Please use this identifier to cite or link to this item:
Title: The optimisation of surface roughness using extreme learning machine and particle swarm optimization
Authors: Janahiraman, T.V. 
Ahmad, N. 
Abdullah, T.A.R.T. 
Issue Date: 2019
Abstract: In recent days, metal cutting has become a highly demanding sector due to growing applications. The turning process is one of the metal cutting processes which produces circular shapes from a cylindrical bar. Currently, turning operation is conducted using computer numerical control machine (CNC). The machinist is required to assign the optimal cutting parameters in CNC turning which have direct influence on the performance of each cutting process. Therefore, it is crucial to achieve the optimal parameters before the process is started. In usual cases, these parameters will be assigned according to machinist's past experience or with reference to the manual handbook provided by the tool supplier. However, this approach can be considered time consuming and do not guarantee that it can produce the desired cutting performance. In light of this issue, a new optimisation technique has been proposed to figure out the optimal cutting parameters. Box Behnken's design is used as the experimental design, while the improved Extreme Learning Machine which is based on Particle Swarm Optimisation is proposed as the prediction model. A powerful and effective, Particle Swarm Optimisation will act as the optimiser of the prediction model. The turning parameters: cutting speed, feed rate and depth of cut, are considered as the input variables to the model. The optimisation results prove that the system is able to predict well and generate optimal cutting parameters to minimize the surface roughness of the machined workpiece. © 2019 Int. J. Mech. Eng. Rob. Res.
DOI: 10.18178/ijmerr.8.1.69-73
Appears in Collections:UNITEN Scholarly Publication

Files in This Item:
File SizeFormat 
The Optimisation of Surface Roughness using.pdf1.36 MBAdobe PDFView/Open
Show full item record

Google ScholarTM



Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.