Tommy Zettervall, SQC - International Atomic Energy Agency · PDF fileTommy Zettervall, SQC...

© Swedish Qualification Centre AB, 17-07-11--14

Effectiveness Evolution of DMW Inspection Techniques and Crack Simulations

Assessed Through the International PARENT Project

Tommy Zettervall, SQC

IAEA TM-55108 in Vienna

Background and Objectives of PARENT

Results and conclusions from the blind trials and reference to NUREC reports

Crack simulation and signal response

Conclusions from cracks simulation project

© Swedish Qualification Centre AB, 17-07-11--14 IAEA TM-55108 in Vienna

Presentation outline

In the outage of year 2000 a number of surface breaking transverse defects were detected during the in-service inspection (ISI) of dissimilar connection welds (Alloy 182) in the RC-loops

The cracks were, with one exception, oriented across the weld at an angle close to 90°

IDSCC appeared in the weld metal, primarily consisting of Alloy 182

The cracks occurred as a result of internal welding repairs

Crack opening displacement COD decreases at the intersection of the crack with the component surface


ISI experience from Sweden

In the USA there have been a number of events related to PWSCC in DMWs There have been 17 events reported for butt weld PWSCC from 1993 through 2012 Nearly one-half of these events involved PWSCC that had circumferential flaw orientation Axial flaws were typically deeper in through wall extent than the circumferential flaws This was one of the driving forces for the PINC and the PARENT programs


ISI experience from USA

Program to Assess the Reliability of Emerging Non-destructive Techniques PARENT (2010-2017) was a continuation of the work begun in PINC (2006-2010) and applies the lessons learned to a series of open and blind international round-robin tests (PISC, PINC, PARENT)

The objectives were:

To pool International expertise and resources

To identify and quantitatively assess NDE methods for accurately detecting, characterizing, and sizing PWSCC/IDSCC cracks

To investigate and document field locations and crack morphologies for PWSCC/IDSCCin an electronic database “ATLAS”

To utilize lessons learned from PINC to employ techniques to manufacture representative NDE mock-ups containing realistic PWSCC/IDSCC


Background and Objectives of PARENT

The following organizations were part of PARENT: USA (NRC, PNNL) Japan (NRA, JAPEIC, MHI, NPP’s) Korea (KINS, KHNP) Europe (SSM, ENSI, VTT, SQC, SVTI)

A total number of 29 laboratories participate and contribute with NDE-data

19 open test assemblies and 13 blind test assemblies were used, representing small-bore piping, large-bore piping and BMIs

The test blocks contained laboratory-grown stress corrosion cracks, weld solidification cracks and repairs


PARENT organizations

SBDMW (BWR Safe-End nozzle)


LBDMW (PWR Safe-End nozzle)


BMI


Conclusions from PARENT

I.D. procedures show better performance over O.D. procedures

Circumferential flaws exhibiting a greater likelihood of detection than axial flaws, as a function of depth

I.D. procedures that include ECT performed better at length sizing

Flaw orientation did not show an influence on depth sizing performance based on RMSE results

PAUT procedures show better performance than conventional UT procedures on O.D



PARENT results indicate substantial improvement in PAUT performance compared with PINC

One of the conventional UT procedures exhibiting the poorest performance for examinations of SBDMWs using only one angle for inspection

One of five procedures applied for length sizing on LBDMW test blocks on O.D. did not meet the intent of ASME Code

One of nine procedures applied for length sizing on SBDMW test blocks on O.D. did not meet the intent of ASME Code

Two of three procedures applied for depth sizing on LBDMW test blocks by I.D. surface access meet the intent of the ASME Code Section XI requirement of RSME within 3.12 mm



Two of nine procedures applied for depth sizing on SBDMW test blocks by O.D. surface access meet the intent of the ASME Code, Section XI requirement of RSME within 3.12 mm

Insufficient data was collected on BMI test blocks to draw firm conclusions regarding detection and length sizing performance, due to the flaws implant technique



One of the significant differences with PARENT versus PINC was that many more procedures combined a variety of techniques In PARENT

Nevertheless, the majority of the results from PINC were confirmed by the PARENT RRT results

Multiple techniques tended to perform a higher POD and smaller sizing RMSE values in PARENT


Procedure families POD for Flaw depth of

False call rate 5 mm 10 mm 15 mm

Eddy Current 0,22 0,72 0,96 0,03

UT 0,15 0,31 0,67 0,07

PAUT 0,23 0,58 0,87 0,06

UT + ECT 0,15 0,30 0,50 0,07

UT + PAUT 0,55 0,98 1,00 0,03

UT + TOFD 0,37 0,90 0,99 0,04

UT + TOFD + ECT 0,21 0,46 0,74 0,08

Brief comparison between PINC and PARENT

The POD results for the DMW round robin show significant variability in detection performance based upon the technique, procedure and inspection team None of the NDE techniques in PISC and PINC demonstrated the capability to accurately depth size flaws to the requirements specified in ASME Section XI, and tended to slightly undersize the flaws The use of a diversity of NDE techniques tended to improve performance for detection, depth sizing and length sizing The advances in the use of PAUT is significant and procedures including this technology tended to perform better than those relying on conventional UT


Lessons learned

Having access to the surface from which IDSCC/PWSCC initiates is preferred All three studies recommend the need for procedure, equipment and personnel qualification All three RRTs support only using correct flaw simulation techniques to obtain NDE responses realistic of service degradation in test assemblies


Lessons learned, cont.


PARENT – Flaw simulation and implant of real flaws

The main objective of this project was to developed a technique to cut out real cracks and weld them into a new test block with correct geometry

Another objective was to verify that solidification cracks SC, that are used in qualification specimens, give a realistic signal response

An idea that emerged was to cut out real cracks and weld them into a new INCONEL-block with a NPP realistic geometry

Phase1 – Defect free coupon’s

Phase 2 – EDM-notches in Inconel 182

Phase 3 – Solidification cracks in Inconel 182 + one fatigue crack from

Phase 4 – Implant of real cracks into a new weld/pipe



Example of a field IDSCC with ECT


TOFD on the same field IDSCC

Part 3 – Solidification + fatigue cracks, ECT


Part 3 – Solidification + fatigue cracks, UT ID


Part 3 – Solidification crack No. 2, P-A OD


Part 4 – Coupon from P32


Part 4 – ENSI cracks, ECT


P28 P29 P31 P32

Part 4 – SC and ENSI cracks, UT ID


SC

ENSI

The height measurements

gave small different

values depending on

probe and scan direction,

and therefore we took the

average values from the

different probes and

directions.

Several tip signals in

depth of a crack depends

on the tightness and

“bridges” between crack

surfaces in depth, and it

will cause a problem to

find the deepest tip

signal.

Part 4 – SC and ENSI crack, TOFD ID


SC

ENSI

Conclusions from crack simulation project


The project developed a technique to cut out real cracks and weld them into a

new test block with correct geometry

Detection

The SC simulations were more conservative compared to the real cracks with

ECT, in the sense that the defects were more closed in the surface

Sizing

The signal pattern for UT is typical for what is considered characteristic for

IDSCC/PWSCC, i.e. multiple and more or less weak signals in depth, which

have an impact on the through wall sizing

Height measurement of IDSCC/PWSCC is a challenging process and should

be a part for future work and projects

Tommy Zettervall, SQC - International Atomic Energy Agency · PDF fileTommy Zettervall, SQC...

Documents

Transcript of Tommy Zettervall, SQC - International Atomic Energy Agency · PDF fileTommy Zettervall, SQC...