Welds Safeguard Plutonium-Bearing Containers
Welds safeguard plutonium-bearing containers
A sound closure welding process was developed for 3013 containers to prevent the release of hazardous materials to the environment.
A key element in the Department of Energy (DOE) strategy for the stabilization, packaging, and storage of plutonium-bearing materials involves closure welding of DOE-STD-3013 outer containers. The 3013 container provides the primary barrier and pressure boundary for preventing release of plutonium-bearing materials to the environment. The final closure weld of the 3013 container must be leaktight, structurally sound, and meet DOE-STD 3013-specified criteria.
In February 2001, the Savannah River Technology Center (SRTC) supplied a welding system to the Hanford Plutonium Finishing Plant (PFP), located near Richland, Wash., for the closure of 3013 containers. Efforts to supply this system included development, qualification, and demonstration of an automatic gas tungsten arc welding (GTAW) process for making the closure weld. GTAW was chosen for its demonstrated history in critical applications, ease of remote operation, and ability to make high-integrity welds.
Figure 1 - Sketch of a 3013 outer container (right) and photograph of the container after welding
Figure 2 - The modified orbital weld head.
Figure 3 - ASME Section VIII, Division 1, Figure UW-13.2 (d), Weld Cross Section Geometry requirements.
The 3013 container, made of Type 316/316L stainless steel and measuring almost 5 inches in diameter and approximately 10 inches in length, is closure-welded at the lid/container interface – Fig. 1. The corner joint is formed by pressing an interference-fit lid into the 3013 container, creating a square groove weld preparation. The closure is made autogenously (without addition of filler) using a modified AMI© 9 Series orbital weld head with a Model 227 power supply. The modified weld head (Fig. 2) includes an integral chill block.
An encoder and shunt were added for weld travel speed and arc current measurements, respectively. In addition, the weld head rotor was adapted to receive a specially designed snap-in tungsten electrode for ease of replacement.
This article focuses on the development, qualification, and demonstration of the welding process for the closure welding of Hanford PFP 3013 containers.
Weld Process Parameter Development
Weld Performance Objectives/Acceptance Criteria
Process development was driven by weld acceptance criteria specified in the DOE standard 3013-99 (Ref. 1) and customer-defined requirements for waste acceptance. These documents specify requirements for weld leak tightness, soundness, strength, and bead geometry. Good weldability and weld quality for 3013 container/lid materials using the GTAW process has been demonstrated many times over the years. The primary challenge posed was developing the process to meet specific ASME Section VIII, as-welded dimensional criteria – Fig. 3.
GTA weld beads are typically characterized by relatively low bead depth-to-width values (aspect ratio). Because of the relative thinness of the lid flange, achieving full penetration without consuming the top corner was considered the primary challenge. Therefore, development efforts have centered on weld bead aspect ratio.
Target Welding Parameter Development
Welding-arc energy (power density) can have a significant affect on bead penetration and shape. As a result, readily controllable variables having a direct influence on arc energy were identified for evaluation. Various tungsten alloy types, tip geometries, and shielding gasses were tested to characterize their relative affect on bead geometry. Initial welding parameters based on standard GTAW practices, shown in Table 1, were utilized for this welding. This weld schedule employed a pulsed-step welding current, with travel occurring on the background pulse.
The Table 2 matrix lists the variables and changes studied, their affect on weld bead geometry, and the final values (parameters) selected.
Welds were made on test pipes and evaluated by metallography: Fig. 4 is typical of such welds. The selected parameters were further evaluated and subsequently modified through a series of test welds using mockups of actual 3013 containers. During the course of this effort, the need for additional weld bead control was recognized and a chill block was added to the process. Chill block design considered material thermal properties, chill block mass, surface finish, and the force applied at the chill block/lid interface. Figure 5 provides a sketch of the chill block and details the pertinent design and operational criteria. Having added the chill block to the process, a set of target welding parameters was established that produced the desired weld bead shape – Fig. 6.
With the target welding parameters identified, the following series of test welds, designed to identify the overall process and operational windows, was conducted.
Figure 4 - Metallography on welded test pipes.
Figure 5 - Chill block sketch and details.
Figure 6 - A weld cross section.
