Inspection of Bottom and Lid Welds for Disposal Canisters - · PDF file ·...

P O S I V A O Y

O l k i l u o t o

F I N - 2 7 1 6 0 E U R A J O K I , F I N L A N D

P h o n e ( 0 2 ) 8 3 7 2 3 1 ( n a t . ) , ( + 3 5 8 - 2 - ) 8 3 7 2 3 1 ( i n t . )

F a x ( 0 2 ) 8 3 7 2 3 7 0 9 ( n a t . ) , ( + 3 5 8 - 2 - ) 8 3 7 2 3 7 0 9 ( i n t . )

POSIVA 2010 -04

Inspection of Bottom and Lid Weldsfor Disposal Canisters

September 2010

J o r m a P i t k ä n e nJ o r m a P i t k ä n e nJ o r m a P i t k ä n e nJ o r m a P i t k ä n e nJ o r m a P i t k ä n e n

POSIVA 2010-04

September 2010

POSIVA OY

O l k i l u o t o

F I - 27160 EURAJOK I , F INLAND

Phone (02 ) 8372 31 (na t . ) , ( +358 -2 - ) 8372 31 ( i n t . )

Fax (02 ) 8372 3709 (na t . ) , ( +358 -2 - ) 8372 3709 ( i n t . )

J o r m a P i t k ä n e nJ o r m a P i t k ä n e nJ o r m a P i t k ä n e nJ o r m a P i t k ä n e nJ o r m a P i t k ä n e n

Pos iva Oy

Inspection of Bottom and Lid Weldsfor Disposal Canisters

ISBN 978-951-652-175-9ISSN 1239-3096

Tekijä(t) – Author(s)

Jorma Pitkänen, Posiva Oy

Toimeksiantaja(t) – Commissioned by

Posiva Oy

Nimeke – Title

INSPECTION OF BOTTOM AND LID WELDS FOR DISPOSAL CANISTERS Tiivistelmä – Abstract

This report presents the inspection techniques of copper electron beam and friction stir welds. Both welding methods are described briefly and a more detailed description of the defects occurring in each welding methods is given. The defect types form a basis for the design of non-destructive testing. The inspection of copper material is challenging due to the anisotropic properties of the weld and local changes in the grain size of the base material.

Four different methods are used for inspection. Ultrasonic and radiographic testing techniques are used for inspection of volume. Eddy current and visual testing techniques are used for inspection of the surface and near surface area. All these methods have some limitations which are related to the physics of the used method. All inspection methods need to be carried out remotely because of the radiation from the spent nuclear fuel.

All methods have been described in detail and the use of the chosen inspection techniques has been justified. Phased array technology has been applied in ultrasonic testing. Ultrasonic phased array technology enables the electrical modification of the sound field during inspection so that the sound field can be adjusted dynamically for different situations and detection of different defect types. The frequency of the phased array probe has been chosen to be 3.5 MHz. It is a compromise between good sizing and defect detectability. It must be taken into account that ultrasonic testing is not suitable for detection of defect types which are in the direction of the beam. Ultrasonic and radiographic testing techniques complement each other in case of planar defects.

Positioning of the indication in the radial direction is rather limited in radiographic testing. Surface inspection has been added to the inspection routine because indications from the outer surface of the canister cannot be distinguished from weld defects in the radiographic image. A 9 MeV linear accelerator has been used in the radiographic testing for this study. Other radiation sources cannot penetrate the wall thickness sufficiently.

Defects detected in visual testing will be sized using eddy current testing. Special inspection techniques have been developed for sizing using low frequency (LF) eddy current testing. High frequency (HF) coils are used to evaluate the area of the surface defects. In the case of copper inspection, 30 kHz is considered a high frequency and 200 Hz a low frequency.

The acceptance criteria define the defect detectability requirements of the inspection methods. The defect detectability of the methods is under further study to ascertain their viability. In the case of a single defect, the inaccuracy of the sizing method is less important. The significance of the inaccuracy in sizing is more pronounced when there are two or more defects in the corrosion path of the canister. The sizing capability will be studied and subsequently inspection techniques qualified on the basis of the recommendation of ENIQ (European Network for Inspection Qualification). This type of qualification is applied in Europe. In the USA, the qualification in nuclear applications is based on the performance demonstration part of ASME XI.

All required information (input information, inspection procedures, technical justifications, etc.) will be gathered in order to prepare for the qualification. At the moment in the NDT reliability study the defect detectability is in focus as well the inaccuracy of sizing of indications. The precise sizing is the clearly more safety-based and industrial acceptance of canisters can be carried out. Sizing will be also verified by a limited metallographic study of detected indications. The work of technical justifications will be started. By qualifying the inspectors and inspection methods, the defined quality requirements for manufacturing disposal canisters for spent nuclear fuel will be met.

Avainsanat - Keywords

NDT of copper weld, visual testing, eddy current testing, radiographic testing, ultrasonic testing, nuclear fuel disposal canister, closure weld, Electron beam welding, Friction stir welding, weld defects. ISBN ISBN 978-951-652-175-9

ISSN ISSN 1239-3096

Sivumäärä – Number of pages 98

Kieli – Language English

Posiva-raportti – Posiva Report Posiva Oy Olkiluoto FI-27160 EURAJOKI, FINLAND Puh. 02-8372 (31) – Int. Tel. +358 2 8372 (31)

Raportin tunnus – Report code

POSIVA 2010-04 Julkaisuaika – Date

September 2010

Tekijä(t) – Author(s)

Jorma Pitkänen, Posiva Oy

Toimeksiantaja(t) – Commissioned by

Posiva Oy Nimeke – Title

LOPPUSIJOITUSKAPSELIN POHJA- JA KANSIHITSIEN TARKASTUS Tiivistelmä – Abstract

Tämä raportti esittelee kuparin elektronisuihkuhitsauksen ja kitkatappihitsauksen tarkastustekniikat. Molemmat hitsausmenetelmät esitetään lyhyesti ja yksityiskohtaisemmin kuvataan vikoja, joita esiintyy hitseissä, jotka ovat valmistettu näillä menetelmillä. Mahdolliset näissä hitseissä esiintyvät vikatyypit, muodostavat perustan rikko-mattoman tarkastuksen suunnittelulle. Kuparin tarkastus on haastavaa johtuen hitsin anisotrooppisista ominaisuuk-sista ja paikallisista perusaineen raekoon vaihteluista. Neljää menetelmää käytetään kuparihitsin tarkastuksissa, Ultraääni- ja radiograafista tarkastusta käytetään hitsin tilavuuden tarkastukseen ja pyörrevirta- ja visuaalista tarkastusta käytetään pinnan ja pinnan lähellä olevan alueen tarkastukseen. Kaikilla näillä menetelmillä on joitakin rajoituksia, jotka liittyvät lähinnä käytetyn mittausmene-telmän fysikaaliseen perustaan. Mittaus kaikki tarkastusmenetelmillä täytyy suorittaa kauko-ohjatusti johtuen ydin-jätteen aiheuttamasta säteilystä. Kaikki menetelmät kuvataan yksityiskohtaisesti ja valittujen tarkastustekniikoiden käyttö perustellaan. Vaiheistettua anturiteknologiaa on käytetty ultraäänitarkastuksessa. Vaiheistettu ultraäänianturi teknologia mahdollistaa ääni-kentän sähköisen muokkauksen tarkastuksen aikana, niin että äänikenttää voidaan sovittaa dynaamisesti erilaisiin tilanteisiin ja erilaisten vikatyyppien havaitsemiseen. Vaiheistetun ryhmäanturin taajuudeksi on valittu 3,5 MHz. Se on kompromissi hyvän vian koon määrityksen ja vian havaittavuuden välillä. Täytyy ottaa huomioon, että ultraääni-tarkastus ei sovellu hyvin vikojen tarkastukseen jotka ovat äänikeilan suuntaisia tasomaisia vikoja. Ultraääni- ja radio-graafinen tarkastus täydentävät toisiaan tasomaisten vikojen tapauksessa. Näyttämien paikallistaminen kapselin halkaisijan suunnassa on melko rajallinen radiograafisessa tarkastuksessa. Pintatarkastus on lisätty tarkastuskäytäntöön johtuen että kapselin ulkopinnalla olevia vikoja ei voi erottaa hitsin sisällä olevista vioista radiograafisessa tarkastuksen kuvasta. Tämän raportin radiograafiseen tarkastukseen on käy-tetty 9 MeV kiihdytintä. Viat, jotka havaitaan visuaalisessa tarkastuksessa, voidaan mitoittaa pyörrevirtatarkas-tuksessa. Erityinen matala-taajuinen pyörrevirtatarkastus-tekniikka (LF) on kehitetty vian mitoitukseen. Korkeataajuuskeloja (HF) käytetään arvioimaan pintavikojen alueen kokoa. Kuparin tarkastuksessa 30 kHz on korkeataajuus ja 200 Hz matalataajuus. Hyväksymisrajat määrittelevät vian havaittavuus vaatimuksen tarkastusmenetelmille. Menetelmien vian havaittavuutta tutkitaan jotta niiden käyttökelpoisuus voidaan varmentaa. Yhden vian tapauksessa vian koon määri-tyksen epätarkkuus ei ole merkittävä. Kuitenkin kun kaksi tai useampi vika on läsnä korroosiotiellä seinä-mänpaksuuden suunnassa vian koon määrityksen epätarkkuuden merkitys kasvaa merkittävästi. vian koon määrityk-sen kykyä tutkitaan ja lopullisesti testataan pätevöinnissä, joka perustuu ENIQin (European Network for Inspection Qualification) suosituksiin. Euroopassa sovelletaan tämän tyyppisiä pätevöintejä. Yhdysvalloissa ydinvoima sovel-luksissa pätevöinnit perustuvat ASME XI:n osaan "Performance demonstration". Kaikki vaadittu tieto (Lähtöaineisto, tarkastusohjeet, tekniset perustelut jne.) kerätään pätevöintiin valmistautumista varten. Tällä hetkellä tarkastustekniikoiden luotettavuuden tutkimuksessa vian havaittavuuteen keskitytään. kuten myös vian koon määrityksen epätarkkuus. Mitä tarkempi vian koon määritys on sitä varmemmin kapseleiden hyväk-syminen voidaan suorittaa. Vian koon määritys tullaan verifioimaan rajoitetulla havaittujen näyttämien metallo-grafisella tutkimuksella. Työt teknisten perustelut tekemiseksi tarkastustekniikoille ollaan aloittamassa. Pätevöit-tämällä tarkastajat ja tarkastusmenetelmät käytetyn polttoaineen loppusijoitus-kapselin valmistuksen laatu-vaatimukset voidaan täyttää. Avainsanat - Keywords Kuparihitsin rikkomaton aineenkoetus, visuaalinen tarkastus, pyörrevirtatarkastus, radiograafinen tarkastus, ydinjätekapseli, sulkuhitsi elektronisuihkuhitsaus, kitkatappihitsaus, hitsausviat. ISBN ISBN 978-951-652-175-9

ISSN ISSN 1239-3096

Sivumäärä – Number of pages 98

Kieli – Language Englanti

Posiva-raportti – Posiva Report Posiva Oy Olkiluoto FI-27160 EURAJOKI, FINLAND Puh. 02-8372 (31) – Int. Tel. +358 2 8372 (31)

Raportin tunnus – Report code

POSIVA 2010-04 Julkaisuaika – Date

Syyskuu 2010

1

TABLE OF CONTENTS

ABSTRACT

TIIVISTELMÄ

1 INTRODUCTION ..................................................................................................... 5 2 WELDING METHODS ............................................................................................. 7

2.1 Electron beam welding (EBW) ......................................................................... 7 2.2 Defect types in electron beam welding ............................................................. 9

2.2.1 Internal root defects .................................................................................. 9 2.2.2 Gas porosity .............................................................................................. 9 2.2.3 Excess of penetration .............................................................................. 10 2.2.4 Cavities ................................................................................................... 10 2.2.5 Gun discharge defects ............................................................................ 10 2.2.6 Run out .................................................................................................... 10 2.2.7 Welding stresses ..................................................................................... 11

2.3 Friction stir welding (FSW) ............................................................................. 11 2.4 Defect types in friction stir welding ................................................................. 13

2.4.1 Pores ....................................................................................................... 14 2.4.2 Voids ....................................................................................................... 14 2.4.3 Entrapped oxide particles ........................................................................ 15 2.4.4 Tool trace material .................................................................................. 15 2.4.5 Incomplete penetration ............................................................................ 16 2.4.6 Joint line hooking .................................................................................... 16 2.4.7 Faying surface flaw ................................................................................. 16

3 REQUIREMENTS FOR WELD INSPECTION AND SET BY NDT METHODS ..... 17 4 NDT OF THE WELD .............................................................................................. 23

4.1 Visual testing .................................................................................................. 24 4.1.1 Principle .................................................................................................. 24 4.1.2 Equipment ............................................................................................... 25 4.1.3 Reference specimens ............................................................................. 26 4.1.4 Adjusting equipment for inspection and data acquisition ........................ 28 4.1.5 Data evaluation using image analysis ..................................................... 28 4.1.6 Defect detection ...................................................................................... 28 4.1.7 Defect sizing ............................................................................................ 29 4.1.8 Reporting of evaluation ........................................................................... 30

4.2 Eddy current testing ....................................................................................... 30 4.2.1 Principle .................................................................................................. 31 4.2.2 Equipment ............................................................................................... 33

2

4.2.3 Sensors ................................................................................................... 34 4.2.4 High frequency sensors .......................................................................... 34 4.2.5 Low frequency sensors ........................................................................... 35 4.2.6 Eddy current testing techniques for weld inspection ............................... 36 4.2.7 Reference specimens ............................................................................. 37 4.2.8 Adjusting equipment for inspection and data acquisition ........................ 38 4.2.9 Data analysis ........................................................................................... 39 4.2.10 Defect detection ...................................................................................... 39 4.2.11 Defect sizing and characterizing ............................................................. 40 4.2.12 Reporting of evaluation ........................................................................... 42

4.3 Radiographic testing ....................................................................................... 43 4.3.1 Principle .................................................................................................. 43 4.3.2 Equipment ............................................................................................... 45 4.3.3 Reference specimens ............................................................................. 46 4.3.4 Data acquisition ....................................................................................... 48 4.3.5 Adjusting equipment for inspection and data acquisition ........................ 49 4.3.6 Data analysis ........................................................................................... 49 4.3.7 Defect detection ...................................................................................... 49 4.3.8 Defect sizing ............................................................................................ 51 4.3.9 Reporting of evaluation ........................................................................... 52

4.4 Ultrasonic testing ............................................................................................ 54 4.4.1 Principle .................................................................................................. 54 4.4.2 Equipment ............................................................................................... 54 4.4.3 Conventional ultrasonic probes ............................................................... 57 4.4.4 Linear phased array probes .................................................................... 58 4.4.5 Matrix phased array probes .................................................................... 59 4.4.6 Ultrasonic Inspection techniques for weld inspection .............................. 59 4.4.7 Reference specimens ............................................................................. 62 4.4.8 Adjusting equipment for inspection and data acquisition ........................ 62 4.4.9 Data analysis ........................................................................................... 63 4.4.10 Raw data evaluation ................................................................................ 67 4.4.11 Specific data evaluation .......................................................................... 68 4.4.12 Reporting of evaluation ........................................................................... 75

5 COMBINING INDICATIONS - DEFECTS AND INSPECTION RESULTS ............ 77 6 DEFECT DETECTABILITY STUDY ...................................................................... 79

6.1 POD - Probability of detection ........................................................................ 79 6.2 Evaluation of inspection procedures - Human factor effect ............................ 80 6.3 Verification of results ...................................................................................... 80

3

7 REQUIREMENTS FOR QUALIFICATION ............................................................ 83 8 RESULTS AND EXPERIENCES FROM WELD INSPECTIONS .......................... 87 9 SUMMARY AND CONCLUSIONS ........................................................................ 89 REFERENCES ............................................................................................................. 91

5

1 INTRODUCTION

To check the quality of a sealing weld, the inspection of the weld will be carried out using non-destructive testing methods. In the case where the canister also contains a bottom weld it must be inspected by applying the same criteria for acceptance. The sealing weld is planned in the encapsulation plant as well as the final inspection of the canister sealing weld. For the encapsulation plant a welding and inspection room for quality checking of the canister sealing weld has been designed as shown in Figure 1 (Suikki & Wendelin 2009).

Figure 1. The figure shows a location environment for the weld inspection station. Some of the equipment in other rooms has been omitted from the figure. Only the floor outlines and penetrations are visible. The locations of a weld inspection room (1), an auxiliary system room (2), a welding room (3), an operation control room (4), as well as a fuel handling cell (5) are shown.

The visual, eddy current, ultrasonic and radiography inspections will be executed in the inspection room. All methods have been designed to be in the same room as shown in Figures 2 and 3 (Suikki & Wendelin 2009). The basis for the inspection methods will be discussed in this report. The basis for the inspection is derived from the following:

� target of inspection � requirements for inspection � inspection methods � reliability of the methods used � experiences from inspections.

6

The target is to find possible weld defects using non-destructive inspections. Also first studies how NDT-process will carried out in the encapsulation plant has been started (Jäppinen, 2009a).

Figure 2. The figure shows on the left a weld inspection station from above: A linear accelerator (1), a calibration and assistance ring manipulator (2), a calibration and assistance ring (3), a swivelling bracket (4) for UT, VT and ET testing equipment, a protective wall for the former during RT testing (5), a hatch (6) closing the floor opening. On the right of the figure cameras equipped with swivelling optics for visual inspection of a surface are shown. Visible in the centre are the tank and sensor of an ultrasonic transducer.

Figure 3. Radiographic inspection is shown in principle. An assistance ring homogenizes the material thickness in beam direction across the weld zone. The beam angle can be adjusted over the range of 0–20 degrees and the examination height can be changed by lifting or lowering the canister.

7

2 WELDING METHODS

In the following, the welding methods (EBW - Electron Beam Welding and FSW - Friction Stir Welding) and the defect types which can occur in these types of welding are shortly described. The detectable defects set requirements for inspection techniques because the specific and postulated defects are used to plan non-destructive inspections. Some defect types are mentioned in the following table.

Table 1. Defect types in copper welding methods using electron beam and friction stir welding.

