Nuclear waste disposal task calls on electron beam welding
TWI Bulletin, July/August 1994
Allan Sanderson is the Technology Manager of the Electron Beam Group which is concerned with design, development and application of electron beam equipment for use in-vacuum, at reduced pressure and at atmospheric pressure. He joined TWI in 1966 and in parallel with his activities there he obtained his doctorate in the generation and control of electron beams. It was this pioneer work which led to the breakthrough in high power electron beam welding in the early 1970s, resulting in single pass welding of steel in thicknesses up to 300mm. Since that time, he has been responsible for further innovative work on gun design, power sources, beam monitoring, magnetic trap devices, real-time seam tracking detectors and more recently development of resonant RF circuits for electron beam gun supplies. He is the author of numerous papers and reports on electron beam welding and currently project leads the electron beam welding work being carried out on thick section copper for the Swedish Nuclear Fuel and Waste Management Company's fuel element encapsulation programme.
Encouraging results have been achieved using EB for sealing copper nuclear waste canisters. Allan Sanderson reports.
Safe long term disposal of high level nuclear waste has been the subject of much discussion and debate in all countries operating nuclear power plants. There are many possible means of disposal, but each route raises numerous short, medium and long term technological, safety and cost issues. Of the wide spectrum of possibilities, burial in stable repositories deep in the earth's crust currently appears to be the most safe and viable means of disposal.
The suitability of burial sites is dependent both on geological factors and long term corrosion considerations, that is whether the environment is oxidising or reducing, and on the canister material intended. In Sweden it has been discovered that the ground water in deep bore holes in granite contains extremely low levels of oxygen. Therefore a canister corrosion barrier material such as copper would not be expected to corrode significantly for periods over 100 000 years. This would provide ample time for the bulk of the materials contained in the canisters to decay to safe radioactivity levels.
The philosophy as regards the disposal of spent nuclear waste in Sweden firmly favours direct disposal with no reprocessing. After a suitable cooling off period in special water storage tanks, the spent fuel still contained in its original fuel elements will be dried before insertion in the disposal canisters.
In countries where reprocessing of spent nuclear fuel is still considered necessary, this further complicates the disposal issues and requires a quite different approach. Also in other countries the geology, stability and chemical conditions in potential burial sites can pose other problems.
The principle of using copper as the main corrosion barrier for long term storage of high level nuclear waste is now well established in Sweden. The Swedish Nuclear Fuel & Waste Management Company (SKB) first approached TWI in 1981 with a request to recommend the best means of sealing such canisters. Originally it was thought that a copper canister wall thickness of approximately 200mm would be required for corrosion resistance and strength.
Of the processes considered ( i.e. electron beam welding, diffusion bonding and friction welding) electron beam welding potentially offered the greatest probability of success. Certainly at this time the proposed wall thickness was well beyond the state-of-the-art capabilities of diffusion bonding and friction welding. However, even in 1983 it proved possible to demonstrate successfully electron beam welding of a mock-up lid assembly involving an effective wall thickness of 100mm with an outside diameter of 600mm.
Much research was conducted in the 1980s and early 1990s on in-vacuum electron beam welding (that is where the chamber vacuum is maintained at a pressure of better than 5 x 10 -3mbar) in attempts to achieve freedom from internal weld defects. Also, over the period 1986 to 1992 work was conducted on non-vacuum electron beam welding with a view to sealing the canisters at atmospheric pressure.
More recently, effort has concentrated on use of electron beams at 'reduced pressure', that is over the pressure range 5 x 10 -2mbar to 1000mbar, and this has indicated certain advantages, especially in the control of weld defects and freedom from outgassing problems associated with the double skinned canister. Indeed in the last year it has proved possible for the first time to produce virtually defect free welds, even in the slope-down region of circumferential lid welds where the weld penetration by necessity passes progressively from full depth to zero depth.
As well as the work on development of electron beam welding sealing procedures for canisters, TWI has also been extensively involved in studies of the means of forming and fabricating canisters by electron beam welding. This article presents a summary of the work conducted on electron beam welding of copper canister materials, together with the associated non-destructive testing requirements and techniques.
Materials and design
In the early 1980s a comparison was made of the weldability of phosphorus deoxidised copper (PDO) and oxygen free high conductivity copper (OFHC). The OFHC copper produced superior welds with greater freedom from microporosity compared with the PDO material. However, SKB's subsequent researches showed that the creep strength of the OFHC was inadequate although the corrosion resistance studies predicted a much greater lifespan than had been first supposed.
Similarly, although it proved possible to achieve single pass welds with over 125mm penetration in copper using electron beam welding, freedom from internal defects and particularly defects in the slope-down region of the weld overlap proved to be problematical for a long time.
The need to achieve adequate canister strength to resist the hydrostatic ground water pressure and encouraging copper corrosion resistance studies later led to the concept of a steel inner canister with a wall thickness of 50mm surrounded by a copper canister of 50mm wall thickness, see Fig.1. Also an alternative copper grade was selected to improve creep strength.