Statistical Evaluation of Target Welding Parameters
A series of test welds was conducted to explore the impact of variation in weld current, arc length, and travel speed. In the first set of these welds, each target parameter was varied over a range judged likely to bound typical process variation based on process tolerances. Specifically, the primary weld current was varied ±5 A, the arc length was varied ±0.007 inches, and the travel speed was varied ±0.02 rpm. The eight possible combinations of these extremes were tested along with three replicate conditions and one at the baseline condition. All of the welds exhibited full penetration and acceptable aspect ratio.
The second set of welds was made after preliminary evaluation showed all welds in the first set were acceptable. The parametric ranges were expanded to identify potential margins that existed with respect to the operating range. In these expanded ranges, the primary current was varied by up to +25/-25 A, the arc gap ±0.010 inches and the travel speed by up to 0.04 rpm. These ranges were deisgned so the change in energy input from the increased primary current approximately offset thast of the increased travel speed. Alternatively, the impact on total energy input from a combination of increased primary cuirrent and slower travel speed is additive. In this set of test welds, ten combinations were tested with two replicates. All of the welds exhibited full penetration and acceptable aspect ratio.
The final set of test welds repeated eight parameter combinations from the second set. Eight welded containers were burst tested. In each case, maximum pressure was reached as the container side-wall began to balloon out. The test was stopped prior to rupture. All containers sustained a maximum internal pressure of 4340 to 4510 lb/in.2.
Figure 7 - Contour plot of penetration in excess of defined full penetration as a function of arc length (in.) and primary current (A).
The three sets of test welds identified two parameter combinations with a potential to produce upset conditions. A combination of high current and large arc length consumed the top edge of the lid in some cases, leading to loss of dimensional control. A subsequent change in the chill block material from UNS C18200 chrome-copper to UNS C15715 aluminum-oxide dispersion-strengthened copper (improved thermal properties) eliminated this problem. A combination of high current and small arc length led to the electrode touching the weld. This particular combination was considered when establishing the final operating parameters by selecting values that provided ample margin to avoid such conditions.
The data from these welds were evaluated using various statistical models/tools to optimize the target parameters within the established ranges. Figure 7 illustrates the ability of one such tool to help identify optimum welding parameter settings. This particular contour plot, developed from the weld data, provides a graphical representation of weld bead penetration as a function of primary arc current and arc length.
Statistical analysis of the weld data led to a shift in the original target parameter settings to what was then identified as the production welding parameters (Table 3).
In addition to development of the welding parameters noted above, other process conditions affecting weld quality were evaluated and are summarized as follows:
Container Internal Pressure and Venting. The lid, when pressed into the container, created a seal by virtue of a designed interference fit at the lid-/plug container-wall interface. Immediately prior to closure welding, the loaded 3013 container is backfilled with helium to facilitate postweld, sensitive leak testing. To evaluate the container’s ability to adequately vent or relieve expanding internal gases during the course of welding, several test welds were performed with can/lid combinations at the maximum design interference fit. In addition, the container’s internal-pressure behavior during welding was characterized by use of a pressure transducer. It was observed pressure built to a value of nearly 0.3 lb/in.2 then vented through the unwelded portion of the weld joint to an equilibrium pressure established between the internal backfill and external shielding gases. This pressure fluctuation was repeated several times throughout the weld. It was concluded from this evaluation container internal-pressure behavior did not adversely affect the quality of the closure weld.
Chill Block Development and Qualification. As noted above, a chill block was added to the process to protect the lid’s top corner from being consumed by the weld. The chill block, placed in direct contact with the top surface of the lid, limited bead width by removing excess welding heat.
Finite element analysis was performed to identify a pressure or compressive stress at the chill block/lid interface sufficient to ensure continuous contact during welding and, hence, maintain good thermal transfer. A minimum contact stress of 43 lb/in.2 was specified along with a maximum surface roughness at the lid/chill block interface of 63 micro-inch. The material selected for the chill block was dispersion-strengthened copper alloy UNS C15715. This material has equivalent mechanical properties to the more common chrome-copper alloys typically utilized for such applications, but considerably better thermal properties.
