CGTech’s Product Marketing Manager,
Bill Hasenjaeger, answers your questions:
I divide the issues into 3 categories: Part, Process, and Machine. A part’s overall shape, and the shapes and directions of the layers of CFRP material that create it, determine the complexity of the process. Dramatically curved part contours limit the direction the material can be successfully laid and also affect how wide a path of material can be laid at once. Additionally, the material has a limit to how closely it can be laid in a particular direction, commonly called the “steering limit.” The patches of material that make-up a layer, known as “plies,” each have their own simple or complex outline shape, called a “ply boundary.” The AFP or tape-laying process can only approximate that boundary.
For a complex-shaped part, the process will consist of many steps of laying material and inspecting each layer before it is covered-up with the next one. If the complexity of each layer is high, then the layer must be closely monitored for failures. The NC program and machine must have a mechanism for dealing with process failures during material lay-down. These recovery schemes can also be complicated.
All machines have limitations, and AFP machines are no exception. Thus the process requirements and machine limitations can sometimes conflict. For example, a curvy part shape with complex ply boundaries may result in the end of a course of material landing on a highly curved region of the part surface. It is very difficult to start a course of material on a highly curved surface, but much easier to end it on the curvy spot, so the NC programmer needs the option to lay the tape course starting at the less curved end. In this case the part complexity, the process requirements, and the machine limitations all come together, and the programming system needs to accommodate them all.
Expensive workpieces manufactured with complex processes must have reliable and robust verification and simulation tools. Today’s business conditions don’t tolerate any shop floor experimentation, nor very many do-overs. Whether laying a composite part with an AFP process, trimming the laid and cured part with a water-jet cutter, or drilling and fastening a finished composite panel to a titanium structure, failure is very expensive in terms of equipment, materials, man-power, and (probably the most costly) time lost in an always tight schedule.
If by “stand alone” you mean software developed to be used on any CNC machine rather than one specific brand of machine, we certainly think so. The main advantage to universal software is it’s broad usage, especially over time. Software developed for one specific brand of machine is only exposed to a small set of users of that machine. And software enhancements are driven by user requirements more than any other factor. Software exposed to a broad range of user experiences, followed by listening to those users and synthesizing their comments, requirements and requests allows us as software developers to infer new product features. We would not invent the new features, at least not as quickly or effectively, if left on our own or with a narrow band of user feedback. All users benefit from a broad implementation.
And off-line AFP programming software dedicated to a single machine brand is simply not fair or economical for manufacturers. This approach has a tendency to lock them into a single machine supplier, rather than allowing them to select the best machine for the job at hand. Universal CAD/CAM software that can be used on any CNC metal cutting machine is what allows a manufacturer the freedom to choose the best production metal-cutting solutions without having to retool his manufacturing engineers to new software each time. Imagine having to implement a new CAD/CAM system each time you buy a different brand machining or turning center. It’s not unusual to have more than 10 different brands of CNC machines in a single mid-sized workshop.
Most of our manufacturing customers are struggling to apply current AFP technology to complex high-curvature part shapes. It often seems like the machine the customer has is not designed to do it. However it also appears there are new AFP machine technologies being developed to specifically apply material over complex shapes. And innovative NC programming approaches are needed to successfully and reliably fiber- place complex parts while achieving the structural requirements of the laminate.
The biggest issue with AFP seems to be production rate vs. part complexity. Fairly complex shapes can be created with a single .125” wide strip (tow) of material. But the production time would be impractically long. So machine builders create 6, 8, 16, even 32 tow AFP systems. But then wide compaction rollers for these systems have difficulty maintaining compaction over curved surfaces. The real practical manufacturing limit is the minimum steering radius of the material, and the software’s ability to deal with it.
Process simulation is a very important component of AFP process development. And by process simulation I mean simulating from the exact NC program that will be sent to the CNC control of the AFP machine, and producing a simulated representation of the laminate as produced by the NC program. This is the only way to ensure the NC program produces the nominal results desired. Of course there are physical process variations and environmental conditions that cannot be simulated. But it is important to be able to verify that the NC program is not one of the process variables introduced on the shop floor.
That’s a tough one because I could not have predicted the last two years, let alone the next five. The AFP industry is changing very quickly and there are a lot of smart and creative people involved in it. I think the solution to economically fiber-placing small and complex parts is critical to the success of AFP technology. And the current complex process has to simplify and stabilize so that it is practical for 2nd and 3rd tier suppliers. Until that happens AFP will remain a boutique manufacturing method only available to the highest-end products and companies, not really very far away from the research lab. I think our software approach is a step in the right direction (away from the lab), helping to de-mystify the AFP programming process and make it more approachable for smaller companies.