|
 |
|
Nuclear Fuel Canister Welding Solutions
Selecting and Applying an Automated Welding System
One utility's experience with an automated welding
system to seal dry-storage spent fuel canisters
By Mario Lento, Berkeley Process Control
 Sealing dry-storage canisters of spent nuclear fuel requires a welding system that precisely, reliably, and efficiently welds canister lids and keyways while minimizing radiation exposures of workers. One U.S. nuclear power plant operator met this challenge in the early 1990s by building a proprietary welding system. Over more than a decade of service and dozens of completed canisters, that system well served the plant operator's needs. By 2000, however, replacement parts had become both very difficult to locate and costly to obtain. The plant operator realized that the time had come to replace its proprietary welder and to take advantage of substantial technological advances in the automation industry to improve operational efficiencies.
The plant operator's team of experts extensively interviewed welding machine vendors and visited other plants using welders to determine what type of system would best meet their needs. They ultimately chose to purchase a highly automated welding system (AWS) from Berkeley Process Control.
The Heart of the Matter
At the heart of the AWS is one of the company's flexible, off-the-shelf, 64-bit RISC motion-and-machine controllers. In the AWS application, that controller precisely manages the wire feed, torch head positioning, voltage and current outputs, and scheduling of all process parameters. Unlike welding systems that consist of multiple disparate processors, the AWS controller is the single unified control element in the welding system. User-defined parameters such as weld start sequence (including prepurge time, current rampup and rampdown, and postpurge time), wire feed schedule (start feed delay, stop, feed delay, rollback distance, primary and secondary speeds, duty cycle and synchronization with oscillation or current), and oscillation (excursion time, dwell, width, and synchronization with current) are all precisely controlled and synchronized to deliver optimum weld quality, and canister-to-canister repeatability, with minimum operator fatigue.
The gas tungsten arc torch head of the AWS makes a complex oscillation pathway from outer to inner diameter welds. The AWS operator defines the amount of oscillation along the weld path. The system accomplishes the oscillating weld perfectly and efficiently because the controller tracks the exact position of the torch head in real time and likewise in real time can pulse the primary and background currents to the torch head based on its position on the canister. Operators configure the AWS to achieve the weld penetration results they desire. Arc voltage is maintained by modulating the height of the torch tip according to real-time process feedback. In this manner, the operator can control the application of heat to the canister based on preset recipes and thus produce repeatable, high-quality welds. The operator can correct the path of the weld on the fly while the AWS maintains tight coordination of all weld parameters in real time.
Controlling Software
Some other welding systems use personal computers (PCs) to manage the control of disparate subsystems. In each case, complex custom software is required to define the unique sets of components installed on the PC. This software becomes a unique control system unto itself, and any change to that system requires development of new control code. All replacement hardware must be identical to the original part, which, as the plant operator in this case learned, often results in difficulty finding those outdated parts at any price.
By comparison, the single integrated control architecture of the AWS manages all motion axes, input/output, scheduling, and logic control. The welding system can be easily maintained and upgraded without development of new software.
Teaching the Machine
Unlike some competing systems that require the welding machine be exactly physically centered on the canister, the AWS requires that the welding machinery be only within a few inches of center. A remote crane effects positioning, and therefore workers' residence time in the welding chamber, as well as their radiation exposure, is kept to an absolute minimum. To accommodate this approximate mechanical positioning, the AWS provides a teaching utility that precisely "learns" the position of the torch head relative to the canister and thus affords a simple one-time remote calibration before welding takes place. This step takes only a few minutes.
The AWS teaching process assists the operator in finding the precise path to be welded prior to beginning the welding process. To assist teaching, the AWS automatically advances through the eight circumferential and three keyway teach points. Using front- and rear-view video monitors on the console, operators can hone in each point precisely before accepting that teach point and allowing the AWS to move to the next point. The AWS then calculates all points in between, including compensation for human error in measurements or for out-of-round canister lids, to provide the optimal weld path to seal the canister. Without further teaching it automatically ties in successive welds for consistent weld quality.
Because heat generated by the welding process may cause small changes in the shape of canisters, the AWS operators may occasionally, and remotely, fine tune the torch head position during the weld. "If the machine has been taught properly, basically all you have to do is sit there and watch it," notes one of the plant's experienced AWS operators.
Staffing Needs
The plant's old proprietary welding system, as is true with some competing commercial welding systems today, required four boiler-makers as operators plus a systems technician for each canister weld. The welding skill of the boilermakers largely determined the end result of the canister weld. The AWS, on the other hand, in substantial field experience, required only one experienced welder to operate the system along with one assistant. In this case, the plant operator decided to use two pairs of AWS operators per shift, alternating every two hours. Relatively little operator input was required once the AWS had been taught its pathway. With the performance of 64-bit processing, the Berkeley controller maintains and controls all values and parameters with 15-digit accuracy across the welding of a six-foot-diameter spent fuel canister.
Prior to initial operational use of the AWS in 2000, four plant AWS operators spent one week at company headquarters to learn AWS procedures and practice welding on a mockup version of the canisters used at the power plant. Back at the plant, virtually no changes were required in the company's operations protocols prior to use of the new welding system.
Plant Experience
The plant operator has subsequently sealed 10 canisters with the AWS. Over the course of sealing those canisters, the utility estimates that worker exposures during the welding operation have been cut a total of 300 millirems. This approximately 50 percent reduction in occupational doses was effected largely because operators need to spend less time in the welding chamber to position and secure the welding system to the canister. Welds are accomplished with operators remotely located 100 ft from the canister, and the welding process itself has taken up to 10 fewer hours per canister than with the previous system.
Preprogrammed operations for the inner cover, outer cover, and qualification welding, combined with the extreme reliability of the AWS, have improved the plant's ability to consistently perform welds remotely and thus reduce exposures to workers.
The plant operator also realized a considerable spare parts costs savings by choosing the AWS. Although one competitor recommends nearly $90,000 in spare parts inventory for their welder, the AWS vendor recommends on-hand spare parts totaling less than $2000. Much of this spare parts reduction is brought about because the AWS' single controller allows a far simpler electro-mechanical design than competing systems. In addition, procurement of parts and supplies for the AWS is considerably easier than for the plant operator's previous welding system. Instead of working with a long list of vendors for welding, software, and hardware parts and service, all welding supplies are purchased through a local welding supplier, while Berkeley supplies all other hardware and software support.
The plant operator's success with the AWS - realized in increased worker safety, exceptional reliability, increased productivity, reduced spare parts outlay, and simplified parts and services procuremen - led managers to recommended it to a sister facility looking to install a new spent fuel canister welding system. Subsequent to a competitive evaluation and bid process, that facility also opted to purchase an AWS.
This article originally appeared in the November/December 2002 issue of the American Nuclear Society's Radwaste Solutions.
|
|
|
BX-300 Wafer Handling (PDF)
Registration Updates
Online Support
