ML18065A310
ML18065A310 | |
Person / Time | |
---|---|
Site: | Palisades ![]() |
Issue date: | 10/31/1994 |
From: | ABB COMBUSTION ENGINEERING NUCLEAR FUEL (FORMERLY, ASEA BROWN BOVERI, INC. |
To: | |
Shared Package | |
ML18065A309 | List: |
References | |
CEN-406-NP, NUDOCS 9512010092 | |
Download: ML18065A310 (23) | |
Text
- ATTACHMENT 1 CONSUMERS POWER COMPANY PALISADES PLANT DOCKET 50-255 CEN-406-NP -A STATUS.REPORT ON CEOG ACTIVITIES CONCERNING PRIMARY WATER STRESS CORROSION CRACKING OF INCONEL-600 PENETRATIONS
. -:*:* .. *.' . ;-, ** _.t 1:
I' . ; ; : . ' .
- -:.. tJ*i:,;:.*.
- COMBUSTION . ENGINEERING-*
OWNERS GROUP ; ..
..
-.:* *-**'** *, -. ... *" .* . . *
- * : . . *1* .... '
..
..
..
f.' r.. .
- ... /}\_;*_ ::; l:-:) -. CEN-406-NP . . A STATUS REPORT ON CEOG ACTIVITIES CONCERNING PRIMARY.WATER STRESS CORROSION CRACKING OF INCONEL-600 PENETRATIONS October 1995 \}.' **.-ABB Combustion Engineering Nuclear Operations
- Jl 1111 ,.,1919 f *., . *. © Copyright 1995, Combustion Engineering, Inc. : __ ... i ... -'. , -r . :*
- LEGAL NO.TICE '!_'his report was prepared as an account of work sponsored by the Combustion Engineering Owners Group and ABB Combustion Engineering.
Neither .Combustion Engineering, Inc. nor any person acting on its behalf: A. makes any warranty or representation, express or implied including the warranties of fitness for a particular purpose or _merchantability, . with respect to the accuracy, completeness, or usefulness of the information contained in this report, or that the use of any information, apparatus, method, or process disclosed in this *report may not infringe privately owned rights; or B. assumes any liabilities with respect to the use of, or for damages resulting from the use of, any information, apparatus, method or process disclosed in this report. Combustion Engineering, Inc. . -... v . ' **,,_. ',. ,' , ',, . ,_ .. ,. * ...... **:, : _. *:-:
- i *, ' . *: ....... . ,. 'i* ,* .. **.:* ' . .-;.
- TABLE OF CONTENTS SECTION TITLE
1.0 INTRODUCTION
2.0 3.0 1.1 Purpose 1.2 Approach 1.3 Scope PRIMARY WATER STRESS CORROSION CRACKING OF INCONEL-600 PENETRATIONS
2.1 Failure
Analysis 2.1.1 Pressurizer Heater Sleeve Examinations
2.1.2 Pressurizer
Instrument Nozzles Evaluation
2.1.3 Heater
Sleeve Thermal Analysis 2.2 Susceptibility Analysis 2.2.1 Evaluation of Pressurizer Heater Sleeve Susceptibility to PWSCC 2.2.2 Evaluation of Instrument Nozzle Susceptibility to PWSCC 2.2.3 Primary Pressure Boundary Penetrations 2.2.4 Eddy Current Imaging Development
2.3 Safety
Significance Analysis 2.3.1 Pressurizer Base Metal Corrosion/Erosion
SUMMARY
1 PAGE 1 2 2 2 3 5 5 11 13 14 14 15 16 17 17 17 19
- *
1.0 INTRODUCTION
During a routine in-service inspection of Calvert Cliffs Unit 2 (CC-2) on May 5, 1989, indications of primary coolant leakage, evidenced by the presence of boric acid deposits, were noted on the bottom head of the pressurizer around some of the Inconel-600 heater sleeves. Visual examination identified 20 of the 120 heater sleeves were leaking. Subsequent in-place dye penetrant and eddy current tests determined that all 20 sleeves had axial indications.
