Neural Network Error Compensation of Machine Tools

Period of Performance: 05/12/1994 - 11/05/1994


Phase 1 SBIR

Recipient Firm

Tetra Precision, Inc.
2335 Laurel Lane
Palm Beach Garden, FL 33410
Principal Investigator

Research Topics


The proposed research involves the development of a system capable of improving machining precision using Artificial Neural Network (ANN) technology. In the last few years, there has been a considerable effort to apply artificial neural networks to different fields of manufacturing due to its favorable features like parallelism, robustness and compactness. Research issuesto be explored include: 1) The ability of ANN's to learn and predict geometric and thermal errors from training data of the tool point error vectors, cutting tool location and strategically located temperature probes; 2) identification of appropriate inputs and outputs for ANN prediction of tool wear, elastic deflections and contouring errors; and 3) interface of trained ANN to provide inputs to a real-time error compensation system.