Commissioning their new production line meant the joining technology experts at Magna had to get extremely creative. Before production could go ahead, externally commissioned system integrators had to find a way to bring components into pre-programmed welding positions using state-of-the-art robot systems. Yet achieving the precise sheet edge position needed for welding still proved to be a challenge during the pre-series stage. Naturally the powers that be at Magna were under no illusions that 100% accuracy is a fundamental prerequisite in joining technology, especially when joining light-gage sheets. Before any customer vehicles could roll of the production line, the perfect join between the upper and lower sheets had to be guaranteed.
“At times we work with robot systems ʽon the fly’,” explains Sonja Schober, joining technology team leader of the Business Unit Painted Body at Magna Steyr. This means that the entire back end of the vehicle is not picked up in the usual way. One robot, known as the geogripper, grabs the component at the specified points on the rear side and moves it into the intended positions. The other robot moves the welding torch into the corresponding welding positions. However, it must be ensured that the exact specified dimensions are observed: “Although there are various different tolerances regarding the precise component position within our company, the robots that are used also have specific tolerance ranges – for example, a welding robot has what we call the axis tolerance. This can lead to difficulties if the most unfavorable conditions arise at once; more specifically, if the component tolerance and robot tolerance are right at their maximum or minimum,” continues Schober. “Under certain circumstances the edge being joined can be up to one millimeter in front of or behind the welding wire stick out position. Fusion errors are the result, such as insufficient penetration or lack of fusion between the upper and lower sheets.” Werner Karner, head of the Joining Technology Center at Magna Steyr, also adds: “Light-gage sheet applications are everywhere in automotive engineering. If you get the weld seam wrong then you’ll quickly burn through them. Before we can start series production, we therefore had to optimize our systems in this area.”
Problems with fully automated joining technology usually come right out of the blue, therefore precise testing is needed for quality assurance in production: “Normally we teach the robot on one component and everything is usually fine from then on. But suddenly a weld seam appears out of nowhere that needs reworking, and in this case we would need to examine what had caused this,” explains Marco Miersch, maintenance mechanic at Magna Steyr. “There are of course numerous other causes of minor deviations in any production setup, which is why preplanned manufacturing tolerances are absolutely necessary. In other words, it’s a matter of planned and therefore foreseeable leeway – this is known as active tolerance compensation”, adds Schober. To design a structure with zero tolerances would be absolutely impossible! “For this reason, we depend on being able to reliably find the sheet edge or any gaps, allowing fusion errors when joining to be avoided.”
For everyday production this means that “we examine all of the components, compare them with one another, and if we have any doubts then we either teach each weld seam process on the robot again, or we adjust the power source parameters,” adds Miersch. “We are willing to accept the additional effort in hours and labor, plus the extra material, in order to guarantee absolute precision. This type of comprehensive quality control is our biggest priority as a premium manufacturer.”
If you want to avoid constant rework – such as random visual inspections and manual re-welding – or if you want to avoid component rejects, then you must ensure the correct position of the component before welding takes place. There are many ways to address this particular robotic welding challenge, but, how can this problem be solved when dealing with high unit numbers in a predominantly fully automated production environment? There are various camera systems that claim to make the robot “visible” for the position to be corrected, but, as Karner says: “The welding torches are very long. If another camera is positioned behind, it will not see far enough in front. Furthermore, all manufacturers face the same challenge when it comes to bodywork construction, as bodies are becoming increasingly narrow and angled! This means that accessibility is increasingly restricted, and optical measuring systems cannot keep up.”
However, Magna not only wants to guarantee the highest quality standards; they also want to set the benchmark when it comes to modern production standards in automotive engineering. Since commissioning the production line in 2018, Karner would like to take a new approach. “It was clear to me that the automotive sector needed something more long-term to support the robots with edge detection. Due to our long-standing and excellent cooperation with Fronius technicians and developers, I asked whether they might already have a solution available. The answer was immediate: we may just do…”
» “…reliably detecting the edge and thus the precise welding position has to be the fundamental prerequisite! «
The Fronius application technology team sprang into action: “Our contact partner knew that it was already possible for the wire electrode to be ʽrepurposed’ as a sensor on the intelligent TPS/i power source. For example, the Fronius R&D department had already conducted successful trials in the past. Furthermore, our power sources are configured for the CMT process and so all pre requisites for a ʽfield test’ were already in place,” adds Miersch on the constructive and solution-orientated cooperation.
“It was clear that Fronius had been thinking about this for a long time, but had never fully developed it until now,” explains Karner. “Due to our long-standing cooperation based on mutual trust, it ultimately made sense for both sides to install a prototype in our production department,” which would help bring the new sensor technology to market maturity.
“Since installing WireSense, the robot is able to find the sheet edge reliably and is able to adjust the welding path accurately to the relevant conditions,” says a delighted Schober. “We now have excellent and consistent penetration stability and are consistently able to achieve the perfect join between the upper and lower sheets,” underlines Miersch. Until now, Magna has only used the WireSense on one individual component in the aforementioned 2018 production line, which was the “cross bar ROPS” (roll-over protection structure): “This is where we had the most leeway and so it made sense to implement the process for testing here. Furthermore, the WireSense needs more time for the scanning process before welding. Since a buffer is factored in for this component, delays in cycle times have no effect on overall production.”
Using WireSense on all weld seams requires an entirely different plan before commissioning new production lines: “Before now it had not been possible for us to use this revolutionary function across the entire production setup. In a fully installed production system, the cycles times of individual stations are precisely calculated, so retrofitting WireSense for widespread use would throw the entire system into disarray. The planned cycle time for this rear component takes around five minutes. Every robot has a certain collision detection area. If one robot is currently working with the WireSense and therefore needs more time, the other robots would also be unable to continue to work. The knock-on effects would accumulate automatically and instead of 5 minutes to produce one part, we’d suddenly be looking at 10 minutes.”
Moreover, the WireSense settings also needed to be optimized, which could first of all be implemented on the test component, the cross bar ROPS: “We reckoned that each scanning process would take around three seconds when we first discussed the WireSense option – in the end it took eight seconds,” explains Schober. The Fronius service technicians were not prepared to let this stand and examined the problem in detail. Ultimately, the robot programming had to be perfected. Expert knowledge is a basic requirement in this field in order to fully benefit from the enormous advantages of a wire sensor. It is a matter of precisely defining the robot speeds of the individual stages of a scanning process. “Thanks to the programming changes, some stages were greatly accelerated, which meant that we reached the proposed three-second mark,” adds Schober.
Despite all the optimization, one principle remains the same: “One system impedes the other,” says Karner. “We have over one hundred weld seams in the rear end of this body alone. A three-second scanning process per seam of course ends up in a huge difference!”. Miersch adds: “In order to be able to profit from the WireSense, it is necessary to probe individual seams before production line planning, both to avoid error and additional rework. After all, these are critical seams due to their safety requirements.”
“Despite the additional time spent at the beginning, the WireSense will lead to significant time and cost savings in the end, as long as the seams are precisely assessed during the selection process. The bottom line is that although individual cycle times are longer, system availability is significantly increased and this is primarily due to a reduction of expensive rework by nearly one hundred percent,” summarizes the head of the Joining Technology Center, Karner.