ML20206H519
| ML20206H519 | |
| Person / Time | |
|---|---|
| Issue date: | 08/19/1985 |
| From: | Cheng C, Hazelton W NRC |
| To: | |
| Shared Package | |
| ML20206H509 | List: |
| References | |
| NUDOCS 8606260163 | |
| Download: ML20206H519 (26) | |
Text
.
Conference on Structural Mechanics in Reactor Technology, 8/19-23/85, Brussels, Belgium D 1/8 THE EFFECT OF NDE UNCERTAINTIES ON BWR PIPING REPAIRS AND SAFETY EVALUATIONS C. Y. Cheng and W. S. Hazelton U.S. Nuclear Regulatory Commission Washington, D.C.
20555 Introduction Leaks and cracks in the heat-affected zones of welds due to intergranular stress corrosion cracking (IGSCC) in the BWR sustenitic stainless steel piping have been observed for almost 20 years.
Although an ever-increasing amount of research and developmental activity related to understanding the causes of the cracking and ways to prevent it has been going on during this time period, relatively little effort has been spent in improving the detection reliability of IGSCC until March 1982.
During a hydrostatic test, slight leakage was detected at two of the furnace-sensitized recirculation system pipe welds in the 28-in line at Nine Mile Point nuclear power plant. When these safe ends were examined ultrasonically using the American Society of Mechanical Engineers Boiler & Pressure Vessel (ASME B&PV) Code procedures nine months earlier, no cracking was reported.
Additional ultrasonic testing (UT) using more sensitive procedures disclosed cracks at many of the 28-in diameter recirculation pipe welds.
This finding confirmed the long term suspicion on the adequacy of the NRC endorsed Code ultrasonic procedures to detect IGSCC cracks in large BWR austenitic stainless steel pipes.
06062gfeW9 p9p h6260147
4 Following the Nine Mile Point event, IE Bulletin 82-03 was issued to specify augmented inspections of large diameter piping in the recircu-lation systems of a number of plants scheduled for refueling shortly after. More importantly, it also specified that inspection teams demonstrate that they could detect and properly identify cracks in large-bore pipe welds removed from Nine Mile Point.
was later issued to continue the inspections of all other operating BWRs, and upgraded the UT performance capability demonstrations required for the inspection teams.
Inspections performed under IE Bulletins 82-03 and 83-02 were often performed by examiners with limited knowledge and experience in detecting and sizing IGSCC.
Even with the precedent personnel training and qualifi-cation programs required by NRC, ultrasonic testing, like any other methods of interpretive nondestructive examination (NDE), will not provide "a 100%
guarantee" of detection and accurate sizing of IGSCC.
This is also true even for the most advanced state-of-the-art UT techniques. Therefore, it is important to recognize the magnitude of these NDE uncertainties and to properly factor these uncertainties into the safety evaluation of cracked piping to ensure safe operation of nuclear power plant.
This paper discusses the NRC staff approach to the problem of NDE uncer-tainties, the evaluation of NDE results, and the effects of NDE uncertainties on flaw evaluation and piping repairs.
Finally, this paper also describes the approach to minimize the NDE uncertainties.
. NDE Uncertainties Most of the NDE uncertainties encountered in IGSCC detection and sizing stem from a combination of the following problems:
An ineffective UT procedure.
For example, ASME Section XI Code uses amplitude of UT signal as a basis for reporting (50% DAC) and sizing (100% DAC) cracks.
Large, tight flaw may have signal amplitudes too low to be reported according to the Code; Many very experienced and competent UT examiners have had little or no experience with detecting and sizing IGSCC.
Still, they are
" qualified" under the governing ASME Code to perform examinations of BWR piping; Inspectability, e.g., geometric reflectors from counterbore or at weld root due to suck-up or drop-through; Accessability - weld crown or piping-to-fitting configurations often limit complete scanning of the suspected area of the welds; Highly attenuative nature of austenitic stainless steel;
9
. i IGSCC characteristics - The crack morphology with many branches and irregular surfaces produces weak, low UT signal amplitude responses; ALARA considerations - The UT examiner is generally working in a high radiation field, while trying to inspect a pipe weld that has several physical obstructions around it to prevent easy access to the weld.