To determine the nature of the leakage, Baltimore Gas and Electric Company (BG&E) removed three sleeve samples from CC-2 for further nondestructive and destructive testing. ABB Combustion Engineering (ABB-CE) performed destructive examinations on two of the three sleeves which resulted in the following conclusions:
... The leakage at the heater sleeve penetrations in CC-2 resulted from primary water stress cracking (PWSCC) which initiated on the inside diameter (ID) of the sleeves and propagated through-wall, The cracks were all axially oriented with no circumferential extent indicating that the major stresses were in the hoop direction, The J-weld between the sleeve outside diameter (OD) and the pressurizer cladding was not defective, ... The sleeve material had high strength, a microstructure near the ID characterized by banding and a lack of correspondence between carbide distribution and grain boundaries.
lnconel-600 with these characteristics is susceptible to PWSCC, ... Contaminants known to promote intergranular stress corrosion cracking (IGSCC) in Inconel-600 were not present on the crack surfaces and the composition of the oxide film on the crack surfaces was not consistent with films that develop under strongly acidic or caustic conditions.
This is a further indication that the failure mechanism was probably PWSCC, and ... A review of the material and .fabrication process records indicated that the CC-2 sleeves received a pre-installation reaming operation to increase the ID by 0.005-0.015 inch. The destructive examination indicated the presence of a layer of smeared metal, higher hardness and possible tensile residual stresses in the reamed areas as 1
- compared to the non-reamed areas. The destructive examination concluded that this operation probably was the key to initiating PWSCC in the CC-2 sleeves. After a presentation of these findings to the NRC in September of 1989, the staff requested additional information concerning the susceptibility of other Combustion Engineering Owners Group (CEOG) plants to PWSCC. 1.1 Purpose The purpose of this report is to document the status of the CEOG's assessment of primary water stress corrosion cracking of Inconel-600 primary pressure boundary penetrations in CEOG plants and to demonstrate that the issue has no safety significance.
1.2 Approach
- To address the concerns of the NRC, the CEOG Inconel-600 Working Group was established and a program initiated to evaluate PWSCC of Inconel-600 primary pressure boundary penetrations in CEOG plants with respect to susceptibility, safety, operations, and maintenance.
To this end, the effort was structured to address three specific questions:
How did the cracking occur? Where else could it potentially occur? And what is the safety significance of its occurrence?
The working group includes expertise in the areas of design, fabrication, metallurgy, and corrosion.
The group has met several times since November 1989, initially to develop the program objectives and later to discuss the status of ABB-CE's work to date. 1.3 Scope The scope of the Working Group's program includes three Inconel-600 component groups: pressurizer heater sleeves, pressurizer instrument nozzles, and all other primary pressure boundary penetrations.
Data concerning fabrication processes and material characteristics of these components were compiled for all domestic Combustion Engineering NSSS Owners. 2
- 2.0 PRIMARY WATER STRESS CORROSION CRACKING OF INCONEL-600
- PENETRATIONS To address the NRC's specific concerns regarding the susceptibility of other CEOG plants to PWSCC, ABB-CE first conducted a review of the design, materials data, and fabrication history of Inconel-600 heater sleeves to identify those elements believed to affect susce£tibility.
[ * . j as with any type of intergranular stress corrosion cracking, PWSCC requires the presence of three elements; an aggressive environment, a susceptible metallurgical condition, and high tensile stresses.
The operating environment is essentially the same for all CEOG pressurizers, consisting of 653°F deaerated primary water (boric acid, LlOH, H 2). ltidustry analysis of this degradation mechanism indicates that this environment, high temperature relatively pure water, will stress corrosion cracking if the other two required elements are also present. As defined by laboratory investigations and actual plant experience, the second element, a susceptible metallurgical condition, is characterized by high yield
- strength, fine grain microstructure, and intragranular carbides.
The final anneal temperature and* cooling rate determine the magnitude and qualities of these' characteristics and differ from plant to plant The heat treatment process, however, was basically the same for all pressurizer heater sleeve materials used in CEOG plants, specifically that the material was cold drawn and annealed.
Given that the yield strength is affected by grain size and carbide distribution, the relationship is such that high strength material normally will also possess the undesirable microstructural characteristics.
