Flexible Automation and Integrated Manufacturing 1998

ISBN Print: 978-1-56700-118-1

AN APPROACH TO TECHNOLOGICAL DATABASE - POSSIBILITIES FOR THE DETERMINATION OF CUTTING VALUES WITH NEURAL NETWORKS

DOI: 10.1615/FAIM1998.520
pages 577-586

Аннотация

Flexibility and intelligence are indispensable factors in automation and manufacturing. For such demands, intelligent information systems are required, especially considering the technological data type. The process setting parameters have to be determined in the best suitable way. A central point of technological data are the cutting values feed and speed, especially because of their influence on cost, time, quality und security of the manufacturing process. The rapid development of new advanced materials, tools etc. cause the lack of qualified, actual machining values.
A creative approach to make available technological data are artifical neural networks (ANN). Neural networks offer a possibility to solve these problems by training networks using known cutting values and the generalizing ability in a later utilization phase. Considering different data sources for cutting processes are possible. In order to improve the quality of a trained network components of existing mathematical process models can be integrated into the network. The authors demonstrate results of a project to solve such problems by means of neural network, mathematical model and technological database system.