An improved neural network model for residual stress prediction in turning

• Published in: 2004 IEEE International Conference on Industrial Technology, 2004. IEEE ICIT '04 Vol. 2; pp. 1012 - 1016 Vol. 2
• Main Authors: Amamou, R., Fredj, N.B., Rhouma, A.B., Fnaiech, F.
• Format: Conference Proceeding
• Language: English
• Published: piscataway NJ IEEE 2004
• Subjects: Applied sciences; Artificial neural networks; Computer science; control theory; systems; Exact sciences and technology; Fundamental areas of phenomenology (including applications); Intelligent networks; Machining; Neural networks; Physics; Predictive models; Residual stresses; Solid mechanics; Structural and continuum mechanics; Surface cracks; Surface resistance; Turning; US Department of Energy; Experimental design; High precision; Neural network; Empirical model; Train; Residual stress; Confidence interval
• ISBN: 0780386620; 9780780386624
• DOI: 10.1109/ICIT.2004.1490215

Results presented in this paper are related to the prediction of the longitudinal residual stress generated by the turning process. The main problems associated with the prediction capability of empirical models developed using the design of experiment (DOE) method are given. Their limited aptitude to calculate an accurate output value constitutes a serious limitation of the application of this method to residual stress prediction. In this study an approach suggesting the combination of DOE method and artificial neural network (ANN) is developed. Data of the DOE were used to train the ANNs and the inputs of the developed ANNs were selected among the factors and interaction between factors of the DOE depending on their significance at different confidence levels, expressed by /spl alpha/. Results have put in evidence the existence of a critical set of inputs for which the best learning results of the ANNs can be realied. A high prediction accuracy of these ANNs was tested through a good agreement with the empirical models developed by previous investigations.