The. third element necessary for PWSCC lo occur is the presence of high tensile stresses.
One source of tensile residual stresses was the J-weld between the sleeve and pressurizer cladding.
[ J ABB-CE believed that if the J-weld was the sole source of stresses, sleeve material of the same heat and lot, in other plants should have cracked as well. Therefore, it was believed that another source of residual stresses must also exist The J-weld resulted in local weld shrinkage.
To allow heater installation, Combustion Engineering reamed the sleeves to a larger ID. For a short time, this practice was modified and the sleeves were reamed prior to installation.
When. this did not eliminate the weld shrinkage the practice was discontinued and the initial practice of machine reaming after welding was reinstituted.
The 3
- pre-installation reaming removes a significant amount of material from the ID of the sleeves. Depending on tool condition, reaming may produce either tensile or compressive residual stresses, tensile if a dull tool is used. At CC-2 the smeared metal on the ID surfaces in the reamed area seemed to suggest that iostal.lation reaming may have been accomplished with a dull tool. Therefore, the cold-work induced in* the material as a result of pre-installation reaming was initially thought to have produced the high tensile residual stresses necessary for PWSCC. [ J an aggressive environment is a common denominator for all CEOG plants, it need not be included as a screening criterium when determining plant susceptibility to PWSCC. The two criteria that were used to determine relative susceptibility weret. ] It should be noted that many other vanaoles exist that may increase . the susceptibility of PWSCC in cold drawn,. annealed material.
[ ** * * * *. ] relative susceptibilities are based on laboratory investigations and actual plant experience to date. * *. Review. of the collected data revealed that[ ] Each of the CEOG plants have inspected their pressurizer beater sleeves with no indication of leakage to date. As discussed in the following pages, ABB-CE no longer believes that the high tensile residual stresses resulted from the practice of pre-installation reaming. 4
- However, the cold-work induced in the material as a result of the pre-installation reaming-did produce -a very hard layer at the ID surface that is capable of supporting higher resid:.ial tensile stresses from the J-weld. Therefore,[
J To further assess primary water stress corrosion cracking of Inconel-600 primary pressure boundary penetrations in CEOG plants, the working group designed a program to address three areas: failure mechanism, susceptibility, and safety significance.
On behalf of the CE Owners Group, ABB Combustion Engineering initiated several tasks to address each of these areas. 2.1 Failure Analysis 2.1.1 Pressurizer Heater Sleeve Examinations As previously discussed, metallurgical conditions and residual stresses in heater sleeves are two key parameters of PWSCC. The purpose of this task was to expand the data base on these parameters for cracked and non-cracked sleeves to more accurately assess the relative susceptibility of heater sleeves. All of the pressurizer heater sleeves from CC-2 not previously used for various destructive examinations (115 total) were shipped to ABB-CE's Windsor facilities for use in this task. Of these, ABB-CE selected three sleeves for examination, which based on field observations were leaking. ABB-CE also selected three sleeves which did not leak and, based on field non-destructive examination (NDE), did not contain part through-wall cracks. ABB-CE also selected 30 sleeves which did not leak and which were not inspected prior to removal from CC-2. Finally, ABB-CE retrieved from storage for reexamination the remnants of two Arkansas Nuclear One Unit 2 (AN0-2) sleeves, one of which in April, 1987 cracked due to PWSCC,[ *
- J Conditions*
All 38 sleeves were examined by dye penetrant tests to determine if any cracks were present in the reamed area below the weld in the CC-2 sleeves, or if cracks were present at any location in the AN0-2 sleeve remnants.
ABB-CE then conducted a series of metallurgical tests and evaluations on the three confirmed s
- leaking cc.:2 sleeves, the three non-leaking sleeves with no NDE indication, of the non-leaking CC-2 sleeves that were not inspected, and both AN0-2 sleeves. These tests included a microstructure evaluation, microhardness evaluation, and metallographic evaluation.
