The objective of this program is to determine the effect, if any, of stainless steel cladding upon the propagation of smallsurface cracks subjected to stress states similar to those produced by thermal shock conditions. Preliminary results fromtesting at temperatures 10 ° and 60°C below the nil-ductility-transition temperature have shown that (1) a tough surface layer(cladding and/or heat-affected zone) has arrested running flaws under conditions where unclad plates have ruptured, and (2)the residual load-bearing capacity of clad plates with large subclad flaws significantly exceeded that of an unclad plate.
Nuclear Engineering and Design 115 (1989) 349-357 349 North-Holland, Amsterdam
E F F E C T O F C O M M E R C I A L C L A D D I N G O N T H E F R A C T U R E B E H A V I O R
O F P R E S S U R E V E S S E L S T E E L P L A T E S *
S.K. I S K A N D E R , D.J. A L E X A N D E R , S.E. BOLT, K.V. C O O K , W.R. C O R W I N , B.C. O L A N D ,
R.K. N A N S T A D and G.C. R O B I N S O N
Oak Ridge National Laboratory, Oak Ridge, TN 37831, USA
Received 3 March 1988
The objective of this program is to determine the effect, if any, of stainless steel cladding upon the propagation of small surface cracks subjected to stress states similar to those produced by thermal shock conditions. Preliminary results from testing at temperatures 10 ° and 60°C below the nil-ductility-transition temperature have shown that (1) a tough surface layer (cladding and/or heat-affected zone) has arrested running flaws under conditions where unclad plates have ruptured, and (2) the residual load-bearing capacity of clad plates with large subclad flaws significantly exceeded that of an unclad plate.
1. Introduction
A small crack near the inner surface of clad nuclear reactor pressure vessels (RPV) is an important consider- ation in the safety assessment of the structural integrity of the vessel. The behavior of such flaws is relevant to the pressurized thermal shock (PTS) scenario and to the plant life extension issue.
There is a dearth of information on the behaviour of small flaws in the presence of cladding. This has led some RPV integrity studies to the assumption of in- finitely long flaws (although small flaws are certainly more Credible). The basis for the assumption of long flaws is experimental results which have shown that, in the absence of cladding, a small surface flaw in an embrittled material subjected to severe thermal shock will become a long flaw [1]. Thus, the question remains about the role a tough surface cladding will play in preventing the propagation of small flaws along the surface. Furthermore, the flaw could tunnel beneath the cladding, in which case the residual strength of the structure needs to be estimated.
* Research sponsored by the Office of Nuclear Regulatory Research, Division of Engineering, U.S. Nuclear Regulatory Commission under Interagency Agreement DOE 0551-0551- A1 with the U.S. Department of Energy under contract DE-AC05-840R21400 with Martin Marietta Energy Sys- tems, Inc.
The objective of the research program described (" the clad plate program") is to determine the behavior of small flaws in or near cladding under stress fields similar to those occurring during a PTS scenario. The potential benefit to the U.S. Nuclear Regulatory Com- mission (NRC) is an improved predictive capability of the fracture strength of a RPV with hypothetical flaws. The objectives of this research are achieved by compar- ing the load-bearing capacity of clad and unclad flawed plates. The clad plate program is conducted as part of the Heavy-Section Steel Technology (HSST) program.
A special plate specimen has been developed to investigate the cladding effects. It was commercially clad using the three-wire series-arc technique and stain- less steel type 308, 309, and 304 weld wires. The base metal on which the cladding was applied, A533 grade B steel, was heat treated to raise its transition temperature so that it would be brittle at temperatures at which the cladding would be tough. An electron-beam (EB) weld is introduced into the base metal to provide a crack initiation site [2]. The plate is loaded in four-point bending to approximate the stresses due to thermal shock and then hydrogen charged [2], while the load is maintained constant. A flaw initiates in the EB weld and propagates along the surface until the flaw encoun- ters the clad region. The running flaw is either arrested or leads to complete rupture of the plate. Plates which do not rupture are heat-tinted to define the arrested flaw shape and are then reloaded until either another pop-in or plate rupture occurs.
350 S.K. lskander et al. / Fracture behavior of pressure vessel steel plates
The first series (Series 1) of such experiments indi- cated that the cladding employed may have had suffi- cient arrest toughness to stop running cracks [3]. The bounding of the contribution of the stainless steel clad- ding on the arrest toughness was not realized.
