A neural network-based approach for part family classification for a reconfigurable manufacturing system

The design of RMS initiates with the classification of parts into families, after which reconfiguration of the system is carried out to cater new part families. It is important that parts must be grouped into logical families based on similarities either in manufacturing or design attributes. Generally, production system maintains a large database of existing part families, and once any new part comes in, the efforts must be focused on deciding upon an appropriate existing part family in which the new part may be grouped with. In literature, most of the approaches are based on part family formation from beginning with no consideration of how the existing part family database can be utilised to decide upon a suitable existing part family for a new part. This paper proposed a neural network classification-based approach for such classification. The developed methodology is explained with the help of a numerical illustration.