Implementation and testing of a genetic algorithm for a self-learning and automated parameterisation of an aerodynamic feeding system

An active aerodynamic feeding system developed at the IFA offers a large potential regarding output rate, reliability and neutrality towards part geometries. In this paper, the procedure of a genetic algorithm's into the feeding system's control is shown. The genetic algorithm automatically identifies optimal values for the feeding system's parameters which need to be adjusted when setting up for new workpieces. The general functioning of the automatic parameter identification is confirmed during tests on the convergence behaviour of the genetic algorithm. Thereby, a trade-off between the adjustment time of the feeding system and the solution quality is revealed.