Fig. 1. Basic canister design
Welding
Conventionally, electron beams are generated in a vacuum and applied in a vacuum. The vacuum avoids high voltage break-down and emission source deterioration in the electron gun, and also prevents beam spreading because of electron collision with gas molecules. As regards application of high power electron beams, this can be done in high vacuum (<5 x 10 -3mbar), reduced pressure ( e.g. 10 -1 to 100mbar) or at atmospheric pressure (~1000mbar). Advanced high power (l00-l50kW) equipment now exists at TWI covering the entire pressure range and this has been used to study effects of chamber pressure on fusion zone profile, weld beam shape, defect occurrence and penetration depth, see Fig.2.
Fig. 2. Electron beam equipment covers high vacuum to atmospheric pressure operation: Fig. 2a) 100kW for operation in-vacuum (10 -5 to 5 x 10 -3 mbar)
Fig. 2b) 100-150kW reduced pressure (10 -1 to 500mbar)
Fig. 2c) 100-150kW for non-vacuum (~1000mbar)
The mechanism by which an electron beam penetrates solid material is well established; the electron beam is not only sufficiently intense to melt material but also so intense that material is vaporised from the surface. The jetting action created by metal volatilisation creates a depression in the weld pool enabling the beam to penetrate more deeply. Once established, the vapour jetting action allows the beam to establish a stable keyhole which is held open by vapour pressure in equilibrium with surface tension and gravitational forces which try to close the capillary, see Fig.3.
Fig. 3. Electron beam weld pool keyhole and welding mechanism
The stability of the keyhole is very much related to the temperature distribution produced by the electron beam, and whether or not pore type defects occur is in turn very dependent on the fusion zone shape dictated by the keyhole shape. Electron beam welding proceeds by translation of the liquid lined vapour filled cavity along the joint plane and the weld is formed by flow of liquid metal around the beam in the wake of the beam.
Electron beam welding offers many outstanding advantages for fabrication and sealing of copper canisters. These are:
- Rapid process, equally applicable to both circumferential and longitudinal seam welding;
- Produces high integrity welds;
- Reproducible process readily controlled and monitored for quality assurance purposes;
- Protective atmosphere (especially for in-vacuum and reduced pressure welding) ensures freedom from oxygen pick-up;
- Welds can be inspected by ultrasonic or X-radiographic techniques.
Electron beams can, in principle, be applied to the workpiece with various combinations of approach angle and workpiece orientation. But for a given heat input per millimetre of joint, to achieve the deepest penetration, electron beams are best applied with the beam axis horizontal and the workpiece standing vertically (HV). This, however, can lead to some difficulties in supporting the liquid pool and often requires use of a 'fronting bar' to avoid metal loss, see Fig.4a. In addition, if the weld is fully penetrating it is normal to use backing material to avoid undercut problems; this material is usually removed after welding.
For so called flat position (FP) welding, where the beam is applied vertically down and is required to penetrate the full component thickness, use of backing material is absolutely essential, for thick section copper, to avoid metal loss, see Fig.4b. Control of top bead profile, in this case, in principle, is easier and generally speaking it is reasonable to expect freedom from undercut, particularly if a low power smoothing pass is applied to consolidate the main weld bead profile.
Fig. 4. Use of 'fronting' and backing bars to improve weld bead profiles for:
a) HV partially penetrating weld
b) Fully penetrating weld
In the case of blind welds, either made in the HV or FP mode, the greatest difficulty is formation of root defects. For very intense pencil beams, typical of high vacuum operation, no difficulty is experienced in achieving very deep (HV mode) penetration ( e.g. 300mm in steel, 450mm in light alloys and 150mm in copper). The narrowness of such penetrations, however, can lead to all manner of defect problems, particularly in light alloys. Fortunately, with copper the only likely defect is porosity or cavity formation. Especially in copper, its high conductivity combined with lack of freezing range tends to promote locked in cavities. This is particularly troublesome in blind (non-penetrating) welds where the narrowness of the fusion zone tip typically causes premature freezing of the small liquid volume leading to spike or root defect formation, see Fig.5a & b.
Fig. 5a) & b) Typical root defect in non-penetrating weld in copper caused by sharpness of fusion zone tip
For the present lid design, see Fig.1, HV rather than FP welding has been chosen, primarily to avoid the possibility of the beam penetrating into the inner nuclear waste container, to reduce weld stresses to prevent solidification cracking and for ease of achieving the necessary penetration depth. Also by shaping the canister lid in the form of a stopper, the smaller diameter material provides backing material to ensure freedom from undercut at the rear of the weld, but it also adds mass, avoiding excessive temperature rise during welding.
The welding procedure will involve tack welding followed by a conventional beam current slope-up, overlap and slope-down sequence. During the main body of the weld, when the excess beam is penetrating into the neck of the 'stopper', root defect formation is less important, but following overlap whilst the beam current is decaying to draw the penetration finger towards the surface, the possibility of root cavities exists. For this reason, much emphasis has been placed on elimination of root defects.
In the case of non-vacuum electron beam welding, the maximum penetration achieved to date has been some 40mm, although weld shape was encouraging in that the tip of the fusion zone was well rounded, avoiding premature freezing of the tip finger. In the case of reduced pressure welding it has proved possible not only to achieve adequate penetration depths but also a well radiused fusion zone tip avoiding root defect formation.
