Optimizing flow through robotic welding workcells
A 'Case' Study Revisited
Robotic welding systems can enhance a company's production and bottom line. However, using these systems requires careful thought and planning, building the right infrastructure, and achieving the right balance between robotic and manual operations. This article presents an overview of one company's successful implementation of robotic welding.
Veteran industrial robotic user Case New Holland (CNH), Fargo, N.D., recently installed three robotic welding systems. Each system consists of an inverted robot on a large, three-axis traveling column and two large, dual-axis (skyhook) positioners. These systems weld front and rear chassis for the company's line of wheel loaders. This article discusses how CNH optimizes the flow through these new, sophisticated workcells to maximize throughput and ensure optimal weld quality.
Bill Moser, CNH manager of manufacturing engineering, is responsible for this and many previous robotic automation projects. According to Moser, robots are not a luxury at CNH but a necessity: "Automation is our key to success ... robots address manpower shortages [a nationwide issue], cost savings initiatives, and quality initiatives."
This project is not your run-of-the-mill robot installation. The weldments are complex, weighing up to 4,000 pounds, and some contain nearly 250 welds. The robot systems use both tandem-wire welding and conventional single-wire welding, with automatic tool changers to switch between the two processes during automatic operation. Installing the systems coincided with a new product introduction, which made the project even more challenging.
Part Design for Automated Welding
The first goal was to design the parts for robotic welding. Moser said, "One of our biggest goals is to weld 90 to 95 percent of the frames with robotics. As we progress further into robotic applications, design for robotics is more critical than ever."
What this means in practical terms is that the parts are simplified, the number of component parts reduced, and when possible, the weld seams are designed for optimum access by a robotic welding torch. This design makes a dramatic difference in the robot's ability to reach more weld seams. More important, the new parts have fewer weld seams, meaning less welding than with the old parts. Welding is an expensive process, so it makes good sense to eliminate it whenever possible. Redesigning the parts for robotic welding had led to significantly shorter cycle times and improved weld quality.
The weld seams also are designed to provide access for the larger, bulkier tandem welding torch. Tandem welding is a process using two independent contact tips inside a common welding torch, powered by two independent welding power supplies (see Using the tandem welding process to your advantage). Tandem welding technology doubles weld deposition rates while actually decreasing overall heat input. Because some weld seams are not accessible by the larger tandem torch, the robot systems include tool changers to automatically change between tandem- and single-wire welding torches during the welding cycle.
Moser said, "Tandem welding has reduced manual welding to one-third the time, and it has reduced single-wire robotic welding time by one-half." But he warned that "tandem welding is a process that offers huge cycle-time reductions in certain but not all applications." The pros and cons of tandem welding versus single-wire welding must be studied to determine which process is most beneficial for a specific application.
Although the robots are responsible for 90 percent to 95 percent of the welding on these parts, certain welds still are done manually, either because the robot cannot reach these welds, or because it simply makes more sense for some difficult welds, wraps, and tie-ins to be welded manually.
Optimizing Work Flow
In the robotic welding industry, it sometimes is preferable to share the welding between robots and people for a number of reasons.
First, the tolerances on heavy components are not as tight as on smaller parts. Heavy-steel characteristics are such that thick plates simply cannot be cut and bent within very tight tolerances (at least not cost-effectively), so the actual location of weld seams will vary to some extent. Available options allow the robot to find weld seams that vary in space and automatically adjust the robot program accordingly, but conditions like gaps and wraps around the ends of plates and gussets can be difficult for a robot to handle. Manual welding in these difficult-to-control areas can increase weld quality and decrease the amount of rework caused by faulty robot welds.
Second, access to some welds may be blocked by the fixture required to clamp the part in the robot. CNH minimizes this problem by pretacking the chassis and presenting a rigid tacked assembly to the robot. This greatly simplifies the robot fixture and minimizes the number of clamps and fixture details that can block access to welds. Nevertheless, some welds still cannot be reached by the robot, so the operator makes these welds manually after the robot is finished.
Third, a manufacturing cell's throughput may be increased by eliminating material handling. CNH specified that each robot system be provided with two workstations so the robot can be working at one station while the operator is unloading, reloading, and doing some prewelding at the other station. Each welding positioner is equipped with a manual jog option that allows the operator to jog the robot positioner while he is welding manually and the robot is welding at the other station to optimize access to the part for manual welding. This eliminates the need to install a manual positioner near the robot cell.
