Batch time optimization for an aerodynamic feeding system under changing ambient conditions

Abstract

In order to meet the demands for flexible feeding technology, a self-learning aerodynamic part feeding system has been developed. The actuated system uses a genetic algorithm to find the optimal parameter set for a high rate of correctly oriented parts. This orientation rate can change due to changes in the ambient conditions (e.g. ambient pressure, coefficient of friction). When the orientation rate in pre-defined interval of parts drops below a determined value, a correction algorithm is triggered. The objective of this work is to develop a mathematical model to define the optimal control interval and limit of the orientation rate for triggering the corrective algorithm depending on the total amount of parts still to be fed at any point in time. To evaluate the mathematical approach, a macroscopic simulation model of the aerodynamic feeding system was developed. It was shown, that the feeding time of a batch of 10,000 parts can be reduced by up to 7% and the number of activations of the corrective algorithm can be reduced by up to 50%. Finally, the mathematical model was implemented in the system control.