PCs Control Welding of Nuclear Fuel Rod Spacers

To ensure that robots safely weld components for a nuclear reactor requires monitoring and control beyond what typical technologies can offer. Advanced Nuclear Fuels (ANF) GmbH, based in Karlstein, Germany, commissioned assembly and micro-assembly technology supplier Rohwedder AG to build a state-of-the-art welding plant to produce steel grids for use as fuel rod spacers in BWR (boiling water r...

By Staff December 1, 2005


High-precision welding chambers

Recipe management

Synchronous drive system

High-speed networks

To ensure that robots safely weld components for a nuclear reactor requires monitoring and control beyond what typical technologies can offer.Advanced Nuclear Fuels (ANF) GmbH, based in Karlstein, Germany, commissioned assembly and micro-assembly technology supplier Rohwedder AG to build a state-of-the-art welding plant to produce steel grids for use as fuel rod spacers in BWR (boiling water reactor) or PWR (pressurized water reactor) nuclear plants.

Critical requirements for the plant were high welding precision and accuracy coupled with high speed. During the welding process, up to 8,000 parameters must be transferred in a cycle time under 1 second for each weld point. Rohwedder chose a PC-based PLC (often referred to as a programmable automation controller or PAC) and motion control technologies from Beckhoff to handle this process.

Welding process detail

Rohwedder had to create a welding plant for joining an initially loose construct of longitudinal and lateral bars via laser spot welding to form a rigid steel grid. The plant consists of two welding chambers, each equipped with a laser for the actual welding, a camera system, and a PC for visualization located at the control unit. A Kuka robot handles steel grid components.

First, parts are mounted on a workpiece carrier. Then, they are transferred via conveyor belt into the welding chamber and positioned by the robot. “Components are made of a zircon alloy and the material has to be welded in an argon gas chamber since in a normal atmosphere it would catch fire,” explains Peter Blomberg, Rohwedder’s control technology manager.

Once the workpiece is positioned in the welding cell, the cell is closed and a vacuum is generated. The chamber is then flooded with argon.

Welding process specifications also are demanding. Highly dynamic linear-drives are used for moving the laser to the individual contact points. Here, the laser melts the material with high precision, creating the required weld quality. A camera mounted in the cell monitors the complete welding process. The camera system surveys the grid before the welding process and defective welds are immediately identified and automatically re-welded.

“The welding process is a high-precision, high-speed process that generates a huge amount of data,” Blomberg says. The plant can process more than 50 steel grids simultaneously, 26 of those can be of differing configurations. Individual grids are asymmetric—some have cut-outs. “This means that the system has to manage more than 50 component variations. Positions, travel paths, laser control, NC [numerical control], and camera programs have to be ‘parameterizable,’ ” or put into measurable units, Blomberg says.

Regarding data volume, each spacer requires up to 1,000 welds, each with a cycle time of under 1 second. During this time, the following steps are executed: Positioning, correction via camera, laser welding according to recipe parameters (manual, via NC program, or spot welding), and weld point analysis.

Due to drive, control technology, and visualization requirements, Rohwedder uses PC-based control technology from Beckhoff running version 2.9 of TwinCat NC I automation software and a 19-in. C5102 slide-in 2.4 GHz PC as the control PC. “With TwinCat, we have decided to use an open PC solution instead of conventional axis modules and CNC controls. Exact positioning, fast motion, high data volume, and measuring tasks necessitate the use of a PC-based controller such as this,” Blomberg says. “We would not have been able to handle the complex control tasks using a conventional controller.”

PC-based control technology also offers recipe-management advantages. All recipe parameters, including NC and laser programs, are handled via Microsoft Excel. Blomberg explains that Rohwedder prefers Excel because “it offers the simplest form of data handling, and no programming knowledge is required for creating the recipes.”

The Excel file contains information about the component (steel grid type), such as size, number of weld points, distance of the webs and of any cut-outs, and a description of the weld sequence, including parameters for each weld. The document also contains the NC program, the laser and image processing program, and associated parameters. “The Excel data are read into the associated PLC data-structure. For each component, the PLC creates a data file containing recipe and process parameters, which is also accessed by the higher-level plant management system,” Blomberg says. “Recipe creation and processing of the different components are determined with the customer, who simulates the procedure using TwinCat’s ScopeView. This enables errors to be detected and rectified in advance, optimizing the process.”

TwinCat NC I’s 3-D axis interpolation enables Rohwedder to meet ANF’s requirement for the NC to have 3-axis interpolation with laser power control. “This means that from position 4 mm to 18 mm, for example, the power has to be controllable between 50 and 60%. Positioning of the linear axes for standard grids should take less than 300 ms, corresponding to a travel path of 12 mm,” Blomberg says.

Motor, controller networks

The ANF plant houses 12 water-cooled linear drives controlled with TwinCat NC I. In combination with a SERCOS drive bus, the result is a drive system that can synchronously move several axes to the target. “High data transmission rates and short cycle times with simultaneous interference immunity are arguments for SERCOS,” says Blomberg. “In our case, where high speed is a significant factor, fast and smooth data exchange between the motor and controller is obviously important.”

For this data exchange, Beckhoff provided Rohwedder with Profibus and SERCOS fieldbus system developed for fast controllers and real-time tasks, such as drive position control.

At the I/O level, Beckhoff bus terminals in protection class IP20 and fieldbus box I/O modules in protection class IP67 with Profibus interface are used. According to Blomberg, the decisive criteria for using the IP67 modules were their compact and robust design and modular expandability through additional modules via the IP-Link system. A coupler box is used for interfacing with the Profibus network. Further I/O modules are integrated via extension box modules. The coupler box collects the I/O data via the IP-Link optical fiber connection from the extension modules and communicates with the higher-level controller.

Rohwedder uses its in-house-developed AMS XP product as a plant management system for central monitoring and management of the process. “Apart from process visualization, AMS XP also deals with process data archiving, order and type management, and statistical analysis. The traceability tool is an increasingly important feature for us,” says Blomberg. “All maintenance is documented. Any errors or reworking, as well as information on third-party products are stored within the program. In the event of damage, the end-user can obtain information about the cause.”

For related information on PC-based control, see cover story, this issue.