ML091210426
| ML091210426 | |
| Person / Time | |
|---|---|
| Site: | North Anna  |
| Issue date: | 05/12/2009 |
| From: | Bruemmer S, Crawford S, Stephen Cumblidge, Doctor S, Harris R, Carol Moyer, Schuster G, Sefferns P, Toloczko M Battelle Memorial Institute, Pacific Northwest National Laboratory, NRC/RES/DE/CMB |
| To: | NRC/RES/DE |
| Moyer Carol E. 301-251-7641, RES | |
| References | |
| Download: ML091210426 (20) | |
Text
A Destructive Validation of NDE Responses of Service-Induced PWSCC Found in North Anna 2 Control Rod Drive Nozzle 31 P
Presented t d by b CCaroll M Moyer, USNRC Stephen Cumblidge, Steven R. Doctor, George Schuster, Susan Crawford, Robert Harris, Rob Seffens, Mychailo Toloczko, Steve Bruemmer 7th International Conference on NDE May 12, 2009
Outline Motivation Control Rod Drive Nozzle Description Eddy Current Analysis of Nozzle 31 Destructive Validation of Eddy Current Results D t ti E Destructive Evaluation l ti off Ult Ultrasonic i Leak L k Path P th Measurements 2
Motivation Some control rod drive mechanism (CRDM) nozzles in PWRs have been vulnerable to pressurized water stress corrosion cracking (PWSCC).
Finding PWSCC in these nozzles using nondestructive evaluation (NDE) techniques has proven to be challenging.
Destructive validation of PWSCC and a comparison of the NDE responses p with the p physical y crack characteristics would be very y
helpful in future evaluations of CRDM nozzles.
The U.S. Nuclear Regulatory Commission and the Electric Power Research Institute formed a joint venture to procure and examine service-induced PWSCC in a CRDM nozzle.
This presentation focuses on the destructive validation/evaluation of two NDE techniques: eddy current testing and ultrasonic leak path measurements.
3
CRDM Nozzle Description The reactor pressure vessel head is the lid to the reactor Control rods are inserted through nozzles drilled and welded into the head.
Penetration Tubes Nozzle Interiors Penetration Tube Top Head (Alloy 600) (Low Alloy Steel)
((Interference fit))
Cladding (Stainless Steel)
Not Water Tight Buttering (Alloy 182)
Weld Metal Water Tight (Alloy 182)
Cut-Away View of an RPV Head CRDM J-groove Weld Diagram 4
J-Groove Weld and the Triple Point (Stainle The triple point is Triple Point where the weld Buttering metal, buttering, (Alloy 182) and penetration tube come together.
Weld Metal Any crack that (Alloy 182) penetrates past the triple point can cause leakage out of the reactor pressure vessel.
5
Eddy Current Examination of NA2 Nozzle 31 The Nozzles have been examined using many NDE techniques.
Eddy current testing (ET) has proven to be the most effective.
The J-groove weld of Nozzle 31 was examined using plus-point ET probes at 150 and 350 kHz.
The ET probes were calibrated using a series of narrow electro discharge machining (EDM) notches.
The J-groove weld was scanned with the ET probes oriented at 0 degrees and 45 degrees (relative to scan direction) to assure good sensitivity to flaws of every orientation.
Test ET Scan of CRDM Nozzle Removed from the Midland RPV head 6
Eddy Current Results on Crown of J-Groove Weld and Buttering Zero Degree Probe Rotation 45 Degree Probe Rotation 1 inch 1 inch 1
Flame Cut 1 Flame Cut Damage Damage 0 0 90 270 90 270 4
4 180 180 3 2 2 3 7
Summary of Eddy Current Indications Centered on 60° Centered on 150° Indication Length Max Voltage Indication Length Max Voltage 1 5 mm 2.1 8 3 2.3 2 3 mm 1.9 9 8 3.2 3 4 mm 3.3 10 6 3.3 4 2 mm 1.8 11 10 4.1 5 5 mm 2.2 12 4 2.6 6 3 mm 2.5 7 4 mm 2.3 0 Degree 45 Degree 0 Degree 45 Degree 1
8 2
3 60 Degrees 60 Degrees 9 9 150 Degrees 4 10 5
11 6
12 7
8
Summary of Eddy Current Indications Centered on 210° Centered on 255° Indication Length Max Voltage Indication Length Max Voltage 16 8 4.2 13 7 4.6 14 7 1.8 15 8 4.6 0 Degree 45 Degree 0 Degree 45 Degree 13 13 16 16 210 Degrees 255 Degrees 14 15 15 9
Destructive Evaluation Nozzle #31 was flame-cut from the retired vessel head for laboratory analysis.
The four regions of interest were cut from the J-groove weld using a band saw.
Nozzle 31 was cut jjust above the triple p p point to find a crack that had penetrated to the annulus.
The leaking crack was identified at 155 degrees, and a second very deep crack was found at 255 degrees.
The regions were then further sectioned to determine how far the cracks had penetrated into the weld metal.
10
Destructive Validation of Indications The four regions of interest were cut 6-8 mm above the wetted surface and at 25 mm above the wetted surface to determine whether the ET indications coincided with cracks.