In addition, the background noise is often high and the examiners have a limited time to conduct the examination.
Under such an unfavorable condition, there is no doubt that large UT uncertainties in IGSCC detection and sizing have often been reported from field inspection.
Generally, the UT uncertainties fall in one of the following categories:*
- Miscall or false call - falsely identifying a real crack as a geometric discontinuity or vice versa.
- Undercall - undersizes the crack length or depth.
Overcall - oversizes the crack length or depth.
All three categories of NDE uncertainties, as will be discussed later in detail, will have significant impact on plant safety, economics, and the ALARA goal.
For example, if a UT examiner calls a geometric reflector as
. a crack or oversizes a real crack, depending on the magnitude of overcall, the specific weld involved might be accepted for continued operation as it is, or replaced, repaired with weld overlay, or subject to augmented inservice inspections.
From the safety standpoint, excessive repairs may appear conservative, but the additional tensile stress due to overlay shrinkage exerted on the neighboring uncracked weld might make the entire piping system less safe than otherwise.
The economic impact and the unnecessary radiation exposure to plant personnel due to this miscall or overcall are obvious.
Evaluation of NDE Results When a crack is detected in the recirculation system piping, the ASME Section XI Code provides a step-by-step guidance to evaluate the flaw indications. The staff approach to the evaluation of IGSCC UT results is generally in accordance with the Code's provisions coupled wito additional guidance designed to conservatively address the IGSCC issue recognizing the UT sizing uncertainties.
An IGSCC indication that is properly characterized in accordance with Section XI Code procedure to have definitive crack depth and length can be evaluated and accepted for continued operation by one of the following criteria:
L
O
. IGSCC indications that do
- Acceptance by Ultrasonic Examination:
not exceed the acceptance standards listed in Table IWB-3514-2 and the crack length is less than about one third of the circumference maybeacceptableforcontinuedservice,butwillbesubjectto subsequent augmented inservice inspections.
- Acceptance by Analytical Evaluation:
IGSCC indications that exceed the acceptance standards listed in Table IWB-3514-2 or have lengths greater than about one third of the circumference may be acceptable for service without flaw removal, repair, or replacement if an analytical evaluation, as described in IWB-3600, meets the acceptance criteria of IWB-3600. However,theindicationswillbesubjectedtosubsequent augmented inservice' inspections. This analytical evaluation involves the crack growth calculations during the subsequent service cycle and the staff approach to this calculation is described in a companion paper D 1/9.
Acceptance by Repair:
IGSCC indications that exceed the acceptance standards listed in Tables IWB-3514-2 and IWB-3640 are unacceptable for continued service unless the flaw is either removed by mechanical methods or the component repaired to the extent necessary to meet the acceptance standards.
The present practice for BWR piping is to repair with weld overlay.
However, the weld overlay repair is considered to be a short term remedial action because of difficulties in reinspecting
)
l the overlaid welds.
1
. Acceptance by Replacement:
As an alternative to the overlay repair, the piping or portion of the piping containing the indication may be replaced.
Thereplacementpipingwillprobablybesubjectedtosomedegreeofsub-sequent augmented inservice inspections depending on whether or not IGSCC resistant materials are used and/or stress mitigation process is used.
Effects of NDE Uncertainties on IGSCC Evaluation and Piping Repairs As discussed earlier, it is obvious that UT uncertainties will have a great impact on whether an IGSCC containing pipe weld can be accepted for continued service as it is, or by repair or replacement.
In fact, the most difficult task facing NDE evaluators or regulators in dealing with the UT uncertainties is to come up with a reliable estimate of this uncertainty. Without a reliable knowledge of the magnitude of the UT uncertainties, the regulators are often forced to take a very conservative approach, which may seem to be unrealistic.
In this section, the direct impact of UT uncertainties on several issues related to pipe flaw evaluation and repairs is discussed.
