Multi objective machining estimation model using orthogonal and neural network

Much hard work has been done to model the machining operations using the neural network (NN). However, the selection of suitable neural network model in machining optimization area especially in multi objective area is unsupervised and resulted in pointless trials. Thus, a combination of Taguchi ort...

Full description

Saved in:
Bibliographic Details
Main Authors: Yusoff, Y., Zain, A. M., Sharif, S., Sallehuddin, R.
Format: Article
Language:English
Published: Penerbit UTM Press 2016
Subjects:
Online Access:http://eprints.utm.my/id/eprint/70023/1/YuslizaYusoff2016_MultiObjectiveMachiningEstimationModel.pdf
http://eprints.utm.my/id/eprint/70023/
http://dx.doi.org/10.11113/jt.v78.10116
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Much hard work has been done to model the machining operations using the neural network (NN). However, the selection of suitable neural network model in machining optimization area especially in multi objective area is unsupervised and resulted in pointless trials. Thus, a combination of Taguchi orthogonal and NN modeling approach is tested on two types of electrical discharge machining (EDM) operations; Cobalt Bonded Tungsten Carbide (WC-Co) and Inconel 718 to observe the efficiency of proposed approach on different numbers of objectives. WC-Co EDM considered two objective functions and Inconel 718 EDM considered four objective functions. It is found that one hidden layer 4-8-2 layer recurrent neural network (LRNN) is the best estimation model for WC-Co machining and one hidden layer 5-14-4 cascade feed forward back propagation (CFBP) is the best estimation model for Inconel 718 EDM. The results are compared with trial-error approach and it is proven that the proposed modeling approach is able to improve the machining performances and works efficiently on two-objective problems.