedwater tanks. At the Naval Research Laboratory (NRL), two single specimen autoclaves have been located in radiation ceils and designed for remote handling, of, and data acquisition from irradiated specimen tests. In the normal test preparation, the specimens are loaded into the autoclave, where the displacement gage is bolted in place, and the specimen is lowered into the testing position. The autoclave head is bolted on, and the autoclave is then filled with deoxygenated water, from the feedwater tank, and when the dissolved oxygen level reaches 5 ppb or less, the load cycling is begun. The dissolved oxygen content continues to decrease during the first few hours of the test, ultimately reaching values approaching zero.

3. Results Data for A533B steel in both the irradiated and unirradiated condition is shown in figs. la and b. Although the materials (codes LB3 and HT) are from different heats, and the initial AK for the HT-material tests is somewhat higher, the trend of the results is very similar. The 1 Hz tests produce results residing on or near the ASME air default line over the higher portion of the AK range which was tested. The 17 mHz results show a substantial increase in growth rate over the 1 Hz results, but since the increase is about the same for both irradiated and unirradiated rnaterial, it appears to be a function of the environment and cyclic period, rather than irradiation. Irradiated specimen L83-1 was tested over a AK range which began with a value sufficiently low that the classicalthree-stage behavior of corrosion fatigue crack-growth rates can be seen [I 11. The first stage, called Type 1 growth, or start-up, occurs at the lowest values of the AK range, with crack growth rates rising to a plateau, or Type II growth, characterized by somewhat more AK-independent growth rates then for Types I or III. Lastly, for the higher AK values, Type III, the growth resumes a strong AK-dependence. The reader should note that Type I growth behavior should not be construed as an indication of “threshold”, or lower limit behavior of these fatigue crack growth rates. The location of Type I for these data sets is simply a function of the initial value of

W.H. Cullen et al. /Fatigue

264

crack growth rates

APPLIED CYCLIC STYESS INTENSITY,ksi-\/'TiK I0@

Ia

I R533BSteel--1rradrated Reactor-grade IV'

c

0

Spec.

L83-6

mHz

slnewave

17

10-s IEn

(550

/

Water

288

APPLIED CYCLIC Sll?ESSINTENSITY, ksl-\T

lE2

ii

f)

lE8

II

’

,

w3

I8

RobIn

Round

! / Ia-'

288

c

(550

Aspqc.

1 1 I ,,J,,I

17

/iI 10'

/

I

/,,1111

II

Ia'

RPPLIEO CYCLIC STRESS INTENSITY,ilPa\K

10-q l0B

10

I

R533BSteel

II /

1

Results

IE”

'I

F)

ii

IHTIG

’

!

,I I

mHz

I

1

i,Ii/,II 10'

I!

11

I /l/!//l lE2

RPPLIEO CYCLIC STRESS INTENSITY, MPaVX

Fig. 1. Fatigue crack growth rates versus applied cyclic stress intensity factor for (a) irradiated and (b) unirradiated A533B steel in high-temperature, pressurized reactorgrade water. Comparison of these two graphs shows the parallel behavior of the irradiated and unirradiated specimens. Since the unirradiated tests began at a rather high initial value of aK, the three-stage behavior, seen clearly in irradiated specimen L83-1, does not manifest itself. The dashed lines in these and the following figures are the default lines found in Appendix A of Section XI of the ASME Boiler and Pressure Vessel Code [ 21.

AK chosen for these tests; true threshold studies on these materials have only been conducted for air environments, with the results detailed in ref. [4]. The onset of Type II, the value of the growth rates throughout Type II, and the transition of Type II into Type III growth behavior are established by the test frequency. In this particular case, the 17 mHz test was terminated before the onset of Type III growth had been established. This test was and, in fact, all tests were deliberately terminated well before fracture of the ligament remaining in the specimen(s) so that a final crack length, as inferred from the LVDT displacement gage could be optically confirmed from the specimen(s) fatigue fracture surface. Fig. 2 presents the results of an effort to separate, if present, the effect of irradiation time at temperature from the effect of irradiation damage itself. Specimen j/483-19 was conditioned at 288°C for 1225