Crack Growth Calculation - As described in the companion paper 01/9, the UT uncertainties will have a significant impact on the assumed residual stress distribution used for flaw evaluation.
Depending on the magnitude of overcall or undercall, it will have a strong influence on the nature of the residual stress to be assumed, i.e., tensile or compressive and the magnitude of the final residual stress.
The magnitude of the residual
. stress and crack sizes (both depth and length) will enter into the equations to derive the magnitude of total applied stress intensity factor which will in turn determine the extent of crack growth during the subsequent service cycle.
Thus the crack size after growth may exceed the acceptance standards specified in IWB-3640 and pipe repair or replacement may be needed.
Weld Overlay Repair - Three overlay designs have been used in the weld repairs of the BWR recirculation system piping.
They are:
Standard Overlay Design:
Assuming the original pipe was cracked completely through the wall for 360.
This design has the thickest overlay, thus probably has induced the largest weld shrinkage stress.
Designed Overlays:
Taking credit for part of the original pipe.
Weldments with a total length of circumferential cracking less than 10% of the circumference, with no more than four axial cracks, are con-sidered appropriate for repair by a designed overlay.
In most cases, the overlay usually acts to prevent leakage.
_g.
Limited Service Overlays: The overlay design not meeting the design j
criteria for standard or designed overlays will be considered suitable for limited service only, not to exceed one fuel cycle of operations.
Again, depending on the magnitude of overcall or undercall, the required overlay repair will fall in one of the above three designs; i.e., the UT uncertainties will have direct influence on the final overlay dimensions.
The thicker is the overlay, the larger is the induced tensile stress due to weld shrinkage which in turn makes the neighboring uncracked weld more susceptible to cracking during the following service cycle, or significantly increase the crack grdwth rate on other welds with minor cracking.
Induction Heating Stress Improvements (IHSI) - IHSI process alters the residual stress pattern, putting the inner part of the pipe wall in compression, thus inhibiting crack initiation or further growth of shallow cracks.
The process will stretch cracks open, but tests have shown that they are not extended in depth by the process. The tips of deeper cracks, part.icularly those penetrating deeper than half way through the pipe wall, may be left in a general tensile stress field.
This could cause such cracks to propagate in the depth direction faster I
n
+ -
l i
j j
- 10'-
.h than without the IHSI treatment.
Short cracks, however, will not be expected to grow longer because of the beneficial residual stress on the
)
inside portion of the pipe.
Therefore, both short cracks of medium depth, and longer shallow cracks are not expected to grow to a significant size.
l It is clear from the above discussion that the effectiveness of IHSI as i
a repair method for the cracked weldments depends strongly on the crack size of the weldments.
Any UT uncertainties in crack depth and length might give a false sense of safety regarding the piping integrity because of the IHSI treatment.
]
j Approaches to Minimize NDE Uncertainties 1
i If the technical basis for continued operation of primary system piping i
susceptible to IGSCC is going to be built on the results of UT inspection, j
further improvements in UT detection and sizing reliability are needed.
I Although it was never expected that the personnel training and qualifica-4 tion programs required under IE Bulletins 82-03 and 83-02 would provide "a 100% guarantee" of detection and accurate sizing of IGSCC, it was expected that the length of the cracks could be defired satisfactorily, and crack depth sizing determined by examiners and procedures qualified by test would be adequate; i.e., not be grossly underestimated or over-estimated.
Several instances of apparent inadequate examinations occurred i
recently strongly suggest that a requalification program must be initiated, j
i These include the following general cases:
4 L
I i i l
(1) Cracks have been found in welds that previous inspections had determined to be crack free.
It is felt unlikely that crack i
initiation and growth to reported sizes could have occurred in one fuel cycle of operation.
(2) Cracks have been found after IHSI in welds determined to be crack-free 1
i before IHSI.
Some of these were significant cracks, and could not have been caused by the IHSI process.
]
I (3) Recent multiple inspections, performed by different qualified teams, have resulted in markedly different determinations.
