A Neural Network Approach for Milling Process Control


Article Preview

One of key approaches to improve the productivity is to control with constant force in the milling process by adjusting the feed rate. In order to overcome the mismatch model occurred in adaptive control and inaccurate deducing regulation in fuzzy logic control, a three-layer BP neural network is designed for tracing reference force. First of all, control arithmetic is given, and a series of simulation work is achieved to determine the study factor. At last, aimed at two working conditions with abrupt and gradual change of cutting depth, the correctness and effectiveness of the neural network controller are proved by experiments.



Materials Science Forum (Volumes 471-472)

Edited by:

Xing Ai, Jianfeng Li and Chuanzhen Huang




Z. Yang et al., "A Neural Network Approach for Milling Process Control", Materials Science Forum, Vols. 471-472, pp. 107-111, 2004

Online since:

December 2004




[1] S.T. Chiang and D.T. Liu: Int. J. Mach. Tools and Manufacture Vol. 34, No. 4 (1994), p.637.

[2] M. Weck: Mach. Tool Task Force Conf. Vol. 4 (1980), p.7.

[3] K. Tomizu: Model Reference Adaptive Control of the Milling Process (Trans Tech Publications, USA 1987).

[4] L.K. Lauderbaugh and A.G. Ulsory: ASME, J. Eng. Ind. Vol. 111 (1989), p.13.

[5] Y. Altinats: Int. J. Mach Tools and Maunfacture Vol. 34 (1994), p.461.

[6] P.L. Hsu and M.Y. Hsieh: Int. J. Mach. Tools. and Manufacture Vol. 34 (1994), p.859.

[7] Y.S. Tarng and S.T. Cheng: Int. J. Mach. Tools and Manufacture Vol. 30 (1993), p.643.