dissimilar metal weld inspection reliability.2009 nondestructive evaluation: guideline for...

Dissimilar metal weld inspection reliability

Greg SelbyDirector, NDENRC / Industry Technical Meeting June 2011

2© 2011 Electric Power Research Institute, Inc. All rights reserved.

Driver

•

NRC has noted several areas of concern in the examination of dissimilar metal welds–

POD

–

Evaluation of embedded flaws


NRC recognizes industry’s efforts and the value of qualification

•

Every year, thousands of nondestructive examinations are successfully performed in nuclear power plants

•

ASME Code, Section XI, Appendix VIII qualifications as implemented through the industry’s PDI program have greatly improved the quality of the inspections performed on critical, pressure boundary welds

•

Improvements still necessary regarding how NDE is applied

NRC has concerns, but the staff always is careful to acknowledge the great majority of NDE that goes right.

Industry appreciates it!

RES presentation, RIC 2011:


The need for continuous improvement

•

Both industry and NRC recognize the need for continuous improvement in NDE as in all aspects of nuclear power generation

•

Most of NRC’s stated concerns aren’t about how we administer PDI–

PDI does what the Code says–

Concern is, is the Code right?•

There are many actions that can provide improvement without changing the rules–

Technical–

Communications and industry response

“Even if you are on the right track, you’ll get run over if you just sit there.”

Will Rogers, 1879 - 1935

http://upload.wikimedia.org/wikipedia/commons/a/a0/Rogers-Will-LOC.jpg


This presentation

•

This presentation will address:–

Each of the major areas of concern

–

The indication evaluation process, focusing on dissimilar metal welds

–

Industry actions for continuous improvement


NRC Concerns: probability of detection (POD)

–

Concern 1 The POD for qualification exams may be artificially high

relative to what it would be if real flaws in realistic mockups were used

–

Concern 2 Some of the POD curves indicate that candidates may

have missed deep flaws during their qualification examination


POD concern 1: The POD for qualification exams may be artificially high relative to what it would be if real flaws in realistic mockups were used

•

Discussion

–

The mockup configurations are realistic•

They are based on surveys of configurations in the fleet•

Diameter, thickness, tapers, materials, butter, clad, weld prep and weld processes

–

Field-removed SCC in dissimilar metal welds is not available–

It would be possible to use laboratory-grown SCC, but the cracking can’t be controlled•

We would not have independent, verifiable knowledge of the true size and configuration of the cracking

–

The qualification flaws that are in use were designed to respond

to ultrasound similarly to actual SCC

–

The responses from field SCC continually are compared to flaws in qualification samples•

To date nothing indicates that they are not representative •

(examples follow)



•

Similar Flaw in PDI Sample Set (Similar Thickness and Diameter)

•

Flaw reported in BWR Inlet Nozzle (Confirmed)

Embedded Weld Geometry Profiles



Response from actual flaw in field

Response from similar flaw in PDI sample



•

Flaw reported in BWR Inlet Nozzle

•

Similar Flaw in PDI Sample Set (Similar Thickness and Diameter)

Inside Diameter GATE

Top of Flaws


POD concern 2: Some of the POD curves indicate that candidates may have missed deep flaws during their qualification examination

•

Discussion–

Data was evaluated to determine which flaws contributed to curve not reaching 100%•

Attributed to two flaws in two samples

•

Both samples contained multiple flaws with one being much more obvious that the other

•

Combined length of both flaws spanned the majority of sample, hard to determine when one flaw ended and another started

•

In almost every case the weld was rejected with the detection of the more obvious flaw

In the field, detection of one of the flaws will lead to additional inspection of the weld


POD concern 2: Some of the POD curves indicate that candidates may have missed deep flaws during their qualification examination

•

Discussion

–

POD curves are published in:

Materials Reliability Program: Development of Probability of Detection Curves for Ultrasonic Examination of Dissimilar Metal Welds (MRP-262); product ID 1019088; public

–

Considerations when evaluating POD curves generated from qualification databases

•

Flaws are placed throughout the examination volume, but primarily in the weld and butter

