Empty or malfunctioning machines can be moved to the maintenance loop and replaced on the production loop, enabling lay-up to continue uninterrupted. “With classic tape laying or fiber placement, when the fibers are exhausted or the machine is defective, production rate drops to zero,” says Kruse. “With GroFi, if a platform is empty we put another platform into the production loop and move the empty one into maintenance, and the drop-down is nearly zero.”
Robots in the maintenance loop also can be used for basic research, by using the platform in manual-driven mode, he says.
The tracks and control system for GroFi are in place, along with four robots with two tape-laying and two fiber-placement heads. “The next step is four more robots. We are discussing whether it will be four more tape-laying or fiber-placement [units]. It is not clear yet.”
DLR ZLP is beginning fully automated lay-up tests after completing a manually controlled trial. The tests involve laying up carbon fiber on a 7-meter (22.9-ft.)-long demonstration wingbox with A350 geometry, the single-sided Invar mold tool mounted in the vertical position. The platforms are standard Kuka industrial robots, modified with a Siemens computer numerical control system capable of running multiple cooperating robots.
Key to GroFi is the optimization software developed by DLR to distribute the fiber lay-up task between multiple cooperating robots. The process begins with the aircraft manufacturer's outer mold line and ply book, which specifies the number and orientation of carbon-fiber layers required at each point on the part. Computer-aided design and machine programming tools divide the ply book into fiber courses for single lay-up heads to follow, and then DLR's optimization software divides those courses between the different robots.
Testing began with tape laying, which is simpler, and will progress to both tape laying and fiber placement on the same tool. A second, full-size tool is planned, for a complete wing skin. DLR's goal is to demonstrate a minimum of three robots working cooperatively to lay up the wing skin. So far, DLR's optimization software can distribute courses between three robots. “By mid-year we will be able to do four or more,” he says.
Also part of GroFi is the development of sensors providing online control and correction of fiber path and monitoring for quality assurance. An edge-detection system senses gaps and overlaps, while a force-and-torque sensor provides online control of compaction of the composite as it is laid down.
Work so far involves conventional pre-impregnated thermoset composites, but in January DLR ZLP began working to adapt the GroFi platform to dry fiber placement for out-of-autoclave processing. Under other programs, the research center has developed the capability to infuse resin into dry-fiber preforms inside the autoclave. “The next step is to do it outside the autoclave,” says Kruse. GroFi can also be adapted to produce composite fuselages by using a rotating tool, he says.