Dye Penetrant Tests Thirty-eight sleeves were dye penetrant tested to determine if linear like) indications were present. There were no indications in any of the CC-2 sleeves that were non-leakers based on visual examination in the field (33 sleeves) or in the non-leaking AN0-2 sleeve. There were linear indications in two CC-2 sleeves and the remaining AN0-2 sleeve, all of which reportedly leaked in service. The remaining CC-2 sleeve which based on field observations was a leaker, showed no linear indications.
Apparently any defects were located_ above the sleeve removal cut. Microstructural Evaluation The objective of this evaluation was to characterize the metallurgical condition of the sleeves (grain shape and size and carbide distribution) to determine if there was a basic difference between the cracked and cracked sleeves. This objective was accomplished using a dual etch technique.
An orthophosphoric acid etch was first performed to reveal the carbides' location.
A nital etch was then performed to reveal the grain boundaries.
The evaluation was performed at three locations for each of the 10 samples: near the ID surface, at approximately mid-wall, and near the OD surface. The evaluation indicated that the carbides were essentially all intragranular, with little, if any, correspondence between the carbides and the grain boundaries.
- In all cases, at least some carbide banding was present, indicating that the final annealing temperature was low. The carbide bands were aligned in the axial direction and were present in both cracked and non-cracked sleeves. In many areas, carbide distribution suggested a grain boundary-like network indicating carbide precipitation on a prior history grain boundary.
Subsequent heat treatment was sufficiently high to recrystallize the cold-6
- worked microstructure, thereby forming new grain boundary networks but not sufficiently high to dissolve carbides.
The available data does not identify the specific temperature used, although, a final heat treatment of [ . ] The samples contained both coarse and fine grain areas. The average grain size of the coarse grain areas ranged from ASTM 65 to 8.5 and in the fine grain areas, size varied from ASTM 8.5 to 11. Again, there was no apparent difference between sleeves that cracked and those that did not. The AN0-2 sleeve microstructures were similar to CC-2 microstructures.
As part of this task, ABB-CE compared the reamed and non-reamed areas of the CC-2 and AN0-2 sleeves to determine the presence/absence of smeared or deformed metal. Photomicrographs revealed that all of the CC-2 sleeves had a layer of smeared metal with a thickness of up to 2 mils around the circumference of the sleeve. The layer of smeared metal was generally continuous, although there were some breaks and areas where the thickness was less. There were no apparent differences in the reamed areas in sleeves which leaked and those which did not The. leaking AN0-2 sleeve which had been reamed over its entire length, had smeared metal present in the area where cracks developed.
The non-leaking AN0-2 sleeve was not reamed. In some of the CC-2 sleeves, smeared metal was present in localized areas remote from the area of pre-installation reaming. This is probably the result of J2,QS.1-installation reaming. Pre-installation reaming removed a relatively large amount of metal, while post-weld reaming used a reamer of smaller diameter, removing small amounts of metal from local high spots.
Eyaluati6n ABB-CE also determined the microhardness of both reamed and reamed areas of the 10 sleeves examined.
Measurements were taken from the ID surface to a depth of approximately 10 mils at 1 mil increments.
Some of the sleeves examined showed significant differences in microhardness for reamed and non-reamed areas. In general, the reamed areas showed notably higher hardness near the ID, which decreased with 7
- depth and eventually reached the same level on the non-reamed area.. However, there were no differences apparent between reamed and reamed areas for some of the sleeves examined.
In these few cases the microhardness in the reamed area was as low as that in the non-reamed area. And, once again, there were no apparent differences in the hardness levels or depth of hardness between the cracked and non-cracked sleeves.
Evaluation The crack surfaces in four CC-2 sleeves and the cracked AN0-2 sleeve were examined.
All of the cracks were intergranular.
There was no evidence of intergranular attack (IGA) or shallow intergranular penetrations on the ID surfaces that could have served as initiation sites for IGSCC. The cracks were branched to some extent but they were all clearly discrete cracks as opposed ta volumetric IGA The cracks were all axial in orientation indicating that the maximum stresses were in the hoop direction.