2.1 Electron beam welding (EBW)

In electron beam welding gases are emitted during melting, causing more requirements to carry out the welding properly (Schultz 1993). The emitted gases cause porosity. As a precaution, deoxidizers must be used in preventing weld porosity. The welding process is controlled over a very wide range of parameters. The electron beam is tightly focused and the total heat input is much lower than that of any arc welding process. As a result, the effect of welding on the surrounding material is minimal, and the heat-affected zone is narrow. Possible distortion is slight and the work piece cools rapidly. In electron beam welding, joining of metal is produced with the heat generated by bombarding it with a high-velocity electron beam. At a typical accelerating voltage of 150 kV used for welding electrons reach a speed of 2 x108 m/s. This speed is equivalent to two-thirds the speed of light. Electrons impinge the surfaces to be joined and nearly all of their kinetic energy is transformed into heat. In Figure 1 on the left the principle of EBW equipment is shown, where the electron beam is controlled and directed with the help of a changing magnetic field produced by annular winding. Heat vaporizes the base metal and allows the electron beam to penetrate until the specified depth is achieved. The electron beam produces a keyhole in the material, which is filled as the beam passes through. In Figure 4 on right the phases of the keyhole formation according to Schultz 1993 are shown.

8

Figure 4. Principle of keyhole formation in electron beam welding.

In electron beam welding, complete joint penetration is called the keyhole welding technique (Schultz 1993, Sun & Karppi 1996, Petrov et. al. 1998), which is shown in Figure 4 on the right. In the majority of the electron beam welding applications, this technique is employed to accomplish complete joint penetration in a single weld pass, clarified in Figure 5, (Ollonqvist 2007). As a liquid-walled keyhole moves along the centre line of the butt joint, liquid metal in the weld pool flows from the front to the back of the advancing keyhole. A solidifying weld pool in electron beam welding produces a fusion weld that is narrow, parallel-sided and completely penetrated (Linnert 1994).

Figure 5. Principle of electron beam welding of thick specimen with keyhole technique.

9

2.2 Defect types in electron beam welding

In this construction the copper lid is welded to copper over pack with electron beam welding (EBW). In EBW the kinetic energy converts to heat when a focused beam of highly-accelerated electrons hits the target material copper. This high voltage (150kV) EBW equipment uses a high vacuum and power of 50 kW in order to ensure a high quality weld and sufficient penetration for this application of thick copper welding. In EBW several defect types can occur (see also Table 1):

� internal root defect � void or cavities caused by spiking � gas porosity � excess of penetration � cavities � incomplete penetration

As the base metal melts and solidifies during the welding process, several undesirable phenomena may occur which cause defects in the weld (Schultz 1993, Ollonqvist 2007, Aalto 1998, Bowyer 2000). Welding of the copper canister is performed with the partial joint penetration technique. The keyhole does not completely pass the copper shell due to the design aspects. The root of the weld is formed into the copper shell. Not only the defect but also the properties like stress state and grain size distribution in the EB-weld can affect on the detectability in NDT inspections (Pitkänen et al 2007a and Pitkänen et al 2007c). 2.2.1 Internal root defects

The weld profile in electron beam welding of thick copper is normally deep and narrow. This may cause the vapour pressure during welding to be insufficient to hold the melt in a stable uniform shape. Unstable melt slides forwards, collapses and falls near to the bottom of the keyhole. The collapsing melt can block a part of the keyhole. The partly blocked keyhole bottom solidifies rapidly and can cause a cold shut. In normal conditions, the weld root has a constantly changing profile. With high penetration values the electron beam is focused very narrow at the root and the profile can vary greatly. If the welding parameters are incorrect, the spiking effect can cause the deep and narrow needles to be filled insufficiently by the molten material, cavities, precisely cold shuts and spiking defects are formed in the root. Cavities can usually be avoided or minimized by reducing the welding speed, broadening the weld and increasing the radius of the weld root. 2.2.2 Gas porosity

Gas pores are formed in the melt either as a result of metallurgical gaseous reactions or during cooling as a result of the decreasing solubility of gases. Because of the rapid solidification of the melt, formed gases cannot escape quickly enough. Gas porosity is usually in the form of small pores, up to 0.5 mm in diameter, even though full weld width pores are possible. The welding speed has the greatest effect on the porosity in copper, whereas impurities present in Cu-OFP copper have hardly any effect on porosity.

10

2.2.3 Excess of penetration

Excess of penetration is not a real defect for acceptance of the copper canister. But after formation of excess of penetration the inspection volume changes in a way that can cause a lack of inspection and defects can be formed in the volume, which is not in the inspection area, and it generates the possible non-detection of defects for NDT. In that sense the penetration depth has to be kept to predefined limits. 2.2.4 Cavities

Cavities formed during welding are generally larger than gas pores. They are formed during solidification as a result of solidification shrinkage and in the direction of solidification towards the centre line. Cavities are typically 3–4 mm in diameter and they may extend for long distances at the root of the weld or in the run-out region. The size makes them more serious than the pores. 2.2.5 Gun discharge defects

Possible discharges in the electron gun are some of the most important problems of electron beam welding. Discharge is not directly a defect, but it can cause severe damage to the weld if it happens. Gun discharge can happen if metal vapour or impurity elements enter the electron gun during the process. As a result of the electrical discharge, the control voltage collapses and the beam current increases until a current limiting relay cuts it. This leads to defects in the weld. If the electron beam is suddenly interrupted, it generally leaves behind an end crater. Discharges that do not lead to welding being stopped can result in a crack-like defect that arises from shrinkage and can extend to the full depth of the weld. Defects resulting from severe gun discharge can usually be repaired subsequently but with difficulty. Developments have taken place to prevent gun discharge. Increasing the working distance can help and using advanced electrical control systems, which prevent an extreme increase in the current by controlling the power supply functions very rapidly. Another possibility is to deflect or bend the electron beam in the electron gun. By tilting the beam, it is possible to ensure that the molten particles ejected from the point of impingement of the beam cannot enter the electron gun. 2.2.6 Run out

In the keyhole welding technique, the residual energy of the electron beam keeps the keyhole open on the underside of the work piece and affects the shape of the root cap of the weld. If an excessively high beam current is set for a particular welding operation, the surface tension developed during melting may not be sufficient to support the too large fusion envelope and weld pool against the force of gravity. Molten metal can flow downwards, causing concavity in the crown and droplets of metal along the root or falling from the root. If the beam current is set too low, it will not completely melt the full thickness of the work pieces to be joined and will result in an excessively convex weld crown.

11

2.2.7 Welding stresses

Heating and cooling during the welding process builds up stresses in the welded structure. The residual stresses are those which remain in the welded structure after cooling to room temperature. All stresses in the welded structure can cause formation of different types of cracking. In copper welding the crack formation has not been detected yet. For nearly planar defect types, see Figure 6. NDT of thick copper weld residual stresses presents an interesting challenge in case of compression.

Figure 6. Root defect in copper EBW on the left (Pitkänen et al. 2007b). Sketches of different defect types in EBW are shown on the right. 2.3 Friction stir welding (FSW)

Friction stir welding (FSW) is a rather new welding technique and it is mainly used for welding aluminium components (Iordachescu et al. 2009a, Iordachescu et al. 2009b, Ding et al. 2009). Friction stir welding was applied successfully for thick copper components (50 mm thickness) in cooperation with TWI and SKB (Cederqvist and Andrew 2003). In these kinds of thick copper components a pilot hole must be drilled to make the plunge sequence possible without probe fracture (Pavlovic et al. 2008). Friction stir welding is a solid state joining method where the temperature is approximately 70–95 % of the melting temperature of the base material (Savolainen 2008).

FSW produces welds with fine grain equiaxed microstructure and minimizes or eliminates defects due to the liquid state. FSW is performed with either converted milling machines or specially designed and made FSW/FSP machines. The machines have to be very rigid to prevent any undesired vibrations. Good capabilities for clamping, adjustment, monitoring and measuring are also required. The process variables which can be used for control of the welding are downforce (upsetting force), rotation and welding (traverse) speeds, shoulder depth, which indicate plunging depth, as well as tool geometry, material and position: Figure 7 (Iordachescu 2009a, Ronneteg et al. 2006, Savolainen 2008).

At the beginning of the weld the tool is generally tilted with a fixed angle (usually 1–4�) and the rotating tool is plunged between the pieces to be welded. The base material is

12

plasticized by the heat generated by friction at the interface between the rotating tool and the base material and by plastic shear deformation of the base material. When the tool has reached the desired plunging depth, the traverse motion begins. During the tool movement along the joint track, the oxide layer between the surfaces breaks up, which are subsequently stirred together. The tool transports the plasticized material from the leading side of the tool to the trailing side in a manner similar to extrusion (Savolainen 2004). According to Guerra et al. 2001, material from the advancing side is rotated around the tool pin and highly deformed, whereas material from the retreating side is extruded between the rotation zone and parent material. For thick copper welding, SKB has realised FSW as shown in Figure 8.

Figure 7. Parameters to control the FSW process and the principle of FSW.

Figure 8. SKB's solution for copper welding using stir welding for component (shown on the right), and geometry of the copper canister closure weld after welding (Ronneteg et al. 2006).

13

2.4 Defect types in friction stir welding

The typical defects of FSW/FSP differ from those detected in welds resulting from fusion welding methods. Even though the friction stir welding method tries to minimise the formation of defects, possible defect types in friction stir weld are:

- pores and clustered porosities - voids - entrapped oxide particles and oxide inclusions - tool trace material - incomplete penetration - joint like hooking - faying surface flaw.

Defects described here were found in metallographic studies based on the NDT indications or random choice of metallographic areas. Some of the indications are not easy to find with NDT methods but all relevant defects can be found using different NDT methods or combinations thereof (Lévesque, 2008, Santos et al. 2008). In the FSW trials two types of discontinuities have been found using relevant process parameter settings. These two defect types are wormholes related to voids and joint line hooking (JLH). The wormholes are located in the outer part of the weld and are more or less volumetric while the JLH is located in the weld root and is non-volumetric (Pavlovic et al. 2008). There are also other types of defects, such as incomplete penetration, entrapped oxide particles, oxide inclusions, pores, faying surface flaws and tool trace material, which have a smaller relevance in the production of the friction stir weld.

Defects are usually located on the advancing side of the joint. The defects can be caused by several incorrect parameters, which are shown in Figure 9. Cederqvist and Andrews (2003) concluded that a major cause for the formation of weld defects was process instability. They also noticed that a welding temperature control system not only improved weld quality, but also simplified weld production. In the FSW defect type characterisation mainly followed the classification given by Savolainen (2004 and 2008).

14

Figure 9. Possible defect types in friction stir welding and defects found in metallographic studies.

2.4.1 Pores

Copper FS welds have been noticed to contain single pores or pore lines in all areas of the weld. Single pores are 0.1–0.5 mm in diameter, pore lines up to 9 mm in length. They are due to incorrect welding parameters, especially too low tool plunge depth. The detectability of pores is low with NDT methods, but in some cases it is possible to detect them, depending on the pore sizes and the number of pores, using ultrasonic and radiographic testing. In 50 mm thick copper weld the detectability is minimal with ultrasonic and radiographic testing, only surface pores can be detected quite easily with visual or dye penetrant testing. Surface pores are rather difficult to detect with the eddy current method.

2.4.2 Voids

Voids can be located in different positions of the weld. Voids can be surface-breaking or sub-surface defects (Savolainen 2004). Typically voids are called wormholes and they are volumetric defects containing no material and aligned with the welding direction. Voids are usually located at the advancing side of the weld. A void is the result of the insufficient material transport around the probe. According to Savolainen 2004, Iordachescu et al. 2009a, Iordachescu et al. 2009b, the reasons for a void can include:

- too low temperature - too low hydrostatic pressure - too high welding speed - too low rotation speed - too low plunge depth - too high thickness variations.

15

Voids of up to 10 mm have been found in 50 mm copper FS welds. According to Ronneteg et al. 2006, defects are caused by low shoulder depth and/or low tool temperature outside the process window. According to SKB 2005 the defect type was named wormhole according to SFS- EN ISO 6520-2:2001, which describes the defect types in fusion welding and only partly adaptable to FSW defect types. In Figure 9, voids detected in copper FSW can be seen (Savolainen 2004, Ronneteg et al. 2006).

Internal voids can be found with radiography and ultrasonic testing, and even with eddy current testing if the ligament is small enough in the limit of detectability. Visual testing as well as penetrant testing detects surface breaking voids, but when using remote visual testing, the inspection speed must be adjusted for reliable detection.

2.4.3 Entrapped oxide particles

Entrapped oxide comprises both the lazy S flaw and the kissing bond (Savolainen 2004). It consists of a semi-continuous layer of oxide particles along the joint line (Bird 2003). According to Savolainen 2004, thicker oxide scales lead to the more pronounced existence of entrapped oxide particles. It was noticed that the amount and connectivity of the micron-sized voids is highest at the root of the weld and diminishes towards the top of the weld show entrapped oxide in a Cu-OF weld (Savolainen 2004).

Entrapped oxide is due to insufficient breaking and mixing of the original oxide layers. Its formation can be prevented by decreasing the traverse speed or by increasing the rotation speed. Improvements in the tool design may disrupt the oxidized layers more efficiently. Entrapped oxide is undesirable, as it may lead to a loss of mechanical properties, although it may be tolerated in certain circumstances. Entrapped oxide is very difficult to detect using NDT methods. According to Bird 2003, it was observed that by evaluation of changes in noise level the depth of penetration can be detected using material noise ratio analysis from measured ultrasonic backscattered amplitudes. The noise level was used in a similar way to also detect entrapped oxide (Joint line remnant, JLR). According to Bird, it is likely to detect a JLR defect when the computed noise level in FSW exceeded the normal noise level in the weld, indicating the presence of entrapped oxide without LOP. In Bird's application the noise level was computed using 3D volume data out of the raw data from the evaluated area. The algorithm performs a volumetric moving average function to smooth out the individual noise amplitude values. It was not presented to which metallurgical facts this analysis is based and verified.

The presence of oxide inclusions smaller than 300 �m has been noticed in FSW of 50 mm thick copper (SKB 2005). They are generally found near the surface in the overlap zone of the weld. They are thought to be caused by oxidation due to welding in air and they can be possibly avoided by using shielding gas during welding. The oxide inclusions can only be seen with metallographic studies, not using NDT (SKB 2005).

2.4.4 Tool trace material

Tool pin material super alloy Nimonic contains nickel and traces of nickel (20 ppm) and has been found in 50 mm thick copper FS welds (Cederqvist 2006). The trace material

16

is usually located close to the surface, but may occur anywhere in the weld. The size of the inclusions is smaller than 300 �m. They originate from tool wear caused by high temperatures and process forces. In small specimens these types of traces can be found. They can be detected using high-sensitivity radiography or chemical analysis for a metallographic specimen (SKB 2005). In normal inspection this type of defect cannot be found using the normal NDT methods described in this report.

2.4.5 Incomplete penetration

Incomplete penetration, lack of penetration (LOP), leaves the plates at the root of the weld loose, though they may have some weak bonding (Savolainen 2004). LOP causes a reduction in tensile strength and a loss in fatigue strength. The severity of the defect depends on its size. The primary reason for LOP is a too short tool pin. It can be caused also by a too low plunge depth or plate thickness variation. Incomplete penetration can be detected with radiographic and ultrasonic testing. Only in the case when the incomplete penetration is surface breaking can it be detected with dye penetrant testing, and with eddy current testing. In a copper FS weld the root is not accessible and therefore defect detection in the root area is more demanding.

2.4.6 Joint line hooking

A special type of weld defect was reported by SKB 2005, Ronneteg et al. 2006 and Cederqvist 2006. Joint line hooking, detected at the root of a copper FS weld, is shown in Figure 9. It forms when the vertical joint line is pulled out in the horizontal direction by the material flow, or because of a too long tool pin or too much plunge depth. Joint line hooking was most pronounced where circular welds overlap. The size of the defect has been reduced by shortening the tool pin and/or by using a mirror-image tool pin from a maximum of 4.5 mm to a minimum of 1 mm. It can be detected normally using ultrasonic testing, but not normally in radiographic testing.

2.4.7 Faying surface flaw

According to Bird 2003, a faying surface flaw is located at the welding tool surface and it is surface breaking. A faying surface flaw contains oxides and it is metallurgically similar to a rolling lap (Savolainen 2004). The origin of the faying surface flaw can be assumed to be from the oxide layer from the joint line which is not properly dispersed near the tool shoulder. In fact, the faying surface flaw consists of oxides. The faying surface flaw may have harmful effects on the corrosion properties of the material. Faying surface flaw defects are machined away after welding so they do not cause any harm to the canister. They can be detected in visual, dye penetrant or eddy current testing.

17

3 REQUIREMENTS FOR WELD INSPECTION AND SET BY NDT METHODS

First of all, the surface properties are one of the main parameters in each NDT method. There are several items in surface properties which can affect NDT inspections:

- surface roughness - distribution of the surface roughness (geometry of the surface roughness) - different type of layers on the surface - foreign particles or small amounts of foreign materials.

Typical physical parameters are the material to be inspected and the geometry, material properties and defect types to be inspected. These items will be discussed in separate sections where each inspection method is studied more thoroughly. Different types of layers cause difficulties mainly in visual and eddy current testing but in special cases also in ultrasonic testing. Surface layers can cause harmful effects in eddy current testing for two reasons:

- The lift of signal increases with the thickness of the layer. - The conductivity and permeability properties of the layer can affect the eddy

current signal.

In addition, written standards control how to carry out NDT inspection, calibration, set up of the equipment and annual checking of the inspection instrument and sensors. This is important in order to produce inspection results which can be repeated with the expected tolerance.

The qualifications of the operators and inspection technique set requirements for the company to carry out the inspections. This will be discussed in a separate section which handles the qualification mainly according to ENIQ. Two main principles are in use in Finland:

- EN473-type qualifications of personnel for single NDT methods are required.

- Specific types of qualifications, which are directly applicable to the components to be inspected are normally used in nuclear power plant qualifications.

The basic principles of NDT methods may set requirements for the inspection, for example, in radiography inspection, the radiation of the equipment. This has to be solved by sufficient thick shielding between the linear accelerator and the personnel carrying out inspections. In ultrasonic testing, the coupling between material and probe, and the probe distance from the surface is an important parameter. The environment of the inspection area sets requirements due to radiation of the nuclear fuel. As a consequence of this radiation, all NDT inspections for the canister sealing weld must be carried out by remote control. In Table 2 there is shown the summary of principle requirements set by different criteria to carry out the inspection task of disposal canisters.

18

Table 2. Requirement set by different items to carry out inspections of disposal canisters.