•

PDI flaw depths are ≥

~10% T and ≤

~90%T (Code requirement)–

Specific proportions within stated flaw depth ranges

(10% -

30%, 31% -

60%, 61% -

90%)–

Analytically difficult to determine the correct shape of the POD

curve between 0% and 10%


NRC Concerns: evaluation of embedded flaws

•

Concern 1

Weld fabrication flaws not adequately addressed by PDI qualification program

•

Concern 2

Qualified inspectors often lack experience to interpret examination results when weld fabrication flaws are detected in a field exam, particularly when manual ultrasonic techniques are employed

•

Concern 3

Discrimination issues have resulted in misinterpretations of data

•

Concern 4

Blank and flawed grading units with little potential for false calls are selected to enhance the ability of the Performance Demonstration

Administrator (PDA) to conduct examinations. Therefore, personnel may have limited opportunity to make false calls during testing


Embedded flaw concern 1: Weld fabrication flaws not adequately addressed by PDI qualification program

•

Discussion–

The PDI program implements the Code as written (as modified by regulation)•

In compliance with approved Code Case N-695 or N-696 as applicable

•

Approved alternatives to Supplement 10•

Code Cases have been incorporated into later Code editions

–

The qualification requirements are specific to the detection and sizing of potentially safety significant inside surface connected

flaws•

Code authors felt that formal qualification requirements for benign fabrication flaws were not merited for this Supplement

•

Code required examination volume is limited to inner 1/3rd

of the thickness


Embedded flaw concern 2: Qualified inspectors often lack experience to interpret examination results when weld fabrication flaws are detected in a field exam, particularly when manual ultrasonic techniques are employed

•

Discussion–

Since 2003, thousands of qualified dissimilar metal weld examinations have been performed

–

During these examinations fabrication flaws were often detected and evaluated

–

In most cases characterization and evaluation of these flaws is straightforward•

Exceptions are when flaws are in close proximity to the inside surface or when multiple stacked flaws are present

•

When possible, alternative or enhanced techniques were often to used to aid in evaluation (ET, encoded examinations)

•

When evaluations were inconclusive conservative evaluations were performed


Embedded flaw concern 2: Qualified inspectors often lack experience to interpret examination results when weld fabrication flaws are detected in a field exam, particularly when manual ultrasonic techniques are employed

•

Discussion–

It is important to note that the initial examination is a qualified, conservative screening

–

After initial detection of a suspected relevant indication (could be fabrication or inservice), the weld receives additional attention, in accordance with the procedure•

Re-scans, often by different examiners•

Review of fabrication radiographs, welding records, prior examination results

•

If needed, supplemental NDE is performed–

e.g. encoded UT or surface examinations from the inside–

Only the initial examination –

the conservative screening –

is

addressed by the qualification; the rest of the process is a team effort


Embedded flaw concern 3: Discrimination issues have resulted in misinterpretations of data

•

Discussion –

The qualification program produces examiners who are sensitized to detect inservice defects

–

If the available data during the initial scan are consistent with cracking, the examiner must identify this in the report•

This is not the end of the interpretation process!

–

This leads to much additional scrutiny of the weld and its history –

it becomes a team effort

•

This is still within the examination procedure; the process isn’t complete until this is done



•

Discussion –

Extremely rare for the completed process to result in an incorrect interpretation

–

The instances of misinterpretation cited by the staff may actually refer to the result of the initial examination, not the final result of the examination

(discuss)


Embedded flaw concern 4: Blank and flawed grading units with little potential for false calls are selected to enhance the ability of the Performance Demonstration Administrator (PDA) to conduct examinations. Therefore, personnel may have limited opportunity to make false calls during testing

•

Discussion–

The industry’s Appendix VIII dissimilar metal weld program satisfies the requirements of CC-695•

Defines grading unit size (minimum of 3 inches)

•

Defines the minimum number of unflawed grading units that must be included in each test set (1.5 times the number of flawed grading units)

•

Number of false calls allowed–

Statistical value developed by PNNL, accepted by Code

•

Typical test: ~171”

of total weld length, ~132”

unflawed –

Qualification failures are due to false calls and to missed detections, in roughly equal measure


Embedded flaw concern 4: Blank and flawed grading units with little potential for false calls are selected to enhance the ability of the Performance Demonstration Administrator (PDA) to conduct examinations. Therefore, personnel may have limited opportunity to make false calls during testing