The additional pressurizer heater sleeve examinations resulted in the following conclusions:
{ .. .. J The failed AN0-2 sleeve was equivalent to the CC-2 sleeves in tel")llS of microstructural characteristics and presence of cold-worked metal. L J 8
- Residual Stresses
- G Determining the source of the stresses which caused sleeve cracking at CC-2 and measurements of those stresses were also objectives of. this task. Five samples were prepared and residual stresses measured by X-ray diffraction.
The first sample was a non-welded section of pressurizer heater sleeve with both reamed and non-reamed areas. Initially, the practice of reaming the sleeves prior to their installation was thought to have produced the high tensile residual stresses necessary for PWSCC. In fact, the X-ray diffraction analysis of the first sample revealed that the circumferential surface stresses were compressive in nature in both reamed and non-reamed areas. Therefore, something other than the installation reaming apparently induced the required tensile stresses.
A second sample was prepared to assess effects of the J-weld on residual stress levels. As in CC-2, the J-weld was made on the reamed section of the sleeve. The results of the analysis showed that the J-weld caused very high circumferential tensile stresses, especially at the surface of the base metal immediately below the heat affected zone. The very high stresses in this area were possible only because the reariiing produced a much higher yield strength as a result of the cold-work induced in the material.
Normally, stresses as high as those measured would cause the metal to deform to relieve the stress. The cold-work, resulting from the reaming, however, so increased the strength of the material that it did not deform. Below the surface of the base metal the residual tensile stresses were reduced, but still high enough to initiate and propagate PWSCC. A third sample was prepared with the J-weld in the non-reamed section of the sleeve. In this manner, the effe.ct of reaming could be eliminated.
The analysis results revealed circumferential tensile stresses of a lesser magnitude than stresses measured in the reamed area. And although the stresses were still sufficient to initiate and propagate PWSCC, in the absence of the cold-work induced by the reaming, the time to initiate cracking should be significantly increased.
The fourth sample was prepared to assess the effects of post-installation reaming. The sleeve was welded and then machine reamed. The surface circumferential stresses in this sample were compressive, except at the weld. Here, the surface tensile stresses were low and became compressive at a depth of 2 mils. This suggests that even if a surface crack were to form at the weld, it would not 9
- through-wall.
The last sample was prepared with a J-weld in the reamed section of the sleeve and then stress relief annealed at 1200°F for 16 hours1.851852e-4 days <br />0.00444 hours <br />2.645503e-5 weeks <br />6.088e-6 months <br />. As expected, this reduced all stresses to either insignificant tensile stresses or to compressive stresses.
Finally, examination of the longitudinal surface stresses revealed either compressive or relatively low tensile stresses, confirming that cracks, that may occur, should be axially oriented.
The residual stress measurements resulted in the following conclusions:
... The non-welded sleeve in the reamed (or non-reamed) area did not have the tensile stresses necessary to initiate PWSCC, The J-weld induced a significant ten5ile stress to the ID of the sleeve in the area adjacent to the heat affected zone. The stresses induced in this region were sufficient to initiate PWSCC given a susceptible microstructure and the presence of high temperature water, ... [ J ... Reaming after welding produces compressive residual surface stresses . which would inhibit development of PWSCC, ... Stress relief annealing reduces tensile circumferential stress to either innocuous levels or develops compressive circumferential stresses, ... Longitudinal residual surface stresses were either compressive or relatively low tensile stresses such as to preclude circumferential cracking of the . sleeves, and The relative susceptirUities defined in[ 10
- J 2.1.2 Pressurizer Instrument Nozzles Evaluation The same inspection at CC-2 in May of 1989 that identified leaking pressurizer heater sleeves, also resulted in the discovery of one leaking pressure tap nozzle. As with the heater sleeves, ABB-CE was requested to determine the failure mechanism of the instrument nozzle. BG&E removed all four upper level instrument nozzles from the CC-2 pressurizer and shipped them to ABB-CE for analysis.
The leaking nozzle and another pressure tap nozzle, near the leaking nozzle, underwent the following evaluations and tests. The leaking nozzle was examined visually to determine the nature, location and extent of the known and any unknown defects. One major crack was visually identified running perpendicular to the pressurizer wall, suggesting that the welding may have influenced the crack. To locate other defects, the nozzle was dye penetrant tested Two through-wall cracks, including the major crack visually identified, and five part through-wall cracks were found with dye penetrant Inside and outside diameters of both nozzles were measured.