The inspection procedures were drafted for all methods for mechanised inspection except for visual testing. The inspection procedures will be described in following chapters. Acceptance criteria for welds in this welding test are presented in Table 3. Any two adjacent defects separated by a distance smaller than the major dimension of the smaller of the defects shall be considered as a single defect. The criterion covering the intact wall thickness requirement of 35 mm in 100 % and 40 mm in 99 % of the canisters is the master requirement for acceptance, especially for combining defects. Penetration of the weld shall be limited to between 42–70 mm.

19

Table 3. Preliminary acceptance criteria for the canister sealing welds in the lid test series. (Finnish/English) SFS EN-ISO 6520-1:2007, SFS EN-ISO 13919-1:1996, SFS EN-ISO 13919-2:2001, Draft PrEN-ISO25239-1:2009, Draft PrEN-ISO25239-5:2009 The acceptance criteria for different defect types are modified according to remaining wall thickness criterion and completed with thick copper weld defects types (EB- and FS-weld)

Defect No. Type of defect Maximum allowable size

100 Cracks l < 10 mm, h < 3 mm*

2011, 200 Gas pore, porosity l<25 mm, h < 6 mm, w < 8 mm

2013 Clustered porosity l<25 mm, h < 6 mm, w < 8 mm

2014 Linear porosity l<25 mm, h < 6 mm, w < 8 mm

2015 Elongated cavity l<25 mm, h < 6 mm, w < 8 mm

5011, 5012

External undercut, defect on the side of the weld originating from machining and welding, possible repair by machining

l < 20 mm, h < 5 mm

402 Incomplete penetration l continuous, h < 8 mm, Intact 42 mm

Cold lap l < 50 mm, h < 10 mm

Joint like hooking l continuous, h < 8 mm, Intact penetration depth 42 mm

401 Lack of fusion l < 50 mm, h< 10 mm

2016 Wormhole / Crater l < 5 mm, w < 3 mm, h < 10 mm

300 Solid inclusions l < 10 mm, w < 3 mm, h < 10 mm

511 Incompletely filled groove l<10 mm, w < 8 mm, h < 5 mm

Scratches Permitted locally

Indentation

< 1 mm depth large diameter indentation (d > 10 mm), small and sharp indentations (scratch-like) are allowed

Symbols in Table 3 are as follows:

l length of defect, w width of defect, h height of defect

20

The indications detected in inspections are first evaluated in the screening phase according to acceptance criteria (Table 3) keeping in mind the master requirement of the intact wall thickness. This acceptance and rejection process is shown in Figure 10. The evaluation of the indications shall be carried out by qualified personnel (detection and sizing qualification). If acceptance criteria presented in Table 3 are exceeded but total defect length is less than 6 % of the total weld length then the weld is acceptable taking into account that the weld thickness requirement of 35 mm in 100 % length of the weld is met. If these first acceptance criteria are exceeded the additional evaluation of indications will be made by more advanced detection and sizing techniques by two persons qualified for detection and sizing. The main evaluation principle is the master requirement - intact material thickness requirement shall be 35 mm in 100 %. When exceeding this value a new expert assessment will be carried out. The expert panel will consist of an NDT specialist, a sizing specialist, a corrosion specialist, a welding specialist and a damage tolerance specialist. They will make an expert assessment taking into account all the detailed information of the defect and actual geometry, location and material condition. The expert decision can make an proposal to accept or reject the case. If the master requirement is not met and the assessment proposes the case be accepted, then a written deviation report shall be made and acceptance shall be applied for the case from the safety authority. In the case of a canister weld that is not accepted, the canister shall either be repaired or unloaded of the nuclear fuel and rejected.

Preliminarily, a 1 mm deep indentation is assessed to be acceptable as surface handling defect. Deeper indentations are classified unacceptable because deep extending plasticity of copper material makes the material cold deformed and thus inhomogeneous (Unosson 2009).

21

Figure 10. Acceptance and rejection process of canister weld according to evaluation of NDT indications.

23

4 NDT OF THE WELD There are four different NDT methods which are applied to the inspection of copper canister welds:

- visual testing - eddy current testing - radiographic testing - ultrasonic testing.

Visual testing and eddy current testing are mainly surface inspection methods and ultrasonic and radiographic testing are volumetric inspection methods. All of these methods and how they will be applied to the inspection of these welds will be described in this report. The NDT programme contains elements shown in Figure 11. Defects to be especially sought in those components are the basis for all NDT. Furthermore, loading concepts as well physical phenomena to decrease a component's ability to resist loading conditions should be well known by planning NDT. Acceptance criteria can be derived on the basis of these assumptions. The acceptance criteria should be adjusted to the requirements of NDT.

Figure 11. NDT programme for reliable inspections.

24

In remote mechanized inspections all items for inspection repeatability as well NDT reliability must be studied carefully. The physical basis of the NDT method is an important fact by understanding the defect detectability, which typically will be verified, evaluated and qualified. These qualifications must be retained and updated by continuous training. The inspection procedures are important in order to always perform inspections in the same manner. The environment can affect the inspection and therefore it must be considered in case of interference, inspection equipment life time or lower defect detectability in inspections. The procedure for interpretation of the inspection results is closely bound to acceptance criteria and also further actions based on the inspections results.

4.1 Visual testing

Remote visual testing is the first non-destructive testing method applied to a finished weld. The primary purpose of this method is to detect surface breaking defects in the weld area. Visual testing also provides support to other non-destructive testing methods; for example, in characterization of indications in radiographic testing and eddy current testing. Visual testing is the most widely used NDT method; it is one of the basic non-destructive testing methods which is attached to many phases in all industrial production and in-service inspections. Visual testing of a disposal canister weld has to be carried out remotely with a radiation tolerant camera system due to the radiation from spent nuclear fuel. The camera must be able to use different viewing angles (pan and tilt device) due to reflections of the copper surface and facility lighting. In addition to examining for surface breaking defects in the weld, visual testing is also used to detect possible handling defects in the whole canister after welding and before transportation

4.1.1 Principle

Remote visual testing is a visual testing technique where there is an interrupted optical path from the observer's eye to the test area. The observer in this case is the camera system (SFS-EN 1330-10, 2003). The image of the surface under test is captured with a camera, transferred to a monitor (computer) and analyzed at a remote location. Inspection data (images) are stored and can be analyzed afterwards. This type of system is suitable for mechanized and remote inspections, especially in hostile environments (for example, high radiation level). The ability of a visual inspection system (remote or human eye) to detect and recognize discontinuities is usually referred to as visual acuity. Visual acuity can be divided into following four elements (Cumblidge 2004):

- visible minimum - the smallest dot the inspection system can detect - separable minimum (resolution) - the smallest gap between two lines

that the system can resolve - visual acuity by vernier - the ability of the inspection system to

perceive spatial variation between two objects - readable minimum (recognition capability) - the ability of the

inspection system to recognize complex shapes.

25

4.1.2 Equipment

The colour camera planned for the inspection can be operated in an environment with high radiation levels (Ahlberg 2009). Extra radiation protection can prolong the life of the camera, which should also be taken into account in planning visual inspections. In the camera there are two individually adjustable lights, a remotely controlled zoom, and iris/auto-iris. The camera is also equipped with high quality close-up characteristics and high resolution colour reproduction. For remote inspection, it is important that the camera is robust and shock resistant. The camera is capable of producing colour images with a horizontal resolution of 720 TV-lines. The nuclear fuel is radioactive, which can often produce distortion to camera images during inspections. For that reason the camera used has to able to function under high radiation levels. The amount of tolerable dose is shown in the characteristic information from the camera (Figure 12 on the right). The camera system also contains lighting and a pan and tilt unit. For this camera system the maximum operation temperature is 50 ºC. The effect of temperature will be studied for encapsulation plant circumstances. The distance between the inspection object and the camera can be between 5–800 mm. The specific characteristics of the camera are shown in the table in Figure 12. Figure 13 shows the manipulator used in weld inspection studies on the left and schematic inspection arrangement on the right.

width 109,5�mm�X�145�mmLenght 337�mmWeight 5,5�kg

Total�dose 500GyDose�rate 0,5�Mrad/h

Maximum�object�distance�(tele�end) 800�mmMinimum�object�distance�(wide�end) 5�mmhorisontal�angel�of�view�(tele�end) 45�horisontal�angel�of�view�(wide�end) 5�Zoom�ratio X10�(optical),�X10(digital),�X100

Sensor Color�Mega�RadHorisontal�resolution PAL��720�Lines,�NTSC��720�linesMinimum�illumination 8

Lights 2�separately�adjustable�external�lightning�units

Maximum�operation�temperature 50�°CShort�term�temprature�up�to 70�°C

Pan�angle 359�Tilt�angle�(from�vertical�plane ±143°Pan�speed 0�10°�/sTilt�speed 0�10°�/s

Temperature�specification

Pan�and�Tilt

Camera�specifications�ColorMegaRad�S�modelCamera�dimensions

Dose�specification

Optical�data

Electrical�data

Lighting

Figure 12. Megarad camera for visual testing of EB weld area and camera related characteristics are shown on the right.

26

� Figure 13. Posiva manipulator used for visual testing experiment (left) and planned test set for the measurement (right).

In order to provide evaluation of the surface condition of the weld, the still images and video will be stored during inspection and evaluation of the stored data will be carried out later. In Figure 14 on the right the equipment used for the first preliminary measurements that were carried out with system are shown.

Figure 14. Remote visual inspection of copper EB weld using a colour camera.

4.1.3 Reference specimens

A reference specimen was designed for visual testing. The visual acuity of the inspection system can be measured with different types of resolution tests and targets (Cumblidge et al. 2004). Different types of reference defects were chosen for the reference specimen. The weld surface area is the same as in a real size component. Similarly, the lid side is similar. The tube is cut 120 mm from the end of the tube after welding. The conditions for testing are almost similar to those in a real size component. Holes of different diameters and notches with different surface openings and lengths were planned. The planned reference specimen is shown in Figure 15, and one area is enlarged (on the right) showing notches with different properties. This reference

27

specimen will be used for testing defect detectability in variable inspection conditions. such as in different lightning, testing at variable speeds and different camera distances from the inspection surface and using different zoom or lenses in the camera. The effect of speed on the image quality has been studied and reported by Cumblidge et al. 2005.

Figure 15. Reference specimen for visual inspection in defect detection experiments.

28

4.1.4 Adjusting equipment for inspection and data acquisition

The resolution specimen will be used in order to check that the visual testing equipment is able to do testing with required limits. This will be tested by checking, according to the procedure, if the required amount of line pairs in the vertical and horizontal direction will be detected. In this resolution the specimen as a line pair will be used as parallel shallow grooves as according to Figure 16. The amount of detected line pairs represents the image resolution. This limit has been tested in 2009 and 2010 and will be taken into the procedure according to tests. Similar testing charts for camera resolution determination are also available and the results of these studies will be compared to those testing charts which are generally accepted for image quality checking.

Figure 16. Resolution piece for visual testing of copper EB weld using a colour camera.

4.1.5 Data evaluation using image analysis

Remote visual testing will be carried out with a colour camera with the possibility to pan and tilt the camera. During inspection the calibration of the image quality will be tested at first with a standard set of lines (reference chart according to EIA 1956). After checking the image quality the reference defects will be measured and the data stored. This type of testing is under development. Some preliminary tests have been carried out with the Ahlberg colour camera, shown in Figure 17 (Ahlberg 2009). Detected defects on the surface of the weld are shown in Figure 17.

4.1.6 Defect detection

In defect detection of copper components with the visual testing method, there are some factors affecting detectability. Reflection conditions on the surface are affected by oxidation layers or illumination coming from the surrounding of the measurement system. Oxidation layers are difficult to observe with a black and white camera because they look just like black spots. Small reference indications on surfaces and crack-like openings on the weld are better observed with a black and white camera because of better image contrast. Surface openings are shown as a black area on the surface, which

29

stands out and causes inspectors to take notice. Crack-like openings of welds are observable but the inspector has to be very concentrated. Machining of the weld surface will close the opening of weld defects in some cases (horse shoe-like defect). Reflection areas from the surface can also improve the probability of detection while the opening of the defect on the surface becomes more visible in the reflection area of the image.

Also, the inspection speed and focus conditions of the camera play an important role as well as the tilting and pan angles of the camera. Figure 17 shows several images from preliminary remote visual testing of an EB weld. As seen in Figure 17 the light reflections from the surface affect detectability in the inspection. Small holes and scratches are seen in the images. The illumination power can be varied and this will also be studied in testing the reference specimen with different parameters.

Figure 17. Images from the visual testing experiment.

Marks, dents, scratches and other surface defects caused by manufacturing or handling are better observed with a colour camera due to 3D (colour) availability. Also, defining the observations with a pan and tilt camera is easier. These effects are not yet studied and will be studied in 2010 and 2011. A preliminary trial with a bright and oxidized copper reference piece showed no significant degree of impairment for the visual examination performance.

While performing the inspections with a black and white and colour camera, there was a problem with a defined reporting limit because everything was visible from the surface.

4.1.7 Defect sizing

The task for visual testing is to define the surface area of the surface breaking defects. In weld inspection it means defect extensions in radial and circumferential directions. In the outer surface of the tube defect size means axial and circumferential extensions of the defect. Determination of depths of defects is limited in visual inspections but in some cases it can be performed using different view angles (pan and tilt device). Comparable information can be achieved from the image (like shallow, sharp, deep imperfection). A black and white camera can also be equipped with a pan and tilt unit and laser holography device. Laser holography will show the uneven areas and imperfections of the surface. The camera should be perpendicular to the surface and with a certain distance to the surface. Using this arrangement, the depth of surface imperfection can be measured using tailored software.

30

4.1.8 Reporting of evaluation

Measurements are reported in table form as shown in Table 4. Reporting follows ENIQ recommendations to report indications in all coordinates which can be given. The coordinates are followed in the same way in all inspection methods - in polar coordinates. The polar coordinates express exactly the physical position of detected detect compared to 0-positions for each coordinate. This is crucially important when indications will be combined after analysis of different NDT methods.

Table 4. Indication record table format for visual testing.

Ind.�no: C1�[�] C2�[�] A1�[mm] A2�[mm] R1�[mm] R2�[mm] l�[mm] w�[mm] h�[mm] Defect�type� Remarks1 68.0 68.7 0 2 475 480 6 5 2 Wormhole Surface�breaking

4.2 Eddy current testing

Eddy current testing is a widely applied non-destructive inspection method for the tube and pipe industry (Http://www.everyspec.com, 2005). Equipment is simple to install and operate, and provides a much needed monitoring facility at moderate cost. Eddy current testing can be applied to all metals, both ferrous and non-ferrous, at several stages of the manufacturing process and also a coating of similar thickness measurement on the electrical conductive material.

The purpose of eddy current inspection is to find defects in different materials and also possibly to provide information from changes in material properties. The ability to quantitatively determine the location and shape of any defect or internal structure within materials is important for both the evaluation of the copper component surface and the near-surface defect state and characteristics of material.

The surface inspection of copper lid weld plays an important role in the acceptance of nuclear fuel disposal. There are two main reasons to inspect these components: manufacturing and handling defects of components. For the EB weld inspection the data acquisition software, visualizing tools for eddy current measurements and eddy current sensors for detection of unwanted defects were developed. The eddy current equipment was manufactured by Fraunhofer-IZFP and the visualizing software in active co-operation with Posiva and Fraunhofer-IZFP during the inspections. The inspection procedure was produced during the development of the inspection techniques. The aim of the inspection method development is to qualify the method for surface and near-surface defect detection and sizing according to ENIQ. The study of technical justification will be carried out in 2010 and 2011. At the same time, the defect catalogue will also be gathered and gaining experience from measurements of Posiva's research manufacturing will be continued.

31

4.2.1 Principle

Eddy current inspection uses a constant alternative current to excite the current flow in the sensor coil. The alternative current in the coil produces an alternating magnetic field around the coil. When the coil is placed in close proximity to metals, it induces within the test material a flow of electrical currents in the reversed direction, known as eddy currents, compared to current flow direction in the coil. These currents produce a magnetic field which in turn affects the original field. If in eddy current testing the induced magnetic field hits a possible defect in the material, it affects the flow of the eddy current. This change of eddy current flow in turn affects the magnetic field and the current in the coil. When the effect is detectable, defects can be detected. The principle of eddy current testing is shown in Figure 18.

In non-conducting and non-magnetic material no eddy currents are induced so eddy current inspection cannot be applied to such materials. If the material is conducting and nonmagnetic, e.g. copper, zinc, aluminium, titanium or stainless steel, the induced eddy currents provide a magnetic field which opposes a change in the net magnetic flux density and these materials can be tested with eddy current testing.

Figure 18. Principle of eddy current testing.

Eddy current testing is affected by the electrical conductivity and permeability of the material. These strongly affect the penetration depth of the eddy current density. This can be estimated with the help of a simple equation (shown in Figure 19, Http://www.everyspec.com, 2005). The frequency of the used coil has a similar effect like conductivity and permeability. The frequency of the coil is restricted by the resonance frequency of the coil. The frequency used should be clearly less than the resonance frequency of the coil. The standard depth of penetration is often given by the curves shown for different metals in Figure 19. The usable range for inspection with the help of those curves can be quickly estimated.

32

Figure 19. The depth of penetration in eddy current testing of copper.

Figure 20. Impedance plane presentation of different metals in eddy current testing.

Eddy current testing is an inspection method which is affected by different effects and parameters. Such parameters are, for instance:

- surface roughness - foreign material on the surface - edges and geometrical changes of component.

The effect of these parameters should minimize and maximize the effect of the defect signal. The eddy current signals in the impedance plane for different conductive materials are characteristic for that material. The impedance curve shown in Figure 20 present behaviour of coil impedance and inductance by inspection of different metals (Buckley 1994). A typical method in eddy current testing is to produce a lift off signal in order to see at which angle the lift off signal goes and the inspection frequency is

33

chosen in the manner that the indication goes in a different direction to the lift off signal. But the sensitivity and depth of penetration should always be sufficiently high for the inspection. The selection of the frequency is dependent on the expected defect types which must be detected.

There are several types of coil systems available. The two main types of coils are the absolute and differential types of coils. In the absolute type the signal is measured as an absolute impedance change. In differential coils the impedance balance between two coils is measured and changes in this balance are detected. In the coils used for copper, measurements are applied to both coil types. The coil type affects the eddy current signal formation, which can be seen in the polar coordinate signals of the coils.

4.2.2 Equipment

The eddy current equipment Posiva used for inspection is in the preliminary stage. This system will be used for development of the inspection techniques and design for the final inspection system in the encapsulation plant.

