Diagnostic System Spurt as Basis for Connectionist and Hybrid Knowledge Processing in Manufacturing Control

Abstract This contribution describes a diagnostic system applied at a CNC-lathe for the prediction of the cutting tool condition and for manufacturing process control. The decision algorithm consists of two different knowledge based systems. An artificial neural net is used to transmit the condition information from the subsymbolic level where the information is represented as values of different features into the symbolic level with well known parameters for the machine condition. On the basis of this information an expert system predicts the remaining time until the next tool exchange and provides the control information for the manufacturing process.