In the region of the J-weld of the leaking nozzle the diameter increased by as much as 15 mils. Both ID and OD were within specifications.
The other nozzle's ID was also within specification.
Unfortunately, it was removed too far from the J-weld to definitely note any significant change in II) in the weld region. The surface of the crack was examined by scanning electron microscopy (SEM) and energy dispersive spectrometry (EDS) to determine the distnbution or concentration of selected elements and to determine if contaminants were present. The SEM showed clean separation of grains and well defined grain facets,
- indicating IGSCC. This analysis also showed how the ID region appeared more smooth and worn than the OD, indicating ID initiation.
An EDS scan revealed the presence of large amounts of silicon and some magnesium, probably introduced by the dye penetrant chemicals.
No potential crack causing contaminants were identified by EDS. 11
- More sophisticated surface analyses using a Scanning Auger Microprobe (SAM) and X-ray Photoelectron Spectroscopy (XPS}, were then performed.
- Contaminants observed on the defect surface were silicon, sulfur, boron, nitrogen, potassium, and sodium, all present in very small quantities.
Ni/Cr ratios varied somewhat depending on the location being analyzed.
All were roughly equal to the Ni/Cr ratio of the ductile region suggesting that the crack surface was exposed to a neutral pH environment.
The microstructure evaluation of the cracked nozzle near the defect revealed a microstructure consisting of medium to coarse grains in all examined regions except the grains near the ID showing some signs of cold-working.
There was good correspondence between grain boundaries and carbides, with relatively few intragranular carbides.
Intragranular carbides were numerous remote from the defect with no evidence of cold-work.
The microstructure of the other examined nozzle was similar to the cracked nozzle remote from the defect. Microhardness evaluations found that the material in the vicinity of the cracks was harder than material in the rest of the nozzle, suggesting a higher yield strength in that area. Tensile tests indicated no differences in the bulk mechanical properties of the two nozzles. Both the mechanical testing and chemical analysis demonstrated that the material was within specified limits. The failure of a pressurizer instrument nozzle is not an unprecedented event. Similar indications have been found at San Onofre Nuclear Generating Station (SONGS) Unit 3 (1986), St Lucie Unit 2 (1987), seven Electricite de France (EdF) pressurizers (1989), and most recently, the Babcock and Wilcox pressurizer at Arkansas Nuclear One Unit 1 (1990). The. SONGS nozzle failure mechanism was identified as PWSCC. The CC-2 and SONGS-3 nozzle failures were compared to determine if the CC-2 nozzle failure was related to gross metallurgical characteristics of the nozzle material or to local characteristics induced during fabrication of the pressurizer.
The records review did not identify any anomalies during the fabrication of the SONGS-3 pressurizer.
The records review for CC-2 identified that the defective nozzle had initially leaked during a shop hydro, was removed and reworked.
There were significant differences in the mechanical properties and microstructural characteristics of the nozzle material used to fabricate the 12
- **
and CC-2 nozzles. The SONGS nozzle had a much greater strength than the CC-2 nozzle. The review of EdF nozzle failures found no correlation of microstructure or mechanical properties to the cracking.
The EdF nozzles were rolled and seal . welded to the cladding, resulting in both axial and circumferential cracks in the rolled region. In contrast, CC-2 and SONGS-3 penetrations were fabricated with a J-weld only. Since there is no axial constraint in this design, circumferential cracking is not expected nor has it been observed.
The nozzle examinations resulted in the following conclusions:
.. The leaking CC-2 nozzle contained seven ID initiated intergranular stress** corrosion cracks, two of which were through-wall, -The through-wall cracks were oriented perpendicular to the pressurizer wall and weld, .. The through-wall cracks extended beyond the J-weld. None of the cracks extended to the in board end of the nozzle, .. .. Analysis of the oxides on crack surfaces indicated that the cracks were exposed to a neutral pH (pure water) environment, There is strong evidence (including crack orientation and the increase in nozzle ID in the vicinity of the weld) that welding and possibly the extensive reworking of the nozzle created a susceptible material condition and a source of stress that promoted PWSCC, and Nozzles that failed in the SONGS Unit 3 and St. Lucie Unit 2 pressurizers were similar to the CC-2 *nozzles in only two respects:
fabrication processes and operating conditions.
