Welding Automation Technologies Join the Old and New
Learn how humans and robots can work together to create an adaptive welding process By Jaakko Heikonen, owner, Pemamek Oy, Loimaa, Finland Reprinted with permission: The AWS Welding Journal Welding automation is not a new concept, but the way it is implemented is constantly evolving. Automation doesn’t eliminate the human element or intelligence — it just uses it in a different way. For example, rather than having the welder climb a scaffolding, balance themselves on a slippery surface, or squeeze into a dangerously tight place to fill a groove, he or she can change the wire feed, heat range, wire diameter, and much more via a computer screen while the robot does the physical labor. This article aims to take the mystery out of automated welding and provide examples of how the process doesn’t have to be threatening to the welder workforce. Getting Ahead with Offline Programming Bottlenecking is a problem that is often experienced in traditional programming. A bottleneck occurs when the capacity of an application or a computer system is limited or slows down. According to Teemu Rusi, robotics manager at Pemamek Oy, Loimaa, Finland, the ability to program robotic applications offline is ideal for low-volume, high-mix jobs because the next welds are preprogrammed while the robot works on another piece simultaneously — Fig. 1. This solves the bottleneck problems where the operator needs to be next to the robot to control its motions with a teach pendant. Offline programming also enables a preanalysis of weldments; the operator can examine which ones will be suitable for robotic welding.
“The customer can have an additional license not only for programming but for the production planning phase,” Rusi noted. “They can import the piece as a 3D model in the software and see whether the robot can reach difficult welds or how it should be attached to the workpiece positioner.”
The simulation enables collision-free paths and an exact-time study already in the preplanning phase.
Large parts and structures can take time to program and often have irregular groove welds due to the nature of their geometry and size. Adaptive welding paths can be programmed offline as well, allowing the software to adapt those paths to fill grooves of different depths and sizes. This is especially suitable when working with one-off or low-volume jobs in succession.
“In the case of a ship bulkhead, which has many intricate parts and pieces, the first part can be programmed well before you ever get it to the point where you can start to weld,” said Michael Bell, national sales director, Pemamek North America. “When you start to weld the first part, the second part or section can be programmed while the first is welding. You can choose to what degree you want to program the operation.”
Creating Weld Paths Using Parametric Inputs and Scanning
Adaptive welding is practically tailor-made for the jumbo-sized parts found in wind towers and foundations, power, and offshore industries. The sheer size of these pieces almost guarantees there will be inconsistent grooves and measurements that are difficult for welders to manually reach with the welding torch.
Rusi recalled a challenging job brought to him by a Norwegian offshore customer who was building a large steel winch for vessel anchors. Previously, the customer welded these pieces manually using 39.37-in.-long welding torches. Hard-to-reach sections inside the piece and exceptionally thin sections were impossible for a human welder to get to, and they experienced numerous quality issues as a result. The solution was a welding cell with two robots on two column- and boom-type gantries with three-axis movement: X = 5 ft (1500 mm), Y = 6 ft (1800 mm), and Z = 10 ft (3000 mm). A 50-ton, two-axis welding positioner sat in the middle, capable of rotating the part 21 ft (6500 mm) in diameter.
“In this case, we had 59.05-in.-long-reach robotic arms that could weld inside the workpiece. Two robots welded simultaneously with the welding positioner. Welding paths were programmed using parametric input methods whereby the operator entered the outer diameter of the piece and the inner core diameter and the number of ribs inside where the robot was welding. Tailor-made welding programs were created based on these parameters entered by the operator,” Rusi explained.
Welding time was reduced from 36 to 6 h, and the welds were spot-on accurate.
“There were hard-to-reach areas on that vessel,” Rusi recalled. “The welder became the operator, preparing the next pieces while the robot was doing the tough welding task. I think one of the biggest advantages here is that the welder could get away from welding fumes, arc illumination, and positions that are hard on the back. The robots can do these hard tasks.”
Another approach to creating custom weld paths for tough-to-weld workpieces is by allowing the robots to scan the workpieces themselves, then generate programs based on the scanned data, automatically adapting each welding pass to those areas where the grooves need to be filled. As with offline and parametric programming, the operator can take a symmetrical 3D computer-aided design image of the workpiece and split it down the middle to create a mirror image.
“The operators can use copy and mirroring tools to create one weld on a piece that they can copy or mirror to the opposite side of the weldment,” Rusi said.
In some cases such as the one described above, welders become operators by trading in their welding torch for a sophisticated software program that controls robotic welding arms to accurately reach and join those tough-to-reach spots where manual processes just won’t do.
Multigenerational Welding Approach
Younger, newly graduated welders have had robots in their lives since birth, and they have most likely used some sort of basic welding path software program. It’s the more experienced welders, those who are nearing retirement age, who might balk at the idea that a robot can do a human’s job. But, what if the two generations combined their experience and knowledge to create the most optimal adaptive welding processes around?
According to Bell, software developers “take that welding knowledge — metallurgy, speeds, feeds, etc. — put it in a program and automate it. Welders still need their education, background, and understanding of the process but can program the robots to do the work.”
Robots are especially beneficial for back-bending work. Manual welding can be a physically demanding job with workpieces ranging from manageable sizes like those found in automotive applications to those requiring scaffolding and precarious positioning to reach a welding area. At some point, the welder can be put in a precarious position that jeopardizes his or her safety. Automated welding technologies take advantage of the welder’s practical experience and knowledge honed over years on the job and give the physically tough tasks to an entity that can’t really get hurt.
Onshoring the Production Process
Manufacturing in the era of the COVID-19 pandemic has highlighted the weak links in the supply chain from country to country. There is much talk about onshoring manufacturing, not just in the U.S. but in other countries in Europe and beyond. Investment in automated technologies goes a long way toward keeping production in-house for many companies.
“It helps you to keep manufacturing close to home. Automating welding cycles can result in big savings during the manufacturing process and increase efficiency,” said Pemamek CEO Juha Mäkitalo. “Some of our customers have realized more than 50% reduction in cycle times by automating their welding operations. In time, you can add another shift using only automated processes (think preprogramming) using the same number of people on the job. I think there are a lot of opportunities to increase efficiency.”
Customization of equipment can help speed along processes even more.
“We always analyze customer production and the products they are planning to use for robotic welding,” Rusi said. “We check their portfolio of products and then determine how big the welding stations should be, how long the robot arm reach should be, how much movement for the external axis do we need, and what the capacity should be of our workpiece welding positioner. All of our robotic systems are modular and can be modified according to customer needs.”
Welding of the Future
In the future, Rusi envisions an all-automated process, from programming to final weld, for seamless, collision-free movements, while Mäkitalo looks to the integration of different types of software such as welding control programs with enterprise resource planning systems so that every aspect of the manufacturing process is documented. One thing is for certain, if accuracy, efficient welding operations, and the ability to have absolute control over the manufacturing process is a priority for your shop, you’ll be adopting automated welding technologies quite soon.
Fig. 1 A ship component is welded by a robot. 
