Robotic Welding Robot Retrofit
Overview
The company here was a major automotive OEM, focused on large contracts, rather than civilian consumer sales. Governments made up around 90% of their customer base. This meant that while volumes were lower, standardization across any one run of products allowed for exceptional efficiency gains.
In this case, the company typically started making components by hand, then piece wise transitioning to automated solutions. The project we are talking about here was to implement a robotic weld cell to streamline manufacturing of a vehicle door.
Challenge
The manufacturer had just won the largest contract, in its history. Not just in number of vehicles to sell, but also in complexity of vehicles. The bill of material for a door weldment on one of their past products was around 15-20 individual pieces. The doors on this new vehicle required 45-50 pieces!
On top of that, the manual welding fixturing was neither robust, nor efficient for the production rate required.
The filler material required to weld these doors also had special safety requirements due to fumes released at the production rate required.
To hit the high production rate of this new contract, the company had hired around 20% more welders, many inexperienced. While they had a training program to teach new employees to weld, these were still green welders.
Finally, the project had an exceptionally tight budget. While the doors were a major pain point, several other weldments were in worse shape. These other parts got first access to funding, leaving the doors team with not enough capital for a traditional integrator-automation project.
Project Approach
The company made clear that their priority was:
Safety
Quality
Cost
Rate
As always though, if the cost wasn’t right, they would just hurt until they could get the capital next year. To make better use of the limited funds, the company opened up it’s storage facility where it kept old tooling from past production runs. There we were able to pick the starting point of the project.
Finding a used robot
While there was several robot cells to pick from, there was one cell that best met the priority list. The weld cell we found was a two arm Fanuc weld cell with two rotary tables and a full enclosure with light curtains/partial walls. It was even a palletized system, meaning transport was straightforward.
By using one of the customers previous cells as a starting point, we already had the basic outline of a safety cell. We knew the robots had a reliable maintenance schedule while in service. Most importantly, we only had to pay outside vendors for the transport of the machine since the in house maintenance team could install this system. In addition, the company would no longer be paying storage fees on the large weld cell.
Project Planning
With the starting point selected, the bones of a project plan were already in place. We set out the following path:
Install the old cell as is.
Install new fume extraction system.
Install new fire suppression system.
New tooling to hold the door weldments.
Program robots.
Train operations team.
Tooling Design
Since the install with fire suppression and fume extraction would take some time, we focused on tooling design. The company made clear they were unhappy with the existing manual tooling. This meant a clean slate tooling design. We opted for a simple design that shared as many parts with the companies existing production lines. This meant that clamping cylinders, shims, hardware and weld consumables could all be stocked and shared across the facility.
The new tooling was both ergonomic and easy to load. By using a preferred water jet vendor, we were able to skip expensive machining operations. Since we started from scratch we were also able to make access for this particular robot a priority.
The cost for this tooling came in at under 10% of what it cost to make the original manual fixtures, all while being more robust and ergonomic. After this project, most new tooling the company made followed our basic design process using water jet and off the shelf parts.
Robot Programming
Programming a whole new sequence for the weld robot allowed us to incorporate some features that helped both cost and safety.
We incorporated the fume extraction system into the program, allowing us to start running it shortly before welding began and keep running shortly after welding finished. By only running when required, it saved energy, reduced noise while the operator was loading/unloading all while ensuring the harmful fumes were exhausted.
The program also made use of probing functionality to offset the weld program in case the fixture was shimmed or improperly loaded.
In addition, we were able to sequence the welds to take advantage of counteracting heat input. Since this cell had two rotary tables, the doors could be flipped around, welding 4 at a time. This allowed us to run the robot faster and weld hotter, improving throughput substantially.
The programming was quick and straightforward, requiring minimal touch ups. We consumed 4 sets of doors in the process, 3 of which were fit for use at assembly.
Operator Training
The final step in commissioning this robotic weld cell was operator training. The manual door weld line had 12 welders at any given time. Operations leadership indicated there were 2 welders most interested in learning to run a robot cell.
We included the welders at the end of the programming phase, having them load the doors and components as we created and tested the program.
We also allowed them to see the program, and understand what the lines of code meant, even though the program was to be locked down, only accessible to the engineering team. This meant, when an issue occurred, the operator was able to provide meaningful information, and even troubleshoot some issues on their own.
Results
Safety was improved dramatically over the old manual welding line. The robot erased the ergonomic issues welders were experiencing, and the new fume extraction system ensured cleaner air for the shop.
The weld quality was also drastically improved. Around 95% of doors manually welded required at least 5 min of rework. The new robotic welded doors dropped the rework rate to 20% requiring 5 min. Most importantly, it allowed the quality team to maintain the weld process as a constant. Allowing them to find issues in the raw component cutting, forming and machining.
The project cost around 5-10% of the total quoted by traditional integrators. While the used robot was a large part of those savings, the tooling and integrated training made the largest difference. By revolutionizing their tool design practices and training operators as we programmed, the project cut major waste.
Finally the project was delivered on time in a fully functional state. There were no return trips or pushed deliverables. The cell functioned as promised, on time.
Conclusion
While this robot integration was highly unusual, it showcased the power of flexibility. We strive to be creative and find the best path. Sometimes the existing way makes the most sense, but in this case, being agile made all the difference. The company even went so far as to order a second (brand new) identical weld cell to double production using duplicates of our tools and programs. This allowed them to skip all the engineering costs, only having to buy the materials.
We are proud of this result and the way this project turned out. We learned a lot along the way and can’t wait to implement these lessons and practices on our next project.