Microstructure and mechanical properties differed completely.
2.1.3 Heater
Sleeve Thermal Analysis Stress corrosion cracking is strongly temperature dependent.
This task was undertaken to determine the temperature profile existing in different, typical insulated beater sleeve configurations to assess the potential for crack initiation or 13 -* j . , J** .. , '* ., ,f'..
- propaga_!ion in different pressurizers.
A 3-D finite element heat transfer analysis was performed for a sleeve with calcium silicate insulation and a sleeve with insulation encased in 24 GA stainless steel. Differences between the temperature profiles for the two insulations were very small. Additionally, temperatures in the region of crack initiation were only slightly reduced from operating temperatures, while temperatures nearer the pressurizer shell dropped significantly.
These results affirm the expectation that cracks will initiate at higher temperature locations within the pressurizer shell, restricting crack expansion.
2.2 Susceptibility
Analysis 2.2.1 Evaluation of Pressurizer Heater Susceptibility to PWSCC As discussed previously, *eXJ>erience to date has taught us that PWSCC of cold drawn, annealed, and air cooled Inconel-600 requires the presence of an aggressive environment, a susceptible metallurgical condition, and high tensile stresses.
All CEOG pressurizers are subjected to an aggressive environment.
Similarly, it is believed that the J-weld between the sleeve and the cladding induced the high tensile stresses needed to initiate PWSCC, and all CEOG pressurizer heater sleeves are installed to the pressurizer with a J-weld. H these two criteria alone were used to determine the relative susceptibility of pressurizers to PWSCC then all CEOG pressurizers would be equally suscepttble because all CEOG pressurizers meet these two criteria.
But PWSCC also requires the presence of a susceptible metallurgical condition, [ . . . ] However, the cold-work induced in the material as a result of the installation reaming does produce a very hard layer at the ID surface that is capable of supporting higher residual tensile stresses from the J-weld. Therefore, 14
- J ' 2.2.2 Evaluation or Instrument Nozzle Susceptibility to PWSCC Although there is an extensive body of laboratory and field data available on which to base the suscepttbility criteria for pressurizer heater sleeves, experience*
to date indicates that these criteria do not apply to hot forged, annealed nozzle materials.
In contrast to cold annealed material, PWSCC in nozzle material has occurred in a region with medium to coarse grains and primarily intergranular carbides.
Most of the observed failures have occurred in material that received significant cold-work.
The CC-2 and EdF failed nozzles, and labqratory heats of material that cracked all received low temperature final anneals, as did various steam generator tube plugs that have* failed repeatedly.
And although the strength level at which forged Inconel-600 becomes susceptible to PWSCC is not defined, it is suspected .that, [ . . . . .. . . J .
- In summary, the evaluation of pressurizer*instrument nozzles concluded that nozzles expected to fail by PWSCC wi1( * .. . *.* :J . . A review of available information resulted in the following conclusions: [ J 15
- * . 2.2.3 *. P.!imary Pressure Boundary Penetrations Fabrication and material identification drawings for each CEOG plant were reviewed to identify other primary pressure boundary penetrations constructed of 11 or part Inconel-600 material.
In general,[
J Certified material test reports from the original material suppliers were reviewed to obtain materials data considered pertinent to the PWSCC process: Data required by ASME or CE material specifications were limited to chemistry and basic mechanical properties.
Hardness data were cataloged where available.
The annealing temperature of the as-produced
µiaterials, not required to be reported,*
was often not obtainable.
[ . * . . J Fabrication drawings and Shop Traveler documents were reviewed to establish the basic fabrication processes used in each application.
Finally, Rejection Notice documents were reviewed to identify conditions not conforming to fabrication drawing. requirements, and the disposition of the non-conformance.