Some have reported significant cracks in welds that others reported to be crack free with no indications other than those attributed to root geometry.
4 l
(4) A few cases are known of welds removed from service in which cracks I
have been found that were not reported on the original examination.
i i
(5) There are also cases where welds reported to have significant cracks 4
l were found to be crack free after removal.
4 i
Although the EPRI NDE Center training program may have been basically 4
adequate, a lack of specimens with a spectrum of weld geometry and crack type severely limited the opportunity for extensive hands-on practice early in the program.
i 4
,.,.. _ _ _.. ~. _ _ _ _ _ _. _..,,. _,
.. _,,.... _., _ _ _, _ _ _ _ _. _ _. _.. _ _ _. ~. -.
._...._ _,.~.-.
=-
a l
Furthermore, the same lack of a broad range of cracked welds may have made If reliance on j
the final practical examination less than comprehensive.
weldinspectionsistobeavalidjustificationforfurtheroperationof piping systems susceptible to IGSCC, a more rigorous training and examination program must be instituted and all NDE examiners performing inspections for IGSCC must be upgraded through this more rigorous program.
l l
In summary, the NRC staff believes that all examination procedures and the I
f specific equipment used in the field inspections, and all Level 1, 2, and i
3 NDE examiners or operatcrs for flaw detection and sizing should demon-strate their field performance capability on cracked, preferably service-induced, samples in a manner acceptable to the NRC. No NDE examiner or l
operator should be per'mitted to perform inspection of BWR piping without f
demonstrating his competence even if he must take special training to gain In specific skills and knowledge required to perform these inspections.
accordance with the NDE Coordination Plan agreed upon recently by NRC, I
EPRI, and BWROG, the EPRI program being conducted at NDE Center in Charlotte j
North Carolina, is acceptable to continue the performance capability demon-stration tests.
]
1
)
Conclusions The effects of NDE uncertainties on BWR piping repairs and safety evaluations are significant.
An unreliable UT operator can often make an irrelevant or i
otherwise innocuous indication, such as geometric reflector, into a relevant i
i
. one and thus result in additional inspection and augmented surveillance.
As a result, the additional radiation exposure to inspection personnel and the economic penalty are enormous because of additional inspections and plant downtime. On the other hand, if a relevant indication is mis-called and left unrepaired, its impact on the safe and leak free operation of primary system piping can be significant.
If reliance on weld inspections is to be the technical basis for continued operation of piping systems susceptible to IGSCC, a more rigorous personnel training and performance capability demonstration test is needed.
i
]
5 Conference on Structural M:chanics in Reactor Technology, 8/19-23/85, Brussels, Belgium D 1/8 THE EFFECT OF NDE UNCERTAINTIES ON BWR PIPING REPAIRS AND SAFETY EVALUATIONS C. Y. Cheng and W. S. Hazelton U.S. Nuclear Regulatory Commission Washington, D.C.
20555 Introduction Leaks and cracks in the heat-affected zones of welds due to intergranular stress corrosion cracking (IGSCC) in the BWR austenitic stainless steel piping have been observed for almost 20 years.
Although an ever-increasing amount of research and developmental activity related to understanding the causes of the cracking and ways to prevent it has been going on during this time period, relatively little effort has been spent in improving the detection reliability of IGSCC until March 1982.
During a hydrostatic test, slight leakage was detected at two of the furnace-sensitized recirculation system pipe welds in the 28-in line at Nine Mile Point nuclear power plant.
When these safe ends were examined ultrasonically using the American Society of Mechanical Engineers Boiler & Pressure Vessel (ASME B&PV) Code procedures nine months earlier, no cracking was reported.
Additional ultrasonic testing (UT) using more sensitive procedures disclosed cracks at many of the 28-in diameter recirculation pipe welds.
This finding confirmed the long term suspicion on the adequacy of the NRC endorsed Code ultrasonic procedures to detect IGSCC cracks in large BWR austenitic stainless steel pipes.