h which, combined with residence time in the autoclave prior to and during the test, closely approximates the time for irradiations of the other L83 specimens described above. Specimen L83-18 was tested in the as-received condition. Results of 1 Hz tests in reactor-grade water environment of all three specimens (L83-1, -18, -19) are shown in fig. 2, allowing the conclusion that the response of the material to fatigue is essentially unaltered by irradiation or by an equivalent time at temperature for these materials as evaluated in this study. A similar set of tests was conducted on A508-2 material, code Q7 1. These results are shown in fig. 3. As for A533B, there is a easily observable increase in growth rates for the 17 mHz waveform as opposed to the 1 Hz waveform, but this increase is nearly identical to the increase shown in figs. la and b and as before, is due to the influence of the environment

265

W.H. Cullen et al. / Fatigue crack growth rates

RPPLIFD CYCLIC ST!ESS INTENSITY, ksl-\/F VT-7

%i33B

I

““‘I,,

C

(550

10' --r-r-r-~-q

108

I

Steel

ieactor-grade 288

/

’ 81

APPLIED CYCLIC STRESS INTENSITY, ksr\/T

ia

!0

!OB

j

”

18-j

ii--/_R50B-2

i

Hater

10-a

Steel--1rradlated

Reactor-grade 288

F)

10

C

Water

(550

F)

Q71-9 L

Spec.

L83-1

r 17

mHz

Irradiated

Q71-1 t 1 Hz

0

Spec.

sinewave

I

s1newave

183-18

Unirradiated Conditioned 3

Spec.

”

Steel--Unirrad.

. 071-4 I r,eWa”e

183-19

As-received

‘- i: c .i

Ik

I iB-5

/u

L

10'

10” WPlIfll

CYCLIC

STRLSS INTENSITY, HPaVX-

d------‘i--L”-! IB8

10’

WPLIED

I I I

I

/

,//l/Il 10?

CYCLIC STRESS INTENSITY, MPa\/K

Fig. 2. Fatigue crack growth rates versus applied cyclic stress intensity factor for as-received, unirradiated but conditioned at reactor time and temperature, and irradiated A533B steel in high-temperature, pressurized reactor-grade water. These three overlapping data sets indicate that irradiation under conditions of this initial study has an essentially insignificant effect on fatigue crack growth rate.

Fig. 3. Fatigue crack growth rates versus applied cyclic stress intensity factor for irradiated and unirradiated, but reactor time- and temperature-conditioned, ASO%2 steel. These results are very similar to those for A533B shown in figs. 1 and 2. In this case, the u~adiated specimen tested at 1 Hz shows a slight, but measureable decrease in crack growth rates.

during the longer period waveform. Since the test on Q7 1-9 was carried out to a longer final crack length, and hence higher AK value than L83-6, the onset and extent of Stage II growth is clearly described in this case. Note that the increase in growth rates between the 1 Hz and 17 mHz results is about a factor of twenty, while the increase in cyclic period is about sixty. The relationship is not, therefore, one-to-one, and in fact, if irradiation has little or no effect, we can expect, on the basis of unirradiated results [ 12,131 that the 17 mHz waveform affords a saturation or maximization of the en~ronmental effect, for PWR environments, and thus, the 17 mHz results of figs. 1 and 3 represent upper limits of fatigue crack growth for the R = 0.2 constant-load-amplitude condition. The comparison of 1 Hz test results for irra-

diated, and unirradiated but time-and-temperatureconditioned specimens, also shown in fig. 3 indicates that the irradiated results are somewhat lower than the unirradiated results, although the difference is not clearly greater than the commonly accepted scatter band (a factor of about two) for similar, intralaboratory fatigue crack growth rate tests. In the unirradiated condition, A508-2 forging code V82 can be characterized, on the basis of figs. 4a and b as being slightly more environmentally susceptible than the code R material which was fully described in earlier studies [12,15]. However, the 17 mHz results for both irradiated and unirradiated specimens, shown in fig. 5, indicates that irradiation does not further aggravate the levels of fatigue crack growth in this material.