The eddy current test instrument performs three basic functions: generating, receiving, and displaying eddy current signals. The generating portion of the unit provides an alternating current to the test coil. The receiving section processes the signal from the test coil to the required form and amplitude for display. The instrument outputs or displays consist of a variety of visual, audible, storage or transfer techniques utilizing meters, video displays, chart recorders, alarms, magnetic tape, computers and electrical or electronic relays.

Figure 21. Principle of the eddy current system used for copper inspection.

Posiva is using a 32-channel eddy current device based on two InspECT® eddy current boards. Each board can multiplex up to 16 frequency channels. With the external multiplexer box one board can drive 16 probes. Frequency and probe multiplexing can be mixed to get multi-frequency data of a sensor array with just one scan. The device

34

has a two axis coordinate interface for scanning with manipulators. Data connection to the PC is over Ethernet. There is no need for a real time operating system because the data is buffered in the device and synchronized with the coordinates of the manipulation system. A sketch of the equipment is shown in Figure 21 (Pitkänen et al. 2009a).

4.2.3 Sensors

Coil selection is the most important part of solving an eddy current application; no instrument can achieve much if it does not get the right signals from the sensors. Coil designs can be divided into three groups:

1 Surface sensors used mostly with the probe axis normal to the surface, in addition to the basic ‘pancake’ coil, this includes pencil coil and special- purpose surface coils such as those used inside a fastener hole.

2 Encircling coils are normally used for in-line inspection of round products. The product to be tested is inserted though a circular coil.

3 ID probes are normally used for in-service inspection of heat exchangers. The sensor is inserted into the tube. Normally, ID coils are wound with the coil axis along the centre of the tube.

These categories are not exhaustive and there are obviously overlaps; for example, between non-circumferential wound ID coils and internal surface sensors. The used sensors for copper inspection are surface coils.

So far, we have only discussed eddy current sensors consisting of a single coil. These are commonly used in many applications and are commonly known as absolute coils because they give an ‘absolute’ value of the condition at the test point. Absolute coils are very good for metal sorting and detection of cracks in many situations; however, they are also sensitive to material variations, temperature changes, etc.

Another commonly used probe type is the ‘differential’ coil. This has two sensing elements looking at different areas of the material being tested. The instrument responds to the difference between the eddy current conditions at the two points. Differential probes should be particularly good for detection of small defects, and they are relatively unaffected by lift-off; although the sensitivity is reduced in just the same way as in absolute coils, they are not so sensitive to temperature changes.

4.2.4 High frequency sensors

A high frequency probe is shown in Figure 22. The frequency used is typically 30 kHz in copper measurement; for high frequency the depth of penetration is about 0.5 mm. A high frequency coil is a bobbing coil. The applied coils are identical "absolute" pan-cake coils. The coils are not shielded. Each coil has an internal reference coil. The coils are sensitive to crack-like defects and to volumetric defects. The sensitivity is independent of the orientation of the defects. The distance between the neighbouring coils is 2.5 mm. The maximum sensitivity drop for circumferential crack-like defects between the coils and scans is 4 dB. To avoid wearing and respective changes in

35

sensitivity, the coil heads must not be in contact with the canister surface during examination. The distance applied between the probe head and the surface is 0.25 mm. The coils must be locked onto the probe head to prevent the changes of the distance between the coils and the surface being examined. The coil is manufactured at Fraunhofer-IZFP in Germany. The eddy current surface inspection is carried out using the absolute method. A four-coil-array probe has been applied to weld inspections. The width of the active area of the array is 10 mm. Each coil is connected to the respective channel of the eddy current unit. The depth of penetration of current density generated by the coils was computed and is shown to be about 1 mm as a maximum.

Figure 22. High frequency coil for copper weld measurement.

4.2.5 Low frequency sensors

During inspection of the EB weld deep needle type volumetric defects were found and for this type of defect a low frequency probe was developed. For copper inspection, a low frequency probe was suggested by Uchanin et al. 2003. This eddy current coil was using four receivers, the transmitter was in the middle of that coil system and the frequency used in the measurements was 2 kHz. In copper EB weld inspections the frequency used was 200 Hz. The low frequency differential coil consisting of one transmitter coil and two receiver coils can be seen on the right in Figure 23. According to the modelling of the current density the depth of penetration was computed to be about a little less than 15 mm by the transmitter and in the receiver coils a little over 10 mm. The modelling software Civa was used to optimize current density distribution in the copper shown in Figure 24. The area for detection is about 40 mm. The distribution could be lowered to 30 mm with help of extra ferrite around the transmitter coil, and the sensitivity was increased about 5 dB.

36

Figure 23. Low frequency probe on the left and modelled eddy current density low eddy current probe on the right.

Figure 24. Modelling of low frequency probe on the left and inspection directions of the low eddy current probe on the right.

4.2.6 Eddy current testing techniques for weld inspection

The eddy current testing procedure for copper weld is a combination of high frequency and low frequency probe measurements. In the preliminary weld inspection low frequency measurements were applied in two directions and they were called 0� and 90� inspection directions. The inspection directions can be seen in Figure 24. In the weld inspection the surface and near-surface area until about a depth of 10 mm will be inspected. The surface breaking defect until 10 mm can be detected and sized and a defect with a ligament of about 5 mm can also be detected. The inspection area is shown in Figure 25. This detectability of defects will be studied in various separate projects.

37

Figure 25. Inspection area of eddy current inspection is shown in the coloured area.

�� Figure 26. C-scans and selected indication in polar coordinate presentation of EB weld using low (on the left) and high (on the right) frequency coils.

The lid specimen is used as a reference specimen for eddy current testing of the weld. Figure 26 shows the eddy current testing results of one real size EB weld. So far, about 20 real size EB welds and two friction stir welds have been inspected with eddy current testing. The welds may contain subsurface defects. Subsurface defects can be distinguished from other defects as the range of low frequency coils is up to a depth of 5 mm, while the ability of high frequency coils to detect defects is limited to 0.5 mm. By combining the above mentioned information, the characterization and sizing of the defects can be carried out simply. The indications of the four screw holes are clearly seen in the C-scans of Figure 26. The surface areas of the indications obtained using high frequency coils are rather small compared to those by low frequency coils. The results of one indication obtained using both techniques are shown in a polar coordinate presentation in Figure 26. The final evaluation of these indications will be conducted after the evaluation software has been completed.


A reference specimen for lid inspection was applied for weld inspection evaluation. The lid specimen contains notches from 5 mm in length and 1 mm in depth to 20 mm long and 20 mm deep notches. All reference defects can be detected with both HF and LF probes. Also, holes are available on the upper surface of the reference lid. The result of a HF frequency study is shown in the C-scan view in Figure 27 and LF measurements in the C-scan view in Figure 28.

38

Figure 27. C-scans from lid inspection with high frequency coils - upper area (left) and lower area (right).

Figure 28. C-scans from lid inspection from low frequency measurements form lower area (left) and lid reference specimen.


For adjusting the eddy current equipment for inspection, 1 mm and 2 mm deep notches are applied (see Figure 29). The signal for each single channel of high frequency four-coil probe will be adjusted separately. All channels are adjusted so that signals are similar in sensitivity as well as in form. The same procedure will also be carried out for the low frequency probe.

39

Figure 29. Reference specimen with 1 and 2 mm notches and corresponding indications in the polar coordinate.

4.2.9 Data analysis

Data analysis can be divided into defect detection and defect sizing. In eddy current inspection it must be assured that the quality of inspection data meets the requirements of the inspection procedure. In data sizing it must be checked that the signals are not saturated and that the right sensitivity level has been used in the inspection.


In defect detection the registration limit is important. The registration limit is fixed to the level of reference defects. The amplitude of a 1 mm deep notch is applied. A registration level of -12 dB from the reference defect has been used. Equipment noise level is about 180 digits and the indication of a 1 mm deep notch corresponds to 20557 digits, and the indication of a 2 mm deep notch corresponds to 27348 digits (see Figure 30). The reference defect is a 1 mm deep notch and the signal noise ratio is about 41 dB. In the plate weld measurement the signal to noise ratio was about 35–38 dB (see Figure 30). So detectability is very good. This is not the same as the real noise, which is affected by the surface roughness and probe holder lifting movement. The registration level is near the real signal to noise ratio. This can be improved by making surface roughness better and probe holder movement smaller. In every measurement the indications are compared to the reference defect. The verification of defect indications is under study.

40

Figure 30. Eddy current noise level and indication from reference defects in copper weld inspection.

4.2.11 Defect sizing and characterizing

In defect sizing the measures of detected indications will be given. This information is one part of comparing inspection results against acceptance criteria. The eddy current indications of known reference defects with different depths shows in the polar coordinate presentation how the angle of indications reveals the depth of the defects. The inaccuracy of depth sizing will be studied. As seen in Figures 31 and 32, the indications will saturate at a depth of approximately 10 mm for surface breaking defects. This means that it is possible to estimate the depths of the indications until 10 mm, after that the reliability clearly decreases. Also it is possible to distinguish volumetric and planar types of defects. These two types of defects have their own defect depth sizing curves. So it is important to first evaluate the type of indication and surface area to improve accuracy in sizing.

5�x�3�mm�Max�amplitude�1390,�

Max�angle�48.5�°

11�x�8�mm�Max�amplitude�1523,Max�angle�61.5�°

Defect depth

3 mm

Figure 31. Reference defect depth variations from a U-type of defect show variation in amplitude and angle in polar coordinate presentation for low frequency coils.

41

Figure 32. Defect sizing principles: change of indication angle depends on the depth of the defect (left) and evaluation bases also on the amplitude of detected indications (in the middle) and U-type of indication (right).

The evaluation can be based either on the measured amplitude or on the angle of indication as shown in Figure 32. The preliminary measurements show that in U-type defects, as shown in Figure 32, saturation is reached at a depth of 10 mm. These types of reference defects were chosen because these types of real defects were detected during inspection of welding. The welding process and machining after welding can also produce defects which do not open to the surface, which means that there is a so-called ligament between the defect and the surface. If this ligament is not too thick, the defects can be detected in eddy current testing. As shown in Figure 33, the defect with a ligament up to 5 mm was still detectable. The ligament thickness can be evaluated from the angle of the indications. The evaluation scheme of eddy current signals is shown in Figure 34.

1�mm2�mm3�mm4�mm5�mm7�mm

Figure 33. Effect of ligament to angle of indications is shown. C-scan and amplitude scans are shown on the left and the dependency of the angle of indications to the ligament is shown on the right.

42

Figure 34. The evaluation scheme for eddy current signals.


Reporting the evaluation is under development but in preliminary inspections the reported indications have been filled in an Excel table similar to that shown in the following table. In this Excel (Table 5), evaluated data of indications connected to position data is reported and this is crucially important when indications will be combined after analysis of different NDT methods.

43

Table 5. Eddy current inspection reporting.

Defect Ampl itude�and�S/N

IndicationAmpl i tude�l inear

Ampl itude�dB

Noise�dB

S/N��[dB�]

R�s tart�[mm]

R�middle�[mm]

R�max�

(mm)R�end�[mm]

Length(R)�[mm]

R�corr�(mm) C�start[°] C�max�(º) C�end[°]

Length�(C)�(mm)

C�corr�(mm)

Defect�type

XK030�ET1 238 �27,9 �39 11,1 479 481,5 483,6 4,6 4,1 8,9 8,8 9,4 6,4 4,1 Volum.XK030�ET2 491 �21,6 �39 17,4 473,8 477,1 479,3 5,5 4,6 86,7 86,8 87,3 4,4 4,6 planarXK030�ET3 1534 �11,7 �39 27,3 476,1 477,3 482,5 6,4 5,4 104,8 105,5 105,7 6,8 5,4 planarXK030�ET4 184 �30,1 �39 8,9 475 477,05 477,1 479,1 4,1 4 126,6 127,1 127,2 4,6 4 vol .XK030�ET5 377 �23,9 �39 15,1 472,2 474,7 474,9 477,2 5 4,5 221,5 222,2 222,3 5,5 4,5 volum.XK030�ET6 332 �25 �39 14 477 479,3 479,4 481,6 4,6 4,4 256,6 256,8 257,2 3,9 4,4 planarXK030�ET7 769 �17,7 �39 21,3 472,3 475,25 478,3 478,2 5,9 4,9 270,8 271,3 271,7 6,4 4,9 planarXK030�ET8 990 �15,5 �39 23,5 472,2 475,2 473,1 478,2 6 5,1 275,1 275,6 275,8 5,8 5,1 planarXK030�ET9 437 �22,6 �39 16,4 473 475,55 475 478,1 5,1 4,6 303,4 303,7 304 4,5 4,6 planar

XK030�ET10 7689 2,3 �39 41,3 473,3 477,7 476,9 482,1 8,8 7,3 316,5 316,9 317,7 10 7,3 VolumXK030�ET11 590 �20 �39 19 473,9 477,15 479,2 480,4 6,5 4,8 342,6 342,8 343,2 5,1 4,8 planar

The�s i ze�of�the�indications �on�the�Cscan�of�ECU�Viewer

R�s tart�[mm]

R�middle�[mm]

R�end�[mm]

Correcte

d�R�maxC�s tart�[mm]

C�middle�[mm]

C�middle�

[°]C�end�[mm]

Corrected�C�max�(mm)

Lenght�C�proj.�[mm]

Length�R�rea l �[mm]

Lenght�C�real �[mm]

A�s tart�[mm]

A�middle�[mm]

A�end�[mm]

42,02 42,27 42,51 42,07 86,3 86,3 9,2 86,4 83,1 0,1 0,5 0,1 0 0,1 0,246,09 46,53 46,98 45,99 797,4 797,6 87,0 797,9 795,8 0,5 0,9 0,4 0 0,1 0,243,18 43,68 44,17 45,67 963,2 964,4 105,3 965,6 966,7 2,4 1,0 2,2 0 0,25 0,545,75 45,79 45,84 45,74 1161,6 1162,2 126,9 1162,8 1164,0 1,1 0,1 1,0 0 0,05 0,147,32 47,56 47,81 47,36 2029,0 2030,2 221,9 2031,4 2032,9 2,4 0,5 2,2 0 0,05 0,142,66 42,76 42,85 42,66 2349,6 2350,0 256,9 2350,3 2349,1 0,7 0,2 0,6 0 0,05 0,146,22 46,71 47,21 43,67 2479,6 2481,1 271,3 2482,5 2481,6 2,9 1,0 2,7 0 0,1 0,246,29 46,74 47,18 48,83 2519,0 2519,5 275,5 2519,9 2520,8 0,9 0,9 0,8 0 0,15 0,345,97 46,22 46,46 46,77 2777,3 2777,6 303,7 2777,8 2777,6 0,5 0,5 0,4 0 0,1 0,2

43,25 43,99 44,73 44,79 2898,5 2900,0 317,1 2901,5 2898,2 3,0 1,5 2,8 0 2,5 543,54 44,38 45,23 42,33 3135,6 3135,7 342,9 3135,9 3134,8 0,3 1,7 0,2 0 0,1 0,2

Estimated�s ize�of�defects :�acc.�projection�R�and�conververs ion�degrees �to�mm�

4.3 Radiographic testing

4.3.1 Principle

X-rays can penetrate solid matter, they are differentially absorbed or scattered by different media and they may be diffracted by crystalline materials. Their largest use is to take images of the inside of objects in radiography testing (RT) and to analyze the crystalline structure of materials. X-rays are produced by interaction of high energy electrons or ions with matter. Bremsstrahlung is radiation given off by the electrons as they are scattered by the strong electric field near the high-Z (proton number) nuclei. These X-rays have a continuous spectrum. The intensity of the X-rays increases with decreasing frequency, from zero at the energy of the incident electrons. The energy corresponds to the voltage on the X-ray tube. The resulting output of a tube consists of a continuous Bremsstrahlung spectrum falling off to zero at the tube voltage, plus several spikes at the characteristic lines (characteristic spectrum, see Figure 35, http://www.everyspec.com on, 2005). The wavelength is decreasing with the increasing energy. The smaller the wavelength is the more efficiently penetrating X-rays are.

44

Figure 35. Typical X-ray spectrum.

When the X-rays penetrate through material several interactions take place. The most important one with regard to NDT is the absorption process. The absorption of the photons is a result of the photon either striking an electron or entering the nuclear field of the atom. The energy lost by a radiation beam as it travels through matter is due to interactions of the photons with matter. In these interactions, the energy of the photon is transferred principally through three processes. These are photoelectric absorption, Compton's effect and pair production. The key parameters in radiation absorption are the density as well as the thickness of the material. Of course, the used energy correlates to the radiation intensity variation as well as the geometry of the object. In industrial radiographic applications the difference in thickness (often due to discontinuities) is the actual parameter from which the interpretation is made. Therefore, the greater the change in the radiation transmitted due to a particular change in material thickness, the more obvious is the thickness change revealed in the final image. This radiation difference due to material thickness change is called the material contrast. The material contrast is a function of the absorption characteristics of the object inspected and the radiation energy level (Figure 36, http://hyperphysics.phy-astr.gsu.edu/Hbase/quantum/ xrayc.html #c2, Sandlin, 2009a).

45

Figure 36. "Bremsstrahlung" and energy dependent X-ray process in radiographic testing.

4.3.2 Equipment

The operating principle of a linear accelerator (or an X-ray tube) is to accelerate electrons in an evacuated tube to high kinetic energy (Figure 37). As these energetic electrons strike a metallic target they lose their kinetic energy in one or usually many collisions with the target atoms. These multiple collisions are responsible for the energy distribution of the X-rays as the electrons produce photons of different energy in each collision. The acceleration voltage of the Varian Linatron 3000 is 9 MV. This means that the electrons hitting the target to produce X-rays have a kinetic energy of 9 MeV; however, only very few of the electrons loose “all” their kinetic energy in a single collision with a target atom, producing a 9 MeV X-ray photon. Instead, most electrons undergo multiple collisions with the target atoms and thus produce several X-ray photons of considerable energy less than 9 MeV. According to an estimation presented in Müller et al. 2006, the intensity spectrum of the X-rays leaving the accelerator will have the maximum intensity of the X-ray photons around 3 MeV.

46

Figure 37. The layout of linear accelerator (left) and measurement arrangement using 9 MeV (right).

