An integrated application of neural network, fuzzy and expert systems for machining operation sequencing

Abstract

A part is described using features. A neuro fuzzy system then determines the machining sequence for each feature. Previous process plans were utilized to build, test, and validate the Neuro Fuzzy Network (NFN). Parts having similar manufacturing sequences are grouped into families, also using an NFN. A standard manufacturing sequence is obtained for each family comprising all the operations applicable to the features of the parts in the family. An expert system then adapts this standard sequence for the particular part being planned. The optimal operation sequence is inherited by the new part. The procedure is demonstrated by an example industrial part.