L i
. Following the Nine Mile Point event, IE Bulletin 82-03 was issued to specify augmented inspections of large diameter piping in the recircu-lation systems of a number of plants scheduled for refueling shortly after. More importantly, it also specified that inspection teams demonstrate that they could detect and properly identify cracks in large-bore pipe welds removed from Nine Mile Point.
IE Bulletin 83-02 was later issued to continue the inspections of all other operating BWRs, and upgraded the UT performance capability demonstrations required for the inspection teams.
Inspections performed under IE Bulletins 82-03 and 83-02 were often performed by examiners with limited knowledge and experience in detecting and sizing IGSCC.
Even with the precedent personnel training and qualifi-cation programs required by NRC, ultrasonic testing, like any other methods of interpretive nondestructive examination (NDE), will not provide "a 100%
guarantee" of detection and accurate sizing of IGSCC.
This is also true even for the most advanced state-of-the-art UT techniques. Therefore, it is important to recognize the magnitude of these NDE uncertainties and to properly factor these uncertainties into the safety evaluation of cracked piping to ensure safe operation of nuclear power plant.
This paper discusses the NRC staff approach to the problem of NDE uncer-tainties, the evaluation of NDE results, and the effects of NDE uncertainties i
on flaw evaluation and piping repairs.
Finally, this paper also describes the approach to minimize the NDE uncertainties.
4 j l
NDE Uncertainties Most of the NDE uncertainties encountered in IGSCC detection and sizing stem from a combination of the following problems:
An ineffective UT procedure.
For example, ASME Section XI Code uses amplitude of UT signal as a basis for reporting (50% DAC) and sizing (100% DAC) cracks.
Large, tight flaw may have signal amplitudes too low to be reported according to the Code; Many very experienced and competent UT examiners have had little or no experience with detecting and sizing IGSCC.
Still, they are
" qualified" under the governing ASME Code to perform examinations of BWR piping; Inspectability, e.g., geometric reflectors from counterbore or at weld root due to suck-up or drop-through; Accessability - weld crown or piping-to-fitting configurations often limit complete scanning of the suspected area of the welds; Highly attenuative nature of austenitic stainless steel;
. - IGSCC characteristics - The crack morphology with many branches and irregular surfaces produces weak, low UT signal amplitude responses; ALARA considerations - The UT examiner is generally working in a high radiation field, while trying to inspect a pipe weld that has several physical obstructions around it to prevent easy access to the weld.
In addition, the background noise is often high and the examiners have a limited time to conduct the examination.
Under such an unfavorable condition, there is no doubt that large UT uncertainties in IGSCC detection and sizing have often been reported from field inspection. Generally, the UT uncertainties fall in one of the following categories:-
Miscall or false call - falsely identifying a real crack as a geometric discontinuity or vice versa.
- Undercall - undersizes the crack length or depth.
Overcall - oversizes the crack length or depth.
All three categories of NDE uncertainties, as will be discussed later in detail, will have significant impact on plant safety, economics, and the ALARA goal.
For example, if a UT examiner calls a geometric reflector as
. a crack or oversizes a real crack, depending on the magnitude of overcall, the specific weld involved might be accepted for continued operation as it is, or replaced, repaired with weld overlay, or subject to augmented inservice inspections.
From the safety standpoint, excessive repairs may appear conservative, but the additional tensile stress due to overlay shrinkage exerted on the neighboring uncracked weld might make the entire piping system less safe than otherwise.
The economic impact and the unnecessary radiation exposure to plant personnel due to this miscall or overcall are obvious.
Evaluation of NDE Results When a crack is detected in the recirculation system' piping, the ASME Section XI Code provides a step-by-step guidance to evaluate the flaw indications. The staff approach to the evaluation of IGSCC UT results is generally in accordance with the Code's provisions coupled with additional guidance designed to conservatively address the IGSCC issue recognizing the UT sizing uncertainties.
An IGSCC indication that is properly characterized in accordance with Section XI Code procedure to have definitive crack depth and length can be evaluated and accepted for continued operation by one of the following 1
criteria:
l
O
. Acceptance by Ultrasonic Examination:
IGSCC indications that do not exceed the acceptance standards listed in Table IWB-3514-2 and the crack length is less than about one third of the circumference may be acceptable for continued service, but will be subject to subsequent augmented inservice inspections.