266

W.H. Cullen et al. /Fatigue

crack growth rates

APPLIED CYCLIC ST$SS INTENSITY, list-\T, Ia8

Ia I I,,,,, Steel

' R508-2 t 288

C

(550

I

1

RPPLIED CYCLIC STYESS INTENSITY, ks'-"JF

Ia ,1'/111,~ / /

F)

Ia8

I0

Ia

:

o

Spec.

K2-05

1m6

ai RPPLIED CYCLIC STRESS INTENSITY, MPa\/m

RPPLIED CYCLIC STRESS INTENSITY, MPa\fi

Fig. 4. Fatigue crack growth rates versus applied cyclic stress intensity factor for unirradiated, as-received A508-2 steel in hightemperature, pressurized reactor-grade water. The code V82 heat shows slightly more susceptibility to corrosion fatigue crack growth, as shown in (a), for 60 s ramp/l s reset waveforms and (lb), for 17 mHz sinusoidal waveforms. RPPLIED CYCLIC STRESS INTENSITY, kw\/T lBB

IB

I0

4. Discussion Reactor-grade Water 18“ 1 -A Spec. VB2-1--1rrad. _o

Spec.

V82-8

mHz

slnewave

-17 u

! /

-R=0.2

-Ia‘*") :: _ ?

;, ? I

c

:

18" -

-

;

-

E

i

1

-

E

::

-

8 :l@

,% 2 & k

Ia-' -

ii 16

/ IBE

/

1 IllJIll

1 I IB'

/ /111111 lE2

RPPLIED CYCLIC STRESS INTENSITY, MPa\K

/

-5:: g

_

Y

-

? e

I

The overall result derived from this study, that irradiation does not enhance crack growth rates for simulated nuclear coolant environment, is parallel to the earlier results for tests in high and low temperature air environments [ 151. In those instances of irradiated specimen results for which the growth rates are depressed (A508, code Q71, 1 Hz tests), this may be due to the increased yield strength of the irradiated specimen. In those cases for which the irradiated results are slightly higher than for comparable tests of similar materials (17 mHz tests-all cases), the increase may be due to a synergistic effect of the environmental influence on the radiation damaged

Fig. 5. Fatigue crack growth rates versus applied cyclic stress intensity factor for irradiated and unirradiated, as received A508-2 steel in high-temperature, pressurized reactorgrade water. There is no observable effect of irradiation from these initial data.

W.H. Cullen et al. / Fatigue crack growth rates

261

inco~orated into a larger test matrix:However, for simulated pressurized water reactor conditions, and constant-load-amplitude tests, R = 0.2, the effects of irradiation do not provide any significant changes to the expected fatigue crack growth rates. In particular: (a) For tests with a 1 Hz sinusoidal waveforms, tests of irradiated and unirradiated but temperatureconditioned specimens showed that irradiation had no discernible effect (A533B) or may tend to slightly decrease the fatigue crack growth rates fA508-2). (b) For tests with a 17 mHz sinusoidal waveform, the results showed the expected, waveform-dependent increase over the 1 Hz results, but the observed growth rates did not exceed the current ASME Section XI water~nvironment default line. The irradiated specimen results did not differ significantly from results for unirradiated specimens of similar steels from different heats. (c)In compa~ng the results of the two different materials, there is little to distinguish them apart, indicating that, at least for conditions of this study, the ASME code will not have to be reformulated with qualifications for the different materials which are in use.

