Design of a Pilot Setup to sort Damaged Returned Empty Beverage Crates in an Automatic Filling Line

Abstract

The inspection of returned beverage crates as well as bottles in industrial automatic filling lines is mainly performed by imaging systems. These systems are not able to detect invisible damages or embrittlement. A powerful novel system based on the principle of mechanical vibration analysis for the detection of small and concealed damages is presented. The selection of individual crates is performed automatically by a pre-trained artificial neural network (ANN). Numerical finite element simulations form a basic insight into the vibration behaviour of the crates and help to plan a pilot setup. This leads to a final recognition rate of more than 99 % over all checked crates in a prototype for industrial use.