Acceptance by Analytical Evaluation:
IGSCC indications that exceed the acceptance standards listed in Table IWB-3514-2 or have lengths greater than about one third of the circumference may be acceptable for service without flaw removal, repair, or replacement if an analytical evaluation, as described in IWB-3600, meets the acceptance criteria of However, the indications will be subjected to subsequent IWB-3600.
augmented inservice' inspections. This analytical evaluation involves the crack growth calculations during the subsequent service cycle and the staff approach to this calculation is described in a companion paper 0 1/9.
- Acceptance by Repair:
IGSCC indications that exceed the acceptance standards listed in Tables IWB-3514-2 and IWB-3640 are unacceptable for continued service unless the flaw is either removed by mechanical methods or the component repaired to the extent necessary to meet the acceptance standards.
The present practice for BWR piping is to repair However, the weld overlay repair is considered to with weld overlay.
be a short term remedial action because of difficulties in reinspecting i
the overlaid welds.
l
. Acceptance by Replacement: As an alternative to the overlay repair, the piping or portion of the piping containing the indication may be replaced.
Thereplacementpipingwillprobablybesubjectedtosomedegreeofsub-sequent augmented inservice inspections depending on whether or not IGSCC resistant materials are used and/or stress mitigation process is used.
Effects of NDE Uncertainties on IGSCC Evaluation and Piping Repairs As discussed earlier, it is obvious that UT uncertainties will have a great impact on whether an IGSCC containing pipe weld can be accepted for continued service as it is, or by repair or replacement.
In fact, the most difficult task facing NDE evaluators or regulators in dealing with the UT uncertainties is to come up with a reliable estimate of this uncertainty. Without a reliable knowledge of the magnitude of the UT uncertainties, the regulators are often forced to take a very conservative approach, which may seem to be unrealistic.
In this section, the direct impact of UT uncertainties on several issues related to pipe flaw evaluation and repairs is discussed.
- Crack Growth Calculation - As described in the companion paper 01/9, the UT uncertainties will have a significant impact on the assumed residual stress distribution used for flaw evaluation.
Depending on the magnitude of overcall or undercall, it will have a strong influence on the nature of the residual stress to be assumed, i.e., tensile or compressive and the magnitude of the final residual stress. The magnitude of the residual
. stress and crack sizes (both depth and length) will enter into the equations to derive the magnitude of total applied stress intensity factor which will in turn determine the extent of crack growth during the subsequent service cycle.
Thus the crack size after growth may exceed the acceptance standards specified in IWB-3640 and pipe repair or replacement may be needed.
Weld Overlay Repair - Three overlay designs have been used in the weld repairs of the BWR recirculation system piping.
They are:
Standard Overlay Design:
Assuming the original pipe was cracked completely through the wall for 360.
This design has the thickest overlay, thus probably has induced the largest weld shrinkage stress.
Designed Overlays:
Taking credit for part of the original pipe.
Weldments with a total length of circumferential cracking less than 10% of the circumference, with no more than four axial cracks, are con-sidered appropriate for repair by a designed overlay.
In most cases, the overlay usually acts to prevent leakage.
=
_g.
Limited Service Overlays: The overlay design not meeting the design criteria for standard or designed overlays will be considered suitable for limited service only, not to exceed one fuel cycle of operations.
Again, depending on the magnitude of overcall or undercall, the required overlay repair will fall in one of the above three designs; i.e., the UT uncertainties will have direct influence on the final overlay dimensions.
The thicker is the overlay, the larger is the induced tensile stress due to weld shrinkage which in turn makes the neighboring uncracked weld more susceptible to cracking during the following service cycle, or significantly increase the crack grdwth rate on other welds with minor cracking.
- Induction Heating Stress Improvements (IHSI) - IHSI process alters the residual stress pattern, putting the inner part of the pipe wall in compression, thus inhibiting crack initiation or further growth of shallow cracks. The process will stretch cracks-open, but tests have shown that they are not extended in depth by the process.