Chill block force is applied to the lid by a specially designed tool that grips the center (pintel) of the lid. The pintel is engaged by three jaws when the tool lever is rotated and force is applied by pulling the cam-action lever over the top of the tool. In addition to its thermal function, the chill block, which is integral to the weld head, helps align or center the can with respect to the orbiting tungsten. It also provides a mechanical stop automatically locating the cross-seam position of the tungsten tip.
Figure 8 - A typical weld tack.
Tacking and Weld Start Conditions. A tacking sequence was implemented to maintain axial and radial fitup of the weld joint during welding. Seven small tacks (approximately 3/16 inch in length and 0.040 inch deep) are deposited evenly around the can at 45-degree intervals with the weld beginning at the eighth octant – Fig. 8.
During weld development it was observed bead penetration was somewhat less at the weld start than in the remainder of the weld. To compensate for this, a preheat was added to the start of the weld by holding the arc stationary for several seconds.
Base Metal Chemistry Evaluation. The effects of low levels of sulfur on bead penetration and shape in fusion-welded austenitic stainless steels are well documented. As noted previously, controlling weld bead geometry was of primary importance, therefore, the effects of base metal chemistry were closely considered. Weld bead penetration and shape were correlated to the different levels of sulfur as supplied by the various heats of container and lid materials used in development testing. The full range of sulfur, as specified in the ASME material specification, was not available for direct evaluation. Results of the tested levels, however, when combined with information from the welding literature, provided sufficient data to prescribe specific sulfur levels for this application. In addition to bead geometry, sulfur levels at the upper end of the material’s specified range were evaluated for potential deleterious effects on weld bead solidification. Sulfur levels specified were Container Shell = 50 to 250 ppm S and Lid = 100 to 250 ppm S.
Qualification of the Process Welding Parameters
Formal Qualification Testing
Having identified the production welding parameters, the process was then subjected to weld qualification testing as required in the DOE 3013 standard. Table 4 lists the various tests performed along with their results; all testing met specified acceptance criteria. Figures 9 and 10 provide photographs of drop and burst testing, respectively.
Figure 9 - Container prepared for the 30-ft. drop (left).
Tested container (right).
Figure 10 - Prepared container (left). Tested container (right) with failure completely outside of the closure weld.
Demonstration of the Qualified Welding Process
Figure 11 - An installed OCW system.
Reliability Testing. Following the qualification activities, a run of 100 test welds was made to test the durability of the equipment and to evaluate various process upset conditions. Upset conditions were selected based on the likelihood such conditions would be encountered during production operations. Table 5 lists the upset conditions, evaluation technique, and results. In general, the process responded well to the various conditions. Wide variations in welding current were easily accommodated. As might be expected, loss of weld shielding gas and residual container internal pressure led to poor welds. Overall, the process was deemed robust and capable of producing acceptable closure welds, even under some off-normal conditions.
Customer Acceptance Testing. Prior to delivery of the closure welding system, five 3013 containers were welded using qualified operators under mock production conditions in accordance with QA-approved procedures. In addition, full oversight was provided by the Hanford authorized inspector for the tests. Completed welds were subjected to leak, radiographic, and metallographic testing. All testing met specified criteria. Figure 11 shows the completed and installed OCW system.
The authors are grateful to many Savannah River Technology Center individuals who contributed to the success and completion of this work. Among them are Paul Korinko, Tina Usry, David Maxwell, and Glenn McKinney. Steve Harris provided statistical analyses. In addition, recognition is given to Scott Breshears, Kurt Peterson, and James Tarpley for the design and fabrication of the equipment. Finally, appreciation is extended to Bob McQuinn, Rob Gregory, and their entire team at the Hanford Plutonium Finishing Plant, where the system is up and running.
1. Stabilization, Packaging, and Storage of Plutonium-Bearing Materials. 1999. DOE Standard 3013-99.
By G. R. Cannell, W. L. Daugherty, and M. W. Stokes
G. R. Cannell and W. L. Daugherty are principal engineers and M. W. Stokes is a Fellow Engineer with the Savannah River Technology Center, Westinghouse Savannah River Co., Aiken, S.C.
Reprinted from the Welding Journal, July 2002.