Residual Stress Prediction by Adaptive Neuro-Fuzzy System in Milling Aluminum Alloy

Abstract:

As a sort of large-scaled structural components in modern aircraft, aluminum part has been widely used nowadays. Its residual stress measurement and prediction are necessary to reduce machining deformation and keep machining precision. By Adaptive Neuro-Fuzzy Inference System (ANFIS), residual stress prediction model is set up based on different cutting parameters. Due to data sample scarcity, input selection and regression are analyzed comparatively to reduce input data dimension. It shows that cutting speed and feed per tooth have major impacts on residual stress, but they do not have better prediction ability in ANFIS model. The combination of cutting speed and radial depth of cut can predict the residual stress better.

• Periodical: Key Engineering Materials (Volumes 392-394)
• Pages: 504-508
• DOI: https://doi.org/10.4028/www.scientific.net/KEM.392-394.504
• Online since: October 2008
• Authors: X.H. Zhang, Qing Long An, Ming Chen
• Keywords: Adaptive Neuro-Fuzzy Inference System (ANFIS), Aluminum (Al), Prediction, Residual Stress

References

[1] S.P. Lo: Journal of Materials Processing Technology, Vol. 142 (2003), pp.665-675.

[2] A. Iqbal, N. He, L. Li and N.U. Dar: Expert Systems with Applications, Vol. 32 (2007), pp.1020-1027.

[3] S. Kumanan, C.P. Jesuthanam and R.A. Kumar: The International Journal of Advanced Manufacturing Technology, Vol. 35 (2008) No. 7-8, pp.778-788.

[4] J.S.R. Jang: IEEE Transactions on Systems, Vol. 23 (1993) No. 3, pp.665-685.

[5] J.S.R. Jang, C.T. Sun and E. Mizutani: Neuro-Fuzzy and Soft Computing (Prentice Hall, U.S. 1997).