Although CNH strives to do most of the welding robotically, other companies or products may benefit from doing more welds manually. This can decrease robot cycle times and increase the robot cell's overall output. Manually welding the most difficult welds for the robot increases overall quality and decreases rework.
Pretacking and Prewelding
An important aspect of the CNH project is not only to share the welding between human and machine, but also to address strategically the type of manual welding being done both before and after the robotic welding process. As mentioned, these large chassis are pretacked so a rigid assembly can be presented to the robot using a simple locating fixture, and they also are prewelded. What's the difference?
Tacking is simply the means of holding the parts together until permanently welded by the robot, but strategically placed tack welds can help to increase robot weld quality and improve part integrity. A large tack weld in the middle of a straight seam can cause lack of welding fusion directly under the tack, or can cause the robotic seam tracker to track out of the joint, at least temporarily. Rather than tacking randomly, CNH places tack welds in strategic locations to minimize robot problems, or replaces tacks by full-fledged wraps. Moser explained:
One of the things we have been working very hard on this year is the tacking process. We are focusing on having the welders strategically tack the parts versus throwing down randomly placed tacks. By doing this, we are able to wrap a lot of corners manually and program the robot to weld into these wraps, thus making a much cleaner weld and reducing manual input in the tacking/finishing process.
The tacking is done in a complex tacking fixture designed to locate a very large number of individual components. Because of the fixture's complexity, the amount of corner wraps and welding that can be done in the tack fixture is limited. Therefore, a process called prewelding also is done before robotic welding. Prewelding is done by the robot operator after the part is loaded into the robot fixture. This allows the operator to use the manual jog function to jog the part into ideal position for wrapping corners and performing other prewelding.
Care is taken not to overburden the robot operators with too much manual welding. The robot operator is responsible for keeping the robots running, performing limited preventive maintenance, and loading and reloading the robot fixtures, so they have only limited time available for welding. The prewelding procedure is kept to about 15 to 20 minutes, which often is not enough time to prewrap every corner. As a result, the prewelding to be done by the robot operator is defined very carefully to provide the biggest bang for the buck and to allow the operator to take best advantage of the dual-axis positioners to manipulate the parts during prewelding. If done properly, strategic tacking and prewelding can eliminate problems such as the robot burning off the end of a plate or dripping off the edge of a weld seam.
Manually making these wraps and tie-ins also can improve the integrity of the welds. Plate and gusset corners and ends often are areas of high stress, and it is good welding practice to avoid weld starts and stops in such areas. By wrapping around the corners manually (a job that is difficult for a robot to do consistently), the weld starts and end craters can be moved away from the corner or end of the plate, thus improving weld strength.
Any wraps or touchups not done during the tacking or prewelding stage are completed by the finish welders, who also add whatever array of small parts (brackets, hose retainers, studs) are required for assembly purposes. The finish welders are responsible for the part's overall weld quality, because they perform the last operation before the parts are sent to the paint line. The finish welders do any tie-ins, wraps, weld repairs, and straightening required for finishing the chassis to print and preparing it for paint and ultimate assembly.
CNH has installed many robots during the past 15 years and is among the best at utilizing robotics. These are complex machines that take advantage of some very useful technologies, including tandem welding, through-arc tracking, automatic tool changing, and digital pulse-arc welding. The company puts a lot of effort into designing its parts and fixtures for robotic welding and into specifying and planning for robotic welding systems. Its newest systems can accommodate about 23 different chassis configurations. Moser said his goal is always to "design a robotic system that will give the most flexibility possible."
CNH has learned by experience the value of sharing the work between robots and people. The company pays close attention to the details, which enables it to take optimum advantage of the unlimited flexibility and adaptability of the human machine, as well as leverage the strengths and the limitations of robotic welding technology. It has a very pragmatic approach to robot projects. "Too many times people purchase automation," cautioned Moser, "and expect to turn the key and watch the money roll in they believe the cost savings begin the minute you install robotic equipment."
Moser said it takes a lot longer than that to "prove out the robotic process, prove out the operator process, and prove out the best practices" for your particular application. When asked to give advice to less experienced robot users, Moser replied, "Robotic welding is not the silver bullet. You need to develop the infrastructure to support the robotic applications, which includes understanding the process, understanding robotics, understanding maintenance and troubleshooting, and understanding the manual process in combination with robotics.
"Robots should be just one of the tools in your toolbox. One of the most important activities we do is line balancing, which is breaking work down into tasks to reflect the output requirements." Moser said that the company's cost savings from line balancing are as significant as its cost savings from robotics.