2. Test specimen
The specimen used in Series 2 has been designed and fabricated to avoid the problems encountered in the first series. One of the problems was the unknown effect of a groove (machined through the stainless steel clad- ding to put an EB weld in the base metal) on the stress intensity factors. The specimen design for Series 2 al- lows the EB weld to be produced on the surface. Ap- proximately, the EB weld zone is semi-elliptic in shape, 70 mm long with the major axis along the surface, and 15 mm deep. A drawing of a clad-plate specimen, together with the load points to produce a constant moment over the EB weld, is shown in fig. 1.
The Charpy impact energies as a function of temper- ature for the LS and LT orientations (reproduced from
ref. [4]) are shown in fig. 2. In the weld overlay clad- ding, the orientations referred to are those of the un- derlying base plate. The LT orientation is relevant to flaws propagating along the surface and normal to the plate roiling direction. With respect to the EB weld, the LT orientation corresponds to that of a flaw initiating in the EB weld and propagating in the direction of the major axis of the EB weld. The LS orientation corre- sponds to flaws propagating in the thickness direction. In particular note that the Charpy energy of the H A Z is significantly higher than that of either cladding or base metal over the temperature range - 2 5 to 2 5 ° C at which the clad plates were tested.
The yield strength and ductility of the clad plate materials, also from ref. [4], are shown in fig. 3. The yield strength and ductility of the HAZ and base metal are very similar, and at clad-plate test temperatures the cladding is significantly more ductile but lower in yield strength than the base metal and HAZ.
* The nomenclature used to designate crack plane orientations such as LS and LT is that recommended in A S T M Standard Terminoh)gy Relating to Fracture Testing (E 616).
f,,~),j,.5~.~O v
J / / - /
P 2
5
P/2
~ . P2
D I M E N S I O N S IN
C E N T I M E T E ' R S
C L A D D I N G ,~ 3-4 mm T H I C K
Fig. 1. A drawing of a clad-plate specimen, together with the load points to produce a constant moment over the flaw.
S,K. Iskander et a L / Fracture behavior of pressure vessel steel plates 351
7 0 0 I I I I 1
A 1 4 T BASE METAL
175 [3 . . . . . . . . . . O H~,Z
O - - - - - ~ CLADDING 150
[3 D
125 TEST M C] i - i
~- T E M P E R A T U R E / . A
,.~ R A N G E - ~ ~ ~" Cc 100
7 5 D." O ,5
0T
200 , , I I I I I
A A 1 4 T BASE METAL
175 O .......... 0 HAZ
150 O - - - - " 0 CLADDING
O
125 TEST 0 . , " ' " " [ 3
T E M P E R A T U R E / / . ' " " " A
R A N G E ~ ' ~ . ' " ~, 100 E °.1"° ? I f
. . . . . . . .
~'""~/=o so , . , / ~
".~-- N D T
, & , , , J , ¥ , , I , , , I , , ~ I ,
0 . 200 100 0 100 200 300 400
TEMPERATURE I °C I
Fig. 2. Charpy impact energy as a function of temperature for the LS and LT orientations.
Both the higher Charpy energy of the H A Z and high ducti l i ty of the cladding have played prominent roles in the results of the clad plate tests.
3. General description of the test
In the six tests performed, the general test procedure was essentially the same. The instrumented plate was mounted in a 1 - M N Instron testing machine. For tests at other than room temperature, the plate was cooled to the specif ied temperature. The variation in temperature at various locations was kept within 3 ° C. The plate was then loaded in f6ur-point bending to induce a pure bending moment over the span including the EB. The
loads were chosen to induce a specif ied strain level on the surface of the base metal. The EB weld was then hydrogen charged, while the load was maintained con- stant using stroke control until a flaw initiated. This portion of the experiment on an initially unf lawed plate was thus essentially an "arrest" experiment, the purpose of which was to study the effect of c ladding on a running flaw.
In cases in which the flaw arrested, the plate was removed from the testing machine and heat-tinted * to define the arrested flaw shape. Some nondestructive
* For a low-allow steel, no significant effects on the crack tip material properties are expected from heat-tinting at 325 °C for 1 h.
352 S.K. Iskander et al. / Fracture behavior of pressure vessel steel plates
Fig. 3. The yield strength and ductility of the clad plate materials.