Six of the sixteen indications were confirmed as cracks deeper than 6-8 mm. The six confirmed cracks all had ET indications with voltages higher than 30% of the calibration notch and a length of 7 mm or greater.
Indication Angle Length Max Voltage % EDM Notch Verified Depth 1 45 2 mm 2.1 20% Less than 6 mm 2 50 5 mm 1.9 18% Less than 6 mm 3 55 4 mm 3.3 32% Less than 6 mm 4 65° 2 mm 1.8 18% Less than 6 mm 5 70° 4 mm 2.2 21% Less than 6 mm 6 75° 3 mm 2.5 24% Less than 6 mm 7 80° 3 mm 2.3 22% Less than 6 mm 8 130° 4 mm 2.3 22% Less than 8 mm 9 145° 10 mm 3.2 31% Between 8 mm and 25 mm 10 155° 8 mm 3.3 32% Through-Weld Leaking 11 160° 14 mm 4.1 40% Between 8 mm and 25 mm 12 170° 5 mm 2.6 25% Less than 8 mm 13 200° 8 mm 4.6 45% Between 6 mm and 25 mm 14 215° 10 mm 1.8 18% Less than 6 mm 15 225° 9 mm 4.6 45% Between 6 mm and 25 mm 16 255° 7 mm 4.2 41% Through-Weld Not Leaking 11
Electron Microscopy of Crack at 155 Degrees SEM of Crack Surface The crack was located at the weld/butter interface.
The crack at the surface was tight and discontinuous discontinuous.
The main crack section was 4 mm long with short and segmented regions at each end.
12
Expanded View of Through-Weld Crack SEM Image of crack on Wetted Surface The through-weld crack had ligaments of metal crossing the crack in several places Crack The meandering nature of the crack below the surface also allowed for electrical contact between the crack faces.
Ligaments This electrical contact between the crack faces is likely responsible for the reduced ET response relative to some of the less severe cracks.
1 mm 13
Crack Near-Surface Profile The crack has a branched SEM of Crack Section and discontinuous profile Location COD (m) into the weld metal. 1 10 2 29 The crack branches largely 3 6 follow the grain boundaries 4 2 in the weld.
5 6 6 6 The crack opening 7 8
2 Closed displacements along the 9 Closed crack depth are very tight 10 4 and prevent penetrant 11 12 2
9 testing from being effective. 13 7 14 3 15 2 14
Crack Propagation into Annulus SEM of Through-Weld Crack in the Annulus The crack was tracked from the wetted surface to the annulus between the penetration tube and the vessel head.
15
Evaluation of the Ultrasonic Leak Path Measurement Technique There is interest by some in Interference Fit the inspection community to Through-Weld apply an ultrasonic Crack Region examination to the annulus It has been thought that the interference fit would transmit ultrasound, while a leak path would create J-Groove Weld d
damage tto th the carbon b steel t l (wastage) that would result in a detectable UT signal.
The ISI performed in the field and UT performed at PNNL both found what appeared to Penetration Tube be a leakage path close to the location of the through-weld leaking crack.
16
UT Leak Path Evaluation- The Effects of Boric Acid Deposits When the nozzle was cut for 2
destructive evaluation, boric acid 1 2 deposits were found in the annulus.
1 The area with boric acid deposits in the annulus was similar to the pattern found by ultrasonic examination.
i ti The possible leak path identified by UT measurements did not correspond to wastage of the carbon steel, but instead showed a region with no boric acid deposits.
17
UT Leak Path Evaluation- No Signal from Shallow Wastage/Steam Cutting 45° A region of shallow wastage on the penetration tube and carbon steel was observed, but the UT examination found no related indications.
The depth of the damage was not measured, but machining marks were still present on the penetration t ti tube t b and d in i th the annulus.
The UT leak path measurement appeared to be more sensitive to the presence and absence of boric acid deposits than to shallow wastage.
No significant wastage of the carbon steel was present in the nozzle, preventing an evaluation of the effects of severe wastage on the UT signals.
18
Conclusions PNNL detected sixteen possible cracks in the J-groove weld of Nozzle 31 using Eddy Current examination.
Eddy Current testing using Plus point probes at 350 kHz is effective at finding PWSCC in the J-groove weld in CRDM nozzles.
The six indications that were confirmed as cracks deeper than 6-8 mm all had surface lengths longer than 7 mm and ET amplitudes greater than 30% of an 8 mm EDM notch.
The amplitudes of the eddy current indications were strongly affected by the crack morphology at the surface and near the surface. The through-weld and leaking PWSCC at 155 degrees had many ligaments connecting the crack faces. Many shallow cracks provided larger ET signals than the through-weld leaking crack.
The UT leak path detection technique is more sensitive to boric acid deposits than to shallow wastage.
19
Acknowledgements Work funded by the U.S. Nuclear Regulatory Commission (NRC) Office of Nuclear Regulatory Research.
NRC Project Managers Wallace Norris for JCN Y6867 and Carol Moyer for Y6534 Cooperative p p program g between NRC and the Electric Power Research Institute (EPRI); PNNL supports the NRC research CRDMs from North Anna 2 supplied by EPRI 20