Root defects and non-destructive testing
Throughout the development work both ultrasonic testing (UT) and X-radiographic techniques have been used extensively. For development of welding conditions, specimens of up to 300 x 150 x 200mm were used to provide reasonable thermal mass, minimising variations in the weld depth and shape because of temperature rise as the weld progressed, see
Fig.6a.
Fig. 6. Configuration of test block for development of electron beam welding procedures:
a) Workpiece
b)Weld slice for X-radiography (dimensions mm)
These were subsequently fully inspected using 0° P-scan UT and time-of-flight D-scan (TOFD) followed by X-ray inspection through the plane of the weld. To achieve maximum contrast on the X-ray film the weld was sliced out in a 20mm thick slab, see Fig.6b. This allowed defects of less than 0.5mm diameter to be detected. Figure 7 shows a typical radiograph facsimile for a defective weld in which both root cavities and an internal defect have occurred from using a very narrow fusion zone shape.
Fig. 7. Radiograph facsimile of electron beam welded joint showing severe root defects in current slope down region of weld caused by sharpness of fusion zone tip
Recently, using a combination of reduced electron beam welding and special equipment it has proved possible to shape the fusion zone to achieve a much wider near parallel form, with a well radiused tip, see Fig.8. In the vast majority of tests conducted to date this type of fusion zone has shown complete freedom from defects as, for example, in the radiograph contact print in Fig.9. Similarly, 0° P-scan and TOFD inspections ( Fig.10) have also confirmed the absence of defects.
Fig. 8. Fusion zone shape providing a defect free weld
Fig. 9. Radiograph contact print of electron beam weld showing freedom from defects
Fig. 10. Ultrasonic TOFD inspection showing freedom from defects
Similarly, detailed metallographic examination of welds in low phosphorus deoxidised copper shows no significant defects. The presence of micropores of typically less than 0.3mm diameter, see Fig.11 is not thought to be detrimental to the long term corrosion resistance of the canister.
Fig. 11. Microporosity in electron beam fusion zone
Because of the high levels of gamma radiation emitted by the spent nuclear fuel, conventional X-radiation inspection techniques cannot be applied, therefore much work has been carried out on the development of P-scan and TOFD UT to inspect the lid weld remotely. During this research it has been shown that the sensitivity of UT inspection is very much dependent on the grain size of the copper. Although some improvements can be achieved for coarse grained copper by using lower frequencies, the improvement is marginal and generally fine grain copper is required, particularly for the canister lid material, to avoid excessive ultrasound attenuation. This imposes constraints on the material type which can be used and the fabrication route. Copper canister production processes which allow the material to be held at high temperature for long periods are not acceptable because of the excessive grain growth which results.
Using the special reduced pressure electron beam welding equipment, it is now possible consistently to control root defect formation. Work is currently being undertaken to extend the linear welding procedures to curved full diameter cylinders to verify the integrity of circumferential weld slope-up, slope-down and weld overlap regions. This will initially be conducted on 1m long cylinders in the equipment shown in Fig.2b and later be applied to a 2.4m and possibly 5m long canister in an extended version of the same chamber.
Canister fabrication
Ideally a canister should contain as few welds as possible. Provided that material can be obtained in large enough pieces, it can be rolled or pressed with the required accuracy and high integrity electron beam welding can be performed to join the component parts together. TWI has been researching material supply, forming and welding possibilities on behalf of SKB. This has led to a range of fabrication schemes, some of which are shown in
Fig.12.
Fig. 12. Possible canister fabrication routes:
a) Three longitudinal curved plates requiring three electron beam welds
b) Two longitudinal curved plates requiring two electron beam welds
c) Three short length cylinders requiring two circumferential welds
The approach which is currently most strongly favoured is as depicted in Fig.13.
Fig. 13. Proposed electron beam welding equipment for longitudinal seam welding
This would involve two or possibly one longitudinal seam weld(s) carried out using an internal mobile high power electron beam gun mounted on a linear traverse. It is proposed that welding be conducted in the horizontal-vertical mode to simplify jigging requirements and to facilitate assembly; this would require use of 'fronting' and possibly backing material. Real-time seam tracking would be applied to ensure accurate beam/joint alignment. For circumferential butt welds, as in the case of lid and bottom welding, the HV welding mode is preferred.
In conclusion
Electron beam welding is now well established as a viable means of welding thick section copper for encapsulation of high level nuclear waste. Electron beam welding offers both a high integrity fabrication approach and the means of effecting the final canister seal. Much research has been conducted on applying electron beam welding to copper over the last 12 years and this has provided very encouraging results, indicating that safe encapsulation can indeed be assured. In particular, the problems which earlier led to defects in the root of circumferential welds with the possible consequence of lines of weakness in the vessel wall have now been largely overcome by development of special equipment. Similarly, with recent advances in ultrasonic techniques it is expected that the integrity of the electron beam welds can be adequately verified.
Acknowledgements
This work was sponsored by SKB, and the author and the Electron Beam Group at TWI are grateful for SKB's permission to publish this work.