This task. resulted in the following findings:
_J * ... Material for most penetrations, other than pressurizer heater sleeves, was furnished in hot forged product * * ... [ __ J 16
- i;: i ,, 2.3 * * .. _
--Current Imaging Development In the past, conventional eddy current inspection has not been reliable in distinguishing flaws from non-flaws in pressurizer heater sleeves. This is especially true in the region of the J-weld where the axial extent of the weld has a tendency to obscure the presence of flaws. To better employ the information accumulated during the susceptibility analyses, a task was undertaken to qualify eddy current imaging techniques for inspecting pressurizer heater sleeves using the IntraSpectTM/ET20 Eddy Current Imaging System developed by AMDATA, Inc. The results of this effort show that the IntraSpectTM
/ET20 with a semiautomatic motorized scanner is capable of detecting and depth sizing ID indications in pressurizer heater sleeves. Length sizing can also be performed by using a spacial sampling interval.
.. Sizing ID indications should be performed using a minimum of three frequencies.
Three frequencies provide corroboration signals for any real flaws to help e1iminate false calls. A comparison of circumferential versus axial flaws shows that there is equal sensitivity in detection of both. The developed equipment and inspection procedure*
have been successfully
.* demonstrated in the field. Safety Significance Analysis 2.3.1 Pressurizer Base Metal Corrosion/Erosion ABB-CE evaluated pressurizer shell material corrosion and erosion rates to assess the safety significance of of Inconel-600 primary pressure boundary penetrations.
Previous field experience and laboratory data all showed high corrosion/erosion rates. However, no data was available for the condition involving a crack in the thick section of a sleeve or nozzle. lnconel-600 steam generator tube material having a small axial through-wall crack was placed inside a cylinder drilled into a large block of pressurizer shell material (SA533B, 01), and welded into place. Four such samples with varying crack sizes were tested. The crevice between the tube and the cylinder wall was 2-8 mils (measured diametrically), representative of most CEOG plant heater shell clearances.
The tube was subjected to simulated primary water ( 600'F, 2250 17
- _ psi,_de.ae..rated with appropriate levels of boric acid and UOH) which passed 0 through the crack and impinged upon the shell material.
The block was initially heated to 600°F, but this could not be maintained once the leakage began to flash to steam in the crevice. Leakage from the bottom to the block was periodically measured.
Leakage rates ranged from .002 gpm to .119 gpm and the testing duration was 250-650 hours. The corrosion/
erosion tests resulted in the following conclusions:
.. Leakage rates, in almost all cases, decreased to less than 0.01 gpm over a period of about one week. The decrease was due to the build up of boric acid crystals or corrosion products in the crevice or crack. One test resulted in an increase in leakage after 300 hours0.00347 days <br />0.0833 hours <br />4.960317e-4 weeks <br />1.1415e-4 months <br /> of testing. The increase in leakage is thought to have resulted due to either boric acid breaking away or to metal loss which increasc;d the crevice clearance, .. The max:imum*depth of corrosion observed in the cylinder wall was 0.128 inches, but the corrosion was very localized (pit-like) and thus, extrapolation to estimate the maximum corrosion rate is neither appropriate nor meaningful.
A more meaningful measurement, the actual volume of material lost, was low (1 in 3 /year), and .. The boric acid deposits tended to channel the flow of the leakage. This meant that most of the cylinder walls were not attacked at all. Attack on the cylinder wall was greatest in the tests having the smallest cracks, where the leakage had primarily flashed to steam. 18 * ..
-., *
- 3.0 SUMMA.RY The failure and susceptibility analyses conducted by ABB-CE resulted in the following conclusions: . [ J Each of the CEOG plants have inspected their pressurizer heater sleeves and, with the exception of CC-2, have found no indication of leakage to date, and [ J Primary water stress corrosion cracking of Inconel-600 primary pressure boundary penetrations is not expected to be safety significant based on the following:
- All observed cracks are axially oriented,
- Fracture mechanics evaluations indicate unstable crack growth is not expected, and
- Corrosion/
erosion testing confirmed acceptable pressurizer shell material loss. 19