A neural network model for selecting machining parameters in fixture design

This paper describes the training and implementation of an artificial neural network model (a multilayer feedforward backpropagated neural network) for the real-time, online estimation of key machining parameters. The neural network model is customized, configured, and trained off-line and then integrated with the automated fixture design (AFD) system for on-line prediction of the machining parameters (according to given workpiece representations and required processing information) as needed in the fixture design synthesis process. The paper reports the implementation of a prototype system and discusses several issues regarding the integration aspects in a case study.