however the va~ability is well within a scatter band defined by previous growth rate evaluations for these materials. In the case of the A508-2 forging, code V82, which seems to be slightly more environmentally susceptible than the code Q71 material, irradiation did not provide any further detrimental effects. It is important to note that for the specific sets of conditions studied (1 Hz and 17 mHz sinusodal waveforms, R = 0.2, PWR conditions) none of the present results exceed the existing ASME default lines for aqueous environment fatigue crack growth. These results do represent the very first tests measuring high-temperature, pressurized water en~ronnlent fatigue crack growth rates for irradiated materials. Accordingly, the scope of the external variables which have been addressed is quite narrow. Constant-load-amplitude tests with R f 0.2 (e.g. R = O-7), or variable amplitude/variable frequency tests may produce different conclusions. The aqueous environment conditions used in this study were intended to be typical of those for an operating pressurized water reactor, but variability in water chemistry, or temperature, as might occur during upset conditions, may produce somewhat different results. The study of fatigue crack growth rates for steels used in boiling water reactor (BWR) pressure containments might constitute a separate study. ~thou~ the irradiation fIuences in BWR’s are substantially Iower, the water environment is of a different chemistry (deionized, 200 ppb dissolved oxygen) and is known to be more aggressive than PWR coolant. These significant differences may lead to effects which are somewhat different from those detailed above.

The investigations reported were sponsored by the Reactor Safety Research Division of the Nuclear Regulatory Commission. The continuing support of this agency is appreciated. The authors also thank J.R. Hawthorne, who provided the irradiation capsule design and irradiation’logistics.

5. S~rna~

References

steel,

and conclusions

Post-irradiation fatigue crack growth rate tests, in a simulated pressurized water reactor coolant environment, have demonstrated that irradiation does not further increase the growth rates for A508-2 and A533B steels beyond those increases which are due to the environment itself. The reader should note that the results presented here are limited in the scope of the external variables which have been treated, and that to be more complete, other waveforms, temperatures, load ratios and water chemistries should be

Acknowledgements

[ 11W.H. Cullen and K. Torronen, NRL Memorandum

Report 4298, NavalResearchLaboratory (1980). [2] Section XI of the ASME Boiler and Pressure Vessel Code, Rules for In-service Inspection of Nuclear Power plant Components ANSTIASMEBPV-XI-1 (American Society of Mechanical Engineers, New York, issued annually). [3] W.G. Clark, Jr., J.Mater. 16 (1971) 134. [4] P.C. Paris, R-J. Bucci, E.T. Wessel, W.G. Clark and T.R. Mager, ASTM-STP 513 (1972) pp. 141-176. [S] T.R. Mager and V.J. McLaughlin, HSST Technical Report 16, Westinghouse Report WCAP-7776.

268

W.H. Cullen et al. /Fatigue crackgrowth rates

(61 W.H. Barnford, in: Proc. IAEA Technical Committee Meeting on Time and Load Dependent Degradation of Pressure Boundary Materials, Innsbruck, Austria (1978). [7] ASTM E647-78T, ASTM Book of Standards, Vol. 10 (issued annually). [8] Heavy Section Steel Technology Program, Semiannual Progress Report, ORNL-4315, Oak Ridge National Laboratory (1968) p. 29. [9] H.E. Watson, F.J. Loss and B.H. Menke, in: Structural Integrity of Water Reactor Pressure Boundary Components, NRL Report 8006, Naval Research Laboratory (1976). [IO] H.E. Watson, B.H. Menke and F.J. Loss, in: Structural Integrity of Water Reactor Pressure Boundary Components, Ed. F.J. Loss, NRL/NUREG Memorandum Report 3512 M (1977).

[ll]

R.P. Wei, S.R. Novak and D.P. Williams, in: Proc. 33rd AGARD Conf. on Structures and Materials, Brussels, Belgium (1971). [12] W.H. Cullen et al., NUREG/CR 0969, NRL Memorandum Report 4063, Naval Research Laboratory (1979). [ 131 W.H. Cullen, H.E. Watson and V. Provenzano, in: Structural Integrity of Water Reactor Pressure Boundary Components, NUREG/CR-112, NRL Memorandum Report 4122 (1979). [14] W.H. Cullen, R.A. Taylor, H.E. Watson and W. Rohrs, in: Structural Integrity of Water Reactor Pressure Boundary Components, NUREG/CR-1472, NRL Memorandum Report 4254 (1980). [15] L.W. James, Nucl. Safety 18 (1973) 791.