The tips of 4
deeper cracks, particularly those penetrating deeper than half way through the pipe wall, may be left in a general tensile stress field.
This could cause such cracks to propagate in the depth direction faster l
l
. than without the IHSI treatment.
Short cracks, however, will not be expected to grow longer because of the beneficial residual stress on the inside portion of the pipe.
Therefore, both short cracks of medium depth, and longer shallow cracks are not expected to grow to a significant size.
It is clear from the above discussion that the effectiveness of IHSI as a repair method for the cracked weldments depends strongly on the crack size of the weldments.
Any UT uncertainties in crack depth and length might give a false sense of safety regarding the piping integrity because of the IHSI treatment.
Approaches to Minimize NDE Uncertainties If the technical basis for continued operation of primary system piping susceptible to IGSCC is going to be built on the results of UT inspection, further improvements in UT detection and sizing reliability are needed.
Although it was never expected that the personnel training and qualifica-tion programs required under IE Bulletins 82-03 and 83-02 would provide "a 100% guarantee" of detection and accurate sizing of IGSCC, it was expected that the length of the cracks could be defined satisfactorily, and crack depth sizing determined by examiners and procedures qualified by test would be adequate; i.e., not be grossly underestimated or over-estimated.
Several instances of apparent inadequate examinations occurred recently strongly suggest that a requalification program must be initiated.
j These include the following general cases:
. (1) Cracks have been found in welds that previous inspections had determined to be crack free.
It is felt unlikely that crack initiation and growth to reported sizes could have occurred in one fuel cycle of operation.
(2) Cracks have been found after IHSI in welds determined to be crack-free before IHSI.
Some of these were significant cracks, and could not have been caused by the IHSI process.
(3) Recent multiple inspections, performed by different qualified teams, have resulted in markedly different determinations.
Some have reported significant cracks in welds that others reported to be crack free with no indications other than those attributed to root geometry.
(4) A few cases are known of welds removed from service in which cracks have been found that were not reported on the original examination.
(5) There are also cases where welds reported to have significant cracks were found to be crack free after removal.
Although the EPRI NDE Center training program may have been basically adequate, a lack of specimens with a spectrum of weld geometry and crack type severely limited the opportunity for extensive hands-on practice early in the program.
. Furthermore, the same lack of a broad range of cracked welds may have made If reliance on the final practical examination less than comprehensive.
weld inspections is to be a valid justification for further operation of piping systems susceptible to IGSCC, a more rigorous training and examination program must be instituted and all NDE examiners performing inspections for IGSCC must be upgraded through this more rigorous program.
In summary, the NRC staff believes that all examination procedures and the specific equipment used in the field inspections, and all Level 1, 2, and 3 NDE examiners or operators for flaw detection and sizing should demon-strate their field performance capability on cracked, preferably service-induced, samples in a manner acceptable to the NRC.
No NDE examiner or operator should be perinitted to perform inspection of BWR piping without demonstrating his competence even if he must take special training to gain In specific skills and knowledge required to perform these inspections.
accordance with the NDE Coordination Plan agreed upon recently by NRC, EPRI, and BWROG, the EPRI program being conducted at NDE Center in Charlotte, North Carolina, is acceptable to continue the performance capability demon-stration tests.
Conclusions The effects of NDE uncertainties on BWR piping repairs and safety evaluations i
are significant. An unreliable UT operator can often make an irrelevant or otherwise innocuous indication, such as geometric reflector, into a relevant
i one and thus result in additional inspection and augmented surveillance.
As a result, the additional radiation exposure to inspection personnel and the economic penalty are enormous because of additional inspections and plant downtime.
On the other hand, if a relevant indication is mis-called and left unrepaired, its impact on the safe and leak free operation of primary system piping can be significant.
If reliance on weld inspections is to be the technical basis for continued operation of piping systems susceptible to IGSCC, a more rigorous personnel training and performance capability demonstration test is needed.
l
__