As mentioned above, the detector is a collimated line detector. The detector was delivered by (BIR) Bio-Imaging Research Inc (Figure 38). The slit of the collimator has a width of 0.4 mm in the horizontal direction. The scintillator elements of the detector have a total length of about 100 mm in the direction parallel to the slit. The detector array is composed of 2048 lines perpendicular to the slit. Each line has a width of 0.05 mm. Eight lines are used to form one pixel, so they produce a pixel size of 0.4 mm in the direction of the slit. In the perpendicular direction the pixel size is determined by the width of the slit, which is also 0.4 mm as already stated.

Figure 38. Detector for radiographic testing.


In order to test radiographic detection capability, the following reference defects were designed and manufactured with the EDM method (Pitkänen et al. 2007b):

Volumetric defects in square and cylindrical form simulating gas porosity cavities and foreign material like wolfram

47

Volumetric defects in J-Form and U-form simulating small cold lap that can be detected easily with ultrasonic but more difficult to detect with radiography

Surface notches on the surface which simulate crack type of defects and run out defect types

Long EDM drill holes (50 mm long) simulating gun discharge defects

Defects made in oval form simulating spiking and root defect type.

The reference specimen (70° from whole circumference), a length of 450 mm, was cut out from a real size copper weld. The outer diameter of the specimen was 1050 mm. The EB weld of the specimen was controlled with ultrasonic testing and radiography before producing reference defects. The reference specimen is shown in Figure 39.

Figure 39. Radiography reference specimen containing about 50 defects for feasibility study of probability of detection (POD).

48

4.3.4 Data acquisition

During measurement 100 pulses are delivered by the 9 MeV accelerator during 400 ms. After that the accelerator rests for 10 ms and after that the process is repeated. This means that 100 pulses or equivalently 400 ms is used for the exposure of one picture element. The picture element is read out from the detector during the 10 ms before the accelerator- delivers the following 100 pulses for the next picture element (Ronneteg 2006).

0

20

40

60

80

100

120

0 10 20 30 40 50 60 70 80 90Detector position [mm]

Thic

knes

s [m

m]

Exposure angle 20

Exposure angle 10

Figure 40. Principle of radiographic setup using 10�.

A collimator is placed at the front of the accelerator focus of the beam. The detector is covered by a housing of tungsten to reduce distortion by scatter radiation. The X-rays pass through the 70 mm housing in a 0.4 mm-wide vertical slit. As the penetrated thickness varies in the X-ray inspection of the weld, a thickness correction is made to calibrate the system (Figure 40). The correction is made by a mean value calibration from 500 samples around the weld circumference. It can be seen in the radiographic images of Figure 41 on the right that the software does not perfectly correct the thickness variations. Figure 40 shows the thickness variation, which is seen by the detector with exposure angles of 10° and 20°. The penetrated wall thickness is at an angle of 10° about 20 mm less than an angle of 20° in the root area. For this reason mainly an angle of 10° has been chosen for this weld inspection. For instance, in the case of FSW, the weld is welded from the side of the canister and the used angle of exposure is about 30°. Figure 41 shows the image size in pixels made by radiographic inspection

Figure 41. Measured radiographic image in pixels.

49


The image quality indicators (IQI) used in this study were wire type (SFS EN 462 W1 Cu 25, SFS EN 462 W6 Cu 25, SFS EN 462 W6 Cu 50), duplex wire type (SFS EN 462-5), hole type and step-hole type indicators (ASME). The duplex wire IQI is used to measure image un-sharpness while the others are used to measure image contrast.

4.3.6 Data analysis

Data analysis can be divided into defect detection and defect sizing. Defect detection in radiography is concentrated on distinguishing intensity differences in a digital image. When the intensity variations in a digital image are sufficiently high, there is a possibility to detect and characterize the defect. The area describes the size of the defect and, on the other hand, the intensity value is a measure for the size of the defect in the X-ray beam direction.


Reliable defect detection is most important in all NDT methods. In different NDT methods the procedure of detection varies. In radiography the local intensity variation behind the penetrated copper weld is measured by a digital image detector. The noise level can be determined according to SFS EN-14784:2005 and defect detection according to SFS EN-1435:1998 and imaging properties according to SFS EN-13068-1:2000. With the help of this normalized SNR the detection limit can be estimated in the digital radiography of copper EB welds. Measurements with reference specimen of the EB-weld have been reported by Sandlin (2009b and 2009c).

Figure 42. Radiographic measurements of reference defects.

Depending on the radiographic contrast, the evaluation of X-ray images often gives locally a higher visual detectability than what the standard offers: for instance, you can detect more wires than you see from the signal to noise values of those wires. This can be seen, for instance, from the high contrast wolfram double wires in Figure 42. The smallest defect in this image is 1x1x1 mm3. In beam direction the depth of defect is 1 mm and for the grey level it gives a value of 237. The grey value of the noise is about 100. The defects in real components can be different to reference defects. The detectability depends on the defect width and length as well depth in beam direction as the main factors. Volumetric defect types can be detected quite easily until the grey level is 200.

50

In Figure 43 a small defect in blue colour is shown and it can be clearly seen, even along the line with considerable grey value changes. These small defects are visualized in the black box area. The thin line indication, where the grey value is a little bit over 200, is hardly detectable. This thin line is shown in the green box (Figure 43). In images single pixel clusters can be found, which are in square form or long lines which have a very high intensity. Both types of indications are inhomogeneities of the image detector caused by electrical disturbances or faulty pixels. They can cause false calls when they are not eliminated by a calibration procedure. The detectability can be improved with different kinds of filtering and clearing of distortions from geometrical and adjusting effects.

Figure 43. Detected defects in EB weld in the area where control parameters are outside of specification width.

One special case has been also studied when a defect is much larger in depth than the detectability limit for the inspection method. One preliminary in radiography is shown in Figure 44. It shows four 50 mm-long holes extending from the surface to the root of the weld. The diameter of the 50 mm holes varies from 0.5 mm to 2 mm. It can be

51

clearly seen that the 0.5 mm hole is near to the limit where the intensity starts to decrease clearly but it is still more than noise level. This means that a 50 mm-long hole with a diameter of 0.5 mm can be distinguished well. The noise level seems to be 100 grey levels in this case. This case will be studied more thoroughly and also with other inspection methods.

Figure 44. Defect detectability of through weld exceeding hole in radiography.

4.3.8 Defect sizing

In radiographic testing the size of the defects for EB welds will be evaluated from the image. We can measure the axial and circumferential sizes of the indication (Figure 45 on the right). After the detection of indications in the radiographic image it will be assured whether or not they originate from surface scratches on the upper outer surface of the tube, or weld surface or lift surface of the lid. The sizing in radial direction is based on the grey value correspondence to defect size in X-ray beam direction (Figure 45 on the left). This will be used as a measure for defect depth in the radial direction. At the moment, the location of the defect in radial direction will be assumed to be in the middle of the EB weld.

Figure 45. Defect sizing in radiographic inspection in axial, circumferential and radial direction.

52


Each indication is identified by a specified number; for instance, Xk035-RT1. From an indication suspected to be a defect, the indication grey value (IGV) will be measured. The average noise level will be estimated (IG-BL=intensity grey value of base line). The difference between the indication grey value and the base line grey value will be the computed value and the variation of the noise will be evaluated (GV-BL-std.dev). From the grey value difference taken into consideration for the noise variation, the defect size in beam direction will be estimated and computed. This base to the size of several reference defects in beam direction and defect will be positioned in the middle of the weld at the moment. This more accurate evaluation for positioning of indication from radiographic image is under study. The other sizes of indication can be easily determined from the image, as shown in Figure 45. These values in circumferential and axial direction will be reported in form of Table 6. The size of indication will be given and this size can be evaluated against the acceptance criteria. The inspection will also evaluate and characterize the type of indication.

Ta

ble

6. R

epor

ting

indi

catio

n in

radi

ogra

phic

insp

ectio

n.

53

54

4.4 Ultrasonic testing

4.4.1 Principle

Ultrasonic Testing (UT) uses high frequency sound energy to produce sound waves and to make measurements. Ultrasonic inspection can be used for defect detection, evaluation, dimensional measurements and material characterization. To illustrate the general inspection principle, a typical pulse/echo inspection configuration, as illustrated below, will be normally used in ultrasonic inspection.

A typical UT inspection system consists of the pulser/receiver, transducer and display. A pulser/receiver is an electronic device that can produce high voltage electrical pulses. Driven by the pulser, the transducer generates high frequency ultrasonic energy. The sound energy will be generated and it propagates through the materials in the form of waves and back to the ultrasonic probe when reflected. When there is a defect in the sound path, part of the energy will be reflected back from the defect surface. The reflected signals will be transformed back into an electrical signal in the transducer and the signal is displayed on a screen of ultrasonic equipment.

Signal travel time can be directly related to the distance that the acoustic waves have travelled. From the measured signals, information from the defect location, size, orientation and other features can be gained. Detailed information on ultrasonic testing is given in Krautkrämer J. and Krautkrämer H. 1990.

4.4.2 Equipment

At the moment, Posiva applies MultiX the full parallel architecture phased array-system (PA-system) of M2M for ultrasonic inspection. This equipment contains 128 parallel channels (Figure 46). This equipment can drive conventional probes, normal linear or matrix PA-probes. The system is checked annually and in each measurement period the calibration check is performed in following way:

- At the beginning of the inspection

- After changing any part of the inspection system

- Whenever alternations in the system performance are suspected

- At the beginning and at the end of each inspection shift

- After 12 h inspection period if a calibration check is otherwise performed at the end of the inspection

55

Encoder pulsesUSB connection

Probe cable

Multi 2000

PCScanner

Figure 46. Measurement equipment for ultrasonic inspection of copper and cast iron insert components and welds.

The analysis of measured data is performed using the latest updated Civa software version. The welds are scanned with Posiva's lid and weld scanner shown in Figure 47. The linear phased array probe is positioned as shown in Figure 48. Ultrasonic inspection of welds is carried out using phased array technology.

56

Figure 47. The manipulator is seen in the background. And a laptop controlling the manipulator is shown at the front on the left, and the ultrasonic equipment MultiX is shown at the front on the right.

1

128

20

Figure 48. Ultrasonic phased array probe (3.5 MHz, 128 elements) and its positioning for weld inspection.

57

Ultrasonic waves fill the volume of the whole weld by scanning the probe shown in the Figure 49 around the whole weld. The scanning step is realized in the axial direction electronically scanning at the size of 1 mm, which is the element size in the phased array probe. Mechanical scanning is performed in the circumferential direction typically using a step of 2 mm.

Rotation

Scan path

Rotation

Scan path

Figure 49. The rotation of the ultrasonic probe around the whole weld. 4.4.3 Conventional ultrasonic probes

Conventional probes are here determined simply that the generated sound field of the probe cannot be modified during data acquisition. There are several types of typical conventional ultrasonic probes (Pitkänen 1991). The shear wave probe, which produces shear waves, which in copper has velocity of 2160 m/s. The usability of a shear wave probe for copper inspection is limited because of the grain size variation in base material and the anisotropy properties of the EB weld. This makes it difficult to use shear waves in the inspection of copper. A shear wave probe can be either a single piezo crystal probe or a dual piezo crystal, which is called a transmitter receiver probe. Acoustically, shear waves can also be produced with phased array probes.

Longitudinal wave probes are better suited for the inspection of copper material. Longitudinal waves can be produced similarly as shear waves. Longitudinal waves possess the velocity in copper of 4700 m/s. The longitudinal waves possess two main advantages for the inspection of copper. The longitudinal wave probes, having the same frequency as shear wave probes, have lower attenuation (Matthies 1984). According his measurements, longitudinal waves and shear waves attenuate similarly if they have the same wavelengths. In our case, by using 3.5 MHz the wavelength is 1.3 mm. With shear wave the wavelength would be about 0.6 mm. So, in any case, the shear would attenuate faster than a longitudinal wave using the same frequency. A shear wave frequency of about 1.8 MHz would have similar attenuation as 3.5 MHz for longitudinal waves.

Another special type of probe used for copper inspection is a mode conversion probe, which is a very complicated probe type. This type of probe is described in the inspection report made for Posiva. The creeping wave propagating along the scanning surface can be used for detection of surface opening defects and the simultaneously created direct longitudinal wave with a large angle of incidence for detection of defects just below the surface (Figure 50). The characteristics of the probe are shown in Figure 50. This type of probe produces longitudinal waves and shear waves at the same time.

58

Additionally, it produces primary creeping waves which can be used for the detection of surface or near-surface defects. The head waves also send the wave front on the other side of wall thickness, which produces secondary creeping waves on the opposite side of the component and further on multiple creeping. In the material, so-called mode conversion of the reflected waves can occur because of the component geometry.

Figure 50. Mode conversion technique for detection of surface and near-surface defects.

4.4.4 Linear phased array probes

The sound field of the ultrasonic probe in the phased array technique is controlled and formed electronically. The phased array probe consists of several piezoelectric elements with a fixed geometric shape.

The operation of each of these single elements can be controlled separately by switching them on or off at any time by the system. The excitation of a single element can be performed with sequential pulses with a time difference or different amplitudes. These methods can also be combined. The typical way is to use a time difference to achieve a different angle of incidence or a different focus depth (Figure 51).

The resulting sound field of the phased array probe depends on the shape of the element groups and on the positions of used elements in the elementary group. Figure 51 shows the principle of a phased array probe compared to a conventional probe.

A phased array probe can have different focusing depths; different angles of incidence can excite different parts of probe, i.e. different areas from the component to be inspected without any need to make mechanical movement. By changing the angle of incidence ultrasonic phased array equipment changes the so-called delay laws and the angle of incidence changes correspondingly. These items are discussed in detail, for instance, in reference to advances in phased array ultrasonic technology applications (Olympus 2008, RD-Tech corp, 2004). Depending on the geometry of the PA probe, delay laws of the linear phased array in the excited sound field vary.

59

Figure 51. Principe of ultrasonic linear phased array.

4.4.5 Matrix phased array probes

Matrix phased array probes have an active aperture divided into a two aperture axis: primary and secondary aperture (Figure 52). This division can be in the different forms of a checkerboard or sectored rings. These probes allow the ultrasonic beam to be driven in 3D volume of inspected components by combining electronic focusing, deflection and steering. By using matrix phased array 3D defects can be detected more easily. The trials of the use of the matrix phased array have been started. This type of matrix phased array (as in Figure 52) is already in use. In the first stage the application has been introduced to ultrasonic inspection of plate welds, which are used to study the welding method as well as welding the reference welding according to the welding standards. This type of sound field forming gives the possibility to detect the most demanding types of defects. But basically the nature of defects must be known before being able to design the best possible matrix probe for inspection. This matrix phased array technique will be applied in 2010 for real weld geometry in order to improve the detectability of defects. The main reason is that the real defect types are three dimensional in copper EB welds. And it seems that in FS welds they are also in the same way three dimensional.

4.4.6 Ultrasonic Inspection techniques for weld inspection

The evaluation of ultrasonic inspection is carried out using the latest Civa software version. A- , B- and C-scan images are used for the visualization of the indications. In C-scan the whole weld as a top view with possible indications is shown, which has to be analyzed. On a colour scale red shows a defect and dark green is a non-indication area. Using zoom, the size of one separate defect can be determined; by positioning the

60

cursor the A-scan will show the amplitude of the indication related to its position in the weld volume. This amplitude of indication will be compared to the amplitude of the reference defect. If the indication exceeds the registration level, it will be analyzed and the size of the indication will be measured from the corresponding positions in the B- and C-scans.

For inspection of EB welds different ultrasonic inspection techniques will be used. All techniques used are shown in Figure 53. The basic inspection technique uses 0 °L, completing inspection techniques using angles of incidence of ±20 °L. For surface and partly for root inspection a technique with an angle of -20° is used. This technique of +20° is also used for sizing purposes.

Figure 52. Layout for matrix phased array probe.

These different techniques are shown in Figure 53. Using an electronic scan in the direction of the weld penetration, sound beam directions 0° and ± 20° and focusing at the weld depth (Figure 53). Gain values for 0° and ± 20° data acquisition will be set for same level with the help of a reference specimen. The signals from the FBHs (Flat Bottom Holes) and SDHs (Side Drill Holes) located in the front and behind the weld are measured and used as a reference. The used reference defect is FBH and its size is �3 mm.

61

Figure 53. Ultrasonic inspection principle of copper EB weld inspection with phased array system.

Figure 54. The inspection techniques for FSW inspection.

62

For inspection evaluation of an FS weld the following techniques will be applied: the basic inspection technique will be performed using 0 °L surface, the root volume will be inspected using angle scanning. Root and surface inspection for curved defects will be carried out with an angle of -20° as shown in Figure 54. All techniques are visualized in fixed circumferential position showing one electronic scan of that area. The mechanical scan is performed in a similar manner as in the EB weld inspection around the weld.

Different focal laws will be applied to perform the inspections in both weld type inspections. To avoid saturated signals from the key reference reflectors low gain setting values must be applied.


The gain setting can be performed using specific reference specimens, which are made using the same geometry as in the real size canisters. The reference specimen for ultrasonic testing of the EB weld is shown in Figure 55. The reference specimen contains FBH defects for the evaluation of defect amplitude of planar type defects and SDH defects will be used for the evaluation of volumetric type defects. Notches on the surface simulate surface breaking type defects. These types of reference defects are inserted in this specimen.

Figure 55. Reference specimen for ultrasonic inspection techniques of EB weld.


To avoid saturated signals from the key reference reflectors low gain setting values must be applied. The signals from the FBHs and SDHs located in front of and behind the weld must be measured and used as reference. The main reference defect size is �3 mm FBH. The gain will be set so that the sensitivity of different techniques (0° and ±

63

20°) will be same. This procedure will be carried out with the help of a reference specimen.

4.4.9 Data analysis

In data analysis it must first be confirmed that data acquisition has been carried out properly. The first item to check is the quality of the data against bad contact, interference peaks from the surroundings and the designed inspection pattern according to procedure has been followed. The next step is to check that no saturation in signals is present. After that it is possible to check the calibration. If all these items are acceptable, the defect detection and defect sizing and characterization can follow.

Defect detection

The estimated size of the beam varies between 3.2 mm and 3.4 mm in the weld volume (Figure 56). The attenuation is about 10 dB in the weld area, which must be considered in the evaluation of defects. This clearly causes the increase of the detectable defect size behind the weld. In front of the weld according to measurement from the signal to noise ratio (SNR) of �3 mm FBH gives SNR of 35 dB and this is measured for minimum detectable defect size in the optimum case. In the estimation of the smallest defect which can be found, the SNR must be more or the same as 6 dB. The grain size variation in the tube material can vary the detectable defect size in front of the weld. The grain size effect on the detectability of defects in the ultrasonic testing has been discussed in study presented 2007 (Pitkänen et al. 2007c).