100
examination was performed on the first and second plates tested to determine the extent of flaw prop- agation. After partial reinstrumentation, the plate was put back into the testing machine and cooled to a specified temperature. The load was increased at a uniform rate until the plate either ruptured or further pop-in occurred. In the case of the latter event, the process was repeated. The purpose of reloading the arrested flaws was to obtain data on the residual load-bearing capacity of flawed clad plates. This por- tion of the experiment was designated an " ini t ia t ion" one, in contrast with the "arres t" portion previously described.
A data acquisition system was used to record the readings from the instrumentation on magnetic tape at
suitable time and load intervals. In addition to the strains, crack opening displacement (COD), and the temperature at various locations in the plate, the load and the displacement of the testing machine ram were also recorded.
Using an assumed flaw shape that corresponded approximately to the EB weld zone, the loading rate was chosen to be within the range prescribed by A S T M Standard Test Method for Plane-Strain Fracture Tough- ness o f Metallic Materials (E 399), 0.55 to 2.75 MPa • V ~ , S - 1 "
For the six plates tested, the target surface strains and corresponding loads are given in table 1. For ap- plicable cases the load that the specimen can support after pop-in, the "arres t" load, is also given. The surface
S.K. Iskander et al. / Fracture behavior of pressure vessel steel plates
Table 1 Target surface strains and corresponding loads for the six plates tested.
353
Plate Condition Test Load (kN)
temperature a Pop-in (°C)
Arrest b
Target surface strain (~)
CP-15 Clad
CP-17
CP-19
CP-21 CP-18
CP-20
Clad Several pop-ins
occurred before rupture
Clad
Unclad Clad
Clad
RT 676 654 0.31 - 25 759 709
- 100 600 R RT 890 823 0.45
- 25 756/725 R
RT 987 689 0.65 - 5 0 703 R
RT 676 R 0.27 - 25 823 649 0.39 - 25 698 R - 25 868 R 0.41
a RT = room temperature, approximately 25 o C. b R = plate ruptured in two pieces.
s t ra in (in the un i fo rm bend ing m o m e n t span of the plate) is the independen t pa ramete r used in the selec- t ion of the target load. All arrest experiments were per formed at ei ther - 2 5 or 25°C. The loads (and strains) were ma in ta ined cons tan t under stroke control dur ing the per iod of hydrogen charging and are a mea- sure of the crack-driving force act ing dur ing the ins tan t the flaw ini t iated in the EB weld.
For the first plate tested, CP-15, the surface s train was chosen to be approximate ly the yield s train of the base metal. The plate did not rupture. The target surface s t ra in was increased for each of plates CP-17 and CP-19. In all three cases the flaw ini t iated and arrested after
1 0 0 0 B
O INIT IAL LOAD
• ARREST LOAD N 0 - -
- - 8 0 0 - -
o • ,J 7 ~ - -
C P - 2 1 O (FAILED)
S 0 0 - -
U N C L A D PLATE
S 0 0 _ _
C P - 1 9
C P - 1 7
l CP-15
C L A D PLATES
Fig. 4. Pop-in, arrest loads, and corresponding crack lengths for the four plates tested at room temperature.
p ropagat ing benea th the cladding, a dis tance tha t in- creased with increasing target strain. A n unclad plate, CP-21, ruptured when loaded to approximate ly the base metal yield s train on the surface. The pop-in, arrest loads, and cor responding crack lengths for the four plates tested at room tempera ture are shown schemati- cally in fig. 4.
The detai led descr ipt ion of the test on only one of the six plates is provided below. For details on the remaining plates the reader is directed to ref. [5].
4 . T e s t i n g o f p l a t e C P - 1 5
Plate CP-15 was the first three-wire clad plate tested. It was loaded at room tempera ture to a surface s t ra in of approximate ly 0.31% (in the un i fo rm m o m e n t span of the plate), cor responding to the yield stress at room tempera ture of the base metal of 590 MPa, requir ing a load of 676 kN. It was then hydrogen charged while the load was main ta ined cons tan t using stroke control . Pop- in occurred within abou t 1 h. The pla te was removed from the test ing machine and hea t - t in ted at 325°C. The plate was reinstalled, cooled to - 2 5 ° C , and loaded at a cons tan t d isp lacement rate. A second pop- in occurred at a load of 759 kN. The plate was removed f rom the test ing machine and hea t - t in ted at 250°C. The plate was reinstal led in the test ing machine , cooled to - 100°C, and the load increased at a un i fo rm rate unt i l
354 S.K. Iskander et al. / Fracture behavior of pressure vessel steel plates
Fig. 5. The broken halves of the 150-kg test specimen CP-15 with heat-tinted shapes of the first and second pop-ins.
the plate broke completely at 600 kN. Fig. 5 shows the two broken halves of the 150-kg test specimen with heat-tinted shapes of the first and second pop-ins. The fracture surface has been examined in detail and is described in a later section.
