Title: RBF neural network model for machining quality prediction in CNC turning process

Authors: J. Edwin Raja Dhas; R.S. Stalin; J. Rajeesh

Addresses: Department of Automobile Engineering, Noorul Islam Centre for Higher Education, Kumaracoil – 629180, Tamilnadu, India ' Tamilnadu State Transport Corporation, Nagercoil – 629807, Tamilnadu, India ' Department of Biomedical Engineering, Noorul Islam Centre for Higher Education, Kumaracoil – 629180, Tamilnadu, India

Abstract: Intelligent manufacturing systems rely on effective and efficient decision making tools. Decision making is increasingly difficult due to the rapid changes in design, parameters and environments due to variety of applications. Mathematical models which are derived based on the assumptions are limited to model the functioning of the real manufacturing system. There is a need to develop generalised models which can dynamically predict a wide variety of process parameters such others to assist the intelligent manufacturing system. Artificial intelligent tools like neural networks are being attempted in decision-making process. This paper addresses the development of neural radial basis function neural network (RBFNN) model for machining quality prediction. Data acquired to develop the proposed RBFNN model is obtained from experimentation on computer numerical control (CNC) machine centre using response surface methodology design matrix. Results obtained are utilised to train the proposed model with an objective of obtaining good surface finish using single tool operations. Results from the RBFNN model are compared with the neural network model trained with back propagation (BPNN) algorithm in terms of computational speed and accuracy. The performance of RBFNN is better than BPNN model and is reasonably more accurate. Confirmatory experiments are done to validate this approach and presented.

Keywords: surface roughness; cutting parameters; machining quality prediction; radial basis function; RBF neural networks; RBFNN; CNC turning; surface quality; intelligent manufacturing; response surface methodology; RSM.

DOI: 10.1504/IJMIC.2013.056190

International Journal of Modelling, Identification and Control, 2013 Vol.20 No.2, pp.174 - 180

Published online: 27 Sep 2014 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article