Figure 56. Size estimation of the phased array sound field in the volume of the weld.

The inspection of the EB weld for this geometry is actually the only possibility to carry out inspections from the outer side of the tube. The grain size is large in the axial and in the circumferential direction. In the radial direction from which direction the ultrasonic inspection is carried out, the grain size distribution makes the ultrasonic inspection

64

possible. There are many possibilities for improving the detectability of defects and Posiva will study some possibilities, such as sampling the phased array, full matrix capture technique and the evaluation of the matrix phased array and TOFD (Time Of Flight Diffraction Technique), but the impact of these special techniques is mainly in defect sizing, although of course also partly in defect detection. The POD of the EB weld will be studied in co-operation with BAM and the preliminary results will be available at the end of 2010.

The behaviour of weld reference defects in ultrasonic inspection has been modelled with the help of Civa. Reflectors were inserted in a copper cylinder (outer diameter 1050 mm) with a ligament of 44 mm from the scanning surface. The diameters of the FBH varied from 0.5 mm to 3 mm. Also, a 3 mm side drilled hole (SDH) was inserted as a reference (Figure 57). Also, hemispherical-bottom holes (HSBH) were applied as defects to better imitate real defects. The �3 mm FBH and SDH were included in the simulation as reference. The different reflector types are shown in Figure 58 as modelled by the simulation software.

The probe applied was a 3.5 MHz 128 element linear phased array probe (Imasonic 6636, same probe characteristics as used in the weld inspection). The delay laws applied were as defined in the inspection procedure. Aperture 22 elements focusing on a depth of 44 mm. 0° delay laws were computed both with single point focusing (depth 44 mm) and direction and depth (0°/44 mm).

Figure 57. Simulation layout: 0° (left) and 20° (right).

Figure 58. The reflector types applied in the simulation. Hemispherical-bottom hole (HSBH) left, flat-bottom hole (FBH) in middle and side drilled hole (SDH) right. The applied beam direction is shown with a green line.

65

According to the simulation, the amplitude drop is very similar at 0° and 20°. The rounded estimate that applies for both cases is given in Table 7. Table 7. Amplitude drop compared to 3 mm FBH using flat-bottom (FBH) and hemispherical-bottom (HSBH) reflectors.

Reflector diameter (mm)

Amplitude drop of FBHs (dB)

Amplitude drop of HSBHs (dB)

3 0 -15 2 -6 -18.5 1 -17 -22.5

0.75 -21 -25 0.5 -28 -30

In ultrasonic inspection analysis the threshold (recording threshold) is set 21 dB below the signal level of the 3 mm FBH, which is a reference defect. Thus, in theory, defects with a flat surface with a diameter of 0.75 mm and oriented perpendicularly to the beam can be detected. On the other hand, if the defect reflecting surface is spherical, only defects with a diameter of 3 mm (15 dB below the reference reflector amplitude) and 2 mm (18.5 dB below the reference reflector amplitude) can be detected. Spherical reflectors with a 1 mm diameter (22.5 dB below the reference reflector amplitude) are already below the analysis threshold. According to the simulation, the 3 mm SDH produces a signal amplitude about 1 dB higher than the 3 mm FBH when a 0° beam angle is used. These signal levels are equal when a 20° beam angle is applied. According to the simulation, the signal levels of a 20° beam are about 3.1–3.5 dB lower compared to signal levels of the 0° beam. In other words, a 3–3.5 dB higher gain should be applied at the 20° angle inspection technique compared to the 0° inspection technique to achieve equal sensitivity in inspection. The simulation has not taken into account the attenuation of the material or the higher noise of the weld zone. Both of these factors will in reality lower the detectability of the flaws. In the data for the reference specimen measurements the signal noise ratio is about 35 dB from the 3 mm FBH. It gives the minimum detectable diameter at about 0.7 mm (Figure 59). A more practical value seems to be 1 mm in front of the weld and 2 mm behind the weld in the optimum case considering the noise level in the measured welds.

Defect sizing

For UT the defect sizing is shown in Figure 60. All three directions will be measured. The axial and circumferential size will be received from the C-scan data and the radial size of the defect will be estimated from the B-scan. The amplitude value and dynamic behaviour of the defect will be one part in sizing using the A-scan. The angle inspection (±20°) must be considered in sizing simultaneously with the result of a 0° angle to give the final result of ultrasonic testing. In FSW the defect has been sized using angular scanning.

66

Figure 59. Defect detectability in ultrasonic testing of the EB weld.

Figure 60. Defect sizing in ultrasonic testing using a phased array system with a 0° angle.

Figure 61. Sizing of different types of defects in ultrasonic testing.

67

Figure 62. Defect sizing in an FSW sample.

4.4.10 Raw data evaluation

Raw data evaluation for the EB weld is based at the moment on using three different techniques mentioned in ultrasonic inspection for weld measurement. These three different techniques are shown in Figure 53. When a defect can be detected using all of these techniques, it is indicating a volumetric defect type as demonstrated in Figure 63. The length seems to be about 12 mm in circumferential direction and axial direction about 8–10 mm. The defect will be analyzed using three different views: C-, B- and A-scans. The defect dimension is given by the C- and B- scans in the circumferential and axial directions, and the A-scan gives the amplitude compared to reference amplitude (FBH 3 mm) and position in the radial direction.

Figure 63. Raw data evaluation for EB weld measurements using three different techniques and combining the indication results from each technique to one defect.

68

4.4.11 Specific data evaluation

More accurate defect sizing needs specific tools for the evaluation of data. The following methods can be used for improving the data:

- analysis using tip diffracted signals - optimized sound field shaping using matrix phased array - sampling phased array technique - Synthetic Aperture Focusing Technique (SAFT) - full matrix capture technique - TOFD - sizing using sub harmonics - modelling.

The more exact characterization of the defects gives information for discrimination of defects to volume or planar defects and allows for interpreting whether a signal is a false call or not. A typical method of using a diffraction technique for sizing purposes will be studied as well the specific TOFD inspection technique. The use of longitudinal waves improves the detectability of the tip diffraction signals as has been stated (Müller 1999 and Erhard 2004). The sizing studies will be carried out in 2010 and 2011. The studies contain all the methods mentioned above and will be reported in 2012. By using the matrix phased array 3D defects detectability can be improved. Focusing is more efficient in the 3D way with a matrix phased array compared to a linear phased array. Figure 64 shows indications from notches behind the EB weld detected with a focused TRL matrix phased array (Transmitter-Receiver-Longitudinal, TRL) using a 50° angle of incidence. The thickness of the copper plate is 60 mm and the weld width is about 10 mm. Figure 64 shows that even a 2.5 mm notch can be detected with 20 dB signal to noise ratio (SNR) (Figure 64). According to this, it can be assumed that even a 1 mm notch could be detected with sufficient SNR (6 dB) behind the weld.

Figure 64. Indications from reference notches behind the weld using a matrix phased array.

69

Tip diffraction is the simplest method for accurate sizing and it can be applied to raw data images or to images using more sophisticated image processing like SAFT, Sampling phased array - SPA (Bulavinov 2005), or full matrix capture technique - FMC (Calmon et al. 2007, Poidevin et al. 2007). The tip information can also be improved using a specific method like subharmonics (Akino et al. 2003, Zagrai et al. 2003 and Yamanaka et al. 2004). The subharmonic signals can be better detected because of the edge of defect transforms sent acoustic wave signals from the ultrasonic probe and the content of the returning signals from the edge is pronounced with subharmonic content (f/2) - half of the original frequency. This can also be applied for tight planar defects as well as for crack type defects. Planar type defects can be a lack of fusion defects or semitransparent defects, such as the kissing bond defect.

The principle of a sampling phased array is shown in Figure 65. It shows that by a transmission the transmitted signal from one single transmitter element can in reception be detected in each separate element, which improves the possibility of detecting a smaller signal and forming a very clear picture of the defects in the material. Even the defects near the surface can be better resolved than in conventional phased array inspection. One result from side drill measurements shows that near the surface the pseudo indication in conventional phased array visualization disappears using the sampling phased array (Figure 66). Posiva used a sampling phased array in the first trial together with Fraunhofer-IzfP for a copper specimen with a different grain size. The system set up is shown in Figure 67 with the results of smaller grain size material and larger grain size material. The effect of grain size is shown with a lower signal level of the reference defects resulting from a larger grain size seen on the right of the image in Figure 67.

Figure 65. Principle of sampling phased array.

70

Figure 66. Comparison between conventional phased array and sampling phased array.

Figure 67. Measurement set up with different grain size specimen using a sampling phased array. Ultrasonic measurements using 3D visualization is shown below. On the lower left is shown a copper specimen with a smaller grain size compared to the lower right.

71

Figure 68. Reference defect measurements (FBH) in copper using a full matrix capture technique. At the moment, the full matrix capture technique uses from all transmitter elements separately transmitted signals. They are received with all receiver elements separately, and this leads to a huge amount of data. Thus, it is possible to measure only locally using the FMC technique. Figure 68 shows measurements from the reference FBH defects in the EB copper weld. The sizes will become smaller using the FMC technique, but it is time consuming to apply. Some improvements can be possibly added to the software. The average size of defects approached about 2 mm smaller closer to the real defect size. The variation of defect sizes was from 3–4 mm and real sizes of 3 mm, and the raw data evaluation using the 6 dB method gave a result of about 5 mm.

The sampling phased array technique and full matrix capture technique are both similar to the SAFT method. SAFT is one tool for improved understanding of detected ultrasonic signals. The imaging system applied to improve the reliability of the assessment is the use of back propagation of elastic waves in a synthetic aperture focusing technique (Doctor et al. 1995, Kröning & Schmitz 1997, Müller et al. 1985, Schmitz 1988, Osetrov 1995). The SAFT technique has been used for testing pipes, turbines, plates, vessels or pump housings of austenitic and ferritic materials.

The features of a SAFT inspection system can be described in the following way:

- The improvement of signal-to-noise ratio (SNR) is one of the main advantages, which is considered in the theoretical background of SAFT.

- SAFT reconstruction is valid in the far field (Fraunhofer Region).

- This technique is mainly used for single crystal probes but applications of tandem probes and TOFD have also been realized in various software.

72

- The main advantage of SAFT is that the focusing possibility, which is independent of the length of the sound path. This gives possibilities which are not yet fully applied; DDF (Dynamic Depth Focusing) in the phased array technique is approaching this idea.

- SAFT reconstruction offers several tools for sizing using mainly crack tip detection, but not necessarily. More exact ultrasonic field visualization makes it easier to interpret the results.

- After SAFT reconstruction, dynamic 3D sizing can improve crack detection, sizing and characterization capabilities in ultrasonic testing.

- Modelling gives information for understanding measured patterns of many ultrasonic inspection tasks also in SAFT.

The measured signal has back propagated into the material. Even though the image is better than the raw data, the calculated image can still contain, for instance, some artefacts which have to be recognized. In the SAFT measurements those artefacts can be seen, for instance, in turbine axel inspection near the surface. Similar artefacts can be seen, for instance, in the phased array field with strong focusing in the near area of the probe. The modelling gives the possibility of improving the SAFT algorithm to take into consideration the changes in the inspection area, such as the austenitic weld or changes in the grain size and anisotropy (Shlivinski et al. 2004).

Figure 69. Principle of SAFT measurement.

When mechanized ultrasonic inspection data is analyzed it can be analyzed conventionally using static A-scan, B-scan and C-scan images or dynamically letting the data run in different images depending on the moving probe position. Dynamic sizing can be carried out either with a phased array using angular scanning or with SAFT or acoustic holography. After reconstruction the sizing is carried out dynamically going through the inspected material layer by layer to find the limits of the defects, as in this case the depth and the length of the crack with the help of the SAFT reconstructed data. Dynamic sizing can also be carried out in mechanized ultrasonic raw data form as done according to Kenefick & Henry 2001. Dynamic B-scan, C-scan or D-scan evaluation methods will, in the future, be the main evaluation methods for defect

73

characterization. Crack tip indication is clearly seen at a depth of 9 mm, it starts to vanish at a depth of 10.4 mm. The dynamic evaluation is carried out visually looking from the indications above the noise level by going through different frames where the indication is detectable. The noise level can be changed during the evaluation, which gives valuable information to the analysis. This measured fatigue crack is in the ferritic specimen. The wall thickness of the material is about 40 mm. The measured crack is surface breaking. Dynamic sizing differs from the static because one can go forward and backward within the reconstructed data and confirm the indication. Of course, in this case it is also the main interest to choose the best possible probe configuration for the measurement. The same idea from the dynamic sound field is in phased array inspections but in the opposite order. The ultrasonic field is dynamic and moving. Indications are moving due to a variable sound field. This gives a better understanding of the nature of the defects. By moving the probe, the crack front starts to move according to the movement of the probe. This type of SAFT software was developed by Fraunhofer-IzfP (Schmitz & Chakhlov 2001). Dynamic evaluation has been used in the evaluation of the indication in Figure 70.

Figure 70. Dynamic sizing using SAFT in different views (Pitkänen 2006).

74

SAFT has not yet been used so often with phased array equipment. In 2004, some sizing results were presented by using phased array raw data for SAFT reconstruction (Erhard et al. 2004). According to the measurements, the tip signals were located and constructed according to the example shown in Figure 71. In the experiments done by Erhard et al. 2004, it was shown that the tip was better detected with longitudinal waves than with shear waves. It was also demonstrated by Müller 1999. For the phased array application, the geometry of the PA probe must be known quite exactly as well as the movement on the surface. By changing the angles during the measurement the index point changes in the PA probe. For a good SAFT reconstruction it is necessary to have exact knowledge of the probe sound field characteristics at different angles.

TOFD technique can be used for evaluation of indications and it may be a solution for the evaluation of specific and difficult defect orientations (Charlesworth, J.P. and Temple J. A., 2001). In TOFD technique transmitter and receiver probes are positioned on the different side of the defect as shown in Figure 72. One of the main design principles is to make the piezo element in the TOFD probe as small as possible in order to produce a sound field which opens very quickly. This is limited of course by the detectability of small signals in the piezo material of the probe. The TOFD technique can be applied using conventional TOFD probes or phased array probes. The inspection technique principle of copper EB weld is shown in Figure 72.

Figure 71. Defect sizing with a phased array SAFT (Erhard 2004).

75

Figure 72. Principle of the TOFD technique for copper EB weld inspection.

Also, modelling as the serving tool for evaluation help for complex indications can considered. Sometimes it is impossible to produce certain defects in the reference specimen which can be produced in manufacturing. By modelling it is possible to simulate defects in the right positions and sizing technique can be derived from the results of the modelling without having any real measurements. Normally this type of procedure needs a clear defect type, which can be modelled for simulation of the inspection. This type of defect can be detected in metallographic studies or the possibility for its occurrence in manufacturing is high enough, that is, it must be taken into account for the design of the inspection.


All detected defects will be reported according to Table 8. In the table the ultrasonic equipment set up for each ultrasonic technique and reference level originating from the reference defect (a flat bottom hole having diameter of 3 mm) will be indicated. Some measures from the weld will be reported in case it is possible, such as the penetration of the weld. The ultrasonic reporting will contain raw data files, evaluated detected indication images, defect tables of indications. The ultrasonic indication report prepared by a qualified ultrasonic inspector will be added to the reporting documents also containing evaluation of the estimation of the acceptance of the weld according to the results of the ultrasonic inspection. The values shown in Table 8 are from basic ultrasonic techniques used for EB weld inspection shown in Figure 53.

76

Table 8. Ultrasonic indication report.

77

5 COMBINING INDICATIONS - DEFECTS AND INSPECTION RESULTS

Detection and evaluation of damage or defects in a copper canister is aimed at corrosion, not fatigue. The remaining wall thickness is the main base for acceptance of a canister. For long-term safety there are mainly three effects which can cause damage to the copper canister: residual stresses, creep behaviour of the material and corrosion or a combination of all three. Defects can occur either on the surface or in the volume of the canister. Even though the effect of surface defects for long-term safety and corrosion is more significant than the effect of internal defects, the effect is still minor compared to loadings based on fatigue. Detected defects in the same cross section will be summed up together in order to calculate the remaining wall thickness. A reason for combining defects is the inaccuracy (tolerance) of the NDT method, as shown in Figure 73. As a result of an evaluation of two indications in Figure 73, they will be combined into one indication:

Figure 73. Combining defects using method uncertainty.

In depth direction, which is the most important direction for the remaining wall thickness, the combining of defects determines the inaccuracy of the used NDT method in depth sizing. The final result for combining defects can be related to the specific defect analysis of the detected defect. In length direction the combining distance in

78

ASME XI varies depending on the defect type and position. The rule for combining in length direction can also be applied similarly.

The evaluation for defect size for each different method will be, as explained already, a combination of different techniques. All results for each individual NDT method will be evaluated separately and afterwards positioned in a 3D real location. There will be acceptance criteria for single NDT methods and for combined results of several NDT methods. The individual indication results, such as defect locations, are compared and the final size of the defect will be determined according to the combined results of all NDT methods. The evaluation scheme is shown in Figure 74 as well as a preliminary visualization of defects in 3D form.

NDT-Inspection

VT

Measurement

Datastorage

Analysisof data

3D positioningof indications

ET

Measurement

Datastorage

Analysisof data


UT

Measurement

Datastorage

Analysisof data


RT

Measurement

Datastorage

Analysisof data

3D positioningof indications Combiningof defects

Figure 74. Combining of evaluation results and visualisation in 3D form.

79

6 DEFECT DETECTABILITY STUDY 6.1 POD - Probability of detection

Earlier studies of POD curves for the inspection of copper welds were reported in several reports (SKB 2006, Müller et al. 2006, 2007 and 2009, Pavlovic et al. 2008, Pitkänen et al, 2009b) The POD curve with the lower 95 % confidence band is a typical way to present the capability of the NDT system to detect a flaw (Berens 1989). These types of curves were applied for simple cases where it is assumed that the POD of the defect depends only on the defect size and no geometrical influence of the component is taken into account. In applications for a thick copper part of the canister, the POD of the defect is influenced by many parameters and, depending on the position within the inspection volume, the POD of the same defect will be different (Figure 75). The POD of combined methods was presented by (Pavlovic et al. 2009a, 2009b, 2009c). The preliminary result for the POD of radiographic testing was evaluated according to the reference specimen reported in (Pitkänen et al. 2007b). The results shown in Figure 76 are calculated from the image made by a 9 MeV linear accelerator. The other PODs of different NDT methods are under study. They will be ready in 2010. The comparison to results from real defects will be accomplished in 2011. The study of C-scan POD was reported by Takashashi 2009. This type of POD analysis can possibly be of value for the evaluation of mechanized ultrasonic inspection when it is developed to take into account all relevant information used by ultrasonic analyzers, such as A-, B- and C-scans.

