Digital Twin Development and Validation for a Tapered Roller Bearing Multi-Stage Production Line

Abstract

The objective of this work is to develop and validate a Digital Twin (DT) for a multistage production line of tapered roller bearings. The manufacturing process consists of ring machining and component assembly, including intensive quality controls. This work proposes the integration of machine learning models associated with the manufacture of the double outer ring and the two inner rings in the DT. The models are trained with real data, so that the DT can predict the behavior of the production process under changing conditions of ongoing processes, machines or materials, and optimal operating conditions can be predicted. The DT has been developed and integrated with the aim of guiding production by proposing optimal machine configurations. To this end, different stations have been modeled and integrated into the DT as independent modules: grinding machines, inner and outer rings pairing module, and a module for calculating the optimal family of rings to be ground. After integrating the DT in the line, results show not only a raise in the line efficiency but also a decrease in the overall scrap ratio.