As we know, Automated Fibre Placement (AFP) is an additive material deposition process using a numerically controlled machine. The end effector or ‘head’ of the machine differentiates this technology from other processes. The head contains complex systems in order to feed out spooled material around a cylindrical roller onto the layup surface. The spooled material, analogous to a roll of tape, is laid in strips onto the form or layup surface via the software programming to build up the composite component.
Robotic arms, gantries, mandrels, and combinations of these are used to support the head of the machine. Preventing crashes between the head and layup surface, or the machine components, becomes a critical factor due to the complexity and support of the machine head, not to mention the high cost. This can be accomplished through machine simulation and collision avoidance. But where should organisations start?
Before digging deeper into the process, there are two definitions we need to address, as they can often be open to misinterpretation:
Making a head start
Equally as important are the methods taken to avoid collisions in the first place. A typical machine head has several attachments required during the layup process. Some of these include a scoop to feed the material underneath the roller, heaters that warm the surface to increase tack of the material, a laser probe for calibration, and the housing to contain the roller itself. Most of these components are located on the front or the sides of the head. As a result, the collision avoidance solver prioritises rotating the head backward and forward over high curvature areas and ramps in the layup surface. If rotating the head forward and backward is insufficient in creating a collision free program, the solver will then rotate the head side to side, or rotate the head in a compound angular fashion. Side to side rotation is used as a last resort due to potentially modifying the compaction of the material during the layup process.
Collision checking and avoidance can be done at different times during the work flow. There is the collision checking and avoidance that is part of the calculation algorithm. This occurs during calculation of the AFP layup, and is done automatically, without user intervention. The other type of collision checking is the simulation of the machine movements, often done by the programmer or operator, after the AFP layup is calculated. This subsequent collision checking calculation is often where users can visually simulate the movements at any point in the program.
Simulation of the machine after calculating the AFP layup provides an additional layer of assurance when programming an AFP machine. During the programming stage, collision avoidance is only calculated at specific points along a course. There may be areas of possible collisions found in areas in between those checked points. Simulating the program allows collision avoidance to be calculated for the entire path, at specified time intervals, giving the program a more refined analysis. Additionally, all the machine motion is interpolated between each time step which allows a complete simulation at any point in the program.
Even with all these preventative measures, collisions may still occur. A decision will need to be made whether to use an entirely different machine head, alter the current machine head and its components, calculate the AFP layup method using a different strategy or modify the collision buffer distance to a smaller value. All options should be reviewed in line with product and organisational requirements. For each team looking to implement simulation and collision avoidance detection, it’s essential to look at each of the considerations, to ensure that each element of the process creates the most streamlined yet accurate process possible. Through taking advantage of compute power and automation, alongside programmer or operator knowledge, significant benefits can be achieved.