Figure 75. Principle of volumetric POD.

80

Figure 76. The preliminary POD for Posiva's EB welding using radiographic testing and ultrasonic testing for FBHs in front of the weld computed by BAM.

6.2 Evaluation of inspection procedures - Human factor effect

EB weld inspection methods are remotely controlled and mechanized inspection techniques. The selection of the measurement techniques, adjustment of the equipment, data acquisition and data evaluation are based on human decisions. These are influenced by the level of experience and qualifications as well as by organizational factors. A human decision is involved throughout the whole process.

A mechanized inspection is a complicated process. All inspection procedures will be separated into different parts and checked for sources of errors, as in the technique called THERP (Technique for Human Error Rate prediction). For the purposes of understanding human factors, all four inspection methods have been analyzed, i.e. broken into specific tasks and subtasks, using a technique called Hierarchical Task Analysis (HTA). The data evaluation process was considered as the most affected by human factors and studied more carefully within the scope of the ongoing BAM-Posiva project. A group of experts, using a customized Failure Mode and Effects Analysis (FMEA) (Fahlbruch 2009), identified the critical errors, their causes and consequences and existing and possible preventive measures. Additionally, the risk priority was calculated for each critical task. This was done for all four techniques for the weld inspection, and for the UT phased array mechanized inspection of copper components and insert in cooperation with SKB. These studies will be finished by 2011. All inspection methods used for EB welding will be analyzed similarly in cooperation with BAM. These studies will be finished in 2011. This FMEA-methodology has been reported by Fahlbruch 2009.

6.3 Verification of results

In NDT inspections it is important to use metallographic studies when verifying and especially when applying typical defects coming from manufacturing or welding, in this case from copper EB welding. In Posiva's weld trial experiments 50 indication positions have been chosen for the metallographic study. For the study, each indication is positioned according to NDT results and the area around the indication will be cut from the copper weld with a typical dimension of 20x20x65 mm3. Each piece will be sliced at

81

depths of about 1.5 mm. Slicing will continue as long as the indication is present in the evaluated volume. The cutting procedure is shown in Figure 77. The slices will be ground and polished and after polishing photographed. The resolution of the image will be kept as good as possible and several images will be combined into a panorama picture. This panorama picture will be evaluated. As a result of the evaluation of defects, they have clear edges in 2D images. These 2D-evaluated image slices will be combined in a 3D image. The gained defect image will be inserted into the weld image drawing and the 3D defect will be later used as a basis of modelling for inspection techniques applying real defects

The results of evaluated indications will be used for an NDT reliability study of each NDT method. The sizing capability will be studied as well as the detectability. No indication areas will be chosen for metallographic study in order to see if there are smaller�defects which are not detected using current inspection techniques. Most of this evaluation will be carried out in 2010 and finalized and reported in 2011.

Figure 77. Metallographic studies of detected indications in XK030 EB weld.

83

7 REQUIREMENTS FOR QUALIFICATION The qualification according to ENIQ recommendations is divided into the following subareas:

- input information - inspection procedure - technical justification based on the study of the inspection procedure - open and/or blind trials.

The qualification body should accompany the qualification already from the beginning. The qualification body should comment on and check that the qualification test specimens for open and blind trials are valid for the qualification purposes. In the technical justification one important part is to study parameters, which are assumed to have an effect on the inspection performance. These parameters can be divided into two groups: essential and influential parameters (Figure 78) (ENIQ 1998).

It is vitally important to define inspection objectives for qualification and the requirements for NDT techniques to be used. Such requirements are, for example, type and minimum size of defects to be detected and also the accuracy that is needed in the defect sizing. During the definition of inspection target values viewpoints of both structural integrity and inspection shall be taken into account.

The volumes to be covered with inspection shall be defined according to the critical locations of defects. The defect nature may further steer the choice of areas that shall be covered by different methods and techniques. The smallest defect size of a certain defect type that has to be detected is defined using the term “detection target”. Usually structural analyses are applied to justify the choices made concerning the detection target. These analyses or other justifications shall be presented as reference material. The final definition of detection target shall also include the defect orientation (tilt and skew angles).

Modern NDT methods also aim to size the detected defects. Therefore the input information should define the tolerances of defect length and height sizing techniques. Usually the defect positioning requirement is also given in the input information.

There are many parameters which can potentially influence the outcome of an inspection. These are the influential parameters. The list of influential parameters to be considered will depend upon the specific inspection to be qualified. Influential parameters qualification, according to ENIQ recommendations for selected inspection technique, is configured for the verification of the applicability of methods in use. The influential parameters are an important part of the technical justification. Identifying the essential parameters for qualification is the main task for technical justification in order to control that they are included in the inspection procedure.

84

Figure 78. Parameters affecting inspection according to ENIQ (ENIQ 1998).

The influential parameters for the NDT inspection system—for example, ultrasonic equipment, manipulators, equipment for data evaluation etc.—has to be determined in such a form that it is possible to make a qualified detection of all defects relevant for a quantitative POD after having changed the testing equipment or having changed parameters within an allowable range. The changed equipment or variation of parameters has to fulfil the same essential parameters and conditions as the equipment used for qualification.

To improve NDT reliability is to find parameters which are essential and study their effect—for instance, on the POD and to find correspondence between parametric change within an allowable range. According to the author’s opinion, it is important to keep in mind two basic points: find a correlation to the POD of the technique used and reveal weak points in the inspection.

The component parameters are as follows: geometry of the component, material of the component (oxygen-free copper, nodular cast iron insert), surface condition (machined with given surface roughness in specification, weld type (Friction steer welding and electron beam welding), weld crown configuration (weld is grinded), wall thickness (wall thicknesses are given in the drawings), diameter of the component (diameter and geometry of the component varies differently in SKB and in Posiva depending on the nuclear fuel disposal component). These parameters are specified in material specifications

In qualifications the component type specific and postulated defect types originating from welding and handling before inspection should be taken into account. It needs be shown that they are detectable in the range given in the input data. In the qualification the inspector learns from the technical justification the reasons for choosing the technique used for the inspection procedure. The inspector’s experience together with knowledge from the origin of the manufacturing defects will improve NDT reliability in order to find all essential defects within the tolerance range of the input data.

85

The qualification process of the spent fuel canister inspections is still at an early stage. The development of the inspection techniques includes several tasks and phases. Preparation for the qualification of different NDT techniques is planned to start in connection with the inspection research and development work. ENIQ based qualifications has been described by Sarkimo and Pitkänen, 2009a. The preliminary planning for scheduling different inspection development phases linked with qualification process is presented in Figure 79. Important part of qualification are reference specimen, which has been reported 2009 (Paussu and Pitkänen).

Figure 79. Overall view of preliminary planning of the inspection development and qualification tasks and their schedule in the case of spent fuel canisters.

87

8 RESULTS AND EXPERIENCES FROM WELD INSPECTIONS

About 40 real size welds and several hundred plate welds have been inspected. Ultrasonic inspection studies of copper EB-weld were started already 1990's (Jeskanen et al., 1997, Jeskanen et al 2001). Posiva has started to also test some FS welds with ultrasonic and eddy current testing methods (Lahdenperä 2009). The inspection techniques have been developed for ultrasonic inspection from conventional probes to phased array probes (Lipponen and Pitkänen 2009a, 2009b, 2009c and 2009d, Pitkänen 2009a, Sarkimo and Pitkänen 2009b). This is due to the fact the phased array probes give more possibilities to change and adjust the beam sound field to different requirements. This phased array technology is also itself going through a change from linear probes to matrix probes. Posiva will test in 2010 matrix probes for weld testing. The main interest is to improve the detectability of curved defects or defects in bad orientation to the linear phased array sound field.

Statistically it seems that ultrasonic inspection finds defects similar to radiography, but in closer study of the indications it was found that the defect indications differ from positions. This is explainable due to the inspection limitations of both methods. The detectability of defects in radiographic testing has been reported 2009b (Sandlin). Eddy current testing improves the surface inspection as well as visual testing. Posiva is verifying now detectability also by metallography of the indications. Replica-method has been applied to characterize detected surface defects (Jäppinen 2009b).

In studies it has been shown that defects have been found and the welds can be categorized according to quality, but some techniques need to be improved. Also, some welding parameters, such as the penetration of depth, can be shown by ultrasonic testing. Larger grain size in the tube makes it more difficult in some cases to distinguish the depth of penetration of the welding. It also seems that a combination of inspection results improves the quality of the inspection. This is under study and will be reported in 2012.

89

9 SUMMARY AND CONCLUSIONS

In this report methods have been described which can be used for the inspection of copper welds. The welding methods have been presented in short. The main interest is in describing the defect types which might occur in this type of weld because it is the basis for designing inspection methods. Each defect type must be considered from the inspection point of view. This mainly means consideration of defect detectability, probing, choice of inspection technique and inspection pattern and also taking into account the material effects on the response of NDT method used.

The main limitations for inspection methods have been described. Every inspection method has been described separately. The physical background for each method is shown and details are given for how they are used for copper weld inspection of a disposal canister. All the methods used are remotely controlled and are so-called mechanized inspection systems.

The used, studied and developed inspection methods are remote visual testing, eddy current testing using multi-element sensors, high energy linear accelerator radiography inspection and phased array ultrasonic inspection. The copper base material as well as the weld itself set high requirements for inspection technology. The main aspect is the grain size variation in base material and especially the anisotropy of the copper EB weld. The copper FS weld has a much lower grain size when compared to the EB weld.

The base system for remote visual inspection has been created and the preliminary inspection procedures are under study. The detectability of surface defects will be studied, taking into account inspection parameters such as inspection speed, focus distance and illumination. All of this study will be based on the possible defect types which are found in different welds and defects which can occur in these types of welds. Some welds has been measured and evaluated with this method.

In eddy current testing methodology the inspection of surface and near-surface defects has been developed. The more accurate study of detectability and defect sizing is under development. The eddy current inspection is divided into two testing techniques: low frequency (200 Hz) and high frequency (30 kHz) inspection. In the evaluation of the indications, the high frequency technique is used to detect and evaluate the surface size for indications, and the low frequency technique will be used to evaluate the indication depth up to 10 mm, when it is a question of a surface breaking defect. This method has been used for inspections of more than 40 real size welds, totalling 120 m of weld. The evaluation of all this data is ongoing.

In radiography testing the evaluation methodology was developed. A specific reference specimen was manufactured in order to evaluate in a short time defect detectability. The size of a volumetric defect is about 1 mm3. This was also found in POD estimation (Probability of detection). The detectability of thin planar defects in an X-ray beam is of course lower and in that case ultrasonic testing is used to overcome this limitation. The evaluation of data must be developed further in order to meet the requirements of the method for approval of inspection results. The detectability limit for diameter for deep penetrating defects exceeding through the whole weld is under study. This study will be ready in 2011.

90

Inspection techniques as well as evaluation methodology were developed for ultrasonic testing of copper welds. In the earlier phase 2 MHz and 5 MHz probes were used in inspections. To reach a compromise between detectability and sizing capability a 3.5 MHz probe was chosen. Ultrasonic testing is based on phased array technology in order to have a flexible inspection technique which can be modified changing the inspection requirement easily. The base technique is divided at the moment by using three angles of incidence for inspection: 0� and ±20�. Data from different inspection angles will be evaluated separately and combined to create a final ultrasonic inspection result. In inspections detected indications are sized according to inspection procedures. Several reference specimens have been created for inspection, mainly for the use of linear phased array technology. First a matrix phased array probe was applied for plate EB weld inspection, with good results, and this approach will be studied for EB weld inspection of real components. The evaluation is at the moment based on the raw data analyses; further study in defect sizing will be carried out using SAFT or a similar approach such as Sampling Phased Array Technique or Full Matrix Capture Technique. Other basic ultrasonic defect evaluation techniques will also be studied.

Each of the inspection methods will result in inspection records which will be the basis for combining the evaluated indications in order to receive the quality of the weld for total acceptance of the canister weld. The combination of the results provides tools for the better understanding of the total quality of the weld. It is also needed to understand the physical nature of each inspection method and their ability to indicate defects and size different types of defects in order to combine the results.

NDT reliability is important for each of these testing methods. Reliability also means the ability to make inspections repeatedly with results which vary as little as possible. For this different items such as inspection procedures are needed, which are the basis for performing inspections in a similar way each time, and the annual checking of equipment and sensors. This is to ascertain the equipment’s ability to carry out the inspections. One of the main items is to study defect detectability, which provides the limit of method used to detect defects. For all of these four inspection methods, the POD curve will be produced. Preliminary curves are already made for radiography and ultrasonic testing. The study to produce POD curves for eddy current and visual testing will be ready in 2011, in cooperation with BAM. To verify NDT indications to real defects a study of 50 indications will be carried out using metallography. Verification of indications will also be partly done continuously by updating the defect catalogue of the weld.

Qualifications of inspections will be based on a typical procedure in the nuclear industry. The basic qualifications of each inspector are based on EN 473 and the qualifications of the inspection target will be based on the recommendations given by ENIQ. Important for qualifications is to produce input information and the needed documents for qualifications such as inspection procedures and technical justifications and blind/open trials. This will be carried out according to Finnish practice for qualifications.

The basic principles of inspecting the closure weld have been produced and the methods will be developed to meet the requirement of the inspection task and requirements for acceptance criteria of the canister.

91

REFERENCES

Aalto H., 1998, EB-welding of the copper canister for the nuclear waste disposal. Final report of the development programme 1994-1997, Posiva Report 98-03, October, 43 p. Ahlberg, 2009, Color Megarad Type S Underwater Camera service manual. Ahlberg Electronics, 63 p. Akino, M., Mihara, T. and Yamanaka, K., 2003, Fatigue crack closure analysis using non-linear ultrasound, Review of progress in quantitative non-destructive evaluation 27th July – 1st August in Green Bay Wisconsin, vol 23, AIP, Melville, New York, pp. 1256 – 1263. Berens A, P., 1989, NDE reliability analysis, Metals Handbook, Vol. 7, 9th Ed. Nondestructive Evaluation and Quality Control, ASM International, OH. Bird, C., 2003, Ultrasonic Phased Array Inspection Technology for the Evaluation of Friction Stir Welds. 4th International Symposium on Friction Stir Welding, Park City, Utah, USA. Bowyer, W.H., 2000, Defects Which Might Occur in the Copper-Iron Canister Classified According to their Likely Effect on Canister Integrity, SKI Report 00:21, Swedish Nuclear Power Inspectorate, Sweden, 2000, 32 p., ISSN1104–1374. Buckley, J., 1994, An introduction to eddy current testing theory and technology, http://www.ndt.net/v04n08.htm, 7p. Bulavinov, A., 2005, Der getaktete Gruppenstrahler. Dissertation des Grades des Doktors der Ingenieurwissenschaften der Naturwissenschaftlich-Technischen Fakultät III Chemie, Pharmazie, Bio- und Werkstoffwissenschaften der Universität des Saarlandes, Saarbrücken am 11. August, 128p. Calmon, P., Chatillon, S., Fidahoussen A. and Iakovleva, E:, 2007 Model base processing for characterization of defects defected by phased array techniques, International Congress on Ultrasonics, Vienna, April 9.-13.2007, Session R17: NDT Modelling and simulation, 4 p. Cederqvist, L., 2006, FSW to Manufacture and Seal 5 cm thick Copper Canisters for Sweden’s Nuclear Waste. 6th

International Symposium on Friction Stir Welding, Saint-Sauveur, Canada. Cederqvist, L. and Andrews, R.E., 2003 A Weld That Lasts for 100,000 Years: FSW of Copper Canisters. 4th International Symposium on Friction Stir Welding. Park City, Utah, USA. Charlesworth, J.P. and Temple J. A., 2001, Engineering applications of ultrasonic time of flight diffraction, second edition, Research studies press Baldock Herdfordshire, England, 254 p.

92

Cumblidge S.E., Anderson M.T. and Doctor S.R., 2004 An Assessment of Visual Testing,.Pacific Northwest National Laboratory, USA. NUREG/CR-6860. U.S. Nuclear Regulatory Commission, Washington, DC, USA, 2004. 59 p Cumblidge, S.E. Anderson, M.T., Doctor, S.R. and Simonen, F.A., 2005, An Assessment of Remote Visual Methods to Detect Cracking in Reactor Components, Pacific Northwest National Laboratory, USA; A.J. Elliott, Oregon State University Corvallis, USA. Ding, J, Carter B., Lawless, K., Nunes, A., Russell, C:, Suites, M. and Schneider J., 2008 A decade of friction stir welding R&D at NASA's Marshall Space flight center and a glance into future. NASA. Doctor, S., Schuster, G. J., Reid, L. D. and Hall, T. E., 1995. Real-Time 3-D SAFT-UT System Evaluation and Validation. NUREG/CR-6344 PNNL-10571, 169 p + App. 78 p http://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/20080009619_2008009118.pdf, 2009, Retrieved 2009-04-14. 7 p. Draft prEN-ISO25239-1, 2009, Friction stir welding - Aluminium - Part 1: Vocabulary (ISO/DIS 25239-5:2009) CEN European committee for standardization, 9 p. Draft prEN ISO 25239-5, 2009, Friction stir welding - Aluminium - Part 5: Quality and inspection requirements (ISO/DIS 25239-5:2009), CEN European committee for standardization, 10 p. ENIQ, 1998, Technical Justification, Pre-trials, ENIQ Report 10 EUR 18114 EN, November 1998, 132 p. Erhard, A., Tscharntke, D., Montag, H.-J., Tessaro, U., Voskamp, R. and Schmidbauer, J.,2004 .SAFT and TOFD Used for Defect Sizing at Steam Generator Feed Water Nozzle. NDE in Relation to Structural Integrity for Nuclear and Pressurised Components, 6th – 8th December, London, UK, pp. 251 – 260. Fahlbruch, B., 2009, Integrating Human Factors in Safety and Reliability Approaches, 4th European-American Workshop on Reliability of NDE, 24th - 26th June, 21 p. Guerra, M., McClure, J.C., Murr, L.E., and Nunes, A.C. (2001) Metal Flow During Friction Stir Welding. In: Jata, K.V., Mahoney, M.W., Mishra, R.S., Semiatin, S.L., and Field, D.P. (eds.), Friction Stir Welding and Processing. Pennsylvania, USA: The Minerals, Metals & Materials Society. Pp. 25-34. ISBN 0-87339-502-6. http://www.everyspec.com on, 2005. Nondestructive inspection methods, basic theory, Technical manual, T.O. 33B-1-1, Navair 01-1A-16, retrieved 2010-01-30 T17, TM 1-1500-335-23.