After the first pop-in, a surface crack in the HAZ of
Fig. 6. Surface crack in the HAZ of the EB weld of plate CP-15 after the second pop-in (top center), extending as dimples into the cladding on either side. The crater on the left of the crack
was formed during EB welding and not during testing.
the EB weld could be discerned, but it did not appear to have extended on the surface. Instead, a small concave dimple extended axially from the ends of the flaw, indicating that perhaps the flaw extended under the surface. The extension of the axial dimple stopped short of the cladding. The crack had actually propagated in the base metal just below the surface until it encoun- tered the HAZ with a very thin layer of metal covering the flaw.
During the second pop-in, the surface flaw appeared to have extended in the base metal along the previously dimpled surface to a length of about 7 cm, but still short of the 10 cm needed to reach the cladding. How- ever, the dimple extended into the cladding about 5 -10 mm. The surface extension of the flaw in the base metal was merely the rupturing of the thin layer mentioned above. Fig. 6 shows the surface crack in the H A Z of the EB weld after the second pop-in (top center), extending as dimples into the cladding on either side. The crater on the left of the crack was formed during EB welding and not during testing. Note the buckled weldable strain gages, indicating the large C O D to which they had been subjected.
After the second pop-in, due to the lack of observa- ble flaw extension along the surface, the plate was X-rayed, and examined ul trasonical ly with dye penetrant. The X-ray and ultrasonic examinations indi- cated a subsurface flaw about 27 cm long. The dye penetrant failed to reveal any extensions of the flaw on the surface beyond those observed visually (see fig. 7). The thin surface layer over the port ion of the flaw residing in the base metal had only partially ruptured, thus blocking the dye penetrant from indicating that portion of the flaw.
S.K. Iskander et al. / Fracture behavior of pressure vessel steel plates 355
Fig. 7. Dye penetrant (examination of plate CP-15 after the second pop-in) failed to reveal any extensions of the flaw on the surface beyond those observed visually.
5. Fraetography of plate CP-15
The fracture surface of plate CP-15 was examined to determine the profile of the crack front at the arrest events. For this specimen, the initial fracture and arrest occurred at room temperature, the second fracture and arrest occurred at - 2 5 o C, and the final fracture of the
plate was at - 1 0 0 ° C . Heat- t int ing was used to mark the crack front after the two arrest events.
Preliminary examinations were made of the fracture surface after the final fracture. Optical micrographs
were also made. The fracture surfaces were then cut off the fractured specimen halves. One of the resulting pieces was then cut into 11 sections to allow port ions of
Fig. 8. Fractography of clad plate CP-15. There is no evidence of ductility at the arrest locations. The second event resulted in crack arrest in the cladding on the right-hand side of the flaw.
356 S.K. Iskander et a L / Fracture behavior of pressure vessel steel plates
the surface to be inserted into the scanning electron microscope (SEM) for further examination.
5.1. First fracture event
The initial event, which occurred at room tempera- ture, initiated in the HAZ of the EB weld and propa- gated down into the plate as well as extending out to either side. Near the surface, the crack was arrested on either end by the HAZ associated with the cladding. The result was the bowed-out shape shown in fig. 5. The fracture mode during this event was cleavage, all the way to the arrest location. There is no evidence of ductile fracture at the arrest point, as is shown in fig. 8.