Performance on Unflawed Grading Units

Category

Attempts

False Calls

Average False Call Rate

A(P) 5020 259 5% A(P+F) 7167 573 8% B1(P) 1467 34 2%

B1(P+F) 1539 43 3% B2(P) 111 18 16%

B2(P+F) 166 32 19%

•

P, P+F: P refers to qualifications by candidates who passed; P+F includes failures too

•

Attempts: number of times a flawed grading unit was inspected

•

False Calls: number of times a crack was identified when none was present

•

Average False Call Rate: ratio of the other two columns



•

Completed actions

–

1) Dissimilar metal weld guideline was developed and published in June 2009

Nondestructive Evaluation: Guideline for Conducting Ultrasonic Examinations of Dissimilar Metal Welds; product ID 1018181; public•

Includes recommendations on how to prepare for examinations –

Surface conditions–

Readiness reviews–

Planning for providing the supporting information for interpretation•

Fabrication records•

Construction radiographs•

Previous inspection reports, including encoded UT data–

Post-job lessons learned



•

Completed actions

–

2) Several similar and dissimilar metal weld samples have been fabricated that contain embedded flaws•

Open samples that are available for practice and training–

Several vendors and utilities have utilized them for training of

personnel

–

Generic procedures have been revised to include additional guidance on how to evaluate embedded flaws

–

Newly qualified procedures are also required to contain additional guidance on embedded flaws



•

Completed actions–

3) The ability to detect, size and characterize embedded fabrication flaws was evaluated and the performance documented in a technical report to support changes to the acceptance criteria contained in Section III

Advanced Nuclear Technology: Reduction of American Society of Mechanical Engineers III Weld Fabrication Repairs – Fitness for Purpose; product ID

1019217; public

–

4) Capturing OE related to examination of dissimilar welds and making it available to the industry in the form of technical reports published annually Nondestructive Evaluation: Dissimilar Metal Piping Weld Examination Guidance Volume 8; product ID 1021149

–

5) Supported the xLPR

project with detailed statistical analysis of the PDI Appendix VIII, Supplement 10 qualification databases

Materials Reliability Program: Development of Probability of Detection Curves for Ultrasonic Examination of Dissimilar Metal Welds (MRP-262); product ID 1019088; public



•

Current actions–

1) Industry is preparing a Guideline for evaluating ultrasonic indications•

Areas of focus–

Embedded defects–

Service-induced cracking (primarily stress corrosion cracking)•

Will include:–

Inspection, interpretation and evaluation timeline–

Reinforcement of items from the 2009 Guideline•

Supporting information and contingencies to have ready•

Guidance for interpretation analysis–

Communication planning•

Series of implementation workshops beginning Fall 2011–

Classroom–

Hands-on, with mockups (already available)



•

Current actions

–

2) Continue to work with industry NRC on evaluating data in PDI databases•

Currently working with RPV and Overlay data•

Future collaborations will be considered if a clear need is identified–

Data is secure and great care is needed to preserve integrity–

Perform research only where a clear need to know is identified

–

3) Continue to gather OE and transmit lessons learned to the fleet•

Updates are published annually



•

Future actions

–

1) Develop an embedded flaw detection sizing and characterization computer based training (CBT) course that will be available to industry•

Will include data collected on actual samples with known embedded flaws

•

Project proposed for 2012

–

2) Develop infrastructure to support future statistical analyses

of qualification databases•

Common database platform and user interface–

Make the analyses faster and easier–

Maintain data security–

Ensure that the data transferred to the platform is verified and

validated

•

Project proposed for 2012


Summary

•

ASME Section XI, Appendix VIII qualified procedures are capable of performing high quality reliable examinations–

Numerous examinations have been performed and the large majority

of flaws detected were effectively evaluated, repaired or mitigated

•

Industry has several completed, executing, and proposed activities to support continuous improvement–

Technical capability of the workforce–

Smoothness of industry response•

In isolated instances industry (utilities, Issue Programs, NRC) have expended significant resources on ultimately inconsequential conditions

•

Industry will continue to use the qualification database to support industry initiatives


Together…Shaping the Future of Electricity

dissimilar metal weld inspection reliability.2009 nondestructive evaluation: guideline for...

Documents