93

Iordachescu, M., Iordachescu, D., Ocana, J. L., Vilaca, P. and Scutelnicu, E., 2009a, Characteristic flaws in aluminium alloys joints, Metallurgia International Vol XIV special issue n:o 2, 2009, pp. 135-138. Iordachescu, M., Iordachescu, D., Ocana, J. L., Vilaca, P. and Scutelnicu, E., 2009b, Contribution to the Classification of the Characteristic Flaws in Friction Stir Welded Aluminium Butt Joints, Welding in the World, Elsevier, ISSN 0043-2288, July 2009, 53(SPEC. ISS.), pp. 259-264. Jeskanen, H., Kauppinen, P. and Pitkänen, J., 2001, Applicability of ultrasonic techniques to inspection of copper canisters used for active waste disposal, Posiva working report 2001-08, March , 27 p. Jeskanen, H. and Kauppinen, P., 1997 Ultrasonic inspection of electron beam welded joints in copper, Posiva working report 97-34e, August 27p+8 appendix. Jäppinen T., 2009a, Process for Non-destructive Testing of the Disposal Canister Lid Welds, Posiva work report 2009-100 (to be published). Jäppinen T., 2009b, Replicas from surface Defects Detected in Disposal Canister Lid Welds, Posiva work report 2009-101, (to be published). Krautkrämer, J & Krautkämer H., 1983, Ultrasonic testing of materials, Springer Verlag Berlin, 667 p. Kenefick, S. and Henry, G. 2001, Ultrasonic Imaging Technique for Characterizing Surface Breaking Flaws. Proceedings of the Third International Conference on NDE in Relation for Nuclear and Pressurised Components, 14th -16th, Seville, Spain, pp. 552-560. Kröning, M. and Schmitz, V., 1997. Contributions of NDT, Risk Based Assessment of Industrial Components and Plants. 15th –17th December, Madras, Chennai, 31 p. Lahdenperä K., Lipponen A. and Pitkänen J., 2009, Eddy Current Inspection of Disposal Canister Lid Weld In Otaniemi June - June October 2008, Inspected EB-welds XK030-XK041 Posiva work report 2009-96 (to be published). Lévesque, D., Mandache, D., Dubourg, L. and Gougeon, P., 2008, Detectability of FSW defects with NDE methods, 7th

International Friction Stir Welding Symposium, Awaji Island, Japan, 20th - 22nd May 12 p. Linnert,G.E., 1994 Welding Metallurgy, Volume I, The American Welding Society, USA, 940 p., ISBN 0-87171-457-4. Lipponen A. and Pitkänen J., 2009a, Phased array ultrasonic inspection of disposal canister lid welds, Inspected welds XK010 - XK012; XK017-XK020, Posiva work report 2009-91 (to be published).

94

Lipponen A. and Pitkänen J., 2009b, Phased array ultrasonic and eddy current inspections of disposal canister lid weld, In Otaniemi 2009, Inspected EB-welds XK020-XK029; XK042-XK045, Posiva work report 2009-97 (to be published). Lipponen A. and Pitkänen J., 2009c, Phased array Ultrasonic Inspection of Disposal Canister Lid Welds, n Otaniemi June - June October 2008, Inspected EB-welds XK030-XK041, Posiva work report 2009-95 (to be published). Lipponen A. and Pitkänen J., 2009d, Phased Array Ultrasonic and Eddy Current Inspections of Disposal Canister Lid Weld, In Otaniemi 2009, Inspected EB-welds XK020-XK029; XK042-XK045 Posiva work report 2009-97 (to be published). Matthies, K., Neumann, E. and Mrasek, H., 1986, Lösungsansätze für die Ultraschallprüfung von Austeniten. Fortbildungsseminar "Ultraschallprüfung von austenistischen Werkstoffen", BAM, Berlin, 29 - 31. October, 36s. MIL-HDBK-1823, 2007, Department of defence handbook, Nondestructive evaluation system reliability assessment, Draft 28.2.2007. Müller, C., Elagin, M., Scharmach, M., Bellon, C., Jaenisch, G.-R., Bär, S., Redmer, B., Goebbels, J., Ewert, U., Zscherpel, U., Böhm, R., Brekow, G., Erhard, A., Heckel, T., Tessaro, U., Tscharntke D. and Ronneteg, U., 2006, Reliability of nondestructive testing (NDT) of the copper canister seal weld, SKB report R-06-08, 158 p. Müller, C., Bertovic, M., Gaal, M., Heidt, H., Pavlovic, M., Rosenthal M., Takahashi, K., Pitkänen J. and Ronneteg, U., 2009, Progress in Evaluating the Reliability of NDE Systems Paradigm Shift, European-American Workshop on Reliability of NDE, June 24-26, 2009, in Berlin. Müller, C., Pavlovic, M., Takahashi, K., Ewert, U., Scharmach, M., Brekow, G., Boehm, R., Ronneteg U., Emilsson G. and Pitkänen J., 2007, POD Evaluation of NDE Techniques for Canister-Components for Risk Assessment of Nuclear Waste Encapsulation, 6th international conference on NDE in relation to structural integrity for nuclear and pressurised components, 8th - 10th, October, 2007 Budapest Hungary. Müller, W., Schmitz, V. and Schäfer, G., 1985. The Synthetic Aperture Focussing Technique and its Application in NDE - Principles – Signal Processing and Measurement. Philadelphia, USA September 25 p. Müller, W. and Bruche, D., 1999. Erkannbarkeit von Risspitzensignalen bei Vorwärts- und Rückstreuung – ein Vergleich. IZFP-FHG Bericht Nr. 990112-TW, March, 140 p. Ollonqvist P., 2007, Micro-structural characterization and mechanical properties of electron beam welded thick phosphorous micro-alloyed oxygen free copper (Cu-OFP), Master of science thesis, Department of Mechanical Engineering, Helsinki University of Technology, 31st of March, 88 p + 8 Appendices. Olympus, 2008 Advances in phased array technology applications, 467 p.

95

Osetrov, A. V., 1995, Complex Flaws Reconstruction by Means of Second Order Synthetic Aperture Focusing Technique. Proceedings International Conference on Computer Methods and Inverse Problems in Nondestructive Testing and Diagnostics, CM NDT-95, 21st - 24th November, Minsk, Belarus, pp. 116 - 119. Paussu, R., and Pitkänen, J., 2009, Reference Specimens of Copper Components and EB Weld for Qualification and Performance Demonstration Purposes, 7th international conference on NDE in relation to structural integrity for nuclear and pressurized components, 12th- 14th. May, 2009 Yokohama Japan. Pavlovic M., Takahashi K. Müller, Boehm, R. and Ronneteg, U., 2008, NDT reliability - Final report Reliability in non-destructive testing (NDT) of the canister components, Svensk kärnebränslehantering AB, SKB Report R-08-129, December, 56 p. Pavlovic, M, Müller, C. and Pitkänen, J., 2009a, The total reliability of different NDT methods for an inspection of the electronic beam welding, ASNT Research Symposium, St. Louis MO USA, 6th - 20th March 2009, 32 p. Pavlovic M., Müller C., Takahashi K., Pitkänen J. and Ronneteg, U., 2009b, Multi-Parameter Influence on the Response of the Flaw to the Phased Array Ultrasonic NDT System, The Volume POD, European-American Workshop on Reliability of NDE, June 24th - 26th, 2009, Berlin. Pavlovic, M., Takahashi, K., Müller, C. and Pitkänen, J., 2009c The Total Reliability of NDT Methods for Inspection of the EB-Weld of the Copper Canister used for a Long-term Storage of Spent Nuclear Fuel, 7th international conference on NDE in relation to structural integrity for nuclear and pressurized components, 12th- 14th. May, 2009 Yokohama Japan. Petrov, P., Georgiev, C. and Petrov, G., 1998, Experimental Investigation of Weld Pool Formation in Electron Beam Welding, Vacuum, 51 (1998), pp. 339-343, ISSN 0042-207X Pitkänen, J., 1991, Ultrasonic probe design and manufacturing, Licentiate’s Thesis, Helsinki university of Technology, Laboratory of Metal working and heat treatment of Metals, 82 p. Pitkänen, J., 2006 SAFT - is it a tool for improved sizing in ultrasonic testing, ECNDT Berlin 25.-29. 2006 - Poster 211 DGzfP, http://www.ndt.net/article/ecndt2006/doc/P211.pdf 13 p. Pitkänen, J., Laukkanen, A., Kemppainen, M. and Virkkunen, I., 2007a, Effect of stress on ultrasonic response in detection and sizing of cracks, MP Materials Testing 49(2007)6 pp. 299 - 309. Pitkänen, J., Salonen, J., Sandlin, S. and Ronneteg, U., 2007b, Defect Detectability in EB-welded Copper Disposal Canister with 9 MeV Accelerator, 6th International

96

Conference on NDE in Relation to Structural Integrity for Nuclear and Pressurized Components, Budapest 12th -14th October 2007, 14 p. Pitkänen J., Arnold W. and Hirsekorn, S.,2007c The Effect of Grain Size on the Defect Detectability in Copper Components in Ultrasonic Testing, 6th international conference on NDE in relation to structural integrity for nuclear and pressurised components, 8th - 10th October, 2007 Budapest Hungary. Pitkänen J., Lipponen A., Lahdenperä K. and Kiselmann I., 2009a The Eddy Current Inspection for Detection of Surface Defects in Copper Components and Copper EB-Weld, 7th international conference on NDE in relation to structural integrity for nuclear and pressurized components, 12th - 14th May, 2009 Yokohama Japan. Pitkänen J., Salonen T., Bertovic M., Müller C. and Pavlovic M., 2009b, NDT Reliability in Risk Minimization during Manufacturing and Welding of Spent Nuclear Fuel Disposal Components – A Realistic Tool for Reliable Inspections, Germany, European-American Workshop on Reliability of NDE, June 24-26, 2009, in Berlin. Poidevin, C., Mahaut S., Dubois P., 2007, Roy O and Le Ber, L. 2007, Advance Matrix phased arrays for inspection, 6th international conference on NDE in relation to structural integrity for nuclear and pressurised components, 8th - 10th October, 2007 Budapest Hungary. RD-Tech corp., 2004, Introduction to phased array ultrasonic technology applications, RD-tech guideline Quebec Canada 356 p. Ronneteg, U, Cederqvist, L., Ryden H., Öberg, T. and Müller, C., 2006, Reliability in sealing of canister for spent nuclear fuel, SKB Report R-06-26, 121 p. Sandlin S., 2009a, High Energy Radiography for Inspection of the Lid Weld in Disposal Canisters, Posiva work report 2009-82 (to be published). Sandlin S., 2009b, Defect Detectability in the Disposal Canister Lid -weld using the 9 Mev Linear Accelerator, Posiva work report 2009-84 (to be published). Sandlin S., 2009c, X-ray Inspection Setups for Disposal Canister Lid Weld, Posiva work report 2009-98 (to be published). Santos, T., Vilaça, P. and Quintino, L:, 2008, Developments in NDT for detecting imperfections in friction stir welds in aluminium alloys, Welding in the World, Vol. 52, No 9/10, 2008, pp. 30-37. Sarkimo M., and Pitkänen J., 2009a, Background information for NDT- qualification for Finnish disposal canisters of spent fuel, Posiva Report 2008-07 (to be published). Sarkimo M. and Pitkänen J., 2009b, Phased array Ultrasonic Inspection of Disposal Canister Lid Welds in Otaniemi 16th - 19th of June 2008, Posiva work report 2009-93 (to be published).

97

Savolainen, K., 2004, Friction Stir Weldability of Copper Alloys. M.Sc. Thesis. Helsinki University of Technology, Laboratory of Engineering Materials. 115 p + 3 Appendices. Savolainen K, 2008, Friction stir welding and processing of copper Licentiate’s Thesis, Helsinki university of Technology, Laboratory of Engineering Materials, 82 p. Schultz, H., 1993, Electron Beam Welding, Abington Publishing, Great Britain, 240 p., ISBN 1-85573-034-0. Schmitz, V. and Chakhlov, 2001, Fehleranalyse durch Ultraschallrekonstruktion 3D-CAD-SAFT-Handbuch v. 1.8. 92 p. Schmitz, V., Müller, W. and Schäfer, G., 1988. Ultraschallprüfung mit synthetischen Aperture-Verfahren (SAFT). Symposium für Werkstoffprüfung, Greifswald, Germany, Juni, 17 p. SFS-EN 1330-10, 2003 Non-destructive testing. Terminology. Part 10: Terms used in visual testing. Finnish Standards Association SFS, 2+36 p. SFS-EN 13068-1, 2000, Standard - Non-destructive testing - Radioscopic testing, Part 1: Quantitative measurement of imaging properties, confirmed 2000-02-14, 22 p. SFS-EN-ISO 13919-1, 1996 Welding, Electron and laser beam welded joints. Guidance on quality level for imperfections. Part 2: Steel, Finnish Standards Association -SFS, 1p + 2 p + Appendix 13 p. SFS-EN-ISO 13919-2, 2001, Welding, Electron and laser beam welded joints. Guidance on quality level for imperfections. Part 2: Aluminium and its weldable alloys Finnish Standards Association -SFS, 1 p + Appendix 14 p. SFS-EN 1435, 1998, Standard – Non-destructive testing - Non-destructive examination of welds. Radiographic examination of welded joints, confirmed 1998-01-05 40 p. SFS-EN 14784-2, 2005, Standard - Non-destructive testing - Industrial computed radiography with storage phosphor imaging plates. Part 2: General principles for testing of metallic materials using X-rays and gamma rays, confirmed 2005-10-17, 17 p. SFS-EN 462-1, 1994, Standard - Non-destructive testing – Image quality of radiographs - Part 1: Image quality indicators (wire type) – Determination of image quality value, confirmed 1994-05- 30, 6 p. SFS-EN 462-4, 1995, Standard - Non-destructive testing – Image quality of radiographs - Part 4: Experimental evaluation of image quality values and image quality tables, confirmed 1995-02-13, 5 p. SFS-EN 462-5, 1996, Standard - Non-destructive testing – Image quality of radiographs - Part 5: Image quality indicators (duplex wire type), determination of image unsharpness value, confirmed 1996-05-06, 8 p.

98

SFS-EN 462-2, 1994, Standard - Non-destructive testing – Image quality of radiographs - Part 2: Image quality indicators (step/hole type) – Determination of image quality value, confirmed 1994- 06-20, 6 p. SFS-EN 462-3, 1996, Standard - Non-destructive testing - Image quality of radiographs - Part 3: Image quality classes for ferrous metals, confirmed 1996-12-02, 13 p. SFS-EN-ISO 6520-1, 2007, Welding and allied processes, Classification of geometrical imperfections in metallic materials, part 1: fusion welding Finnish Standards Association -SFS, 1 p + Appendix 27 p. SKB, 2005, Friktionssvetsning av kopparkapslar rapport 1, Friction Stir Welding of Copper Capsules, Report 1, SKB Report R-05-73. SKB, 2006, Kapsel för använt kärnbränsle, Oforstörande provning av svetsar, R-06-06, September, 69 p. Shlivinski, A., Langenberg, K. J., Marklein, R., Schmitz, V., Müller, W. and Mletzko, U., 2004. Modelling of Elastic Wave Propagation and Imaging of Defects in Dissimilar Welds. NDE in Relation to Structural Integrity for Nuclear and Pressurised Components, 6th - 8th December, London UK, pp. 900 – 909. Suikki, M. and Wendelin T., 2009, Inspection station design for canister sealing weld, Posiva Working report 2009-71, December, 34 p. Wu, P. & Stepinski, T:, 1998 Inspection of copper canisters for spent nuclear fuel by means of ultrasonic array system -Modelling , defect detection and grain noise estimation, SKB Technical report, TR-99-12, July 1998, 109 p. Sun, Z. and Karppi, R., 1996, The Application of Electron Beam Welding for the Joining of Dissimilar Metals, Journal of Materials Processing Technology, 59(1996), pp. 257-267, ISSN 0924-0136. Takahashi K., Müller C., Pavlovic M., Pitkänen J. and Ronneteg U., 2009, Probability of Detection from C-Scan Data for Reliability Analysis of Ultrasonic Non-Destructive Testing, European-American Workshop on Reliability of NDE, June 24-26, 2009, in Berlin. Uchanin, V., Mook G. and T. Stepinski, 2003, The Investigation of Deep Penetrating High Resolution EC Probes For Subsurface Flaw Detection and Sizing, NDT.net - February 2003, Vol. 8 No.2. Unosson, M., 2009, Intryck i koppar, SKBdoc 1205273 rev 2.0, 26 p. Zagrai, A., Donskoy, D. and Lottiaux, J.-L., 2003, N-SCAN: New Vibro –modulation system for crack detection, monitoring and characterisation, Review of progress in quantitative non-destructive evaluation 27th July – 1st August in Green Bay Wisconsin, Vol. 23, AIP, Melville, New York, pp. 1414 – 1421. Yamanaka, K., Mihara, T. and Tsuji, T., 2004, Evaluation of closed cracks by analysis of subharmonic ultrasound, Insight, 46(2004)11, pp. 666 – 670.

LIST OF REPORTS

POSIVA-REPORTS 2010

_______________________________________________________________________________________

POSIVA 2010-01 Models and Data Report 2010 Barbara Pastina, Saanio & Riekkola Oy Pirjo Hellä, Pöyry Finland Oy ISBN 978-951-652-172-8 March 2010 POSIVA 2010-02 Interim Summary Report of the Safety Case 2009 Posiva Oy ISBN 978-951-652-173-5 March 2010 POSIVA 2010-03 Biosphere Assessment Report 2009 Hjerpe Thomas, Saanio & Riekkola Oy Ikonen Ari T. K., Posiva Oy Broed Robert, Facilia AB ISBN 978-951-652-174-2 March 2010 POSIVA 2010-04 Inspection of Bottom and Lid Welds for Disposal Canisters Jorma Pitkänen, Posiva Oy ISBN 978-951-652-175-9 September 2010

Inspection of Bottom and Lid Welds for Disposal Canisters - · PDF file ·...

Documents

Transcript of Inspection of Bottom and Lid Welds for Disposal Canisters - · PDF file ·...