5.2. Second fracture event
The second fracture and arrest occurred at - 2 5 o C. The crack extended from the first arrest point into the plate, extending much further to either side and also through the thickness of the plate. SEM examination showed that the crack was again arrested by the HAZ boundary on one end of the crack, here termed the "left-hand side". Again, the fracture mode was clea- vage, all the way to the arrest point. In the center of the plate the crack also extended by cleavage to its arrest location. However, at the other end of the flaw, here referred to as the "right-hand side", the flaw was able to penetrate the HAZ and did actually come into con- tact with the cladding. In this case, the HAZ was fractured over a distance of roughly 2.54 cm, at which point the flaw deviated and returned to follow the HAZ boundary. The distance over the which the HAZ was penetrated matched the width of the initial strip of cladding. The crack front deflection occurred at the point where the HAZ of the second strip of cladding encountered the HAZ from the first strip. The crack grew through the first strip of HAZ by a cleavage mechanism, and then hit the stainless steel cladding. Near the edge of the cladding, the flaw penetrated into the cladding for as much as 0.8 mm. The amount of penetration decreased further from the edge of the cladding until the flaw followed the cladding-base metal interface. Fracture of the stainless steel cladding oc- curred by a ductile mechanism, as the primarily austenitic material will not cleave.
HAZ. Fracture in the cladding tended to follow the ferrite phase, resulting in a relatively brittle fracture appearance. In this case, however, there was no indica- tion that the crack had stopped at the interface between the cladding and the HAZ, or at the HAZ-base metal boundary.
6. Discussion of results
The tough surface layer of cladding and HAZ seemed to have contributed significantly to the load-bearing capacity of the plates by arresting flaws at loads and temperatures that have ruptured unclad plates, as seen by comparing the results of the tests given in table 1 for plates CP-15 and CP-21. In fact, the clad plate CP-19 arrested a flaw subjected to a driving force (as measured by the target load) almost 50% high than that which broke an unclad plate. Moreover, the residual load- bearing capacity of plates, as measured by the critical loads in initiation experiments with fairly large flaws, was generally greater than required to break the unclad plate, even though the test temperatures were lower by 50°C.
The tests indicate a propensity of propagating flaws to tunnel, even without the aid of the tough surface layer composed of cladding/HAZ, as has occurred in the base metal portion of the clad plates and in the unclad plate. This potential for tunneling is due to the location of the maximum stress intensity factor for short flaws occurring somewhat below the surface when the flaw is in a stress gradient, as in these tests.
The ductility of cladding appears to have been a necessary ingredient in increasing the load-bearing capacity of clad plates. However, it is not clear at this time whether cladding alone, without benefit of the tough HAZ which played a pronounced role in arresting propagating flaws, would have also elevated the load- bearing capacity. In the case of radiation embrittled reactor pressure vessels, the HAZ will most likely un- dergo toughness degradation similar to that of base metal, and would therefore not play such a prominent role in arresting propagating flaws.
Nomenclature
5.3. Final structure
The final fracture at - 1 0 0 ° C resulted in the com- plete fracture of the plate. Fracture occurred by clea- vage through the remaining areas of base metal and
COD CP-n EB HAZ HSST
Crack opening displacement Clad plate number n Electron beam Heat-affected zone Heavy Section Steel Technology
S.K. lskander et al. / Fracture behavior of pressure vessel steel plates 357
N R C O R N L PTS RPV SEM
U.S. Nuclear Regulatory Commission Oak Ridge National Laboratory Pressurized thermal shock Reactor pressure vessel Scanning electron microscope
References
[1] S.K. Iskander, A method of LEFM analysis of RPV during SBLOCA, Int. J. Pressure Vessels Piping 25 (1986) 279-298.
[2] P.P. Holz, Flaw preparations for HSST program vessel fracture mechanics testing: Mechanical cyclic pumping and electron beam weld-hydrogen-charge cracking schemes, NUREG/CR-1274 (ORNL/NUREG/TM-369) , Oak Ridge National Laboratory (May 1980).
[3] W,R. Corwin et al., Effect of stainless steel weld overlay cladding on the structural integrity of flawed steel plates in bending - Series 1, NUREG/CR-4015 (ORNL/TM-9390), Oak Ridge National Laboratory (April 1985).
[4] W.R. Corwin, Stainless steel cladding investigations, in: Heavy-Section Steel Technology Program Semiannual Prog. Rep. October 1985-March 1986, NUREG/CR-4219, Vol. 3, No. 1 (ORNL/TM-9593/V3&N1), Oak Ridge National Laboratory (June 1986) pp. 70-87.
[5] S.K. Iskander et al., Crack arrest behavior in clad plates, in: Heavy-Section Steel Technology Program Semiann. Prog. Rep. April-September 1987, NUREG/CR-4219, Vol. 4, No. 2 (ORNL/TM-9593/V4&N2), Oak Ridge National Laboratory (April 1988) pp. 222-242.
